Datasets:

TaylorAI

/

pubmed_author_manuscripts

like 3

Follow

Taylor 16

PMC005xxxxxx/PMC5125930.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 2985088R 4967 J Mol Biol J. Mol. Biol. Journal of molecular biology 0022-2836 1089-8638 27019298 5125930 10.1016/j.jmb.2016.03.018 NIHMS829249 Article Heme Synthesis and Acquisition in Bacterial Pathogens Choby Jacob E. 1 Skaar Eric P. 12 1 Department of Pathology, Microbiology, & Immunology, Vanderbilt University School of Medicine, Nashville, TN, USA 2 Tennessee Valley Healthcare System, U.S. Department of Veterans Affairs, Nashville, TN, USA Correspondence to Eric P. Skaar: Department of Pathology, Microbiology, & Immunology, Vanderbilt University School of Medicine, 1161 21st Ave S., Nashville, TN 37232, USA. eric.skaar@vanderbilt.edu 15 11 2016 24 3 2016 28 8 2016 28 11 2016 428 17 34083428 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. Bacterial pathogens require the iron-containing cofactor heme to cause disease. Heme is essential to the function of hemoproteins, which are involved in energy generation by the electron transport chain, detoxification of host immune effectors, and other processes. During infection, bacterial pathogens must synthesize heme or acquire heme from the host; however, host heme is sequestered in high-affinity hemoproteins. Pathogens have evolved elaborate strategies to acquire heme from host sources, particularly hemoglobin, and both heme acquisition and synthesis are important for pathogenesis. Paradoxically, excess heme is toxic to bacteria and pathogens must rely on heme detoxification strategies. Heme is a key nutrient in the struggle for survival between host and pathogen, and its study has offered significant insight into the molecular mechanisms of bacterial pathogenesis. iron uptake hemoglobin Isd system heme toxicity NEAT domain Introduction Heme and iron are essential for life The tetrapyrrole cofactor heme is important for the cellular processes of most organisms and essential to many lifeforms across domains of life. Heme, a porphyrin ring complexed with iron, serves as a redox active moiety required for the function of many cellular proteins. Heme functions as an electron shuttle in enzymes of the electron transport chain and is required for cellular respiration. Additionally, cells rely on heme for the function of many widely conserved enzymes including catalase, nitric oxide synthase, hemoglobin (Hb), and others. Heme is also an important molecule involved in diverse cellular processes including signaling, gas sensing, microRNA processing, and cellular differentiation [1–4]. Thus, nearly all organisms must satisfy the requirement for heme through either synthesis or acquisition. Heme coordinates an iron atom at its center which is vital for heme's electron transfer abilities and redox activity. Like heme, iron is nearly universally required for life, and only a few exceptions have been identified [5,6]. As an inorganic cofactor, iron can act alone or in iron–sulfur clusters as a prosthetic moiety for members of the oxidoreductase, nitrogenase, hydrogenase, dehydrogenase, and hydratase enzyme families [7–12]. Therefore, organisms have evolved elaborate strategies to acquire, store, and regulate intracellular iron for heme-dependent and other iron-dependent enzymes. Nutritional immunity limits host iron availability Nutritional immunity, a concept articulated primarily by Eugene Weinberg in the 1970s, describes the processes by which humans and other organisms sequester iron to limit acquisition by bacterial pathogens [13,14]. Nutritional immunity has since been expanded to include the host processes that manipulate local levels of manganese, zinc, and other transition metals in order to metal starve or intoxicate the invading pathogens (reviewed previously in Refs. [15,16–18]). The limited access of pathogens to metals serves as an antimicrobial strategy and limits bacterial replication. For instance, free iron rarely exists in the mammalian host. The solubility of ferric iron in aerobic solution is exceedingly low, and high-affinity iron-binding proteins, including transferrin, lactoferrin, albumin, and ferritin, sequester iron extracellularly and intracellularly. Iron-binding proteins function to transport iron, protect host cells from iron-mediated oxidative damage, and keep iron from pathogens. However, bacterial pathogens have developed exquisite tactics to overcome iron limitation and elaborate high-affinity iron receptors and chelators. In this regard, an evolutionary arms race has developed at the host–pathogen interface involving host iron-binding proteins and the mechanisms bacteria encode to steal iron. Heme is an important host iron source Heme makes up the greatest reservoir of iron in the host and serves as an iron source for many bacterial pathogens. Humans and other metazoa synthesize heme through a variety of steps in the mitochondria and cytosol. This pathway, called the Shemin or four-carbon pathway, begins with the condensation of glycine and succinyl-CoA to form the committed precursor δ-aminolevulinic acid (ALA) [19–21]. A series of enzymes produces protoporphyrin IX from ALA and iron is inserted, forming protoheme IX. For the sake of simplicity in this review, heme will refer to ferrous and ferric iron forms of protoheme IX. Heme is then bound by hemoproteins to serve a variety of intracellular and extracellular tasks. Catalase, peroxidase, and myeloperoxidase rely on heme to catalyze the hydrolysis of peroxide molecules. Energy generation by the electron transport chain relies on heme-dependent c- and b-type cytochromes of the ubiquinol–ferricytochrome c oxidoreductase (Complex III) family [22,23]. Hemoproteins involved in tissue oxygen homeostasis include myoglobin and neuroglobin. Perhaps the most well-known hemoprotein is the oxygen transporter Hb. Its abundance and location in erythrocytes make Hb a rich heme source for pathogens. Hb contains about two-thirds of the body's iron, and a single erythrocyte contains more than 280 million molecules of Hb [15,24]. Bacterial pathogens have evolved high-affinity Hb-binding proteins for the acquisition of heme, and these proteins will be described below. Owing in part to the reactive nature of heme-iron, free heme and Hb are toxic to the human host and bacterial pathogens alike [25,26]. To prevent excess heme toxicity, eukaryotic heme synthesis is highly regulated and heme homeostasis and sequestration are well orchestrated. When Hb is released from erythrocytes or otherwise exists extracellularly, it is rapidly bound by haptoglobin (Hp) [27]. The abundance of cell-free Hb is thought to be very low in healthy adults, but a variety of genetic disorders, infections, and other disease states can increase the concentration of free Hb [28]. Free Hb and its modified forms, in the presence of reactive oxygen species, exhibit cytotoxic effects toward endothelial cells [29]. However, the relevance of these in vivo studies is unclear, and a comprehensive understanding of concentrations to achieve Hb toxicity in healthy humans has not been achieved [25]. On the other hand, in the absence of infection-free heme that has been liberated from its hemoprotein likely only exists transiently in serum or in cells. In serum, heme is immediately bound by the highly abundant albumin (kd ≈ 10 nM) then transferred to hemopexin (kd < 1 pM) [30]. The heme is delivered to cells expressing the hemopexin receptor; these cells then degrade the heme using heme oxygenases [30]. The rapid sequestration and degradation of free heme in the blood is vital to the survival of erythrocytes, as heme in the presence of reactive oxygen species exhibits cytotoxicity and lipid peroxidation at micromolar concentrations [31,32]. During infection of host heme- and Hb-replete niches, bacterial pathogens experience heme toxicity and encode systems to protect from heme toxicity as well [33]. Therefore, heme is at the center of a complex interplay between host and pathogen for survival. Bacterial Heme Synthesis Divergent heme synthesis pathways in Gramnegative and Gram-positive organisms While both humans and bacteria share the early heme precursor ALA, most bacteria (Alphaproteo-bacteria are the exception), archaea, and plants synthesize ALA from charged glutamyl-tRNAGlu via the “C5 pathway” (Fig. 1) [21,34,35]. The gluta-myl-tRNA reductase HemA produces the highly reactive intermediate glutamate-1-semialdehyde, which HemL, a glutamate-1-semialdehyde amino-mutase, converts to ALA [36,37]. The three steps, from ALA to uroporphyrinogen, are well conserved and thought to be the evolutionary core of heme biosynthesis. ALA dehydratase (also called porphobilinogen synthase; annotated as HemB) is responsible for the condensation of two ALA to porphobilinogen (PBG) [38]. The linear tetrapyrrole hydroxymethylbilane (HMB) is produced by a head-to-tail condensation and deamination of four PBG molecules, catalyzed by HMB synthase (alternatively called PBG deaminase, annotated as HemC) [39,40]. Under physiological conditions, HMB will spontaneously cyclize to form the uroporphyrinogen I isomer, a biosynthetic deadend. Therefore, most bacteria utilize uroporphyrinogen III synthase (HemD) to catalyze the cyclization of HMB through a spiro-intermediate to form uroporphyrinogen III [41]. Uroporphyrinogen III can be utilized for the synthesis of several tetrapyrrole-based cofactors. Uroporphyrinogen III decarboxylase (HemE) decarboxylates the four acetate side chains to methyl groups, producing coproporphyrinogen III, the next step in heme synthesis [42]. Additionally, uroporphyrinogen III can be shunted from heme synthesis and converted to precorrin-2 to synthesize vitamin B12, coenzyme F430, and siroheme [43]. The Ahb enzymes of some archaea and sulfur-reducing bacteria can convert siroheme (produced from uroporphyrinogen) to heme [44,45]. The contribution of the Ahb alternative heme pathway has not been demonstrated in bacterial pathogens. In Gram-negative organisms, as well as eukaryotes, coproporphyrinogen III is converted to proto-porphyrinogen IX by coproporphyrinogen III oxidase. This step is the first of the terminal three steps in the classical heme synthesis pathway (in blue in Fig. 1) and is catalyzed by oxygen dependent HemF or by oxygen independent HemN [46,47]. Protoporphyrinogen IX is subsequently oxidized to form proto-porphyrin IX, by a six-electron oxidation catalyzed by one of three protoporphyrinogen oxidase enzymes. HemG, in Gammaproteobacteria and some Alphaproteobactera and Deltaproteobacteria, uses the respiratory chain as its electron acceptor and is not dependent on oxygen [48]. HemJ is poorly characterized but represents the most common protoporphyrinogen oxidase among Alphaproteobactera and Deltaproteobacteria [49]. The third protoporphyrinogen oxidase is HemY, an FAD- and oxygen-dependent protoporphyrinogen oxidase found in some Proteobacteria as well as eukaryotes [50]. The final step of the classical pathway is the insertion of ferrous iron by protoporphyrin ferrochelatase (HemH) to form protoheme IX, called heme [51]. From ALA to heme, the steps of the classical synthesis pathway are shared by eukaryotes and Gram-negative bacteria. The terminal steps of the classical pathway were considered universally conserved for all heme synthesizing organisms. However, just in the last few years, the terminal steps of heme synthesis in the Gram-positive phyla Firmicutes and Actinobacteria have been described with genomic and biochemical analysis and termed the non-canonical or transitional pathway [34,52]. Very few HemF or HemN coproporphyrinogen oxidases can be identified in Gram-positive genomes; instead it has been realized that the annotated HemY in these organisms functions as a coproporphyrinogen oxidase to form coproporphyrin III [34,53]. The Gram-positive HemH, a coproporphyrin ferrochelatase, inserts ferrous iron to form coproheme [52]. Finally, coproheme is decarboxylated by HemQ, an enzyme unique to members of the Firmicutes and Actinobacteria to form protoheme IX [54–57]. It is now clear that Grampositive organisms utilize a unique series of terminal steps to synthesize heme (in green in Fig. 1). Regulation of heme synthesis Despite the vital role of heme to bacterial physiology, the regulation of heme biosynthesis has not been well studied outside of a few model organisms. In bacteria, regulation has been recognized to occur largely at two steps, abundance of the initial enzyme HemA and transcription of the coproporphyrinogen oxidase enzymes. Regulation of HemA is typically heme-dependent, indicating that bacteria reduce synthesis of heme and all intermediates in heme-replete conditions. This process has been extensively studied in Escherichia coli and Salmonella enterica serovar. Typhimurium. The addition of heme to cell extracts of E. coli reduces total HemA activity, without inhibiting the activity of the purified enzyme [58,59]. This was explained by the observation that excess heme results in the proteolytic degradation of HemA in Salmonella, suggesting that HemA might bind excess heme [60]. The Clp and Lon proteases are responsible for this reduction in HemA levels [61]. Furthermore, mutations in HemA have been described that render HemA resistant to heme- and protease-mediated degradation, indicating that HemA binds excess heme, and holo-HemA but not apo-HemA is a substrate for proteolytic degradation [62,63]. In this manner, cellular levels of heme can regulate the first step of heme synthesis and limit the unnecessary synthesis of heme intermediates as well as the consumption of iron. Recent metabolic engineering efforts to enhance ALA production in E. coli suggest that protoporphyrin IX post-translationally inhibits HemB, an additional example of feedback inhibition [64]. It is likely that for many organisms, heme and terminal heme intermediates can have post-translational regulatory effects on heme synthesis enzymes. Like Salmonella and E. coli, the Gram-positive bacterium Bacillus subtilis regulates levels of HemA. While a mechanistic explanation has not been described, the membrane protein HemX post-transcriptionally regulates HemA abundance in B. subtilis [38,65]. Homologs of B. subtilis HemX exist in multiple Gram-positive pathogens; however, the function of HemX and HemA regulation has yet to be detailed. In addition to the regulation of HemA enzyme levels, the transcription of hemA is also a point of control for heme biosynthesis. Two promoters exist upstream of hemA in the Gram-negative pathogen Pseudomonas aeruginosa, and these promoters contain binding sites for the regulators Anr (oxygen sensing), Dnr (redox regulator), IHF (integration host factor), and NarL (nitrate regulator) [66,67]. Therefore, hemA expression is induced in the presence of oxygen or when oxygen is lacking but an alternative electron acceptor such as nitrate is present for utilization of heme-dependent respiration. In B. subtilis, hemEHY is induced anaerobically and hemAXCBL is induced by peroxide through de-repression of PerR [38,68]. As in B. subtilis, PerR has been implicated as a regulator of the hemEHY operon in Staphylococcus aureus; yet recent work has demonstrated that major differences exist between B. subtilis and S. aureus PerR orthologs, and therefore, it is difficult to conclude that PerR plays a role in S. aureus heme synthesis [69,70]. Corynebacterium diphtheriae, a member of the Actinobacteria phylum, encodes two heme-responsive two-component systems (TCS). The response regulator HrrA directly binds the promoters of hemA, hemE, and hemH to repress their transcription in heme-replete conditions [71]. Similarly, ChrA can repress transcription of hemA in heme replete conditions [72,73]. These data suggest that in C. diphtheriae, heme utilization is preferred over synthesis when exogenous heme is available. Together, these examples point to the transcriptional and post-translational control of HemA as a central step in heme synthesis regulation. The expression of coproporphyrinogen oxidase genes is the second major point of heme synthesis regulation. In several species, hemF and hemN are regulated by different oxygen- or anaerobic-responsive regulators to ensure proper expression of oxygen-dependent or oxygen-independent coproporphyrinogen oxidases. OxyR, a global regulator in E. coli, is responsible for the induction of oxygen-dependent hemF expression in hydrogen peroxide stress. It has been suggested that the [Fe-S] cluster in oxygen-independent HemN is vulnerable to peroxide damage, so HemF is produced to take the place of HemN [74]. In B. subtilis, the transcription of coproporphyrinogen III oxidases hemN and hemZ (a second coproporphyrinogen oxidase, not to be confused with oxygen-dependent HemY) is induced anaerobically by the regulatory cascade of ResDE, Fnr, and YwiD to replace the oxygen-dependent HemY [75–78]. Similarly, Pseudomonas hemF and hemN are expressed anaerobically under the control of Anr and Dnr, while Anr induces the expression of only hemN aerobically [79]. It has been suggested, but not validated, that the expression of oxygen-dependent hemF in oxygen limited conditions by Anr and Dnr serves to consume residual oxygen during the transition to anaerobiosis, which would protect other anaerobically induced oxygen-sensitive proteins [79]. Thus, oxygen is a key regulatory of expression of coproporphyrinogen oxidase genes. Contribution of heme synthesis to pathogenesis With a few notable exceptions including Bartonella hensaela, Enterococcus faecalis, Haemophilus influenzae, and Streptococcus spp., most human pathogens encode complete heme biosynthetic pathways [80–83]. However, the contribution of heme synthesis to the pathogenesis of bacterial pathogens is largely understudied. For S. aureus, whose reliance on heme acquisition during infection has been well established, it is now clear that heme biosynthesis is vital to cause disease in murine models of infection [84–86]. Inactivation of hemA, which renders S. aureus heme deficient, causes the small-colony variant (SCV) phenotype [87]. During systemic infection, this mutant is highly defective at colonizing the murine heart and liver relative to wildtype S. aureus [87]. A mutant lacking hemB, also a heme-deficient SCV, demonstrates reduced colonization and bone destruction in a murine model of osteomyelitis [88,89]. These data demonstrate that for S. aureus, heme acquisition is insufficient to support organ colonization and therefore heme biosynthesis is critical to pathogenesis. Importantly, the SCV phenotype is encountered clinically. Despite their reduced virulence, SCVs are generally more resistant to antibiotics and oxidative stress, more equipped to evade the immune system by living intracellularly, and are likely the etiological agent of persistent staphylococcal infections [89–92] (reviewed in Ref. [93]). Less evidence for the role of heme synthesis during infection is available for other pathogens. For the intracellular pathogen Brucella abortus, hemH is required for virulence in a murine model of brucellosis [94]. Therefore, like S. aureus, host heme utilization is insufficient and synthesis is required for full virulence. In addition to B. abortus and S. aureus, the advent of whole genome in vivo analysis of mutants using techniques such as transposon-sequencing and signature tagged mutagenesis has highlighted the role of heme synthesis. In these infections, genes with marked mutations that are recovered at a lower frequency from the infected tissue relative to growth in vitro are considered important to infection. These types of experiments have demonstrated a role for different heme synthesis genes during infection. Transposon mutantsdisrupted in hemY were found to be defective for P. aeruginosa colonization of the murine gastrointestinal tract [95]. hemN was found to be important for Yersinia pestis infection of deep tissue [96]. Transposon mutants lacking hemE in Acinetobacter baumannii were less effective at colonizing the murine lung [97]. Finally, hemG was found to be important for Listeria mono-cytogenes oral infection [98]. Based on these trans-poson library infections, heme synthesis is vital to the fitness of a variety of pathogens. Current challenges and opportunities The divergence between the terminal steps of Gram-positive heme synthesis and the classical pathway utilized by Gram-negative organisms as well as humans presents the opportunity for targeted small molecule interventions to inhibit or activate Gram-positive heme synthesis. The terminal Gram-positive enzymes HemQ, which exists only in Actinobacteria and Firmicutes, as well as HemY and HemH, which recognize different substrates than the eukaryotic host enzymes, present three potential targets. Small molecules have been described that modulate heme synthesis in vivo, while in vitro inhibitors of S. aureus HemY have recently been reported, suggesting that Gram-positive heme synthesis is an attractive drug target [52,99,100]. Outside of a few model pathogens, very little is understood regarding the regulation of heme synthesis, particularly during pathogenesis. Regulation is a central question in understanding the role of heme synthesis in infection. Considering that in some niches host heme is available and can reach toxic levels, pathogens with the capacity to both steal and synthesize heme must regulate both pathways. For S. aureus, in which heme synthesis and acquisition are vital during infection, the regulation of heme synthesis is unknown. This is despite the observation over half a century ago that the rate of staphylococcal heme synthesis is modulated by exogenous heme [101]. For other pathogens, the contribution of heme synthesis to disease is still unclear, but whole-genome in vivo fitness experiments like transposon-sequencing suggest many bacterial pathogens rely on heme biosynthesis to cause disease, and this field of research provides ample opportunity for further exploration. Gram-Positive Heme Acquisition Strategies Bacterial pathogens utilize a variety of heme acquisition strategies during infection, ranging from surface receptors to secreted proteins that bind either heme or hemoproteins. Heme acquired from the host is used fully intact or degraded to liberate heme-iron and both processes are important during bacterial pathogenesis. Gram-positive pathogens, including S. aureus, Bacillus anthracis, and C. diphtheriae rely on heme acquisition during infection. The heme uptake pathways of these three pathogens will be presented as models for theGram-positive processes, along with the regulation of the pathway and evidence for the role of heme uptake during pathogenesis. The S. aureus Isd paradigm The Iron-regulated surface determinant system (Isd), first described in S. aureus, is the paradigm for Gram-positive heme acquisition [102]. During infection, S. aureus utilizes the leukocidins HlgAB and LukED to lyse erythrocytes and liberate Hb into the bloodstream [103]. This results in accessible free heme, heme bound by hemopexin (Hx), free Hb (Hb), and Hb bound by Hp to form the Hp–Hbcomplex. The Isd system enables utilization of free heme, or heme bound to Hb and Hp–Hb complexes. Isd proteins bind heme and Hb at the cell wall surface with conserved near transporter (NEAT) domains. The NEAT domains are 120–125 aa domains that constitute a conserved eight-stranded β-sandwich fold [104,105]. Heme is bound in a hydrophobic pocket with critical coordination by tyrosine residues in a YXXXY motif. These NEAT-containing surface proteins (IsdB, IsdH, IsdA in S. aureus) shuttle heme to NEAT-containing IsdC. IsdC transfers heme to the membrane-associated transporter IsdDEF for transit across the membrane. To access host heme and hemoproteins, IsdB, IsdH, and IsdA are covalently attached to the peptidoglycan by the standard Sortase A cysteine transpeptidase [106]. IsdB contains two NEAT domains, NEAT1 (N1) binds Hb and Hb–Hp, but not Hp and N2 binds heme; as such IsdB is believed to be the primary Hb-binding protein [85,107,108]. IsdH contains three NEAT domains, N1 and N2 bind both Hb and Hp, andN3bindsheme [109,110]. IsdA, which is partially surface exposed, contains a single heme-binding NEAT domain [102]. The current model (Fig. 2), supported by strong structural evidence, suggests that IsdB-N1 binds Hb, and IsdB-N2 extracts heme [111]. Similarly, IsdHN1 and N2 bind Hb and Hp, and IsdH-N3 extracts the heme. The heme is then transferred either directly to IsdC or shuttled via IsdA to IsdC. S. aureus encodes an iron-regulated Sortase B (SrtB) for which IsdC is the only substrate, and SrtB attaches IsdC to peptidoglycan in such a way that IsdC is not surface exposed but rather buried in the cell wall, which is 15–30 nm thick [112,113]. This organization allows heme transferred from surface Isd proteins to pass through the cell wall to the membrane by IsdC's single heme-binding NEAT domain. IsdC alone transfers heme to the IsdE of the IsdDEF transporter [114]. At the membrane, IsdDEF transit heme across the membrane and into the cytosol. Upon import, heme is incorporated into staphylococcal proteins or degraded. Exogenous heme accumulates in the membrane and is also capable of complementing the growth of heme-deficient mutants [84]. Alternatively, the heme oxygenases IsdG and IsdI degrade heme to release iron [115] (reviewed in Ref. [116]). IsdG and IsdI are structurally similar and are the first described members of the Isd heme oxygenase family, which catabolizes heme to staphylobilin instead of biliverdin [117–119]. IsdG and IsdI are required for growth using heme as a sole iron source and are expressed during infection [115,120]. The widely conserved ferric uptake regulator (Fur) is the principle regulator of the expression of heme acquisition systems in S. aureus. In iron-deplete conditions, Fur no longer represses its regulon, allowing the transcription of the isdB, isdA, isdC-DEFsrtBisdG, and isdI loci [102]. During infection of iron-deplete niches, the heme acquisition system and associated iron-liberating heme oxygenases are expressed. Further regulation of the heme oxygenases exists; IsdG abundance increases in the presence of heme and IsdG half-life is increased when heme-bound [120]. Also, the Clp proteases have a role in Hb acquisition by modulating IsdB levels [121]. Additional heme-dependent regulation likely exists but has not been described. Isd-mediated heme acquisition is vital to the virulence of S. aureus. Heme is the preferred iron-source during systemic infection, in part because a heme-responsive transcriptional regulator activates iron siderophore synthesis only when heme-iron is unavailable [84,122]. The role of the Isd system has been extensively demonstrated in murine infection models. Mutants lacking components of the Isd system are highly defective inpathogenesis, highlighting the importance of heme acquisition to staphylo-coccal disease [84–86,108,120,123,124]. Isd-dependent heme uptake by B. anthracis B. anthracis encodes a heme uptake system that shares the core of the S. aureus Isd, but with additional unique proteins. B. anthracis encodes two secreted hemophores termed IsdX1 and IsdX2 [125]. These are the first described Gram-positive hemophores and bind heme, Hb, and methemoglobin [125– 129]. IsdX1 contains one NEAT domain, while IsdX2 contains five NEAT domains; both are secreted past the cell wall as they lack sortase signals or membrane spanning domains [125]. B. anthracis also encodes other NEAT contain proteins; Hal contains a single NEAT domain and leucine-rich repeats, which extract heme from Hb [130]. Unlike IsdX1/2, Hal is sortase anchored to the cell wall [131]. A second, recently described NEAT protein is BslK, which is non-covalently attached to the cell wall and transfers heme to IsdC [132]. The current proposed model (Fig. 2) is that IsdX1 is secreted, binds heme, and transfers heme to wall-anchored IsdC. IsdX2 can bind free heme, accept heme from IsdX1, and transfer heme to IsdC. The multiple NEAT domains ofIsdX2 have been proposed to be important for these multiple functions, and it has been suggested that IsdX2 can serve as a heme storage protein. IsdDEF transports heme across the membrane for utilization by IsdG, an orthologue of the S. aureus heme oxygenase [133]. The diversity of heme and Hb-binding proteins relative to S. aureus may be the result of the greater variety of environmental niches that germinant and sporulent B. anthracis inhabits. The role of B. anthracis heme acquisition during infection is not clear. A guinea pig infection model demonstrated that ΔisdCX1X2 was as virulent as wild type, yet these proteins are expressed during infection [134]. Also, a mutant of B. anthracis lacking Hal demonstrated reduced virulence in a model of inhalational anthrax [135]. It is likely that the IsdX1/X2 hemophores, BslK, and Hal are partially redundant, and a mutant lacking all four proteins would be highly defective in causing anthrax. In addition to S. aureus and B. anthracis, many other pathogens have evolved NEAT-containing heme acquisition systems, including Staphylococcus lugdunensis, L. monocytogenes, and Streptococcus pyogenes [136–143]. The conservation of NEAT-mediated heme uptake highlights the contribution of host heme to bacterial infection. C. diphtheriae heme uptake C. diphtheriae utilizes non-NEAT-mediated heme uptake systems for heme-iron acquisition, termed HmuTUV, HtaABC, and ChtABC/CirA. The Hmu (hemin-uptake) system was the first heme acquisition system described in Gram-positive organisms. The associated heme oxygenase, HmuO, was discovered and described first, and then HmuTUV was discovered for the ability of a plasmid encoding hmuTUV to complement a Corynebacterium ulcerans strain that cannot grow on Hb as a sole iron source [144,145]. Sequence analysis suggests that HmuTUV acts asan ABC transporter that shuttles heme across the cell membrane [146]. It was later discovered that an additional gene is encoded within the hmuTUV operon, termed htaA (heme-transport associated) [147]. Adjacent to this locus are the genes htaB and htaC. Unlike the sortase anchoring of other Gram-positive uptake systems, HtaA and HtaB contain N-terminal secretion signals as well as C-terminal intermembrane domains. This results in surface exposure of HtaA and HtaB, which both bind heme. Interestingly, a portion of HtaA is secreted and not anchored to the cell envelope. HtaA isolated from cell culture is unable to complement the growth of an htaA mutant, suggesting that surface bound HtaA may serve as a heme receptor and secreted HtaA may serve as a hemophore [147,148]. However, heme transfer between HtaA molecules and further description of the function of HtaA on the surface have not been reported. In addition to heme, HtaA can acquire heme from Hb and transfer heme to HtaB, suggesting a heme shuttle from HtaA toHtaB toHmuT; HmuT isa surface-anchored lipoprotein, which then transfers heme to the cognate ABC transporter HmuUV [148]. While the Isd NEAT domains rely on tyrosine alone as the axial ligand for heme binding, HmuT relies on an N-terminal histidine and a C-terminal tyrosine to coordinate heme [149]. Inactivation of the Hmu/Hta systems does not completely eliminate growth with heme as a sole iron source, suggesting the involvement of an additional heme uptake system [147]. This led to the characterization of the ChtAB and CirAChrC operons, which are regulated by iron levels via DtxR. DtxR is the Diphtheria Toxin regulator which activates the expression of Diphtheria Toxin as well as HmuTUV and HtaABC [150,151]. ChtAB and ChtC appear to be the result of gene duplication of HtaAB, as all three groups of proteins have sequence similarity, N-terminal secretion signals, and C-terminal transmembrane domains, and contain the same heme-binding domain [152]. Like HtaAB, ChtAB and CirAChtC are surface exposed and ChtAB and ChtC bind heme and Hb. It appears that these heme-binding proteins serve redundant functions, and as such, a mutant lacking both HtaB and ChtB is deficient at utilizing Hb as an iron source [152]. Recently, it has been shown that ChtA and ChtC are both capable of binding Hp–Hb for heme extraction, and acquisition of heme from Hp–Hb requires HtaA [153]. The current model (Fig. 2) for Hp–Hb heme acquisition involves binding of Hp–Hb by a combination of HtaA and ChtA or ChtC, heme extraction either actively or passively, and transfer to HtaB, HmuT, and HmuUV [153]. Gram-Negative Heme Acquisition Strategies The outer membrane of the cellular envelope of Gram-negative organisms presents an additional barrier to heme acquisition. Therefore, Gram-negative heme uptake systems consist of outer-membrane receptors that either bind heme and hemoproteins directly, or bind heme-bound secreted hemophores. Heme then transits the periplasm and is brought into the cell via ABC transporters at the inner membrane. The versatile opportunistic pathogen P. aeruginosa encodes direct heme uptake and hemophore systems at the outer membrane, H. influenzae uses a hemophore uptake system, and Neisseria meningitidis uses a unique bipartite receptor for heme acquisition from host hemoproteins . These pathogens are presented as models for Gram-negative heme uptake systems. P. aeruginosa P. aeruginosa encodes direct and indirect systems for heme uptake. The Phu (Pseudomonas heme uptake) consists of a TonB-dependent PhuR which binds heme and transports it to the periplasm. PhuR activity is representative of Gram-negative TonB-dependent outer-membrane receptors. These β-barrel proteins bind substrates (often iron containing molecules) with high affinity, and rely on proton motive force and TonB for transport across the outer membrane [154]. TonB is an inner-membrane protein with a substantial periplasmic portion for direct interaction with periplasmic domains of the outer membrane proteins. Upon PhuR translocation of heme into the periplasm, the soluble periplasmic protein PhuT binds heme and brings it to PhuUV, an ABC transporter at the inner membrane. In addition, HasA/HasR (heme assimilation system) is utilized for heme uptake. HasA is a secreted hemophore that binds heme and transfers it to a second TonB-dependent transporter, HasR. Like other Gram-negative heme-binding motifs, HasA coordinates heme using histidine and tyrosine residues with picomolar affinity. Data from the orthologous HasA hemophore of Serratia marcescens suggest that HasA binds Hb and extracts heme, then HasA transfers heme to HasR [155,156]. The present model (Fig. 3) for these two heme uptake systems suggests that Phu is the principle heme acquisition system but full heme utilization requires HasA/HasR. HasA/HasR may be more relevant as a heme sensing system; under low heme conditions, the inner-membrane HasS binds the sigma factor inhibitor HasI. When heme is available, HasS instead binds HasR, and HasI is free to recruit RNA polymerase to activate the transcription of hasAR, hasSI, phuSTUV, and phuR [157]. The P. aeruginosa heme uptake system PhuSTUV/PhuR is regulated by Fur in addition to the HasI sigma factor detailed above. Recently, small regulatory RNAs have been described that impact phuS mRNA levels, suggesting another layer of heme-responsive regulation [158,159] In contrast to many other organisms, Pseudomonas encodes a soluble cytoplasmic heme-binding protein that is not a heme oxygenase. This protein, PhuS, transfers heme to the heme-oxygenase HemO for iron liberation. PhuS, unlike many hemoproteins, binds ferric-iron heme and subsequently transfers it to HemOunder iron-deplete conditions [160]. The dissociation constant of the heme–PhuS–HemO complex is in the nanomolar range, suggesting that PhuS transfers heme to HemO specifically and not to the second Pseudomonas heme oxygense, BphO [160]. While the PhuS heme transfer has not been described completely, PhuS has been shown to bind heme as a monomer utilizing one of two histidine residues (His209 and His212), and a third binding site exists when PhuS is in dimeric form [161]. Further in vitro characterization and structural analysis has led to a model whereby heme coordination occurs primarily at the His212 ligand and induces a conformational change required for interaction with HemO [162,163]. Additionally, in vitro heme oxygenase activity has been attributed to PhuS; however, the in vivo relevance of this function is unclear as no biliverdin-β (the product of HemO heme catabolism) is detected in a mutant lacking hemO [164,165]. A recent clinical evaluation of genetic changes to P. aeruginosa during infection of cystic fibrosis lungs revealed the importance of heme acquisition during infection [166]. Long-term infection led to the selection of mutations in the promoters of the phuSTUWV and phuR loci, resulting in greater Phu expression. These changes to phu transcription confer a growth advantage enabling the utilization of heme from Hb as the sole iron source and suggest that heme is an important iron source during chronic Pseudomonas infection. The infections also selected for mutants that demonstrate enhanced expression of the feo ferrous-iron acquisition genes, indicating that ferrous iron is also a source of bioavailable iron. These clinical data confirm experimental findings suggesting that P. aeruginosa heme acquisition contributes to chronic infection. H. influenzae H. influenzae is a notable exception to the other pathogens outlined here, as it is incapable of synthesizing heme and therefore requires heme uptake for aerobic respiration [167]. It is capable of acquiring heme from diverse host sources (Fig. 3), including hemopexin, free heme, albumin-bound heme, myoglobin, and Hb; the variety of heme sources is in accordance with its absolute reliance on exogenous heme [168]. H. influenzae has evolved a variety of heme uptake systems important for growth in vitro using various host heme sources. While some systems are well described, less is known about others, and a global understanding of the utilization of these heme uptake systems during infection is lacking. The HxuCBA system, described primarily in H. influenzae type B, is capable of heme acquisition from free heme and heme-hemopexin (Hx). HxuA is a secreted hemophore that is released from the outer membrane by its transporter HxuB [169–171]. HxuA exhibits no heme-binding motif but rather demonstrates high-affinity binding specifically to Hx with little distinction between apo- and holo-Hx [172]. HxuC is a TonB-dependent transporter that binds heme after release from the Hx–heme–HxuA complex and imports it into the periplasm [173]. Additionally, HxuC is capable of acquiring heme from serum albumin (Alb) independent of HxuA [174]. HpbA is another heme acquisition protein identified in nontypeable and type B H. influenzae. A lipoprotein, HbpA is important for growth using Hb, Hp–Hb, and heme-human serum albumin as heme sources [175,176]. The inner-membrane heme transporter has not been definitively identified, but the Hip proteins have been implicated [174]. Additionally, H. influenzae encodes three receptors, HgpA, HgpB, and HgpC, that can acquire heme from Hp–Hb and Hp bound myoglobin, albeit it at greater concentrations than thought to be physiologically relevant [177,178]. While the contribution of the Hgps seems redundant, HgpB has been demonstrated to be most important for utilization of Hp–Hb and Hp–myoglobin. There are many outstanding questions regarding H. influenzae heme uptake. Many proteins have been attributed to be involved in heme uptake, but their function requires further investigation [179–183]. The regulation of the heme uptake system expression is not well described, except that hxuCBA and the hgp genes are expressed under in vitro iron/heme deplete conditions during experimental infection of the chinchilla ear [184,185]. Lastly, a heme oxygenase of Haemophilus has not been described, suggesting that acquired heme is utilized intact and that other iron acquisition pathways, from transferrin and lactoferrin sources, are sufficient for cellular iron needs. However, it is also possible that a heme oxygenase exists and has not yet been identified. Genetic evidence from clinical isolates suggests that heme uptake is vital to pathogenic strains of H. influenzae. Isolates from otitis media infection in children relative to commensal throat isolates exhibit greater rates of hxuA, hxuB, hxuC, and hgpB gene prevalence, indicating that heme uptake may be a virulence determinant [186,187]. Several animal models have been used to demonstrate the role of heme uptake during H. influenzae infection. In a model of H. influenzae bacteremia, infant rats infected with a mutant lacking HbpA completely clear the infection after one week while rats infected with wildtype remain infected [176]. Likewise, a mutant lacking both HxuC and HgpABC uptake proteins is unable to cause bacteremia in the same rat model [188]. Additionally, the Hgp proteins are required to cause otitis media in a chinchilla model [189]. It is clear that for H. influenzae pathogenesis, heme uptake is a critical virulence determinant. N. meningitidis N. meningitidis encodes a bipartite heme uptake system consisting of HpuAB and HmbR (Fig. 3). HpuAB is expressed from an iron-repressed operon and consists of the HpuA lipoprotein and HpuB, the TonB-dependent receptor capable of binding Hb, apo-Hp, and Hp–Hb [190,191]. Upon heme transport into the cytoplasm, the HemO heme oxygenase degrades heme to biliverdin and liberates iron. As such, HemO is required for survival using heme, Hb, or Hp–Hb as a sole iron source [192,193]. Heme is extracted from these hemoproteins and is imported intact, as Hb can complement the deficiencies of a heme synthesis mutant in an HpuAB-dependent manner [194]. The inner-membrane transporter has not yet been identified, but a zinc transporter has been implicated [195]. Initial studies of the individual function of HpuA and HpuB failed to describe the role of HpuA in heme acquisition. HpuB is sufficient to bind Hb, but a high-affinity HpuB-Hb complex requires the presence of HpuA, even though HpuA-Hb binding was not detected by a flow cytometry assay [196,197]. Additionally, HpuA is required for growth with Hb as a sole iron source and heme import [198]. However, a recent structural characterization has described a direct, albeit weak, interaction between HpuA and Hb, and a co-crystal structure of Hb and an HpuA homolog from Kingella denitrificans has been solved [199]. While these data are not conclusive, they suggest that HpuA and HpuA homologs interact with Hb, and this interaction is required for HpuAB-mediate heme uptake. HmbR (Hb receptor) is an additional N. meningitidis heme uptake protein that binds host Hb with species specificity, exhibiting a greater utilization of human Hb but is unable to bind the Hp-Hb complex and therefore likely binds free Hb only [200,201]. Like HpuAB, it is subject to phase variation [202]. HmbR, based on spectroscopy and mutational analysis, also coordinates heme with a Tyr residue, which further confirms that diverse heme-binding domains have evolved to utilize tyrosine as the axial ligand [203]. The mechanism of heme extraction by HmbR, the associated inner-membrane heme transporter that partners with HmbR extraction, and structural descriptions of ligand binding are still undescribed for HmbR heme uptake. In N. meningitidis, expression of hemO and hmbR is regulated by Fur as well as the MisRS TCS [204,205]. MisRS activates the expression of hemO and hmbR independent of Hb and iron concentration, which suggests an additional layer of regulation for Hb acquisition. However, the activating signal of MisRS has not yet been described. The genetic diversity of N. meningitidis clinical isolates has highlighted the importance of heme uptake to meningococcal virulence. While not all N. meningitidis strains express both the HmbR and HpuAB systems, most express at least one. Most pathogenic isolates express at least HmbR, but HpuAB expression is equally associated with disease and carriage isolates, which indicates that HmbR is an indicator of pathogenesis [206,207]. N. meningitidis serotype B isolates associated with disease also exhibit “on” phase variation of HmbR, correlating virulence with the expression of HmbR [208]. Additionally, HmbR is required for virulence in an infant rat model of meningitis [200]. These data implicate heme uptake, particularly HmbR, as an important component of Neisseria infection. Current challenges and opportunities Study of heme uptake strategies has offered great insight into bacterial pathogenesis and nutrient acquisition. There is still great opportunity for discovery. For most bacterial heme-binding motifs, the transfer from host hemoprotein has not been demonstrated as either passive dissociation or active extraction. The redundancy of heme uptake systems in pathogens like B. anthracis, P. aeruginosa, and H. influenzae is well appreciated, but the role of each system during infection of various niches or commensal environments has yet to be fully elucidated. The relative contribution of host heme to iron acquisition by bacterial pathogens during infection is understudied. It is unclear if pathogens rely on heme for iron in unique spatiotemporal niches and rely on ferrous iron and siderophore acquisition systems in other niches. Opportunity abounds to understand the role of heme-iron utilization across time and tissues during infection. Finally, while global abundance of heme and Hb in the host has been measured, the local availability of heme and hemoproteins during infection has not been described and presents an opportunity to understand the microenvironment of an infectious niche as well as the host response to infection. In terms of clinical application, heme uptake systems may be attractive therapeutic targets. S. aureus Isd proteins have been the target of vaccine development with mixed success and monoclonal antibodies against IsdB have been studied for therapeutic use [209–213]. Considering the importance of heme acquisition to infection, using surface-exposed heme uptake proteins as targets for vaccine and antibodies should continue to be investigated. Additionally, the Mycobacterium tuberculosis heme uptake system, which comprises three unique proteins and is sufficient to rescue the growth of a heme auxotroph, has been proposed as a new mycobacteria-specific antimicrobial target to be explored [214–216] The interactions between host hemoproteins and bacterial hemoprotein binding proteins offer an excellent opportunity to study host–pathogen co-evolution. It has been recently demonstrated that the human and primate iron-binding protein transferrin has undergone positive selection at the interface of binding by bacterial transferrin receptors, suggesting that the co-evolution of humans and pathogens has produced an evolutionary arms race in the context of nutritional immunity [217,218]. In the same vein, the Hb-binding IsdB of S. aureus exhibits species specificity and more efficiently utilizes human Hb relative to mouse Hb [86]. In keeping with this, transgenic mice expressing human Hb are more susceptible to S. aureus disease [86]. The contribution of bacterial heme acquisition to human evolution presents ample opportunity to further investigate co-evolution and nutritional immunity. Heme Toxicity and Tolerance Bacterial pathogens dedicate extensive cellular machinery to the synthesis and acquisition of heme. Paradoxically, excess heme is toxic and thus during infection, invading pathogens must contend with heme toxicity as a component of pathogenesis. While heme toxicity is well studied in eukaryotes, less is known in bacteria [25,219,220]. A brief description of heme toxicity in bacteria and strategies utilized to combat toxicity follow. Multi-faceted mechanism of heme toxicity The reactive nature of heme that makes it such a versatile cofactor also results in toxicity at excess concentrations. While the toxicity of heme toward bacteria has been observed for over 60 years, a complete understanding of the mechanisms of heme toxicity is lacking [26,221]. Free heme is rapidly bactericidal toward various Gram-positive and Gram-negative pathogens in low- to mid-micromolar concentrations [33,222-225]. However, investigation of heme toxicity in a variety of bacterial species has led to a model of heme inducing iron- and non-iron-related damage to the cell. The accumulation of heme results in excess iron by one of two mechanisms, both of which are likely at play under aerobic conditions. First, a portion of iron is freed by the heme oxygenases. Second, iron itself may be liberated from the porphyrin ring upon reaction with reactive oxygen species (ROS). Irrespective of the source, iron can cycle between ferrous and ferric states via Fenton chemistry and the Haber-Weiss reaction (reaction 1), yielding a regenerating supply of ROS. (1) Fe2++H2O2→Fe3++HO•+OH− (2) Fe3++H2O2→Fe2++HOO•+H+ Iron-mediated production of ROS can damage DNA, lipids, and proteins [226,227]. Further evidence for the contribution of oxidative stress to heme toxicity comes from S. aureus. In conditions of excess heme toxicity, membrane proteins are highly oxidized and superoxide is formed by redox cycling of heme-iron through membrane menaquinone [228]. Superoxide production is a separate source of oxidative damage from ferrous iron-mediated ROS and is a major component of heme damage in S. aureus [228]. In addition to experimentally validating that heme-mediated ROS is a key to heme toxicity, this work also localized heme toxicity primarily to the membrane. The lipophilic nature of heme suggests that it partitions to the membrane of bacteria, and this has been demonstrated in S. aureus, likely resulting in damage to membrane proteins and lipids [84]. Further evidence suggests that iron-mediated ROS production and subsequent membrane damage are an insufficient description of heme toxicity. First, heme is toxic in anaerobic conditions, and second, non-iron protoporphyrins are toxic to bacteria and activate the cellular response to heme toxicity [229–231]. Also, porphyrins cause significant damage to bacterial DNA [232]. Finally, resistance to heme toxicity is in part mediated in N. meningitidis by Ght (gene of hydrophobic agent tolerance), suggesting that damage by heme is similar to other hydrophobic molecules and may disrupt the Gram-negative outer membrane [224,233]. The toxicity of heme is likely the result of a combination of membrane disruption, membrane protein and lipid oxidation, and DNA damage. However, a total understanding of heme-mediated damage is far from complete. Strategies to overcome heme toxicity While the direct result of excess heme is unclear, it is evident that bacteria must contend with heme damage and have evolved a variety of strategies to overcome heme toxicity (Fig. 4). These systems consist primarily of efflux and sequestration. Additionally, the heme oxygenase outlined as part of heme acquisition strategies may contribute to the reduction of heme toxicity by cleaving the porphyrin ring and liberating iron for use. Heme efflux strategies have been primarily characterized in Gram-positive organisms, potentially because efflux across a single-membrane barrier is simpler to achieve than in Gram-negative pathogens. Three systems have been described, HrtAB, PefAB/CD, and MtrCDE. The S. aureus heme-regulated transporter HrtAB is required for survival in toxic concentrations of heme. hrtAB expression is activated by the HssRS heme sensing TCS [222,234,235]. While the ligand of the HssS histidine kinase has remained elusive, excess exogenous or endogenous heme leads to activation, either directly or indirectly [99]. HrtA is an ATPase that drives efflux by HrtB permease of its ligand, likely heme. Orthologues of HrtAB have been described in B. anthracis and Lactococcus lactis and are required for resistance to heme toxicity in these organisms [236]. When the Hrt efflux pump is inactivated in both S. aureus and L. lactis, levels of intracellular heme increase, suggesting that heme is the substrate of HrtAB export [229,237]. In B. anthracis, an HssRS orthologue controls the expression of HrtAB and cross-talks with a second TCS that responds to cellular envelope stresses, further implicating membrane damage as a component of heme stress [238]. HrtAB is actively expressed during murine anthrax, suggesting that organisms that replicate in the bloodstream must tolerate heme toxicity [33]. Additional efflux systems exist, suggesting that this strategy is well conserved. Streptococcus agalactiae encodes an orthologue of HrtAB, as well as a dual efflux system PefAB and PefRCD [223]. In heme stress, hrtAB and pefAB/RCD are expressed at high levels, and the Pef systems are required for resistance to heme toxicity [223]. The Gram-negative N. gonorrhoeae encodes an efflux pump, MtrCDE, for hydrophobic molecules that is required for resistance to heme stress [239]. Heme sequestration and storage is a second theme in strategies to resist heme toxicity. The conserved HemS family has been described in Yersinia enterocolitica, Y. pestis, Shigella dysenteriae (termed ShuS), P. aeruginosa (called PhuS, detailed above), and E. coli (ChuS, which also has heme oxygenase activity) [146,160,225,240–244]. While a variety of heme storage, transfer, and degradation properties have been assigned to these proteins, their involvement in resisting heme toxicity is clear. Additionally, non-HemS family proteins have been found to bind heme and play a role in heme homeostasis, including the small outer-membrane Protein E of H. influenzae and the Cu,Zn superoxide dismutase of Haemophilus ducreyi [245,246]. Current challenges and opportunities While numerous systems are involved in detoxifying heme, there are many outstanding questions. The efflux systems have been described genetically, but a complete understanding of the ligands exported is still murky. For Gram-positive pathogens, the efflux systems may provide an additional therapeutic target for infection. Inhibition of efflux may offer a treatment option for bloodstream infections by S. aureus and B. anthracis; presumably the effects of heme toxicity would be deadly to the bacterium if the HrtAB pump were pharmacologically inactivated. This strategy could also pair well with small molecule activation of heme synthesis, which has been developed [99]. In terms of heme sequestration proteins, it has been difficult to fully interpret the contribution of heme sequestration because additional properties like oxygenase (PhuS and ChuS) and DNA binding (ShuS) have been observed. Finally, the role of heme oxygenases in resisting heme stress has not been well studied. Concluding Remarks Heme synthesis, uptake, utilization, and toxicity have been an area of intense investigation in bacterial pathogenesis. As outlined throughout, there are many additional questions in this field. Some of the most fundamental aspects of heme homeostasis have not been studied in detail. For most pathogens, regulation of heme synthesis is unclear and the contribution of heme synthesis to infection has not been investigated. For organisms that acquire and synthesize heme, a full model of preference between exogenous and endogenous sources is unknown. Based on limited evidence, exogenous heme is preferred when available, but is there a division between exogenous and endogenous heme in the partitioning to hemoproteins and heme oxygenases? Also, when heme enters the cell or is synthesized, does it exist in the free state or are there heme chaperones analogous to metallochaperones? As the heme uptake, synthesis, and toxicity processes are well conserved and vital to bacterial pathogenesis, they present an opportunity for therapeutic intervention. As the field gains further insight into these processes, hopefully academia and industry will pursue small molecule interventions and vaccine candidates for the treatment of the bacterial pathogens outlined in this review, many of which are recalcitrant current treatment options. We thank members of the Skaar laboratory for critical review of this manuscript. J.E.C. is supported by NIH T32 GM065086. Heme acquisition, synthesis, and toxicity research in the Skaar laboratory is supported by NIH grants AI069233 and AI073843. Abbreviations Hb hemoglobin Hp haptoglobin ALA δ-aminolevulinic acid PBG porphobilinogen HMB hydroxymethylbilane Hx hemopexin SCV small colony variant PPIX protoporphyrin IX NEAT near-transporter domain Isd iron-regulated surface determinants TCS two-component system Fig. 1 Bacterial heme biosynthesis The heme synthesis pathway of most bacteria begins with charged glutamyl-tRNAGlu to form the universal precursor ALA, and coproporphyrinogen III isformed through a series of conserved enzymatic steps. The classical pathway (blue) forms heme through the protoporphyrinogen IX intermediate; most organisms including Gram-negative bacteria and eukaryotes use this pathway. The noncanonical pathway (green), performed by most Gram-positive bacteria, produces heme through the coproporphyrin III intermediate. Shown for each step is the enzyme name followed by the common protein annotation in bold. Fig. 2 Gram-positive heme uptake systems The iron-regulated surface determinant (Isd) systems for heme acquisition in S. aureus and B. anthracis, as well as the non-Isd systems of C. diphtheriae are diagrammed. Host Hb, Hp-bound Hb, and free heme (Fe-containing ring) can serve as heme sources during infection. (a) In S. aureus, IsdH is the primary Hp–Hb receptor and IsdB is the principle Hb receptor. Both are sortase-linked on the surface of the cell wall, bind host hemoproteins with NEAT domains, and extract heme using additional NEAT domains. IsdA can bind free heme or accept heme from IsdB and IsdH. Heme is transferred to IsdC, which is embedded in the cell wall and transits heme to the membrane complex IsdDEF. IsdDEF transports heme to the cytoplasm for utilization intact or for degradation by the heme oxygenases IsdG/I. (b) Similarly, B. anthracis uses Isd proteins to acquire heme. IsdX1 and IsdX2 are secreted hemophores that bind Hb, Hp–Hb, or free heme as depicted. IsdX2, which has five NEAT domains, may also serve as a heme storage protein. Additionally, the sortase anchored Hal serves as a Hb receptor on the cell surface and uses its NEAT and leucine-rich repeat domains to acquire heme. BslK is cell wall associated and binds heme via its NEAT domain. IsdC transports heme to the IsdDEF membrane importer for utilization or degradation by IsdG. (c) C. diphtheriae utilizes a unique set of heme uptake proteins for heme utilization. HtaA is a cell wall spanning lipoprotein that can acquire heme from Hp–Hb in conjunction with ChtA or ChtC. HtaB can bind free heme or accept heme transfer from HtaA and transfers heme to the HmuTUV membrane transporter. A portion of HtaA may also serve as a secreted hemophore. C. diphtheriae HmuO heme oxygenase can liberate iron from imported heme. Fig. 3 Gram-negative heme acquisition The heme uptake systems as described in the text are depicted. (a) P. aeruginosa PhuR binds heme at the outer membrane and imports heme into the periplasm in a TonB-dependent manner. Heme is transferred to PhuT, which subsequently transfers heme to the PhuUV inner-membrane transporter for transit into the cytoplasm. There, PhuS binds and stores heme or transfers heme to the heme oxygenase HemO for iron utilization. P. aeruginosa also secretes the hemophore HasA which binds Hb or free heme, and transfers heme to the TonB-dependent outer-membrane receptor HasR. The fate of HasR imported heme is not fully understood, but may be trafficked to PhuTUV for import. HasS serves as an inner-membrane sensor and regulates expression of the has and phu systems through the sigma factor HasI (not shown). (b) H. influenzae can utilize a variety of host heme sources. Secreted HxuA specifically binds hemopexin (Hx), and heme from Hx is transferred into the periplasm when HxuA interacts with HxuBC at the outer membrane. Independent of HxuA, HxuC can also import heme from serum albumin (Alb). HgpA, HgpB, and HgpC are highly similar outer-membrane receptors for heme acquisition from Hb complexed with Hp, free Hb, and Hp bound to myoglobin (not shown). The inner-membrane heme transporter has not been clearly defined, but the Hip system has been implicated for heme transit into the cytoplasm. Interestingly, all imported heme may be utilized intact, as no heme oxygenase has been identified yet. (c) The N. meningitidis outer-membrane, TonB-dependent complex of HpuAB can acquire heme from Hb and Hp–Hb and bring heme into the periplasm. Additionally, the HmbR outer-membrane receptor specifically extracts heme from Hb for transport. The identity of the inner-membrane heme transporter is unclear at this time, but heme somehow enters the cytoplasm where it can be utilized or degraded by the HemO heme oxygenase. Fig. 4 Strategies to avoid heme toxicity Heme toxicity (center) is a combination of heme damage to membrane lipids, membrane proteins, and DNA, and oxidative damage. Oxidative damage is mediated by the production of superoxide dismutase ( O2−•), hydroxyl radical (HO•), and hydroperoxyl radical (HOO•). To reduce heme damage, many Gram-positive organisms (the S. aureus system is diagrammed here) encode the HrtAB efflux pump. The HssRS two-component system responds to excess heme and activates the transcription of the hrtAB system, thus preventing the accumulation of toxic levels of heme. Alternatively, Gram-negative organisms rely on intracellular heme sequestration proteins (PhuS of P. aeruginosa, HemS of Yersinia), the periplasmic heme-binding, copper and zinc dependent superoxide dismutase (Cu,Zn SOD, of H. ducreyi), and systems that respond to hydrophobic molecules, including heme (MtrCDE efflux and Ght of Neisseria). 1 Chen JJ London IM Hemin enhances the differentiation of mouse 3T3 cells to adipocytes Cell 26 1981 117 122 6799206 2 Shelver D Kerby RL He Y Roberts GP CooA, a CO-sensing transcription factor from Rhodospirillum rubrum, is a CO-binding heme protein Proc Natl Acad Sci USA 94 1997 11216 11220 9326589 3 Faller M Matsunaga M Yin S Loo JA Guo F Heme is involved in microRNA processing Nat Struct Mol Biol 14 2007 23 29 17159994 4 Bonyhady RE Hendry IA Hill CE McLennan IS Effects of haemin on neurones derived from the neural crest Dev Neurosci 5 1982 125 129 6127199 5 Posey JE Gherardini FC Lack of a role for iron in the Lyme disease pathogen Science 288 2000 1651 1654 10834845 6 Archibald F Lactobacillus plantarum, an organism not requiring iron FEMS Microbiol Lett 19 1983 29 32 7 Kessler D Leibrecht I Knappe J Pyruvate-formate-lyasedeactivase and acetyl-CoA reductase activities of Escherichia coli reside on a polymeric protein particle encoded by adhE FEBS Lett 281 1991 59 63 2015910 8 Yamamoto I Saiki T Liu SM Ljungdahl LG Purification and properties of NADP-dependent formate dehydrogenase from Clostridium thermoaceticum, a tungsten–selenium–iron protein J Biol Chem 258 1983 1826 1832 6822536 9 Peters JW Lanzilotta WN Lemon BJ Seefeldt LC X-ray crystal structure of the Fe-only hydrogenase (CpI) from Clostridium pasteurianum to 1.8 Angstrom resolution Science 282 1998 1853 1858 9836629 10 Nagashima S Nakasako M Dohmae N Tsujimura M Takio K Odaka M Novel non-heme iron center of nitrile hydratase with a claw setting of oxygen atoms Nat Struct Biol 5 1998 347 351 9586994 11 Johnson DC Dean DR Smith AD Johnson MK Structure, function, and formation of biological iron–sulfur clusters Annu Rev Biochem 74 2005 247 281 15952888 12 Shomura Y Yoon KS Nishihara H Higuchi Y Structural basis for a [4Fe–3S] cluster in the oxygen-tolerant membrane-bound [NiFe]-hydrogenase Nature 479 2011 253 256 22002607 13 Weinberg ED Iron and susceptibility to infectious disease Science 184 1974 952 956 4596821 14 Weinberg ED Nutritional immunity: host's attempt to withold iron from microbial invaders JAMA 231 1975 39 41 1243565 15 Cassat JE Skaar EP Iron in infection and immunity Cell Host Microbe 13 2013 509 519 23684303 16 Hood MI Skaar EP Nutritional immunity: transition metals at the pathogen-host interface Nat Rev Microbiol 10 2012 525 537 22796883 17 Zackular JP Chazin WJ Skaar EP Nutritional immunity: S100 proteins at the host–pathogen interface J Biol Chem 290 2015 18991 18998 26055713 18 Juttukonda LJ Skaar EP Manganese homeostasis and utilization in pathogenic bacteria Mol Microbiol 97 2015 216 228 25898914 19 Wriston JC Jr Lack L Shemin D The mechanism of porphyrin formation; further evidence on the relationship of the citric acid cycle and porphyrin formation J Biol Chem 215 1955 603 611 13242557 20 Kikuchi G Shemin D Bachmann BJ The enzymic synthesis of delta-aminolevulinic acid Biochim Biophys Acta 28 1958 219 220 13535711 21 Hamza I Dailey HA One ring to rule them all: trafficking of heme and heme synthesis intermediates in the metazoans Biochim Biophys Acta 2012 1823 1617 1632 22 Rieske JS Composition, structure and function of complex III of the respiratory chain Biochim Biophys Acta 456 1976 195 247 788795 23 Omura T Sato R The carbon monoxide-binding pigment of liver microsomes. I. Evidence for its hemoprotein nature J Biol Chem 239 1964 2370 2378 14209971 24 Bridges KR Seligman PA Disorders of Iron Metabolism J.B Lippincott Company New York 1995 25 Everse J Hsia N The toxicities of native and modified hemoglobins Free Radic Biol Med 22 1997 1075 1099 9034247 26 Anzaldi LL Skaar EP Overcoming the heme paradox: heme toxicity and tolerance in bacterial pathogens Infect Immun 78 2010 4977 4989 20679437 27 Smith A McCulloh RJ Hemopexin and haptoglobin: allies against heme toxicity from hemoglobin not contenders Front Physiol 6 2015 187 26175690 28 Schaer DJ Buehler PW Cell-free hemoglobin and its scavenger proteins: new disease models leading the way to targeted therapies Cold Spring Harb Perspect Med 3 2013 29 Jeney V Balla J Yachie A Varga Z Vercellotti GM Eaton JW Pro-oxidant and cytotoxic effects of circulating heme Blood 100 2002 879 887 12130498 30 Tolosano E Altruda F Hemopexin: structure, function, and regulation DNA Cell Biol 21 2002 297 306 12042069 31 Lin H McFaul SJ Brady JC Everse J The mechanism of peroxidase-mediated cytotoxicity. II. Role of the heme moiety Proc Soc Exp Biol Med 187 1988 7 13 3340620 32 Balla G Jacob HS Eaton JW Belcher JD Vercellotti GM Hemin: a possible physiological mediator of low density lipoprotein oxidation and endothelial injury Arter-ioscler Thromb 11 1991 1700 1711 33 Stauff DL Skaar EP Bacillus anthracis HssRS signalling to HrtAB regulates haem resistance during infection Mol Microbiol 72 2009 763 778 19400785 34 Dailey HA Gerdes S Dailey TA Burch JS Phillips JD Noncanonical coproporphyrin-dependent bacterial heme biosynthesis pathway that does not use protoporphyrin Proc Natl Acad Sci USA 112 2015 2210 2215 25646457 35 Beale SI Gough SP Granick S Biosynthesis of delta-aminolevulinic acid from the intact carbon skeleton of glutamic acid in greening barley Proc Natl Acad Sci USA 72 1975 2719 2723 1058487 36 Hoober JK Kahn A Ash DE Gough S Kannangara CG Biosynthesis of delta-aminolevulinate in greening barley leaves. IX. Structure of the substrate, mode of gabaculine inhibition, and the catalytic mechanism of glutamate 1-semialdehyde aminotransferase Carlsb Res Commun 53 1988 37 Kannangara CG Gough SP Bruyant P Hoober JK Kahn A von Wettstein D tRNA(Glu) as a cofactor in delta-aminolevulinate biosynthesis: steps that regulate chlorophyll synthesis Trends Biochem Sci 13 1988 139 143 3075378 38 Hansson M Rutberg L Schroder I Hederstedt L The Bacillus subtilis hemAXCDBL gene cluster, which encodes enzymes of the biosynthetic pathway from glutamate to uroporphyrinogen III J Bacteriol 173 1991 2590 2599 1672867 39 Louie GV Brownlie PD Lambert R Cooper JB Blundell TL Wood SP Structure of porphobilinogen deaminase reveals a flexible multidomain polymerase with a single catalytic site Nature 359 1992 33 39 1522882 40 Bung N Pradhan M Srinivasan H Bulusu G Structural insights into E. coli porphobilinogen deaminase during synthesis and exit of 1-hydroxymethylbilane PLoS Comput Biol 10 2014 e1003484 24603363 41 Jordan PM The biosynthesis of tetrapyrroles Elsevier New York, NY 1991 42 Hansson M Hederstedt L Cloning and characterization of the Bacillus subtilis hemEHY gene cluster, which encodes protoheme IX biosynthetic enzymes J Bacteriol 174 1992 8081 8093 1459957 43 Leeper FJ The biosynthesis of porphyrins, chlorophylls, and vitamin B12 Nat Prod Rep 4 1987 441 469 3325863 44 Bali S Lawrence AD Lobo SA Saraiva LM Golding BT Palmer DJ Molecular hijacking of siroheme for the synthesis of heme and d1 heme Proc Natl Acad Sci USA 108 2011 18260 18265 21969545 45 Bali S Palmer DJ Schroeder S Ferguson SJ Warren MJ Recent advances in the biosynthesis of modified tetrapyrroles: the discovery of an alternative pathway for the formation of heme and heme d 1 Cell Mol Life Sci 71 2014 2837 2863 24515122 46 Troup B Hungerer C Jahn D Cloning and characterization of the Escherichia coli hemN gene encoding the oxygen-independent coproporphyrinogen III oxidase J Bacteriol 177 1995 3326 3331 7768836 47 Seehra JS Jordan PM Akhtar M Anaerobic and aerobic coproporphyrinogen III oxidases of Rhodopseudomonas spheroides. Mechanism and stereochemistry of vinyl group formation Biochem J 209 1983 709 718 6603215 48 Boynton TO Daugherty LE Dailey TA Dailey HA Identification of Escherichia coli HemG as a novel, menadione-dependent flavodoxin with protoporphyrinogen oxidase activity Biochemistry 48 2009 6705 6711 19583219 49 Kobayashi K Masuda T Tajima N Wada H Sato N Molecular phylogeny and intricate evolutionary history of the three isofunctional enzymes involved in the oxidation of protoporphyrinogen IX Genome Biol Evol 6 2014 2141 2155 25108393 50 Porra RJ Falk JE The enzymic conversion of coproporphyrinogen III into protoporphyrin IX Biochem J 90 1964 69 75 5832302 51 Camadro JM Labbe P Purification and properties of ferrochelatase from the yeast Saccharomyces cerevisiae. Evidence for a precursor form of the protein J Biol Chem 263 1988 11675 11682 3042776 52 Lobo SA Scott A Videira MA Winpenny D Gardner M Palmer MJ Staphylococcus aureus haem biosynthesis: characterisation of the enzymes involved in final steps of the pathway Mol Microbiol 97 2015 472 487 25908396 53 Hansson M Gustafsson MC Kannangara CG Hederstedt L Isolated Bacillus subtilis HemY has copropor-phyrinogen III to coproporphyrin III oxidase activity Biochim Biophys Acta 1340 1997 97 104 9217019 54 Mayfield JA Hammer ND Kurker RC Chen TK Ojha S Skaar EP The chlorite dismutase (HemQ) from Staphylococcus aureus has a redox-sensitive heme and is associated with the small colony variant phenotype J Biol Chem 288 2013 23488 23504 23737523 55 Dailey HA Gerdes S HemQ: an iron-coproporphyrin oxidative decarboxylase for protoheme synthesis in Firmicutes and Actinobacteria Arch Biochem Biophys 574 2015 27 35 25711532 56 Dailey TA Boynton TO Albetel AN Gerdes S Johnson MK Dailey HA Discovery and characterization of HemQ: an essential heme biosynthetic pathway component J Biol Chem 285 2010 25978 25986 20543190 57 Celis AI Streit BR Moraski GC Kant R Lash TD Lukat-Rodgers GS Unusual peroxide-dependent, heme-transforming reaction catalyzed by HemQ Biochemistry 54 2015 4022 4032 26083961 58 Jahn D Michelsen U Soll D Two glutamyl-tRNA reductase activities in Escherichia coli J Biol Chem 266 1991 2542 2548 1990004 59 Javor GT Febre EF Enzymatic basis of thiol-stimulated secretion of porphyrins by Escherichia coli J Bacteriol 174 1992 1072 1075 1732201 60 Wang LY Brown L Elliott M Elliott T Regulation of heme biosynthesis in Salmonella typhimurium: activity of glutamyl-tRNA reductase (HemA) is greatly elevated during heme limitation by a mechanism which increases abundance of the protein J Bacteriol 179 1997 2907 2914 9139907 61 Wang L Elliott M Elliott T Conditional stability of the HemA protein (glutamyl-tRNA reductase) regulates heme biosynthesis in Salmonella typhimurium J Bacteriol 181 1999 1211 1219 9973348 62 Jones AM Elliott T A purified mutant HemA protein from Salmonella enterica serovar Typhimurium lacks bound heme and is defective for heme-mediated regulation in vivo FEMS Microbiol Lett 307 2010 41 47 20412302 63 Wang L Wilson S Elliott T A mutant HemA protein with positive charge close to the N terminus is stabilized against heme-regulated proteolysis in Salmonella typhimurium J Bacteriol 181 1999 6033 6041 10498716 64 Zhang J Kang Z Chen J Du G Optimization of the heme biosynthesis pathway for the production of 5-aminolevulinic acid in Escherichia coli Sci Rep 5 2015 8584 25716896 65 Schroder I Johansson P Rutberg L Hederstedt L The hemX gene of the Bacillus subtilis hemAXCDBL operon encodes a membrane protein, negatively affecting the steady-state cellular concentration of HemA (glutamyl-tRNA reductase) Microbiology 140 1994 731 740 8012594 66 Krieger R Rompf A Schobert M Jahn D The Pseudomonas aeruginosa hemA promoterisregulatedby Anr, Dnr, NarL and Integration Host Factor Mol Gen Genomics 267 2002 409 417 67 Hungerer C Troup B Romling U Jahn D Regulation of the hemA gene during 5-aminolevulinic acid formation in Pseudomonas aeruginosa J Bacteriol 177 1995 1435 1443 7883699 68 Ye RW Tao W Bedzyk L Young T Chen M Li L Global gene expression profiles of Bacillus subtilis grown under anaerobic conditions J Bacteriol 182 2000 4458 4465 10913079 69 Ji CJ Kim JH Won YB Lee YE Choi TW Ju SY Staphylococcus aureus PerR Is a hypersensitive hydrogen peroxide sensor using iron-mediated histidine oxidation J Biol Chem 290 2015 20374 20386 26134568 70 Ravcheev DA Best AA Tintle N Dejongh M Osterman AL Novichkov PS Inference of the transcriptional regulatory network in Staphylococcus aureus by integration of experimental and genomics-based evidence J Bacteriol 193 2011 3228 3240 21531804 71 Frunzke J Gatgens C Brocker M Bott M Control of heme homeostasis in Corynebacterium glutamicum by the two-component system HrrSA J Bacteriol 193 2011 1212 1221 21217007 72 Bibb LA Kunkle CA Schmitt MP The ChrA–ChrS and HrrA–HrrS signal transduction systems are required for activation of the hmuO promoter and repression of the hemA promoter in Corynebacterium diphtheriae Infect Immun 75 2007 2421 2431 17353293 73 Bibb LA Schmitt MP The ABC transporter HrtAB confers resistance to hemin toxicity and is regulated in a hemin-dependent manner by the ChrAS two-component system in Corynebacterium diphtheriae J Bacteriol 192 2010 4606 4617 20639324 74 Mancini S Imlay JA The induction of two biosynthetic enzymes helps Escherichia coli sustain heme synthesis and activate catalase during hydrogen peroxide stress Mol Microbiol 96 2015 744 763 25664592 75 Marino M Ramos HC Hoffmann T Glaser P Jahn D Modulation of anaerobic energy metabolism of Bacillus subtilis by arfM (ywiD) J Bacteriol 183 2001 6815 6821 11698370 76 Hippler B Homuth G Hoffmann T Hungerer C Schumann W Jahn D Characterization of Bacillus subtilis hemN J Bacteriol 179 1997 7181 7185 9371469 77 Homuth G Rompf A Schumann W Jahn D Transcriptional control of Bacillus subtilis hemN and hemZ J Bacteriol 181 1999 5922 5929 10498703 78 Chen L Keramati L Helmann JD Coordinate regulation of Bacillus subtilis peroxide stress genes by hydrogen peroxide and metal ions Proc Natl Acad Sci USA 92 1995 8190 8194 7667267 79 Rompf A Hungerer C Hoffmann T Lindenmeyer M Romling U Gross U Regulation of Pseudomonas aeruginosa hemF and hemN by the dual action of the redox response regulators Anr and Dnr Mol Microbiol 29 1998 985 997 9767567 80 Sander A Kretzer S Bredt W Oberle K Bereswill S Hemin-dependent growth and hemin binding of Bartonella henselae FEMS Microbiol Lett 189 2000 55 59 10913865 81 Frankenberg L Brugna M Hederstedt L Enterococcus faecalis heme-dependent catalase J Bacteriol 184 2002 6351 6356 12399505 82 Biberstein EL Mini PD Gills MG Action of Haemophilus cultures on δ-aminolevulinic acid J Bacteriol 86 1963 814 819 14066479 83 Baureder M Hederstedt L Heme proteins in lactic acid bacteria Adv Microb Physiol 62 2013 1 43 23481334 84 Skaar EP Humayun M Bae T DeBord KL Schneewind O Iron-source preference of Staphylococcus aureus infections Science 305 2004 1626 1628 15361626 85 Pishchany G Sheldon JR Dickson CF Alam MT Read TD Gell DA IsdB-dependent hemoglobin binding is required for acquisition of heme by Staphylococcus aureus J Infect Dis 209 2014 1764 1772 24338348 86 Pishchany G McCoy AL Torres VJ Krause JC Crowe JE Jr Fabry ME Specificity for human hemoglobin enhances Staphylococcus aureus infection Cell Host Microbe 8 2010 544 550 21147468 87 Hammer ND Reniere ML Cassat JE Zhang Y Hirsch AO Hood MI Two heme-dependent terminal oxidases power Staphylococcus aureus organ-specific colonization of the vertebrate host mBio 4 2013 88 Hammer ND Cassat JE Noto MJ Lojek LJ Chadha AD Schmitz JE Inter- and intraspecies metabolite exchange promotes virulence of antibiotic-resistant Staphylococcus aureus Cell Host Microbe 16 2014 531 537 25299336 89 von Eiff C Heilmann C Proctor RA Woltz C Peters G Gotz F A site-directed Staphylococcus aureus hemB mutant is a small-colony variant which persists intracellularly J Bacteriol 179 1997 4706 4712 9244256 90 Balwit JM van Langevelde P Vann JM Proctor RA Gentamicin-resistant menadione and hemin auxotrophic Staphylococcus aureus persist within cultured endothelial cells J Infect Dis 170 1994 1033 1037 7930701 91 Painter KL Strange E Parkhill J Bamford KB Armstrong-James D Edwards AM Staphylococcus aureus adapts to oxidative stress by producing H2O2-resistant small-colony variants via the SOS response Infect Immun 83 2015 1830 1844 25690100 92 Sadowska B Bonar A von Eiff C Proctor RA Chmiela M Rudnicka W Characteristics of Staphylococcus aureus, isolated from airways of cystic fibrosis patients, and their small colony variants FEMS Immunol Med Microbiol 32 2002 191 197 11934563 93 Proctor RA von Eiff C Kahl BC Becker K McNamara P Herrmann M Small colony variants: a pathogenic form of bacteria that facilitates persistent and recurrent infections Nat Rev Microbiol 4 2006 295 305 16541137 94 Almiron M Martinez M Sanjuan N Ugalde RA Ferrochelatase is present in Brucella abortus and is critical for its intracellular survival and virulence Infect Immun 69 2001 6225 6230 11553564 95 Skurnik D Roux D Aschard H Cattoir V Yoder-Himes D Lory S A comprehensive analysis of in vitro and in vivo genetic fitness of Pseudomonas aeruginosa using high-throughput sequencing of transposon libraries PLoS Pathog 9 2013 e1003582 24039572 96 Palace SG Proulx MK Lu S Baker RE Goguen JD Genome-wide mutant fitness profiling identifies nutritional requirements for optimal growth of Yersinia pestis in deep tissue mBio 2014 5 97 Wang N Ozer EA Mandel MJ Hauser AR Genomewide identification of Acinetobacter baumannii genes necessary for persistence in the lung mBio 5 2014 e01163 14 24895306 98 Cummins J Casey PG Joyce SA Gahan CG A mariner transposon-based signature-tagged mutagenesis system for the analysis of oral infection by Listeria monocytogenes PLoS ONE 8 2013 e75437 24069416 99 Mike LA Dutter BF Stauff DL Moore JL Vitko NP Aranmolate O Activation of heme biosynthesis by a small molecule that is toxic to fermenting Staphylococcus aureus Proc Natl Acad Sci USA 110 2013 9206 8211 100 Dutter BF Mike LA Reid PR Chong KM Ramos-Hunter SJ Skaar EP Decoupling activation of heme biosynthesis from anaerobic toxicity in a molecule active in Staphylococcus aureus ACS Chem Biol 2016 PMID: 26890615, Epub ahead of print http://dx.doi.org/10.1021/acschembio.5b00934 101 Jensen J The effect of heme on tetrapyrrol synthesis in a heme requiring Staphylococcus aureus Biochem Biophys Res Commun 8 1962 271 277 14451646 102 Mazmanian SK Skaar EP Gaspar AH Humayun M Gornicki P Jelenska J Passage of heme-iron across the envelope of Staphylococcus aureus Science 299 2003 906 909 12574635 103 Spaan AN Reyes-Robles T Badiou C Cochet S Boguslawski KM Yoong P Staphylococcus aureus targets the Duffy Antigen Receptor for Chemokines (DARC) to lyse erythrocytes Cell Host Microbe 18 2015 363 370 26320997 104 Grigg JC Vermeiren CL Heinrichs DE Murphy ME Haem recognition by a Staphylococcus aureus NEAT domain Mol Microbiol 63 2007 139 149 17229211 105 Sharp KH Schneider S Cockayne A Paoli M Crystal structure of the heme-IsdC complex, the central conduit of the Isd iron/heme uptake system in Staphylococcus aureus J Biol Chem 282 2007 10625 10631 17287214 106 Mazmanian SK Liu G Ton-That H Schneewind O Staphylococcus aureus sortase, an enzyme that anchors surface proteins to the cell wall Science 285 1999 760 763 10427003 107 Fonner BA Tripet BP Eilers BJ Stanisich J Sullivan-Springhetti RK Moore R Solution structure and molecular determinants of hemoglobin binding of the first NEAT domain of IsdB in Staphylococcus aureus Biochemistry 53 2014 3922 3933 24871270 108 Torres VJ Pishchany G Humayun M Schneewind O Skaar EP Staphylococcus aureus IsdB is a hemoglobin receptor required for heme iron utilization J Bacteriol 188 2006 8421 8429 17041042 109 Dryla A Gelbmann D Von Gabain A Nagy E Identification of a novel iron regulated staphylococcal surface protein with haptoglobin-haemoglobin binding activity Mol Microbiol 49 2003 37 53 12823809 110 Pilpa RM Robson SA Villareal VA Wong ML Phillips M Clubb RT Functionally distinct NEAT (NEAr Transporter) domains within the Staphylococcus aureus IsdH/HarA protein extract heme from methemoglobin J Biol Chem 284 2009 1166 1176 18984582 111 Muryoi N Tiedemann MT Pluym M Cheung J Heinrichs DE Stillman MJ Demonstration of the iron-regulated surface determinant (Isd) heme transfer pathway in Staphylococcus aureus J Biol Chem 283 2008 28125 28136 18676371 112 Mazmanian SK Ton-That H Su K Schneewind O An iron-regulated sortase anchors a class of surface protein during Staphylococcus aureus pathogenesis Proc Natl Acad Sci USA 99 2002 2293 2298 11830639 113 Beveridge J Ultrastructure of Gram-positive cell walls Fishetti VA Novick RP Feretti JJ Portnoy DA Rood JI Gram-Positive Pathogens second ASM Press Washington, DC 2000 3 10 114 Zhu H Xie G Liu M Olson JS Fabian M Dooley DM Pathway for heme uptake from human methemoglobin by the iron-regulated surface determinants system of Staphylococcus aureus J Biol Chem 283 2008 18450 18460 18467329 115 Skaar EP Gaspar AH Schneewind O IsdG and IsdI, heme-degrading enzymes in the cytoplasm of Staphylococcus aureus J Biol Chem 279 2004 436 443 14570922 116 Reniere ML Torres VJ Skaar EP Intracellular metalloporphyrin metabolism in Staphylococcus aureus Biometals 20 2007 333 345 17387580 117 Reniere ML Ukpabi GN Harry SR Stec DF Krull R Wright DW The IsdG-family of haem oxygenases degrades haem to a novel chromophore Mol Microbiol 75 2010 1529 1538 20180905 118 Lee WC Reniere ML Skaar EP Murphy ME Ruffling of metalloporphyrins bound to IsdG and IsdI, two hemed-egrading enzymes in Staphylococcus aureus J Biol Chem 283 2008 30957 30963 18713745 119 Wu R Skaar EP Zhang R Joachimiak G Gornicki P Schneewind O Staphylococcus aureus IsdG and IsdI, heme-degrading enzymes with structural similarity to monooxygenases J Biol Chem 280 2005 2840 2846 15520015 120 Reniere ML Skaar EP Staphylococcus aureus haem oxygenases are differentially regulated by iron and haem Mol Microbiol 69 2008 1304 1315 18643935 121 Farrand AJ Reniere ML Ingmer H Frees D Skaar EP Regulation of host hemoglobin binding by the Staphylococcus aureus Clp proteolytic system J Bacteriol 195 2013 5041 5050 23995637 122 Laakso HA Marolda CL Pinter TB Stillman MJ Heinrichs DE A heme-responsive regulator controls synthesis of Staphyloferrin B in Staphylococcus aureus J Biol Chem 291 2016 29 40 26534960 123 Visai L Yanagisawa N Josefsson E Tarkowski A Pezzali I Rooijakkers SH Immune evasion by Staphylococcus aureus conferred by iron-regulated surface determinant protein IsdH Microbiology 155 2009 667 679 19246738 124 Cheng AG Kim HK Burts ML Krausz T Schneewind O Missiakas DM Genetic requirements for Staphylococcus aureus abscess formation and persistence in host tissues FASEB J 23 2009 3393 3404 19525403 125 Maresso AW Garufi G Schneewind O Bacillus anthracis secretes proteins that mediate heme acquisition from hemoglobin PLoS Pathog 4 2008 e1000132 18725935 126 Fabian M Solomaha E Olson JS Maresso AW Heme transfer to the bacterial cell envelope occurs via a secreted hemophore in the Gram-positive pathogen Bacillus anthracis J Biol Chem 284 2009 32138 32146 19759022 127 Honsa ES Fabian M Cardenas AM Olson JS Maresso AW The five near-iron transporter (NEAT) domain anthrax hemophore, IsdX2, scavenges heme from hemoglobin and transfers heme to the surface protein IsdC J Biol Chem 286 2011 33652 33660 21808055 128 Maresso AW Chapa TJ Schneewind O Surface protein IsdC and Sortase B are required for heme-iron scavenging of Bacillus anthracis J Bacteriol 188 2006 8145 8152 17012401 129 Honsa ES Owens CP Goulding CW Maresso AW The near-iron transporter (NEAT) domains of the anthrax hemophore IsdX2 require a critical glutamine to extract heme from methemoglobin J Biol Chem 288 2013 8479 8490 23364793 130 Balderas MA Nobles CL Honsa ES Alicki ER Maresso AW Hal Is a Bacillus anthracis heme acquisition protein J Bacteriol 194 2012 5513 5521 22865843 131 Gaspar AH Marraffini LA Glass EM Debord KL Ton-That H Schneewind O Bacillus anthracis sortase A (SrtA) anchors LPXTG motif-containing surface proteins to the cell wall envelope J Bacteriol 187 2005 4646 4655 15968076 132 Tarlovsky Y Fabian M Solomaha E Honsa E Olson JS Maresso AW A Bacillus anthracis S-layer homology protein that binds heme and mediates heme delivery to IsdC J Bacteriol 192 2010 3503 3511 20435727 133 Skaar EP Gaspar AH Schneewind O Bacillus anthracis IsdG, a heme-degrading monooxygenase J Bacteriol 188 2006 1071 1080 16428411 134 Gat O Zaide G Inbar I Grosfeld H Chitlaru T Levy H Characterization of Bacillus anthracis iron-regulated surface determinant (Isd) proteins containing NEAT domains Mol Microbiol 70 2008 983 999 18826411 135 Carlson PE Jr Carr KA Janes BK Anderson EC Hanna PC Transcriptional profiling of Bacillus anthracis Sterne (34F2) during iron starvation PLoS One 4 2009 e6988 19768119 136 Farrand AJ Haley KP Lareau NM Heilbronner S McLean JA Foster T An iron-regulated autolysin remodels the cell wall to facilitate heme acquisition in Staphylococcus lugdunensis Infect Immun 83 2015 3578 3589 26123800 137 Zapotoczna M Heilbronner S Speziale P Foster TJ Iron-regulated surface determinant (Isd) proteins of Staphylococcus lugdunensis J Bacteriol 194 2012 6453 6467 23002220 138 Haley KP Janson EM Heilbronner S Foster TJ Skaar EP Staphylococcus lugdunensis IsdG liberates iron from host heme J Bacteriol 193 2011 4749 4757 21764939 139 Malmirchegini GR Sjodt M Shnitkind S Sawaya MR Rosinski J Newton SM Novel mechanism of hemin capture by Hbp2, the hemoglobin-binding hemophore from Listeria monocytogenes J Biol Chem 289 2014 34886 34899 25315777 140 Xiao Q Jiang X Moore KJ Shao Y Pi H Dubail I Sortase independent and dependent systems for acquisition of haem and haemoglobin in Listeria monocytogenes Mol Microbiol 80 2011 1581 1597 21545655 141 Dahesh S Nizet V Cole JN Study of streptococcal hemoprotein receptor (Shr) in iron acquisition and virulence of M1T1 group A streptococcus Virulence 3 2012 566 575 23076332 142 Ouattara M Cunha EB Li X Huang YS Dixon D Eichenbaum Z Shr of group A streptococcus is a new type of composite NEAT protein involved in sequestering haem from methaemoglobin Mol Microbiol 78 2010 739 756 20807204 143 Ouattara M Pennati A Devlin DJ Huang YS Gadda G Eichenbaum Z Kinetics of heme transfer by the Shr NEAT domains of Group A Streptococcus Arch Biochem Biophys 538 2013 71 79 23993953 144 Schmitt MP Utilization of host iron sources by Coryne-bacterium diphtheriae: identification of a gene whose product is homologous to eukaryotic heme oxygenases and is required for acquisition of iron from heme and hemoglobin J Bacteriol 179 1997 838 845 9006041 145 Drazek ES Hammack CA Schmitt MP Corynebacterium diphtheriae genes required for acquisition of iron from haemin and haemoglobin are homologous to ABC haemin transporters Mol Microbiol 36 2000 68 84 10760164 146 Thompson JM Jones HA Perry RD Molecular characterization of the hemin uptake locus (hmu) from Yersinia pestis and analysis of hmu mutants for hemin and hemoprotein utilization Infect Immun 67 1999 3879 3892 10417152 147 Allen CE Schmitt MP HtaA is an iron-regulated hemin binding protein involved in the utilization of heme iron in Corynebacterium diphtheriae J Bacteriol 191 2009 2638 2648 19201805 148 Allen CE Schmitt MP Novel hemin binding domains in the Corynebacterium diphtheriae HtaA protein interact with hemoglobin and are critical for heme iron utilization by HtaA J Bacteriol 193 2011 5374 5385 21803991 149 Draganova EB Akbas N Adrian SA Lukat-Rodgers GS Collins DP Dawson JH Heme binding by Corynebacterium diphtheriae HmuT: function and heme environment Biochemistry 54 2015 6598 6609 26478504 150 Kunkle CA Schmitt MP Analysis of a DtxR-regulated iron transport and siderophore biosynthesis gene cluster in Corynebacterium diphtheriae J Bacteriol 187 2005 422 433 15629913 151 Kunkle CA Schmitt MP Analysis of the Corynebacterium diphtheriae DtxR regulon: identification of a putative siderophore synthesis and transport system that is similar to the Yersinia high-pathogenicity island-encoded yersiniabactin synthesis and uptake system J Bacteriol 185 2003 6826 6840 14617647 152 Allen CE Burgos JM Schmitt MP Analysis of novel iron-regulated, surface-anchored hemin-binding proteins in Corynebacterium diphtheriae J Bacteriol 195 2013 2852 2863 23585541 153 Allen CE Schmitt MP Utilization of host iron sources by Corynebacterium diphtheriae: multiple hemoglobin-binding proteins are essential for the use of iron from the hemoglobin– haptoglobin complex J Bacteriol 197 2015 553 562 25404705 154 Noinaj N Guillier M Barnard TJ Buchanan SK TonB-dependent transporters: regulation, structure, and function Annu Rev Microbiol 64 2010 43 60 20420522 155 Letoffe S Nato F Goldberg ME Wandersman C Interactions of HasA, a bacterial haemophore, with haemoglobin and with its outer membrane receptor HasR Mol Microbiol 33 1999 546 555 10417645 156 Letoffe S Deniau C Wolff N Dassa E Delepelaire P Lecroisey A Haemophore-mediated bacterial haem transport: evidence for a common or overlapping site for haem-free and haem-loaded haemophore on its specific outer membrane receptor Mol Microbiol 41 2001 439 450 11489129 157 Smith AD Wilks A Differential contributions of the outer membrane receptors PhuR and HasR to heme acquisition in Pseudomonas aeruginosa J Biol Chem 290 2015 7756 7766 25616666 158 Oglesby-Sherrouse AG Vasil ML Characterization of a heme-regulated non-coding RNA encoded by the prrF locus of Pseudomonas aeruginosa PLoS One 5 2010 e9930 20386693 159 Reinhart AA Powell DA Nguyen AT O'Neill M Djapgne L Wilks A The prrF-encoded small regulatory RNAs are required for iron homeostasis and virulence of Pseudomonas aeruginosa Infect Immun 83 2015 863 875 25510881 160 Lansky IB Lukat-Rodgers GS Block D Rodgers KR Ratliff M Wilks A The cytoplasmic heme-binding protein (PhuS) from the heme uptake system of Pseudomonas aeruginosa is an intracellular heme-trafficking protein to the delta-regioselective heme oxygenase J Biol Chem 281 2006 13652 13662 16533806 161 Block DR Lukat-Rodgers GS Rodgers KR Wilks A Bhakta MN Lansky IB Identification of two heme-binding sites in the cytoplasmic heme-trafficking protein PhuS from Pseudomonas aeruginosa and their relevance to function Biochemistry 46 2007 14391 14402 18020455 162 Tripathi S O'Neill MJ Wilks A Poulos TL Crystal structure of the Pseudomonas aeruginosa cytoplasmic heme binding protein, Apo-PhuS J Inorg Biochem 128 2013 131 136 23973453 163 O'Neill MJ Bhakta MN Fleming KG Wilks A Induced fit on heme binding to the Pseudomonas aeruginosa cytoplasmic protein (PhuS) drives interaction with heme oxygenase (HemO) Proc Natl Acad Sci USA 109 2012 5639 5644 22451925 164 Lee MJ Schep D McLaughlin B Kaufmann M Jia Z Structural analysis and identification of PhuS as a heme-degrading enzyme from Pseudomonas aeruginosa J Mol Biol 426 2014 1936 1946 24560694 165 O'Neill MJ Wilks A The P. aeruginosa heme binding protein PhuS is a heme oxygenase titratable regulator of heme uptake ACS Chem Biol 8 2013 1794 1802 23947366 166 Marvig RL Damkiaer S Khademi SM Markussen TM Molin S Jelsbak L Within-host evolution of Pseudomonas aeruginosa reveals adaptation toward iron acquisition from hemoglobin mBio 5 2014 e00966 14 24803516 167 Granick S Gilder H The porphyrin requirements of Haemophilus influenzae and some functions of the vinyl and propionic acid side chains of heme J Gen Physiol 30 1946 1 13 168 Stull TL Protein sources of heme for Haemophilus influenzae Infect Immun 55 1987 148 153 3025098 169 Cope LD Thomas SE Hrkal Z Hansen EJ Binding of heme-hemopexin complexes by soluble HxuA protein allows utilization of this complexed heme by Haemophilus influenzae Infect Immun 66 1998 4511 4516 9712810 170 Cope LD Yogev R Muller-Eberhard U Hansen EJ A gene cluster involved in the utilization of both free heme and heme: hemopexin by Haemophilus influenzae type b J Bacteriol 177 1995 2644 2653 7751272 171 Baelen S Dewitte F Clantin B Villeret V Structure of the secretion domain of HxuA from Haemophilus influenzae Acta Crystallogr Sect F: Struct Biol Cryst Commun 69 2013 1322 1327 172 Fournier C Smith A Delepelaire P Haem release from haemopexin by HxuA allows Haemophilus influenzae to escape host nutritional immunity Mol Microbiol 80 2011 133 148 21276097 173 Cope LD Love RP Guinn SE Gilep A Usanov S Estabrook RW Involvement of HxuC outer membrane protein in utilization of hemoglobin by Haemophilus influenzae Infect Immun 69 2001 2353 2363 11254593 174 Morton DJ Seale TW Madore LL VanWagoner TM Whitby PW Stull TL The haem-haemopexin utilization gene cluster (hxuCBA) as a virulence factor of Haemophilus influenzae Microbiology 153 2007 215 224 17185550 175 Morton DJ Madore LL Smith A Vanwagoner TM Seale TW Whitby PW The heme-binding lipoprotein (HbpA) of Haemophilus influenzae: role in heme utilization FEMS Microbiol Lett 253 2005 193 199 16289530 176 Morton DJ Seale TW Bakaletz LO Jurcisek JA Smith A VanWagoner TM The heme-binding protein (HbpA) of Haemophilus influenzae as a virulence determinant Int J Med Microbiol 299 2009 479 488 19451029 177 Morton DJ Van Wagoner TM Seale TW Whitby PW Stull TL Utilization of myoglobin as a heme source by Haemophilus influenzae requires binding of myoglobin to haptoglobin FEMS Microbiol Lett 258 2006 235 240 16640579 178 Morton DJ VanWagoner TM Seale TW Whitby PW Stull TL Differential utilization by Haemophilus influenzae of haemoglobin complexed to the three human haptoglobin phenotypes FEMS Immunol Med Microbiol 46 2006 426 432 16553817 179 Morton DJ Smith A Ren Z Madore LL VanWagoner TM Seale TW Identification of a haem-utilization protein (Hup) in Haemophilus influenzae Microbiology 150 2004 3923 3933 15583146 180 Morton DJ Seale TW Vanwagoner TM Whitby PW Stull TL The dppBCDF gene cluster of Haemophilus influenzae: role in heme utilization BMC Res Notes 2 2009 166 19703293 181 Morton DJ Smith A VanWagoner TM Seale TW Whitby PW Stull TL Lipoprotein e (P4) of Haemophilus influenzae: role in heme utilization and pathogenesis Microbes Infect 9 2007 932 939 17548224 182 LaCross NC Marrs CF Gilsdorf JR Otitis media associated polymorphisms in the hemin receptor HemR of nontypeable Haemophilus influenzae Infect Genet Evol 26 2014 47 57 24820341 183 Vogel AR Szelestey BR Raffel FK Sharpe SW Gearinger RL Justice SS SapF-mediated heme-iron utilization enhances persistence and coordinates biofilm architecture of Haemophilus Front Cell Infect Microbiol 2 2012 42 22919633 184 Whitby PW VanWagoner TM Seale TW Morton DJ Stull TL Comparison of transcription of the Haemophilus influenzae iron/heme modulon genes in vitro and in vivo in the chinchilla middle ear BMC Genomics 14 2013 925 24373462 185 Whitby PW Seale TW VanWagoner TM Morton DJ Stull TL The iron/heme regulated genes of Haemophilus influenzae: comparative transcriptional profiling as a tool to define the species core modulon BMC Genomics 10 2009 6 19128474 186 Xie J Juliao PC Gilsdorf JR Ghosh D Patel M Marrs CF Identification of new genetic regions more prevalent in nontypeable Haemophilus influenzae otitis media strains than in throat strains J Clin Microbiol 44 2006 4316 4325 17005745 187 Hariadi NI Zhang L Patel M Sandstedt SA Davis GS Marrs CF comparative profile of heme acquisition genes in disease-causing and colonizing nontypeable Haemophilus influenzae and Haemophilus haemolyticus J Clin Microbiol 53 2015 2132 2137 25903577 188 Seale TW Morton DJ Whitby PW Wolf R Kosanke SD VanWagoner TM Complex role of hemoglobin and hemoglobin–haptoglobin binding proteins in Haemophilus influenzae virulence in the infant rat model of invasive infection Infect Immun 74 2006 6213 6225 16966415 189 Morton DJ Bakaletz LO Jurcisek JA VanWagoner TM Seale TW Whitby PW Reduced severity of middle ear infection caused by nontypeable Haemophilus influenzae lacking the hemoglobin/hemoglobin–haptoglobin binding proteins (Hgp) in a chinchilla model of otitis media Microb Pathog 36 2004 25 33 14643637 190 Lewis LA Gray E Wang YP Roe BA Dyer DW Molecular characterization of hpuAB, the haemoglobin–hapto-globin-utilization operon of Neisseria meningitidis Mol Microbiol 23 1997 737 749 9157245 191 Lewis LA Dyer DW Identification of an iron-regulated outer membrane protein of Neisseria meningitidis involved in the utilization of hemoglobin complexed to haptoglobin J Bacteriol 177 1995 1299 1306 7868605 192 Zhu W Wilks A Stojiljkovic I Degradation of heme in gram-negative bacteria: the product of the hemO gene of Neisseriae is a heme oxygenase J Bacteriol 182 2000 6783 6790 11073924 193 Zhu W Hunt DJ Richardson AR Stojiljkovic I Use of heme compounds as iron sources by pathogenic neisseriae requires the product of the hemO gene J Bacteriol 182 2000 439 447 10629191 194 Lewis LA Sung MH Gipson M Hartman K Dyer DW Transport of intact porphyrin by HpuAB, the hemoglobin– haptoglobin utilization system of Neisseria meningitidis J Bacteriol 180 1998 6043 6047 9811666 195 Kumar P Sannigrahi S Tzeng YL The Neisseria meningitidis ZnuD zinc receptor contributes to interactions with epithelial cells and supports heme utilization when expressed in Escherichia coli Infect Immun 80 2012 657 667 22083713 196 Chen CJ Elkins C Sparling PF Phase variation of hemoglobin utilization in Neisseria gonorrhoeae Infect Immun 66 1998 987 993 9488386 197 Rohde KH Dyer DW Analysis of haptoglobin and hemoglobin –haptoglobin interactions with the Neisseria meningitidis TonB-dependent receptor HpuAB by flow cytometry Infect Immun 72 2004 2494 2506 15102756 198 Rohde KH Gillaspy AF Hatfield MD Lewis LA Dyer DW Interactions of haemoglobin with the Neisseria meningitidis receptor HpuAB: the role of TonB and an intact proton motive force Mol Microbiol 43 2002 335 354 11985713 199 Wong CT Xu Y Gupta A Garnett JA Matthews SJ Hare SA Structural analysis of haemoglobin binding by HpuA from the Neisseriaceae family Nat Commun 6 2015 10172 26671256 200 Stojiljkovic I Hwa V de Saint Martin L O'Gaora P Nassif X Heffron F The Neisseria meningitidis haemoglobin receptor: its role in iron utilization and virulence Mol Microbiol 15 1995 531 541 7783623 201 Stojiljkovic I Larson J Hwa V Anic S So M HmbR outer membrane receptors of pathogenic Neisseria spp.: iron-regulated, hemoglobin-binding proteins with a high level of primary structure conservation J Bacteriology 178 1996 4670 4678 202 Richardson AR Stojiljkovic I HmbR, a hemoglobin-binding outer membrane protein of Neisseria meningitidis, undergoes phase variation J Bacteriol 181 1999 2067 2074 10094683 203 Mokry DZ Nadia-Albete A Johnson MK Lukat-Rodgers GS Rodgers KR Lanzilotta WN Spectroscopic evidence for a 5-coordinate oxygenic ligated high spin ferric heme moiety in the Neisseria meningitidis hemoglobin binding receptor Biochim Biophys Acta 2014 1840 3058 3066 24968987 204 Sebastian S Agarwal S Murphy JR Genco CA The gonococcal Fur regulon: identification of additional genes involved in major catabolic, recombination, and secretory pathways J Bacteriol 184 2002 3965 3974 12081969 205 Zhao S Montanez GE Kumar P Sannigrahi S Tzeng YL Regulatory role of the MisR/S two-component system in hemoglobin utilization in Neisseria meningitidis Infect Immun 78 2010 1109 1122 20008531 206 Harrison OB Evans NJ Blair JM Grimes HS Tinsley CR Nassif X Epidemiological evidence for the role of the hemoglobin receptor, MmbR, in meningococcal virulence J Infect Dis 200 2009 94 98 19476432 207 Harrison OB Bennett JS Derrick JP Maiden MC Bayliss CD Distribution and diversity of the haemoglobin-haptoglobin iron-acquisition systems in pathogenic and non-pathogenic Neisseria Microbiology 159 2013 1920 1930 23813677 208 Lucidarme J Findlow J Chan H Feavers IM Gray SJ Kaczmarski EB The distribution and ‘in vivo’ phase variation status of haemoglobin receptors in invasive meningococcal serogroup B disease: genotypic and phenotypic analysis PLoS One 8 2013 e76932 24098814 209 Fowler VG Jr Proctor RA Where does a Staphylococcus aureus vaccine stand? Clin Microbiol Infect 20 5 2014 66 75 24476315 210 Kuklin NA Clark DJ Secore S Cook J Cope LD McNeely T A novel Staphylococcus aureus vaccine: iron surface determinant B induces rapid antibody responses in rhesus macaques and specific increased survival in a murine S. aureus sepsis model Infect Immun 74 2006 2215 2223 16552052 211 McNeely TB Shah NA Fridman A Joshi A Hartzel JS Keshari RS Mortality among recipients of the Merck V710 Staphylococcus aureus vaccine after postoperative S. aureus infections: an analysis of possible contributing host factors Hum Vaccin Immunother 10 2014 3513 3516 25483690 212 Ebert T Smith S Pancari G Clark D Hampton R Secore S A fully human monoclonal antibody to Staphylococcus aureus iron regulated surface determinant B (IsdB) with functional activity in vitro and in vivo Hum Antibodies 19 2010 113 128 21178283 213 Pancari G Fan H Smith S Joshi A Haimbach R Clark D Characterization of the mechanism of protection mediated by CS-D7, a monoclonal antibody to Staphylococcus aureus iron regulated surface determinant B (IsdB) Front Cell Infect Microbiol 2 2012 36 22919628 214 Parish T Schaeffer M Roberts G Duncan K HemZ is essential for heme biosynthesis in Mycobacterium tuberculosis Tuberculosis (Edinb) 85 2005 197 204 15850757 215 Tullius MV Harmston CA Owens CP Chim N Morse RP McMath LM Discovery and characterization of a unique mycobacterial heme acquisition system Proc Natl Acad Sci USA 108 2011 5051 5056 21383189 216 Owens CP Chim N Goulding CW Insights on how the Mycobacterium tuberculosis heme uptake pathway can be used as a drug target Future Med Chem 5 2013 1391 1403 23919550 217 Barber MF Elde NC Nutritional immunity. Escape from bacterial iron piracy through rapid evolution of transferrin Science 346 2014 1362 1366 25504720 218 Barber MF Elde NC Buried treasure: evolutionary perspectives on microbial iron piracy Trends Genet 31 2015 627 636 26431675 219 Chou AC Fitch CD Mechanism of hemolysis induced by ferriprotoporphyrin IX J Clin Invest 68 1981 672 677 7276166 220 Schmitt TH Frezzatti WA Jr Schreier S Hemin-induced lipid membrane disorder and increased permeability: a molecular model for the mechanism of cell lysis Arch Biochem Biophys 307 1993 96 103 8239671 221 Van Heyningen WE Inhibition of aerobic sporing bacilli by haematin Nature 162 1948 114 222 Stauff DL Torres VJ Skaar EP Signaling and DNA-binding activities of the Staphylococcus aureus HssR–HssS two-component system required for heme sensing J Biol Chem 282 2007 26111 26121 17635909 223 Fernandez A Lechardeur D Derre-Bobillot A Couve E Gaudu P Gruss A Two coregulated efflux transporters modulate intracellular heme and protoporphyrin IX availability in Streptococcus agalactiae PLoS Pathog 6 2010 e1000860 20421944 224 Rasmussen AW Alexander HL Perkins-Balding D Shafer WM Stojiljkovic I Resistance of Neisseria meningitidis to the toxic effects of heme iron and other hydrophobic agents requires expression of ght J Bacteriol 187 2005 5214 5223 16030215 225 Wyckoff EE Lopreato GF Tipton KA Payne SM Shigella dysenteriae ShuS promotes utilization of heme as an iron source and protects against heme toxicity J Bacteriol 187 2005 5658 5664 16077111 226 Imlay JA Chin SM Linn S Toxic DNA damage by hydrogen peroxide through the Fenton reaction in vivo and in vitro Science 240 1988 640 642 2834821 227 Imlay JA Linn S DNA damage and oxygen radical toxicity Science 240 1988 1302 1309 3287616 228 Wakeman CA Hammer ND Stauff DL Attia AS Anzaldi LL Dikalov SI Menaquinone biosynthesis potentiates haem toxicity in Staphylococcus aureus Mol Microbiol 86 2012 1376 1392 23043465 229 Wakeman CA Stauff DL Zhang Y Skaar EP Differential activation of Staphylococcus aureus heme detoxification machinery by heme analogues J Bacteriol 196 2014 1335 1342 24443529 230 Stojiljkovic I Kumar V Srinivasan N Non-iron metalloporphyrins: potent antibacterial compounds that exploit haem/Hb uptake systems of pathogenic bacteria Mol Microbiol 31 1999 429 442 10027961 231 Nitzan Y Wexler HM Finegold SM Inactivation of anaerobic bacteria by various photosensitized porphyrins or by hemin Curr Microbiol 29 1994 125 131 7765091 232 Nir U Ladan H Malik Z Nitzan Y In vivo effects of porphyrins on bacterial DNA J Photochem Photobiol B 11 1991 295 306 1816365 233 Putker F Grutsch A Tommassen J Bos MP Ght protein of Neisseria meningitidis is involved in the regulation of lipopoly-saccharide biosynthesis J Bacteriol 196 2014 780 789 24296673 234 Stauff DL Bagaley D Torres VJ Joyce R Anderson KL Kuechenmeister L Staphylococcus aureus HrtA is an ATPase required for protection against heme toxicity and prevention of a transcriptional heme stress response J Bacteriol 190 2008 3588 3596 18326576 235 Torres VJ Stauff DL Pishchany G Bezbradica JS Gordy LE Iturregui J A Staphylococcus aureus regulatory system that responds to host heme and modulates virulence Cell Host Microbe 1 2007 109 119 18005689 236 Lechardeur D Cesselin B Liebl U Vos MH Fernandez A Brun C Discovery of intracellular heme-binding protein HrtR, which controls heme efflux by the conserved HrtB–HrtA transporter in Lactococcus lactis J Biol Chem 287 2012 4752 4758 22084241 237 Joubert L Derre-Bobillot A Gaudu P Gruss A Lechardeur D HrtBA and menaquinones control haem homeostasis in Lactococcus lactis Mol Microbiol 93 2014 823 833 25040434 238 Mike LA Choby JE Brinkman PR Olive LQ Dutter BF Ivan SJ Two-component system cross-regulation integrates Bacillus anthracis response to heme and cell envelope stress PLoS Pathog 10 2014 e1004044 24675902 239 Hagman KE Pan W Spratt BG Balthazar JT Judd RC Shafer WM Resistance of Neisseria gonorrhoeae to antimicrobial hydrophobic agents is modulated by the mtrRCDE efflux system Microbiology 141 Pt 3 1995 611 622 7711899 240 Stojiljkovic I Hantke K Transport of haemin across the cytoplasmic membrane through a haemin-specific periplasmic binding-protein-dependent transport system in Yersinia enter-ocolitica Mol Microbiol 13 1994 719 732 7997183 241 Suits MD Pal GP Nakatsu K Matte A Cygler M Jia Z Identification of an Escherichia coli O157:H7 heme oxygenase with tandem functional repeats Proc Natl Acad Sci USA 102 2005 16955 16960 16275907 242 Ouellet YH Ndiaye CT Gagne SM Sebilo A Suits MD Jubinville E An alternative reaction for heme degradation catalyzed by the Escherichia coli O157:H7 ChuS protein: release of hematinic acid, tripyrrole and Fe(III) J Inorg Biochem 154 2016 103 113 26598215 243 Kaur AP Lansky IB Wilks A The role of the cytoplasmic heme-binding protein (PhuS) of Pseudomonas aeruginosa in intracellular heme trafficking and iron homeostasis J Biol Chem 284 2009 56 66 18990702 244 Kaur AP Wilks A Heme inhibits the DNA binding properties of the cytoplasmic heme binding protein of Shigella dysenteriae (ShuS) Biochemistry 46 2007 2994 3000 17323920 245 Al Jubair T Singh B Fleury C Blom AM Morgelin M Thunnissen MM Haemophilus influenzae stores and distributes hemin by using protein E Int J Med Microbiol 304 2014 662 668 24863527 246 Negari S Sulpher J Pacello F Ingrey K Battistoni A Lee BC A role for Haemophilus ducreyi Cu, ZnSOD in resistance to heme toxicity Biometals 21 2008 249 258 17704897

PMC005xxxxxx/PMC5126229.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 7602745 7988 Vet Parasitol Vet. Parasitol. Veterinary parasitology 0304-4017 1873-2550 27884445 5126229 10.1016/j.vetpar.2016.10.022 NIHMS825869 Article Taenia hydatigena in pigs in Burkina Faso: a cross-sectional abattoir study Dermauw Veronique 1 Ganaba Rasmané 2 Cissé Assana 3 Ouedraogo Boubacar 2 Millogo Athanase 4 Tarnagda Zékiba 3 Van Hul Anke 1 Gabriël Sarah 15 Carabin Hélène 6 Dorny Pierre 1 1 Institute of Tropical Medicine, Nationalestraat 155, B-2000, Antwerp, Belgium 2 AFRICSanté, 01 BP 298 Bobo Dioulasso 01, Burkina Faso 3 Institut de Recherche en Sciences de la Santé (IRSS), Avenue de la Liberté, BP 545, Bobo-Dioulasso, Burkina Faso 4 University of Ouagadougou, Ouagadougou, Burkina Faso 5 Ghent University, Faculty of Veterinary Medicine, Salisburylaan 133, 9820 Merelbeke, Belgium 6 University of Oklahoma Health Sciences Center, Oklahoma City, OK, United States sgabriel@itg.be, avanhul@itg.be, pdorny@itg.be, rganaba@hotmail.com, assanacisse@yahoo.fr, zekiba@hotmail.com, athanase.millogo@gmail.com, sarah.gabriel@ugent.be, helene-carabin@ouhsc.edu Corresponding author: Veronique Dermauw, telephone: +3232476447, fax: + 3232476268, vdermauw@itg.be 29 10 2016 24 10 2016 30 10 2016 30 10 2017 230 913 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. Taenia hydatigena is a non-zoonotic cestode that has canines as definitive hosts and ruminants and pigs as intermediate hosts. In pigs, its presence causes cross-reactivity in serological testing for Taenia solium cysticercosis. Therefore, knowledge on the occurrence of T. hydatigena is paramount for validly estimating the seroprevalence of T. solium cysticercosis in pigs. In a cross-sectional abattoir study, we estimated the prevalence of T. hydatigena in pigs slaughtered in Koudougou, Burkina Faso. Carcasses of 452 pigs were examined by investigators for perceived and suspected T. hydatigena cysticercus lesions in the abdominal cavity or on the surface of abdominal organs. Routine meat inspection was performed by local inspectors to identify T. solium cysticerci. All lesions were subjected to PCR-RFLP analysis in order to differentiate Taenia spp. Additionally, individual blood samples were examined for the presence of circulating cysticercus antigens using the B158/B60 Ag-ELISA. Perceived T. hydatigena cysticerci were found in 13 pigs, whereas meat inspectors found seven carcasses infected with T. solium cysticerci. All were confirmed by molecular analysis. Of pigs with other suspected lesions, mostly located in the liver, 27 and six were found to harbour T. hydatigena and T. solium cysticerci, respectively. Overall, 8.8% of pigs (40/452) were found infected with T. hydatigena and 2.9% (13/452) with T. solium. Of these positive pigs, one was found infected with both Taenia spp. (0.2%, 1/452). Blood samples of 48.5% of pigs (219/452) were positive in the Ag-ELISA. Pigs with confirmed cysts of T. hydatigena and T. solium had a positive Ag-ELISA result in 57.5% (23/40) and 61.5% (8/13) of cases, respectively. The observed T. hydatigena prevalence in this study is relatively high in comparison to other studies in Africa. Estimates of the occurrence of active porcine T. solium infection using the B158/B60 Ag-ELISA should therefore be adjusted for the presence of T. hydatigena. The low level of T. solium infection detected upon meat inspection in this study is likely an underestimation of the true prevalence since routine meat inspection shows poor sensitivity and pigs perceived to be infected based on tongue palpation are rarely sent to official abattoirs. Taenia hydatigena Taenia solium cysticercosis pigs Burkina Faso 1. Introduction Pigs may act as the intermediate hosts of Taenia hydatigena, a non-zoonotic cestode that has canines as definitive hosts. Infection with this parasite rarely leads to clinical signs in pigs, but is causing cross-reactions in serological tests developed for the detection of Taenia solium cysticercosis (Dorny et al., 2004a). Worldwide, the occurrence of T. hydatigena in pigs varies widely. In Asia, studies have reported prevalences as high as 25.7% in China (Yin et al., 2013), and 22.4% in Laos (Conlan et al., 2012). In contrast, the reported prevalence in Africa seems to be much lower, ranging from 1.4% to 6.7% (Ngowi et al., 2004; Permin et al., 1999). Pigs are also the intermediate hosts of T. solium that has humans as definitive hosts. Whereas humans are not susceptible to infection with T. hydatigena, they can acquire T. solium through the consumption of undercooked infected pork, leading to the development of an intestinal adult tapeworm infection (taeniosis). In addition, ingestion of tapeworm eggs by humans and pigs can lead to cysticercosis. In humans, cysticerci can establish in the central nervous system causing a condition called neurocysticercosis (NCC). NCC can be accompanied by severe signs and symptoms such as epilepsy and chronic headaches (Carabin et al., 2011). Considering its zoonotic character, it is important to monitor the presence of T. solium in humans and pigs as part of disease control and surveillance (WHO, 2016). In pigs, various diagnostic tools have been described to detect T. solium cysticercosis (Lightowlers et al., 2016). Tongue palpation and carcass inspection are widely used for ante and post-mortem diagnosis, respectively. However, these tools show poor sensitivities of 16.1% (95% confidence interval (95%CI): 5-34) and 38.7% (95%CI: 22-58), respectively (Dorny et al., 2004b). Serological assays, whether aimed at detection of antigens (Ag) or antibodies (Ab), provide a better alternative as they show higher test performances. The B158/B60 Ag-ELISA is a serological assay that detects circulating antigens of Taenia spp., of which the presence is an indication of infection with viable cysticerci. The test has a reported sensitivity and specificity to detect both viable and dead T. solium cysticerci in pigs of 64.5% (95% CI: 45-81) and 91.2% (95%CI: 76-98), respectively (Dorny et al., 2004b). However, cross reactivity exists in this test between T. solium and other Taenia spp., such as T. hydatigena, due to its genus but not species-specific nature (Dorny et al., 2004a). In Burkina Faso, a large on-going community randomized-control trial (EFECAB) aims to estimate the impact of an educational intervention on human and porcine cysticercosis (Carabin et al., 2015). In this trial, the B158/B60 Ag-ELISA is used as a diagnostic tool to estimate levels of porcine cysticercosis. Currently, no studies have investigated the presence of T. hydatigena in pigs in Burkina Faso, although knowledge on the occurrence of this parasite is essential to interpret levels of T. solium cysticercosis estimated through the B158/B60 Ag-ELISA. Therefore, the aim of the present study was to estimate the prevalence of T. hydatigena among pigs slaughtered at the Koudougou abattoir, Burkina Faso. 2. Materials & methods 2.1. Study area & sampling strategy A cross-sectional abattoir study was conducted in Koudougou, a small city located about 100 km west of the capital Ouagadougou. Pigs slaughtered in the Koudougou abattoir typically originate from Koudougou and nearby villages. In these areas, traditional pig management is practiced, meaning pigs are left roaming to scavenge for their food during the dry season. In 2010, the pig population in the province of Boulkiemdé, where Koudougou is located, was estimated at 191,438, the largest in the country (Ministère des Ressources Animales, 2010). Yet, not all households in the region consume pork or own a pig (Ganaba et al., 2011). The planned sample size for the current study was 384, assuming a large population size with unknown T. hydatigena prevalence and setting the required confidence level at 95% with ±5% precision (Cochran, 1977). Ultimately, 452 pigs were randomly sampled at the Koudougou abattoir in March and April 2015. Upon slaughter, a jugular blood sample was collected in an additive-free blood tube from each pig. The age, sex and origin of sampled pigs were noted. Blood samples were allowed to clot, after which serum was obtained. Serum aliquots were stored at −20°C until further analysis. One researcher subjected pigs to detailed post-mortem inspection for the presence of T. hydatigena cysticerci in the abdominal cavity and on the surface of the liver and other abdominal organs. Samples were categorized as “perceived T. hydatigena cysts” (i.e., large loose-hanging cysts attached to an organ surface) or “suspected lesions” (i.e., other smaller lesions). Two local meat inspectors conducted their routine examination according to national regulations (Conseils des Ministres, 1989). The examination involved detailed inspection of all abdominal and thoracic organs (including incision of the heart and lymph nodes), as well as inspection of the directly visible muscular surfaces, in particular of the thigh, psoas, intercostal muscles, abdominal wall, pillars of the diaphragm, heart, tongue and larynx (Conseil des Ministres, 1989). Lesions deemed positive for T. solium by meat inspectors were defined as “lesions identified as T. solium upon meat inspection”. All lesions were collected in 70% ethanol and kept at ambient temperature until further analysis. 2.2. Laboratory analyses Molecular methods were used to identify and differentiate Taenia spp. in sampled lesions. First, genomic DNA was extracted using the DNeasyBlood and Tissue Extraction Kit according to the manufacturers’ instructions (QIAGEN, Hilden, Germany). PCR was used to amplify a mitochondrial 12s rDNA gene fragment with the primer set ITM TnR-TaenF (Rodriguez-Hidalgo et al., 2002). Afterwards, restriction fragment length polymorphism (RFLP) was used to differentiate the Taenia spp. The restriction enzymes DdeI and HinfI were used to identify T. solium (Rodriguez-Hidalgo et al., 2002) and HpaI was used to identify T. hydatigena (Devleesschauwer et al., 2013). In case the observed pattern was equivocal, sequencing was performed on the PCR product to obtain full confirmation of the species in sampled lesions. Sequencing was performed at the VIB Genetic Service Facility (University of Antwerp, Belgium). BioEdit was used to edit and align obtained sequences (Hall, 1999) and BLAST was performed on NCBI. A monoclonal antibody-based B158/B60 Ag-ELISA (Brandt et al., 1992; Dorny et al., 2004b) was used to detect the presence of circulating cysticercus antigens in serum. The Ag-ELISA was performed as described by Dorny et al. (2000), with slight modifications described by Dorny et al (2004b). Briefly, the monoclonal antibody B158C11A10, diluted at 5 μg/ml was coated on the wells of Nunc Maxisorp microtitre plates. After washing and blocking of the wells, serum samples pre-treated with 5% trichloro acetic acid were added. Next, after washing, biotinylated monoclonal antibody B60H8A4 diluted at 1.25 μg/ml in phosphate buffered saline-Tween 20 (PBS-T20) + 1% new born calf serum (NBCS) was used to detect the circulating cysticercus antigens. Streptavidin-horseradish peroxidase (Jackson Immunoresearch Lab, Inc.) diluted at 1/10,000 in PBS-T20+1% NBCS was used as the conjugate. Incubations were done at 37°C on a shaking plate and followed for each step by washing five times with PBS-T20 using an automated ELISA washer. Finally, ortho-phenylenediamine (Dako, Glostrup, Denmark) was used as the substrate. After 15 minutes of incubation in the dark, the reaction was stopped with 50 μl 4 N H2SO4. Optical densities in the wells were read using a spectrophotometer (Titertek Multiskan EIA reader, Vienna, Virginia, USA) at 492 nm and a reference at 655 nm. Serum samples from two pigs confirmed to be heavily infected were used as positive controls. The cut-off was calculated for each plate based on the optical density (OD) of eight negative reference sera based on a variation of the student t-test distribution at a probability of p = 0.001 (Sokal and Rohlf, 1991). The OD of each sample was divided by the cut-off to obtain a ratio. Ag-ELISA ratios of positive pigs were classified as follows: low: 1 ≤ ratio < 2; medium: 2 ≤ ratio < 5; high: 5 ≤ ratio < 10; very high: ratio ≥ 10 (Praet et al., 2010). 2.3. Statistical analyses All collected data were entered into Excel (Microsoft Office Excel 2010). Descriptive statistical analyses (i.e., calculation of percentages) were conducted using R, version 3.2.4 (R Core Team, 2016). 3. Results A total of 452 pigs were inspected at the Koudougou abattoir, Burkina Faso. The majority of these pigs originated from Koudougou (311/452, 68.8%), while the others came from 12 surrounding villages. Half of the sampled pigs were female (225/452, 49.8%) and most were younger than two years (405/452, 89.6%). A total of 58 suspected lesions were collected among 57 of the 452 inspected pigs (Table 1). Thirteen of these lesions were perceived to be cysticerci of T. hydatigena, whereas local meat inspectors identified seven carcasses as being infected with T. solium cysticerci. Suspected lesions without clear diagnosis were found in livers of an additional 37 pigs (of which one pig also had another cyst perceived to be T. hydatigena). One pig had a suspected lesion in the spleen. Molecular analysis of sampled lesions confirmed the diagnosis of all perceived T. hydatigena cysts (n = 13) as well as the identification of T. solium positive carcasses by local meat inspectors (n = 7). Suspected liver lesions of 27 pigs were identified as T. hydatigena, and in six pigs, liver lesions were identified as T. solium. No Taenia spp. could be identified in four pigs with suspected liver lesions or in the pig with the suspected spleen sample. Overall, based on molecular analysis of sampled lesions, 8.8% of pigs (40/452) were found infected with T. hydatigena and 2.9% (13/452) with T. solium. One of these positive pigs was found infected with both Taenia spp. (overall 0.2%, 1/452). The Ag-ELISA identified 48.5% of pigs (219/452) as positive for circulating cysticercus Taenia spp. antigens (Table 2). Cysticerci of T. hydatigena or T. solium were identified in 13.7% of sero-positive pigs (30/219), with T. hydatigena being identified in 10.5% (23/219). Conversely, the Ag-ELISA was negative in 42.3% (22/52) of pigs found to harbour T. solium and/or T. hydatigena cysts. Overall, T. hydatigena and T. solium confirmed cases were Ag-ELISA positive in 23 out of 40 (57.5%) and eight out of 13 (61.5%) pigs, respectively. Approximately half of T. hydatigena positive pigs (19/40) had an Ag-ELISA ratio below 2 (Table 3), whereas almost all of the remaining pigs had a very high ratio (≥10) (19/40). All pigs with samples perceived to be obvious T. hydatigena cysts (n = 13) had very high ratios. 4. Discussion This is the first study reporting the prevalence of T. hydatigena in pigs in Burkina Faso and the first study on T. hydatigena prevalence in pigs to include molecular confirmation. The prevalence of T. hydatigena found in this study is relatively high (8.8%) in comparison to other African studies, where the reported prevalence in pigs ranged between 1.4% (Tanzania: Ngowi et al., 2004) and 6.6 to 6.7% (6.6% (16/243) in Tanzania (Braae et al., 2015), 6.7% (4/60) in Ghana (Permin et al., 1999)) (reviewed by Nguyen et al., 2016). Pigs in those studies were sampled in small abattoirs, in villages or in both, and prevalence estimates were based on post-mortem carcass inspection only. The proportion of T. solium infected carcasses identified through meat inspection (1.5%, 7/452) in the current study was higher than the national prevalence of 0.6% reported by public authorities in 1997 (Coulibaly and Yameogo, 2000). The reported proportions of T. solium infected carcasses identified upon meat inspection in other West African countries varied widely from 1.2% (Senegal) to 20.5% (Nigeria) (Zoli et al., 2003). Such a large range in estimates may be explained by variation in meat inspection regulations between countries and disparate adherence to inspection regulations, even within countries (WHO, 2015; Zoli et al., 2003). Furthermore, it is known that routine meat inspection is largely insufficient to detect all T. solium positive carcasses, especially in the case of light infections, due to its low sensitivity (Boa et al., 2002; Dorny et al., 2004b; Gonzalez et al., 1990). For instance, in a study in Zambia, 31 out of 65 pigs were found to harbour T. solium cysts upon complete dissection whereas only 12 were identified as positive upon meat inspection (Dorny et al., 2004b). Hence, the proportion of T. solium positive pigs found through meat inspection in Koudougou cannot be considered as a valid estimate of the prevalence of porcine cysticercosis in the region. In the current study, 48.5% of pigs slaughtered at the Koudougou abattoir were positive in the Ag-ELISA, indicating a high level of active T. solium infections among pigs slaughtered in an official abattoir in this area. In two villages in the centre of Burkina Faso, where pig breeding and pork consumption are very common, Ganaba et al. (2011) found that 32.5% and 39.6% of the pigs were positive in the Ag-ELISA, respectively. In the Democratic Republic of the Congo, 38.4% and 41.2% of pigs were found positive in the Ag-ELISA in markets and in villages, respectively (Praet et al., 2010). In free-ranging pigs in Zambia, a sero-prevalence of 23.3% was observed (Sikasunge et al., 2008), whereas at a local slaughter place, 57.1% of sampled pigs were found positive in the Ag-ELISA (Dorny et al., 2004b). The level of Ag-ELISA positive pigs was thus high in all these countries, suggesting a widespread presence of porcine cysticercosis. Tongue palpation is often performed by pig owners to screen their own animals (Gonzalez et al., 1990), as it will detect pigs heavily infected with T. solium (Dorny et al., 2004b). In case an infected pig is identified, it is commonly home slaughtered or sold at a local market, and thus de facto excluded from the official pig supply chain (Morales et al., 2006; Ngowi et al., 2004; Praet et al., 2010). If this practice is also common in Burkina Faso, the true proportion of T. solium positive pigs in the area may even be higher. A false negative Ag-ELISA result was observed in five out of 13 confirmed T. solium positive pigs in the current study. As the Ag-ELISA is designed to detect active infections, and has a reported sensitivity for overall infections (i.e., for both viable and dead cysticerci) of around 64.5% (Dorny et al., 2004b), our data confirm that it will even underestimate prevalence of T. solium at the slaughterhouse level. Estimates of T. solium prevalence based on Ag-ELISA results should therefore be adjusted for its test performance characteristics. On the other hand, T. hydatigena is known to cross-react in the Ag-ELISA (Dorny et al., 2004a). Considering the relatively high prevalence of T. hydatigena in this study, it is suggested that estimates of current T. solium infection prevalence based on Ag-ELISA results in this region should equally be adjusted for the presence of T. hydatigena. Almost 70% of the pigs sampled in this study originated from the town of Koudougou. Sampling of pigs in the abattoir of Koudougou was preferred over sampling in villages due to practical constraints. However, this raises the question of whether it is possible to generalise the estimated prevalence of T. hydatigena found in this study to all study villages included in EFECAB. Nonetheless, livestock management practices are not believed to vary much between Koudougou and villages in the area, suggesting that the estimated prevalence of T. hydatigena may actually be a reasonable estimate to adjust for the validity of the Ag-ELISA in this region. Conclusions In the present study, a high proportion of T. hydatigena infections was found among pigs slaughtered at the Koudougou abattoir in Burkina Faso, relative to what had been previously reported in other African countries. This finding has important consequences for the interpretation of the results of the B158/B60 Ag-ELISA when used in studies in the region. Hence, estimates of the occurrence of active porcine T. solium infection in this area should be adjusted for the presence of T. hydatigena. Nevertheless, the inherent underestimation of the T. solium prevalence in pigs by the B158/B60 should also be taken into account. Supplementary Material SM Acknowledgements The authors wish to thank the local meat inspectors and butchers for their assistance during pig sampling. Financial support for this work was provided by the National Institute of Neurological Disorders and Stroke (NINDS) and of the Fogarty International Center (FIC) of the National Institutes of Health (NIH) under the Brain in the Developing World: Research across the life span program, grant R01NS064901. Table 1 Number of pigs with suspected Taenia hydatigena and Taenia solium lesions sampled for molecular confirmation in a cross-sectional study conducted in March-April 2015 at the Koudougou abattoir, Burkina Faso (n = 452). Molecularly confirmed species Sample Total T. hydatigena T. solium None Perceived T. hydatigena cysts 13 13 0 0 Lesions identified as T. solium upon meat inspection 7 0 7 0 Suspected liver lesions 37 27 6 4 Suspected spleen lesions 1 0 0 1 Total 58 40 13 5 Table 2 Proportion of Ag-ELISA and Taenia hydatigena negative and positive pigs sampled at the Koudougou abattoir, Burkina Faso (n = 452). Ag-ELISA serum T. hydatigena n Percentage (%) Negative Negative 216 47.8 Negative Positive 17 3.8 Positive Negative 196 43.4 Positive Positive 23a 5.1 a This group included one pig with a confirmed T. solium cyst as well Table 3 Ag-ELISA ratio in Taenia hydatigena positive pigs sampled at the Koudougou abattoir, Burkina Faso (n = 40). Ratioa n Percentage (%) Low 19 47.5 Medium 1 2.5 High 1 2.5 Very high 19b,c 47.5 a Classes based on Ag-ELISA ratio: Low: 1 ≤ ratio < 2, Medium: 2 ≤ ratio < 5, , High: 5 ≤ ratio < 10, Very high: ratio ≥ 10 (Praet et al., 2010) b This class included all pigs with cysts perceived to be obvious T. hydatigena cysts (n = 13) upon post-mortem examination c This group included one pig with a confirmed T. solium cyst as well Highlights Taenia hydatigena, a non-zoonotic cestode, can have pigs as intermediate hosts It causes cross-reactivity in serological tests for active T. solium cysticercosis In Koudougou, Burkina Faso, a prevalence of 8.8% was found at the abattoir Estimates of active porcine cysticercosis should be adjusted for T. hydatigena This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. The authors declare that there are no conflicts of interest. References Boa ME Kassuku AA Willingham AL Keyyu JD Phiri IK Nansen P Distribution and density of cysticerci of Taenia solium by muscle groups and organs in naturally infected local finished pigs in Tanzania. Vet. Parasitol 2002 106 155 164 12031817 Braae UC Kabululu M Nørmark ME Nejsum P Ngowi HA Johansen MV Taenia hydatigena cysticercosis in slaughtered pigs, goats, and sheep in Tanzania. Trop. Anim. Health Prod 2015 47 1523 1530 26210397 Brandt JR Geerts S De Deken R Kumar V Ceulemans F Brijs L Falla N A monoclonal antibody-based ELISA for the detection of circulating excretory-secretory antigens in Taenia saginata cysticercosis. Int. J. Parasitol 1992 22 471 477 1644522 Carabin H Millogo A Cissé A Gabriël S Sahlu I Dorny P Bauer C Tarnagda Z Cowan LD Ganaba R Prevalence of and factors associated with human cysticercosis in 60 villages in three provinces of Burkina Faso. PLoS Negl. Trop. Dis 2015 9 1 20 Carabin H Ndimubanzi PC Budke CM Nguyen H Qian Y Cowan LD Stoner JA Rainwater E Dickey M Clinical manifestations associated with neurocysticercosis: A systematic review. PLoS Negl. Trop. Dis 2011 5 e1152 21629722 Cochran WG Sampling techniques 1977 John Wiley & Sons New York 3rd Edition. Conlan JV Vongxay K Khamlome B Dorny P Sripa B Elliot A Blacksell SD Fenwick S Thompson RCA A cross-sectional study of Taenia solium in a multiple taeniid-endemic region reveals competition may be protective. Am. J. Trop. Med. Hyg 2012 87 281 291 22855759 Conseil des Ministres Le Kiti n° AN VII-0114/FP/AGRI-EL du 22 novembre 1989 instituant une inspection sanitaire pour les animaux, produits et sous-produits d’origine animale destinés à la consommation humaine 1989 Coulibaly ND Yameogo KR Prevalence and control of zoonotic diseases: Collaboration between public health workers and veterinarians in Burkina Faso. Acta Trop 2000 76 53 57 10913767 Devleesschauwer B Aryal A Tharmalingam J Joshi DD Rijal S Speybroeck N Gabriël S Victor B Dorny P Complexities in using sentinel pigs to study Taenia solium transmission dynamics under field conditions. Vet. Parasitol 2013 193 172 178 23298565 Dorny P Brandt J Geerts S Immunodiagnostic approaches for detecting Taenia solium. Trends Parasitol 2004a 20 259 261 15147673 Dorny P Phiri I Vercruysse J Gabriel S Willingham A Brandt J Victor B Speybroeck N Berkvens D A Bayesian approach for estimating values for prevalence and diagnostic test characteristics of porcine cysticercosis. Int. J. Parasitol 2004b 34 569 576 15064121 Ganaba R Praet N Carabin H Millogo A Tarnagda Z Dorny P Hounton S Sow A Nitiéma P Cowan LD Factors associated with the prevalence of circulating antigens to porcine cysticercosis in three villages of Burkina Faso. PLoS Negl. Trop. Dis 2011 5 e927 21245913 Gonzalez AE Cama V Oilman RH Tsang VCW Pilcher JOYB Chavera A Castro M Montenegro T Verastegui M Bazalar H Prevalence and comparison of serological assays, necroscopy, and tongue examination for the diagnosis of porcine cysticercosis in Peru. Am. J. Trop. Med. Hyg 1990 43 194 199 2389823 Hall TA BioEdit: a user-friendly biological sequence alignment program for windows 95/98/NT. Nucl. Acids Symp. Series 1999 41 95 98 Lightowlers MW Garcia HH Gauci CG Donadeu M Abela-Ridder B Monitoring the outcomes of interventions against Taenia solium: Options and suggestions. Parasite Immunol 2016 38 158 169 26538513 Ministère des Ressources Animales Statistique du secteur de l'élévage 2010 http://cns.bf/IMG/pdf/mra_annuaire_2010.pdf Morales J Martínez JJ Garcia-Castella J Peña N Maza V Villalobos N Aluja AS Fleury A Fragoso G Larralde C Sciutto E Taenia solium: the complex interactions, of biological, social, geographical and commercial factors, involved in the transmission dynamics of pig cysticercosis in highly endemic areas. Ann. Trop. Med. Parasitol 2006 100 123 35 16492360 Ngowi HA Kassuku AA Maeda GEM Boa ME Willingham AL A slaughter slab survey for extra-intestinal porcine helminth infections in Northern Tanzania. Trop. Anim. Health Prod 2004 36 335 340 15241967 Nguyen MTT Gabriël S Abatih EN Dorny P A systematic review on the global occurrence of Taenia hydatigena. Vet. Parasitol 2016 226 97 103 27514893 Permin A Yelifari L Bloch P Steenhard N Hansen NP Nansen P Parasites in cross-bred pigs in the Upper East Region of Ghana. Vet. Parasitol 1999 87 63 71 10628701 Praet N Kanobana K Kabwe C Maketa V Lukanu P Lutumba P Polman K Matondo P Speybroeck N Dorny P Sumbu J Taenia solium cysticercosis in the democratic Republic of Congo: How does pork trade affect the transmission of the parasite? PLoS Negl. Trop. Dis 2010 4 e817 20838646 R Core Team R: A language and environment for statistical computing 2016 http://www.r-project.org/ Rodriguez-Hidalgo R Geysen D Benítez-Ortiz W Geerts S Brandt J Comparison of conventional techniques to differentiate between Taenia solium and Taenia saginata and an improved polymerase chain reaction-restriction fragment length polymorphism assay using a mitochondrial 12S rDNA fragment. J. Parasitol 2002 88 1007 1011 12435145 Sikasunge CS Phiri IK Phiri AM Siziya S Dorny P Willingham AL Prevalence of Taenia solium porcine cysticercosis in the Eastern, Southern and Western provinces of Zambia. Vet. J 2008 176 240 4 17468023 Sokal RR Rohlf JF Biometry: the principals and practice of statistics in biological research 1991 Freedman and Company New York 4rd Edition. WHO Taenia solium taeniasis/cysticercosis diagnostic tools. 2016 Report of a stakeholder meeting Geneva 17-18 December 2015 Geneva WHO Landscape analysis: control of Taenia solium. 2015 Geneva Yin D-M Wang X-J Lin Y Liu Z-Z Xia N-B Sheng X-F Wang T Liu Y Liu W Prevalence of helminths in adult pigs in Hunan province, China. J. Anim. Vet. Adv 2013 12 1123 1125 Zoli A Shey-Njila O Assana E Nguekam JP Dorny P Brandt J Geerts S Regional status, epidemiology and impact of Taenia solium cysticercosis in Western and Central Africa. Acta Trop 2003 87 35 42 12781376

PMC005xxxxxx/PMC5126643.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 0376343 5302 J Thorac Cardiovasc Surg J. Thorac. Cardiovasc. Surg. The Journal of thoracic and cardiovascular surgery 0022-5223 1097-685X 26298871 5126643 10.1016/j.jtcvs.2015.07.037 NIHMS830994 Article The diabetes epidemic and its effect on cardiac surgery practice Raza Sajjad MD a Blackstone Eugene H. MD ab Sabik Joseph F. III MD a a Department of Thoracic and Cardiovascular Surgery, Heart and Vascular Institute, Cleveland Clinic, Cleveland, Ohio b Department of Quantitative Health Sciences, Research Institute, Cleveland Clinic, Cleveland, Ohio Address for reprints: Joseph F. Sabik III, MD, Department of Thoracic and Cardiovascular Surgery, Cleveland Clinic, 9500 Euclid Ave/Desk J4-1, Cleveland, OH 44195, (sabikj@ccf.org) 21 11 2016 17 7 2015 10 2015 29 11 2016 150 4 783784 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. coronary artery bypass grafting diabetes cost THE EPIDEMIC The diabetes epidemic is one of the most challenging public health issues of the 21st century, responsible for 4.9 million deaths worldwide in 2014—one every 7 seconds.1 According to the World Health Organization, excess body weight and physical inactivity are the 2 main culprits leading to diabetes, which is now developing at younger ages.2 The difficulty of altering diet and lifestyle habits portends a future of ever-increasing prevalence. With diabetes comes development of coronary artery disease, which is not only common in diabetic patients, but also their major cause of death.3,4 Therefore, it is not surprising that the proportion of patients presenting for coronary artery bypass grafting (CABG) who have diabetes is also on the rise, as confirmed by our study5 published in the August 2015 issue of the Journal, nor is the fact that they need to be treated optimally to maximize survival. EFFECT ON PRACTICE We investigated outcomes of 10,362 medically treated diabetic patients and 45,139 nondiabetic patients who underwent primary isolated CABG from 1972 to 2011. The proportion of patients undergoing coronary surgery who had diabetes skyrocketed, from 7% in the 1970s to 40% in 2010. This experience is not isolated. Zhang and colleagues6 examined outcomes of 9240 patients undergoing primary isolated CABG at Fuwai Hospital in Beijing between 1999 and 2008; of these, 2682 had diabetes. Diabetic patients comprised 20% of the CABG population in 1999, but 32% in 2008. An analysis7 of the Society of Thoracic Surgeons database also showed an increase in the proportion of patients undergoing CABG who had diabetes, from 33% in 2000 to 40% in 2009. EFFECT ON EARLY OUTCOMES AND COST OF TREATMENT In our experience, occurrence of in-hospital complications and long-term survival (up to 20 years) was worse in diabetic patients, and the direct technical cost (the sum of direct preoperative, operative, and postoperative expenditures) of CABG was 9% higher than that for nondiabetic patients, a difference driven mainly by clinical and laboratory tests, imaging studies, medicines, nursing costs, and longer intensive care unit and postoperative stays. Again, ours are not isolated findings. China has the largest number of people with diabetes in the world—more than 100 million.6 The report of Zhang and colleagues6 discusses this population. Adjusted hospital death after primary isolated CABG was similar for diabetic and nondiabetic patients. In-hospital costs, medication costs, and total costs of CABG at 2 years were higher in diabetic patients ($66,000 vs $77,000; P<.001). EFFECT ON MANAGING CORONARY ARTERY DISEASE Diabetic patients represent an important population presenting for treatment of coronary artery disease. The BARI (Bypass Angioplasty Revascularization Investigation) 2D trial8 demonstrated that CABG is better than medical therapy, and the FREEDOM (Future Revascularization Evaluation in Patients With Diabetes Mellitus: Optimal Management of Multivessel Disease) trial9 demonstrated that CABG, rather than percutaneous coronary intervention, is the revascularization strategy of choice for diabetic patients with multivessel disease. However, although fewer adverse events occurred after CABG than after percutaneous coronary intervention in diabetic patients, their long-term survival after coronary surgery was worse than that of nondiabetic patients. In addition, we observed worse long-term survival among diabetic patients after CABG, and Zhang and colleagues6 documented that long-term outcomes (up to 10 years), including mortality, stroke, and major adverse cardiovascular and cerebrovascular events, were worse in diabetic patients, as was occurrence of rehospitalization for heart failure and stroke. These adverse long-term results after CABG could be due to at least 2 factors. First, coronary surgery may be less effective in diabetic than in nondiabetic patients, owing to worse graft patency. However, angiographic data from the BARI trial showed similar patency of internal thoracic artery and saphenous vein grafts in diabetic and nondiabetic patients.10 Second, diabetic patients have more comorbid conditions, either caused by or associated with diabetes, than do nondiabetic patients.5 Thus, the prevalence of cardiovascular risk factors, such as obesity and hypertension in our study, has increased across time in general, but more so in diabetic than nondiabetic patients. When the cause of death was compared 5 years after coronary revascularization in the BARI trial,11 cardiac mortality was similar—5.8% in diabetic versus 4.7% in nondiabetic patients—just as it was in the study of Zhang and colleagues in China, but noncardiac mortality was strikingly higher (12.2% vs 4.8%, respectively). Thus, worse long-term survival after CABG in diabetic patients is more likely due to their greater comorbidity burden than to ineffectiveness of CABG. EFFECT ON SOCIETY Coupled with the greater expense of bypass surgery in diabetic patients, this upward trend in prevalence of diabetes contributes importantly to the growing healthcare cost crisis. Currently, 25.8 million people in the United States have diabetes, and by 2034, that number is projected to reach 44.1 million.12,13 Even more alarming, the Centers for Disease Control and Prevention projects14 that if current trends continue, 1 in 3 US adults could have diabetes by 2050! This increase will lead to a further rise in the prevalence of coronary artery disease in the population, and to the need for revascularization. Clearly, policies and programs focused on controlling the factors that promote diabetes are critical to improving global public health and reining in the rising cost of healthcare. In the meantime, cardiac surgeons can play an important role in extending the lives of patients with diabetes by optimizing coronary revascularization, performing bilateral internal thoracic artery grafting with complete revascularization whenever feasible.15 Support for this work was provided by the Cleveland Clinic Department of Thoracic and Cardiovascular Surgery. Sajjad Raza, MD Joseph F. Sabik III, MD Eugene H. Blackstone, MD Abbreviations and Acronyms BARI Bypass Angioplasty Revascularization Investigation CABG coronary artery bypass grafting FREEDOM Future Revascularization Evaluation in Patients With Diabetes Mellitus: Optimal Management of Multivessel Disease Central Message Bilateral internal thoracic artery grafting with complete revascularization should be used to maximize long-term survival after CABG in diabetic patients. Perspective Controlling the factors that promote diabetes is critical to improving global public health and reining in the cost of healthcare. Cardiac surgeons can play an important role in extending the lives of patients with diabetes by optimizing coronary revascularization, performing bilateral internal thoracic artery grafting with complete revascularization whenever possible. Conflict of Interest Statement Dr Sabik is the North American principal investigator for the Abbott Laboratories–sponsored left main coronary disease randomized trial (EXCEL), is on the Society for Thoracic Surgeons Board of Directors, and is on the scientific advisory board of Medtronic. All other authors have nothing to disclose with regard to commercial support. 1 International Diabetes Federation (IDF) IDF Diabetes Atlas 6th Brussels IDF 2013 Available at: http://www.idf.org/diabetesatlas. Accessed July 6, 2015 2 World Health Organization WHO diabetes fact sheet Available at: http://www.who.int/mediacentre/factsheets/fs312/en/. July 7, 2015 3 Grundy SM Benjamin IJ Burke GL Chait A Eckel RH Howard BV Diabetes and cardiovascular disease: a statement for healthcare professionals from the American Heart Association Circulation 1999 100 1134 46 10477542 4 Bax JJ Young LH Frye RL Bonow RO Steinberg HO Barrett EJ Screening for coronary artery disease in patients with diabetes Diabetes Care 2007 30 2729 36 17901530 5 Raza S Sabik JF III Ainkaran P Blackstone EH Coronary artery bypass grafting in diabetics: a growing health care cost crisis J Thorac Cardiovasc Surg 2015 150 304 12 26027913 6 Zhang H Yuan X Osnabrugge RL Meng D Gao H Zhang S Influence of diabetes mellitus on long-term clinical and economic outcomes after coronary artery bypass grafting Ann Thorac Surg 2014 97 2073 9 24751154 7 El Bardissi AW Aranki SF Sheng S O’Brien SM Greenberg CC Gammie JS Trends in isolated coronary artery bypass grafting: an analysis of the Society of Thoracic Surgeons adult cardiac surgery database J Thorac Cardiovasc Surg 2012 143 273 81 22248680 8 A randomized trial of therapies for type 2 diabetes and coronary artery disease N Engl J Med 2009 360 2503 15 19502645 9 Farkouh ME Domanski M Sleeper LA Siami FS Dangas G Mack M Strategies for multivessel revascularization in patients with diabetes N Engl J Med 2012 367 2375 84 23121323 10 Schwartz L Kip KE Frye RL Alderman EL Schaff HV Detre KM Coronary bypass graft patency in patients with diabetes in the Bypass Angioplasty Revascularization Investigation (BARI) Circulation 2002 106 2652 8 12438289 11 Influence of diabetes on 5-year mortality and morbidity in a randomized trial comparing CABG and PTCA in patients with multivessel disease: the Bypass Angioplasty Revascularization Investigation (BARI) Circulation 1997 96 1761 9 9323059 12 Centers for Disease Control and Prevention (CDC) National diabetes fact sheet: national estimates and general information on diabetes and prediabetes in the United States, 2011 Atlanta US Department of Health and Human Services, CDC 2011 13 Huang ES Basu A O’Grady M Capretta JC Projecting the future diabetes population size and related costs for the U.S Diabetes Care 2009 32 2225 9 19940225 14 Centers for Disease Control and Prevention Number of Americans with diabetes projected to double or triple by 2050 Available at: http://www.cdc.gov/media/pressrel/2010/r101022.html. Accessed July 6, 2015 15 Raza S Sabik JF III Masabni K Ainkaran P Lytle BW Blackstone EH Surgical revascularization techniques that minimize surgical risk and maximize late survival after coronary artery bypass grafting in patients with diabetes mellitus J Thorac Cardiovasc Surg 2014 148 1257 64 discussion 64–6 25260269

PMC005xxxxxx/PMC5126758.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 7802221 6476 Plasmid Plasmid Plasmid 0147-619X 1095-9890 27234933 5126758 10.1016/j.plasmid.2016.05.001 NIHMS793195 Article New high-cloning-efficiency vectors for complementation studies and recombinant protein overproduction in Escherichia coli and Salmonella enterica VanDrisse C.M. Escalante-Semerena J.C. * Department of Microbiology, University of, 120 Cedar Street, Athens, GA 306022605 Georgia * Corresponding author: jescala@uga.edu (J.C. Escalante-Semerena) 5 11 2016 24 5 2016 7 2016 01 7 2017 86 16 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. Galloway et al. recently described a method to alter vectors to include Type IIS restriction enzymes for high efficiency cloning. Utilizing this method, the multiple cloning sites of complementation and overexpression vectors commonly used in our laboratory were altered to contain recognition sequences of the Type IIS restriction enzyme, BspQI. Use of this enzyme increased the rate of cloning success to >97% efficiency. L(+)-Arabinose-inducible complementation vectors and overexpression vectors encoding N-terminal recombinant tobacco etch virus protease (rTEV)-cleavable H6-tags were altered to contain BspQI sites that allowed for cloning into all vectors using identical primer overhangs. Additionally, a vector used for directing the synthesis of proteins with a C-terminal, rTEV-cleavable H6-tag was engineered to contain BspQI sites, albeit with different over-hangs from that of the previously mentioned vectors. Here we apply a method used to engineer cloning vectors to contain BspQI sites and the use of each vector in either in vivo complementation studies or in vitro protein purifications. Type IIS restriction cloning High-efficiency cloning vectors Complementation vectors Overexpression vectors Escherichia coli S. enterica 1. Introduction Increasing cloning efficiency is desirable for the rapid completion of experiments where complementation of function or protein overproduction is the objective. Engler et al. developed a method termed Golden Gate cloning; in which a Type IIS restriction enzyme (BsaI) is utilized for high efficiency cloning into expression constructs (Engler et al., 2008). Additionally, Galloway et al. (Galloway et al., 2013) reported an approach for the introduction of BspQI sites into a vector’s multiple-cloning site (MCS). Type IIS restriction enzymes (e.g., BspQI, BsaI, BpiI) have been used in cloning methods for the construction of gene expression reporters (Oster and Phillips, 2011), recombinant protein expression (Engler et al., 2008; Galloway et al., 2013), assembly of multiple DNA fragments (Engler et al., 2009) or gene fusions and promoter shuffling (Engler and Marillonnet, 2014). Type IIS restriction enzymes cleave at least one strand of DNA located outside of the recognition sequence, with BspQI cleaving both strands of DNA (Roberts et al., 2003). Type IIS restriction enzymes act as dimers, with one domain binding to a 4–7 base pair recognition sequence and another domain interacting with a cleavage site 1–20 nucleotides away from the recognition site (Pingoud et al., 2005; Szybalski et al., 1991). Having a separate cut site from the recognition sequence is valuable for vector design because the cleavage site can be engineered to have its own unique overhangs. An enzyme, such as BspQI, has a recognition sequence [GCTCTTC (1/4)] where the first number in parentheses corresponds to the position of cleavage on the coding strand and the second number to the cleavage site on the complementary strand (Roberts et al., 2003). This creates a three base pair overhang, resulting in 81 different possible overhangs (Fig. 1). The method described by Galloway et al. simplifies cloning efforts because digestion and ligation occur in the same reaction mixture, substantially reducing the time needed to clone genes of interest (Galloway et al., 2013). Additionally, because Type IIS restriction enzymes cut outside of their recognition sequence, ligated insert-vector products no longer contain the recognition site, and any empty vector remaining in the reaction mixture can be linearized by addition of the Type IIS restriction enzyme being used, substantially reducing false-positive background (Engler et al., 2008). In this study, we used the method described by Galloway et al. to introduce BspQI sites into the MCSs of plasmids used for arabinose-inducible expression of genes of interest, or for the overproduction of proteins. Here, we modified the MCSs of three complementation vectors (pBAD24, pBAD30, and pBAD33-SD1 (Guzman et al., 1995a)). Specifically, each MCS was modified to contain two BspQI sites. We also modified two overexpression vectors to contain BspQI sites matching those present in the complementation vectors. The resulting vectors directed the synthesis of proteins of interest fused to an N-terminal, re-combinant tobacco etch virus (rTEV)-cleavable hexahistidine (H6-tag (pTEV5, (Rocco et al., 2008)) or an N-terminal, rTEV-cleavable H6, maltose binding protein tag (pTEV6, (Rocco et al., 2008)). Conveniently, cloning into all of the above-mentioned vectors was done using the same primer overhangs. Lastly, an overexpression vector with a pTEV backbone was also altered to contain BspQI sites. In this case, however, the protein whose synthesis was directed by this vector was fused to a C-terminal, rTEV-cleavable H6-tag. Cloning into all of the BspQI-containing vectors was ≥97% efficient. The main goal of this work is to provide investigators with information regarding the specific aforementioned vectors. 2. Materials and methods 2.1. Bacterial strains, culture media, chemicals, and sequencing methods All strains used in this study are listed in Table S1. Escherichia coli C41 (λDE3) (Miroux and Walker, 1996) and DH5α (New England Biolabs) strains were grown at 37 °C in lysogeny broth (LB, Difco). Strains used for growth analysis were derivatives of Salmonella enterica sv Typhimurium LT2 and grown at 37 °C in nutrient broth (NB, Difco) containing NaCl (85 mM), or no-carbon essential (NCE) minimal medium (Berkowitz et al., 1968) supplemented with sodium acetate (10 mM), L-methionine (0.5 mM), MgSO4 (1 mM), and trace minerals (Atlas, 1995). Antibiotics were used at the following concentrations: ampicillin, 100 μg mL−1; chloramphenicol, 20 μg mL−1. All chemicals were purchased from Fischer unless noted otherwise; chloramphenicol, L(+)-arabinose (Sigma-Aldrich); ethylenediaminetetra-acetic acid (EDTA, VWR); and isopropylβ-D-1-thiogalactopyranoside (IPTG, IBI Scientific). All restriction enzymes were purchased from Thermo Scientific™ with the exception of BspQI (New England Biolabs). DNA sequencing was performed using Big Dye® Terminator v3.1 protocols (Applied Biosystems). DNA sequencing was performed at the Georgia Genomics Facility. 2.2. Construction of vectors containing BspQI sites Native BspQI sites (GCTCTTC) in the backbone of pBAD24, pBAD30, pBAD33-SD1, pTEV5, and pTEV6 were eliminated via site-directed mutagenesis and new BspQI sites were introduced into the MCS as described previously (Galloway et al., 2013). For a detailed procedure, see Supplemental Information or reference 2. Newly engineered vectors are listed in Table S3 and are hereafter referred to as pTEV16 (derived from pTEV5), pTEV17 (derived from pTEV6), and pCV1 (for Complementation Vector, derived from pBAD24), pCV2 (derived from pBAD30), and pCV3 (derived pBAD33-SD1). 2.3. Site-directed mutagenesis of pTEV16 and pTEV17 To reduce the number of primers and to facilitate cloning into over-expression and complementation vectors, DNA containing the cut site (pTEV16: AGC, pTEV17: ACC) was mutated to the pCV cut site (i.e., TTC) by site-directed mutagenesis. Polymerase chain reaction was performed using Pfu Ultra II DNA polymerase using primers listed in Table S2. Modifications included an anneal time of 60 s, an extension temperature of 68 °C, and an extension time of 2.5 min kb−1. DNA changes were confirmed by sequencing. 2.4. Construction of C-terminal rTEV cleavable H6-tag BspQI vector The pTEV18 vector was amplified with Pfu Ultra II DNA polymerase using primers outside of the N-terminal H6-tag and MCS (Table S2). PCR product was purified using Wizard® SV DNA Clean-Up System (Promega). Linear blunt-end fragments were ligated using the Fast-Link-DNA Ligation Kit (Epicentre), and transformed into DH5α chemically competent cells (Maniatis et al., 1982). Cells were plated on LB + ampicillin, incubated overnight at 37 °C, and 10 individual ApR colonies were used to inoculate fresh LB + ampicillin medium. Plasmids were isolated from each overnight culture using the Wizard® Plus SV Miniprep DNA Purification System (Promega). The presence of BspQI sites in the MCS was determined by cutting with BspQI and an enzyme outside of the MCS (e.g., ScaI). The resulting MCS was sequenced to confirm the presence of the BspQI site and C-terminal H6 rTEV cleavable tag. 2.5. Cloning of SeAcs into newly designed BspQI vectors To illustrate the efficiency of the new vectors, the S. enterica acs+ gene encoding acetyl-CoA synthetase was cloned. For this purpose, primers were designed with BspQI sites on the 5′-ends such that when amplicons were cut, a three base pair overhang was complementary to overhangs in the corresponding digested vector (Table S2). The acs+ gene was PCR amplified from S. enterica LT2 genomic DNA using Pfu Ultra II DNA polymerase. PCR products corresponding to the correct size were verified via gel electrophoresis and extracted as described above. Gene products were cloned into pCV1, pCV2, pCV3 pTEV16, pTEV17, pTEV18, pTEV19 and pTEV20 as described elsewhere (Galloway et al., 2013). For a more detailed procedure, refer to supplemental materials or reference 2. A sample of the ligation reaction (1 μL) was used to transform E. coli DH5α competent cells (Maniatis et al., 1982), and cells were plated on LB containing the appropriate antibiotic. The presence of inserts in the cloning vectors was confirmed by colony PCR with Go Taq® Green Master Mix (Promega) using primers that annealed to the plasmid outside of the MCS (Table S2). PCR products were analyzed on a 1% agarose gel with Tris/Borate/EDTA buffer for 40 min at 115 V. Cells harboring plasmids containing the expected inserts were grown overnight in LB supplemented with the appropriate antibiotic, and plasmids were isolated as described above. The presence of acs+ in each plasmid was confirmed by DNA sequencing. 2.6. Growth studies Complementation vectors carrying the acs+ allele or empty vector were electroporated into acs+ and Δacs strains (O’Toole et al., 1993) (Table S1). Starter cultures of each strain were grown overnight at 37 °C with shaking in NB containing the appropriate antibiotic. Fresh minimal medium (198 μl) containing sodium acetate (10 mM), appropriate antibiotics, and L(+)arabinose (250 μM) was dispensed into each well of a 96-well microtiter dish. Each well was inoculated with 2 μl (1% v/v) of the aforementioned overnight cultures. Arabinose was included in the medium to induce expression of acs+ from the ParaBAD promoter. Growth was monitored at 630 nm with shaking at 37 °C for 24 h (BioTek ELx808-1 Ultra microplate reader). Three technical and three biological replicates were analyzed; a representative growth curve is shown. Data were analyzed using Prism v6 software (GraphPad) and error bars represent standard deviation. 2.7. Purification of SeAcs from pTEV16-20 Plasmids encoding H6-SeAcs (pACS65 and pACS67), H6-MBP-tagged SeAcs (pACS66 and pACS68), and SeAcs-H6 (pACS69) were electroporated using a protocol described elsewhere (Seidman et al., 1997) into Escherichia coli strain C41 (λDE3) (Miroux and Walker, 1996) Δpat (strain JE9314). Cultures of cells containing plasmids were grown to stationary phase (OD650 ~ 1.3) and sub-cultured (1:100 v/v) into 1 L of LB + ampicillin. Cultures were grown shaking at 37 °C to an OD650 of 0.5, after which ectopic gene expression was induced with IPTG (0.5 mM). Cultures were grown overnight at 25 °C, cells were harvested by centrifugation at 6000 ×g for 15 min at 4 °C, and cell pellets were stored at −80 °C until used Acs proteins were purified from cell pellets as described (Hentchel and Escalante-Semerena, 2015). The amount of Acs in fractions was quantified on a NanoDrop™ 1000 Spectrophotometer (Thermo Scientific) using the molecular weight (72.15 kDa) and extinction coefficient (138,770 M−1 cm−1; ExPASy ProtParam) of the protein. Percent purity was calculated using ImageQuant™ TL. 3. Results 3.1. Cloning efficiencies of newly designed vectors The S. enterica acs+ gene was cloned into each vector (Materials and methods Section 2.6) to test complementation (pCV1–3) and overexpression (pTEV16-20) using primers listed in Table S2. For each vector, 20 colonies were screened using colony PCR and cloning efficiencies were calculated (Table 1). A 2.2-kb band was indicative of a positive result, as shown in Fig. S2. The cloning efficiency varied between 95 and 100%. Information on troubleshooting cloning issues can be found in Table 3 of the Conclusions section. 3.2. Description of complementation vectors (pCVs) Plasmids pCV1, pCV2, and pCV3 contained the promoter region of the E. coli araBAD operon (ParaBAD) for the purpose of inducing expression of genes of interest by the addition of L(+)-arabinose (Cronan, 2006; Guzman et al., 1995a). All vectors contained an rrnB terminator, and either a bla+ gene for ampicillin resistance (pCV1, pCV2) or a cat+ gene for chloramphenicol resistance (pCV3) (Guzman et al., 1995b). Plasmid pCV1 contained a pBR322 origin of replication, while pCV2 and pCV3 contain a p15A origin of replication, making pCV1 compatible in vivo with either pCV2 or pCV3. All plasmids contained M13 intergenic regions for ssDNA packaging into phage capsids. The MCS and plasmid maps of each complementation vector are shown in Fig. 2. Noteworthy is the presence of the Shine Dalgarno (SD) sequence in plasmids pCV1 and pCV3. Plasmid pCV2 did not contain a SD sequence, allowing for the cloning of a gene and its native ribosome-binding site. A stop codon was added after the second BspQI site and when combined with a gene’s native stop codon, the error of read through was reduced. NheI, EcoRI and HindIII restriction sites surrounded the BspQI sites for verification of inserts by restriction analysis. It was necessary to add BspQI recognition and cut sites to both forward and reverse primers. This allowed for cutting of amplified gene products so they could anneal to the corresponding digested overhangs for each vector. In the case of pCV plasmids, the primer overhangs were identical for all three vectors. A simplified list of overhangs to be added to primers for each vector can be found in Table 2. 3.3. Use of complementation vectors for growth behavior analysis S. enterica Δacs strains grow poorly on 10 mM acetate as the sole source of carbon and energy, and such growth defect can be corrected by the ectopic expression of the acs+ allele (Chan et al., 2011). We used the above-mentioned phenotype of S. enterica acs strains to verify the functionality of the acs+ allele in the newly described complementation vectors as described in the Materials and Methods Section 2.7. Plasmids pCV1, pCV2, or pCV3 encoding S.e. acs+ (pACS62-pACS64) were induced with L(+)-arabinose (250 μM), a concentration of inducer that effectively compensated for the absence of the Acs protein in a S. enterica Δacs strain (Fig. 3). This result indicated that introduction of BspQI sites into the MCS did not disrupt the intended function of these plasmids. 3.4. Overexpression vectors containing rTEV protease cleaving sites All of the TEV vectors described herein contained the E. coli lacI+ allele, whose protein product acts as a repressor to the T7 promoter of the plasmids. Addition of IPTG had the dual effect of relieving LacI repression and inducing transcription of genome-encoded T7 RNA polymerase. All vectors included a bla+ gene for the synthesis of β-lactamase, which provided resistance to ampicillin, an f1 origin for ssDNA packaging into phage capsids, and a pBR322 origin of replication (Bolivar, 1978; Soberon et al., 1980). The plasmid maps and MCSs of plasmids pTEV16-20 are shown in Fig. 4. 3.4.1. Plasmids pTEV16, pTEV18 Genes expressed from plasmids pTEV16 and pTEV18 resulted in proteins with an N-terminal, rTEV-cleavable H6-tag. Plasmid pTEV18 was constructed by mutagenizing the NheI site of plasmid pTEV16 to GCTTTC, rendering this site inactive, but allowed for the design of one set of primers that could be used for cloning into the complementation and overexpression vectors. Tag removal resulted in proteins with three additional residues (i.e., Gly-Ala-Ser for pTEV16; or Gly-Ala-Phe for pTEV18) on the N-terminus. A stop codon was engineered immediately after the BspQI sites to reduce read through when combined with the stop codon of the gene of interest. Plasmids pTEV16 and pTEV18 also contained multiple restriction sites in the MCS for alternative cloning methods or verification via restriction analysis. An XbaI site in plasmid pTEV18 is present upstream of the H6-tag, if needed for insert verification by restriction analysis. 3.4.2. Plasmids pTEV17, pTEV19 Genes expressed from plasmids pTEV17 and pTEV19 directed the synthesis of proteins with an N-terminal rTEV-cleavable H6-tag fused to a maltose binding protein (MBP). The MBP-tag was included with the goal of increasing protein solubility. Plasmid pTEV19 was constructed by mutagenizing the KpnI site of pTEV17 to GGTTTC, rendering this site inactive but allowing for the design of primers that could be used with complementation and overexpression vectors in this study. A stop codon was also engineered in these vectors to reduce read through. Tag removal resulted in proteins with two additional residues (i.e., Gly-Thr for pTEV17; or Gly-Phe for pTEV19). Additional restriction sites different from those found in plasmids pTEV16 and pTEV18 were present in plasmids pTEV17 and pTEV19 for alternative cloning options or verification via restriction analysis. An XbaI site in pTEV19 lies upstream of the H6-tag, if needed for insert verification by restriction analysis. 3.4.3. Vectors engineered for the synthesis of C-terminally tagged proteins Plasmid pTEV20 was also engineered for the overproduction of proteins. However, pTEV20 differed from the aforementioned overexpression vectors by the location of the rTEV-cleavage site and the primer overhangs. Proteins overproduced in cells harboring plasmid pTEV20 have a C-terminal H6-tag that can be removed by rTEV protease. Plasmid pTEV20 was constructed by amplifying plasmid pTEV18 outside of the H6-tag and MCS; consequently pTEV20 maintained the backbone features of plasmids pTEV16 and pTEV18. Differences in the MCS of pTEV20 can be seen in Fig. 4. Designed primer overhangs also differed for this vector due to the fact that there is no stop codon engineered immediately downstream of the cloned gene. When designing primers for pTEV20, it was necessary to remove the stop codon of the native gene to ensure read through of the rTEV cleavage site and the H6-tag. After cleavage, proteins obtained from pTEV20 retained six additional amino acids (i.e., Glu-Asn-Leu-Tyr-Phe-Gln) at the C-terminus. A simplified list of overhangs to add to primers for each TEV-vector can be seen in Table 2. 3.5. Isolation of Salmonella enterica acetyl-CoA synthetase (Acs) from pTEV16-20 To assess the efficacy of the overexpression vectors, S.e. acs+ was cloned into pTEV16-20, overexpressed in E. coli strain C41 (ΔDE3), and proteins were purified as described in the Materials and Methods Section 2.8. Purity of protein purified from cells overexpressing acs+ from pTEV16, pTEV17, pTEV18, pTEV19, and pTEV20 was assessed and final protein products are shown in Fig. 5. Yields of SeAcs protein purified from overexpression from each vector were 8.35 mg L−1 (pTEV16), 0.9 mg L−1 (pTEV17), 8.55 mg L−1 (pTEV18), 0.9 mg L−1 (pTEV19), 5.32 mg L−1 (pTEV20). 4. Conclusions Here we describe modified cloning vectors containing BspQI restriction sites that allow for rapid and efficient cloning of genes of interest. Common issues that may arise during cloning may be resolved by solutions provided in Table 3. Using these vectors paired with a previously described cloning method (Galloway et al., 2013), the number of cells carrying empty vectors is greatly reduced. We also have designed the complementation vectors (pCV1–3) and N-terminal rTEV cleavable H6-tag vectors (pTEV18–19) to carry the same primer overhangs so that only one primer set is needed for in vivo and in vitro studies. Because MBP is a 42-kDA protein, yields of the cleaved protein once fused to MBP were lower. For this reason, it may be necessary to increase expression volumes to obtain similar yields compared to the pTEV16/18 vectors. It is important to note that when purifying proteins encoded by plasmids pTEV17 or pTEV19, some MBP contamination may occur, which if necessary may be removed using an MBPTrap HP column (GE Healthcare Life Sciences). Additionally, an N-terminal tag may be problematic for some proteins due to folding issues or interference with the protein’s activity. For this reason we have provided a C-terminal rTEV-cleavable H6-tag vector (pTEV20). Due to the nature of the rTEV cleavage site, recombinant proteins purified and cleaved from the cells harboring this vector will contain additional amino acids compared to the N-terminally tagged vectors. Overall, we have established a set of vectors that should facilitate complementation studies or protein production. Supplementary Material sup sup1 This work was supported by USPHS grant R01 GM062203 to J.C.E.-S. The authors thank Nicholas Galloway in the laboratory of Dr. Cory Momany for the technical assistance. Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx.doi.org/10.1016/j.plasmid.2016.05.001. Fig. 1 Visual representation of BspQI vector digestion. BspQI binds to the recognition site [GCTCTTC (1/4), highlighted in grey] and will cut both strands at indicated arrows. This creates a three base pair overhang. The cut site does not interfere with recognition and can be custom designed. Shown is an example of vectors constructed in this study (pCV1–3, pTEV18, 19). For a more detailed figure, see supplemental. Fig. 2 Plasmid maps and MCSs of complementation vectors. Plasmid maps show notable genetic features. Plasmid MCS sequences of each vector are shown between NheI and HindIII restriction sites. S.D. in MCSs represents Shine-Dalgarno (ribosome-binding site) sequence. Asterisks identify location of BspQI sites in original vectors that were mutated from GCTCTTC to WCTWTTC, as explained in Materials and methods Section 2.3. Fig. 3 Growth analysis and complementation of S. enterica acs phenotype. S.e. acs+ was cloned into pCV1-pCV3 and transformed into a acs strain to assess the effectiveness of the vectors. Cells were grown in NCE medium with acetate (10 mM) and transcription of acs+ in each vector was induced with L(+)-arabinose (250 μM). Growth curves were obtained using a microplate reader (BioTek Instruments). Fig. 4 Plasmid maps and MCSs of overexpression vectors. Plasmid maps show notable genetic features. Plasmid MCS sequences of each vector are shown. Nucleotides inside black boxes of pTEV16 were changed to TTC resulting in pTEV18 and nucleotides in black boxes in pTEV17 were changed to TTC resulting in pTEV19. Therefore pTEV18 lacks an NheI site and pTEV19 lacks a KpnI site. An XbaI site lies upstream of the H6-tag. Plasmid pTEV20 was constructed by amplifying plasmid pTEV18 outside of the H6-tag and NotI restriction site, resulting in an rTEV-cleavable C-terminal H6-tag. Arrows mark the rTEV cleavage site. Asterisks identify location of BspQI sites in original vectors that were mutated from GCTCTTC to WCTWTTC, as explained in Materials and methods Section 2.3. Fig. 5 SDS-PAGE gel of purified Acs A 12.5% SDS-PAGE gel shows molecular weight standards (BioRad Precision Plus Protein™, left) and Acs (72 kDa) purified from cells harboring pACS65 (pTEV16), pACS66 (pTEV17), pACS67 (pTEV18), pACS68 (pTEV19), and pACS69 (pTEV20). Numbers above each well correspond to protein purified from cells harboring respective pTEV vectors (e.g., well labeled 16 corresponds to pTEV16). Table 1 Cloning efficiencies of newly designed vectors. Vector GenBank accession # No. of positive clones (20 screened)a Efficiency (%)b pCV1 KU974153 19 95 pCV2 KU974154 20 100 pCV3 KU974155 19 95 pTEV16 KU974156 20 100 pTEV17 KU974157 20 100 pTEV18 KU974158 18 90 pTEV19 KU974159 19 95 pTEV20 KU974160 20 100 A positive result was indicated by a 2.2 kbband as shown in Fig. S2. a Colony PCR was used to screen 20 colonies for each vector. b Cloning efficiencies were calculated as described under Materials & methods Section 2.6. Table 2 Designing primers for vectors used in this study. Plasmid Sequence to add to 5′ of primer pTEV16 forward primer 5′NNGCTCTTCNAGC pTEV17 forward primer 5′NNGCTCTTCNACC pCV1, pCV2, pCV3, pTEV18, pTEV19 forward primer 5′NNGCTCTTCNTTC pCV1, pCV2, pCV3, pTEV16, pTEV17, pTEV18, pTEV19 reverse primer 5′NNGCTCTTCNTAA pTEV20 forward primer 5′NNGCTCTTCNTAC pTEV20 reverse primer 5′NNGCTCTTCNTTC Table 3 Troubleshooting. Problem Potential explanations or solutions Cloning yielding all empty vector Potential primer dimers cloning into vectors, gel extract PCR reaction Check activity of BspQI Cloning yielding incorrect insert sizes Verify competent cells are DH5α and not contaminant Contaminants in PCR reaction, gel extract correct band size Cloning yielding zero colonies Confirm primer overhangs correspond correctly to vector overhangs Check for BspQI sites in gene, if present eliminate second BspQI digestion step, then inactivate ligase at 70 °C before transformation Re-PCR amplify insert to rule out mutations in overhangs Verify activity of ligase or stability of ATP Atlas R 1995 Handbook of Media for Environmental Microbiology CRC Press Boca Raton Berkowitz D 1968 Procedure for identifying nonsense mutations J Bacteriol 96 215 220 4874308 Bolivar F 1978 Construction and characterization of new cloning vehicles. III Derivatives of plasmid pBR322 carrying unique Eco RI sites for selection of Eco RI generated recombinant DNA molecules Gene 4 121 136 363519 Chan CH 2011 In Salmonella enterica, the sirtuin-dependent protein acylation/ deacylation system (SDPADS) maintains energy homeostasis during growth on low concentrations of acetate Mol Microbiol 80 168 183 21306440 Cronan JE 2006 A family of arabinose-inducible Escherichia coli expression vectors having pBR322 copy control Plasmid 55 152 157 16139359 Engler C Marillonnet S 2014 Golden gate cloning Methods Mol Biol 1116 119 131 24395361 Engler C 2008 A one pot, one step, precision cloning method with high throughput capability PLoS ONE 3 e3647 18985154 Engler C 2009 Golden gate shuffling: a one-pot DNA shuffling method based on type IIs restriction enzymes PLoS ONE 4 e5553 19436741 Galloway NR 2013 Rapid cloning for protein crystallography using type IIS restriction enzymes Crystal Growth & Design 13 2833 2839 Guzman LM 1995a Tight regulation, modulation, and high-level expression by vectors containing the arabinose PBAD promoter J Bacteriol 177 4121 4130 7608087 Guzman LM 1995b Tight regulation, modulation, and high-level expression by vectors containing the arabinose PBAD promoter J Bacteriol 177 4121 4130 7608087 Hentchel KL Escalante-Semerena JC 2015 In Salmonella enterica, the Gcn5-related acetyltransferase MddA (formerly YncA) acetylates methionine sulfoximine and methionine sulfone, blocking their toxic effects J Bacteriol 197 314 325 25368301 Maniatis T 1982 Introduction of plasmid and bacteriophage lambda into Escherichia coli Maniatis T Molecular Cloning: A Laboratory Manual Cold Spring Harbor Laboratory Cold Spring Harbor, New York Miroux B Walker JE 1996 Over-production of proteins in Escherichia coli: mutant hosts that allow synthesis of some membrane proteins and globular proteins at high levels J Mol Biol 260 289 298 8757792 Oster CJ Phillips GJ 2011 Vectors for ligation-independent construction of lacZ gene fusions and cloning of PCR products using a nicking endonuclease Plasmid 66 180 185 21854804 O’Toole GA 1993 Analysis of mutants of defective in the synthesis of the nucleotide loop of cobalamin J Bacteriol 175 3317 3326 8501035 Pingoud A 2005 Type II restriction endonucleases: structure and mechanism Cell Mol Life Sci 62 685 707 15770420 Roberts RJ 2003 A nomenclature for restriction enzymes, DNA methyltransferases, homing endonucleases and their genes Nucleic Acids Res 31 1805 1812 12654995 Rocco CJ 2008 Construction and use of new cloning vectors for the rapid isolation of recombinant proteins from Escherichia coli Plasmid 59 231 237 18295882 Seidman CG 1997 Introduction of plasmid DNA into cells Ausubel FM Current Protocols in Molecular Biology 1 Wiley Interscience New York Soberon X 1980 Construction and characterization of new cloning vehicles. IV Deletion derivatives of pBR322 and pBR325 Gene 9 287 305 6248430 Szybalski W 1991 Class-IIS restriction enzymes—a review Gene 100 13 26 2055464

PMC005xxxxxx/PMC5127197.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 0412041 133 Acta Neuropathol Acta Neuropathol. Acta neuropathologica 0001-6322 1432-0533 23109048 5127197 10.1007/s00401-012-1054-9 NIHMS830852 Article Filamin C-related myopathies: pathology and mechanisms Fürst Dieter O. Institute for Cell Biology, University of Bonn, Ulrich-Haberland-Str. 61a, 53121 Bonn, Germany dfuerst@uni-bonn.de Goldfarb Lev G. Clinical Neurogenetics, National Institutes of Health, Bethesda, MD, USA GoldfarbL@ninds.nih.gov Kley Rudolf A. Department of Neurology, Neuromuscular Center Ruhrgebiet, University Hospital Bergmannsheil, Ruhr-University Bochum, Bochum, Germany rudolf.kley@rub.de Vorgerd Matthias Department of Neurology, Neuromuscular Center Ruhrgebiet, University Hospital Bergmannsheil, Ruhr-University Bochum, Bochum, Germany matthias.vorgerd@bergmannsheil.de Olivé Montse Institute of Neuropathology, Department of Pathology, and Neuromuscular Unit, Department of Neurology, IDIBELL-Hospital Universitari de Bellvitge, Hospitalet de Llobregat, Barcelona, Spain 25169mop@comb.cat van der Ven Peter F. M. Institute for Cell Biology, University of Bonn, Ulrich-Haberland-Str. 61a, 53121 Bonn, Germany pvdven@uni-bonn.de 19 11 2016 30 10 2012 1 2013 29 11 2016 125 1 3346 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. The term filaminopathy was introduced after a truncating mutation in the dimerization domain of filamin C (FLNc) was shown to be responsible for a devastating muscle disease. Subsequently, the same mutation was found in patients from diverse ethnical origins, indicating that this specific alteration is a mutational hot spot. Patients initially present with proximal muscle weakness, while distal and respiratory muscles become affected with disease progression. Muscle biopsies of these patients show typical signs of myofibrillar myopathy, including disintegration of myofibrils and aggregation of several proteins into distinct intracellular deposits. Highly similar phenotypes were observed in patients with other mutations in Ig-like domains of FLNc that result in expression of a noxious protein. Biochemical and biophysical studies showed that the mutated domains acquire an abnormal structure causing decreased stability and eventually becoming a seed for abnormal aggregation with other proteins. The disease usually presents only after the fourth decade of life possibly as a result of ageing-related impairments in the machinery that is responsible for disposal of damaged proteins. This is confirmed by mutations in components of this machinery that cause a highly similar phenotype. Transfection studies of cultured muscle cells reflect the events observed in patient muscles and, therefore, may provide a helpful model for testing future dedicated therapeutic strategies. More recently, FLNC mutations were also found in families with a distal myopathy phenotype, caused either by mutations in the actin-binding domain of FLNc that result in increased actin-binding and non-specific myopathic abnormalities without myofibrillar myopathy pathology, or a nonsense mutation in the rod domain that leads to RNA instability, haploinsufficiency with decreased expression levels of FLNc in the muscle fibers and myofibrillar abnormalities, but not to the formation of desmin-positive protein aggregates required for the diagnosis of myofibrillar myopathy. Filamin C Filaminopathy Myofibrillar myopathy Distal myopathy Limb-girdle muscular dystrophy Pathomechanism Introduction Filaminopathies are recently identified progressive skeletal myopathies manifesting initially by bilateral weakness in either proximal leg muscles or in distal upper limb muscles spreading to other muscle groups and in some forms eventually resulting in tetraparesis and wheelchair dependence [6, 18, 22, 57]. Three distinct types of filaminopathy are recognized. The disease caused by mutations resulting in protein aggregation (so far found at various locations in the FLNc rod domain) presents in the fourth-to-sixth decade of life with slowly progressive predominantly proximal muscle weakness. Associated cardiac and respiratory muscle involvement severely complicate the course of illness [22, 57]. In contrast, mutations in the actin-binding domain (ABD) of FLNc are responsible for the second disease variant that is initially characterized by weakness and wasting of distal muscles, especially intrinsic hand muscles, manifesting in the third decade of life [6]. An intermediate filaminopathy phenotype affecting primarily distal muscles of the upper and lower limbs has recently been described [18]. Muscle biopsies of patients with aggregation-causing FLNC mutations show disintegration of myofibrils and formation of desmin-positive protein aggregates within muscle fibers [30, 39]. These are typical findings in myofibrillar myopathies (MFM), a clinically and genetically diverse group of progressive devastating hereditary skeletal and cardiac myopathies. Thus far MFM has been associated with mutations in seven genes (DES, MYOT, LDB3/ZASP, CRYAB, BAG3, FLNC and FHL1 [33, 40–42]). Muscle biopsies from patients with the second filaminopathy variant show non-specific myopathic abnormalities without MFM pathology [6], while histological evaluation in cases with the intermediate variant indicated disease-associated myofibrillar abnormalities, but desmin-positive protein aggregates required for the diagnosis of MFM were not detected [18]. As a pathological entity, the first variant of filaminopathy related to FLNc rod mutations classifies with a group named protein aggregate myopathy (PAM). PAM is a general term for neuromuscular conditions marked by aggregation of proteins within muscle fibers. This is a diverse group of disorders that in addition to MFM includes among others, nemaline myopathy, myosin storage myopathy, cytoplasmic body myopathies, and reducing body myopathy [15, 16]. The other filaminopathy variants belong to autosomal dominant distal myopathies, of which nine types have been assigned to known genes [50]. The list includes disorders caused by TTID, LDB3, CRYAB, FLNC and DES mutations, genes that have also been associated with MFM, thus indicating a close relationship between these classification units. This review presents a comparative analysis of the contrasting FLNc mutation-related phenotypes of filaminopathy and their distinct underlying pathomechanisms. It also offers practical considerations regarding diagnostic procedures, severely complicated due to the existence of a FLNC-related pseudogene, as well as health implications and therapeutic strategies. Clinical aspects MFM-type FLNc myopathy Evaluation of about 70 MFM patients with different FLNC mutations has revealed a markedly homogenous clinical phenotype ([2, 22, 23, 28, 44, 48] and our unpublished data). Muscle weakness mostly starts in the fourth or fifth decade of life (range 24–60 years). Proximal muscle weakness leading to difficulty walking uphill and climbing stairs is the initial sign. In the course of illness, most patients develop slowly progressive weakness in both distal and proximal leg and arm muscles (Fig. 1a). Winging of the scapula is a frequent phenomenon. Muscle weakness slowly progresses to the inability to walk. An involvement of respiratory muscles, often requiring nocturnal ventilation, usually occurs with disease progression and contributes to reduced life expectancy. About one-third of the patients showed cardiac abnormalities, including conduction blocks, left ventricular hypertrophy, and diastolic dysfunction. Sudden cardiac arrest as the cause of death was presumed in at least five patients. Creatine kinase (CK) levels were mostly elevated up to tenfold of the upper limit. Electromyography regularly showed typical myopathic changes. Although histological findings indicated neurogenic changes in about one-half of skeletal muscle biopsies [22], clinical examinations and neurophysiological measurements did not reveal a relevant involvement of the peripheral nervous system in any of the FLNc patients. Chronic gastrointestinal complaints were reported by a few patients with p.W2710X and p.K899_V904del/V899_C900ins mutations and may indicate an involvement of smooth muscle [23, 28]. Late-onset cerebellar ataxia with atrophy of cerebellum and vermis was observed in one sporadic patient [48], but the causal relationship with the detected mutation in FLNc Ig-like domain 22 is unclear. A diagnostic challenge is to discriminate FLNc-based MFM-type myopathy from other myopathies including MFM subtypes, limb-girdle muscular dystrophies (LGMDs), X-chromosomal muscular dystrophy Becker type, myotonic dystrophy type 2 (PROMM), acid maltase deficiency (late-onset Pompe disease), and inclusion body myositis/myopathy. All these diseases are late-onset myopathies that typically present with proximal weakness, slow disease progression and mild to moderate CK elevation. The features of FLNc-based MFM-type myopathy most useful for differential diagnosis appear to be a symmetrical involvement of proximal muscles in the lower extremities, respiratory weakness during the disease course, an autosomal dominant inheritance pattern, MFM-typical histological changes and characteristic muscle imaging findings (see below). Distal FLNc myopathy Distal myopathy was associated with missense mutations (p.A193T; p.M251T) located in the N-terminal actin-binding domain of FLNc in families from Australia and Italy [6]. The illness developed in the third decade of life. Intrinsic hand muscles were primarily affected and led to reduced grip strength, followed by leg muscle weakness resulting in difficulties with running and jumping. CK levels were only mildly elevated up to threefold of the normal upper limit. Two patients displayed signs of cardiomyopathy and none had respiratory insufficiency. In a recently reported Bulgarian family with a frame-shifting deletion mutation in exon 30 of FLNC (p.F1720LfsX63) leading to haploinsufficiency, the disease was associated with distal muscle weakness primarily in the upper limbs with lower limb involvement upon disease progression [18]. It manifests in adulthood between the ages of 20–57 years. Initial symptoms were distal muscle weakness mostly in the upper limbs with subsequent lower limb involvement upon disease progression. CK levels ranged from normal to sixfold elevated. Neurophysiological studies revealed normal motor and sensory nerve conduction velocities. None of the patients had respiratory disturbances and cardiomyopathy was documented in only a single patient. The main clinical differential diagnoses are adult-onset distal myopathies, a group of muscle diseases which share the clinical findings of predominant weakness in lower leg and/or hand muscles. Typical clinical findings characterizing distal FLNc myopathy are weakness in hand and calf muscles with an onset in early adulthood and a family history compatible with an autosomal dominant trait. All mutations are further specified below in the genetics section (Fig. 3; Table 1). Muscle imaging in FLNc myopathy Magnetic resonance imaging (MRI) is a powerful and non-invasive tool in the diagnostic workup, evaluation of therapeutic efficacy and disease follow-up in neuromuscular disorders. MRI of lower limbs showed a rather homogenous pattern of symmetrical muscle involvement in MFM-type disease caused by FLNC mutations in Ig-like domain 7 and 24 ([10, 22, 23] and our unpublished data). Non fat-saturated T1-weighted images showed a reticular pattern of hyperintensity in less affected patients, whereas homogenous lipomatous alterations were visible in individuals with a more advanced disease. In proximal lower limbs, gluteal muscles, semimembranosus, adductor magnus and longus, long head of biceps femoris, vastus intermedius and vastus medialis were most affected (Fig. 1b, c). The rectus femoris seemed to be more affected in patients carrying the p.V930_T933del mutation than in those with Ig-like domain 24 mutation [23]. The sartorius and gracilis muscles appeared almost normal, even in patients with more advanced clinical course. In lower legs, soleus and the medial head of gastrocnemius were most affected, followed by tibialis anterior, extensor hallucis longus, extensor digitorum longus and peroneal muscles (Fig. 1d). The lateral head of the gastrocnemius was relatively spared. Muscular signal intensities on T2-weighted TIRM images were only mildly elevated, indicating an absence of distinct intramuscular edema. The pattern of muscle involvement in patients with MFM caused by FLNC mutations is similar to that observed in MFM caused by MYOT or ZASP mutations, with only subtle differences detectable by statistical analysis, but is sharply different from that observed in desminopathy or αB-crystallinopathy. Indeed, comparison with other genetically classified MFM subtypes revealed that the combination of the following findings was highly specific for MFM caused by FLNC mutations [10]: (1) semitendinosus and long head of biceps femoris at least equally affected as sartorius, (2) semimembranosus at least equally affected as adductor magnus, (3) medial head of the gastrocnemius more affected than the lateral head. These criteria developed initially in a retrospective study were validated in subsequent MRI analyses of newly identified MFM patients ([23] and our unpublished data). Muscle imaging data of distal myopathies caused by FLNC mutations are rather limited. In patients with ABD mutations [6], fatty degeneration of semimembranosus and, in contrast to MFM-associated filaminopathy, semitendinosus was described as an early change in thigh muscles, followed by involvement of all hamstring muscles and adductor magnus. Upon further progression of the disease, vastii muscles of the quadriceps showed lipomatous alterations. At the lower leg level, severe fatty degeneration of soleus muscles, asymmetrical involvement of peroneal muscles and a slightly lesser involvement of medial and lateral gastrocnemius muscle were described in a patient with p.M251T mutation. A computer tomography scan in a patient with p.A193T mutation showed severe fatty degeneration of soleus and gastrocnemii muscles and a less severe involvement of peroneal muscles. The anterior compartment and tibialis posterior were spared in both patients. Compared to filaminopathy patients with MFM phenotype, semitendinosus and lateral gastrocnemius seems to be more and tibialis anterior and adductor magnus less affected. Interestingly, the early involvement of the semitendinosus observed in patients with ABD mutations is also typically seen in MFM resulting from mutations in DES or CRYAB [10] and in the recently described hereditary myopathy with early respiratory failure (HMERF) resulting from mutations in the A-band portion of titin [32, 34]. In contrast, the pattern of muscle involvement in the lower legs shows clear differences between these diseases [10, 32, 34]. Muscle MRI of lower limbs in one patient with FLNC haploinsufficiency showed similarities to MFM-associated filaminopathy regarding the most severely affected muscles: gluteus maximus, long head of biceps femoris and semimembranosus at the thigh level and the soleus, medial head of gastrocnemius and the tibialis anterior in the lower legs. In comparison with the MFM subtype of filaminopathy, lipomatous muscle alterations were not only more distinct in lower legs but also markedly more pronounced in distal parts of muscles [18]. Muscle biopsy findings Muscle biopsy features in filamin C myopathy largely depend on the site of mutation within the FLNc molecule and, more importantly, on the impact that different mutations have on its biophysical and biochemical properties [6, 18, 23]. Affected muscles from patients carrying mutations in the Ig-like domains of FLNc that lead to the expression of a toxic protein show typical features of MFM [23, 26, 28, 44, 48, 57]. General myopathological abnormalities vary from mild variation in fiber size and increased numbers of internal nuclei to more advanced degenerative abnormalities comprising muscle fiber atrophy and hypertrophy, fiber splitting, and fibro-fatty tissue proliferation depending on the stage of illness and the muscle examined (Fig. 2). Muscle fibers undergoing necrosis and phagocytosis are observed, but usually not as a prominent phenomenon. Rimmed and non-rimmed vacuoles are commonly seen. Non-rimmed vacuoles are often marked with strong PAS-positivity. Additionally, increased acid phosphatase activity is observed in some fiber areas. Oxidative activity is partially reduced in some fiber areas resulting in core-like lesions, but rubbed-out fibers are rarely seen. ATPase staining reveals type 1 fiber predominance in the majority of cases. Typically, muscle fibers contain polymorphous cytoplasmic inclusions that correspond to protein aggregates. These are observed as single or multiple plaque-like formations within the cytoplasm, convoluted serpentine inclusions of varying thickness, granular deposits and spheroid bodies. The aggregates are eosinophilic on HE stain, dark-blue to purple on modified trichrome stain and mostly display strong congophilia when congo red stained sections are visualized in rhodamine optics. Oxidative and ATPase activities are partially decreased in fiber regions containing inclusions but oxidative activity is usually enhanced at the periphery. This reflects the absence of mitochondria within the inclusions and increased numbers of them at the periphery. The inclusions can be focal or diffuse occupying the entire cross-section of the fiber; furthermore, abnormal fibers usually show an uneven distribution across the fascicles. Immunohistochemical and immunofluorescence analyses (Fig. 2) show strong positivity for FLNc, myotilin, desmin, the four sarcoglycans, αB-crystallin, BAG3, Xin and multiple other proteins in areas corresponding to protein aggregates [22, 23, 44, 57]. Moreover, proteins involved in protein degradation pathways including heat shock proteins, subunits of the ubiquitin proteasome system and markers of autophagy such as LAMP2 accumulate in areas corresponding to protein deposits (Fig. 2) [23]. Ultrastructural analyses show widespread myofibrillar abnormalities, including Z-disc streaming and accumulation of fine thin filaments that initially emanate at the level of the Z-disc and later coalesce into electron dense inclusions often surrounded by groups of mitochondria. Additionally, nemaline bodies and collections of 15–18 nm tubulofilaments and granulofilamentous material are seen in severely damaged fibers. Autophagic vacuoles containing myelin-like figures and cellular debris are usually present [22, 23, 28, 44]. Two families carrying mutations in the actin-binding domain of FLNc have been reported so far [6]. Muscle biopsies in four affected patients only showed non-specific myopathic features that varied from mild variation of fiber size to more severe dystrophic changes with prominent fibro fatty tissue proliferation. Many fibers showed an uneven distribution of oxidative enzyme activity, but no vacuoles and no protein aggregates were observed. Ultra-structural analysis performed in a single patient revealed no abnormalities [58]. Finally, muscle biopsy from a patient with FLNC haploinsufficiency showed increased variability of fiber size, fiber splitting and pyknotic nuclear clumps. ATPase revealed type I fiber predominance. Oxidative enzyme activity was partially reduced in some fiber areas. Although a few fibers displayed few fine myotilin granular deposits, no definite protein aggregates suggestive of MFM were detected, probably because the truncated mutant protein is not expressed. Ultrastructural analysis revealed some unspecific myofibrillar abnormalities including Z-disc streaming, nemaline bodies, and dappled dense bodies all of which are also observed in patients with the MFM type of filaminopathy [18]. Genetics Affected gene and its structure FLNc is a filamin isoform mainly expressed in striated muscles; it contains 2,725 amino acids and has a molecular mass of 291 kDa (GenBank isoform a: NP_001449.3). The FLNC gene is located in 7q32-q35 chromosome band, comprises ~29.5 kb of genomic DNA and contains 49 coding exons [4, 13, 29]. Since the exon encoding the FLNC-specific unique insert in Ig-like domain 20 was numbered 40a, the last FLNC exon carries number 48, not 49 [4]. Conversely, the splice variant that is predominantly expressed in skeletal and cardiac muscles (GenBank isoform b: NP_001120959) lacks exon 32 that encodes the hinge region between Ig-like domains 15 and 16, resulting in a protein of 2,692 amino acids with a molecular mass of 287 kDa [60]. Molecular diagnosis of filaminopathy is hampered by the presence of a pseudogene (pseFLNC) located approximately 53.6 kilobases downstream of the functional FLNC gene in inverted orientation. It is 1,178 base pairs in length and >98 % identical to the functional FLNC exons 46, 47, 48 (including part of the 3′ untranslated region), as well as introns 45 (partly), 46 and 47. In a recent work [31], DNA sequence mismatches between the functional FLNC and pseFLNC have been fully characterized, and an optimized strategy was devised enabling the differentiation of mutations occurring in FLNC from those accumulating in pseFLNC. Reflecting on the difficulty of differentiating between mutations in the functional gene and the pseudogene, some results of FLNC gene studies have been erroneous, as for example a report implicating a c.8107delG variant as the cause of filaminopathy in six patients [24]. The authors tested FLNC exon 48 with primers that amplify both the functional gene and the pseudogene, and the c.8107delG that is present in the pseudogene was misinterpreted as the cause of illness. This mistake could have been avoided, if the functional gene and the pseudogene were sequenced separately [31, 55]. Mutation spectrum The first FLNc-related disease was described in 2005 when a nonsense mutation (c.G8130A, p.W2710X) in the FLNc dimerization domain was shown to cause skeletal and cardiac myopathy in a large German MFM family [57]. A haplotype-sharing set of further German families also carrying the p.W2710X FLNc mutation was described soon after the first report [22], and the identical mutation was found in three kinships of the Mayo MFM cohort that were not described in detail [41], as well as in two further families from Macedonia and China [23]. These observations established that the p.W2710X mutation is the cause of filaminopathy in genetically unrelated families originating from different ethnic groups, implying that FLNC codon 2710 is a mutational hotspot. Two families with filamin C myopathy harboring mutations in FLNc Ig-like domain 7 of rod 1 segment have been reported: one harboring an internal 12-nucleotide deletion (c.2997_3008del, p.V930_T933del) [44] and a second exhibiting an 18-nucleotide deletion/6 nucleotide insertion (c.2695–2712del/GTTTGT ins, p.K899_V904del/V899_C900ins) [28]. In addition, an MFM family with a p.Y1216N mutation in Ig-like domain 10 and a single patient with proximal weakness at presentation and MFM-type pathology harboring a c.C7256T, p.T2419M mutation located in FLNc Ig-like domain 22 were recently described [2, 48]. In three distantly related Bulgarian distal myopathy families, a deletion (c.5160delC, p.F1720LfsX63) in exon 30 encoding FLNc Ig-like domain 15 triggers a frameshift, nonsense-mediated decay and haploinsufficiency [18]. Finally, a disorder caused by two different point mutations located in the ABD domain of FLNc (c.577G>A, p.A193T and c.752T>C, p.M251T) has also been associated with a distal myopathy with non-specific myopathic abnormalities on muscle biopsy [6]. Currently known FLNc mutations are shown on mutation chart of Fig. 3 and in Table 1. This clearly differentiates two FLNc associated phenotypes, one with involvement of predominantly limb-girdle muscles, cardiomyopathy, respiratory failure, and MFM-type pathologies caused by mutations occurring in FLNc Ig-like domains 7, 10, 22 and 24; the other characterized by myopathy seen in distal muscles, no cardiomyopathy or respiratory disturbances, and no typical MFM-type pathology with mutations in the ABD or Ig-like domain 15. This indicates that, as has previously been shown for FLNa and FLNb [3, 7, 37], mutations in different functional domains of filamins can lead to distinct disease phenotypes. Molecular diagnostics Timely molecular diagnosis of filaminopathy is important for the prediction and prevention of life-threatening cardiac arrhythmias and respiratory failure that may occur in these patients. A precise diagnosis is also crucial for appropriate counseling. Routine testing of patients for FLNC mutations should be recommended in cases showing limb-girdle distribution of weakness and MFM-type pathological phenomena. Since exon 48 is a hot spot for mutations causing this type of disease, it should be analyzed first. A mutation in the ABD domain of FLNc (exons 1–3 and part of 4) needs to be considered in patients with distal myopathy, especially if thenar and intrinsic hand muscle atrophy are the first clinical symptoms and the family history is consistent with an autosomal dominant pattern of inheritance. This will have to be followed by a full FLNC sequencing in case of a negative result. Pathophysiology Protein expression and function The mammalian filamin family includes three members, A, B and C (FLNa, FLNb and FLNc). They exhibit about 70 % amino acid identity [35]. Whereas Northern blots only detect FLNC mRNA in striated muscles [29], the more sensitive RT-PCR analysis reveal low levels of FLNC expression in multiple other tissues [60]. The lack of FLNc-specific antibodies hampered analysis at the protein level. Staining of multi-tissue slides with an antibody against the carboxy-terminus of FLNc by “the Human Protein Atlas” (http://www.proteinatlas.org/) [51] showed that apart from skeletal and cardiac myocytes, smooth muscle cells, glandular cells and neuronal cells in several tissues are also stained. Filamins are large proteins that bind to actin and many other proteins (Fig. 4) having diverse physiological functions. Through these connections, filamins stabilize delicate three-dimensional actin filament networks and link it to cellular membranes, thus integrating cell architectural and signaling functions. All three filamin variants bind a plethora of proteins, in particular via their carboxy-terminal Ig-like domains [9, 35, 46, 52]. Some of these interactions may be irrelevant in vivo because of differential expression patterns or significantly differing binding affinities. A prediction of interactions based on simple extrapolation is, therefore, highly questionable. FLNc binds essentially two groups of ligands: (1) proteins of the Z-disc including myotilin [17, 56], myopodin [25], the calsarcins [8, 11, 47] and nebulette [21]; (2) sarcolemma-associated proteins such as dystrophin-associated proteins γ-and δ-sarcoglycan [49], the NRAP-talin complex [27], the ponsin-Nck2 complex [14, 61] and β1A integrin [17]. Although often suggested to occur, a direct interaction with the costameric β1D-integrin isoform has been excluded [17]. This implies that only during early developmental stages, FLNc–β1A-integrin interaction may be involved in membrane anchorage of (pre)myofibrils. In mature muscle cells, FLNc most likely indirectly links myofibrils to sarcolemma-associated integrins via the above-mentioned protein complexes. At the same time, FLNc mediates assembly of Z-discs through its interaction with several Z-disc components. Early expression of FLNc during myofibril assembly and its localization to Z-bodies are in line with the proposed role for FLNc in this process [54]. It was also suggested that by shuttling between the sarcolemma and the Z-disc FLNc is involved in signal transduction processes [49, 56]. Protein structure Filamins consist of an aminoterminal actin-binding domain composed of two calponin homology (CH) domains followed by 24 immunoglobulin-like (Ig-like) domains of 93–103 amino acid residues each (Fig. 3). Two filamin molecules form a homodimer via self-association of their Ig-like domains 24, thus giving rise to large, elongated, Y-shaped molecules that form flexible bridges between two actin filaments [20, 36, 45]. Ig-like domains form an extended rod separated into two segments by hinge regions located between Ig-like domains 15 and 16 and Ig-like domains 23 and 24. However, the FLNc variant that is predominantly expressed in cross-striated muscles lacks the first hinge, implying that this isoform is less flexible than FLNa, and FLNb and FLNc isoforms containing this hinge region [60]. A conspicuous difference between FLNc and the other filamins is a unique insertion in Ig-like domain 20 that is involved in interaction with the Xin-repeat proteins Xin [53] and XIRP2 (unpublished data). Biochemical and biophysical analysis of mutant proteins Studies performed in animal models and the data obtained from human patients with haploinsufficiency for FLNc (see below) have demonstrated that the precise stoichiometry of FLNc is of critical importance for muscle function and maintenance. Of particular interest is, therefore, a study that has unraveled a new pathway essential for muscle maintenance that was termed ‘chaperone-assisted selective autophagy’ (CASA). This pathway, which includes the co-chaperone BAG3, the ubiquitin ligase CHIP, the autophagy adaptor p62 and DNAJB6 [1, 38], is essential for the homeostasis of certain proteins, including FLNc [1]. CASA continuously operates at the Z-disc to dispose of mechanically damaged proteins, which distinguishes it from the atrophy-driven degradation pathways [59]. Impairment of this pathway leads to the formation of FLNc-containing protein aggregates, Z-disc disintegration and progressive muscle weakness [1, 19, 38, 43]. In the case of FLNc mutations causing partial protein destabilization, incorrectly folded and damaged mutant FLNc directly drives protein aggregation, thus aggravating CASA (Fig. 5). Indeed, molecular components of CASA were found to be increased in biopsies from such patients (Fig. 2) [23]. Recently, mutations in the DNAJB6 gene have been identified as a cause of limb-girdle muscular dystrophy (LGMD1D) [19, 38]. Pathomechanisms Since the first description of a family with a FLNC mutation causative for a muscle disease, several additional mutations have been found in different parts of the gene. Grossly these mutations can be subdivided in three classes (Fig. 6): Mutations that lead to the expression of misfolded FLNc, thereby overstraining the ubiquitin proteasome and autophagy pathways in the long run; Mutations that do not affect protein solubility properties but give rise to a toxic gain of function by altering ligand binding properties; Mutations causing a premature stop codon and concomitant nonsense-mediated decay, resulting in haploinsufficiency. While the first type of mutations results in protein aggregation and subsequent impairment of protein homeostasis, giving rise to the typical MFM phenotype, the other two types of mutations result in distal myopathy with no protein aggregates. Although until now aggregation-causing mutations have only been found in Ig-like domains, such mutations may also occur in other portions of the molecule. Thus far, MFM-causing mutations in Ig-like domains 7 (p.V930_T933del) and 24 (p.W2710X) have been analyzed at the biochemical and cellular level [23, 26, 57]. Since the latter mutation is localized in the last exon (exon 48) of FLNC [57], the mutant mRNA is stable and not prone to degradation by nonsense-mediated decay. The affected part of FLNc is its dimerization domain that is truncated and lacks the carboxyterminal 16 amino acids. Circular dichroism spectroscopy showed that the mutant domain is improperly folded, making it less stable and more susceptible to proteolysis. Hence, the p.W2710X mutation in FLNc impedes its ability to dimerize [26, 57] and instead, the mutant protein acquires a strong tendency for uncontrolled aggregation (Fig. 6a). This results in the deposition of massive protein aggregates that attract multiple other proteins including desmin and other Z-disc-associated proteins. These events ultimately lead to disintegration of myofibrils [26, 57]. The deletion of four amino acids in the β-strand of Ig-like domain 7 that is involved in interactions stabilizing the fold, also causes significant changes in the three-dimensional structure of the mutant domain, as illustrated by a higher proportion of unfolded or disordered structures, reduced stability and increased protease sensitivity (Fig. 6b) [23]. The mutations in the ABD of FLNc that were found in two distal myopathy families apparently cause only minimal structural alterations. Amino acid substitutions, however, are predicted to alter intradomain interactions, thereby facilitating binding to actin and increasing its binding constant (Fig. 6c) [6]. In contrast to the nonsense mutation in Ig-like domain 24, the frameshift deletion mutation in Ig-like domain 15 (c.5160delC, p.F1720LfsX63) does not occur in the last exon and thus activates nonsense-mediated decay of the mutant mRNA [18]. Since no truncated protein could be detected in the patients’ muscles, the 50 % reduction of FLNC mRNA and protein levels is the most probable reason for the disease phenotype (Fig. 6d). The lack of expression of mutant protein precludes the development of major protein aggregates that are a hallmark of MFM. Cell and animal models Cell models The effects of the expression of truncated and full length mutated FLNc constructs were analyzed in tissue culture. Initially, the expression of mutant p.W2710X “mini-filamins” consisting of the ABD and Ig-like domains 15–24 was shown to be sufficient for spontaneous aggregation of the mutant protein in cultured cells [26]. The same effect was found upon transfection of full-length p.W2710X and p.V930_T933del FLNc in C2C12 mouse myoblasts [23]. The up to ten times higher number of transfected cells showing mutant FLNc aggregates in cells transfected with p.W2710X protein indicated that this mutant makes the cells more vulnerable to spontaneous aggregation. Transfection of the FLNc variants p.M251T and p.A193T also resulted in the development of protein aggregates in transfected non-muscle and muscle cells. Many of these aggregates also contained F-actin [6]. These cell models might become valuable tools to study the mechanisms of protein aggregation and evaluate treatments that prevent or reverse this phenomenon. shRNA constructs in lentiviral vectors were used to generate a C2C12 cell line with a reduction in the level of Flnc mRNA of 93 % [5]. These cells, that expressed only very low levels of FLNc protein, proliferate and fuse normally, but instead of developing long myotubes, cells round up and form multinucleate myoballs upon fusion. This process is associated with a decrease in the expression of myogenin and muscle-specific genes, indicating a direct effect on myogenesis. For these experiments, a cell line showing the highest reduction in Flnc expression was selected. A variant with a more modest knock-down efficiency of Flnc expression could provide a model system for human diseases associated with Flnc haploinsufficiency. Animal models Mouse model The only Flnc mouse model that has been created thus far is B6;129-Flnctm1Lmk/J. In these mice, the last eight exons (exons 41–48) of the Flnc gene were deleted by targeted mutation [5]. This mutation results in the expression of reduced levels of truncated mRNA and very low levels of truncated FLNc protein consisting of the ABD and Ig-like domains 1–19 and part of Ig-like domain 20 that is truncated after the FLNc-unique insertion. Mice homozygous for the knockout allele die at birth due to the inability to breath caused by severe abnormalities of their skeletal muscles. Whereas the heart has a normal appearance, the development of skeletal muscles is grossly disturbed, leading to reduced numbers of muscle fibers, often containing centrally located nuclei. Specifically, intercostal muscles and the diaphragm showed infiltration of connective tissue. Heterozygous mice were viable and fertile, and no abnormalities were reported, indicating that neither the low level of truncated FLNc nor the reduction of the level of wildtype FLNc results in an obvious phenotype [5]. Unfortunately, these mice apparently were not analyzed at older age. Since in man FLNC haploinsufficiency leads to distal myopathy at an average age of approximately 40 years [18], these mice might be a valuable model for this disease. Medaka A mutation in one of two flnc genes of medaka (Oryzias latipes, a teleost fish) was identified in zacro (zac) mutants [12]. This strain, which was obtained by N-ethyl-N-nitrosourea treatment, is characterized by disorganization of skeletal muscle fibers and abnormal development of the heart associated with a rupture of the myocardial layer. The causative mutation was found to be a nucleotide substitution in one of the flnc genes, leading to the introduction of a stop codon that would result in the expression of an FLNc variant that is truncated in Ig-like domain 15. Fish heterozygous for the mutation developed normally. All embryos showing the zac phenotype were homozygous for the mutant allele. Instability of the mutant mRNA most likely leads to nonsense-mediated decay and significantly lower the level of flnc mRNA. Although expression of a truncated protein was not analyzed, the lack of FLNc and not the expression of a toxic protein seems to be the most probable explanation for the observed phenotype. This was supported by morpholino-based antisense RNA experiments that resulted in a similar phenotype, at least in the heart [12] which explains why no alterations typical for MFM were found in the muscle of the zacro mutant fish. Heterozygous fish that are expected to be haploinsufficient for flnc should be studied at late adult age to conform to a model for distal myopathy caused by haploinsufficiency. Future perspectives for research and therapy The continuing search for mutations in FLNC will certainly result in the identification of more and more disease-associated mutations. Whereas the pathomechanisms of the mutations described thus far are roughly explained, it will be interesting to learn whether mutations in the rod2 segment have an impact on the association of FLNc with its many ligands, and how such a defect would influence the disease phenotype. The main goal of future research should be the search for approaches to prevent the formation of aggregates in the muscle fibers of filaminopathy (and other MFM) patients. A promising strategy may be the induction of chaperones. For this purpose, appropriate cell and animal models are needed. Preliminary studies in cultured muscle cells indicate that transfection with mutant FLNc leads to aggregate formation. Cell lines stably transfected with constructs expressing mutant FLNc might, therefore, be an invaluable tool. Alternatively, skeletal muscle satellite cells from filaminopathy patients could be immortalized for such studies. Finally, patient fibroblasts could be converted to embryonic stem cells and be forced to develop into skeletal muscle cells. The currently existing FLNC-related animal models only represent the type of filaminopathy associated with reduced expression levels but not the protein aggregation phenotype. To allow for testing therapeutic interventions in patients with the MFM type of filaminopathy, the development of an animal model would be of great value. A prime candidate would be a knock-in mouse carrying, e.g., the c.G8130A, p.W2710X mutation in one allele, since this represents the most frequent type of human filaminopathy. This research was supported in part by the Intramural Research Program of the National Institute of Neurological Disorders and Stroke, NIH [L.G.G], the German Research foundation [KL 2487/1-1 to R.A.K., FOR1228 to M.V., D.O.F., FOR1352 to D.O.F.], the German Ministry of Education and Research [01GM0887 to R.A.K., P.F.M.v.d.V., M.V., D.O.F.], the Ruhr-University Bochum [FoRUM K042-09 to R.A.K.], and the Spanish Instituto de Salud Carlos III [PI08-574 to M.O.]. Fig. 1 Image and transverse T1-weighted muscle MRI of MFM patients harboring mutations in the dimerization domain of FLNc. The patient (a) has predominantly proximal muscle atrophy in the upper and lower limbs and winged scapula. MRI images demonstrate a typical pattern of muscle involvement (hyperintensities reflect lipomatous alterations). On the thigh level (b, c), semimembranosus (SM), adductor magnus (AM) and longus (AL), long head of biceps femoris (BF), vastus intermedius (VI) and medialis (VM) are most affected. Sartorius (SA) and gracilis (GR) appear normal and the semitendinosus (ST) shows only mild lipomatous alterations. In lower legs (d), the soleus muscle (SO) shows pronounced fatty changes. The medial head of the gastrocnemius (GM), the tibialis anterior (TA), extensor digitorum longus, extensor hallucis longus (ED/EH), and peroneal muscles show mild to moderate lipomatous alterations whereas the lateral head of the gastrocnemius (GL) is almost spared Fig. 2 Histochemical and immunofluorescence findings in MFM-type FLNc myopathy. Myopathic changes vary from mild variability in fiber size (a), to more pronounced fiber size variation, with atrophic and hypertrophic fibers, and moderate (b) to severe (c) fibro-fatty tissue proliferation. Abnormal fibers show darkly stained areas and small vacuoles (d); ATPase staining reveals type I fiber predominance (e); COX activity is partially reduced in some fiber regions (f). Immunofluorescence analysis showing accumulation of FLNc (g), myotilin (h), δ-sarcoglycan (i), Hsp22 (j), ubiquitin (k) and LAMP-2 (l) in fiber regions corresponding to protein aggregates. a–c HE; d modified trichrome; e ATPase at pH 4,65; f COX. Scale bar in c (also applies to a, b) 100 μm, scale bar in l (also applies to d–k) 50 μm Fig. 3 Schematic diagram representing the structure of FLNc and the distribution of muscle-disease-associated mutations in FLNc. The aminoterminus consists of two calponin homology domains that together constitute the actin-binding domain (ABD). The ABD is followed by 24 Ig-like domains, the most carboxy terminal of which is responsible for dimerization. Ig-like domain 20 is colored differently since it contains a unique insertion. The positions of the published mutations within FLNc are depicted at the top Fig. 4 Schematic diagram illustrating the complex interactome of FLNc within the Z-disc and at the sarcolemma. Each protein is depicted as an ellipse and direct protein interactions are depicted as connecting lines Fig. 5 FLNc protein homeostasis. Upon increased muscle activity, damaged FLNc is degraded by BAG3 and CHIP-regulated chaperone-assisted selective autophagy (CASA). a The scheme summarizes the mechanism of selective FLNc release from the Z-disc by BAG3, resulting in ubiquitination, subsequent autophagosome formation and lysosomal degradation. b In LAMP2−/− mice autophagy is blocked and FLNc no longer localizes at Z-discs but instead forms massive aggregates (arrows). Reproduced with permission from [1] Fig. 6 Pathomechanisms of FLNc mutations. a Top panel disturbed dimerization of pW2710X FLNc (top) revealed by chemical cross-linking experiments using wild type (d23–24) and mutant (d23–24mut) filamin constructs. In the presence (+) of the cross-linker EGS, the wild type protein was mainly found in dimer form (d), while the mutant construct was detected in higher molecular mass complexes representing trimers (t) and aggregated (a) oligomers. a Bottom panel analytical gel chromatography shows greatly decreased retention times for the mutant FLNc d23–24 protein indicating aggregation of the mutant but not the wild-type protein. b Decreased stability and increased protease-sensitivity of the p.V930_T933del mutant FLNc. The top panel shows the digestion of wild type and mutant FLNc d5-9ΔVKYT with the protease thermolysin, resulting in complete digestion of the mutant protein after 30 min, while a significant portion of the wild type variant was still intact after 60 min of incubation, indicating less stable folding of the mutant protein. The bottom panel shows temperature denaturation experiments for FLNc d7-8 and deletion mutant FLNc d7-8ΔVKYT. The melting temperature Tm is determined at the inflection point of the fluorescence signal, the first derivative of which is reported here. Solid line represents the wild type, dashed line the deletion mutant, scaled to the wild-type. Note that the mutant protein has a significantly reduced melting temperature in comparison to wild type protein. c Mutant FLNc ABDs have similar structure but stronger actin-binding affinity. The top panel gives circular dichroism spectra obtained with wild-type and mutant filamin ABD constructs, demonstrating that wild-type and p.A193T mutant spectra are almost identical and only minor changes for the p.M251T mutant, indicating only minor structural effects of these mutations. The bottom panel gives high-speed F-actin cosedimentation assays with FLNc wild-type and mutant ABDs, showing increased actin binding activity of the mutant proteins. d Decreased FLNC mRNA and protein levels in patients with the p.F1720LfsX63 mutation. Quantitative FLNC transcript and protein analysis in muscle tissue from patients (Pt) and controls (Ctrl) reveals an approximately 50 % reduction in FLNC mRNA and protein levels in muscle samples from patients. An antibody directed against N-terminal FLNc did not detect expression of a truncated FLNc protein (~190 kDa). Reproduced with permission or adapted from a [26], b [23], c [6] and d [18] Table 1 Summary of the reported mutations in FLNc Mutation Domain Type Disease References p.A193T ABD Point mutation Distal myopathy Duff et al. [6] p.M251T ABD Point mutation Distal myopathy Duff et al. [6] p.K899_V904del/V899_C900ins Ig-like 7 Deletion/insertion MFM, gastrointestinal complaints Luan et al. [28] p.V930_T933del Ig-like 7 Deletion MFM Shatunov et al. [44] p.Y1216N Ig-like 10 Point mutation MFM Avila-Smirnov et al. [2] p.F1720LfsX63 Ig-like 15 Frameshift deletion Distal myopathy Guergueltcheva et al. [18] p.T2419M Ig-like 22 Point mutation MFM, ataxia Tasca et al. [48] p.W2710X Ig-like 24 Nonsense MFM Vorgerd et al. [57] Conflict of interest The authors declare that they have no conflict of interest. 1 Arndt V Dick N Tawo R Dreiseidler M Wenzel D Hesse M Fürst DO Saftig P Saint R Fleischmann BK Hoch M Höhfeld J 2010 Chaperone-assisted selective autophagy is essential for muscle maintenance Curr Biol 20 143 148 20060297 2 Avila-Smirnov D Béhin A Gueneau L Claeys K Beuvin M Goudeau B Richard P Ben Yaou R Romero NB Mathis S Voit T Eymard B Gil R Fardeau M Bonne G 2010 A novel missense FLNC mutation causes arrhythmia and late onset myofibrillar myopathy with particular histopathology features Neuromuscul Disord 20 623 624 3 Bicknell LS Farrington-Rock C Shafeghati Y Rump P Alanay Y Alembik Y Al-Madani N Firth H Karimi-Nejad MH Kim CA Leask K Maisenbacher M Moran E Pappas JG Prontera P de Ravel T Fryns JP Sweeney E Fryer A Unger S Wilson LC Lachman RS Rimoin DL Cohn DH Krakow D Robertson SP 2007 A molecular and clinical study of Larsen syndrome caused by mutations in FLNB J Med Genet 44 89 98 16801345 4 Chakarova C Wehnert MS Uhl K Sakthivel S Vosberg HP van der Ven PFM Fürst DO 2000 Genomic structure and fine mapping of the two human filamin gene paralogues FLNB and FLNC and comparative analysis of the filamin gene family Hum Genet 107 597 611 11153914 5 Dalkilic I Schienda J Thompson TG Kunkel LM 2006 Loss of FilaminC (FLNc) results in severe defects in myogenesis and myotube structure Mol Cell Biol 26 6522 6534 16914736 6 Duff RM Tay V Hackman P Ravenscroft G McLean C Kennedy P Steinbach A Schöffler W van der Ven PFM Fürst DO Song J Djinović-Carugo K Penttilä S Raheem O Reardon K Malandrini A Gambelli S Villanova M Nowak KJ Williams DR Landers JE Brown RH Jr Udd B Laing NG 2011 Mutations in the N-terminal actin-binding domain of filamin C cause a distal myopathy Am J Hum Genet 88 729 740 21620354 7 Farrington-Rock C Firestein MH Bicknell LS Superti-Furga A Bacino CA Cormier-Daire V Le Merrer M Baumann C Roume J Rump P Verheij JB Sweeney E Rimoin DL Lachman RS Robertson SP Cohn DH Krakow D 2006 Mutations in two regions of FLNB result in atelosteogenesis I and III Hum Mutat 27 705 710 16752402 8 Faulkner G Pallavicini A Comelli A Salamon M Bortoletto G Ievolella C Trevisan S Kojic S Dalla VF Laveder P Valle G Lanfranchi G 2000 FATZ: a filamin, actinin, and telethonin binding protein of the Z-disk of skeletal muscle J Biol Chem 275 41234 41242 10984498 9 Feng Y Walsh CA 2004 The many faces of filamin: a versatile molecular scaffold for cell motility and signalling Nat Cell Biol 6 1034 1038 15516996 10 Fischer D Kley RA Strach K Meyer C Sommer T Eger K Rolfs A Meyer W Pou A Pradas J Heyer CM Grossmann A Huebner A Kress W Reimann J Schröder R Eymard B Fardeau M Udd B Goldfarb L Vorgerd M Olivé M 2008 Distinct muscle imaging patterns in myofibrillar myopathies Neurology 71 758 765 18765652 11 Frey N Olson EN 2002 Calsarcin-3, a novel skeletal muscle-specific member of the calsarcin family, interacts with multiple Z-disc proteins J Biol Chem 277 13998 14004 11842093 12 Fujita M Mitsuhashi H Isogai S Nakata T Kawakami A Nonaka I Noguchi S Hayashi YK Nishino I Kudo A 2012 Filamin C plays an essential role in the maintenance of the structural integrity of cardiac and skeletal muscles, revealed by the medaka mutant zacro Dev Biol 361 79 89 22020047 13 Gariboldi M Maestrini E Canzian F Manenti G De Gregorio L Rivella S Chatterjee A Herman GE Archidiacono N Antonacci R 1994 Comparative mapping of the actin-binding protein 280 genes in human and mouse Genomics 21 428 430 8088838 14 Gehmlich K Hayeß K Legler C Haebel S van der Ven PFM Ehler E Fürst DO 2010 Ponsin interacts with Nck adapter proteins: implications for a role in cytoskeletal remodelling during differentiation of skeletal muscle cells Eur J Cell Biol 89 351 364 20129698 15 Goebel HH 2003 Congenital myopathies at their molecular dawning Muscle Nerve 27 527 548 12707973 16 Goebel HH 2009 Protein aggregate myopathies. Introduction Brain Pathol 19 480 482 19563539 17 Gontier Y Taivainen A Fontao L Sonnenberg A van der Flier A Carpén O Faulkner G Borradori L 2005 The Z-disc proteins myotilin and FATZ-1 interact with each other and are connected to the sarcolemma via muscle-specific filamins J Cell Sci 118 3739 3749 16076904 18 Guergueltcheva V Peeters K Baets J Ceuterick-de Groote C Martin JJ Suls A De Vriendt E Mihaylova V Chamova T Almeida-Souza L Ydens E Tzekov C Hadjidekov G Gospodinova M Storm K Reyniers E Bichev S van der Ven PFM Fürst DO Mitev V Lochmüller H Timmerman V Tournev I De Jonghe P Jordanova A 2011 Distal myopathy with upper limb predominance caused by filamin C haploinsufficiency Neurology 77 2105 2114 22131542 19 Harms MB Sommerville RB Allred P Bell S Ma D Cooper P Lopate G Pestronk A Weihl CC Baloh RH 2012 Exome sequencing reveals DNAJB6 mutations in dominantly-inherited myopathy Ann Neurol 71 407 416 22334415 20 Himmel M van der Ven PFM Stöcklein W Fürst DO 2003 The limits of promiscuity: isoform-specific dimerization of filamins Biochemistry 42 430 439 12525170 21 Holmes WB Moncman CL 2008 Nebulette interacts with filamin C Cell Motil Cytoskeleton 65 130 142 17987659 22 Kley RA Hellenbroich Y van der Ven PFM Fürst DO Huebner A Bruchertseifer V Peters SA Heyer CM Kirschner J Schröder R Fischer D Müller K Tolksdorf K Eger K Germing A Brodherr T Reum C Walter MC Lochmüller H Ketelsen UP Vorgerd M 2007 Clinical and morphological phenotype of the filamin myopathy: a study of 31 German patients Brain 130 3250 3264 18055494 23 Kley RA Serdaroglu-Oflazer P Leber Y Odgerel Z van der Ven PFM Olivé M Ferrer I Onipe A Mihaylov M Bilbao JM Lee HS Höhfeld J Djinović-Carugo K Kong K Tegenthoff M Peters SA Stenzel W Vorgerd M Goldfarb LG Fürst DO 2012 Pathophysiology of protein aggregation and extended phenotyping in filaminopathy Brain 135 2642 2660 22961544 24 Kono S Nishio T Takahashi Y Goto-Inoue N Kinoshita M Zaima N Suzuki H Fukutoku-Otsuji A Setou M Miyajima H 2010 Dominant-negative effects of a novel mutation in the filamin myopathy Neurology 75 547 554 20697107 25 Linnemann A van der Ven PFM Vakeel P Albinus B Simonis D Bendas G Schenk JA Micheel B Kley RA Fürst DO 2010 The sarcomeric Z-disc component myopodin is a multiadapter protein that interacts with filamin and alpha-actinin Eur J Cell Biol 89 681 692 20554076 26 Löwe T Kley RA van der Ven PFM Himmel M Huebner A Vorgerd M Fürst DO 2007 The pathomechanism of filaminopathy: altered biochemical properties explain the cellular phenotype of a protein aggregation myopathy Hum Mol Genet 16 1351 1358 17412757 27 Lu S Carroll SL Herrera AH Ozanne B Horowits R 2003 New N-RAP-binding partners alpha-actinin, filamin and Krp1 detected by yeast two-hybrid screening: implications for myofibril assembly J Cell Sci 116 2169 2178 12692149 28 Luan X Hong D Zhang W Wang Z Yuan Y 2010 A novel heterozygous deletion-insertion mutation (2695–2712 del/GTTTGT ins) in exon 18 of the filamin C gene causes filaminopathy in a large Chinese family Neuromuscul Disord 20 390 396 20417099 29 Maestrini E Patrosso C Mancini M Rivella S Rocchi M Repetto M Villa A Frattini A Zoppe M Vezzoni P 1993 Mapping of two genes encoding isoforms of the actin binding protein ABP- 280, a dystrophin like protein, to Xq28 and to chromosome 7 Hum Mol Genet 2 761 766 7689010 30 Nakano S Engel AG Waclawik AJ Emslie-Smith AM Busis NA 1996 Myofibrillar myopathy with abnormal foci of desmin positivity. I. Light and electron microscopy analysis of 10 cases J Neuropathol Exp Neurol 55 549 562 8627346 31 Odgerel Z van der Ven PFM Fürst DO Goldfarb LG 2010 DNA sequencing errors in molecular diagnostics of filamin myopathy Clin Chem Lab Med 48 1409 1414 20578970 32 Ohlsson M Hedberg C Bradvik B Lindberg C Tajsharghi H Danielsson O Melberg A Udd B Martinsson T Oldfors A 2012 Hereditary myopathy with early respiratory failure associated with a mutation in A-band titin Brain 135 1682 1694 22577218 33 Olivé M Odgerel Z Martinez A Poza JJ Bragado FG Zabalza RJ Jerico I Gonzalez-Mera L Shatunov A Lee HS Armstrong J Maravi E Arroyo MR Pascual-Calvet J Navarro C Paradas C Huerta M Marquez F Rivas EG Pou A Ferrer I Goldfarb LG 2011 Clinical and myopathological evaluation of early- and late-onset subtypes of myofibrillar myopathy Neuromuscul Disord 21 533 542 21676617 34 Pfeffer G Elliott HR Griffin H Barresi R Miller J Marsh J Evila A Vihola A Hackman P Straub V Dick DJ Horvath R Santibanez-Koref M Udd B Chinnery PF 2012 Titin mutation segregates with hereditary myopathy with early respiratory failure Brain 135 1695 1713 22577215 35 Popowicz GM Schleicher M Noegel AA Holak TA 2006 Filamins: promiscuous organizers of the cytoskeleton Trends Biochem Sci 31 411 419 16781869 36 Pudas R Kiema TR Butler PJ Stewart M Ylänne J 2005 Structural basis for vertebrate filamin dimerization Structure (Camb) 13 111 119 15642266 37 Robertson SP Twigg SR Sutherland-Smith AJ Biancalana V Gorlin RJ Horn D Kenwrick SJ Kim CA Morava E Newbury-Ecob R Orstavik KH Quarrell OW Schwartz CE Shears DJ Suri M Kendrick-Jones J Wilkie AO 2003 Localized mutations in the gene encoding the cytoskeletal protein filamin A cause diverse malformations in humans Nat Genet 33 487 491 12612583 38 Sarparanta J Jonson PH Golzio C Sandell S Luque H Screen M McDonald K Stajich JM Mahjneh I Vihola A Raheem O Penttilä S Lehtinen S Huovinen S Palmio J Tasca G Ricci E Hackman P Hauser M Katsanis N Udd B 2012 Mutations affecting the cytoplasmic functions of the co-chaperone DNAJB6 cause limb-girdle muscular dystrophy Nat Genet 44 450 452 22366786 39 Schröder R Schoser B 2009 Myofibrillar myopathies: a clinical and myopathological guide Brain Pathol 19 483 492 19563540 40 Selcen D 2008 Myofibrillar myopathies Curr Opin Neurol 21 585 589 18769253 41 Selcen D 2011 Myofibrillar myopathies Neuromuscul Disord 21 161 171 21256014 42 Selcen D Bromberg MB Chin SS Engel AG 2011 Reducing bodies and myofibrillar myopathy features in FHL1 muscular dystrophy Neurology 77 1951 1959 22094483 43 Selcen D Muntoni F Burton BK Pegoraro E Sewry C Bite AV Engel AG 2009 Mutation in BAG3 causes severe dominant childhood muscular dystrophy Ann Neurol 65 83 89 19085932 44 Shatunov A Olivé M Odgerel Z Stadelmann-Nessler C Irlbacher K van Landeghem F Bayarsaikhan M Lee HS Goudeau B Chinnery PF Straub V Hilton-Jones D Damian MS Kaminska A Vicart P Bushby K Dalakas MC Sambuughin N Ferrer I Goebel HH Goldfarb LG 2009 In-frame deletion in the seventh immunoglobulin-like repeat of filamin C in a family with myofibrillar myopathy Eur J Hum Genet 17 656 663 19050726 45 Sjekloca L Pudas R Sjöblom B Konarev P Carugo O Rybin V Kiema TR Svergun D Ylänne J Djinović-Carugo K 2007 Crystal structure of human filamin C domain 23 and small angle scattering model for filamin C 23–24 dimer J Mol Biol 368 1011 1023 17379241 46 Stossel TP Condeelis J Cooley L Hartwig JH Noegel A Schleicher M Shapiro SS 2001 Filamins as integrators of cell mechanics and signalling Nat Rev Mol Cell Biol 2 138 145 11252955 47 Takada F Vander Woude DL Tong HQ Thompson TG Watkins SC Beggs AH Kunkel LM 2001 Myozenin: an α-actinin- and γ-filamin-binding protein of skeletal muscle Z lines Proc Natl Acad Sci USA 98 1595 1600 11171996 48 Tasca G Odgerel Z Monforte M Aurino S Clarke NF Waddell LB Udd B Ricci E Goldfarb LG 2012 Novel FLNC mutation in a patient with myofibrillar myopathy in combination with late-onset cerebellar ataxia Muscle Nerve 46 275 282 22806379 49 Thompson TG Chan YM Hack AA Brosius M Rajala M Lidov HG McNally EM Watkins S Kunkel LM 2000 Filamin 2 (FLN2). A muscle-specific sarcoglycan interacting protein J Cell Biol 148 115 126 10629222 50 Udd B 2011 Distal muscular dystrophies Handb Clin Neurol 101 239 262 21496636 51 Uhlén M Oksvold P Fagerberg L Lundberg E Jonasson K Forsberg M Zwahlen M Kampf C Wester K Hober S Wernerus H Björling L Ponten F 2010 Towards a knowledge-based Human Protein Atlas Nat Biotechnol 28 1248 1250 21139605 52 van der Flier A Sonnenberg A 2001 Structural and functional aspects of filamins Biochim Biophys Acta 1538 99 117 11336782 53 van der Ven PFM Ehler E Vakeel P Eulitz S Schenk JA Milting H Micheel B Fürst DO 2006 Unusual splicing events result in distinct Xin isoforms that associate differentially with filamin C and Mena/VASP Exp Cell Res 312 2154 2167 16631741 54 van der Ven PFM Obermann WMJ Lemke B Gautel M Weber K Fürst DO 2000 Characterization of muscle filamin isoforms suggests a possible role of γ-filamin/ABP-L in sarcomeric Z-disc formation Cell Motil Cytoskeleton 45 149 162 10658210 55 van der Ven PFM Odgerel Z Fürst DO Goldfarb LG Kono S Miyajima H 2010 Dominant-negative effects of a novel mutation in the filamin myopathy Neurology 75 2137 2138 21135393 56 van der Ven PFM Wiesner S Salmikangas P Auerbach D Himmel M Kempa S Hayeß K Pacholsky D Taivainen A Schröder R Carpén O Fürst DO 2000 Indications for a novel muscular dystrophy pathway. γ-filamin, the muscle-specific filamin isoform, interacts with myotilin J Cell Biol 151 235 248 11038172 57 Vorgerd M van der Ven PFM Bruchertseifer V Löwe T Kley RA Schröder R Lochmüller H Himmel M Koehler K Fürst DO Huebner A 2005 A mutation in the dimerization domain of filamin c causes a novel type of autosomal-dominant myofibrillar myopathy Am J Hum Genet 7 297 304 58 Williams DR Reardon K Roberts L Dennet X Duff R Laing NG Byrne E 2005 A new dominant distal myopathy affecting posterior leg and anterior upper limb muscles Neurology 64 1245 1254 15824355 59 Willis MS Schisler JC Portbury AL Patterson C 2009 Build it up-Tear it down: protein quality control in the cardiac sarcomere Cardiovasc Res 81 439 448 18974044 60 Xie Z Xu W Davie EW Chung DW 1998 Molecular cloning of human ABPL, an actin-binding protein homologue Biochem Biophys Res Commun 251 914 919 9791010 61 Zhang M Liu J Cheng A Deyoung SM Saltiel AR 2007 Identification of CAP as a costameric protein that interacts with filamin C Mol Biol Cell 18 4731 4740 17898075

PMC005xxxxxx/PMC5127263.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 9422762 20221 Magn Reson Imaging Clin N Am Magn Reson Imaging Clin N Am Magnetic resonance imaging clinics of North America 1064-9689 1557-9786 25476670 5127263 10.1016/j.mric.2014.08.006 NIHMS831102 Article 4D Flow MRI Applications for Aortic Disease Burris Nicholas S. MD Hope Michael D. MD * Department of Radiology and Biomedical Imaging, University of California, San Francisco, 505 Parnassus Avenue, Box 0628, San Francisco, CA 94143-0628, USA * Corresponding author. michael.hope@ucsf.edu 21 11 2016 5 10 2014 2 2015 29 11 2016 23 1 1523 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 4D flow Aorta Aortic aneurysm Bicuspid aortic valve Hemodynamic imaging Phase contrast MRI Flow imaging OVERVIEW OF AORTIC IMAGING Aortic disease is a broad term that encompasses related and sometimes overlapping conditions associated with substantial morbidity and mortality, including aortic aneurysm and dissection.1,2 Imaging has long been used to diagnose and monitor these processes, and along with advances in surgical technique, has markedly improved mortality over the last 50 years.3 Advances in MRI have led to a more sophisticated understanding of intrinsic and valve-related aortic pathology. These developments not only provide insight into underlying pathophysiology, but also may allow imaging to better predict disease progression and inform the timing of interventions. Hemodynamic imaging with MRI is among the most recent advances in aortic evaluation. The hemodynamic environment of the aorta is unique with extreme pressure variations, high flow rates, and complex flow patterns that exist in both normal and pathologic states. Phase contrast MRI (PC-MRI) has been utilized in routine conventional cardiac MRI for more than 2 decades to accurately measure these flow dynamics. However, the recent optimization of volumetric, time-resolved (cine) and 3-directional PC-MRI (4D Flow), has led to striking visualization of dynamic flow patterns and quantification of associated abnormal hemodynamics. In this article, we briefly review common clinical applications of 2-directional (2D) PC-MRI, and discuss the emerging applications of 4D Flow for the thoracic aorta. We aim to highlight ongoing research in the field of aortic 4D Flow imaging by focusing on promising quantitative hemodynamic markers of aortic disease. CURRENT CLINICAL 2-DIRECTIONAL FLOW APPLICATIONS Aortic blood flow imaging with MRI is conventionally performed in 2 dimensions (2D). This means that single-directional velocity data is captured with PC-MRI through or parallel to a prescribed imaging plane. In contrast, 4D Flow imaging captures 3-directional (3D) velocity data through a volume on interest. Currently, conventional 2D evaluation is routinely used in several clinical scenarios: aortic valve disease, cardiac shunt lesions and aortic coarctation. Echocardiography is commonly used for assessing aortic valvular disease. However, over the past decade, MRI has become increasingly popular because it provides several distinct advantages, including better evaluation of the ascending aorta, more reproducible measurements of regurgitation fraction, and quantitative measurement left ventricular size, function, and mass.4,5 For evaluation of aortic stenosis, transvalvular gradient is a key parameter for determining disease severity and need for intervention. Transvalvular gradient can be estimated with 2D PC-MRI by imaging parallel the aortic centerline, just beyond the valve, to determine peak jet velocity, which can then be used for estimation of the pressure gradient using the modified Bernoulli equation. However, low temporal resolution and suboptimal imaging plane placement can lead to underestimation of true pressure gradients. For aortic regurgitation, measurement of forward and backward flow volume through the ascending aorta can be used to calculate the regurgitant fraction. Conventional 2D PC-MRI can also be used to quantify cardiac shunt fractions and to assess the hemodynamic impact of aortic coarctation. Similar to aortic regurgitation, through plane flow volumes are measured in the aorta and main pulmonary artery, and the shunt ratio or Qp/Qs can be calculated. In the case of coarctation, acquisition planes are placed across the aorta just distal to the coarctation and at the diaphragm to quantify collateral flow through intercostal vessels, which increases with worsening coarctation severity.6 4-DIRECTIONAL FLOW Multidimensional MRI (eg, 4D Flow) has become an increasingly popular research tool over the past decade. There are several advantages compared with conventional 2D PC sequences: Free breathing, acquisition of 3D velocity data, no need for prospective 2D plane placement, and powerful flow visualization and quantification software. Currently, the main disadvantages of 4D Flow are the need for labor-intensive data after processing, greater susceptibility to motion artifacts, and longer scan times, although scan times have improved dramatically over the last 5 years. Data Acquisition, Reconstruction, and Fidelity 4D Flow datasets are acquired over multiple cardiac cycles: Using electrocardiographic gating, data are collected over many hundreds of heartbeats and then compiled to create a cine image of a complete, representative cardiac cycle. Without the use of scan acceleration techniques, datasets can take 1 hour to fully acquire, a prohibitively long time for routine applications. When clinical patients are considered, rather than a population of healthy research volunteers, scan time becomes a dominant concern. A significant amount of research effort has been dedicated to reducing 4D Flow scan times, with promising results reported using spiral (rather than conventional Cartesian) acquisitions and data undersampling techniques for acceleration, including compressed sensing. Many studies report scan times of less than 15 minutes, with some groups acquiring data in as few as 5 minutes.7,8 Data errors with PC-MRI acquisitions have been reported,9 but can be minimized by routine correction of eddy currents, gradient fields, and Maxwell phase effects. Data fidelity is an important consideration because 4D Flow data are subject to multiple artifacts. In addition to the technical issues listed, there are other sources of error related to acquisitions spanning numerous heart beats and the presence of underlying pathologic flow characteristics (eg, turbulence, helicity). Confirmation of 4D Flow data fidelity has largely relied on in vitro, in silico, and animal models.10,11 One in vivo approach for data verification relies on the pathline technique of flow visualization. Pathlines tend to be sensitive to artifact accumulation, because they are calculated by integration over time. Applying conservation-of-mass principles in a specific region of interest, pathlines can be used for data quality verification. Another metric for internal data quality control relies in measurement of aortic (Qs) and pulmonary (Qp) flow rates, which should be equal in the absence of shunts, and can help in the identification of erroneous data. Of note, several important flow parameters are relative to other internal measurements (eg, Qp/Qs, collateral flow, flow displacement) so that the absolute values of the velocity data are less important than relative internal consistency. Flow Visualization Several approaches can be used for 4D Flow data visualization, including vector plots and particle traces (ie, streamlines and pathlines), which are typically color coded for velocity data.12 Whereas a vector plot represents the actual velocity data at a given moment in time, streamlines are imaginary lines that smoothly connect together these vectors for a depiction of an instantaneous flow field. They are visually appealing, but do not represent actual blood flow because they only reflect 1 time point. Pathlines, on the other hand, do represent blood flow through time. They are calculated by releasing imaginary particles into the flow field and then using the dataset to determine where particles will travel in an iterative process through the cardiac cycle. Despite the need for time-intensive post processing, these visualization methods have shown promise for several potential clinical applications, including investigation of valve-related abnormal flow patterns,13 blood compartmentalization in the ventricles,14 intimal entry tears and flow patterns in chronic dissection,15 corrected postrepair flow patterns,16 and identification of embolic pathways.17 QUANTITATIVE HEMODYNAMIC MARKERS Dynamic flow visualization is an obvious appeal of 4D Flow. It affords an intuitive representation of flow and a qualitative assessment of abnormal flow patterns. Quantitative markers, however, can be derived from 4D Flow datasets to more precisely characterize the hemodynamic consequences of pathologic flow disturbances. Although the true clinical utility of aortic 4D Flow remains to be determined, the ability to measure a diverse set of novel hemodynamic markers is likely to be its most clinically applicable feature. Clinical guidelines for the management of aortic disease use maximal diameter to risk stratify patients and to set thresholds for operative intervention. Aortic dissection and increased mortality, however, are reported with normal-sized and mildly dilated aortas.18 The topic of disease progression and aortic diameter is particularly problematic for patients with bicuspid aortic valves (BAV), with marked variability in management reported in routine surgical practice.19 Markers beyond aortic dimension that reflect the degree of an individual’s risk would be useful to better anticipate disease progression and complication. Herein we review promising new flow-related markers in 3 general contexts: Aortic valve disease, valve-related aortic disease, and aortic wall disease. Aortic Valve Disease Flow assessment has long been central to the clinical evaluation of the aortic valve. Three main flow-derived parameters are currently measured to determine the severity of aortic stenosis and guide surgical intervention: Transvalvular gradient, valve area, and peak velocity. Although these measurements generally perform well as markers of aortic disease severity, they are not always accurate. For example, gradient estimates using the modified Bernoulli equation do not take into account pressure recovery, and severe aortic stenosis with a low ejection fraction can have low gradients (ie, low-flow/low-gradient severe aortic stenosis).20,21 A more accurate composite measurement of the increased overall ventricular workload, “valvuloarterial impedance,” has been developed, and takes into account the degree of valve obstruction, the ventricular response, and the systemic vascular impedance.22 Valvuloarterial impedance may better represent the pathologic burden placed on the left ventricle that leads to overload and failure.20,23 4D Flow affords a complimentary means of calculating the overall ventricular workload experienced with aortic stenosis. Energy loss can be directly assessed with the technique by 2 recently described methods: Estimation of vicious energy loss and turbulent kinetic energy. Viscous dissipation of energy is a normal feature of aortic flow. For normal laminar flow, it is caused by friction between adjacent fluid layers with different velocities. This friction increases with the abnormal flow features that are seen with aortic valve disease, resulting in substantially elevated viscous energy losses (Fig. 1).24 A limitation of measuring viscous energy loss, however, is that it does not take into account turbulence. Turbulence is a common feature of poststenotic flow and a significant contributor to total irreversible pressure loss owing to the dissipation of mechanical energy into heat. Turbulent kinetic energy (TKE) can be estimated with 4D Flow by measuring the distribution of velocities within each imaging voxel: The greater the standard deviation of velocities, the higher the TKE. Recent studies have shown TKE is significantly increased in aortic stenosis, and that TKE measurements by 4D Flow correlate well with established methods of calculating irreversible pressure loss.25 As a direct imaging measurement of irreversible pressure loss, TKE may best reflect the increased workload placed on the ventricle with aortic stenosis. Valve-Related Aortic Disease Valve-related aortic disease has become one of the principal areas of aortic 4D Flow research. Disease of the aortic valve is frequently associated with abnormal dilation of the ascending aorta (i.e., post-stenotic dilation), especially in the case of congenital BAV. Abnormal hemodynamics and intrinsic aortic wall disease both likely play a role in the development of aortic dilation with BAV, with the relative contribution of each factor debated in the literature.26 The asymmetry of ascending aortic dilation that is typically seen with BAV, where there is disproportionate dilation of the aortic convexity, suggests an underlying asymmetric driver of disease.27 4D Flow research has focused on identifying hemodynamic markers that may be responsible for this dilation pattern and be used to predict disease progression. Research efforts have been invigorated by recent data that shows elevated aortic growth rates with restricted cusp opening angle with BAV, which presumably drives abnormal aortic flow.28 Abnormal systolic aortic flow patterns with BAV were first demonstrated with 4D Flow in 2008.29 Since then, numerous studies have demonstrated similar abnormal flow patterns in the ascending aorta of patients with acquired and congenital aortic valve disease.30–32 Qualitative visualization of disturbed flow has led to the development of qualitative measures of flow abnormality (Fig. 2). Wall sheer stress (WSS) is an often-reported quantitative marker of abnormal flow. The parameter can be estimated from near-wall velocity gradients with 4D Flow, and represents the frictional force experienced by the endothelium owing to flow viscosity. When deviated from a normal, intermediate range, WSS can adversely affect endothelial activation and signaling. High WSS states have been associated with vascular dilation and remodeling. Using 4D Flow, focally elevated WSS has been demonstrated in patients with BAV at the aortic convexity,30,33,34 a mechanism that may contribute to asymmetric dilation in this region. Although promising, WSS measurements are challenging. Greater differences between subgroups of patients are demonstrated if peak WSS values are reported,33 but these peak values are more subject to noise than the mean values that other groups have reported.34 Furthermore, the accuracy and reliability of WSS estimates derived from 4D Flow data have been called into question, given the difficulty in vessel wall segmentation and limited spatiotemporal resolution.10,35 Flow displacement is another quantitative parameter that has shown promise for characterizing BAV-related aortic disease. It measures the displacement of peak systolic flow from the vessel center caused by the presence of a BAV.36 In the more common BAV leaflet fusion pattern (right–left aortic leaflet fusion [RL]), this displacement is toward the aortic convexity, whereas with the other common fusion (right–noncoronary leaflet fusion [RN]), the displacement is typically more posterior. In some cases, differences in the direction of flow displacement result in completely different orientations of helical flow with RL versus RN fusions (Fig. 3). Intriguingly, the differences in flow displacement between RL and RN fusions have recently been associated with often reported differences in patterns of aortopathy; RL is associated with dilation of the tubular portion of the ascending aorta, whereas RN is associated with dilation of the aortic arch.32 Flow displacement is a very reproducible parameter that can clearly distinguish between aortic valve phenotypes.32,37 It is also the only 4D Flow parameter to date to be correlated with aortic growth in a small cohort study.37 Furthermore, data suggest that using 2D PC-MRI at the typically location for flow analysis in the ascending aorta can identify peak flow displacement values (Fig. 4). Taken together, these findings suggest that the simple parameter of flow displacement could be calculated from conventional 2D PC-MRI for the risk stratification of patients with BAV for aortic disease progression. Aortic Wall Disease Diseases that stiffen the aortic wall can affect aortic flow. Aortic wall stiffening is commonly seen with aging and is an independent predictor of mortality both in the general population and for patients with common chronic illnesses.38 Flow imaging can be used to estimate aortic wall stiffness by determining the speed of the systolic impulse through the aorta, or pulse wave velocity (PWV). Elevated PWV reflects wall stiffness and has been demonstrated using PC-MRI with hypertension and Marfan disease.39,40 For Marfan patients, PWV has long been investigated as means of predicting disease progression, with normal PWV values associated with regional aortic stability.41 4D Flow offers a more comprehensive assessment of aortic PWV than 2D PC-MRI (Fig. 5), and may improve the detection of regional differences that prior studies with Marfan syndrome have reported.41,42 In addition to PWV, 4D Flow is capable of producing accurate, noninvasive measurements of aortic pressure waveforms.43 Relative pressure can then be subdivided into inertial and viscous components, which show unique patterns in different types of aortopathy.44 Beside aortic wall stiffening, altered aortic flow patterns have been linked to the development of atherosclerosis. Low and oscillatory WSS have been shown induce endothelial changes that lead to plaque formation.45,46 Using 4D Flow estimates of WSS, further evidence has been gathered showing the interrelationship of altered aortic flow and atherosclerotic plaque formation.47,48 This suggests that 4D Flow could be used for identification of abnormal flow markers to predict an atherogenic-prone region, and thus guide risk stratification and medical management. SUMMARY Aortic disease is routinely monitored with anatomic imaging, but until the recent advent of 4D Flow imaging, associated blood flow abnormalities have gone largely undetected. 4D Flow is a rapidly evolving technique that is currently able to measure a range of aortic hemodynamic markers in less than 15 minutes. Initial qualitative flow visualization has spurred the investigation of new quantitative flow markers. For example, eccentric systolic flow with BAV has led to the application of WSS and flow displacement for assessment of related ascending aortic disease. Many promising 4D Flow markers of aortic disease have been proposed, although larger prospective studies are needed to validate their clinical relevance. Within the next decade, 4D sequences may be commonly acquired during routine clinical cardiac MR studies, and provide valuable information to guide the medical and surgical management of patients with aortic disease. Funding: Covidien/Radiological Society of North America Research Scholar Grant RSCH1215 (M.D. Hope). Fig. 1 Systolic velocity streamlines (left), maximum intensity projection (MIP) of the 3-dimensinal (3D) velocity field (middle), and viscous dissipation (right) in the thoracic aorta of a patient with borderline severe stenosis (peak systolic velocity of 4.0 m/s). Regionally high velocity gradients (double asterisk, black arrow) result in elevated energy loss. Flow jet impingement at the aortic wall (double asterisk, white arrow) is co-located with a region of substantial energy loss. EL', cumulative peak systolic energy loss in the ascending aorta. (Courtesy of A. Barker, PhD, and P. van Ooij, PhD, Northwestern University, Chicago, IL.) Fig. 2 Conventional aortic valve anatomy (A) is associated with a normal valve opening angle (~75°), flow jet angle (θ1), wall shear stress (WSS), and flow displacement. Flow displacement measures the displacement of peak systolic flow (arrowhead) from the vessel centerline, and is normalized to aortic size. It is calculated by dividing the distance from centerline of peak systolic flow (d), by the aortic diameter (dotted line with brackets). The abnormal valve anatomy seen with a bicuspid aortic valve (B), here typical of right-left aortic leaflet fusion, leads to a reduced valve opening angle, increased flow jet angle (θ1), and increased flow displacement. The increased near wall velocity gradient (region denoted by star) results in asymmetrically increased systolic WSS at the aortic convexity. Fig. 3 Right-handed, nested helical flow in a patient with a bicuspid aortic valve involving fusion of the right and left leaflets (R-L Fusion), and normal aortic dimensions. (A) Streamline analysis reveals greater than 180° curvature of peak systolic streamlines in a right-handed twist around slower, central helical flow in the ascending thoracic aorta. (B) Vector analysis reveals a right-anterior eccentric systolic flow jet. Left-handed nested helical systolic flow in a patient with a bicuspid aortic valve involving fusion of the right and noncoronary leaflets (R-N Fusion) and normal aortic dimensions. (A) Streamline analysis reveals greater than 180° curvature of peak systolic streamlines in a left-handed twist around slower, central helical flow in the ascending thoracic aorta. (B) Vector analysis reveals a left-posterior eccentric flow jet. Ant, anterior. (From Hope MD, Hope TA, Meadows AK, et al. Bicuspid aortic valve: four-dimensional MR evaluation of ascending aortic systolic flow patterns. Radiology 2010;255(1):59, 60; with permission.) Fig. 4 Representative plots of flow displacement from each of the cross-sections (red dots, red curve-spline fit) as a function of distance along the ascending thoracic aorta (AsAo) for each of the 6 groups labeled. The dotted blue line indicates the location of the 2-dimensional plane, which was placed just distal to the sinotubular junction. No high-flow displacement was observed for the healthy control group (A). Generally, the presence of aortic stenosis resulted in the highest flow displacement values that remained high for longer distances along the AsAo (B-D). Patients with bicuspid aortic valves (BAV) had characteristic 3-dimensional displacement plots with a sharp increase to proximal high peak values, followed by a smooth tapering distally through the ascending aorta (E, F). TAV, tricuspid aortic valve. (Courtesy of M. Sigovan, PhD, University of Lyon, Lyon, France.) Fig. 5 (A): Visualization of pulse wave propagation within the thoracic aorta, based on flow-sensitive 3-directional phase contrast MRI (4D) data. Equidistant analysis planes with an interanalysis plane gap of 10 mm were positioned upstream (negative analysis plane numbers) and downstream (positive analysis plane numbers) along the thoracic aorta, starting with analysis plane #0 (outlet of the left subclavian artery). (B–D) Spatially varying flow profiles from the proximal to the descending thoracic aorta (DAo) in analysis planes can be appreciated in successive systolic time frames: during early systole (B), profiles in the proximal DAo are already fully developed, whereas velocities are continuously lower further downstream. During peak systole (C), flow profiles reach their maxima in the entire DAo, whereas during late systole (D), flow profiles in the proximal DAo are already reduced compared with flow further downstream. AAo, ascending thoracic aorta. (From Markl M, Wallis W, Brendecke S, et al. Estimation of global aortic pulse wave velocity by flow-sensitive 4D MRI. Magn Reson Med 2010;63(6):1579; with permission.) KEY POINTS Three-directional phase contrast MRI (4D flow MRI) is able to visualize and quantify abnormal flow related to a wide variety of aortic pathologies. Limitations of the technique, including scan time and artifacts, have been greatly reduced making 4D flow a clinically feasible technique. Novel quantitative hemodynamic markers have been developed to characterize abnormal flow, and to investigate underlying mechanisms of disease. Markers beyond aortic diameter could improve risk stratification of patients with aortic disease and better determine the timing of intervention. Larger, prospective studies are needed to validate the clinical relevance of 4D flow. Disclosures: No conflicts to disclose. REFERENCES 1 Olsson C Thelin S Stahle E Thoracic aortic aneurysm and dissection: increasing prevalence and improved outcomes reported in a nationwide population-based study of more than 14,000 cases from 1987 to 2002 Circulation 2006 114 24 2611 8 17145990 2 Michelena HI Khanna AD Mahoney D Incidence of aortic complications in patients with bicuspid aortic valves JAMA 2011 306 10 1104 12 21917581 3 Clouse WD Hallett JW Jr Schaff HV Improved prognosis of thoracic aortic aneurysms: a population-based study JAMA 1998 280 22 1926 9 9851478 4 Tsai SF Trivedi M Daniels CJ Comparing imaging modalities for screening aortic complications in patients with bicuspid aortic valve Congenit Heart Dis 2012 7 4 372 7 22691072 5 Bonow RO Carabello BA Chatterjee K ACC/ AHA 2006 guidelines for the management of patients with valvular heart disease: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (writing Committee to Revise the 1998 guidelines for the management of patients with valvular heart disease) developed in collaboration with the Society of Cardiovascular Anesthesiologists endorsed by the Society for Cardiovascular Angiography and Interventions and the Society of Thoracic Surgeons J Am Coll Cardiol 2006 48 3 e1 148 16875962 6 Steffens JC Bourne MW Sakuma H Quantification of collateral blood flow in coarctation of the aorta by velocity encoded cine magnetic resonance imaging Circulation 1994 90 2 937 43 8044965 7 Schnell S Markl M Entezari P k-t GRAPPA accelerated four-dimensional flow MRI in the aorta: effect on scan time, image quality, and quantification of flow and wall shear stress Magn Reson Med 2014 72 2 522 33 24006309 8 Hsiao A Lustig M Alley MT Rapid pediatric cardiac assessment of flow and ventricular volume with compressed sensing parallel imaging volumetric cine phase-contrast MRI AJR Am J Roentgenol 2012 198 3 W250 9 22358022 9 Gatehouse PD Rolf MP Graves MJ Flow measurement by cardiovascular magnetic resonance: a multi-centre multi-vendor study of background phase offset errors that can compromise the accuracy of derived regurgitant or shunt flow measurements J Cardiovasc Magn Reson 2010 12 5 20074359 10 Petersson S Dyverfeldt P Ebbers T Assessment of the accuracy of MRI wall shear stress estimation using numerical simulations J Magn Reson Imaging 2012 36 1 128 38 22336966 11 Boussel L Rayz V Martin A Phase-contrast magnetic resonance imaging measurements in intracranial aneurysms in vivo of flow patterns, velocity fields, and wall shear stress: comparison with computational fluid dynamics Magn Reson Med 2009 61 2 409 17 19161132 12 Sengupta PP Pedrizzetti G Kilner PJ Emerging trends in CV flow visualization JACC Cardiovasc Imaging 2012 5 3 305 16 22421178 13 Hope MD Wrenn SJ Sigovan M Imaging Biomarkers of Aortic Disease: increased growth rates with eccentric systolic flow J Am Coll Cardiol 2012 60 4 356 7 22813616 14 Eriksson J Carlhall CJ Dyverfeldt P Semiautomatic quantification of 4D left ventricular blood flow J Cardiovasc Magn Reson 2010 12 9 20152026 15 Clough RE Waltham M Giese D A new imaging method for assessment of aortic dissection using four-dimensional phase contrast magnetic resonance imaging J Vasc Surg 2012 55 4 914 23 22386146 16 Allen BD Barker AJ Kansal P Impact of aneurysm repair on thoracic aorta hemodynamics Circulation 2013 128 17 e341 3 24146124 17 Harloff A Strecker C Frydrychowicz AP Plaques in the descending aorta: a new risk factor for stroke? Visualization of potential embolization pathways by 4D MRI J Magn Reson Imaging 2007 26 6 1651 5 18022846 18 Chau KH Elefteriades JA Natural history of thoracic aortic aneurysms: size matters, plus moving beyond size Prog Cardiovasc Dis 2013 56 1 74 80 23993240 19 Verma S Yanagawa B Kalra S Knowledge, attitudes, and practice patterns in surgical management of bicuspid aortopathy: a survey of 100 cardiac surgeons J Thorac Cardiovasc Surg 2013 146 5 1033 40 e4 23988289 20 Otto CM Valvular aortic stenosis: disease severity and timing of intervention J Am Coll Cardiol 2006 47 11 2141 51 16750677 21 Lancellotti P Magne J Valvuloarterial impedance in aortic stenosis: look at the load, but do not forget the flow Eur J Echocardiogr 2011 12 5 354 7 21555456 22 Briand M Dumesnil JG Kadem L Reduced systemic arterial compliance impacts significantly on left ventricular afterload and function in aortic stenosis: implications for diagnosis and treatment J Am Coll Cardiol 2005 46 2 291 8 16022957 23 Hachicha Z Dumesnil JG Pibarot P Usefulness of the valvuloarterial impedance to predict adverse outcome in asymptomatic aortic stenosis J Am Coll Cardiol 2009 54 11 1003 11 19729117 24 Barker AJ van Ooij P Bandi K Viscous energy loss in the presence of abnormal aortic flow Magn Reson Med 2014 72 3 620 8 24122967 25 Dyverfeldt P Hope M Tseng EE Noninvasive magnetic resonance measurement of turbulent kinetic energy for the estimation of irreversible pressure loss in aortic stenosis JACC Cardiovasc Imaging 2013 6 1 64 71 23328563 26 Girdauskas E Borger MA Secknus MA Is aortopathy in bicuspid aortic valve disease a congenital defect or a result of abnormal hemodynamics? A critical reappraisal of a one-sided argument Eur J Cardiothorac Surg 2011 39 6 809 14 21342769 27 Lu MT Thadani SR Hope MD Quantitative assessment of asymmetric aortic dilation with valve-related aortic disease Acad Radiol 2013 20 1 10 5 22951111 28 Della Corte A Bancone C Conti CA Restricted cusp motion in right-left type of bicuspid aortic valves: a new risk marker for aortopathy J Thorac Cardiovasc Surg 2012 144 2 360 9 22050982 29 Hope MD Meadows AK Hope TA Images in cardiovascular medicine. Evaluation of bicuspid aortic valve and aortic coarctation with 4D flow magnetic resonance imaging Circulation 2008 117 21 2818 9 18506021 30 Bissell MM Hess AT Biasiolli L Aortic dilation in bicuspid aortic valve disease: flow pattern is a major contributor and differs with valve fusion type Circ Cardiovasc Imaging 2013 6 4 499 507 23771987 31 Hope MD Hope TA Meadows AK Bicuspid aortic valve: four-dimensional MR evaluation of ascending aortic systolic flow patterns Radiology 2010 255 1 53 61 20308444 32 Mahadevia R Barker AJ Schnell S Bicuspid aortic cusp fusion morphology alters aortic 3D outflow patterns, wall shear stress and expression of aortopathy Circulation 2014 129 6 673 82 24345403 33 Hope MD Hope TA Crook SE 4D flow CMR in assessment of valve-related ascending aortic disease JACC Cardiovasc Imaging 2011 4 7 781 7 21757170 34 Barker AJ Markl M Burk J Bicuspid aortic valve is associated with altered wall shear stress in the ascending aorta Circ Cardiovasc Imaging 2012 5 4 457 66 22730420 35 Stalder AF Russe MF Frydrychowicz A Quantitative 2D and 3D phase contrast MRI: optimized analysis of blood flow and vessel wall parameters Magn Reson Med 2008 60 5 1218 31 18956416 36 Sigovan M Hope MD Dyverfeldt P Comparison of four-dimensional flow parameters for quantification of flow eccentricity in the ascending aorta J Magn Reson Imaging 2011 34 5 1226 30 21928387 37 Hope MD Sigovan M Wrenn SJ MRI hemodynamic markers of progressive bicuspid aortic valverelated aortic disease J Magn Reson Imaging 2014 40 1 140 5 24788592 38 Cavalcante JL Lima JA Redheuil A Aortic stiffness: current understanding and future directions J Am Coll Cardiol 2011 57 14 1511 22 21453829 39 Westenberg JJ Scholte AJ Vaskova Z Age-related and regional changes of aortic stiffness in the Marfan syndrome: assessment with velocity-encoded MRI J Magn Reson Imaging 2011 34 3 526 31 21761466 40 Brandts A Westenberg JJ van Elderen SG Site-specific coupling between vascular wall thickness and function: an observational MRI study of vessel wall thickening and stiffening in hypertension Invest Radiol 2013 48 2 86 91 23262794 41 Kroner ES Scholte AJ de Koning PJ MRI-assessed regional pulse wave velocity for predicting absence of regional aorta luminal growth in Marfan syndrome Int J Cardiol 2013 167 6 2977 82 23000269 42 Markl M Wallis W Brendecke S Estimation of global aortic pulse wave velocity by flow-sensitive 4D MRI Magn Reson Med 2010 63 6 1575 82 20512861 43 Delles M Rengier F Jeong YJ Estimation of aortic pressure waveforms from 4D phase-contrast MRI Conf Proc IEEE Eng Med Biol Soc 2013 2013 731 4 24109791 44 Lamata P Pitcher A Krittian S Aortic relative pressure components derived from four-dimensional flow cardiovascular magnetic resonance Magn Reson Med 2014 72 4 1162 9 24243444 45 Cheng C Tempel D van Haperen R Atherosclerotic lesion size and vulnerability are determined by patterns of fluid shear stress Circulation 2006 113 23 2744 53 16754802 46 Slager CJ Wentzel JJ Gijsen FJ The role of shear stress in the generation of rupture-prone vulnerable plaques Nat Clin Pract Cardiovasc Med 2005 2 8 401 7 16119702 47 Markl M Brendecke SM Simon J Co-registration of the distribution of wall shear stress and 140 complex plaques of the aorta Magn Reson Imaging 2013 31 7 1156 62 23773622 48 Frydrychowicz A Stalder AF Russe MF Three-dimensional analysis of segmental wall shear stress in the aorta by flow-sensitive four-dimensional-MRI J Magn Reson Imaging 2009 30 1 77 84 19557849

PMC005xxxxxx/PMC5127273.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 8706257 4240 Gastroenterol Clin North Am Gastroenterol. Clin. North Am. Gastroenterology clinics of North America 0889-8553 1558-1942 27837775 5127273 10.1016/j.gtc.2016.07.001 NIHMS823246 Article The Gut Microbiota: The Gateway to Improved Metabolism Martinez Kristina B. PhD, RD Pierre Joseph F. PhD Chang Eugene B. MD Section of Gastroenterology, Hepatology, and Nutrition, Department of Medicine, University of Chicago, Chicago IL Corresponding Author: Eugene B. Chang, MD, Martin Boyer Professor of Medicine, Department of Medicine, Knapp Center for Biomedical Discovery, Rm. 9130, 900 E 57th Street, University of Chicago, Chicago, IL 60637, Phone number: 773-702-6458, FAX Number: 773-702-2281, echang@medicine.bsd.uchicago.edu 15 10 2016 12 2016 01 12 2017 45 4 601614 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. Obesity is an emerging global epidemic with profound challenges to world healthcare economies and societies. Traditional approaches to fight obesity have not shown promise in promoting a decline in obesity prevalence. The gut microbiota are becoming widely appreciated for its role in regulating metabolism and thus represents a target for new therapies to combat obesity and associated co-morbidities. In this review, we provide an overview of altered microbial community structure in obesity, dietary impact on the gut microbiota, host-microbe interactions contributing to the disease, and lastly improvements in microbial assemblage after bariatric surgery and with therapies targeting the gut microbiome. Microbiota Microbiome Obesity Metabolism Bariatric Surgery RYGB Enteroendocrine Hormones Lipid Absorption Circadian Rhythm Probiotics Prebiotics Synbiotics Introduction: Obesity and the Gut Microbiota The rise of obesity and its related comorbidities in ‘westernized’ countries over the past four decades presents an emerging global epidemic with profound challenges to world healthcare economies and societies. In the past 35 years, the rate of adult obesity has risen by 75% globally 1,2. This number was greater among children 3,4. Stratified assessment of Body Mass Index (BMI) further demonstrates disproportionate increases among the most severely obese (≥35 kg/m2), compared with the lesser obese (≥30 kg/m2), illustrating the scale of the problem. Unfortunately, obesity and its comorbidities 5, including metabolic syndrome, diabetes, and heart disease, have detrimental effects on quality of life and substantial costs to individuals and societies. Thus, the need for understanding the complexity of pathophysiological events and elucidating effective interventions remain urgent. The etiology of obesity is multifactorial, including the complex interaction of genetics and environment, which encompasses diet, developmental factors, lifestyle (eg. hedonistic tendencies, altered sleep patterns), and antibiotic use. Intestinal microbes are impacted by all of these factors in their community structure and function and in turn initiate host-microbe interactions that may disrupt metabolic and immune homeostasis. Strikingly, fecal microbiota transplant (FMT) of microbes under environmental stressors, like diet and obesity, can induce a similar phenotype in recipients 6. Indeed, the gut microbiome is by definition a microbial organ - vital to intestinal and systemic functions – one that we cannot live without, but also an organ that is transplantable (i.e., via FMT). This technique is commonly used for Clostridium difficile infection and has only recently been studied for use in other conditions, including obesity. However, other therapies targeting the microbiome, such as pre- and pro-biotics, may confer modest, yet positive, improvements for symptoms associated with obesity and its comorbidities. Although an extreme measure reserved for the morbidly obese, one of the most effective strategies to decrease obesity is bariatric surgery which profoundly changes the gut microbiota, energy balance, and alters physiological and endocrine metabolic states 7. It is expected that by changing metabolic set points, desired weight can be achieved. Therefore, understanding the mechanisms behind bariatric surgery and associated changes in the gut microbiota may be leveraged to develop new therapies to fight the obesity epidemic. In this review, we explore these concepts by providing an overview of altered microbial structure and function in obesity, host-microbe interactions driving obesity, dietary influence on the microbiome, improvements in metabolism and microbial structure with RYGB, the host-microbe interactions driving obesity, and lastly current therapies targeting the gut microbiome to facilitate positive metabolic outcomes. Paradigms in Gut Microbiota During Obesity Obesity-Driven Alterations in Gut Microbiota The human body contains staggering numbers of microbes, including thousands of bacterial species, in addition to many eukaryotes, Achaea, protists, and viruses, which collectively contain an estimated 5 million genes that have profound metabolic and immunomodulatory effects upon the mammalian host 8. The community of microbes is termed the microbiota, while their collective genes are called the microbiome. Both the state of obesity and westernized diets are associated with microbial dysbiosis, which is a deviation from microbial organization that would otherwise promote optimal metabolic homeostasis. Dysbiotic microbiota in obesity is characterized by decreased diversity in the microbial community and by an increased ratio of the phylum Firmicutes to the phylum Bacteroidetes 9. The change in Firmicutes and Bacteriodetes ratio occurs in both mice and humans, and weight loss restores microbial composition 9–11. Interestingly, 3 genuses of bacteria are often overrepresented under obesity in humans, including Bacteroides and Prevotella (both Bacteroidetes) or Ruminococcus (Firmicutes) 12. In addition to composition, major functional differences are observed in metabolic capacity of the microbial community. For instance, decreases in SCFA producers, such as from the phylum Actinobacteria and blooms in pathogenic bacteria from the phylum Proteobacteria, occur in obesity 13. In addition to bacteria, recent work demonstrates the microbiota metabolic networks include yeast and archaea, which synergistically produce and utilize metabolites collectively with bacteria 14. While this area is still relatively unexplored, recent work suggests yeast species abundance is lower under obesity, and supplementation with Saccharomyces cerevisea improves metabolic parameters and adiposity 15–17. When dysbiotic communities of bacteria are transferred to naïve germ-free mice, the recipient mice develop increased adiposity, demonstrating a direct impact of the microbes on advancing fat storage in the mammalian host 6. While some bacteria are associated with excess adiposity, others have been directly implicated in improving metabolic syndrome and atherosclerosis, such as Akkermansia muciniphila 18,19, which are often found to be underrepresented in obesity. Administration of A. muciniphila during obesity was shown to improve glucose tolerance 19. Schneeberger et al found that among 27 genes that regulate inflammation and metabolism in white adipose tissue under high fat feeding, 20 genes negatively correlate with the relative abundance of A. muciniphila. Additionally, Bifidobacterium spp. also negatively correlated with 6 of 27 genes. Positive correlations were observed with Bilophila wadsworthia in 14 of 27 genes, and this microbe is known to expand under high milk fat diets and subsequently stimulate inflammatory responses 20. Together these observations suggest certain microbes might regulate aspects of peripheral metabolism related to obesity and metabolism, but further investigations to determine strong proof of causality are required. Another line of evidence that supports the notion microbes might closely regulate host metabolism and body weight is found through the study of acute malnutrition in childhood. In contrast to overfed and obese individuals, work by Gordon and colleagues 21 followed severely malnourished children for 2 years. Through compositional modeling, they demonstrated the microbiota normally develop with the growing child, but with malnutrition, the microbiota “maturity” remains stunted and lags behind host development. Even after common therapeutic food interventions, the immaturity of the microbiota persisted. Intriguingly, it was found that the microbiota remains immature even under less severe malnourished states and that microbiota maturity correlated with anthropometric measurements of the children 21. These findings strongly support the notion of the gut microbiota functioning as a vital organ, and its development and growth throughout life may have important unknown implications for human health. Understanding the role of microbial development under states of hyper-alimentation may provide insights into dysfunctional microbial-host metabolic interactions that lead to excessive fat storage. Dietary Impact on the Gut Microbiota While host phenotype influences the composition of the microbial communities, diet also notably has an immediate and dramatic impact on microbial structure that mimics communities seen in obese individuals. For instance, Turnbaugh et al observed in humans that a diet rich in animal-derived fat and protein resulted in significant changes in the gut microbiota in as little as a day, and of particular note, blooms in hydrogen-sulfide producing bacteria such as Bilophila wadsworthia were also observed 22. Later his group found that diet has a greater impact on altering microbial assemblage than genetic background in mice 23. In this study, five different inbred mouse strains, four genetic knockout strains relevant to host-microbe interactions (e.g., ob/ob, NOD2, MyD88−/−, and Rag1−/−), and 200 outbred mice were placed on high-fat, high-sugar diets or diets rich in plant polysaccharides. In each experiment, the western diet had profoundly altered community structure, regardless of strain differences or gene deficiencies 24. More recently, Sonnenburg’s group nicely demonstrated that diets low in microbe-accessible carbohydrate (MAC) and high in simple sugars result in loss of bacterial diversity and extinction of specific microbial groups which is compounded over generations 25. The generational loss of bacterial diversity could only be remedied with fecal microbiota transplant from control mice maintained on the MAC-rich diet, but not by diet alone. This study provides a model for the rapid and drastic impact of our food supply, containing readily-available processed high-fat and high-sugar food, on the progressive loss of bacterial diversity over the past several decades. This theory postulates that our bodies are not equipped to reciprocate and adapt to the sudden insult on our gut microbiota, thereby leading to the development of obesity. Based on the results from this study, suitable therapies to combat the loss of bacterial diversity might include probiotic supplementation or fecal microbiota transplant. High fat (HF) diets also transform the metagenomes of the bacteriophage community, also known as the “phageome.” It was demonstrated by Howe and Ringus et al. 26 that HF diets can shift the “phageome” independent of observed alterations in their bacterial host pattern. The impact of diet on viral communities was also rapid, occurring within 24 hours. In addition, the change in the viral metagenomes by high fat diet was not reversible after washout, suggesting that diet-mediated changes in the phage community are persistent, similar to the aforementioned findings in bacteria 26. Further research in this area is needed to better understand the regulation and function of the phageome, its impact on gut microbial ecology, and more importantly consequences for the host. Restructuring of Gut Microbiota in Bariatric Surgery Surgical intervention, although largely invasive, is the most effective weight loss strategy for obesity. Roux-en-Y (RYGB) is the most common and the most effective, promoting a 20–40% weight loss compared to 15–30% loss of body weight with gastric banding 27. RYGB includes the formation of a small pouch, made of the upper stomach that is then attached to a region of jejunum approximately 75 cm distal to the stomach (termed a gastrojejunostomy). The resulting limb (including the distal stomach) carries bile, gastric juice, and pancreatic juices alone, without nutrients, another 125 cm distal from the gastrojejunostomy, collectively delaying the mixture of digestive juices from nutrients for approximately 200 cm up upper GI tract. It is becoming increasingly apparent that bariatric surgery, particularly RYGB, may involve multiple mechanisms beyond simply physical restrictions to nutrient intake and absorption through reduced stomach size and decreased absorptive capacity. It is plausible that new treatments will be discovered based on the mechanisms underlying bariatric surgery efficaciousness. Intriguing data suggests the mechanisms involved may include altered gut microbial function 7 and interactions of the microbiome with the hosts bile acid pool 28,29. Anatomical rearrangement triggers the dramatic restructuring of the intestinal microbiota and host-microbe interactions that may contribute to weight loss after bariatric surgery. For example, RYGB in mice resulted in the rapid restructuring of gut microbiota as early as one week compared to sham controls 7. The early changes in microbial composition occur under the same time frames that improvement in glucose tolerance and reduced insulin resistance are observed 30, in contrast to body weight and adiposity changes that occur over weeks and months, suggesting microbial alterations may be involved in the resetting of metabolic set points that are distinct from adiposity. Specifically, RYGB results in a decrease in the Firmicutes to Bacteroidetes ratio, which includes increases in Bacteroidales, Enterbacteriales, as well as increases in Gammaproteobacteria (E. coli) and Verrucomicrobia 7 as a relative percentage of the microbial community. Interestingly, the Verrucomicrobia genus Akkermansia utilizes host secretion of mucin as a fuel source and has been inversely correlated with body weight 19. As previously discussed, oral administration of live Akkermansia muciniphila restores insulin sensitivity in high fat fed animals 19. Following RYGB in diabetic rodents, the level of A. muciniphila in the small bowel increased significantly compared with sham obese controls 31. Intriguingly, in that study, the increase in A. mucinipila was positively related to the release of GLP-1, an important intestinal incretin, suggesting this microbe could be modulating peripheral glucose handling through modulated insulin tolerance. In humans following RYGB, an inverse correlation in the relative percentage of E. coli, Bacteroides, and Prevoltella with circulating leptin levels were found, an important adipose factor released at higher levels under obesity 32. Finally, the increases in Proteobacteria observed following RYGB have reached 50-fold, and together with other models suggest Proteobacteria may influence insulin sensitivity 33,34. A direct role for the microbial community in mediating host metabolism following RYGB was confirmed with fecal microbiota transplant from bariatric surgery donors into recipients, which conferred protection from obesity 35. Altogether, these studies provide strong evidence that the gut microbiota may significantly contribute to the effectiveness of RYGB surgery, paving the way for focused investigations into altering the microbiota in a similar manner for the treatment of obesity. The direct role for microbes in improving metabolism following RYGB remain under investigation, but other indirect roles include microbial changes to the bile acid composition, bile-acid activation of the ileal and colonic bile acid receptors, and regulation of gut peptide enteroendocrine hormones, such as GLP-1 and PYY. Current evidence suggests the composition of bile acids influence the microbiota assemblage through antimicrobial function 28,29, since bile acids are detergents and influence the membrane chemistry. Reciprocally, bacteria influence bile acid composition by deconjugation and fermentation of primary bile acids into secondary and tertiary bile acids, which have differential effects upon host metabolism. Primary bile acids are associated with improved metabolism, while secondary bile acids are potentially carcinogenic and not associated with metabolic improvement 36,37. Therefore, bile acids and microbial compositions are inseparably associated and continually interacting in the gut. Novel work demonstrated that bile acid-altered microbial communities in turn influence host metabolism, establishing a cross-talk between bile acids and the intestinal microbiome that influences host metabolism 38. In addition to altering bacterial viability and growth, and aiding in the absorption of luminal dietary lipids and lipid vitamins, bile acids directly modulate host metabolism through host bile acid receptors 28,39. Bile acid interactions with the G protein-coupled bile acid receptor 1 (GPBAR1 or TGR5) and farnesoid X Receptor (FXR) regulate peripheral energy expenditure and counteract obesity and diabetes. Upon activation by bile acids, TGR5 specifically stimulates the release of GLP-1, GLP-2 and PYY from enteroendocrine cells as well as expression of various transport proteins and biosynthetics, resulting in improved glycemic control. Enteroendocrine cells also contain toll-like receptors and sense bacteria in the intestinal lumen 40. Following RYGB, elevated circulating levels of GLP-1 and PYY are reported 41,42. PYY is normally released postprandial to increase energy expenditure and decrease food intake. In healthy individuals, PYY release following feeding is proportionate to caloric consumption, acting directly on the hypothalamus and vagal afferents to slow feeding behavior 43. It remains unclear if elevated PYY and GLP-1 following RYGB are in response to altered microbial populations, bile acid pools, or a combination. Regardless, changes in intestinal enteroendocrine signaling following RYGB have profound effects on host metabolism and multiple lines of evidence now strongly support the involvement of the microbiota. Host-Microbe Interactions Driving Obesity The observation that GF mice are resistant to diet-induced obesity has created a foundation for understanding the contribution of microbes and host-microbe interactions to the development of obesity and its co-morbidities. Several mechanisms to explain microbe-mediated obesity have been proposed, including 1) short chain fatty acid (SCFA) production, 2) regulation of food intake and sensory perception of food, 3) nutrient absorption, 4) circulation of microbe-derived enterotoxins like LPS and reduced production of angiopoietin like 4 (angptl4) resulting in increased fatty acid uptake in liver and adipose tissue 13, 5) and peripheral control of circadian rhythm which is intimately links metabolic coordination between the brain and peripheral organs 44,45. One metabolic function of microbes is the production of SCFAs, including acetate, propionate, and butyrate from otherwise indigestible fibers. SCFAs can act as energy sources for the intestinal epithelium and liver and mouse models of obesity demonstrate elevated SCFA in luminal content and lower energy content in feces 46. However, SCFAs have many reported beneficial effects on metabolism and improved glucose tolerance. For instance, diets supplemented with fructooligosaccharides, butyrate, and propionate decreased weight gain and improved glucose tolerance in rats compared to controls. It was shown that these positive effects were mediated through stimulation of intestinal gluconeogenesis, as mice deficient in the catalytic subunit of glucose-6 phosphatase, displayed impaired glucose tolerance 47. Thus, conflicting evidence exists regarding the negative consequences of increased energy availability through SCFA production given the potential positive impact of SCFAs on metabolism. Obesity is also related to peripheral inflammation, especially in adipose tissue. The gut microbiota influences gut permeability, which may lead to entry of microbial ligands, including lipopolysaccharides (LPS), into the blood stream and periphery, where they can induce insulin resistance and prevent peripheral uptake of fat 48–50. Intriguingly, bioactive dietary components such as omega 3 fatty acids and polyphenols that are reported to improve adipose inflammation also impact microbial structure. For example, Backhed’s group demonstrated that the gut microbiota exacerbate adipose inflammation through TLR signaling upon saturated fat feeding, as has been suspected for some time in the adipose biology field. Notably, microbiota transplant from fish oil-fed mice attenuated weight gain in antibiotic mice that were maintained on a lard diet 51. Cranberry and grape polyphenols have also been reported to alter microbial structure, specifically via increasing the abundance of Akkermansia muciniphila, as well as improve glucose tolerance and adipose inflammation 52,53. Microbial regulation of metabolism is mediated in part through the sensory perception of food, gastrointestinal motility, and nutrient absorption, as these are altered in GF mice. For instance, GF mice have increased preference for sugar-sweetened liquids and fat emulsions but lack the machinery to process the nutrients. Swartz et al. 54 found that GF mice consume more sucrose solution concurrent with increased expression of type 1 taste receptor 3 (TIR3) expression and sodium glucose luminal transporter 1 (SGLT1) in the small intestinal epithelium compared to conventional mice 54. Anorexingenic gut peptide hormones, including PYY and CCK, are also regulated by gut microbes facilitating control of food intake. While GF mice have increased expression of lingual fatty acid translocase (CD36), expression of gut peptide hormones including PYY, CCK, and GLP-1 were reduced in the intestinal epithelium, as well as decreased numbers of EECs in the ileum 55. In contrast, to these gut peptide hormones, Backhed’s group reported that GF mice have elevated GLP-1 which slows gastric motility and increases intestinal transit time as a compensatory mechanism to allow for enhanced nutrient absorption 56. Given the role of enteroendocrine hormone signaling in nutrient absorption, dysregulation of these hormones may explain why germ free mice have elevated levels of triglycerides and total lipids in their stool after high fat diet feeding 57. Interestingly, conventionalization of GF zebrafish increases lipid accumulation in the intestinal epithelium 44. Taken together, these findings suggest that gut microbes facilitate hormonal cues to regulate sensory perception of food, dietary intake, as well as nutrient absorption. However, the exact mechanisms behind microbial regulation of carbohydrate and lipid absorption and the extent to which microbe-induced nutrient absorption significantly contributes to obesity have not been well characterized. Differences between GF and conventional mice also involve dysregulation of bile production. Due to the lack of microbes in GF mice, there is little to no deconjugation of conjugated bile acids entering the GI lumen, thereby resulting in high levels of taurine-conjugated bile acids in GF mice compared to conventional mice 38. Backhed’s group reported that elevated taurine-conjugated bile acids block FXR-mediated induction of FGF-15, which would otherwise decrease bile acid synthesis in the liver. Thus, germ free mice have increased bile acid production. This study implicates the role of gut microbiota in regulating bile acid metabolism through a gut-liver axis 58. The difference in bile ‘1acid metabolism speaks to the marked difference in liver function between GF and conventional mice. Indeed, GF mice have decreased liver lipid content and altered expression of gene networks including those involving xenobiotic metabolism and circadian rhythm 59. At the hub of xenobiotic gene networks are two nuclear hormone receptors, constitutive androstane receptor (CAR) and pregnane X receptor (PXR), which are implicated in regulating whole body metabolism. Activation of CAR has been shown to decrease body weight and improve insulin sensitivity, whereas PXR activation has been positively associated with obesity 60–62. Thus, it is tempting to speculate that CAR-mediated metabolic activity in GF mice may contribute to their resistance to high fat diet-induced obesity. However, this connection has not been thoroughly investigated in the current literature. Gut microbes have been found to control circadian function. This has important implications for fighting obesity, as the disruption of the natural cycle of day and night (e.g., jet lag, shift work and sleep apnea) contributes to the increasing prevalence of metabolic disorders 13. Circadian rhythms are regulated by molecular clocks that coordinate regularly timed events (i.e., states of feeding vs fasting) and the necessary physiological responses to enhance metabolic efficiency. Thus, circadian rhythm is intimately linked to the regulation of food intake, activity, and whole body metabolism involving clocks located in the brain as well as peripheral metabolic tissues. The circadian transcriptional program is under the control of two major transcriptional activators, Bmal and Clock, which are counter-regulated by repressors, Period 1–3 and Cryptochrome 1/2. Consumption of high-fat diets represses diurnal variation of these gene transcripts and impairs normal circadian function 63. It has recently been shown that these changes are dependent on the gut microbiota, as GF mice and antibiotic-treated mice have reduced expression of Bmal and Clock and increased expression of Per1-3, and Cry1/2 in the intestinal epithelium 64. It was later demonstrated by Leone et al. 45 that diurnal variation in the circadian gene program is also blunted in the liver of GF compared to SPF mice. In addition to host circadian rhythm, microbes themselves display circadian behavior 45. Strikingly, community structure of the gut microbiota as well as butyrate exhibits diurnal variation over a 24-hour period under normal feeding conditions and is diminished under high fat feeding. To ensure these changes were not due to times of feeding, stool was collected from mice on total parenteral nutrition (TPN) and compared to mice fed enterally. Although differences existed in the relative abundance of specific microbes (e.g., increase in Verrucomicrobia in TPN group), diurnal shifts were still evident, indicating that microbial abundance may fluctuate based on host cues, such as the release of mucin or other epithelial proteins and secretions. Altogether these findings suggest that the regulation of host circadian function is dependent on the activity of the gut microbiota and conversely, the circadian behavior of the gut microbiota is dependent on host physiology 13,45. Identifying the host-microbe interactions that facilitate microbial control of circadian rhythm may lead to therapies targeting the gut microbiota to restore the metabolic consequences of disrupted sleep, common in obesity. Treatments targeting microbiome to fight obesity and metabolic syndrome The host-microbiome field is moving toward improving metabolism and weight maintenance through modulating gut microbial communities using a variety of supplements such as pre- and pro-biotics, synbiotics, FMT, and postbiotics. Prebiotics are foods or dietary supplements that encourage the growth of saccharolytic bacteria that metabolize non-digestible carbohydrates such as inulin and oligofructose. Several criteria must be met for a supplement to be considered a prebiotic and these include: resistance to gastric acidity, non-digestible by the host in the small intestine, bacterial fermentation, and promotion of beneficial bacteria 65. Prebiotics Prebiotics have recently been shown to improve complications associated with metabolic disorders including obesity and insulin resistance 66. Various mechanisms have been identified to explain these beneficial effects including SCFA production, stimulation of intestinal gluconeogenesis, epithelial integrity, release of hormones PYY and GLP1 to promote satiety and insulin sensitivity, increased expression of antimicrobial peptides, and alteration of gut microbial community structure 66. Gene expression of the antimicrobial peptide, Reg3y, was reduced after HFD feeding but restored upon delivery of oligofructose 35. Prebiotic supplementation also increased intectin expression, which promotes epithelial cell turnover and maintenance. Fructooligosaccharide (FOS) treatment in mice fed a Western diet, exhibited improved glucose and insulin tolerance compared to controls. The therapeutic effect of FOS was lost in mice deficient in glucose-6-phosphatase catalytic subunit (G6Pc), thereby shutting down intestinal gluconeogenesis. These findings implicate that intestinal gluconeogenesis is necessary for FOS-mediated glucose and insulin sensitivity 47. Similar results have been shown in humans. For example, participants fed brown beans 67 or prebiotics containing wheat fiber and soluble fiber. 68 displayed improved insulin sensitivity. Taken together, these findings support the use of prebiotic therapy in both animals and humans for improved metabolic health. Probiotics Another commonly used approach and widely studied supplement is the use of probiotics, that are live microorganisms delivered individually or in combinations such as VSL#3, that positively impact health outcomes in the host 65. It is important to consider the composition of probiotic formulations as each strain may have a different impact on microbial structure/function or on the host immune response, for instance. Wang et al. 66 demonstrated in mice that three strains of bacteria including Lactobacillus Paracasei CNCM I-4270, L. rhamnosus I-3690 and Bifidobacterium animalis subsp Lactis I-2494 independently decreased body weight and improved glucose tolerance but through different mechanisms (reviewed in 13). Daily gavage of the probiotic yeast Saccharomyces boulardii or Biocodex, elicited changes in gut microbiota, reflecting a less obesogenic state, as well as improved the metabolic profile of genetically obese and diabetic db/db mice 17. A common concern with probiotics use is the lack of colonization following supplementation 69. In addition, mixed strain probiotics like VSL#3 or a symbiotic, which is the combination of a probiotic and prebiotic, may be more effective than single microbial isolates alone. VSL#3 contains 7 different strains belonging to the genus Bifidobacterium and Lactobacillus and has been shown to improve NAFLD in children 70 and reduce the risk of hepatic encephalopathy in patients with cirrhosis 71. Due to the complexity in formulating pro- and prebiotic supplements, more research is needed to maximize their effectiveness for ameliorating metabolic disorders associated with obesity. Fecal Microbiota Transplant (FMT) Other alternative therapies include fecal microbiota transplant (FMT) and post-biotics. FMT is the transfer of fecal slurries to a recipient from an approved donor. While FMT is effective in ~90% cases of Clostridium difficile infection 72, it’s use for other diseases in humans is still under investigation. Intriguingly, FMT from bariatric patients results in an improved metabolic profile in mice 73. Unfortunately, few reports exist for use of FMT in relation to metabolic disease. However it was shown by Vrieze et al. that FMT improved symptoms related to insulin resistance in men with metabolic syndrome 74. Along with further study for metabolic disease, well-defined safety practices are needed for the use of FMT (reviewed in 75). Postbiotics More recently, research has focused largely on metabolomics and the introduction of “postbiotics” which are new formulations containing purified microbial metabolites or bacterial components that have a defined benefit to the host, as opposed to live bacteria in probiotics. Postbiotics may become a popular treatment option, because this targeted approach involves small, bioactive molecules that have a defined and specific function, without the potential adverse side effects live bacteria may promote. For example, ex vivo culture with the probiotic Lactobacillus plantarum NCIMB8826 elicited an undesired immune response, but the culture media protected against Salmonella-mediated TNF secretion from intestinal mucosal explants 76. The use of postbiotics would bypass adverse effects promoted by unknown processes triggered by probiotic formulations or potential pathogens delivered via FMT. Summary Obesity and metabolic disease delve from various underlying causes, including genetics and environmental factors, making appropriate and effective treatments difficult to identify. The emergency of high throughput sequencing has recently made it possible to examine the intestinal microbiome in the context of obesity. Understanding how the microbiota structure and function changes under states of obesity as well as bariatric surgery may resolve the role of the microbiome in regulating host metabolic set-points, likely including interactions with the endocrine and nervous systems. The use of 16s rRNA amplicon sequencing is now routinely performed by many labs, but offers only limited information regarding the microbial members present under certain conditions such as obesity, malnutrition, and RYGB surgery. It does not provide information regarding the function of key microbial species that drive host outcome. Without this information it is difficult to determine which strains to examine for potentially vital host-microbe interactions. The complex and individualized nature of obesity presents another obstacle in understanding how we can utilize the microbial organ. Which organ and what pathways do we target and how do we determine this on an individual basis? However, with further advances in the field and employment of available technologies, such as metagenomics and metabolomics, keystone microbes should be better identified and interaction with the host understood. This will allow for the creation of a database of potential pathobionts to target in order to modulate the microbial community. Conversely, these techniques can also be applied to beneficial microbes to understand how we can utilize them for developing more effective prebiotic, probiotic, or postbiotic therapies. Key Points Shifts in the gut microbiome are inseparably associated with the development of obesity and comorbidities. Transfer of dysbiotic microbial communities confers disease phenotypes in recipients, supporting a central role for microbe-mediated regulation of metabolism. Bariatric surgery, the most effective treatment for morbid obesity, results in rapid changes in the gut microbiota, with concurrent improvements in metabolic parameters Deeper understanding of host-microbe interactions may hold promise in the treatment of obesity, which remains a global epidemic. Disclosure Statement: This work was supported by NIH NIDDK DK42086 (DDRCC), DK097268 T32DK07074 to KBM, F32DK105728-01A1 to JFP. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. 1 Flegal KM Epidemiologic aspects of overweight and obesity in the United States Physiol Behav 2005 86 5 599 602 10.1016/j.physbeh.2005.08.050 16242735 2 Sturm R Hattori A Morbid obesity rates continue to rise rapidly in the United States Int J Obes (Lond) 2013 37 6 889 891 10.1038/ijo.2012.159 22986681 3 Bray GA Obesity: a time bomb to be defused Lancet (London, England) 1998 352 9123 160 161 10.1016/S0140-6736(98)22029-0 4 de Onis M Blossner M Borghi E Global prevalence and trends of overweight and obesity among preschool children Am J Clin Nutr 2010 92 5 1257 1264 10.3945/ajcn.2010.29786 20861173 5 Ogden CL Carroll MD Curtin LR McDowell MA Tabak CJ Flegal KM Prevalence of overweight and obesity in the United States, 1999–2004 JAMA 2006 295 13 1549 1555 10.1001/jama.295.13.1549 16595758 6 Turnbaugh PJ Ridaura VK Faith JJ Rey FE Knight R Gordon JI The effect of diet on the human gut microbiome: a metagenomic analysis in humanized gnotobiotic mice Sci Transl Med 2009 1 6 6ra14 10.1126/scitranslmed.3000322 7 Liou AP Paziuk M Luevano J-M Machineni S Turnbaugh PJ Kaplan LM Conserved shifts in the gut microbiota due to gastric bypass reduce host weight and adiposity Sci Transl Med 2013 5 178 178ra41 10.1126/scitranslmed.3005687 8 Amar J Serino M Lange C Involvement of tissue bacteria in the onset of diabetes in humans: evidence for a concept Diabetologia 2011 54 12 3055 3061 10.1007/s00125-011-2329-8 21976140 9 Ley RE Bäckhed F Turnbaugh P Lozupone CA Knight RD Gordon JI Obesity alters gut microbial ecology Proc Natl Acad Sci U S A 2005 102 31 11070 11075 10.1073/pnas.0504978102 16033867 10 Shen W Gaskins HR McIntosh MK Influence of dietary fat on intestinal microbes, inflammation, barrier function and metabolic outcomes J Nutr Biochem 2014 25 3 270 280 10.1016/j.jnutbio.2013.09.009 24355793 11 Ley RE Turnbaugh PJ Klein S Gordon JI Microbial ecology: human gut microbes associated with obesity Nature 2006 444 7122 1022 1023 10.1038/4441022a 17183309 12 Gill SR Pop M Deboy RT Metagenomic analysis of the human distal gut microbiome Science 2006 312 5778 1355 1359 10.1126/science.1124234 16741115 13 Ojeda P Bobe A Dolan K Leone V Martinez K Nutritional modulation of gut microbiota - the impact on metabolic disease pathophysiology J Nutr Biochem 2016 28 191 200 10.1016/j.jnutbio.2015.08.013 26372091 14 Hoffmann C Dollive S Grunberg S Archaea and fungi of the human gut microbiome: correlations with diet and bacterial residents PLoS One 2013 8 6 e66019 10.1371/journal.pone.0066019 23799070 15 Borgo F Verduci E Riva A Relative Abundance in Bacterial and Fungal Gut Microbes in Obese Children: A Case Control Study Child Obes 3 2016 10.1089/chi.2015.0194. 16 de Araújo TV Andrade EF Lobato RV Effects of beta-glucans ingestion (Saccharomyces cerevisiae) on metabolism of rats receiving high-fat diet J Anim Physiol Anim Nutr (Berl) 3 2016 10.1111/jpn.12452. 17 Everard A Matamoros S Geurts L Delzenne NM Cani PD Saccharomyces boulardii administration changes gut microbiota and reduces hepatic steatosis, low-grade inflammation, and fat mass in obese and type 2 diabetic db/db mice MBio 2014 5 3 e01011 e01014 10.1128/mBio.01011-14 24917595 18 Schneeberger M Everard A Gómez-Valadés AG Akkermansia muciniphila inversely correlates with the onset of inflammation, altered adipose tissue metabolism and metabolic disorders during obesity in mice Sci Rep 2015 5 16643 10.1038/srep16643 26563823 19 Everard A Belzer C Geurts L Cross-talk between Akkermansia muciniphila and intestinal epithelium controls diet-induced obesity Proc Natl Acad Sci U S A 2013 110 22 9066 9071 10.1073/pnas.1219451110 23671105 20 Devkota S Wang Y Musch MW Dietary-fat-induced taurocholic acid promotes pathobiont expansion and colitis in Il10−/− mice Nature 2012 487 7405 104 108 10.1038/nature11225 22722865 21 Subramanian S Huq S Yatsunenko T Persistent gut microbiota immaturity in malnourished Bangladeshi children Nature 2014 510 7505 417 421 10.1038/nature13421 24896187 22 David LA Maurice CF Carmody RN Diet rapidly and reproducibly alters the human gut microbiome Nature 2014 505 7484 559 563 10.1038/nature12820 24336217 23 Carmody RN Gerber GK Luevano JM Diet Dominates Host Genotype in Shaping the Murine Gut Microbiota Cell Host Microbe 2014 17 1 72 84 10.1016/j.chom.2014.11.010 25532804 24 Carmody RN Gerber GK Luevano JM Diet dominates host genotype in shaping the murine gut microbiota Cell Host Microbe 2015 17 1 72 84 10.1016/j.chom.2014.11.010 25532804 25 Sonnenburg ED Smits SA Tikhonov M Higginbottom SK Wingreen NS Sonnenburg JL Diet-induced extinctions in the gut microbiota compound over generations Nature 2016 529 7585 212 215 10.1038/nature16504 26762459 26 Howe A Ringus DL Williams RJ Divergent responses of viral and bacterial communities in the gut microbiome to dietary disturbances in mice ISME J 2016 10 5 1217 1227 10.1038/ismej.2015.183 26473721 27 Tadross JA le Roux CW The mechanisms of weight loss after bariatric surgery Int J Obes (Lond) 2009 33 Suppl 1 S28 S32 10.1038/ijo.2009.14 19363504 28 Islam KBMS Fukiya S Hagio M Bile acid is a host factor that regulates the composition of the cecal microbiota in rats Gastroenterology 2011 141 5 1773 1781 10.1053/j.gastro.2011.07.046 21839040 29 Begley M Gahan CGM Hill C The interaction between bacteria and bile FEMS Microbiol Rev 2005 29 4 625 651 10.1016/j.femsre.2004.09.003 16102595 30 Kashyap SR Daud S Kelly KR Acute effects of gastric bypass versus gastric restrictive surgery on beta-cell function and insulinotropic hormones in severely obese patients with type 2 diabetes Int J Obes (Lond) 2010 34 3 462 471 10.1038/ijo.2009.254 20029383 31 Yan M Song M-M Bai R-X Cheng S Yan W-M Effect of Roux-en-Y gastric bypass surgery on intestinal Akkermansia muciniphila World J Gastrointest Surg 2016 8 4 301 307 10.4240/wjgs.v8.i4.301 27152136 32 Furet J-P Kong L-C Tap J Differential adaptation of human gut microbiota to bariatric surgery-induced weight loss: links with metabolic and low-grade inflammation markers Diabetes 2010 59 12 3049 3057 10.2337/db10-0253 20876719 33 Zhang H DiBaise JK Zuccolo A Human gut microbiota in obesity and after gastric bypass Proc Natl Acad Sci U S A 2009 106 7 2365 2370 10.1073/pnas.0812600106 19164560 34 Carvalho BM Guadagnini D Tsukumo DML Modulation of gut microbiota by antibiotics improves insulin signalling in high-fat fed mice Diabetologia 2012 55 10 2823 2834 10.1007/s00125-012-2648-4 22828956 35 Everard A Lazarevic V Gaïa N Microbiome of prebiotic-treated mice reveals novel targets involved in host response during obesity ISME J 2014 8 10 2116 2130 10.1038/ismej.2014.45 24694712 36 Baptissart M Vega A Maqdasy S Bile acids: from digestion to cancers Biochimie 2013 95 3 504 517 10.1016/j.biochi.2012.06.022 22766017 37 Watanabe M Houten SM Mataki C Bile acids induce energy expenditure by promoting intracellular thyroid hormone activation Nature 2006 439 7075 484 489 10.1038/nature04330 16400329 38 Swann JR Want EJ Geier FM Systemic gut microbial modulation of bile acid metabolism in host tissue compartments Proc Natl Acad Sci U S A 2011 108 Suppl 4523 4530 10.1073/pnas.1006734107 20837534 39 Inagaki T Moschetta A Lee Y-K Regulation of antibacterial defense in the small intestine by the nuclear bile acid receptor Proc Natl Acad Sci U S A 2006 103 10 3920 3925 10.1073/pnas.0509592103 16473946 40 Palazzo M Balsari A Rossini A Activation of enteroendocrine cells via TLRs induces hormone, chemokine, and defensin secretion J Immunol 2007 178 7 4296 4303 http://www.ncbi.nlm.nih.gov/pubmed/17371986 Accessed May 26, 2016 17371986 41 Dirksen C Jørgensen NB Bojsen-Møller KN Gut hormones, early dumping and resting energy expenditure in patients with good and poor weight loss response after Roux-en-Y gastric bypass Int J Obes (Lond) 2013 37 11 1452 1459 10.1038/ijo.2013.15 23419600 42 le Roux CW Borg C Wallis K Gut hypertrophy after gastric bypass is associated with increased glucagon-like peptide 2 and intestinal crypt cell proliferation Ann Surg 2010 252 1 50 56 10.1097/SLA.0b013e3181d3d21f 20562614 43 Batterham RL Cowley MA Small CJ Gut hormone PYY3-36 physiologically inhibits food intake Nature 2002 418 6898 650 654 10.1038/nature00887 12167864 44 Semova I Carten JD Stombaugh J Microbiota regulate intestinal absorption and metabolism of fatty acids in the zebrafish Cell Host Microbe 2012 12 3 277 288 10.1016/j.chom.2012.08.003 22980325 45 Leone V Gibbons SM Martinez K Effects of Diurnal Variation of Gut Microbes and High-Fat Feeding on Host Circadian Clock Function and Metabolism Cell Host Microbe 2015 17 5 681 689 10.1016/j.chom.2015.03.006 25891358 46 Schwiertz A Taras D Schäfer K Microbiota and SCFA in Lean and Overweight Healthy Subjects Obesity 2010 18 1 190 195 10.1038/oby.2009.167 19498350 47 De Vadder F Kovatcheva-Datchary P Goncalves D Microbiota-generated metabolites promote metabolic benefits via gut-brain neural circuits Cell 2014 156 1–2 84 96 10.1016/j.cell.2013.12.016 24412651 48 Karlsson FH Tremaroli V Nookaew I Gut metagenome in European women with normal, impaired and diabetic glucose control Nature 2013 498 7452 99 103 10.1038/nature12198 23719380 49 Karastergiou K Fried SK Xie H Distinct developmental signatures of human abdominal and gluteal subcutaneous adipose tissue depots J Clin Endocrinol Metab 2013 98 1 362 371 10.1210/jc.2012-2953 23150689 50 Cani PD Delzenne NM Gut microflora as a target for energy and metabolic homeostasis Curr Opin Clin Nutr Metab Care 2007 10 6 729 734 10.1097/MCO.0b013e3282efdebb 18089955 51 Caesar R Tremaroli V Kovatcheva-Datchary P Cani PD Bäckhed F Crosstalk between Gut Microbiota and Dietary Lipids Aggravates WAT Inflammation through TLR Signaling Cell Metab 2015 22 4 658 668 10.1016/j.cmet.2015.07.026 26321659 52 Anhê FF Roy D Pilon G A polyphenol-rich cranberry extract protects from diet-induced obesity, insulin resistance and intestinal inflammation in association with increased Akkermansia spp. population in the gut microbiota of mice Gut 2015 64 6 872 883 10.1136/gutjnl-2014-307142 25080446 53 Collins B Hoffman J Martinez K A polyphenol-rich fraction obtained from table grapes decreases adiposity, insulin resistance and markers of inflammation and impacts gut microbiota in high-fat-fed mice J Nutr Biochem 2016 31 150 165 10.1016/j.jnutbio.2015.12.021 27133434 54 Swartz TD Duca FA de Wouters T Sakar Y Covasa M Up-regulation of intestinal type 1 taste receptor 3 and sodium glucose luminal transporter-1 expression and increased sucrose intake in mice lacking gut microbiota Br J Nutr 2012 107 5 621 630 10.1017/S0007114511003412 21781379 55 Duca FA Swartz TD Sakar Y Covasa M Increased oral detection, but decreased intestinal signaling for fats in mice lacking gut microbiota PLoS One 2012 7 6 e39748 10.1371/journal.pone.0039748 22768116 56 Wichmann A Allahyar A Greiner TU Microbial modulation of energy availability in the colon regulates intestinal transit Cell Host Microbe 2013 14 5 582 590 10.1016/j.chom.2013.09.012 24237703 57 Rabot S Membrez M Bruneau A Germ-free C57BL/6J mice are resistant to high-fat-diet-induced insulin resistance and have altered cholesterol metabolism FASEB J 2010 24 12 4948 4959 10.1096/fj.10-164921 20724524 58 Sayin SI Wahlström A Felin J Gut microbiota regulates bile acid metabolism by reducing the levels of tauro-beta-muricholic acid, a naturally occurring FXR antagonist Cell Metab 2013 17 2 225 235 10.1016/j.cmet.2013.01.003 23395169 59 Björkholm B Bok CM Lundin A Rafter J Hibberd ML Pettersson S Intestinal microbiota regulate xenobiotic metabolism in the liver PLoS One 2009 4 9 e6958 10.1371/journal.pone.0006958 19742318 60 Gao J He J Zhai Y Wada T Xie W The constitutive androstane receptor is an anti-obesity nuclear receptor that improves insulin sensitivity J Biol Chem 2009 284 38 25984 25992 10.1074/jbc.M109.016808 19617349 61 Dong B Saha PK Huang W Activation of nuclear receptor CAR ameliorates diabetes and fatty liver disease Proc Natl Acad Sci U S A 2009 106 44 18831 18836 10.1073/pnas.0909731106 19850873 62 Moore DD Physiology: A metabolic minuet Nature 2013 502 7472 454 455 10.1038/502454a 24153294 63 Kohsaka A Laposky AD Ramsey KM High-fat diet disrupts behavioral and molecular circadian rhythms in mice Cell Metab 2007 6 5 414 421 10.1016/j.cmet.2007.09.006 17983587 64 Mukherji A Kobiita A Ye T Chambon P Homeostasis in intestinal epithelium is orchestrated by the circadian clock and microbiota cues transduced by TLRs Cell 2013 153 4 812 827 10.1016/j.cell.2013.04.020 23663780 65 Mennigen R Nolte K Rijcken E Probiotic mixture VSL#3 protects the epithelial barrier by maintaining tight junction protein expression and preventing apoptosis in a murine model of colitis Am J Physiol Gastrointest Liver Physiol 2009 296 5 G1140 G1149 10.1152/ajpgi.90534.2008 19221015 66 Wang J Tang H Zhang C Modulation of gut microbiota during probiotic-mediated attenuation of metabolic syndrome in high fat diet-fed mice ISME J 2015 9 1 1 15 10.1038/ismej.2014.99 24936764 67 Nilsson A Johansson E Ekström L Björck I Effects of a brown beans evening meal on metabolic risk markers and appetite regulating hormones at a subsequent standardized breakfast: a randomized cross-over study PLoS One 2013 8 4 e59985 10.1371/journal.pone.0059985 23577078 68 Xiao S Fei N Pang X A gut microbiota-targeted dietary intervention for amelioration of chronic inflammation underlying metabolic syndrome FEMS Microbiol Ecol 2014 87 2 357 367 10.1111/1574-6941.12228 24117923 69 Ettinger G MacDonald K Reid G Burton JP The influence of the human microbiome and probiotics on cardiovascular health Gut Microbes 2014 5 6 719 728 10.4161/19490976.2014.983775 25529048 70 Alisi A Bedogni G Baviera G Randomised clinical trial: The beneficial effects of VSL#3 in obese children with non-alcoholic steatohepatitis Aliment Pharmacol Ther 2014 39 11 1276 1285 10.1111/apt.12758 24738701 71 Dhiman RK Rana B Agrawal S Probiotic VSL#3 reduces liver disease severity and hospitalization in patients with cirrhosis: a randomized, controlled trial Gastroenterology 2014 147 6 1327 1337e3 10.1053/j.gastro.2014.08.031 25450083 72 Bakken JS Fecal bacteriotherapy for recurrent Clostridium difficile infection Anaerobe 2009 15 6 285 289 10.1016/j.anaerobe.2009.09.007 19778623 73 Tremaroli V Karlsson F Werling M Roux-en-Y Gastric Bypass and Vertical Banded Gastroplasty Induce Long-Term Changes on the Human Gut Microbiome Contributing to Fat Mass Regulation Cell Metab 2015 22 2 228 238 10.1016/j.cmet.2015.07.009 26244932 74 Vrieze A Van Nood E Holleman F Transfer of intestinal microbiota from lean donors increases insulin sensitivity in individuals with metabolic syndrome Gastroenterology 2012 143 4 913 916e7 10.1053/j.gastro.2012.06.031 22728514 75 Bakken JS Borody T Brandt LJ Treating Clostridium difficile Infection With Fecal Microbiota Transplantation Clin Gastroenterol Hepatol 2011 9 12 1044 1049 10.1016/j.cgh.2011.08.014 21871249 76 Tsilingiri K Barbosa T Penna G Probiotic and postbiotic activity in health and disease: comparison on a novel polarised ex-vivo organ culture model Gut 2012 61 7 1007 1015 10.1136/gutjnl-2011-300971 22301383

PMC005xxxxxx/PMC5127275.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 101130617 29778 Cancer Cell Cancer Cell Cancer cell 1535-6108 1878-3686 27846390 5127275 10.1016/j.ccell.2016.10.001 NIHMS821678 Article ERK activation globally downregulates miRNAs through phosphorylating exportin-5 Sun Hui-Lung 12 Cui Ri 1 Zhou JianKang 3 Teng Kun-yu 4 Hsiao Yung-Hsuan 5 Nakanishi Kotaro 6 Fassan Matteo 17 Luo Zhenghua 1 Shi Guqin 8 Tili Esmerina 19 Kutay Huban 4 Lovat Francesca 1 Vicentini Caterina 7 Huang Han-Li 10 Wang Shih-Wei 11 Kim Taewan 12 Zanesi Nicola 1 Jeon Young-Jun 1 Lee Tae Jin 1 Guh Jih-Hwa 13 Hung Mien-Chie 121415 Ghoshal Kalpana 4 Teng Che-Ming 2 Peng Yong 3* Croce Carlo M. 116* 1 Department of Cancer Biology and Genetics, Ohio State University, Columbus, OH 43210 2 Pharmacological Institute, College of Medicine, National Taiwan University, Taipei 10051, Taiwan 3 Department of Thoracic Surgery, State Key Laboratory of Biotherapy/ Collaborative Innovation Center for Biotherapy, West China Hospital, Sichuan University, Chengdu 610041, PR China 4 Department of Pathology, Ohio State University, Columbus, OH 43210 5 Department of Human Sciences, Human Nutrition Program, College of Education and Human Ecology, Ohio State University, OH 43210 6 Department of Chemistry and Biochemistry, Ohio State University, OH 43210 7 ARC-NET Research Centre, University and Hospital Trust of Verona, Verona 37126, Italy 8 Division of Medicinal Chemistry and Pharmacognosy, College of Pharmacy, Ohio State University, Columbus, OH 43210 9 Department of Anesthesiology, Ohio State University, Columbus, OH 43210 10 The Ph.D. Program for Cancer Biology and Drug Discovery, College of Medical Science and Technology, Taipei Medical University, Taipei 11031, Taiwan 11 Department of Medicine, Mackay Medical College, New Taipei City 25245, Taiwan 12 Department of Molecular and Cellular Oncology, University of Texas MD Anderson Cancer Center, Houston, TX 77030 13 School of Pharmacy, National Taiwan University, Taipei 10051, Taiwan 14 Graduate Institute of Cancer Biology and Center for Molecular Medicine, China Medical University, Taichung 40402, Taiwan 15 Department of Biotechnology, Asia University, Taichung 41354, Taiwan * Correspondence: carlo.croce@osumc.edu and yongpeng@scu.edu.cn 16 Lead Contact. 9 10 2016 14 11 2016 14 11 2017 30 5 723736 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. SUMMARY MicroRNAs (miRNA) are mostly downregulated in cancer. However, the mechanism underlying this phenomenon and the precise consequence in tumorigenesis remain obscure. Here we show that ERK suppresses pre-miRNA export from the nucleus through phosphorylation of exportin-5 (XPO5) at T345/S416/S497. After phosphorylation by ERK, conformation of XPO5 is altered by prolyl isomerase Pin1, resulting in reduction of pre-miRNA loading. In liver cancer, the ERK-mediated XPO5 suppression reduces miR-122, increases microtubule dynamics, and results in tumor development and drug resistance. Analysis of clinical specimens further showed that XPO5 phosphorylation is associated with poor prognosis for liver cancer patients. Our study reveals a function of ERK in miRNA biogenesis and suggests that modulation of miRNA export has potential clinical implications. In Brief Sun et al. find that ERK phosphorylates XPO5, which induces a Pin1-mediated conformational change that inhibits the ability of XPO5 to load and export pre-miRNA from the nucleus. Phosphorylation of XPO5 is associated with global miRNA downregulation and correlates with poor survival in hepatocellular carcinoma. INTRODUCTION MicroRNAs (miRNA) are a class of small non-coding RNAs that regulate gene expression through inhibition of translation or stability of target mRNAs. The biogenesis of miRNA involves multiple steps, including transcription of primary miRNA (pri-miRNA) by RNA polymerase II, cleavage of pri-mRNA to precursor miRNA (pre-miRNA) by Drosha, nucleocytoplasmic export of pre-miRNA by exportin-5 (XPO5), processing of pre-miRNA to mature miRNA by Dicer, and formation of functional RNA-induced Silencing Complex containing Argonaute (Krol et al., 2010). Depending on the cellular context, a miRNA can function as an oncogene or a tumor suppressor (Croce, 2009). However, global downregulation of miRNA expression has been observed in many tumors (Lu et al., 2005). Interestingly, systematic evaluation of miRNA levels in cancer cell lines demonstrated that many pre-miRNAs are retained in the nucleus (Lee et al., 2008), implying that function of the nuclear-cytoplasmic export machinery may be compromised in tumors. XPO5 is currently considered the indispensable transport receptor for pre-miRNA in most organisms examined (Katahira and Yoneda, 2011). Upon binding to Ran-GTP, XPO5 uses a baseball mitt-like structure to shield the pre-miRNA stem region and the tunnel-like structure to recognize the 2-nucleotide 3′-overhang (Okada et al., 2009). Following GTP hydrolysis, XPO5 releases the pre-miRNA into the cytoplasm. It has been proposed that XPO5 is unique among classical haploinsufficient tumor suppressor genes since partial, but not complete, loss promotes tumor development (Melo et al., 2010). However, the exact role of XPO5 in tumor progression remains elusive. XPO5 is downregulated in lung cancer (Chiosea et al., 2007), but upregulated in breast and prostate cancer (Leaderer et al., 2011). Furthermore, the C/C genotype of rs11077 in the XPO5 3′UTR, which decreases XPO5 protein expression, is associated with poorer prognosis of esophageal and renal cancer (Horikawa et al., 2008) but better prognosis of non-small-cell lung cancer, multiple myeloma, and liver cancer (Liu et al., 2014). Development of hepatocellular carcinoma (HCC) is a multistep sequential process going from chronic inflammation, cirrhosis, primary HCC to metastatic HCC. It has been reported that aberrant miRNA expression drives HCC development and global miRNA downregulation at a later stage promotes metastasis (Wong et al., 2012). Since the reduced miRNA expression in liver cancer is attributed to defective miRNA biogenesis (Lee et al., 2008) and the activity of XPO5 is the rate-limiting step for the production of mature miRNA (Yi et al., 2005), we decided to investigate the role of XPO5 in liver cancer. RESULTS ERK Activation Suppresses Nuclear Export Activity of XPO5 To determine the role of XPO5 in the development of HCC, we first investigated the expression of XPO5 in liver cancer patients. Although the expression level of XPO5 was similar in normal and tumor tissues (Figure S1A), we noticed that more XPO5 was retained in the nucleus of tumor cells (Figures 1A and 1B). Consistent with the findings from IHC staining, Western blot analysis showed nuclear localization of XPO5 in liver tumor lysates (Figure 1C). Because phosphorylation-dependent mobility shift could be detected by phos-tag (Kinoshita et al., 2006) in tumors, we further compared phosphorylation status of XPO5. Interestingly, serine phosphorylation of XPO5 was significantly increased in tumors (Figure 1C). XPO5 overexpression is known to increase nuclear export of pre-miRNA, leading to the enhanced inhibition of miRNA-targeted gene expression. Therefore, we assessed the effects of XPO5 phosphorylation on its function using an indicator luciferase reporter containing eight perfect miR-30a binding sites. Consistent with previous reports (Yi et al., 2005), overexpression of XPO5 exports the endogenous miR-30 to inhibit luciferase expression (Figure 1D). Interestingly, when we co-transfected five different oncogenic serine/threonine kinases, including IKKα, IKKβ, myristoylated-AKT, MEKDD, or CDC2; only constitutively active MEK (MEKDD), which activates ERK, significantly reversed the effects of XPO5 (Figure 1D). In addition, ERK activation increased the slow-migrating and serine phosphorylated form of XPO5 (Figure S1B). The activities of these five kinases were confirmed using previously reported substrates (Figure S1C). Distribution of pre-miR-30 in the cytoplasm was reduced by knockdown of XPO5 and restored by overexpression of shRNA-resistant XPO5 (R-WT) (Figure 1E). In contrast, ERK activation by phorbol 12-myristate 13-acetate (PMA) or co-transfection with MEKDD abrogated the cytoplasmic accumulation of pre-miR-30 (Figures 1E and S1D). Treatment of cells with MEK1 inhibitor, PD98059, was found to reverse the decrease in pre-miR-30 triggered by PMA. Northern blot analysis confirmed inefficient pre-miRNA export in cells with activated ERK (Figure S1E). Eukaryotic translation elongation factor 1A (eEF1A), a nuclear export substrate of XPO5 (Bohnsack et al., 2002), was also used as a readout of XPO5 function. Consistently, MEK activation, whose effect can be reversed by dominant negative (DN) ERK, also blocks XPO5-mediated eEF1A nuclear export (Figure 1F). In summary, these results suggest that the activation of ERK inhibits nuclear export activity of XPO5. ERK Phosphorylates XPO5 at T345, S416 and S497 To determine how ERK inhibits XPO5 activity, we first tested whether ERK interacts with XPO5. Co-immunoprecipitation experiments showed that XPO5 was associated with ERKs upon MEK activation (Figure S2A). Although ERK1 and ERK2 are highly similar and possess identical substrate specificity in vitro, we focused on the ERK2 because its expression exceeds that of ERK1 in most cells, and Erk2 knockout results in embryonic lethality while Erk1 knockout does not (Pages et al., 1999; Yao et al., 2003). The interaction between endogenous ERK2 and XPO5 was observed after PMA stimulation (Figure S2B). After MEK stimulation, the phosphorylated ERK translocates into the nucleus to activate nuclear substrates or forms a dimer to activate cytoplasmic substrates (Casar et al., 2009). As shown in Figure 2A, XPO5 does not interact with phosphorylation site mutant TAYF but still binds to dimerization mutant HEL4A of ERK, suggesting that XPO5 could be a nuclear substrate of ERK. Indeed, ERK co-localizes with XPO5 in the nucleus (Figure 2B). ERK displays specificity for phosphorylation at the serine/threonine-proline (S/T-P) motif. Since the S/T-P motif is found in many proteins, ERK uses a docking motif to ensure its substrate specificity. The best characterized docking sites on ERK are the F-site recruitment site (FRS) and common docking (CD) domain, which responds to the F-site (FX-F/Y-P) and D domain (K/R0–2-X1–6-ϕ-X-ϕ) on substrates (Roskoski, 2012). We found that CD mutant (321N), but not FRS mutant (263A) of ERK abolishes interaction with XPO5 (Figure 2C). Mutational analysis of three potential D domains of XPO5, identified by Eukaryotic Linear Motif database, further revealed that XPO5 residues 284–291 are critical for the association (Figures 2C and S2C). Given the physical interaction between ERK and XPO5 and that calf intestinal alkaline phosphatase can eliminate the mobility shift of XPO5 induced by ERK (Figure S2D), we examined whether XPO5 is a physiological substrate of ERK. Phospho-S/T-P antibody detected XPO5 phosphorylation correlated with ERK activation (Figures S2E and S2F). In vitro kinase assays and mutational analyses suggested that three highly conserved residues T345/S416/S497 of XPO5 are major ERK phosphorylation sites (Figures 2D, S2G and S2H). Mass spectrometry analysis showed that ERK phosphorylated XPO5 at T345/S416 in vitro and S416 in vivo (Figures 2E and S2I). Although phosphorylation at S497 was not detected by mass spectrometry, it was confirmed by a phosphorylation site specific polyclonal antibody (Figure 2F). The original intention was to generate all three phospho-specific antibodies, but only those against p-S416 and p-S497 were successful. Considering the nucleus-enriched specificity and stronger signal intensity (Figure 2F), p-S416 antibody was used as a tool to monitor XPO5 phosphorylation hereafter (Figure S2J). In conclusion, we demonstrate that ERK interacts with XPO5 and phosphorylates it at T345/S416/S497. XPO5 Phosphorylation Globally Downregulates miRNA Expression Before investigating how phosphorylation inhibits XPO5 activity, we first examined the functional importance of XPO5 phosphorylation in miRNA regulation. Comparison of the phosphorylation of ERK and XPO5 in a panel of liver cancer cell lines showed that XPO5 phosphorylation correlated positively with ERK phosphorylation (Figure 3A) and negatively with sensitivity to sorafenib (Figure S3A), a RAF inhibitor approved for the treatment of advanced liver cancer (Bai et al., 2009). To show the global trend of miRNA expression, dot density plot was used to compare expression of 798 miRNAs identified by Nanostring nCounter miRNA expression assay. We found that XPO5-knockdown induced global downregulation of miRNA in the low ERK phosphorylated cell lines, Huh-7 and HepG2, which could be rescued by wild-type XPO5, but not a phosphomimetic 3D mutant (Figures 3B and S3B). Moreover, non-phosphorylatable 3A mutant upregulated miRNA expression more than wild-type XPO5 in the high ERK phosphorylated cell lines, SNU-423 and SK-Hep-1 (Figures S3C and S3D). Based on relative expression levels, we classified the miRNA into three sets: downregulated (<0.67-fold), unchanged (0.67- to 1.5-fold), and upregulated (>1.5-fold) (Figure S3E). The 40 most abundantly expressed miRNAs were also selected to generate the heat map showing the trend of miRNA expression (Figures 3C and S3F–H). The Huh-7 cell line was chosen as a model to study the effect of XPO5 phosphorylation hereafter because it exhibits the lowest ERK and XPO5 phosphorylation levels (Figure 3A). ERK activation by PMA, TGF-α, or MEKDD induced global downregulation of miRNA in the presence of wild-type XPO5, but this effect of ERK activation was attenuated in cells expressing 3A XPO5 (Figure 3D). These results support the notion that ERK activation leads to globally downregulated miRNA levels by phosphorylating XPO5 at T345/S416/S497. Considering that abundance of miRNA is crucial for miRNA function, expression levels of the top ten abundantly expressed miRNAs were validated by qRT-PCR (Figure 3E). MiR-122 is known to dominate the miRNA content in the normal adult liver (Jopling, 2012). In Huh-7 cells, we noticed that miR-4454 and miR-720 are expressed at higher levels than miR-122, but are minimally affected by XPO5. MiRNA precursors fold into stem-loop structures and are recognized by XPO5; however, a large number of similar hairpins that are not pre-miRNAs can also be found in the genome and are called pseudo-hairpins. RNA-fold (Hofacker, 2003) shows these two pre-miRNAs have significantly higher minimum free-energy (Figure 3F), and MiPred (Jiang et al., 2007) does not classify them as pre-miRNA like hairpins (Figure S3I). Consistent with the in silico prediction, RNA-binding protein immunoprecipitation (RIP) and RNA pull-down suggest miR-4454 and miR-720 minimally interact with XPO5, compared to miR-122 (Figures 3G and S3J). Considering that Huh-7 cells resemble normal hepatocytes in expressing high levels of miR-122 (Chang et al., 2004), miR-122 was chosen hereafter as a functional indicator for XPO5 phosphorylation-induced miRNA downregulation. Defective in pre-miR-122 export in 3D mutant expressing Huh-7 cells (Figures 3H and S3K) resulted in the de-repression of miR-122 targets as measured by miR-122 reporter luciferase or EGFP harboring miR-122 binding sites in the respective 3′-UTR (Figures 3I and S3L), indicating that miR-122 function is inhibited upon XPO5 phosphorylation. Suppression of miR-122 Confers Taxol Resistance by Upregulating Septin-9 Because XPO5 phosphorylation decreased miR-122 expression and loss of miR-122 reduced chemosensitivity (Xu et al., 2011; Yang et al., 2011; Yin et al., 2011), we examined whether phosphorylation of XPO5 confers chemoresistance. As shown in Figure 4A, taxol resistance was induced by 3D XPO5 and reversed by miR-122 mimetic. Knockdown of miR-122 (simiR-122) also increased taxol resistance, both in vitro and in vivo (Figures 4A and S4A). We then investigated whether inhibition of the ERK pathway could increase miR-122 and attenuate taxol resistance. As expected, sorafenib, an inhibitor to reduce ERK activation, increased miR-122 expression and sensitized tumor cells to taxol (Figures 4B and 4C and S4B–S4E). However, the 3D XPO5 mutant was resistant to sorafenib-induced miR-122 accumulation and taxol sensitization, as confirmed by the upward shift of the combination index (CI) plot (Figure 4D). The importance of XPO5 phosphorylation in the synergism of ERK inhibition and taxol was further substantiated by DN-ERK and other pharmacological inhibitors of the RAF/MEK/ERK pathway (Figure S4F). miR-122 plays a central role in diverse hepatic functions, as highlighted by its direct role in fat metabolism, tumor suppression, inflammation and fibrosis (Hsu et al., 2012). To find a therapeutic target to circumvent loss-of-miR-122-induced taxol resistance, we noticed that 5 of 30 genes identified by Gene ontology and biological association analysis of the miR-122 targets (Boutz et al., 2011) are associated with microtubule dynamics, including septin-2 (SEPT2), septin-9 (SEPT9), MAP1B, MAP4, and vimentin. Western blot analysis showed that SEPT9 was significantly increased in cells depleted of miR-122 (Figure 4E) and luciferase reporter assay indicated that miR-122 negatively regulates SEPT9 expression by interacting with its 3′ UTR (Figure S4G). Because septins may increase microtubule dynamics through scaffolding microtubule-associated proteins (MAP) and microtubule affinity regulating kinase (MARK) (Spiliotis, 2010), we compared MAP4 phosphorylation in XPO5 stable transfectants and examined the role of MARK. MAP4 phosphorylation was higher in 3D XPO5 cells, resulting in susceptibility to MARK inhibitor (MARKI) (Figures 4F, 4G and S4H). Consistently, miR-122 KO hepatocytes that express more SEPT9 are resistant to sorafenib, but sensitive to the combination of MARKI and taxol (Figures 4H and S4I). We further knocked down four isoforms of MARK in 3D XPO5 cells and found MARK4 depletion significantly inhibited MAP4 phosphorylation to restore its association with tubulin, thereby conferring taxol sensitivity (Figure 4I). Combined with the finding that MARK4 is localized to microtubules (Trinczek et al., 2004) and upregulated in liver cancer (Figure S4J) (Kato et al., 2001), our results suggest that MARK4 could be a therapeutic target to overcome taxol resistance caused by depletion of miR-122. Pin1-Induced Conformational Changes in XPO5 Decrease Pre-miRNA Loading To gain insight on how phosphorylation impairs XPO5 activity, we analyzed the structure of XPO5. Based on WESA (weighted ensemble solvent accessibility) algorithm (Chen and Zhou, 2005) and the secondary structure of XPO5 (Figures S5A and S5B), three ERK phosphorylation sites are fairly accessible and not part of defined helix structures. The structure of XPO5 resembles a flexible wound spring; therefore, small changes in the relative orientation of successive HEAT repeats could cumulatively generate substantial changes in the helicoidal pitch (Okada et al., 2009). It has been shown that prolyl isomerase Pin1 interacts with proteins phosphorylated at S/T-P motifs, thereby controlling their activity by promoting cis-trans isomerization (Liou et al., 2011). Because ERK phosphorylates substrate at S/T-P motifs, we investigated whether Pin1 plays a role in ERK-mediated impairment of XPO5 functions. Treatment of cells with two structurally distinct Pin1 inhibitors, juglone or PiB, was found to reverse the decrease in miR-122 triggered by PMA (Figure 5A). GST pull-down assay shows that ERK activation by ERK2-L4A-MEK1 fusion, a constitutively active and nuclear form of the ERK2, enhances wild-type, but not 3A, XPO5 to associate with Pin-1 (Figure 5B). The interaction between endogenous XPO5 and Pin-1 also increased when ERK was activated (Figure S5C). To investigate if Pin1 changes the conformation of XPO5, we performed partial proteolysis assay with subtilisin, a protease particularly sensitive to substrate conformation (Lu and Zhou, 2007). Incubation with wild-type, but not with catalytically inactive K63A (Zhou et al., 2000) Pin1 mutant protects XPO5 from proteolytic cleavage (Figure S5D). Notably, Pin1 fails to protect 3A XPO5 from proteolysis. Given that Pin1 induces conformational changes in XPO5, we used PP2A, a trans-pSer/Thr-Pro isomer specific phosphatase (Zhou et al., 2000), to examine the cis/trans status of XPO5. Wild-type, but not K63A, Pin1 prevents dephosphorylation of XPO5 by PP2A, indicating that trans-conformation of pSer/Thr-Pro motif on XPO5 was isomerized to cis-conformation in the presence of Pin1 (Figure 5C). EMBOSS analysis (Jones et al., 2002) with WATER algorithm revealed that 416S at HEAT9 and 497S at HEAT10 of XPO5 aligned better with WFYS*PR, an optimal binding sequence for Pin1 (Yaffe et al., 1997) (Figure S5E). Considering the inner helix of HEAT9 provides the closest contact with the stem of pre-miRNA (Okada et al., 2009), we asked whether Pin1-induced conformational changes interfere pre-miRNA binding. Despite the availability of the crystal structure of XPO5, the disorder region in the helix of HEAT10 hampered accurate in silico simulation (Figure S5B). The structure, however, indicates that the possible rearrangement of helices 9 and 10 upon trans-to-cis conversion of P417 and P498, will obstruct pre-miRNA loading (Figure 5D). We therefore investigated whether the Pin1-induced XPO5 conformational changes contribute to ERK-mediated global miRNA downregulation. As shown in Figure 5E, miRNA expression was globally downregulated in cells expressing wild-type Pin1, but this effect was attenuated in the presence of K63A Pin1 or Pin1-binding-defective 3A XPO5. The assembly of the pre-miRNA/XPO5/Ran-GTP complex after Pin1-mediated isomerization was further examined. Because changes in accessibility to the MPM2 antibody have been suggested to reflect conformational changes induced by Pin1 (Stukenberg and Kirschner, 2001), we used it as a tool to evaluate the conformation of XPO5. Consistent with the in silico predictions, the conformation-altered XPO5 was not able to load pre-miR-122 (Figures 5F and S5F), resulting in the inhibition of pre-miR-122 export (Figure S5G), de-repression of miR-122 targets, and resistance to taxol-induced tubulin polymerization (Figures S5H–J). The involvement of Pin1-mediated conformational changes in pre-miRNA loading can also be observed in HepG2 and SK-Hep-1 cells (Figure S5K). Considering that pancreatic cancer exhibits the highest incidence of KRAS mutations and that melanoma has the highest rate of BRAF mutations (Cheng et al., 2013; Neuzillet et al., 2013), we investigated whether ERK-mediated impairment of XPO5 functions also occurs in these two types of cancer. Our results demonstrated that Pin1-induced XPO5 conformational changes contribute to KRASG12V and BRAFV600E-induced impairment of XPO5 function in pancreatic cancer and melanoma, respectively (Figure S5L and S5M). Because phosphorylated XPO5 is enriched in the nucleus (Figure 2F), we also examined the role of Pin1 in XPO5 localization. As shown in Figure 5G, Pin1 loss-of-function partially restored nuclear export of XPO5 and interaction with NUP153, which is essential in the nuclear export of the pre-miRNA-XPO5 complex (Wan et al., 2013). Interestingly, it has been reported that ERK is a physiological nucleoporin kinase that targets the FG repeat regions to disrupt karopherin-nucleoporin interactions. ERK phosphorylates nucleoporin NUP153 at Ser257, Ser320, Ser334, Ser338, Thr369, Thr388, Thr413, Ser516, Ser522 and Ser529. Through generating phospho-specific antibodies, it was shown that ERK directly phosphorylates NUP153 at least at Ser338 and Ser529 (Kosako et al., 2009). NUP153 was not found to directly bind Pin1 (Figure S5N). We found non-phosphorylatable S257A/S320A/S334A/S338A/T369A/T388A/T413A/S516A/S522A/S529A (10A) NUP153 further restored the interaction with XPO5 in cells expressing catalytically inactive Pin1 with activated ERK, resulting in the export of XPO5 and pre-miR-122 (Figures 5G, 5H and S5O). These results support the conclusion that ERK phosphorylates both XPO5 and NUP153 to trap XPO5 and pre-miRNA in the nucleus. XPO5 Phosphorylation Promotes Tumor Development and Associates with Poor Prognosis We showed above that depleting Pin1 or MARK4 may increase taxol sensitivity in XPO5-phosphorylated cells (Figures 4I and S5J). Indeed, inhibition of Pin1 or MARK4 restored taxol-induced tubulin polymerization (Figure S6A), reduced cell viability (Figure 6A), and decreased tumor burden (Figures 6B and S6B). These data support the idea that Pin1 and MARK4 could be therapeutic targets in XPO5 phosphorylation-associated drug resistance. Although the XPO5 phosphomimetic did not significantly increase Huh-7 cell proliferation in vitro (Figure S6C) in 2 days, we noticed that in vivo tumor growth was increased when 3D XPO5 was expressed (Figures 6C and S6D). Interestingly, the cancer stem cell marker CD133 and EPCAM is detected in the liver of young miR-122 KO mice, followed by further increase in the tumors (Hsu et al., 2012; Tsai et al., 2012). Consistently, we found cell surface markers CD133 and EPCAM, functional marker aldehyde dehydrogenase (ALDH) activity, and tumor sphere formation was increased in 3D XPO5 cells, but attenuated when miR-122 was overexpressed or Pin1 was knocked down (Figures 6D, 6E, S6E and S6F). Taken together, our data supports the conclusion that XPO5-dependent downregulation of miRNA contributes to tumor development through increasing cancer stem cell formation and chemoresistance. To examine whether these conclusions are in accord with findings in human tumors, we performed nanostring miRNA profiling and immunoblotting using 12 paired HCC tumor specimens and adjacent normal tissues. XPO5 phosphorylation was found to be associated with global miRNA downregulation (Figures 6F, 6G and S6G). Because miR-122 dominates the miRNA content in the normal liver, we choose it as an indicator and studied its expression using liver tumor tissue array that contained 59 HCC specimens. Consistent with our findings, XPO5 phosphorylation was associated with ERK phosphorylation, miR-122 downregulation, and SEPT9 expression (Figures 6H and S6H). We next correlated our findings with patient survival. The Kaplan-Meier survival curves showed that high p-XPO5 (S416) were associated with poor survival (Figure 6I). Moreover, p-XPO5/p-ERK double positive patients had reduced overall survival relative to the p-XPO5 negative/p-ERK positive group (Figure S6I), supporting that ERK-mediated phosphorylation of XPO5 is associated with worse prognosis. Consistent with the discovery cohort, we also observed that XPO5 phosphorylation correlated with poor survival in the validation cohort (Figure S6J). A sex, age, stage, and adjacent liver status (normal, viral hepatitis, viral hepatitis related cirrhosis)-adjusted multivariate regression analysis suggested that p-XPO5 was independently associated with patient survival (Figure S6K). Taken together, analyses of tumor specimens further strengthened the notion that phosphorylation of XPO5 by ERK downregulates miRNA expression and is associated with poor clinical outcome of liver cancer patients. DISCUSSION Hepatocellular carcinoma is often diagnosed at advanced stages when surgery is not feasible and is also highly resistant to conventional chemotherapy. Therefore, identification of signaling pathways regulating liver carcinogenesis is important in developing targeted therapy. In this study, we demonstrated that ERK phosphorylates XPO5 to impair its ability to export pre-miRNA. However, ERK is also known to increase Drosha and Dicer by phosphorylating DGCR8 and TRBP, a consequence of increasing pro-growth miRNA levels (Herbert et al., 2013; Paroo et al., 2009). Further investigation is therefore needed to establish how ERK affects miRNA biogenesis. It is important to mention that nuclear export by XPO5 has been suggested as a rate-limiting step in miRNA biogenesis. Only XPO5, but not Drosha or Dicer, overexpression enhances inhibition of gene expression by miRNA (Yi et al., 2005). Concordant with our results, most miRNAs were downregulated in PMA-treated myeloid leukemia (Wang et al., 2013) and upregulated in sorafenib-treated HCC (Zhou et al., 2011). But in colon cancer, which has been reported to express an XPO5 frameshift mutant (Melo et al., 2010), expression of miRNAs was upregulated by KRAS and inhibited by MEK/ERK inhibitors (Ragusa et al., 2012). Based on our results and the published literature, an interesting model can be proposed: in cells with wild-type XPO5, activated ERK phosphorylates XPO5, which results in decreased miRNA production. However, if XPO5 is mutated, the effect of ERK on Drosha and Dicer becomes apparent, meaning ERK can globally increase miRNA expression. This model suggests that XPO5 plays a critical role in deciding how ERK reprograms miRNA biogenesis. Shuttling of RNAs and proteins out of the nucleus is essential in maintaining normal cellular function. Cancer cells can usurp this process to stimulate tumor growth and evade apoptosis (Gravina et al., 2014). However, among all of the potential targets in nucleo-cytoplasmic transport, only XPO1 is better understood and is inhibited by selective inhibitor of nuclear export (SINE), which is currently in phase 2 clinical trial (Gerecitano, 2014). An in-depth understanding of miRNA export machinery is needed for the successful design of clinical therapeutics. Our study showed that ERK phosphorylation followed by Pin1-mediated isomerization impairs XPO5 activity. These findings suggest that XPO5 and global miRNA downregulation may become druggable targets, although systems biology and network analysis are necessary to determine the effect of XPO5 restoration in a holistic manner. Pin1 is an attractive target for designing small molecule inhibitors because of its well-defined active site. However, the available Pin1 inhibitors still lack the required specificity and potency (Moore and Potter, 2013). Interestingly, a recent report showed that all-trans retinoic acid (ATRA) directly binds and degrades the active form of Pin1 that is overexpressed in cancer (Wei et al., 2015). Thus, development of more potent and specific Pin1-targeted ATRA derivatives may be a strategy to restore XPO5 function. XPO5 is the only known transporter to export pre-miRNA to the cytoplasm. In our study, we found that cells with XPO5 knockdown or phosphomimetic XPO5 were characterized as having global down-regulation of miRNA expression. Nonetheless, nanostring nCounter analysis revealed that some miRNAs levels were not altered by the impairment of the XPO5. Two splicing-independent mirtrons, miR-1225 and 1228, are also known to be independent of XPO5 for their biogenesis (Havens et al., 2012). We plan to use affected and unaffected pre-miRNA, such as pre-miR-122 and pre-miR-221, as probes to seek additional nuclear export pathways. Among the miRNAs downregulated after XPO5 phosphorylation, we further explored the consequence of miR-122 reduction because it is an abundant liver-specific miRNA and is essential for the maintenance of liver homeostasis. Previous reports suggested that restoring miR-122 may be a strategy to limit HBV/leishmania infection, suppress HCV negative HCC, and prevent the development of nonalcoholic steatohepatitis and cirrhosis (Thakral and Ghoshal, 2015). However, the efficacy of the miRNA mimetic and delivery method still needs to be resolved before advancing to clinical trial. An alternative method to restore miRNA function is to attack downstream targets of the miRNA. Our results suggest that MARK4 may be a therapeutic target to overcome miR-122 loss-induced tumorigenesis and drug resistance. Although the crystal structure of MARK4 is still unknown, combined molecular dynamic simulation and pharmacophore-based virtual screening recently identified six potent compounds specific for MARK4 (Jenardhanan et al., 2014). These lead compounds deserve further testing as potential candidates for miR-122-related diseases. In summary, we propose a model based on the current findings for ERK-induced tumorigenesis in liver cancer cells (Figure 7). In the absence of ERK activation, pre-miR-122 is exported by XPO5 from nucleus to cytoplasm through the nuclear pore complex to produce mature miR-122 to inhibit SEPT9 expression. Un-phosphorylated MAP4 binds to tubulin, causing cells to respond to taxol (left panel). When ERK is activated, XPO5 is phosphorylated at T345/S416/S497, followed by isomerization by Pin1, which impairs its ability to load pre-miR-122. ERK also phosphorylates NUP153 to further inhibit pre-miRNA-XPO5 complex export. SEPT9 then acts as a scaffold for MAP4 and MARK4 and causes phosphorylated MAP4 to detach from microtubules to impart taxol resistance (right panel). The identification of XPO5 as a downstream substrate for ERK links the mitogenic signaling and miRNA nuclear export machinery, and suggests that strategies toward restoring the activity of XPO5 may have therapeutic value. EXPERIMENTAL PROCEDURES Mouse model for tumorigenesis All mice were handled according to the procedures approved by Ohio State University Institutional Laboratory Animal Care and Use Committee. Human liver tumor samples The Ohio State University Institutional Review Board approved this study. Primary human liver cancer and adjacent normal tissues were obtained from the Cooperative Human Tissue Network at the Ohio State University James Cancer Hospital. Informed consent was obtained from all patients. Statistical analysis Each experiment was performed at least three times, and representative data are shown. Data in bar graph are given as the mean ± SEM. Means were checked for statistical difference using Student’s t test or Pearson’s chi-square test, and p values less than 0.05 were considered significant (* p<0.05, ** p<0.01, *** p<0.001). For dot density plot, Mann-Whitney rank sum test was applied. For survival analysis, Kalpan-Meier analysis and log-rank test were applied. Supplementary Material supplement We thank the Ohio State University Comprehensive Cancer Center microscopy, genomics, nucleic acids, flow cytometry, mass spectrometry, and comparative pathology core facility. We thank Dr. M.-C. Hung, J. Dahlberg, P. Sarnow, A. Deiters, Y. Zheng, C.-S. Chen, T. Williams for providing reagents; Dr. K. Huebner, W.-L. Shih, T. Banh for editing; Dr. P. Fadda for nanostring profiling technical support; Dr. S. Palko and Dr. D. Wernicke-Jameson for administrative support. This study was supported by the US National Institutes of Health (R35CA197706 to C.-M. Croce; R01CA193244 to K. Ghoshal); Taiwan Ministry of Science and Technology (NSC99-2320-B400-008-MY3 and NSC 99-2628-B002-024-MY3 to C.-M. Teng; 101-2917-I-564-052 to H.-L. Sun; MOST 105-2632-B-715-001 and MOST 104-2320-B-715-002 to S.-W. Wang); Pelotonia Graduate Student Fellowship (to K.-Y. T); National Key Research & Development Program of China (2016YFA0502204) and National Natural Science Foundation of China (81572739) to Y. Peng. Figure 1 ERK Activation Suppresses XPO5 Function (A) Representative specimens of low and high XPO5 expression in paired human normal and liver tumor tissues on Human HistoArray IMH-342 and IMH-318 by IHC staining of XPO5. Scale bar: 50 μm. (B) Percentage of human normal liver and tumor specimens with the majority of XPO5 localized in the nucleus (N) or cytoplasm (C) was determined. (n=35 per group, p values were calculated by Pearson’s chi-square test). (C) Comparison of localization and phosphorylation of XPO5 in paired human HCC (T) and adjacent benign liver (N). Tissue lysates were subjected to immunoprecipitation (IP) with p-serine or p-tyrosine antibody, or separation of cytoslic and nuclear fraction. Proteins were separated by regular SDS-PAGE or Phos-tag SDS-PAGE then detected by immunoblotting (IB) with XPO5 antibody. (D) 293T cells were transfected with pCMV-Luc-miR-30, pKmyc-XPO5, and the indicated kinase expression vectors for 48 hr before luciferase assay. Data were presented as relative to the cells transfected with the kinase expression vector without XPO5 overexpression. (n=3 per group, data represents mean ± SEM, p values were calculated by Student’s t test). (E) Control (siCTL) cells were pre-treated with or without 30 μM PD98059 for 1 hr then treated with 100 nM PMA for 8 hr, while XPO5-depleted (siXPO5) cells were transfected with siRNA-resistant XPO5 (R-WT) with or without MEKDD for 48 hr. All cells were co-transfected with pCMV-miR-30 because 293T cells do not express Northern blot-detectable miR-30 (Zeng et al., 2002). Cytoplasmic RNA was extracted using PARIS kit then detected by qRT-PCR. Data were presented as relative to the cytoplasmic pre-miR-30 expression in siCTL 293T cells transfected with pCMV-miR-30 (n=3 per group, data represents mean ± SEM, p values were calculated by Student’s t test). (F) siCTL cells were transfected with MEKDD with or without DN-ERK, while siXPO5 cells were transfected with siRNA-resistant XPO5 (R-WT) with or without MEKDD. All cells were co-transfected with GFP-NLS-eEF1A and observed by florescence microscopy after 48 hr. Scale bars: 50 μm. See also Figure S1. Figure 2 ERK Interacts with and Phosphorylates XPO5 (A) Lysates of 293T cells transfected as indicated were subjected to IP with anti-Flag antibody and IB. Interaction was assessed between XPO5 and WT ERK2, dimerization mutant ERK2-H178E/L335A/L338A/L343A/L346A (HEL4A), and MEK-phosphorylation site mutant ERK2-T185A/Y187F (TAYF). (B) Immunofluorescence analysis of XPO5 (green) and ERK (red) of 293T cells untreated or treated with 100 nM PMA for 30 min. DAPI (blue) was used to mark the nucleus. Scale bars: 10 μm. (C) Interaction between WT or mutant ERK2 and XPO5 in lysates from MEKDD expressing 293T cells transfected as indicated was examined by co-IP. EE: R284E/K285E. (D) Lysates of 293T cells transfected with WT or non-phosphorylatable alanine mutants of XPO5 were subjected to IP and IB. 2A: T345A/S416A, 3A: T345A/S416A/S497A. (E) Mass spectrometry detected T345 and S416 phosphorylation in GST-XPO5 (251–550) incubated with active recombinant ERK2. (F) Characterization of p-XPO5 (S416) and p-XPO5 (S497) antibodies by IB analysis of cytosolic, nuclear, and total lysates of 293T cells transfected as indicated with MEKDD, ERK2, and WT or non-phosphorylatable alanine mutant (3A) XPO5. See also Figure S2. Figure 3 Phosphorylation of XPO5 by ERK Globally Downregulates miRNA Expression (A) Basal XPO5 expression, localization, and phosphorylation in a panel of human liver cancer cell lines were analyzed by IB. (B) Dot density plot depicting the expression changes of miRNA in siXPO5 Huh-7 cells stably overexpressing siXPO5-resistant WT (R-WT) or phosphomimetic mutant (R-3D) XPO5. Data were presented as relative to the miRNA expression in siXPO5 cells (n=2). Inset: IB of XPO5. (C) Heat map generated using R program shows relative miRNA expression in Huh-7 stable transfectants as indicated by the green to red key bar at the top of the map. NanoString counts >350 are shown. (D) Dot density plot was used to examine the effect of ERK activation on global miRNA expression. WT or non-phosphorylatable 3A XPO5 expressing Huh-7 cells were treated with 100 nM PMA, 100 ng/ml TGF-α or transfected with MEKDD for 48 hr. Data were presented as relative to the miRNA expression in untreated R-WT Huh-7 cells (n=2). Inset: IB of ERK phosphorylation across treatment groups. WT or non-phosphorylatable 3A XPO5 expressing Huh-7 cells were treated with 100 nM PMA, 100 ng/ml TGF-α for 30 min or transfected with MEKDD for 48 hr. (E) Top ten highly expressed miRNAs in Huh-7 stable transfectants were quantified by qRT-PCR. miRNA expression was normalized to RNU48 (n=3 per group, data represents mean ± SEM, p values were calculated by Student’s t test). (F) Secondary structure of pre-miR-122, pre-miR-720 and pre-miR-4454. MFE structure drawing encoding positional entropy was predicted by RNAfold. (G) Lysates of Huh-7 R-WT stable transfectant were subjected to RNA-binding protein immunoprecipitation (RIP) analysis. Extracts of Huh-7 cells were subjected to IP with IgG or anti-myc tag antibody. Pull-down RNA was analyzed by qRT-PCR using specific primers for pre-miR-122, pre-miR-720, and pre-miR-4454 (n=3 per group, data represents mean ± SEM, p values were calculated by Student’s t test). (H) Comparison of pri-miR-122 and pre-miR-122 expression in Huh-7 stable transfectant (n=3 per group, data represents mean ± SEM, p values were calculated by Student’s t test). Left panel: RNA was isolated by TRIzol for Taqman pri-miR-122 assay. Data were presented as relative to CTL Huh-7 cells. Right panel: Cytoplasmic and nuclear RNA fractions were isolated using PARIS kit then subjected to pre-miR-122 qPCR analysis. Data were presented as relative to the cytoplasmic pre-miR-122 expression in CTL Huh-7 cells. (I) Huh-7 stable transfectants were transfected with a reporter with miR-122-binding sites for 48 hr before luciferase assay (n=3 per group, data represents mean ± SEM, p values were calculated by Student’s t test). See also Figure S3. Figure 4 Depletion of miR-122 Upregulates SEPT9 to Induce Taxol Resistance (A) Vector-transfected and simiR-122-expressing R-WT Huh-7 cells or vector-transfected and miR-122-expressing R-3D Huh-7 cells were treated with 10 μM cisplatin, 50 μM 5-fluorouracil, 1 μM doxorubicin, 0.1 μM taxol, or 0.1 μM vincristine for 48 hr before measuring cell viability by MTT assay. Data were presented as relative to the cells without drug treatment (n=3 per group, data represents mean ± SEM, p values were calculated by Student’s t test). (B) Stable Huh-7 transfectants were incubated with 5 μM sorafenib for 24 hr followed by miR-122 quantification by qRT-PCR. Data were presented as relative to the miR-122 expression in the R-WT cells without sorafenib treatment (n=3 per group, data represents mean ± SEM, p values were calculated by Student’s t test). (C) Representative photograph of engrafted tumors of Huh-7 xenograft treated with taxol with or without sorafenib. (D) The combination index plot for the Huh-7 stable transfectants exposed to fixed molar ratios, 1:100, of taxol to sorafenib for 48 hr (n=3 per group, data represents mean ± SEM, p values were calculated by Student’s t test) (E) IB of four proteins identified as miR-122 targets in Huh-7 stable transfectants. (F) After 24 hr treatment with 30 nM taxol with or without 30 μM MARK inhibitor (MARKI), total cell lysates or tubulin fractions (soluble or polymerized tubulin) from Huh-7 stable transfectants were analyzed by IB. (G) Cell viability of Huh-7 stable transfectants after treatment with 30 nM taxol and 30 μM MARKI for 48 hr was measured by MTT assay. Data were presented as relative to the Huh-7 stable transfectants without drug treatment (n=3 per group, data represents mean ± SEM, p values were calculated by Student’s t test). (H) Primary hepatocytes isolated from WT or miR-122 liver-specific knockout (LKO) mice were treated with 30 nM taxol, 3 μM sorafenib, and 30 μM MARKI for 48 hr before determining cell viability by MTT assay. Data were presented as relative to the hepatocytes without drug treatment (n=3 per group, data represents mean ± SEM, p values were calculated by Student’s t test). (I) Total cell lysates or tubulin fractions of siMARK Huh-7 R-3D stable transfectants treated with 100 nM taxol for 24 hr were analyzed by IB. See also Figure S4. Figure 5 Pin1 Inhibits the Loading of Pre-miRNA to XPO5 (A) Huh-7 cells were pre-treated with or without 5 μM juglone or 5 μM PiB for 1 hr then treated with 100nM PMA for 24 hr before measuring miR-122 by qRT-PCR. Data were presented as relative to the miR-122 expression in the Huh-7 cells without drug treatment (n=3 per group, data represents mean ± SEM, p values were calculated by Student’s t test). (B) Lysates of 293T cells transfected as indicated were incubated with GST-Pin1 then subjected to GST pull-down and IB. (C) Ectopically expressed myc-XPO5 was purified from Pin1 knocked down 293T cells followed by phosphorylation with active ERK2. Purified XPO5 was then incubated with GST, GST-Pin1, or GST-Pin1 (K63A) for 30 min before PP2A treatment for 5 min. (D) Relative position of P417 and P498 (red) in the crystal structure of XPO5 (gray, PDB 3A6P) in complex with RanGTP (cyan) and pre-miR30a (green and pink). Possible movements of helices 9 and 10 when changed to cis form are shown as red arrows. (E) miRNA of Huh-7 Flag-ERK2-MEK1 stable transfectant with Pin1-depleted (siPin1), siRNA-resistant Pin1 (R-WT), catalytic inactive K63A Pin1, or Pin1-binding-defective 3A XPO5 were profiled by nanostring assay. Data were presented as relative to the miRNA expression in R-WT XPO5 Huh-7 cells though dot density plot (n=2) (F) Lysates of Huh-7 stable transfectants expressing Flag-ERK2-MEK1 were incubated with 3′biotin-labeled pre-miR-122, GST-RanQ69L, MPM2 antibody, or HA antibody. Pull-down proteins were subjected to IB. (G) Lysates of Huh-7 stable transfectants were subjected to separation of cytoslic and nuclear fraction or IP with myc or HA antibody before IB. For detection of the NUP153 phosphorylation, cells were labeled with 32P orthophosphate. (H) Immunofluorescence analysis of XPO5 (green) in Huh-7 stable transfectants. DAPI (blue) was used to mark the nucleus. Scale bars: 10 μm. See also Figure S5 Figure 6 XPO5 phosphorylation promotes tumor development and associates with poor prognosis (A) Huh-7 stable transfectants were treated with 100 nM taxol for 48 hr before MTT assay. Data were presented as relative to the cells without taxol treatment (n=3 per group, data represents mean ± SEM, p values were calculated by Student’s t test). (B) Taxol (20 mg/kg once every 3 days, i.v.) was injected into mice for 4 cycles when tumors volume reached 150 mm3. Data shown represent percentage tumor volume relative to xenograft without taxol treatment (n=5 per group, data represents mean ± SEM, p values were calculated by Student’s t test). Arrowheads indicate time points of taxol injection. (C) Tumor volume of engrafted Huh-7 stable transfectants as indicated over time (n=5 per group, data represents mean ± SEM, p values were calculated by Student’s t test). (D) Representative flow cytometry analyses of Huh-7 stable transfectants for ALDH activity using an Aldefluor kit. WT cells treated with the ALDH inhibitor diethylaminobenzaldehyde (DEAB) served as the negative control. The gated cells represent the subpopulation of cells that are positive for ALDH activity. (E) Representative tumor spheres of Huh-7 stable transfectants were imaged under a phase contrast microscope. Scale bars: 50 μm. (F) Dot density plot illustrates the miRNA level in tumor relative to adjacent normal tissue (n=2). The overall average of miRNA levels is marked as mean. Expression levels lower than 0.67 were classified as downregulated miRNA. (G) Paired human normal (N) and liver tumor (T) tissues were analyzed by IB. (H) Correlation between p-ERK, p-XPO5, miR-122, and SEPT9 in 59 human tumor specimens on Human HistoArray IMH-318. p values were calculated by Pearson’s chi-square test. (I) Kaplan-Meier survival curves of liver cancer patients on Human HistoArray IMH-318 grouped by p-XPO5 expression. See also Figure S6. Figure 7 XPO5 Phosphorylation by ERK Leads to Taxol Resistance in Liver Cancer by Decreasing miR-122 Export Under basal condition, pre-miR-122 is exported by XPO5 to produce mature miR-122 to inhibit SEPT9 expression. Un-phosphorylated MAP4 binds to tubulin to maintain taxol sensitivity. When ERK is activated, XPO5 is phosphorylated followed by isomerization by Pin1, which impairs its ability to load pre-miR-122. SEPT9 then acts as a scaffold for MAP4 and MARK4 to cause phosphorylated MAP4 to detach from microtubules therefore conferring taxol resistance. Significance Systemic analysis of miRNAs has revealed that many pre-miRNA are retained in the nucleus of cancer cells. In this study, we have identified that ERK phosphorylation coupled with Pin1-mediated conformational changes in XPO5 inhibits miRNA export. Global downregulation of miRNA expression including that of miR-122 is observed in liver cancer when XPO5 is phosphorylated. Depletion of miR-122 activates MARK4 to increase microtubule dynamics, thereby inducing drug resistance and tumorigenesis. Furthermore, XPO5 phosphorylation correlates with poor prognosis in liver cancer patients. Our findings identified a mechanism of global downregulation of miRNA in liver cancer and targets, Pin1 and MARK4, for therapeutic intervention. Highlights ERK phosphorylation followed by Pin1-mediated isomerization impairs XPO5 activity. Down-regulation of miR-122 leads to taxol resistance through septin-9 and MARK4. XPO5 phosphorylation correlates with poor prognosis in HCC patients. Pin1 and MARK4 are potential targets for clinical intervention in liver cancer. SUPPLEMENTAL INFORMATION Supplemental information, including supplemental experimental procedures and six figures can be found with this article online. AUTHOR CONTRIBUTIONS C. M. C, Y. P, C.-M. T and M.-C. H conceived the project. H.-L. S, R. C., and Y. P. designed experiment. R. C, Y.-H. H, Z. L, N. Z. performed in vivo experiments. K.-Y. T, H. K, and K. G. provided RNA and tissue extracts from HCC patients’ specimens and miR-122 KO mouse livers and hepatocytes. K. N. and G. S. analyzed the crystal structure of XPO5. M. F. and C. V. performed miRNA in situ hybridization and pathology study. J. Z, H.-L. H., T. K., T. J. L, and Y.-J. J provided technical assistance. K.N., E.T., F. L., S.-W. W, J.-H. G, M.-C. H, C.-M. T., Y.P., K.G., and C. M. C edited the manuscript. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. Bai S Nasser MW Wang B Hsu SH Datta J Kutay H Yadav A Nuovo G Kumar P Ghoshal K 2009 MicroRNA-122 inhibits tumorigenic properties of hepatocellular carcinoma cells and sensitizes these cells to sorafenib J Biol Chem 284 32015 32027 19726678 Bohnsack MT Regener K Schwappach B Saffrich R Paraskeva E Hartmann E Gorlich D 2002 Exp5 exports eEF1A via tRNA from nuclei and synergizes with other transport pathways to confine translation to the cytoplasm EMBO J 21 6205 6215 12426392 Boutz DR Collins PJ Suresh U Lu M Ramirez CM Fernandez-Hernando C Huang Y de Abreu RS Le SY Shapiro BA 2011 Two-tiered approach identifies a network of cancer and liver disease-related genes regulated by miR-122 J Biol Chem 286 18066 18078 21402708 Casar B Pinto A Crespo P 2009 ERK dimers and scaffold proteins: unexpected partners for a forgotten (cytoplasmic) task Cell Cycle 8 1007 1013 19279408 Chang J Nicolas E Marks D Sander C Lerro A Buendia MA Xu C Mason WS Moloshok T Bort R 2004 miR-122, a mammalian liver-specific microRNA, is processed from hcr mRNA and may downregulate the high affinity cationic amino acid transporter CAT-1 RNA Biol 1 106 113 17179747 Chen H Zhou HX 2005 Prediction of solvent accessibility and sites of deleterious mutations from protein sequence Nucleic Acids Res 33 3193 3199 15937195 Cheng Y Zhang G Li G 2013 Targeting MAPK pathway in melanoma therapy Cancer Metastasis Rev 32 567 584 23584575 Chiosea S Jelezcova E Chandran U Luo J Mantha G Sobol RW Dacic S 2007 Overexpression of Dicer in precursor lesions of lung adenocarcinoma Cancer Res 67 2345 2350 17332367 Croce CM 2009 Causes and consequences of microRNA dysregulation in cancer Nat Rev Genet 10 704 714 19763153 Gerecitano J 2014 SINE (selective inhibitor of nuclear export)--translational science in a new class of anti-cancer agents J Hematol Oncol 7 67 25281264 Gravina GL Senapedis W McCauley D Baloglu E Shacham S Festuccia C 2014 Nucleo-cytoplasmic transport as a therapeutic target of cancer J Hematol Oncol 7 85 25476752 Havens MA Reich AA Duelli DM Hastings ML 2012 Biogenesis of mammalian microRNAs by a non-canonical processing pathway Nucleic Acids Res 40 4626 4640 22270084 Herbert KM Pimienta G DeGregorio SJ Alexandrov A Steitz JA 2013 Phosphorylation of DGCR8 increases its intracellular stability and induces a progrowth miRNA profile Cell Rep 5 1070 1081 24239349 Hofacker IL 2003 Vienna RNA secondary structure server Nucleic Acids Res 31 3429 3431 12824340 Horikawa Y Wood CG Yang H Zhao H Ye Y Gu J Lin J Habuchi T Wu X 2008 Single nucleotide polymorphisms of microRNA machinery genes modify the risk of renal cell carcinoma Clin Cancer Res 14 7956 7962 19047128 Hsu SH Wang B Kota J Yu J Costinean S Kutay H Yu L Bai S La Perle K Chivukula RR 2012 Essential metabolic, anti-inflammatory, and anti-tumorigenic functions of miR-122 in liver J Clin Invest 122 2871 2883 22820288 Jenardhanan P Mannu J Mathur PP 2014 The structural analysis of MARK4 and the exploration of specific inhibitors for the MARK family: a computational approach to obstruct the role of MARK4 in prostate cancer progression Mol Biosyst 10 1845 1868 24763618 Jiang P Wu H Wang W Ma W Sun X Lu Z 2007 MiPred: classification of real and pseudo microRNA precursors using random forest prediction model with combined features Nucleic Acids Res 35 W339 344 17553836 Jones J Field JK Risk JM 2002 A comparative guide to gene prediction tools for the bioinformatics amateur Int J Oncol 20 697 705 11894112 Jopling C 2012 Liver-specific microRNA-122: Biogenesis and function RNA Biol 9 137 142 22258222 Katahira J Yoneda Y 2011 Nucleocytoplasmic Transport of MicroRNAs and Related Small RNAs Traffic Kato T Satoh S Okabe H Kitahara O Ono K Kihara C Tanaka T Tsunoda T Yamaoka Y Nakamura Y 2001 Isolation of a novel human gene, MARKL1, homologous to MARK3 and its involvement in hepatocellular carcinogenesis Neoplasia 3 4 9 11326310 Kinoshita E Kinoshita-Kikuta E Takiyama K Koike T 2006 Phosphate-binding tag, a new tool to visualize phosphorylated proteins Mol Cell Proteomics 5 749 757 16340016 Kosako H Yamaguchi N Aranami C Ushiyama M Kose S Imamoto N Taniguchi H Nishida E Hattori S 2009 Phosphoproteomics reveals new ERK MAP kinase targets and links ERK to nucleoporin-mediated nuclear transport Nat Struct Mol Biol 16 1026 1035 19767751 Krol J Loedige I Filipowicz W 2010 The widespread regulation of microRNA biogenesis, function and decay Nat Rev Genet 11 597 610 20661255 Leaderer D Hoffman AE Zheng T Fu A Weidhaas J Paranjape T Zhu Y 2011 Genetic and epigenetic association studies suggest a role of microRNA biogenesis gene exportin-5 (XPO5) in breast tumorigenesis Int J Mol Epidemiol Genet 2 9 18 21552306 Lee EJ Baek M Gusev Y Brackett DJ Nuovo GJ Schmittgen TD 2008 Systematic evaluation of microRNA processing patterns in tissues, cell lines, and tumors RNA 14 35 42 18025253 Liou YC Zhou XZ Lu KP 2011 Prolyl isomerase Pin1 as a molecular switch to determine the fate of phosphoproteins Trends Biochem Sci 36 501 514 21852138 Liu S An J Lin J Liu Y Bao L Zhang W Zhao JJ 2014 Single Nucleotide Polymorphisms of MicroRNA Processing Machinery Genes and Outcome of Hepatocellular Carcinoma PLoS One 9 e92791 24676133 Lu J Getz G Miska EA Alvarez-Saavedra E Lamb J Peck D Sweet-Cordero A Ebert BL Mak RH Ferrando AA 2005 MicroRNA expression profiles classify human cancers Nature 435 834 838 15944708 Lu KP Zhou XZ 2007 The prolyl isomerase PIN1: a pivotal new twist in phosphorylation signalling and disease Nature reviews Molecular cell biology 8 904 916 17878917 Melo SA Moutinho C Ropero S Calin GA Rossi S Spizzo R Fernandez AF Davalos V Villanueva A Montoya G 2010 A genetic defect in exportin-5 traps precursor microRNAs in the nucleus of cancer cells Cancer Cell 18 303 315 20951941 Moore JD Potter A 2013 Pin1 inhibitors: Pitfalls, progress and cellular pharmacology Bioorg Med Chem Lett 23 4283 4291 23796453 Neuzillet C Hammel P Tijeras-Raballand A Couvelard A Raymond E 2013 Targeting the Ras-ERK pathway in pancreatic adenocarcinoma Cancer Metastasis Rev 32 147 162 23085856 Okada C Yamashita E Lee SJ Shibata S Katahira J Nakagawa A Yoneda Y Tsukihara T 2009 A high-resolution structure of the pre-microRNA nuclear export machinery Science 326 1275 1279 19965479 Pages G Guerin S Grall D Bonino F Smith A Anjuere F Auberger P Pouyssegur J 1999 Defective thymocyte maturation in p44 MAP kinase (Erk 1) knockout mice Science 286 1374 1377 10558995 Paroo Z Ye X Chen S Liu Q 2009 Phosphorylation of the human microRNA-generating complex mediates MAPK/Erk signaling Cell 139 112 122 19804757 Ragusa M Statello L Maugeri M Majorana A Barbagallo D Salito L Sammito M Santonocito M Angelica R Cavallaro A 2012 Specific alterations of the microRNA transcriptome and global network structure in colorectal cancer after treatment with MAPK/ERK inhibitors J Mol Med (Berl) 90 1421 1438 22660396 Roskoski R Jr 2012 ERK1/2 MAP kinases: structure, function, and regulation Pharmacol Res 66 105 143 22569528 Spiliotis ET 2010 Regulation of microtubule organization and functions by septin GTPases Cytoskeleton (Hoboken) 67 339 345 20517923 Stukenberg PT Kirschner MW 2001 Pin1 acts catalytically to promote a conformational change in Cdc25 Mol Cell 7 1071 1083 11389853 Thakral S Ghoshal K 2015 miR-122 is a unique molecule with great potential in diagnosis, prognosis of liver disease, and therapy both as miRNA mimic and antimir Curr Gene Ther 15 142 150 25537773 Trinczek B Brajenovic M Ebneth A Drewes G 2004 MARK4 is a novel microtubule-associated proteins/microtubule affinity-regulating kinase that binds to the cellular microtubule network and to centrosomes J Biol Chem 279 5915 5923 14594945 Tsai WC Hsu SD Hsu CS Lai TC Chen SJ Shen R Huang Y Chen HC Lee CH Tsai TF 2012 MicroRNA-122 plays a critical role in liver homeostasis and hepatocarcinogenesis J Clin Invest 122 2884 2897 22820290 Wan G Zhang X Langley RR Liu Y Hu X Han C Peng G Ellis LM Jones SN Lu X 2013 DNA-damage-induced nuclear export of precursor microRNAs is regulated by the ATM-AKT pathway Cell Rep 3 2100 2112 23791529 Wang J Liu L Xie L Xiang G Zhou Y 2013 Induction of differentiation-specific miRNAs in TPA-induced myeloid leukemia cells through MEK/ERK activation Int J Mol Med 31 59 66 23175175 Wei S Kozono S Kats L Nechama M Li W Guarnerio J Luo M You MH Yao Y Kondo A 2015 Active Pin1 is a key target of all-trans retinoic acid in acute promyelocytic leukemia and breast cancer Nat Med 21 457 466 25849135 Wong CM Wong CC Lee JM Fan DN Au SL Ng IO 2012 Sequential alterations of microRNA expression in hepatocellular carcinoma development and venous metastasis Hepatology 55 1453 1461 22135159 Xu Y Xia F Ma L Shan J Shen J Yang Z Liu J Cui Y Bian X Bie P 2011 MicroRNA-122 sensitizes HCC cancer cells to adriamycin and vincristine through modulating expression of MDR and inducing cell cycle arrest Cancer Lett 310 160 169 21802841 Yaffe MB Schutkowski M Shen M Zhou XZ Stukenberg PT Rahfeld JU Xu J Kuang J Kirschner MW Fischer G 1997 Sequence-specific and phosphorylation-dependent proline isomerization: a potential mitotic regulatory mechanism Science 278 1957 1960 9395400 Yang F Zhang L Wang F Wang Y Huo XS Yin YX Wang YQ Sun SH 2011 Modulation of the unfolded protein response is the core of microRNA-122-involved sensitivity to chemotherapy in hepatocellular carcinoma Neoplasia 13 590 600 21750653 Yao Y Li W Wu J Germann UA Su MS Kuida K Boucher DM 2003 Extracellular signal-regulated kinase 2 is necessary for mesoderm differentiation Proceedings of the National Academy of Sciences of the United States of America 100 12759 12764 14566055 Yi R Doehle BP Qin Y Macara IG Cullen BR 2005 Overexpression of exportin 5 enhances RNA interference mediated by short hairpin RNAs and microRNAs RNA 11 220 226 15613540 Yin J Tang HF Xiang Q Yu J Yang XY Hu N Lei XY 2011 MiR-122 increases sensitivity of drug-resistant BEL-7402/5-FU cells to 5-fluorouracil via down-regulation of bcl-2 family proteins Pharmazie 66 975 981 22312705 Zeng Y Wagner EJ Cullen BR 2002 Both natural and designed micro RNAs can inhibit the expression of cognate mRNAs when expressed in human cells Mol Cell 9 1327 1333 12086629 Zhou C Liu J Li Y Liu L Zhang X Ma CY Hua SC Yang M Yuan Q 2011 microRNA-1274a, a modulator of sorafenib induced a disintegrin and metalloproteinase 9 (ADAM9) down-regulation in hepatocellular carcinoma FEBS Lett 585 1828 1834 21530512 Zhou XZ Kops O Werner A Lu PJ Shen M Stoller G Kullertz G Stark M Fischer G Lu KP 2000 Pin1-dependent prolyl isomerization regulates dephosphorylation of Cdc25C and tau proteins Molecular cell 6 873 883 11090625

PMC005xxxxxx/PMC5127391.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 0412041 133 Acta Neuropathol Acta Neuropathol. Acta neuropathologica 0001-6322 1432-0533 26197969 5127391 10.1007/s00401-015-1460-x NIHMS830527 Article Functional recovery in new mouse models of ALS/FTLD after clearance of pathological cytoplasmic TDP-43 Walker Adam K. 1 Spiller Krista J. 1 Ge Guanghui 1 Zheng Allen 1 Xu Yan 1 Zhou Melissa 1 Tripathy Kalyan 1 Kwong Linda K. 1 Trojanowski John Q. 12 Lee Virginia M.-Y. vmylee@upenn.edu 12 1 Department of Pathology and Laboratory Medicine, Center for Neurodegenerative Disease Research, Perelman School of Medicine, University of Pennsylvania, Philadelphia, PA 19104, USA 2 Institute on Aging, Perelman School of Medicine, University of Pennsylvania, Philadelphia, PA 19104, USA 22 11 2016 22 7 2015 11 2015 29 11 2016 130 5 643660 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. Accumulation of phosphorylated cytoplasmic TDP-43 inclusions accompanied by loss of normal nuclear TDP-43 in neurons and glia of the brain and spinal cord are the molecular hallmarks of amyotrophic lateral sclerosis (ALS) and frontotemporal lobar degeneration (FTLD-TDP). However, the role of cytoplasmic TDP-43 in the pathogenesis of these neurodegenerative TDP-43 proteinopathies remains unclear, due in part to a lack of valid mouse models. We therefore generated new mice with doxycycline (Dox)-suppressible expression of human TDP-43 (hTDP-43) harboring a defective nuclear localization signal (ΔNLS) under the control of the neurofilament heavy chain promoter. Expression of hTDP-43ΔNLS in these ‘regu-latable NLS’ (rNLS) mice resulted in the accumulation of insoluble, phosphorylated cytoplasmic TDP-43 in brain and spinal cord, loss of endogenous nuclear mouse TDP-43 (mTDP-43), brain atrophy, muscle denervation, dramatic motor neuron loss, and progressive motor impairments leading to death. Notably, suppression of hTDP-43ΔNLS expression by return of Dox to rNLS mice after disease onset caused a dramatic decrease in phosphorylated TDP-43 pathology, an increase in nuclear mTDP-43 to control levels, and the prevention of further motor neuron loss. rNLS mice back on Dox also showed a significant increase in muscle innervation, a rescue of motor impairments, and a dramatic extension of lifespan. Thus, the rNLS mice are new TDP-43 mouse models that delineate the timeline of pathology development, muscle denervation and neuron loss in ALS/FTLD-TDP. Importantly, even after neurodegeneration and onset of motor dysfunction, removal of cytoplasmic TDP-43 and the concomitant return of nuclear TDP-43 led to neuron preservation, muscle re-innervation and functional recovery. Amyotrophic lateral sclerosis (ALS) Frontotemporal dementia (FTD) Frontotemporal lobar degeneration (FTLD) TDP-43 Neurodegeneration Motor neuron Spinal cord Mouse model Introduction Amyotrophic lateral sclerosis (ALS) is a devastating neurodegenerative disease primarily defined by the presence of TAR DNA-binding protein of 43 kDa (TDP-43) pathology in the brain and spinal cord [31]. Moreover, ∼50 % of frontotemporal lobar degeneration (FTLD) patients also develop TDP-43 pathology, with this form of disease known as FTLD-TDP [10, 31]. ALS and FTLD-TDP are therefore now considered to form a spectrum of disorders [12, 13]. Although only ∼10 % of ALS and ∼30 % of FTLD patients have a family history of disease, the majority of available mouse models rely on gene mutations identified in these small minority of ALS and FTLD patients. For example, mutant human SOD1 mice have been used to model ALS for >20 years and although they appear to have a clinical disease course that mirrors human ALS [17], they do not develop robust TDP-43 pathology [36, 41, 45]. Furthermore, no disease modifying treatments using the SOD1 mice have succeeded in clinical trials [14]. New valid mouse models that recapitulate TDP-43 pathology are therefore required for investigation of upstream disease pathogenesis and preclinical testing for both ALS and FTLD-TDP. TDP-43 is a widely expressed and highly conserved protein with diverse functions in RNA metabolism, including RNA translation, splicing and transport [26]. Although primarily located in the nucleus, a small proportion of TDP-43 is normally present in the cytoplasm of cells [4, 47]. However, in postmortem tissue from patients with ALS, FTLD-TDP, or both, affected neurons and glia in the brain and spinal cord lose nuclear TDP-43 and accumulate insoluble phosphorylated TDP-43 in the cytoplasm [31]. Multiple attempts have been made to generate transgenic mouse models that recapitulate these neuropathological features of ALS and FTLD-TDP. For example, expression of human wild type, ALS-linked mutant, or cytoplasmic TDP-43 under the control of broad or brain-specific promoters, including PrP, Thy1, and Camk2a, resulted in rodent models which recapitulate some, but not all, features of ALS/FTLD-TDP [2, 20, 22, 44, 48, 49]. Notably, no or only rare TDP-43 inclusions were detected in these models and no or only modest loss of spinal cord motor neurons occurred. Moreover, many of these animals do not develop a progressive ALS-like phenotype, which is a requirement for future preclinical testing of ALS therapeutics [29]. In addition, since previous rodent models did not successfully recapitulate the key features of disease, the timeline for the progressive TDP-43 nuclear clearance, cytoplasmic accumulation, muscle denervation and neuron loss in vivo has been difficult to define. Our previous efforts to generate transgenic mouse models that develop TDP-43 pathology included crossing doxycycline (Dox)-suppressible Camk2a promoter mice with transgenic mice expressing TDP-43 in which the nuclear localization sequence (NLS) was genetically ablated [22]. These mice accumulated cytoplasmic TDP-43 in the brain but with little pathological TDP-43 aggregates and did not develop an ALS-like phenotype due to lack of expression in the spinal cord [22]. To overcome these limitations, we generated a second generation of TDP-43 mouse models with cytoplasmic TDP-43 pathology and a progressive disease course similar to ALS, using newly generated mice with Dox-suppressible neurofilament heavy chain (NEFH) promoter-driven brain and spinal cord expression. Since it remains uncertain if modulating TDP-43 sub-cellular location and pathology is a valid avenue for therapeutic development, we further used these new mouse models of ALS/FTLD-TDP to investigate if decreasing cytoplasmic hTDP-43 levels, clearance of hTDP-43 pathology and return of nuclear mTDP-43 could ameliorate ALS-like phenotypes after the onset of neuron loss, muscle denervation and motor dysfunction. Materials and methods Generation of NEFH-tTA transgenic mice and bigenic rNLS mice NEFH-tTA transgenic (Tg) mice were generated by injection of a plasmid with the human NEFH promoter isolated from BAC clone (CHORI: RP11-91J21) to drive expression of the tetracycline transactivator (tTA) gene [20] (kindly provided by Xu-Gang Xia, Thomas Jefferson University) into the pronucleus of fertilized eggs from C57BL/6J × C3HeJ F1 matings (Transgenic and Chimeric Mouse Facility of the University of Pennsylvania). Mice were maintained on a mixed C57BL/6J × C3HeJ F1 background. Monogenic NEFH-tTA mice were bred to previously described tetO-hTDP-43-ΔNLS line 4 mice [22], which allow expression of K82A/R83A/K84A mutant human TDP-43 (hTDP-43ΔNLS) and were also maintained on a mixed C57BL/6J × C3HeJ F1 background, to produce non-transgenic (nTg), NEFH-tTA monogenic, hTDP-43ΔNLS monogenic and NEFH-tTA × hTDP-43ΔNLS bigenic mice (referred to as rNLS mice). Similarly, monogenic NEFH-tTA mice were bred to previously described tetO-hTDP-43-WT line 4 and tetO-hTDP-43-WT line 12 mice [22]. Seven original NEFH-tTA Tg founders were produced but four lines were discontinued. The remaining three lines (designated 8, 9B and 43) were studied extensively as described here. Breeding mice and weaned mice on Dox were provided with chow containing 200 mg/kg Dox (Dox Diet #3888, Bio-Serv). To induce expression of hTDP-43ΔNLS, mice were switched to standard chow lacking Dox (Rodent Diet 20 #5053, PicoLab). To re-suppress expression of hTDP-43ΔNLS, Rodent Diet 20 was replaced with Dox Diet. Genotyping was performed using tail DNA [52] as described previously for tetO-hTDP-43ΔNLS mice [22], and with the following primers for all three lines of NEFH-tTA mice: NEFH-tTA forward (5′-TCCACTTTGAGGGGTCTCTG-3′) and NEFH-tTA reverse (5′-AGCATCTCATCACTTCCCTG-3′) [20]. All procedures were performed in accordance with the NIH Guide for the Care and Use of Experimental Animals and studies were approved by the Institutional Animal Care and Use Committee of the University of Pennsylvania. Mouse phenotype monitoring and behavioral analysis Mice were weighed and monitored for tremor and clasping phenotype three times per week. For observation of tremor, mice were held on their backs in the palm of the observer's hand gripped gently between thumb and index finger and both forelimb and hindlimb movements were observed for ∼30 s. The presence of either fast fine tremor at any point in this observation period or resting tremor when at rest in the home cage was recorded as a positive response. Tremor of this nature was never observed in nTg and monogenic littermate control mice. For observation of hindlimb clasping, mice were suspended by the tail ∼30 cm above the cage and slowly lowered. The presence of both hindlimbs held together within 5 s of being raised and maintained for ∼30 s was recorded as a positive response. Persistent hindlimb clasping of this nature was never observed in nTg and monogenic littermate control mice. Mice were tested once weekly for wirehang and rotarod performance. For the wirehang test of grip strength, mice were placed on a wire cage top and gently flipped to be suspended upside-down ∼30 cm above a clean cage. The time to fall was recorded up to 180 s; if still hanging at the end of a session a time of 180 s was recorded. One training session and two test sessions were performed each with >5 min rest between sessions, with the final score being the average time of the two test sessions. For the rotarod test of motor coordination and balance, mice were placed on a rotarod apparatus (Model 7650, Ugo Basile) at a speed of 4 rpm with acceleration up to 40 rpm over 300 s. The time to fall was recorded up to 300 s; if still running at the end of the session a time of 300 s was recorded. One training session and two test sessions were performed each with >30 min rest between sessions, with the final score being the average time of the two test sessions. Upon signs of movement difficulty, mice were provided with moistened chow on the cage floor. Disease end stage in mice was defined as complete hindlimb paralysis, a loss of >30 % body weight from peak weight or the inability to right within 15 s when placed on their backs. Generally, rNLS bigenic mice did not develop complete hindlimb paralysis or persistent loss of righting reflex prior to reaching the weight cut-off for end stage using the simple supportive measures reported here. Mouse tissue collection for NMJ and spinal cord motor neuron quantification Mice were deeply anesthetized using ketamine/xylazine and intracardially perfused with ∼15 mL room temperature phosphate-buffered saline (PBS) followed by ∼30 mL of 10 % formalin. The tibialis anterior (TA) muscles were washed in PBS overnight, and spinal cords were post-fixed in 10 % formalin overnight and then all were processed for cryoprotective embedding. For visualization of neuromuscular junctions (NMJs) in whole TA muscles, BTX conjugated to AlexaFluor-488 (Invitrogen, Carlsbad, CA, USA) 1:500 was added with rabbit anti-VAChT 1:16,000 (gift of C. Henderson). The percentage of NMJ innervation was determined by dividing the total number of areas of overlap between VAChT and BTX signals (total number innervated endplates) by the number of areas of BTX signal (total number of endplates) on consecutive longitudinal 35 μm cryosections of muscles. We observed 700–1000 NMJs per TA. For the quantification of motor neuron numbers, motor neurons were counted in every fifth transverse 20 μm cryo-sections of the L4 and L5 levels of the spinal cord stained for VAChT and a mouse anti-human TDP-43 monoclonal antibody (MAb) (0.06 μg/mL, clone 5104, CNDR) [25]. Mouse tissue collection for biochemistry and IHC Mice were deeply anesthetized using ketamine/xylazine and intracardially perfused with PBS as above. Tissue was rapidly dissected and weighed and either immediately frozen on dry ice and stored at −80 °C for use in biochemistry or post-fixed overnight in 10 % formalin. Post-fixed tissues were rinsed in TBS, pH 7.4 (50 mM Tris–HCl and 150 mM NaCl), embedded in paraffin, and sectioned at 6 μm. Generally, each mouse brain was bisected and the left hemisphere was dissected into different regions for biochemistry before freezing, and the right hemisphere used for immunohistochemistry (IHC) and immunofluorescence (IF). Human tissue samples Frozen and formalin-fixed, paraffin-embedded brain and spinal cord tissues were from the Center for Neurodegenerative Disease Research Brain Bank at the University of Pennsylvania. Human tissues were obtained at autopsy and processed for neuropathological assessment as described previously [43]. Informed consent was obtained in accordance with the University of Pennsylvania Institutional Review Board guidelines. Cortical and spinal cord tissues of ALS and FTLD patients with a high burden of TDP-43 pathology (FTLD-TDP) were used in experiments. Human tissue for biochemistry was sequentially extracted as described previously [37]. The urea-soluble fraction from frontal cortex of a 68 year-old female patient with a neuropathological diagnosis of FTLD-TDP was used for immunoblotting. Paraffin 6 μm sections of motor cortex and lumbar spinal cord from a 66 year-old female patient with a neuropathological diagnosis of ALS were used for IHC. IHC and IF For IHC and IF on paraffin-embedded sections, sections were deparaffinized in xylene and rehydrated through a series of decreasing concentration of ethanol. For IHC only, sections were incubated in 5 % H2O2 in methanol for 30 min to block endogenous peroxidases. Sections for all antibodies except anti-ubiquitin were subjected to antigen retrieval by microwaving at 95 °C for 15 min in 1 % antigen unmasking solution pH 6.0 (Vector Laboratories). Sections were allowed to cool and then blocked in 2 % FBS in 0.1 M Tris and incubated in primary antibodies for ∼16 h. Primary antibodies used for IHC and IF were a rabbit anti-C-terminal TDP-43 polyclonal antibody (PAb) for detection of both human and mouse TDP-43 (0.15 μg/mL, CNDR C1039) [23], a rabbit anti-N-terminal TDP-43 PAb for detection of both human and mouse TDP-43 (0.22 μg/mL, CNDR N1065) [23], a mouse anti-internal TDP-43 MAb for detection of both human and mouse TDP-43 (0.03 μg/mL, clone 5117, CNDR) [25], a mouse anti-human TDP-43 MAb (0.06 μg/mL, clone 5104, CNDR) [25], a mouse anti-human TDP-43 MAb (0.06 μg/ mL, clone 5123, CNDR) [25], a rabbit anti-mouse TDP-43 (0.10 μg/mL, CNDR 2340-181, affinity-purified IgG raised by immunizing rabbits with a peptide corresponding to amino acids 181-190 of mTDP-43), a rabbit anti-phospho-S403/404 TDP-43 PAb (1:2000, CosmoBio TIP-PTD-P05), a rat anti-phospho-S409/410 TDP-43 MAb (1:200, TAR5P-1D3, Ascenion, Munich, Germany) [30], a rabbit anti-phospho-S409/410 TDP-43 PAb (1:5000, CosmoBio TIP-PTD-P01), a rat anti-phospho-NFM + H MAb (1:200, clone TA51, CNDR) [42], a mouse anti-ubiquitin MAb (1:5000, clone Ubi-1, Millipore 1510), a rabbit anti-NeuN MAb (1:1000, clone EPR12763, Abcam 177487), a rabbit anti-GFAP PAb (1:2000, Dako z0334), a rat anti-GFAP MAb (1:2000, clone 2.2B10, CNDR) [28], a rabbit anti-Iba1 PAb (1:1000, Wako 019-19741), and a rabbit anti-Olig-2 PAb (1:250, Millipore 9610). For IHC, a Vectastain ABC kit was used with biotinylated anti-mouse, anti-rabbit or anti-rat secondary antibodies for detection with 0.05 % 3,3′-diaminobenzidine tetrahydrochloride hydrate (Sigma D5637) with 0.003 % H2O2 in 0.1 M phosphate buffer pH 7.4. Sections were counterstained with haemotoxylin, dehydrated through ethanol and xylenes, and mounted with Cytoseal 60 mounting medium (Vector Laboratories). For IF, sections were incubated with goat anti-mouse, anti-rabbit or anti-rat 488- or 594-conjugated AlexaFluor secondary antibodies (1:1000, Molecular Probes) and mounted using Fluoromount Gold plus DAPI (Southern Biotech). Images were acquired using a Nikon Eclipse Ni inverted microscope with DS-Fi2 camera (brightfield) or DS-Qi1Mc camera (fluorescence). Preparation of mouse brain lysates Tissues were thawed on ice and then sonicated in 5× v/w RIPA buffer (50 mM Tris, 150 mM NaCl, 1 % NP-40, 5 mM EDTA, 0.5 % sodium deoxycholate, and 0.1 % SDS, pH 8.0) containing 1 mM PMSF and protease and phosphatase inhibitor cocktails (Sigma). Samples were centrifuged at 4 °C, 100,000g for 30 min and the supernatant taken as the RIPA-soluble fraction. The pellet was washed by sonication with RIPA buffer as above. This supernatant was discarded and the pellet sonicated in 2× v/w urea buffer (7 M urea, 2 M thiourea, 4 % CHAPS, and 30 mM Tris, pH 8.5) and centrifuged at 22 °C, 100,000g for 30 min. This supernatant was taken as the RIPA-insoluble/ urea-soluble fraction. Protein concentrations of the RIPA-soluble fractions were determined using the bicinchoninic acid protein assay (Pierce). Immunoblotting Twenty micrograms of RIPA-soluble protein and a volume equivalent to 7.5-times the amount of the corresponding urea-soluble fraction (normalized to non-specific bands detected by Ponceau S staining of membranes for urea-soluble spinal cord samples) were analyzed by 10 or 12 % SDS-PAGE with nitrocellulose membrane (Bio-Rad). Antibodies used for immunoblotting were C1039, 5104, 2340-181, anti-phospho-S403/404 TDP-43 PAb and TAR5P-1D3, a rabbit anti-neurofilament light chain PAb (1:500, phospho-independent NFL, CNDR) [5], a mouse anti-neurofilament medium chain MAb (1:500, phospho-independent NFM, clone RMO44, CNDR) [27] and a mouse anti-GAPDH MAb (1:10,000, clone 6C5, Advanced Immunochemical). Signals were detected using goat anti-mouse, anti-rabbit or anti-rat IRDye-680 or IRDye-800 conjugated secondary PAbs (Li-Cor or Rockland) with imaging using a Li-Cor Odyssey imaging system. Band infrared fluorescent signals were quantified using Li-Cor Image Studio Version 2.0. Measurement of cortical thickness Cortical thickness was measured from the pial surface to the dorsal corpus callosum from images of DAPI-labeled or haemotoxylin and eosin-stained brain sections at the level of the motor cortex (Bregma 1.10) according to the standard mouse brain atlas [33]. Statistical analyses and experimental design Statistical significance between two values was determined using a two-tailed t test, analyses of three or more values were performed by one-way ANOVA with Bonferroni's post hoc test using Prism 4 (GraphPad). P < 0.05 was considered statistically significant. All results are presented as mean ± standard error of the mean (SEM) and in figures *p < 0.05, **p < 0.01, and ***p < 0.001. The investigators were blinded for genotype for mouse wirehang and rotarod testing, and for NMJ and neuron counts. Mice were randomly allocated to groups for time point analysis and for Dox recovery experiments. Both sexes were analyzed. Littermate nTg and monogenic animals were used as controls. Littermate control and bigenic groups were matched to include equal sexes. Mice were excluded from all analyses when non-neurological disease was present: three rNLS8 mice and two rNLS9B mice developed skin lesions/ fighting wounds, one rNLS9B mouse was born microcephalic, and one rNLS8 mouse died of unknown cause at 3 weeks off Dox. Results The NEFH promoter drives expression of hTDP-43ΔNLS in brain and spinal cord To drive the expression of hTDP-43 in neurons with large-caliber axons of the brain and spinal cord, we generated lines of mice in which the tetracycline transactivator protein (tTA) is expressed under the control of the human NEFH promoter [18, 20]. Three NEFH-tTA mouse lines (designated lines 8, 9B and 43) were crossed with our previously developed tetO-hTDP-43ΔNLS line 4 mice, in which the nuclear localization sequence (NLS) is genetically ablated leading to cytoplasmic accumulation of hTDP-43 [22]. Therefore, in the resulting NEFH-tTA/tetO-hTDP-43ΔNLS bigenic animals (designated as ‘regulatable’ or rNLS mice), hTDP-43ΔNLS expression is controlled by the tTA gene product in a doxycycline (Dox)-dependent manner, such that hTDP-43ΔNLS expression is suppressed in the presence of Dox (Fig. 1a). rNLS mice used in all experiments were bred and maintained in the presence of Dox-containing chow until approximately 5 weeks of age, when Dox was removed to allow hTDP-43ΔNLS expression. rNLS mice showed widespread, high levels of hTDP-43ΔNLS expression in brain and spinal cord as early as 1 week off Dox, but not in peripheral tissues including spleen, kidney and liver (Fig. 1b–e). Line rNLS8 was chosen for the most thorough characterization due to the higher percentage of neurons expressing hTDP-43, although similar results were also confirmed in lines rNLS9B and rNLS43 (Fig. 1c; Fig. S1a). rNLS8 mice expressed full length hTDP-43ΔNLS in the cytoplasm with frequent structures resembling pathological TDP-43 inclusions, as detected by antibodies against the N-terminus and C-terminus as well as human specific TDP-43 (Fig. 1f–h). We quantified expression in layer V of the motor cortex in rNLS8 mice, where greater than 95 % of NeuN-positive neurons were positive for hTDP-43 by 1 week off Dox, and this level of expression was maintained over time (Fig. S1b). Despite NEFH being a neuron-specific promoter, unexpectedly 5–10 % of GFAP-positive astrocytes also expressed hTDP-43ΔNLS, although Iba1-positive micro-glia and Olig2-positive oligodendrocytes rarely expressed the transgene (Fig. 1i; Fig. S1c–e). Insoluble phosphorylated TDP-43 accumulates in brain and spinal cord of rNLS mice To confirm the presence of insoluble TDP-43 aggregates and to investigate the consequences of mislocalized cytoplasmic hTDP-43 in the rNLS mice, brain and spinal cord samples were separated into RIPA-soluble and RIPA-insoluble (urea-soluble) fractions for biochemical analysis. hTDP-43ΔNLS was robustly detected in the RIPA-soluble protein fraction of cortex of rNLS8 mice after 1 week off Dox, with maximal expression 2–4 weeks after Dox removal (Fig. 2a). Quantitative analysis revealed total TDP-43 levels as more than 10-fold higher in rNLS mice at 2–4 weeks compared to non-transgenic (nTg) and monogenic littermates, with a concomitant ∼50 % decrease in endogenous mTDP-43 levels (Fig. 2a; Fig. S2), indicating the previously described phenomenon of auto-regulation of endogenous TDP-43 upon TDP-43 over-expression [3, 22]. Notably, hTDP-43 was readily detected in the RIPA-insoluble (urea-soluble) fraction from 1 week off Dox with a concurrent decrease in mTDP-43 levels, suggesting that the decrease in soluble mTDP-43 levels was indeed due to auto-regulatory down-regulation of the endogenous mTDP-43 gene rather than recruitment of mTDP-43 to the insoluble protein fraction (Fig. 2b). Likewise, RIPA-insoluble phosphorylated TDP-43 (pTDP-43; ∼45 kDa) was also detected in rNLS8 mouse cortex from 2 weeks off Dox, which accumulated at high levels from 4 weeks off Dox, with both phosphorylated S403/404-TDP-43 (p403/404) and S409/410-TDP-43 (p409/410) specific antibodies. These findings indicate that accumulation of insoluble TDP-43 occurs prior to the phosphorylation of the protein. In addition, the detection of ∼45 kDa pTDP-43 from 4 weeks accompanied by decreased levels of ∼43 kDa TDP-43 after 2 weeks indicates a shift in apparent molecular mass and accumulation of the TDP-43 phosphorylated species, as described previously [53]. The RIPA-insoluble ∼25 kDa C-terminal fragments typically found in FTLD-TDP patients were not detected in any rNLS mice with any of the anti-TDP-43 antibodies used here (Fig. 2b), suggesting a biochemical signature more akin to ALS, in which TDP-43 C-terminal fragments are less abundant. Furthermore, we also detected p403/404 and p409/410 TDP-43 in the RIPA-insoluble fraction of spinal cord of rNLS8, rNLS9B and rNLS43 mice (Fig. 2c; Figs. S3, S4). A decrease in both RIPA-soluble and -insoluble h, m and h+mTDP-43 with continued robust pTDP-43 levels at later time points off Dox in rNLS8 mice suggests potential neuron loss with disease progression. In addition to the biochemical signature of pathological TDP-43 in rNLS mice, pTDP-43-positive inclusions were detected by IHC with both p403/404 and p409/410 antibodies (Fig. 2d–i). Numerous speckled and large pTDP-43 cytoplasmic inclusions were detected in many brain regions, including motor cortex (Fig. 2d, e), spinal cord (Fig. 2f, g; Fig. S5a), striatum (Fig. 2h, i), visual, entorhinal, and somatosensory cortex, cerebellum and hippocampus (Fig. S5a) in all rNLS lines, resembling those detected in motor cortex and spinal cord of ALS/FTLD-TDP patients (Fig. 2j, k) [30]. Rare cytoplasmic pTDP-43-positive inclusions were detected as early as 1 week off Dox in layer V of the motor cortex (Fig. S5b), and the percentage of NeuN-positive cells with hTDP-43-positive puncta increased over time to approximately 28 % of neurons at disease end stage (8–18 weeks off Dox) in rNLS8 mice (Fig. S5c), but TDP-43-positive inclusions were detected in fewer than 2 % of spinal cord motor neurons. As in ALS and FTLD-TDP [31], there was also an accumulation of cytoplasmic ubiquitin with pTDP-43-positive inclusions in the brains and spinal cords of rNLS mice (Fig. S6). To further investigate whether or not the TDP-43 pathology resulted from high levels of cytoplasmic hTDP-43ΔNLS or was due simply to over-expression of hTDP-43 driven by the NEFH promoter, we crossed NEFH-tTA line 8 mice with previously reported tetO-hTDP-43-WT line 4 or line 12 mice [22]. As expected, nuclear hTDP-43-WT was detected in neurons of the brain and spinal cord of bigenic mice from both crosses (n = 7 NEFH-tTA line 8/tetO-hTDP-43-WT line 4 mice analyzed between 1 week and 6 months off Dox, and n = 6 NEFH-tTA line 8/tetO-hTDP-43-WT line 12 mice analyzed between 2 weeks and 6 months off Dox, Fig. S7). hTDP-43- or pTDP-43-positive cytoplasmic inclusions were not detected in any of the NEFH-tTA/tetO-hTDP-43-WT mice. Extremely rare small intranuclear pTDP-43 inclusions were occasionally detected, in several brain regions including the motor cortex and striatum but not the spinal cord (Fig. S7). These findings indicate that pathology in the rNLS mice is due to the combination of high levels of over-expression driven by the NEFH promoter and the presence of the cytoplasmic targeting of the hTDP-43ΔNLS. Thus, similar to human FTLD-TDP and ALS cases, all three rNLS lines accumulated abundant TDP-43 cytoplasmic inclusions in cortex and spinal cord that correlated with the presence of RIPA-insoluble TDP-43. Our longitudinal analyses indicate that endogenous mTDP-43 is rapidly down-regulated, and the accumulation of insoluble TDP-43 occurs prior to robust TDP-43 phosphorylation in both brain and spinal cord in rNLS mice. rNLS mice develop progressive cortical atrophy and neuromuscular junction denervation followed by spinal cord motor neuron loss with dramatic muscle atrophy To sequentially characterize the ALS-like phenotype in the rNLS8 mice, we measured changes in cortical thickness, brain mass, tibialis anterior (TA) neuromuscular junction (NMJ) innervation, motor neuron (MN) number and muscle integrity. rNLS8 mice displayed decreased cortical thickness indicative of neuronal degeneration beginning at 4 weeks off Dox (Fig. 3a–c), which was accompanied by astrogliosis in many regions, including layer V of the motor cortex (Fig. S8). By disease end stage at 10–18 weeks off Dox, rNLS8 mice had significantly smaller brains when compared to littermate controls, indicating dramatic brain atrophy (Fig. 3d). To investigate the effect of hTDP-43ΔNLS expression on innervation of NMJs, we analyzed hindlimb TA muscle using the cholinergic synaptic marker vesicular acetylcholine transporter (VAChT) to label motor terminals, and fluorophore-conjugated alpha-bungarotoxin (BTX) to mark motor endplates. Significant denervation occurred after only 4 weeks of transgene expression in rNLS8 mice, and by 6 weeks of transgene expression the TA showed massive denervation with only 39.6 ± 6.6 % (mean ± SEM, n = 4) of NMJs still intact (Fig. 3e–h). Interestingly, axonal dieback occurred before MN loss, since the number of spinal MNs at the lumbar L4-L5 level (innervating hindlimb muscles) were unchanged at this 4 week time point but were decreased by ∼28 % after 6 weeks off Dox, and by 8 weeks off Dox only ∼50 % of lumbar MNs remained (Fig. 3i–l). Finally, following loss of MNs at disease end stage, there was grouped fiber atrophy and centralized nuclei indicative of attempted regeneration in muscle, as well as dramatic decreases in the mass of isolated TA and gastrocnemius muscles (Fig. 3m–p). rNLS mice develop a progressive motor phenotype leading to death To determine if hTDP-43ΔNLS expression also initiated motor dysfunction in rNLS mice, we monitored several behavioral phenotypes over time. rNLS8 mice developed early-onset hindlimb clasping and fine forelimb and/or hindlimb tremor (Fig. 4a, b; Video S1), progressive loss of grip strength as measured by wirehang test (Fig. 4c), and progressive decline in coordinated movement and balance as measured by impairment on the accelerating rotarod test (Fig. 4d). In addition, rNLS8 mice showed progressively decreasing body mass from a peak at 2 weeks off Dox (Fig. 4e), and early death with median survival of 10.3 weeks off Dox (Fig. 4f). Similar findings of brain and muscle atrophy, and progressive motor phenotype, were obtained in rNLS9B and rNLS43 mice, albeit with differences in the time of onset and length of survival (Figs. S9, S10). Since males and females of all lines developed consistent pathology and disease, they were both included in analyses. Notably, rNLS8 mice had a more protracted disease course than the rapid progression observed in rNLS9B and rNLS43 mice, consistent with their lower spinal cord hTDP-43 expression levels (Fig. 1b, c; Fig. S1a). Thus, the pathology and disease course in rNLS8 mice progressed with several ALS-relevant features, including early cytoplasmic accumulation of hTDP-43 along with down-regulation of endogenous mTDP-43 from 1 week off Dox, the formation of cytoplasmic pTDP-43 inclusions accompanied by onset of motor symptoms from 2 weeks off Dox, brain atrophy, astrogliosis and muscle NMJ denervation at 4 weeks off Dox, which were followed by lower spinal cord MN loss and an inability to perform several motor tasks from 6 weeks off Dox until disease end stage at a median of 10.3 weeks from initial Dox removal (Fig. 5). This sequence of events is likely the same for lines rNLS9B and rNLS43, although with compressed timing. Therefore, these rNLS mice faithfully recapitulate many of the pathological and phenotypic features of ALS. TDP-43 pathology is rapidly cleared and nuclear mTDP-43 returns following suppression of hTDP-43ΔANLS expression in rNLS mice Since most patients with ALS and/or FTLD are not diagnosed until well into the disease course when neuron loss has already occurred, we next asked whether suppression of hTDP-43ΔNLS expression would eliminate existing TDP-43 pathology and allow functional recovery even after significant motor deficits had developed. We hypothesized that this paradigm would allow us to determine whether or not therapies which drive clearance of cytoplasmic TDP-43 and allow return of nuclear TDP-43 could be clinically relevant, and would establish the rNLS mice as a model for future preclinical testing of ALS and FTLD-TDP therapeutics. After 6 weeks of transgene expression, when significant TDP-43 pathology was present with overt neuron loss and dramatic motor decline (Figs. 2d-i, 3, 4), we suppressed hTDP43ΔNLS expression by re-introducing Dox (Fig. 6a). After only 2 weeks of Dox re-introduction, a dramatic decrease in accumulated hTDP-43ΔNLS and pTDP-43 protein was observed in the cortex and spinal cord (Fig. 6b–d). Additionally, pTDP-43 pathology was no longer detected in the motor cortex or in any other brain region, including the thalamus and genu of the corpus callosum, of rNLS8 mice after 12+ weeks back on Dox (Fig. 6e; Fig. S11a–c). As has been seen with previous TDP-43 transgenic mice [22], over-expression of hTDP-43ΔNLS caused a loss of normal nuclear mTDP-43 protein (Fig. 7a; Fig. S11d). However, when mice were re-introduced to Dox at 6 weeks, nuclear mTDP-43 returned (Fig. 7a; Fig. S11d) and levels of mTDP-43 detected biochemically were significantly increased as early as 2 weeks back on Dox (Fig. 7b, c). Furthermore, re-introduction of Dox led to clearance of cytoplasmic TDP-43 (Fig. S11a, d) and diminished astro-gliosis (Fig. S11e). Thus, suppression of hTDP-43ΔNLS transgene expression resulted in elimination of cytoplasmic TDP-43, clearance of pTDP-43 pathology and return of endogenous mTDP-43 to the nucleus. Similar to our previous findings [22], we also detected a decrease in neurofilament light (NFL) chain but not neurofilament medium (NFM) chain protein in the cortex of rNLS8 mice at 6 weeks off Dox (Fig. 7d–f). Although NFL levels showed some increase upon re-introduction of Dox, these remained significantly decreased compared to controls after 18–32 weeks back on Dox (Fig. 7d–f). Furthermore, double labeling IF of pTDP-43 with phospho-NFM and neurofilament heavy chain proteins showed little or no co-localization of NF proteins with TDP-43 pathology in the cortex and spinal cord of rNLS8 mice, suggesting that mislocalization of NF proteins is not involved in driving neuronal death (Fig. S12). rNLS mice functionally recover following hTDP-43ΔNLS transgene suppression We next asked whether clearing hTDP-43ΔNLS would lead to symptom alleviation and functional recovery in mice with overt ALS-like motor impairments. Remarkably, mice that had expressed hTDP-43ΔNLS for 6 weeks, with associated pTDP-43 pathology, neuron loss and motor dysfunction, showed a rapid functional improvement as early as 2 weeks back on Dox, with partial recovery of hindlimb clasping phenotype and rapid weight gain (Fig. 8a, b; Video S1). Importantly, rNLS8 mice at 8 weeks off Dox had significantly fewer lumbar spinal cord MNs compared to those at 6 weeks off Dox (Fig. 3l), however there was no statistically significant difference in MN numbers between mice at 6 weeks off Dox and mice at 12–18 weeks back on Dox (Fig. 8c), suggesting prevention of continued MN loss in mice back on Dox. Likewise, there was no significant difference in motor cortex cortical thickness in rNLS8 mice at 6 weeks off Dox compared to rNLS8 mice back on Dox for 12–32 weeks (Fig. S11f, g). Indeed, despite the significant muscle denervation and even with the ∼25 % reduction in the total number of MNs at the TA-innervation L4 level of the spinal cord (Fig. 3h), we detected a significantly higher percentage of innervated NMJs in mice back on Dox, suggesting that the remaining MNs were able to re-innervate the vacated TA motor endplates following suppression of hTDP-43ΔNLS expression (Fig. 8d, e). rNLS8 mice also displayed a remarkable return of grip strength and improvement in performance in the rotarod test beginning as early as 1 week back on Dox (Fig. 8f, g). Finally, although rNLS8 mice had a median survival of 10.3 weeks off Dox and a maximum lifespan of 18 weeks, rNLS8 mice returned to Dox at 6 weeks showed a dramatic extension of lifespan (Fig. 8h). Of n = 14 mice returned to Dox at 6 weeks, three were killed for early analysis, whereas 11 mice lived up to or past the maximum rNLS8 lifespan without sign of disease progression (Fig. 8h). In addition to the phenotypic recovery in rNLS8 mice returned to Dox at 6 weeks, we also observed a similar functional recovery in rNLS9B mice returned to Dox at the time point of disease when these mice had lost 25 % of peak body weight (5.0–6.0 weeks off Dox). Similar to rNLS8 mice, rNLS9B mice back on Dox showed no change in MN number and a significantly higher percentage of innervated NMJs (Fig. S13a,b). In addition, although 3 mice progressed rapidly to disease end stage, 12 of 15 rNLS9B mice back on Dox showed a partial recovery of hindlimb clasping phenotype, improvement in the wirehang test, progressive weight gain, and extended survival (Fig. S13c–f). Notably, although 3 of the 12 recovered rNLS9B mice were killed for analysis after 4 months back on Dox, the 9 remaining recovered rNLS9B mice remained alive without progression of disease at between 5 and 8 months back on Dox, far past when all 17 of the rNLS9B littermate mice which were not returned to Dox had succumbed to disease (Fig. S13f). Discussion Our newly generated rNLS mice faithfully recapitulate many of the pathological features of ALS and FTLD-TDP. The previous development of numerous TDP-43 in vivo models has largely failed to recapitulate the combined key features of TDP-43 pathology including neuron loss, muscle denervation/atrophy and a progressive motor phenotype leading to premature death. The major reasons for our success include the use of the NEFH promoter to direct hTDP-43ΔNLS expression not only to brain but also to spinal cord, accounting for the spinal motor neuron loss and muscle denervation leading to a more robust ALS-like phenotype. Our results also indicate that the NEFH promoter drives higher levels of expression than the Camk2a promoter (>10 fold versus 8 fold, respectively, in the cortex), which in turn increased hTDP-43 and resulted in the accumulation of abundant pathological pTDP-43 detected by both biochemical and histological methods [22]. Although high levels of over-expression of TDP-43 artificially targeted to the cytoplasm produces a model which diverges from the circumstances in human disease, given that the initial upstream initiation of disease has been hypothesized to require multiple pathogenic ‘hits’ [26, 34], this strategy has nevertheless allowed us to overcome the need for such multiple triggers and resulted in the most disease-relevant TDP-43 model produced to date. Moreover, our strategy of targeting hTDP-43 to the cytoplasm, the location of accumulation in human disease, instead of to the nucleus, allowed for higher hTDP-43ΔNLS accumulation before the onset of neurotoxicity and brought about pathology more reminiscent of ALS/FTLD-TDP than was seen with expression of nuclear hTDP-43-WT. It should be noted that although TDP-43 pathology accumulates in both brain and spinal cord in the rNLS mice, testing of cognitive function relevant to FTLD-TDP phenotypes is confounded by the early and robust progressive motor impairment. However, the common pathology in ALS and FTLD-TDP indicate that these diseases are syndromic variants of neurodegenerative TDP-43 proteinopathies, and suggest that our results could be generalized to both diseases. Indeed, our data provide insights into the role of neuron loss and TDP-43 phosphorylation in disease pathogenesis. Specifically, the motor phenotypes in the rNLS8 mice, which begin with hindlimb clasping, are detected at 1–2.5 weeks off Dox (Fig. 4a), prior to the detection of overt brain atrophy, NMJ denervation and neuron loss, thereby suggesting that the initial dysfunction seen in these animals is not due to simple neuron death but instead by an earlier disruption to neuronal function. Despite high levels of hTDP-43ΔNLS expression from 1 week off Dox, pTDP-43 is biochemically detected at robust levels only after 4 weeks off Dox in rNLS8 mice, at the point when overt cell loss and muscle denervation begins, suggesting that phosphorylation of TDP-43 may be a signal in an attempt to clear the accumulated insoluble protein. Notably, although both clearance of TDP-43 pathology from the rNLS8 mice and recovery of motor function occurred rapidly following suppression of hTDP-43ΔNLS expression, the halting of MN loss cannot be attributed to the removal of inclusions alone, since although they were readily detected, large pTDP-43 inclusions were observed in only a minority of MNs. Despite this, up to 50 % of MNs were lost by 8 weeks off Dox (Fig. 3l), indicating that although expressing cytoplasmic hTDP-43, most MNs that degenerated did not contain pTDP-43 inclusions. These findings suggest that large pTDP-43 inclusions, despite being hallmarks of disease, are not necessarily the cause of neuronal demise. Rather, our findings indicate a complex cascade of events leading to disease development and progression to death. It will be interesting to delineate the identity of the surviving neurons that allow for muscle re-innervation and recovery. Furthermore, it will be important to identify the molecular determinants of those neurons that succumb early in disease versus those that survive and are able to functionally recover following hTDP-43ΔTDP-43 suppression. Another key question that remains to be addressed is the contribution of loss of nuclear TDP-43 versus the accumulation of cytoplasmic TDP-43 pathology, which are both seen in rNLS mice, to the development of disease. Complete knockout of TDP-43 in mice is embryonic lethal [24, 40, 50], and ubiquitous inducible post-natal Cre recombinase-mediated deletion of TDP-43 in mice causes alterations in fat metabolism and death within 9 days [11]. In addition, the targeted depletion of TDP-43 from spinal cord motor neurons using an HB9 promoter-driven Cre-Lox system caused progressive motor phenotypes with motor neuron loss [51]. These studies highlight the importance of maintaining TDP-43 levels for neuronal viability and animal survival, and suggest that loss of normal nuclear TDP-43 function is detrimental to motor neurons and can cause ALS/FTLD phenotypes. However, loss of function is unlikely to be the sole cause of neurodegeneration because over-expression of wild-type TDP-43 resulted in elevated levels of nuclear TDP-43 but also caused disease phenotypes and neuron death in vivo [22, 49]. Thus, no single mechanism may be responsible for the full gamut of pathology and dysfunction in ALS/FTLD-TDP. To address this more specifically, future studies should distinguish the contributions to pathogenesis of a toxic gain of function or a loss of normal function by employing transgenic animals with genetic ablation of auto-regulatory sequences in the TDP-43 3′UTR to investigate the effect of cytoplasmic accumulation of TDP-43 separate from nuclear depletion. Interestingly, in all three lines of NEFH-tTA mice analyzed, a small minority of astrocytes was found to express the hTDP-43ΔNLS transgene (Fig. 1; Fig. S1). The reason for this unexpected non-neuronal expression from the neuronal-specific NEFH promoter remains unclear, although this result is unlikely to be caused by the random chromosomal site of transgene insertion due to each line being derived from individual founder mice. However, since this is the first report of the use of the human NEFH promoter in transgenic mice, we speculate that there may be differences in the transcriptional recognition of the promoter in mouse compared to human. Regardless, the NEFH-tTA mice showed high levels of Dox-suppressible expression in neurons of the brain and spinal cord, and therefore offer a new model system to allow further investigation not only of TDP-43 but also other proteins involved in neurodegenerative diseases. Recently, the role of cell-to-cell transmission and propagation of misfolded proteins through the central nervous system (CNS) has suggested a novel mechanism for neurodegenerative disease progression [15, 16, 19, 21, 38]. For example, misfolded tau and α-synuclein induce disease spread in vivo and cause pathology reminiscent of Alzheimer's and Parkinson's diseases, respectively [15, 16, 19, 21, 38]. Isolated pathological TDP-43 from ALS and FTLD-TDP patient tissue was also shown to seed aggregation of endogenous TDP-43 following direct transduction of cultured cells [32]. Furthermore, post mortem studies of TDP-43 pathology in ALS and FTLD-TDP patients demonstrate a progressive sequential spread of TDP-43 pathology in brain and spinal cord, suggesting focal initiation of disease with subsequent involvement of interconnected brain and spinal cord regions [7–9]. Our findings here in the rNLS mice indicate that intracellular TDP-43 pathology can be rapidly cleared in vivo, which halts further neurodegeneration, thus suggesting that the presence of intracellular pathological TDP-43 is insufficient for cell-to-cell spread and propagation of disease pathology throughout the neuraxis in the absence of continued cytoplasmic TDP-43 accumulation. It is possible, however, that in the extended disease course of human patients minute quantities of transmitted pathological TDP-43 may lead to amplification of disease over time. Further studies will be required in order to delineate the contribution of spreading of pathology in ALS and FTLD-TDP. No effective therapy currently exists for either ALS or FTLD-TDP, and the most exciting aspect of this study is the demonstration of functional recovery following suppression of cytoplasmic TDP-43 expression even after the time point when TDP-43 pathology has formed, nuclear TDP-43 has decreased and neurodegeneration has begun. Although these findings in the rNLS mice cannot be directly applied to human ALS/FTLD-TDP therapy due to the reliance on an artificial genetic system in which transgene transcription can be rapidly prevented, they nevertheless show that functional recovery from TDP-43-related disease progression is possible after symptom onset in vivo. TDP-43 is an RNA/DNA binding protein that regulates its own expression via a direct auto-regulatory mechanism involving binding of TDP-43 to the 3′UTR of its own transcript [3, 22, 35]. For this reason, over-expression of exogenous TDP-43 (lacking the 3′UTR required for auto-regulation) causes down-regulation of endogenous TDP-43, and in the case of over-expression of cytoplasmically targeted hTDP-43 in rNLS mice, this leads to a loss of nuclear mTDP-43 (Fig. 2a; Fig. S11a, d) [22]. Therefore, in rNLS mice returned to Dox, suppression of cytoplasmic hTDP-43 expression would release the auto-regulatory inhibition of mTDP-43 transcription, leading to the return of nuclear mTDP-43 by renormalization of homeostatic TDP-43 expression. We cannot distinguish whether the clearance of cytoplasmic TDP-43 or the return of nuclear TDP-43 is responsible for the dramatic recovery seen, and indeed the combination of both phenomena may be required for this effect. However, accumulation of cytoplasmic TDP-43 occurs only in disease and directly leads to a decrease in nuclear TDP-43 levels. We therefore hypothesize that clearance of cytoplasmic TDP-43, which allows for the return of normal nuclear TDP-43, could be a therapeutic strategy for ALS and FTLD-TDP, even when disease is very advanced. Aggregated TDP-43 is cleared primarily by autophagy in cell culture [39], and thus autophagy is a likely mechanism by which the rapid clearance of hTDP-43ΔNLS occurs following transgene suppression. Indeed, small molecule enhancers of autophagy have shown promise in previous TDP-43 cell and mouse models [6, 46], suggesting that modulation of degradation may be one avenue for removal of insoluble and pathological TDP-43. Previous attempts to delineate the effects of reversibility of phenotype due to hTDP-43 expression in vivo have produced conflicting results, likely due at least in part to the shortcomings of the previously available models. For example, although clearance of cytoplasmic hTDP-43ΔNLS from the brain of Camk2a-tTA/tetO-hTDP-43ΔNLS mice, which develop some forebrain neuron loss but lack a degenerative motor phenotype, allows for behavioral recovery when mice are returned to Dox at a young age, older mice showed no recovery at 2 weeks post Dox re-introduction [1]. Furthermore, NEFH-tTA/tetO-hTDP-43M337V rats showed a partial recovery of motor function and extension of lifespan upon transgene suppression, but due to the expression of primarily nuclear-localized mutant TDP-43M337V and the development of a very rapid disease course without notable TDP-43 accumulation or overt neuron loss in these animals, longitudinal analysis is difficult and their relevance to human disease is unclear [20]. Our findings now show marked muscle re-innervation in rNLS mice returned to Dox, likely as a result of axonal sprouting, even after significant neuron loss and muscle denervation had occurred (Fig. 8d, e). Thus, suppression of continued cytoplasmic hTDP-43 expression prevents further neuron loss, and surviving motor neurons can re-innervate muscle leading to functional recovery. We here report new rNLS mouse models which for the first time recapitulate cytoplasmic TDP-43 pathology reminiscent of ALS and FTLD-TDP in brain and spinal cord, accompanied by a progressive neurodegenerative ALS-like phenotype. Our findings indicate that in symptomatic rNLS mice, clearance of cytoplasmic TDP-43 and normalization of nuclear TDP-43 levels prevents continued neurodegeneration and allows for muscle re-innervation, leading to functional recovery and dramatic extension of lifespan. Our results therefore highlight the exquisite ability of the CNS to recover from disease-associated dysfunction even at advanced stages of disease in new models of ALS/FTLD-TDP. Supplementary Material supplemental We thank Drs. Todd Cohen, Edward B. Lee, Síl-via Porta and Kurt Brunden for helpful discussion and input on the manuscript, Chi Li and Clark Restrepo for technical assistance, Drs. Manuela Neumann and Elizabeth Kremmer for providing the phosphorylation specific TDP-43 rat monoclonal antibody TAR5P-1D3, Dr. Xu-Gang Xia, Thomas Jefferson University for the NEFH-tTA construct, Dr. Chris Henderson, Columbia University for the VAChT antibody, and Dr. Jean Richa of the University of Pennsylvania Transgenic and Chimeric Mouse Facility for transgenic mouse production. This work was supported by NIH/NIA AG032953 and AG17586, and by Australian National Health & Medical Research Council C.J. Martin Biomedical Early Career Fellowship 1036835 (to A.W.). Fig 1 Expression of hTDP-43ΔNLS in brain and spinal cord of rNLS8 and rNLS9B mice. a Schematic for Dox-regulatable expression of hTDP-43ΔNLS in bigenic mice under the control of the NEFH promoter. Expression of hTDP-43 and total (h+m) TDP-43 protein in olfactory bulb (Ob), cerebellum (Cb), hippocampus (Hp), brainstem and remainder of the brain (Bs), cortex (Cx), spinal cord (Sc), spleen (Sp), kidney (Kd) and liver (Lv) of b rNLS8 and c rNLS9B mice at 4 weeks off Dox. Approximate molecular weight markers in kDa are shown on the left and GAPDH is a loading control. Representative immunoblots of n = 3. Representative images at 1 week off Dox show widespread expression of hTDP-43ΔNLS in d Hp, Cx, and e Sc of rNLS8 mice. IHC in rNLS8 motor cortex showing cytoplasmic TDP-43 inclusions (arrows) detected with both f an N-terminal (Nt)-TDP-43 antibody, and g a C-terminal (Ct)-TDP-43 antibody, at 2 weeks off Dox. h, i There is widespread hTDP-43 expression (green) in NeuN+ neurons (red), shown at 4 weeks off Dox. Arrowhead in i indicates rare non-neuronal hTDP-43-positive glial cell, and nuclei are shown by DAPI (blue). Scale bars d 500 μm, e 200 μm, f, g 50 μm, h, i 50 μm Fig 2 Accumulation of insoluble pTDP-43 over time in brain and spinal cord, and formation of pTDP-43 inclusions in rNLS mice. a RIPA-soluble hTDP-43ΔNLS, mouse (m) TDP-43 and total (h+m) TDP-43 in rNLS8 cortex at various time points compared to non-transgenic (nTg) and monogenic (tTA only) controls at 1 week off Dox. Quantification is shown below each representative immunoblot with n = 3 mice for nTg, tTa and 1, 2, 4 and 6 weeks off Dox, and n = 9 for end stage mice (7–18 weeks); see also Fig. S2. Approximate molecular weight markers in kDa are shown on the left and GAPDH is shown as a loading control. b RIPA-insoluble hTDP-43ΔNLS, mTDP-43 and h+mTDP-43 from 1 week off Dox in rNLS8 cortex. Higher molecular weight and phosphorylated (p) TDP-43 are detected from 2 weeks off Dox. Extract from a human FTLD patient is shown as a positive control. Asterisks indicate pTDP-43, arrowhead indicates 43 kDa TDP-43, and hash mark indicates 20–25 kDa C-terminal TDP-43 fragments. c RIPA-insoluble pTDP-43 in rNLS8 spinal cord from 4 weeks off Dox. Time points of 7–18 weeks in a–c are disease end stage. pTDP-43 pathology was detected with antibodies to both pS409/410- and pS403/404-TDP-43 in rNLS8 at 4 weeks off Dox in d motor cortex and f spinal cord (Sc), rNLS8 at 6 weeks off Dox in h striatum, in rNLS9B at 4 weeks off Dox in e motor cortex, g Sc, and i striatum, and in the j motor cortex and k Sc of human ALS patients. Arrows indicate examples of large cytoplasmic inclusions, and asterisks in e and i indicate small pTDP-43-positive puncta. Scale bar d–k 50 μm Fig 3 rNLS8 mice develop progressive cortical atrophy, muscle denervation, MN loss and muscle atrophy. a, b Representative brain sections from nTg and rNLS8 mice 18 weeks off Dox at Bregma 1.10, stained with haemotoxylin and eosin (H&E). Thickness is measured from the edge of the brain section to the white matter below (indicated with a bar). c Measurement of cortical thickness in nTg and rNLS8 mice at different ages, n = 3 per time point. d Brain mass of rNLS8 mice compared to littermate nTg or tTA monogenic controls at disease end stage of 10–18 weeks off Dox, n = 9 per group. e–h Representative images and quantification of the overlap of VAChT-positive motor terminals (red) with acetylcholine receptors stained using BTX (green) as an indicator of innervated motor endplates in the tibialis anterior (TA) muscle of rNLS8 mice. The TA muscle showed marked denervation at 6 and 8 weeks off Dox (f–g, vacated NMJs noted with asterisks), n = 4 mice per time point. i–k IF for MN marker VAChT (red) with nuclear marker DAPI (blue) revealed loss of MNs in the lumbar spinal cord at 6 and 8 weeks off Dox in rNLS8 mice. I Quantification of number of VAChT-positive lumbar MNs in rNLS8 mice at different times off Dox. n = 4 mice per time point. m, n H&E staining of TA muscle showed gross muscle atrophy in rNLS8 mice at disease end stage compared to littermate nTg control. Examples of central nuclei are shown by an arrow and atrophic fibers by an asterisk. o, p TA and gastrocnemius muscle masses of rNLS8 mice compared to littermate nTg or tTA monogenic controls at disease end stage of 10–18 weeks off Dox, n = 7 per group. Scale bars a, b 500 μm, e–g 50 μm, i–k, m–n 100 μm; *p < 0.05, **p < 0.01, ***p < 0.001 versus control by one-way ANOVA with Bonferroni's post hoc test; ###p < 0.001 by paired two-tailed t test Fig 4 Expression of hTDP-43ΔNLS results in dramatic motor impairments, weight loss, and death in rNLS8 mice. Onset of a hindlimb clasping, b tremor, and progressive declines in c wirehang and d rotarod performance in rNLS8 mice, n = 10 controls, n = 14 bigenic. e rNLS8 mice also showed significant weight loss relative to controls beginning 3 weeks after Dox removal, n = 19 per group. f Kaplan–Meier survival curve for rNLS8 mice and littermate controls shows a marked decrease in survival time in rNLS8 mice, n = 19 per group Fig 5 Schematic showing the time course of major neurodegenerative events in the rNLS8 mouse model of ALS/FTLD-TDP. The earliest events are the accumulation of cytoplasmic hTDP-43 and down-regulation of mTDP-43. Next, mice begin to show motor impairments including clasping, tremor, and impairments in motor tasks, and in some cells hTDP-43 begins to aggregate and becomes phosphorylated at pathological sites. By 4 weeks after initiation of hTDP-43ΔNLS expression, there is a significant brain atrophy and marked axonal dieback from hindlimb muscle. Cell loss and axonal dieback, and their functional consequences, continue to progress until disease end stage Fig 6 Soluble and insoluble TDP-43 are rapidly cleared from brain and spinal cord of rNLS8 mice upon re-introduction of Dox. a Schematic of experimental design. b, c Levels of hTDP-43 and h+mTDP-43 in RIPA-soluble (R) and RIPA-insoluble/urea-soluble (U) protein fractions of rNLS8 mouse cortex and spinal cord at 6 weeks and at subsequent time points back on Dox (+Dox), with tetO-hTDP-43-ΔNLS monogenic (Mono) littermate control. p403/404-TDP-43 and p409/410-TDP-43 were rapidly eliminated from the U fraction. Approximate molecular weight markers in kDa are shown on the left and GAPDH is the loading control. d hTDP-43 was eliminated from spinal cord MNs in rNLS8 mice +Dox, and e p409/410-TDP-43 inclusions were detected in the motor cortex in rNLS8 mice at 6 weeks, but were not detected in rNLS8 mice +Dox (images of 6 weeks off Dox +18 weeks back on Dox are shown). Scale bars d 100 μm, e 50 μm Fig 7 Endogenous nuclear mTDP-43 rapidly returns in cortex of rNLS8 mice upon re-introduction of Dox. a Nuclear mTDP-43 levels in rNLS8 mice at 6 weeks off Dox and after Dox re-introduction (+18 weeks Dox shown), compared to tetO-hTDP-43-ΔNLS monogenic (Mono) littermate control. b, c Endogenous mTDP-43 levels are decreased in rNLS8 mice at 6 weeks off Dox, but are significantly increased by +2 weeks Dox; ***p < 0.001 versus control by one-way ANOVA with Bonferroni's post hoc test, n = 3 per group. d–f Levels of NFL protein, but not NFM, are significantly decreased in rNLS8 mice at 6 weeks off Dox. Representative immunoblots are shown; **p < 0.01 and ***p < 0.001 versus control by one-way ANOVA with Bonferroni's post hoc test, n = 3 per group, +Dox indicates +18–32 weeks back on Dox. Scale bar a 50 μm Fig 8 Suppression of hTDP-43ΔNLS expression rescues motor phenotypes and restores lifespan in rNLS8 mice. a Hindlimb splaying phenotype of nTg control, rNLS8 mouse after 6 weeks of hTDP-43ΔNLS expression and rNLS8 mouse back on Dox after 6 weeks of hTDP-43ΔNLS expression (+Dox, 2 weeks back on Dox shown). b Weight loss reached a nadir at 6 weeks in rNLS8 mice back on Dox (n = 11 rNLS8 +Dox and n = 8 littermate nTg/monogenic controls). c MN number in lumbar SC, and d, e analysis of NMJ innervation in rNLS8 mice after 6 weeks of hTDP-43ΔNLS expression and at 3–4 months back on Dox (n = 3–4 per group), *p < 0.05 versus 6 weeks off Dox, image shows rNLS8 mouse at 6 weeks off Dox +12 weeks back on Dox. f Wirehang performance of rNLS8 mice back on Dox at 6 weeks, p < 0.001 from 3 weeks back on Dox versus 6 weeks off Dox, and g rotarod performance of rNLS8 mice back on Dox at 6 weeks, p < 0.05 at 1 week back on Dox and p < 0.001 from 2 weeks back on Dox versus 6 weeks off Dox, by repeated measures ANOVA with Bonferroni's post hoc test, n = 8. h Survival in rNLS8 mice back on Dox at 6 weeks (rNLS8 + Dox—blue line, n = 14, littermate nTg/monogenic controls—black line, n = 11) and in rNLS8 off Dox (rNLS—dotted red line, n = 19, also included in Fig. 4f). Triangles/circles indicate censored animals either killed for analysis or remaining alive Electronic supplementary material The online version of this article (doi:10.1007/s00401-015-1460-x) contains supplementary material, which is available to authorized users. Compliance with ethical standards: Conflict of interest The authors declare no competing financial interests. Ethical approval All procedures performed in studies involving animals were in accordance with the ethical standards of the institution at which the studies were conducted. 1 Alferi JA Pino NS Igaz LM 2014 Reversible Behavioral Phenotypes in a Conditional Mouse Model of TDP-43 Proteinopathies J Neurosci 34 15244 15259 10.1523/jneurosci.1918-14.2014 25392493 2 Arnold ES Ling SC Huelga SC Lagier-Tourenne C Polymenidou M Ditsworth D 2013 ALS-linked TDP-43 mutations produce aberrant RNA splicing and adult-onset motor neuron disease without aggregation or loss of nuclear TDP-43 Proc Natl Acad Sci 110 E736 E745 10.1073/pnas.1222809110 23382207 3 Ayala YM De Conti L Avendano-Vazquez SE Dhir A Romano M D'Ambrogio A 2011 TDP-43 regulates its mRNA levels through a negative feedback loop EMBO J 30 277 288 10.1038/emboj.2010.310 21131904 4 Ayala YM Zago P D'Ambrogio A Xu YF Petrucelli L Buratti E 2008 Structural determinants of the cellular localization and shuttling of TDP-43 J Cell Sci 121 3778 3785 10.1242/jcs.038950 18957508 5 Balin BJ Clark EA Trojanowski JQ Lee VM 1991 Neurofilament reassembly in vitro: biochemical, morphological and immuno-electron microscopic studies employing monoclonal antibodies to defined epitopes Brain Res 556 181 195 1718561 6 Barmada SJ Serio A Arjun A Bilican B Daub A Ando DM 2014 Autophagy induction enhances TDP43 turnover and survival in neuronal ALS models Nat Chem Biol 10 677 685 10.1038/nchembio.1563 24974230 7 Braak H Brettschneider J Ludolph AC Lee VM Trojanowski JQ Del Tredici K 2013 Amyotrophic lateral sclerosis–a model of corticofugal axonal spread Nat Rev Neurol 9 708 714 10.1038/nrneurol.2013.221 24217521 8 Brettschneider J Del Tredici K Irwin DJ Grossman M Robinson JL Toledo JB 2014 Sequential distribution of pTDP-43 pathology in behavioral variant frontotemporal dementia (bvFTD) Acta Neuropathol 127 423 439 10.1007/s00401-013-1238-y 24407427 9 Brettschneider J Del Tredici K Toledo JB Robinson JL Irwin DJ Grossman M 2013 Stages of pTDP-43 pathology in amyotrophic lateral sclerosis Ann Neurol 74 20 38 10.1002/ana.23937 23686809 10 Chen-Plotkin AS Lee VM Trojanowski JQ 2010 TAR DNA-binding protein 43 in neurodegenerative disease Nat Rev Neurol 6 211 220 10.1038/nrneurol.2010.18 20234357 11 Chiang PM Ling J Jeong YH Price DL Aja SM Wong PC 2010 Deletion of TDP-43 down-regulates Tbc1d1, a gene linked to obesity, and alters body fat metabolism Proc Natl Acad Sci 107 16320 16324 10.1073/pnas.1002176107 20660762 12 Geser F Brandmeir NJ Kwong LK Martinez-Lage M Elman L McCluskey L 2008 Evidence of multisystem disorder in whole-brain map of pathological TDP-43 in amyotrophic lateral sclerosis Arch Neurol 65 636 641 10.1001/archneur.65.5.636 18474740 13 Geser F Martinez-Lage M Robinson J Uryu K Neumann M Brandmeir NJ 2009 Clinical and pathological continuum of multisystem TDP-43 proteinopathies Arch Neurol 66 180 189 10.1001/archneurol.2008.558 19204154 14 Gordon PH Meininger V 2011 How can we improve clinical trials in amyotrophic lateral sclerosis? Nat Rev Neurol 7 650 654 10.1038/nrneurol.2011.147 21947135 15 Guo JL Covell DJ Daniels JP Iba M Stieber A Zhang B 2013 Distinct alpha-synuclein strains differentially promote tau inclusions in neurons Cell 154 103 117 10.1016/j.cell.2013.05.057 23827677 16 Guo JL Lee VM 2014 Cell-to-cell transmission of pathogenic proteins in neurodegenerative diseases Nat Med 20 130 138 10.1038/nm.3457 24504409 17 Gurney ME Pu H Chiu AY Dal Canto MC Polchow CY Alexander DD 1994 Motor neuron degeneration in mice that express a human Cu, Zn superoxide dismutase mutation Science 264 1772 1775 8209258 18 Hirasawa M Cho A Sreenath T Sauer B Julien JP Kulkarni AB 2001 Neuron-specific expression of Cre recombinase during the late phase of brain development Neurosci Res 40 125 132 11377750 19 Holmes BB Furman JL Mahan TE Yamasaki TR Mirbaha H Eades WC 2014 Proteopathic tau seeding predicts tauopathy in vivo Proc Natl Acad Sci 111 E4376 E4385 10.1073/pnas.1411649111 25261551 20 Huang C Tong J Bi F Zhou H Xia XG 2012 Mutant TDP-43 in motor neurons promotes the onset and progression of ALS in rats J Clin Invest 122 107 118 10.1172/jci59130 22156203 21 Iba M Guo JL McBride JD Zhang B Trojanowski JQ Lee VM 2013 Synthetic tau fibrils mediate transmission of neurofibrillary tangles in a transgenic mouse model of Alzheimer's-like tauopathy J Neurosci 33 1024 1037 10.1523/jneurosci.2642-12.2013 23325240 22 Igaz LM Kwong LK Lee EB Chen-Plotkin A Swanson E Unger T 2011 Dysregulation of the ALS-associated gene TDP-43 leads to neuronal death and degeneration in mice J Clin Invest 121 726 738 10.1172/jci44867 21206091 23 Igaz LM Kwong LK Xu Y Truax AC Uryu K Neumann M 2008 Enrichment of C-terminal fragments in TAR DNA-binding protein-43 cytoplasmic inclusions in brain but not in spinal cord of frontotemporal lobar degeneration and amyotrophic lateral sclerosis Am J Pathol 173 182 194 10.2353/ajpath.2008.080003 18535185 24 Kraemer BC Schuck T Wheeler JM Robinson LC Trojanowski JQ Lee VM 2010 Loss of murine TDP-43 disrupts motor function and plays an essential role in embryogenesis Acta Neuropathol 119 409 419 10.1007/s00401-010-0659-0 20198480 25 Kwong LK Irwin DJ Walker AK Xu Y Riddle DM Trojanowski JQ 2014 Novel monoclonal antibodies to normal and pathologically altered human TDP-43 proteins Acta Neuropathol Commun 2 33 10.1186/2051-5960-2-33 24690345 26 Lee EB Lee VM Trojanowski JQ 2012 Gains or losses: molecular mechanisms of TDP43-mediated neurodegeneration Nat Rev Neurosci 13 38 50 10.1038/nrn3121 27 Lee VM Carden MJ Schlaepfer WW Trojanowski JQ 1987 Monoclonal antibodies distinguish several differentially phosphorylated states of the two largest rat neurofilament subunits (NF-H and NF-M) and demonstrate their existence in the normal nervous system of adult rats J Neurosci 7 3474 3488 3119789 28 Lee VM Page CD Wu HL Schlaepfer WW 1984 Monoclonal antibodies to gel-excised glial filament protein and their reactivities with other intermediate filament proteins J Neurochem 42 25 32 6358415 29 Ling SC Polymenidou M Cleveland DW 2013 Converging mechanisms in ALS and FTD: disrupted RNA and protein homeostasis Neuron 79 416 438 10.1016/j.neuron.2013.07.033 23931993 30 Neumann M Kwong LK Lee EB Kremmer E Flatley A Xu Y 2009 Phosphorylation of S409/410 of TDP-43 is a consistent feature in all sporadic and familial forms of TDP-43 proteinopathies Acta Neuropathol 117 137 149 10.1007/s00401-008-0477-9 19125255 31 Neumann M Sampathu DM Kwong LK Truax AC Micsenyi MC Chou TT 2006 Ubiquitinated TDP-43 in frontotemporal lobar degeneration and amyotrophic lateral sclerosis Science 314 130 133 10.1126/science.1134108 17023659 32 Nonaka T Masuda-Suzukake M Arai T Hasegawa Y Akatsu H Obi T 2013 Prion-like properties of pathological TDP-43 aggregates from diseased brains Cell Rep 4 124 134 10.1016/j.celrep.2013.06.007 23831027 33 Paxinos G Franklin KBJ 2001 The mouse brain in stereotaxic coordinates 3rd Academic Press San Diego 34 Pesiridis GS Tripathy K Tanik S Trojanowski JQ Lee VM 2011 A “two-hit” hypothesis for inclusion formation by carboxyl-terminal fragments of TDP-43 protein linked to RNA depletion and impaired microtubule-dependent transport J BiolChem 286 18845 18855 10.1074/jbc.M111.231118 35 Polymenidou M Lagier-Tourenne C Hutt KR Huelga SC Moran J Liang TY 2011 Long pre-mRNA depletion and RNA missplicing contribute to neuronal vulnerability from loss of TDP-43 Nat Neurosci 14 459 468 10.1038/nn.2779 21358643 36 Robertson J Sanelli T Xiao S Yang W Horne P Hammond R 2007 Lack of TDP-43 abnormalities in mutant SOD1 transgenic mice shows disparity with ALS Neurosci Lett 420 128 132 10.1016/j.neulet.2007.03.066 17543992 37 Sampathu DM Neumann M Kwong LK Chou TT Micsenyi M Truax A 2006 Pathological heterogeneity of frontotemporal lobar degeneration with ubiquitin-positive inclusions delineated by ubiquitin immunohistochemistry and novel monoclonal antibodies Am J Pathol 169 1343 1352 10.2353/ajpath.2006.060438 17003490 38 Sanders DW Kaufman SK DeVos SL Sharma AM Mirbaha H Li A 2014 Distinct tau prion strains propagate in cells and mice and define different tauopathies Neuron 82 1271 1288 10.1016/j.neuron.2014.04.047 24857020 39 Scotter EL Vance C Nishimura AL Lee YB Chen HJ Urwin H 2014 Differential roles of the ubiquitin proteasome system and autophagy in the clearance of soluble and aggregated TDP-43 species J Cell Sci 127 1263 1278 10.1242/jcs.140087 24424030 40 Sephton CF Good SK Atkin S Dewey CM Mayer P 3rd Herz J 2010 TDP-43 is a developmentally regulated protein essential for early embryonic development J Biol Chem 285 6826 6834 10.1074/jbc.M109.061846 20040602 41 Shan X Vocadlo D Krieger C 2009 Mislocalization of TDP-43 in the G93A mutant SOD1 transgenic mouse model of ALS Neurosci Lett 458 70 74 10.1016/j.neulet.2009.04.031 19379791 42 Sobue G Hashizume Y Yasuda T Mukai E Kumagai T Mitsuma T 1990 Phosphorylated high molecular weight neurofilament protein in lower motor neurons in amyotrophic lateral sclerosis and other neurodegenerative diseases involving ventral horn cells Acta Neuropathol 79 402 408 2111074 43 Toledo JB Van Deerlin VM Lee EB Suh E Baek Y Robinson JL 2014 A platform for discovery: the University of Pennsylvania Integrated Neurodegenerative Disease Biobank Alzhei-mers Dement 10 477–484 e471 10.1016/j.jalz.2013.06.003 44 Tsai KJ Yang CH Fang YH Cho KH Chien WL Wang WT 2010 Elevated expression of TDP-43 in the forebrain of mice is sufficient to cause neurological and pathological phenotypes mimicking FTLD-U J Exp Med 207 1661 1673 10.1084/jem.20092164 20660618 45 Turner BJ Baumer D Parkinson NJ Scaber J Ansorge O Talbot K 2008 TDP-43 expression in mouse models of amyotrophic lateral sclerosis and spinal muscular atrophy BMC Neurosci 9 104 10.1186/1471-2202-9-104 18957104 46 Wang IF Guo BS Liu YC Wu CC Yang CH Tsai KJ 2012 Autophagy activators rescue and alleviate pathogenesis of a mouse model with proteinopathies of the TAR DNA-binding protein 43 Proc Natl Acad Sci 109 15024 15029 10.1073/pnas.1206362109 22932872 47 Wang IF Reddy NM Shen CK 2002 Higher order arrangement of the eukaryotic nuclear bodies Proc Natl Acad Sci 99 13583 13588 10.1073/pnas.212483099 12361981 48 Wegorzewska I Bell S Cairns NJ Miller TM Baloh RH 2009 TDP-43 mutant transgenic mice develop features of ALS and frontotemporal lobar degeneration Proc Natl Acad Sci 106 18809 18814 10.1073/pnas.0908767106 19833869 49 Wils H Kleinberger G Janssens J Pereson S Joris G Cuijt I 2010 TDP-43 transgenic mice develop spastic paralysis and neuronal inclusions characteristic of ALS and frontotemporal lobar degeneration Proc Natl Acad Sci 107 3858 3863 10.1073/pnas.0912417107 20133711 50 Wu LS Cheng WC Hou SC Yan YT Jiang ST Shen CK 2010 TDP-43, a neuro-pathosignature factor, is essential for early mouse embryogenesis Genesis 48 56 62 10.1002/dvg.20584 20014337 51 Wu LS Cheng WC Shen CK 2012 Targeted depletion of TDP-43 expression in the spinal cord motor neurons leads to the development of amyotrophic lateral sclerosis-like phenotypes in mice J Biol Chem 287 27335 27344 10.1074/jbc.M112.359000 22718760 52 Yoshiyama Y Higuchi M Zhang B Huang SM Iwata N Saido TC 2007 Synapse loss and microglial activation precede tangles in a P301S tauopathy mouse model Neuron 53 337 351 10.1016/j.neuron.2007.01.010 17270732 53 Zhang YJ Gendron TF Xu YF Ko LW Yen SH Petrucelli L 2010 Phosphorylation regulates proteasomal-mediated degradation and solubility of TAR DNA binding protein-43 C-terminal fragments Mol Neurodegener 5 33 10.1186/1750-1326-5-33 20804554

PMC005xxxxxx/PMC5127401.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 101676030 44856 Cell Chem Biol Cell Chem Biol Cell chemical biology 2451-9456 27818300 5127401 10.1016/j.chembiol.2016.10.006 NIHMS823588 Article Multicolor electron microscopy for simultaneous visualization of multiple molecular species Adams Stephen R. 16 Mackey Mason R. 2 Ramachandra Ranjan 2 Palida Lemieux Sakina F. 1 Steinbach Paul 3 Bushong Eric A. 2 Butko Margaret T. 17 Giepmans Ben N.G. 28 Ellisman Mark H. 24 Tsien Roger Y. 135 1 Department of Pharmacology, University of California, San Diego, La Jolla CA 92093-0647 USA 2 National Center for Microscopy and Imaging Research, University of California, San Diego, La Jolla CA 92093-0647 USA 3 Howard Hughes Medical Institute, University of California, San Diego, La Jolla CA 92093-0647 USA 4 Department of Neurosciences, University of California, San Diego, La Jolla CA 92093-0647 USA 5 Department of Chemistry & Biochemistry, University of California, San Diego, La Jolla CA 92093-0647 USA * Correspondence: sadams@ucsd.edu 6 Lead Contact 7 Current address, Biodesy Inc, 384 Oyster Point Blvd Suite #8, South San Francisco, CA 94080, USA and 8 Department of Cell Biology, University Medical Center Groningen, A. Deusinglaan 1, 9713 AV Groningen, The Netherlands. 21 10 2016 3 11 2016 17 11 2016 17 11 2017 23 11 14171427 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. Summary Electron microscopy (EM) remains the primary method for imaging cellular and tissue ultrastructure, although simultaneous localization of multiple specific molecules continues to be a challenge for EM. We present a method for obtaining multicolor EM views of multiple subcellular components. The method uses sequential, localized deposition of different lanthanides by photosensitizers, small molecule probes or peroxidases. Detailed view of biological structures is created by overlaying conventional electron micrographs with pseudocolor lanthanide elemental maps derived from distinctive electron energy-loss spectra (EELS) of each lanthanide deposit via energy-filtered transmission electron microscopy (EFTEM). This results in multicolor EM images analogous to multicolor fluorescence but with the benefit of the full spatial resolution of EM. We illustrate the power of this methodology by visualizing hippocampal astrocytes to show that processes from two astrocytes can share a single synapse. We also show that polyarginine-based cell-penetrating peptides enter the cell via endocytosis, and that newly synthesized PKMζ in cultured neurons preferentially localize to the post-synaptic membrane. eTOC blurb Multicolor electron microscopy is introduced by Adams et al. for imaging multiple, specific cellular components by locally depositing lanthanides and using electron energy-loss spectroscopy and energy-filtering EM. Analogous to multicolor fluorescence this method offers full spatial resolution of EM. Introduction Electron microscopy (EM) of biological samples remains the ultimate method for imaging cellular ultrastructure despite the recent advances in super-resolution microscopy (Betzig, et al., 2006; Hell, 2007; Huang, et al., 2009). Contrast in standard EM of epoxy-embedded samples is dependent upon deposition of heavy metals such as osmium, uranium or lead to highlight cellular components including protein, lipid or nucleic acid using long established and poorly selective stains. Selective visualization of specific proteins or macromolecules can be achieved using antibodies conjugated to gold particles or quantum dots of distinctive size, but poor penetrability of such labels in fixed cells or tissues limits the use of optimal fixation methods that preserve ultrastructure (Schnell, et al., 2012). This limitation can be avoided by the in situ oxidation of diaminobenzidine (DAB) generating a localized osmiophilic precipitate by photosensitizing dyes conjugated to antibodies or ligands (Deerinck, et al., 1994; Maranto, 1982), genetically targeted biarsenicals (Gaietta, et al., 2002) and genetically encoded chimeras of miniSOG (Shu, et al., 2011). Peroxidases (such as horse radish peroxidase, HRP) also generate similar precipitates from DAB on treatment with hydrogen peroxide (H2O2) and robust genetically-encoded versions have been developed recently (Kuipers, et al., 2015; Lam, et al., 2015; Martell, et al., 2012). The high penetrability of small dyes, DAB, and oxygen or H2O2 into optimally fixed cells or tissues enables specific labeling with preservation of cellular ultrastructure. The target protein becomes negatively-stained by the surrounding oxidized DAB precipitate, which may not be readily distinguishable from heavy staining from endogenous cellular structures such as membranes and the post-synaptic density. In this work, we demonstrate a method (Figure 1A) that can differentiate the DAB precipitate from the general staining of endogenous cellular material and permits identification and imaging of successively deposited DAB, each at a targeted or specified protein or cellular target. By precipitating DAB conjugated to lanthanide chelates rather than DAB itself, a specific metal such as Ce3+ is locally deposited. After washing out the unreacted DAB-chelator-Ln, a further round of deposition of DAB-chelator bound to another lanthanide ion such as Pr3+ is carried out by photooxidation of a second photosensitizer targeted to another cellular site or protein, at wavelengths that do not excite the first fluorophore. Alternatively, peroxidases can be used to generate the second precipitate. Following conventional postfixation staining with osmium tetroxide and electron microscopy of sections from embedded samples, the two precipitates containing different tightly-bound lanthanide ions can be spectrally separated using spatially-resolved electron energy-loss spectroscopy (EELS), which is implemented by energy-filtered transmission electron microscopy (EFTEM). Elemental distribution maps for the two metals obtained by EFTEM reveal their spatial distribution, and can be overlaid as pseudocolors on the conventional black-and white electron micrograph to give a multicolor image superimposed on cellular ultrastructure. The method is also useful for only a single deposited lanthanide, because the EELS signal is not obscured by staining of endogenous cellular structures by osmium or other heavy metals used for contrast in EM. Results Synthesis of DAB-metal chelate conjugates In designing metals complexed to DAB that would precipitate on oxidation, we considered the following requirements. The metal ions should have strong, distinct EELS peaks that are simultaneously quantifiable, and must form high affinity chelates to prevent any loss of metal ions during DAB oxidation and subsequent processing leading to a decreased EELS signal or a false positive signal. The lanthanide series have a similar charge (3+), ionic radii, and suitable EELS signals and should bind to a conjugate of DTPA with two DAB (Figure 1B) with three carboxylates to form a high-affinity complex (dissociation constants of 0.1–1 fM for Gd3+ have been reported for related DTPA-bisanilides) (Geze, et al., 1996) with overall neutral charge to facilitate precipitation. The DTPA-DAB2 was synthesized by the reaction of DTPA anhydride with a five-fold excess of DAB to hinder the formation of polymers (Figure S1). Following removal of most of the unreacted DAB by extraction, the product DTPA-DAB2 was precipitated. This solid was used for all subsequent photooxidation experiments with DTPA-DAB2 despite containing some unreacted DAB and monomer DTPA-DAB as measured and quantified by LC-MS (Figure S2). When free DAB was removed from the material by HPLC, the purified DTPA-DAB2 generated less precipitate in cuvette experiments and failed to generate the expected localized precipitate in cells (data not shown). Metal binding to lanthanides and precipitation of metal complex on oxidation The complexation of DTPA-DAB2 to lanthanide ions, Ln3+, was measured by titration using arsenazo III as a colorimetric indicator to the end point and by comparison to an equal concentration of DTPA. Each batch synthesized gave between 75–85% purity by weight from titration assuming the expected 1:1 stoichiometry and closely matched the percent purity of DTPA-DAB2 measured by HPLC (Figure S2). Precipitation of this so formed Ln DTPA-DAB2 (Ln-DAB2) following photooxidation by photosensitization of eosin at 480 nm was measured by monitoring the absorbance from increasing scattering at 600 nm (Deerinck, et al., 1994; Natera, et al., 2011). Typical time courses for La-, Ce-, Pr-, Nd-, and Sm-DAB2 were similar to that of DAB (Figure 1C) and were in contrast to greatly decreased precipitate when no lanthanide was present, confirming the requirement of charge neutralization for efficient precipitation (data not shown). All the Ln-DAB2 tested showed limited solubility in 100 mM sodium cacodylate pH 7.4, the buffer conventionally used for photooxidation in fixed cells and tissues and optimal for preserving ultrastructure in EM. To achieve a concentration close to the 2.5 mM value typically used for DAB, 2.5% DMF was added as a co-solvent and the cacodylate buffer concentration was decreased to 50 mM. The final metal ion concentrations following filtration were determined by inductively coupled plasma mass spectroscopy to be about 0.8 mM with about 2 mM total DAB content (DTPA-DAB2 and DAB) as measured by absorbance at 309 nm using an extinction coefficient of 14200 M−1cm−1. Two-color EM of tissue culture cells We next tested whether two Ln-DAB2 could be orthogonally precipitated in cells and whether the specific EELS signals of the two metals could be detected and separated as elemental images. Madine-Darby canine kidney (MDCK) cells stably expressing green fluorescent protein fused to epithelial cell adhesion molecule (GFP-EpCAM), were initially stained with NBD-ceramide, a Golgi-selective fluorescent probe capable of photosensitizing DAB (Pagano, et al., 1991; Pagano, et al., 1989; Takizawa, et al., 1993), and then subsequently an antibody to the cell surface marker, EpCAM (Schnell, et al., 2013), followed by a biotinylated secondary antibody. Following mild fixation, irradiation at 535 nm in the presence of La-DAB2 and oxygen gave faint darkening from formation of reaction product in cell regions corresponding to the fluorescence image of NBD-ceramide. The cells were treated with acetic anhydride to block any unreacted amines of the DAB moiety of the reaction product to prevent further reaction of the deposited precipitate. Ce-DAB2 was then precipitated after further labeling of EpCAM sites by HRP-streptavidin and incubation with hydrogen peroxide. Following osmification, resin embedding and sectioning, a low- magnification unfiltered electron micrograph (Figure 2A) of a typical cell reveals the expected intracellular and plasma membrane staining from deposited precipitates at the Golgi and cell surface. However, EELS of regions at the Golgi or plasma membrane (circled in Figure 2A) revealed characteristic peaks from predominately La or Ce (Figure 2B) respectively. The small contaminating Ce signal at the Golgi probably resulted from unwanted deposition of Ce-DAB2 at the site of previously photooxidized Ln-DAB2 despite acetylation of any residual free amines, and could be mathematically subtracted in the EELS spectra (Figure 2B′), and the elemental maps for La and Ce (Figure 2C, D, D′). To do this we selected a region in the field being observed that was expected to only contain La such as the Golgi and integrated the La and Ce peaks in this area’s EELS spectrum to give the fraction of contaminating Ce in the La channel that was then subtracted from the La elemental map. These core-loss elemental maps were generated by subtraction of four images (two pre-edge and two after-peak images, see experimental procedures for details) rather than two with the traditional three-window method (Egerton, 1996), to minimize any effects of signal bleed-through of the lanthanides resulting from inaccurate background extrapolation. To reduce sample warping and drift, we found taking multiple short image acquisitions and carbon coating the sample were greatly beneficial. This five-window method shows distinct signal from the appropriate cell regions with La at the Golgi and Ce at the plasma membrane. How should the two energy-filtered lanthanide maps and conventional TEM be visually combined? We first tried “mixing” in Photoshop the conventional TEM in grayscale (normal or inverted) with the lanthanide maps in red and green respectively. However, the grayscale image tended to drown out the color images because the black pixels stayed black regardless of any colors mixed in (Figure S3A–E). Next we tried displaying the conventional TEM in blue, so that a region with strong La-DAB2 or Ce-DAB2 would appear cyan or magenta respectively (Figure S3F, G). Unfortunately, regions with conventional TEM only tend to suffer because of the low psychophysical visibility of the blue channel. Assigning green to the conventional TEM and blue to the Ce-DAB2 de-emphasized the latter too much and merely shifted the problem. Finally, we realized that the conventional TEM image has high spatial frequencies, resolution, and S/N, analogous to the luminance channel in television, whereas the colors should be displayed as lower-resolution overlays, effectively modulating the alpha channel for transparency vs. opacity. Therefore, we used a custom algorithm to generate pseudocolored overlays of the La and Ce elemental maps on the monochrome unfiltered osmium image (conventional EM image) to yield a two-hue representation of marker distribution with the resolution of an electron micrograph (Figures 2E, S3H). An advantage of this algorithm is that it can be generalized to three or more pseudocolor channels. Hippocampal astrocyte cell tracing in brain slices Following this proof of principle, we tested this method’s application to biological questions that required the ultrastructural resolution of EM and labeling of two cell markers. Protoplasmic astrocytes in the mouse hippocampus establish distinct territories with limited overlap between peripheral processes. These fine peripheral processes intimately contact and modulate neuronal synapses (Haydon, 2001; Haydon and Carmignoto, 2006). Whether synapses located at domain boundaries are shared by two astrocytes is unknown as both synaptic profiles and the fine astrocytic processes near synapses are generally beyond the resolution limit of light microscopy (Bushong, et al., 2004; Bushong, et al., 2002; Halassa, et al., 2007). We injected two adjacent astrocytes in fixed hippocampal slices with either Lucifer yellow or a combination of Alexa-568 and neurobiotin (Figure 3A). Ce-DAB2 was photo-oxidized by Lucifer yellow at 470 nm. Acetyl imidazole was used to passivate the Ce-DAB2 precipitate instead of acetic anhydride because of its greater stability at neutral pH, higher solubility in water and self-buffering at pH 5 that favors reaction with the aromatic amines of DAB (Oakenfull and Jencks, 1971). Then neurobiotin was captured with HRP-streptavidin, which in turn was reacted with Pr-DAB2 and H2O2. After osmification and embedding in resin, sections were examined by EM for synapses with surrounding densely-stained astrocyte processes containing both Ce and Pr signals by EELS. An example of a perforated synapse with clearly defined synaptic cleft and pre- and post-synaptic components marked by synaptic vesicles and postsynaptic densities respectively is shown in Figure 3B. EELS of sub-regions of each of the two astrocytic processes contacting the synapse revealed predominately Ce or Pr signals (Figure 3C). Some signal from Pr is still present in the Ce astrocyte, perhaps from incomplete inactivation of the first Ce-DAB2 precipitate by limited penetration of the acetyl imidazole into the fixed brain slice. The individual elemental maps (Figure 3D, E, and corrected Pr map, 3E′) and their overlay with an unfiltered EM (Figure 3F) also indicate the processes from two astrocytes can share a single synapse. Endosomal uptake of cell-penetrating peptides We next explored whether it is feasible to precipitate Ce- and Pr-DAB by successive irradiation of two spectrally distinct photo-sensitizers rather than photooxidation followed by HRP-mediated oxidation. We first generated Ce-DAB2 precipitate by 480 nm irradiation of nuclear targeted miniSOG (Shu, et al., 2011) and then tested the ability of ReAsH-labeled tetracysteine-tagged connexin 43 that form gap junctional plaques (Gaietta, et al., 2002), to photo-oxidize DAB2 when excited at 560 nm. Pr-DAB2 was precipitated as expected at the plasma membrane but also in the nucleus suggesting the initial precipitate itself could act as a photosensitizer of DAB (data not shown). Ce-DAB2 photo-oxidized in a cuvette shows a broad absorbance centered at 500 nm that extends to 600 nm, and correspondingly in cells we found that after nuclear deposition of Ce-DAB2 by mini-SOG at 480 nm, irradiation at >630 nm would not precipitate Pr-DAB2. ReAsH does not absorb at 630 nm, so we tested the ability of far-red photosensitizers such as the phthalocyanine dye, IRDye 700DX (Mitsunaga, et al., 2011; Peng, et al., 2006) to photo-oxidize Pr-DAB2. Cell penetrating peptides (CPP), including oligomers of cationic amino acids like arginine (n=9–14), have been extensively used to deliver membrane-impermeant molecules or particles into the cytoplasm of cells (Copolovici, et al., 2014). CPPs rapidly bind to the plasma membrane and are hypothesized to enter cells via endocytosis (Brock, 2014; Kaplan, et al., 2005; Richard, et al., 2003). To determine whether poly-arginine CPPs enter cells via the endocytic pathway, we examined the localization of an internalized Arg10 peptide compared to Rab5a, a small GTPase that localizes to endosomal membranes, at a scale too small to resolve using conventional light microscopy. We incubated HeLa cells expressing Rab5a fused to miniSOG (Shu, et al., 2011) with an Arg10 peptide conjugated to IRDye 700DX photosensitizer (Figure S4) that can polymerize DAB. After the cells were incubated with Arg10-IRDye 700DX peptide for two hours, they were fixed and imaged by light microscopy (Figure 4). We detected bright intracellular puncta from the endocytosed Arg10-IRdye 700DX peptide (Figure 4C) that partially co-localized with miniSOG fluorescence (Figure 4B, D). We then irradiated the sample at 480 nm to photo-oxidize Ce-DAB2 catalyzed by miniSOG, removed unreacted Ce-DAB2 by washing, blocked amines with acetyl imidazole, and then illuminated at 680 nm to photo-oxidize Pr-DAB2 catalyzed by IRdye 700DX. Both endosomes and multivesicular bodies (MVB) were photo-oxidized and visible in the unfiltered EM and both contain precipitated Ce and Pr by EELS (Figure 4E and I respectively). The elemental maps and overlay with conventional EM image (Figure 4F–H) indicate that Ce is concentrated on the endosome periphery in accordance with the expected cytoplasmic localization of Rab5a, whereas Pr is predominately in the endosome lumen. In this example, we could not correct the Ce channel for contaminating Pr as EELS spectra with adequate signal-to noise could not be obtained solely for endosomal lumen or the periphery. The corresponding images of MVB (Fig. 4J–L) also show a similar distribution of Ce and Pr but with less cytosolic Ce, and several densely Ce-stained luminal vesicles formed by inward budding of the endosomal membrane. Endosomal localization of Rab5a has been shown to progressively decrease during early endosome maturation to MVB (Rink, et al., 2005) which is in agreement with our results. In addition, Arg10-IRdye 700DX colocalization with Rab5a in intracellular vesicles confirms that polyarginine-based CPPs enter the cell via endocytosis. Tracking newly-synthesized PKMζin cultured neurons Finally, we used EELS analysis of a single lanthanide-conjugated DAB to confirm DAB-based labeling that is not readily distinguishable from background with conventional TEM, particularly in regions that are normally electron dense such as the neuronal post-synaptic density. The kinase PKMζ has been implicated in long-term memory maintenance and is upregulated following neuronal activity (Shao, et al., 2012), but the function and precise sub-synaptic localization of these new PKMζcopies is unclear. We fused PKMζcDNA to a TimeSTAMP reporter, TS:YSOG3 (Palida, et al., 2015) that contains both YFP and miniSOG and allows newly synthesized proteins to be labeled in a drug-dependent manner using the small molecule BILN-2061. New copies can be visualized by correlated light and electron microscopy in a manner similar to previous TimeSTAMP reporters incorporating a split YFP and miniSOG (Butko, et al., 2012) (Figure S5). We induced chemical long term potentiation (LTP) by stimulating TS:YSOG3-PKMζ-transfected rat neurons in culture with forskolin and rolipram and then immediately added BILN-2061 for 24 hours to label newly-synthesized copies of PKMζ. These new copies were visible by YFP fluorescence and then were illuminated so that miniSOG would catalyze photooxidation of Ce-DAB2. After osmification, darkening was visible throughout the neuron at low magnification (Figure 5A) and labeling appeared at postsynaptic membranes in TEM (Figure 5B), yet it was unclear whether this signal was derived from DAB deposition or endogenous synaptic electron density. To confirm that the apparent signal was representative of newly produced PKMζprotein, we used EELS to visualize Ce at the same synapse. We found that the Ce signal was enhanced at the postsynaptic membrane (Figure 5C and D at lower magnification, E and F at higher magnification) confirming that new copies of PKMζpreferentially localize to the post-synaptic membrane, consistent with previous reports for PKMζlocalization (Hernández, et al., 2014). Unstimulated neurons that were treated with BILN-2061 and analyzed similarly (Figure 5G–I) showed only a minimal signal for Ce by EELS in synapses (Figure 5J) despite comparable post-synaptic density staining in conventional EM. Discussion In summary, we have developed a method that permits concurrent and selective visualization by EELS and EFTEM of two cellular components that can be labeled by photosensitizing fluorescent tracers or by peroxidases that can oxidize Ln-DAB2. Many fluorescent dyes are known to generate sufficient singlet oxygen via triplet sensitization of molecular oxygen so the method should be of wide scope. The use of genetically-encoded singlet oxygen generators, such as miniSOG when fused to proteins of interest, also permits the selective visualization by fluorescence and correlated EFTEM of their cellular localization in samples ranging from tissue culture cells to complex tissues such as the mammalian brain. Ln-DAB2 is efficiently precipitated by HRP for immunoperoxidase labeling but with some limitations. Unexpectedly, we found that HRP or its genetically-encoded versions and the more recently developed APEX (Lam, et al., 2015; Martell, et al., 2012), a mutated ascorbate peroxidase, act as photosensitizers and precipitate Ln-DAB2 (or DAB) when illuminated between 400–650 nm (data not shown). This property limits their use with multicolor EM unless they are introduced after the first round of Ln-DAB2 photooxidation as described above (Figures 2 and 3). Both peroxidases contain a heme prosthetic group that has not been reported to photosensitize molecular oxygen but if insufficient heme is available in the cell, protoporphyrin IX (Durner and Klessig, 1995) is known to bind both HRP and APEX (Jullian, et al., 1989; Lam, et al., 2015; Martell, et al., 2012). We speculate that trace bound amounts of this efficient photosensitizer (Fernandez, et al., 1997) are responsible for the photooxidation of Ln-DAB2 by these enzymes over the range of illuminating wavelengths that match that of protoporphyrin IX absorbance. APEX2 was engineered to improve heme binding but still deposited DAB upon illumination. Chemical inactivation (Durner and Klessig, 1995) of APEX2 to permit a second Ln-DAB2 to be photo-oxidized by another photosensitizer was ineffective. Further mutation of APEX or methods for increasing cellular heme availability during APEX expression will probably be required for their use in multicolor EM. Despite these limitations, multicolor EM has demonstrated the feasibility of selective deposition of metals by DAB oxidizers that are organelle stains, genetically-targetable or encodable, or immunoreactive. EFTEM although predominately used today for chemical analysis of materials, has been applied to biological samples, using the generally weak signals from endogenous elements (Aronova and Leapman, 2012), but is sufficiently sensitive to detect and distinguish many of the fourteen stable lanthanides when precipitated as Ln-DAB2. The sensitivity of the method is probably not limited by Ln-DAB oxidation as it photooxidizes at a comparable rate to DAB which can yield close to single-molecule detectability after extended illumination, staining with osmium and conventional TEM (unpublished results). There is negligible Ln background signal in EELS spectra of non-photooxidized cells so similar sensitivity might be expected, but EELS is an inherently insensitive technique. The major current limitation is probably the noise introduced by sample drift during the long energy-filtered exposures required for the images. Improvements are underway to boost the sensitivity of multicolor EM by increasing the amount of lanthanide that is deposited during the oxidation of Ln-DAB chelates. Further development should lead to a greater understanding of the relationship between structure and metal deposition and will improve signal-to-noise, decrease acquisition time and sample damage, and potentially permit greater resolution through tomography. The use of DAB to precipitate metals limits the scope of photooxidation because the polymer itself acts as a photosensitizer of further DAB oxidation up to about 600 nm, and limits the number of different Ln-DABS that can be successively deposited by spectrally distinct photosensitizers. Acylation of unreacted amino groups in the Ln-DAB precipitate diminished its undesired reaction with a subsequent Ln-DAB Efforts are underway to completely chemically block it and the photosensitizing effects of precipitated Ln-DAB, and thereby eliminate the present requirement for deconvolution that can be problematic when the elemental signals are all co-localized Significance Major improvements in multicolor and super-resolution fluorescence microscopy over the last two decades’ have dramatically improved our understanding of cellular microarchitecture and function. Comparable progress in electron microscopy has been achieved in throughput and automation but methods for marking multiple molecules of interest has been more limited. This work describes a new methodology for such selective detection or painting by sequential localized oxidation and precipitation of diaminobenzidine conjugates of Ln chelates by genetically-encoded photosensitizers, small molecule probes or peroxidases. Electron energy- loss spectroscopy of these orthogonally deposited lanthanide metals and their imaging by energy- filtered transmission electron microscopy yields elemental maps that can be displayed on conventional electron micrographs as color overlays. Experimental Procedures Reagents and solvents were from Sigma-Aldrich and cell culture reagents and probes were obtained from Life Technologies except where noted. All animal procedures were approved by the Institutional Animal Care and Use Committee of the University of California, San Diego. Synthesis of Ln-DAB2 Diethylene-triamine-N,N′,N″-triacetic acid bis(diaminobenzidine)amide, DTPA-DAB2: DTPA bis-anhydride (3.33 g, 9.33 mmol) suspended in dry DMF (33 ml) with triethylamine (1.30 ml, 9.33 mmol) under N2 was gently heated and bath sonicated until dissolved. After cooling to room temperature the solution was added dropwise over 30 min with stirring under N2, to diaminobenzidine (10 g, 46.65 mmol) and triethylamine (18.66 mmol, 2.60 ml) dissolved in dry DMF (33 ml). After stirring overnight at room temperature, the reaction mix was evaporated, dissolved in water (100 ml) and adjusted to pH 8 with 1N-NaOH until the pH stabilized (about 15 min). The mixture was stirred under N2 for one hour and then unreacted DAB was removed by filtration followed by extraction with EtOAc (3 × 100 ml). The aqueous layer was partially evaporated to remove EtOAc and then acidified to pH 5. 4 with conc. HCl to give the product as a gray precipitate which was collected by filtration and washed with water. Drying over P2O5 in vacuo overnight gave 2.75 g (38%) of a gray solid that was used without further purification. LC-MS indicated 70–80% purity with unreacted DAB as the remainder (Figure S2.) Ln-DAB2 solutions Ln-DAB2 solutions in cacodylate buffer were prepared immediately before use at room temperature. To make 10 ml of a 2 mM Ln, Ce or Pr-DAB2 solution, 15.6 mg (20 μmol) of DTPA-DAB2 were suspended in DMF (0.25 ml) and sonicated/vortexed to disperse. Water (8.33 ml) was added to give a cloudy solution that cleared on addition of LnCl3 aqueous solution (0.1 M of LaCl3.6H2O, CeCl3.6H2O, PrCl3.xH2O; the latter stock solution was dissolved in 0.1M HCl) with 120 μl (of La or Ce solutions) or 140 μl (of Pr solution) followed by vortexing and bath sonication to give clear light brown solutions. Aqueous NaOH solution (1 M) was added sequentially in six equal portions (6 × 10 μl) with vortexing after each addition. A precipitate was initially formed during the early steps of this neutralization but a mostly clear solution was present by the end. Cacodylate buffer (1.67 ml of 0.3 M of sodium cacodylate pH 7.4) was added, mixed and centrifuged (3000g, 10 min) to remove any precipitate. Solutions were syringe-filtered (0.22 μm, Millipore) immediately prior to addition to cells. Metal ion concentrations were measured by inductively-coupled plasma mass spectroscopy (Agilent 7700) Synthesis of Arg10-IRDye 700DX Arg10-IRDye 700DX was prepared by reaction of H2N-GGRRRRRRRRRR-CONH2 (where G and R are L-glycine and L-arginine respectively; synthesized by standard Fmoc chemistry with a Protein Technologies Prelude peptide synthesizer), with IRDye 700DX NHS ester (LICOR) in DMSO and N-methylmorpholine as base. The conjugate was purified by reverse-phase HPLC and characterized by LC-MS. Found, 575.0 (M+6H+), 689.6, (M+5H+), 861.9 (M+4H+), 1148.8 (M+3H+). Deconvolved to 3343.6, calc’d 3443.5. Eosin-sensitized photooxidation A solution of DAB or Ln-DAB2 (0.4 mM, diluted from freshly-prepared 100 mM stock solutions in DMF) and eosin (20 μM) in 100 mM sodium MOPS pH 7.2, or 0.1M sodium cacodylate pH 7.4 in a 3-ml cuvette was irradiated at 480 nm (30 nm band pass) using a solar simulator (Spectra-Physics 92191–1000 solar simulator with 1600 W mercury arc lamp and two Spectra-Physics SP66239–3767 dichroic mirrors to remove infrared and ultraviolet wavelengths). Remaining light was filtered through 10-cm square bandpass filters (Chroma Technology Corp.) with a deflector mirror set at 45° while bubbling with air. At set time points, the absorbance of the reaction was measured at 650 nm until the increase was complete. Cell culturing, labeling and transfection MDCK and HEK293 cells were cultured on poly-d-lysine coated 35-mmm glass bottom dishes (MatTEK) in Dulbecco’s modified Eagle’s medium supplemented with 10% FBS. MDCK Cells were labeled with 5 μM NBD C6-ceramide in media containing 10% fetal calf serum for 30 min, washed, and incubated for 30 mins in new culture media, all at 37°C with 5% carbon dioxide, then washed (x5) with HBSS at 37°C. Cells were then incubated with mouse monoclonal EpCAM antibody (KS1/4) (Santa Cruz Biotechnology) at 1:1000 dilution for 12 min at 37°C in HBSS, washed with HBSS, incubated for 60 min in secondary antibody biotinylated goat anti-mouse IgG (Jackson, 115–065–003) at 1:250 dilution at 37°C in HBSS and then washed. HEK293 cells were transfected with a miniSOG-Rab5a plasmid at 60–80% confluence using Lipofectamine 2000, which was removed after 8 hours. Cells were treated 48 hours after transfection with 2 μM Arg10-IRDye 700DX (prepared in distilled water) and added to the culture media for 2 hours at 37 °C, after which time the media was removed and cells were rinsed once in HBSS. Primary mouse cortical neurons were cultured, transfected with TS:YSOG3-PKMζ, and chemically stimulated with 50 μM forskolin and 0.1 μM rolipram for 10 min, then incubated with 1 μM BILN-2061 as previously described (Palida, et al., 2015). General procedure for photooxidation and HRP reaction of Ln-DAB2 Labeled cells were fixed and blocked (Shu, et al., 2011). Samples were then transferred either to a Bio-Rad MRC1024 with a Zeiss Axiovert 35M microscope or a Leica SPE microscope. MDCK cells stained with NBD C6-ceramide and transfected HEK293 cells exhibiting peptide uptake were identified either by NBD C6-ceramide, or miniSOG, and IRDye 700DX fluorescence respectively, and imaged by confocal microscopy. Freshly prepared and filtered La-DAB2 (2mM) or Ce-DAB2 (2 mM) for the first photooxidation reaction was added to the dish of cells for five min while a stream of pure oxygen was gently blown continuously over the solution. Cells were irradiated to excite NBD C6-ceramide depositing La-DAB2 or miniSOG depositing Ce-DAB2 reaction product both using 450–490 nm excitation (Ex) and 515 nm emission (Em) long-pass (LP) filters with a 580 nm dichroic mirror. Reaction product formation was monitored by transmitted light microscopy and illumination was stopped as soon as a light brown reaction product appeared. Acetic anhydride (MDCK cells) was added (20 × 20 mM freshly prepared) for 1 min each to block precipitated La-DAB2. Alternatively, MDCK cells were rinsed 3×5 min with fresh 100 mM acetyl imidazole in 0.15 M NaCl to prevent further polymerization of either La-DAB2 or Ce-DAB2, then treated with freshly prepared Ce-DAB2 (2 mM) for HRP enzymatic reaction or irradiated to excite IRDye 700DX (Ex 675/67nm, Em, 736 LP) depositing Pr-DAB2 reaction product. Cells were rinsed 5 × 2 min, post-fixed, dehydrated, infiltrated and embedded as previously described (Shu, et al., 2011). Hippocampal astrocyte filling with Lucifer yellow and neurobiotin Intracellular astrocyte filling with fluorescent dyes in fixed tissue A mouse (two-month old BALB-c male) was perfused with Ringers, followed by 4% paraformaldehyde, 0.2% glutaraldehyde in 0.1M PBS (Bushong, et al., 2002). 100 micron-thick coronal slices were cut through the hippocampus using a vibratome. In CA1 stratum radiatum, one astrocyte was iontophoretically injected with 5% Lucifer yellow-CH in water and an adjacent astrocyte was injected with 2.5% Alexa Fluor 568 / 2% neurobiotin (Vector, SP-1120) in 200 mM KCl (Bushong, et al., 2002). The tissue slices were then post fixed with 4% paraformaldehyde / 0.2% glutaraldehyde in 0.1M PBS. Confocal volumes were taken of the filled astrocytes with a Leica SPE inverted confocal microscope and the slices were further fixed with 2.0% glutaraldehyde in 0.15M sodium cacodylate buffer for 10 min on ice, followed by washing several times with 0.15M sodium cacodylate buffer pH 7.4. Photooxidation of Lucifer-yellow filled astrocytes Tissues were treated for 15 min in blocking buffer (50 mM glycine, 5 mM KCN, and 5 mM aminotriazole) to reduce nonspecific background reaction of diaminobenzidine derivatives, filtered Ce-DAB2 solution was added to a tissue slice at room temperature and incubated for 10 mins before photooxidation. A stream of pure oxygen was gently blown continuously over the solution. The Lucifer yellow was then excited using a standard FITC filter set (Ex 470/40, DM 510, Em BA520) with intense light from a 150W-xenon lamp. Illumination was stopped as soon as a light brown reaction product appeared within the filled astrocyte (8–10 min), as monitored by transmitted light. Blocking of first Ln-DAB2 product Following photooxidation each tissue slice was washed several times with cold 0.15 M sodium cacodylate buffer (pH 7.4) and then blocked with freshly prepared 100 mM acetyl imidazole in 0.15 M sodium chloride (3 × 5 min). HRP labeling and enzymatic reaction of photooxidized cells The tissues were incubated with cryoprotectant, then freeze-thawed to permeabilize the tissues (Knott, et al., 2009). The tissues were washed several times with 0.15 M sodium cacodylate buffer pH 7.4, incubated with 1% BSA in the cacodylate buffer for 30 min followed by overnight incubation with Vectorstain Elite ABC staining kit (Vector, PK6100). After rinsing several times for one hour with cold 0.1M cacodylate buffer pH 7.4, 20 ml of Pr-DAB2 solution and 5 μl 30% H2O2 were added to each tissue. After the neurobiotin-filled astrocyte turned brown, the tissue was washed several times with 0.15 M sodium cacodylate buffer pH 7.4. Tissue processing for TEM Tissues were fixed with 2% glutaraldehyde in cacodylate buffer for 20 min, washed several times with cacodylate buffer, post-fixed with 0.5% of OsO4 in cacodylate buffer for 30 min, dehydrated in an ethanol series of 0%, 20%, 50%, 70%, 90% and 100% on ice for 5 min each, 100% ethanol twice for five mins each at room temperature, 1:1 100% dry ethanol : dry acetone for five min, 100% dry acetone for five min, 50:50 dry acetone: Durcupan ACM for 30 min, four changes of Durcupan for one hour each, and embedded in a 60 °C oven for 48 hours. Section preparation from cells and tissue 100nm thick sections were cut by an Ultra 45° Diatome diamond knife using a Leica Ultracut UCT ultramicrotome and sections were picked up on a 50 mesh copper grid (Ted Pella, G50). Sections were carbon coated on both sides by a Cressington 208 carbon coater to prevent charging of the plastic which can cause drift and thermal damage. Electron microscopy EFTEM was performed with a JEOL JEM-3200EF transmission electron microscope operating at either 200 or 300 KV, equipped with an in-column Omega filter and a LaB6 electron source. The samples were pre-irradiated at a low magnification of 100x for about 30 min to stabilize the sample and minimize contamination (Egerton, et al., 2004). The elemental maps were obtained at the M4,5 core-loss edge, the onset of which occurs at 832, 883 and 931 for lanthanum (La), cerium (Ce) and praseodymium, (Pr) respectively (Ahn and Krivanek, 1983). The EFTEM images of the pre and post-edges were obtained using a slit of 30 eV width. The electron energy-loss spectrum was acquired using the Ultrascan 4000 CCD detector from Gatan (Pleasanton, CA, USA). The conventional images and elemental maps were acquired using both the Ultrascan 4000 detector and the direct detection device (DDD) DE-12 from Direct Electron LP. (San Diego, CA). See supplemental experimental procedures for details. Two-color elemental map merge and overlay on TEM Elemental maps and TEM were pre-aligned in Photoshop (Adobe) and merged pixel-by-pixel using the following custom algorithm running in C++. (DisplayPixel)R,G,B=(1-T)·(PixelTEM)R,G,B+T·(PixelOVR)R,G,B Where PixelTEM is the 24-bit RGB coordinate for the gray scale TEM image, and PixelOVR is the 24-bit RGB coordinate for the overlay hue at maximum saturation and brightness. T is a “transparency factor” whose value determines what percentage of the overlay color coordinate contributes to the final display pixel. The color coordinate for PixelOVR is calculated in the HSL (Hue, Saturation and Lightness) color space. In the following example, the hue (H) for PixelOVR is determined using two EELS channels: S1=I1/(I1+I2)S2=I2/(I1+I2)H=(S1·H1)+(S2·H2) Where S is a scale factor and I is the intensity of the signal for each respective channel and H1 and H2 are the arbitrary hues selected for each channels (e.g., red and green). The lightness (I) for PixelOVR is calculated as follows: I=(S1·I1)+(S2·I2) Finally, the RGB coordinate for PixelOVR is determined using an HSL to RGB color space conversion algorithm where S is set to maximum. The transparency factor is calculated as follows: T=I/255 The raw EELS data needs to be scaled between 0 and 255 (8-bits) and should be done prior to implementing the algorithm. Supplementary Material supplement We thank David Mastronarde (University of Colorado Boulder) and Liang Jin (Direct Electron) for help with scripting in Serial EM, and DE-12 direct detection device respectively, and James Bouwer, Thomas Deerinck, Junru Hu for advice and technical assistance. This work was supported by UCSD Graduate Training Programs in Cellular and Molecular Pharmacology (T32 GM007752)(SP) and in Neuroplasticity of Aging (T32 AG000216)(SP), NIH GM103412 (ME), the W.M. Keck Foundation (ME) and NIH GM086197 (RT). Figure 1 Two-color EM using EELS and EFTEM (A) Scheme of process applied to cells with stained mitochondrial (red) and nuclear membranes (green) are first selectively irradiated to photooxidize the red photosensitizer and precipitate Ce-DAB2 (brown ring). After washing and replacement with Pr-DAB2, illumination at an orthogonal wavelength generates precipitate at the nuclear membrane. Alternatively, the Pr-DAB2 can be oxidized by hydrogen peroxide following immunoperoxidase labeling. Following conventional osmification (black ring), embedding, sectioning and TEM, EFTEM yields pseudocolored elemental maps for Ce and Pr that are overlaid on the conventional osmium image (B) Structure of Ln-DAB2 (C) La- and Ce-DAB2 are precipitated at a similar rate to DAB by photosensitization of eosin Figure 2 Two color EM of Golgi and plasma membrane in tissue culture cells (A) Conventional TEM image (x 6000, 1.4 nm/pixel) of MDCK cell following photooxidation with La-DAB2 by NBD ceramide-labeled Golgi and subsequent oxidation of Ce-DAB2 on plasma membrane (PM) with HRP-labeled antibody to EpCAM. (B) Spectra obtained at Golgi region (green spectra), and at PM containing EpCAM (red spectra); the respective regions are shown as circles in A. The Golgi region shows a strong La signal with weak Ce signal, the PM region shows only Ce. (B′) The cross-talk in the spectra shown in B, due to Ce-DAB2 attaching to regions labeled with La-DAB2 has been mathematically subtracted. (C) & (D) La and Ce elemental map obtained by the five-window method on the CCD (bin by four pixels), each energy window image was a sum of nine individual drift corrected images, each acquired for a 60 s exposure. A Gaussian smoothing of blur radius 1 was applied to the images. (D′) The Ce map shown in D, mathematically corrected to remove the cross-talk due to Ce-DAB2 attaching to regions labeled with La-DAB2. (E) Two color merge of the elemental maps (La in green and Ce in red), overlaid on the conventional TEM image. Figure 3 Two-color EM of hippocampal astrocytes in brain slices (A) Correlative fluorescent image of adjacent hippocampal astrocytes injected with Lucifer yellow or neurobiotin/Alexa 568. The white box is the approximate region of the EM acquisition, shown in the following panels. (B) Conventional TEM image (x 15000, 0.56 nm/pixel) showing astrocyte processes containing precipitated Ce- and Pr-DAB2 complexes contacting two spines synapsing (asterisk) with a bouton (PSDs indicated with arrows). (C) EELS spectra obtained at the lower astrocyte (green spectra) photooxidized with Ce-DAB2 and upper astrocyte (red spectra) HRP enzymatically reacted Pr-DAB2; the respective regions are shown as circles in B. The upper astrocyte contains predominately Pr, whereas the bottom has Pr and Ce signals. (C′) The cross-talk in the spectra shown in C, due to Pr-DAB2 attaching to regions labeled with Ce-DAB2 has been mathematically subtracted. (D) & (E) Ce and Pr elemental maps (x 25000; five window method using a sum of 19 drift-corrected 50 s exposures per window). (E′) Corrected Pr map, removing Pr-DAB from regions with Ce-DAB. (F) Two-color merge of the spectrally-separated elemental maps (green for Ce and red for Pr) overlaid on a conventional image, showing the two different astrocyte processes contacting the same synapse. Figure 4 Two-color EM of endosomal uptake of CPPs (A) Transmitted light image of miniSOG-Rab5a transfected HEK293 cell treated with Arg10-IRDye 700DX showing endosomes (arrows) following photooxidation. Corresponding fluorescent images of (B) miniSOG-Rab5a, (C) Arg10-IRDye 700DX labeling and (D) two-color overlay of fluorescent images with typical endosomes arrowed. (E) Conventional TEM image (x 20000, 0.45 nm/pixel) of an endosome after consecutive photo-oxidization of Ce-(miniSOG-Rab5a) and Pr-DAB2 (Arg10-IRDye 700DX). The inset is the EELS spectrum from the circled region and the rectangular box is the region of elemental map acquisitions, shown in the subsequent panels. (F) and (G) Ce and Pr elemental maps respectively of endosome (x 25000, 0.2nm/pixel; five-window method using a sum of 12 drift-corrected 50 s exposure per window, smoothed with Gaussian blur radius 3). (H) Two-color merge of the Ce (green) and Pr (red) elemental maps overlaid on the TEM image reveals the Arg10 is lumenal and Rab5a is on the cytoplasmic face of the endosome. (I) TEM image (x 20000, 0.45 nm/pixel) of a MVB from the same cell with the EELS spectrum of the circled region inset. (J) & (K) Ce and Pr elemental maps respectively of the MVB (x 20000, 1.8 nm/pixel; five-window method on CCD detector (bin by four pixels) using a sum of six drift-corrected 40 s exposures, smoothed with Gaussian blur radius 1). (L) Two-color merge of the Ce (green) and Pr (red) elemental maps), overlaid on TEM image showing vesicular but no cytoplasmic Rab5a. Figure 5 Chemical LTP results in PKMζlocalization to the postsynaptic membrane (A) Transmitted light image of photooxidized cultured neurons expressing TS:YSOG3-PKMζ, stimulated with forskolin and rolipram, treated with BILN-2061 for 24 hours, and photo-oxidized with Ce-DAB2. (B) TEM image (x 6000, 0.9nm/pixel) of a neuron (white box in A) showing enhanced post-synaptic electron density. (C) Ce elemental map (x 8000, 4.4 nm/pixel; three-window method using a sum of seven drift-corrected 40 s exposures per window, smoothed with a Gaussian blur 1) showing that Ce localization corresponds to photooxidation and confirming that PKMζstrongly localizes to the postsynaptic membrane after stimulation. (D) Overlay of Ce map (green) on the TEM image. (E) TEM image at a higher magnification of the region shown in B (white box). (F) Overlay of Ce map at higher magnification (x 25000, 0.2 nm/pixel; three-window method using a sum of 25 drift-corrected 50 s exposures per window, smoothed with Gaussian blur 3) on the TEM image. (G) Transmitted light image of unstimulated neurons expressing TS:YSOG3-PKMζtreated with BILN-2061 for 24 hours and photo-oxidized with Ce-DAB2. (H) TEM image of neuron shown in G (white box). (I) Ce elemental map (x 8000, 4.4nm/pixel; three-window method using a sum of seven drift-corrected 40 s exposures per window, smoothed with a Gaussian blur 1, of the region in H scaled similarly to C indicating minimal Ce deposition and that basally produced TS:YSOG3-PKMζis not post-synaptically localized in the image. J) EELS quantification of the relative Ce signal spectra of the circular regions shown in B (green circle) and H (black circle), normalized to the same background counts. Highlights Multicolor EM paints multiple cellular markers by locally depositing specific Ln3+ Each Ln3+ visualized by electron energy-loss spectroscopy and energy-filtered EM Elemental maps overlaid on conventional EM give multicolor EM Applicable to immuno, genetically-encoded, and molecular probes in cells and tissue Author Contributions S.A., R.T., and M.E. conceived and designed the experiments, S.A., M.M., R.R., S.P., E.B., B.G., M.B., and P.S. performed all experiments and analyzed data, and S.A., M.M., R.R., and R.T. wrote the manuscript. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. Ahn CC Krivanek OL 1983 EELS Atlas (Gatan) Aronova MA Leapman RD 2012 Development of Electron Energy Loss Spectroscopy in the Biological Sciences MRS Bull 37 53 62 23049161 Betzig E Patterson GH Sougrat R Lindwasser OW Olenych S Bonifacino JS Davidson MW Lippincott-Schwartz J Hess HF 2006 Imaging intracellular fluorescent proteins at nanometer resolution Science 313 1642 1645 16902090 Brock R 2014 The uptake of arginine-rich cell-penetrating peptides: putting the puzzle together Bioconjug Chem 25 863 868 24679171 Bushong EA Martone ME Ellisman MH 2004 Maturation of astrocyte morphology and the establishment of astrocyte domains during postnatal hippocampal development Int J Dev Neurosci 22 73 86 15036382 Bushong EA Martone ME Jones YZ Ellisman MH 2002 Protoplasmic astrocytes in CA1 stratum radiatum occupy separate anatomical domains J Neurosci 22 183 192 11756501 Butko MT Yang J Geng Y Kim HJ Jeon NL Shu X Mackey MR Ellisman MH Tsien RY Lin MZ 2012 Fluorescent and photo-oxidizing TimeSTAMP tags track protein fates in light and electron microscopy Nat Neurosci 15 1742 1751 23103964 Copolovici DM Langel K Eriste E Langel Ü 2014 Cell-penetrating peptides: design, synthesis, and applications ACS Nano 8 1972 1994 24559246 Deerinck TJ Martone ME Levram V Green DPL Tsien RY Spector DL Huang S Ellisman MH 1994 Fluorescence photooxidation with Eosin - A method of high-resolution immunolocalization and in-situ hybridization detection for light and electron-microscopy Journal of Cell Biology 126 901 910 7519623 Durner J Klessig DF 1995 Inhibition of ascorbate peroxidase by salicylic acid and 2,6-dichloroisonicotinic acid, two inducers of plant defense responses Proc Natl Acad Sci U S A 92 11312 11316 7479986 Egerton RF 1996 Electron Energy-Loss Spectroscopy Plenum Press Egerton RF Li P Malac M 2004 Radiation damage in the TEM and SEM Micron 35 399 409 15120123 Fernandez JM Bilgin MD Grossweiner LI 1997 Singlet oxygen generation by photodynamic agents Journal of Photochemistry and Photobiology B: Biology 37 131 140 Gaietta G Deerinck TJ Adams SR Bouwer J Tour O Laird DW Sosinsky GE Tsien RY Ellisman MH 2002 Multicolor and electron microscopic imaging of connexin trafficking Science 296 503 507 11964472 Geze C Mouro C Hindre F LePlouzennec M Moinet C Rolland R Alderighi L Vacca A Simonneaux G 1996 Synthesis, characterization and relaxivity of functionalized aromatic amide DTPA-lanthanide complexes Bulletin De La Societe Chimique De France 133 267 272 Halassa MM Fellin T Takano H Dong JH Haydon PG 2007 Synaptic islands defined by the territory of a single astrocyte J Neurosci 27 6473 6477 17567808 Haydon PG 2001 GLIA: listening and talking to the synapse Nat Rev Neurosci 2 185 193 11256079 Haydon PG Carmignoto G 2006 Astrocyte control of synaptic transmission and neurovascular coupling Physiol Rev 86 1009 1031 16816144 Hell SW 2007 Far-field optical nanoscopy Science 316 1153 1158 17525330 Hernández AI Oxberry WC Crary JF Mirra SS Sacktor TC 2014 Cellular and subcellular localization of PKMζ Philos Trans R Soc Lond B Biol Sci 369 20130140 24298142 Huang B Bates M Zhuang X 2009 Super-resolution fluorescence microscopy Annu Rev Biochem 78 993 1016 19489737 Jullian C Brunet JE Thomas V Jameson DM 1989 Time-resolved fluorescence studies on protoporphyrin IX-apohorseradish peroxidase Biochim Biophys Acta 997 206 210 2765557 Kaplan IM Wadia JS Dowdy SF 2005 Cationic TAT peptide transduction domain enters cells by macropinocytosis J Control Release 102 247 253 15653149 Knott GW Holtmaat A Trachtenberg JT Svoboda K Welker E 2009 A protocol for preparing GFP-labeled neurons previously imaged in vivo and in slice preparations for light and electron microscopic analysis Nat Protocols 4 1145 1156 19617886 Kuipers J van Ham TJ Kalicharan RD Veenstra-Algra A Sjollema KA Dijk F Schnell U Giepmans BN 2015 FLIPPER, a combinatorial probe for correlated live imaging and electron microscopy, allows identification and quantitative analysis of various cells and organelles Cell Tissue Res 360 61 70 25786736 Lam SS Martell JD Kamer KJ Deerinck TJ Ellisman MH Mootha VK Ting AY 2015 Directed evolution of APEX2 for electron microscopy and proximity labeling Nat Methods 12 51 54 25419960 Maranto AR 1982 Neuronal mapping: a photooxidation reaction makes Lucifer yellow useful for electron microscopy Science 217 953 955 7112109 Martell JD Deerinck TJ Sancak Y Poulos TL Mootha VK Sosinsky GE Ellisman MH Ting AY 2012 Engineered ascorbate peroxidase as a genetically encoded reporter for electron microscopy Nat Biotechnol 30 1143 1148 23086203 Mitsunaga M Ogawa M Kosaka N Rosenblum LT Choyke PL Kobayashi H 2011 Cancer cell-selective in vivo near infrared photoimmunotherapy targeting specific membrane molecules Nat Med 17 1685 1691 22057348 Natera J Massad W Amat-Guerri F Garcia N 2011 Elementary processes in the eosin-sensitized photooxidation of 3,3 ′-diaminobenzidine for correlative fluorescence and electron microscopy Journal of Photochemistry and Photobiology a-Chemistry 220 25 30 Oakenfull D Jencks W 1971 Reactions of Acetylimidazole and Acetylimidazolium Ion with Nucleophilic Reagents - Structure-Reactivity Relationships Journal of the American Chemical Society 93 178 188 Pagano RE Martin OC Kang HC Haugland RP 1991 A novel fluorescent ceramide analogue for studying membrane traffic in animal cells: accumulation at the Golgi apparatus results in altered spectral properties of the sphingolipid precursor J Cell Biol 113 1267 1279 2045412 Pagano RE Sepanski MA Martin OC 1989 Molecular trapping of a fluorescent ceramide analogue at the Golgi apparatus of fixed cells: interaction with endogenous lipids provides a trans-Golgi marker for both light and electron microscopy J Cell Biol 109 2067 2079 2478562 Palida SF Butko MT Ngo JT Mackey MR Gross LA Ellisman MH Tsien RY 2015 PKMζ, but not PKCζ, is rapidly synthesized and degraded at the neuronal synapse J Neurosci 35 7736 7749 25995463 Peng X Draney D Volcheck W Bashford G Lamb D Grone D Zhang Y Johnson C Achilefu S Bornhop D 2006 Phthalocyanine dye as an extremely photostable and highly fluorescent near-infrared labeling reagent - art. no. 60970E Optical Molecular Probes for Biomedical Applications 6097 E970 E970 Richard JP Melikov K Vives E Ramos C Verbeure B Gait MJ Chernomordik LV Lebleu B 2003 Cell-penetrating peptides. A reevaluation of the mechanism of cellular uptake J Biol Chem 278 585 590 12411431 Rink J Ghigo E Kalaidzidis Y Zerial M 2005 Rab conversion as a mechanism of progression from early to late endosomes Cell 122 735 749 16143105 Schnell U Cirulli V Giepmans BN 2013 EpCAM: structure and function in health and disease Biochim Biophys Acta 1828 1989 2001 23618806 Schnell U Dijk F Sjollema KA Giepmans BN 2012 Immunolabeling artifacts and the need for live-cell imaging Nat Methods 9 152 158 22290187 Shao CY Sondhi R van de Nes PS Sacktor TC 2012 PKMζis necessary and sufficient for synaptic clustering of PSD-95 Hippocampus 22 1501 1507 22378468 Shu XK Lev-Ram V Deerinck TJ Qi YC Ramko EB Davidson MW Jin YS Ellisman MH Tsien RY 2011 A Genetically Encoded Tag for Correlated Light and Electron Microscopy of Intact Cells, Tissues, and Organisms Plos Biology 9 10 Takizawa PA Yucel JK Veit B Faulkner DJ Deerinck T Soto G Ellisman M Malhotra V 1993 Complete vesiculation of Golgi membranes and inhibition of protein transport by a novel sea sponge metabolite, ilimaquinone Cell 73 1079 1090 8513494

PMC005xxxxxx/PMC5127403.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 0413066 2830 Cell Cell Cell 0092-8674 1097-4172 27814521 5127403 10.1016/j.cell.2016.10.027 NIHMS824282 Article The Central Nervous System and the Gut Microbiome Sharon Gil 1* Sampson Timothy R. 1 Geschwind Daniel H. 23456 Mazmanian Sarkis K. 1* 1 Division of Biology and Biological Engineering, California Institute of Technology, Pasadena, California 91125, USA 2 Program in Neurobehavioral Genetics, Semel Institute, and Program in Neurogenetics, University of California, Los Angeles, California 90095, USA 3 Department of Neurology, David Geffen School of Medicine, University of California, Los Angeles, California 90095, USA 4 Interdepartmental Program in Neuroscience, University of California, Los Angeles, California 90095, USA 5 Center for Autism Treatment and Research, Semel Institute, David Geffen School of Medicine, University of California, Los Angeles, California 90095, USA 6 Department of Human Genetics, David Geffen School of Medicine, University of California, Los Angeles, California 90095, USA Correspondences: gsharon@caltech.edu & sarkis@caltech.edu 22 10 2016 3 11 2016 03 11 2017 167 4 915932 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. Summary Neurodevelopment is a complex process governed by both intrinsic and extrinsic signals. While historically studied by researching the brain, inputs from the periphery impact many neurological conditions. Indeed, emerging data suggests communication between the gut and the brain in anxiety, depression, cognition and autism spectrum disorder (ASD). The development of a healthy, functional brain depends on key pre- and post-natal events that integrate environmental cues, such as molecular signals from the gut. These cues largely originate from the microbiome, the consortium of symbiotic bacteria that reside within all animals. Research over the past few years reveals that the gut microbiome plays a role in basic neurogenerative processes such as the formation of the blood-brain-barrier, myelination, neurogenesis, and microglia maturation, and also modulates many aspects of animal behavior. Herein, we discuss the biological intersection of neurodevelopment and the microbiome, and explore the hypothesis that gut bacteria are integral contributors to development and function of the nervous system, and the balance between mental health and disease. Introduction The development of the mammalian brain is an intricate process that lasts through adolescence and into early adulthood in humans. Further, the process of brain development involves extraordinary, large-scale long distance migration of cells during fetal development to specific regions or layers, as well as navigation of their processes across even longer distances (often hundreds of cell body diameters) to build the specific circuits that underlie behavior (Geschwind and Rakic, 2013; Marín and Rubenstein, 2003). The complexity and protracted pre- and post-natal time course over which these events occur, makes them highly sensitive and even vulnerable to environmental factors. In fact, many of the processes governing brain development are driven by extrinsic cues and experiences that shape the developing brain through both generative and regressive events. As the gut is our largest portal to the molecular universe, various dietary components have been shown to interact directly with the developing brain and to induce functional alterations in the mature brain (Chang et al., 2009; Zeisel, 2004), and there is now mounting evidence for a role by the gut microbiome in directing and facilitating developmental processes in the brain with long term implications to health (Figure 1, Table 1). The mammalian microbiome consists of unique assemblages of microorganisms (i.e., bacteria, archaea, fungi, and viruses) associated with various niches in and on the body. Research in animal models and humans has inextricably linked gut bacteria to the development and function of the immune system. The presence of entire immune cell types requires the microbiome, and specific microbes have been discovered that either promote or ameliorate immunologic disorders such as type 1 diabetes, asthma and inflammatory bowel disease (Round and Mazmanian, 2009). If the gut microbiome can so profoundly impact the immune system, why would its influence not reach the nervous system? Indeed, germ-free (GF) mice, devoid of all associated microorganisms, exhibit increased risk-taking behaviors and hyperactivity, while also displaying learning and memory deficits compared to conventional (specific pathogen free; SPF) mice (Clarke et al., 2013; Gareau et al., 2011; Heijtz et al., 2011; Neufeld et al., 2011) (Table 1). Further, GF mice show changes in expression of the 5-hydroxytryptamine receptor (5-HT1A), neurotrophic factors (e.g., BDNF), and NMDA receptor subunits in the hippocampus (Bercik et al., 2011a; Heijtz et al., 2011; Sudo et al., 2004), while also displaying impaired blood-brain barrier function, as well as increased myelination in the prefrontal cortex (Braniste et al., 2014; Hoban et al., 2016)(Table 1; Figure 1). There is also evidence, albeit preliminary and mostly from animal models, for a potential role for the microbiome in neuropsychiatric conditions, including depression and anxiety (Foster and McVey Neufeld, 2013), autism spectrum disorder (ASD) (Krajmalnik-Brown et al., 2015), schizophrenia (Severance et al., 2014) and even Parkinson’s (PD) and Alzheimer’s disease (AD) (Keshavarzian et al., 2015). In this Review, we discuss the intersection between the mammalian microbiome and the brain in both humans and animal models. We explore developmental trajectories and the outcomes of interactions between microbes and the brain during prenatal development, as well as postnatally and to adulthood where microbial communities are established. Further, we highlight potential paradigms by which host-associated microorganisms may play an active role in both supporting health and potentiating disease states. By correlating microbial activities to progressive structural and functional events in the brain in mouse models and humans, we propose pathways whereby the gut microbiome may contribute to neurodevelopment and neurodegeneration. Uncovering microbial and host pathways that regulate these connections may provide novel approaches for addressing behavioral, psychiatric and neurodegenerative disorders. Major processes in neurodevelopment coincide with changes in the maternal and neonatal gut microbiome Prenatal brain development On the third week after conception in humans, after gastrulation is complete, neural stem cells differentiate from the epiblast, marking the first event in a sequence that would eventually result in the adult brain. In the cerebral cortex, which is most elaborated in humans (Geschwind and Rakic, 2013), neural progenitor cells proliferate in the ventricular zone (NPCs; also known as radial glial cells) proliferate in the ventricular zone. Committed progenitors or neurons that migrate over a distance of many cell body diameters through the intermediate zone to the cortical plate, while NPCs remain in the proliferative zone (Kriegstein and Alvarez-Buylla, 2009; Rakic, 1988). While cortical neurogenesis is complete (for the most part) by mid gestation, gliogenesis is primarily a postnatal process. Anterior-posterior and dorsal-ventral patterning of the nervous system occurs via the same basic rules, factors and pathways that pattern the body, and the signature of the regulatory factor gradients that govern the “Protomap” that underlies the spatial and structural differentiation of the brain are present in germinal zones of the developing embryonic brain (Marín and Rubenstein, 2003; Rakic, 1988; Stiles and Jernigan, 2010). Following a massive neuronal expansion and migration during cortical development, approximately 50% of neurons undergo apoptosis during the final weeks of gestation (38–41) and only those that have been integrated into networks, and supported by neurotrophic signals, survive (Ceni et al., 2014). Neurotrophins, amongst them brain-derived neurotrophic factor (BDNF), serve as signals for neuron survival, promote maintenance and differentiation of various cell populations (Ichim et al., 2012), and mediate various stages of neuronal circuitry establishment (Park and Poo, 2013). According to the neurotrophic hypothesis, the more connections a neuron makes, the higher the concentration of neurotrophins around it, and hence the higher chance of survival (Oppenheim, 1989). Neurogenesis is influenced by the presence of microorganisms. Specifically, neurogenesis in the dorsal hippocampus of adult GF mice is increased compared to conventional mice (Ogbonnaya et al., 2015). Interestingly, colonization of GF mice at weaning could not reverse this phenotype, indicating that microbial signals very early in life reduce rates of neurogenesis in the hippocampus. Moreover, adult GF mice exhibit increased volume of the amygdala and hippocampus (specifically CA2/3), and differ in dendrite morphology, while no differences in total brain volume were recorded between GF and SPF animals (Luczynski et al., 2016a). By tapping into pathways that govern neuronal differentiation and survival, via neurotrophins and their receptors, gut microbes can influence the fate of neurons in various regions of the brain and subsequently neurodevelopment and health. In contrast to neurons, astroglia, and oligodendrocytes, microglia are central nervous system (CNS)-resident innate immune cells derived from a primitive subset of macrophages in the yolk sac. Although still controversial, recent reports indicate that microglia are not derived from monocytes, but rather develop earlier and express distinct cell markers, different from bone marrow-derived macrophages (specifically Ly-6ChiCCR2+ monocytes) that replenish the CNS following brain injury (Bennett et al., 2016; Ginhoux et al., 2010; Nayak et al., 2014; Varvel et al., 2012). In contrast, microglia were shown to originate from CD45-c-kit+ erythromyeloid progenitor cells and mature as CD45+c-kit-CX3CR1+ cells in the CNS. In mice, studies show that microglia enter the brain through circulation by embryonic day 8.5 (E8.5), start expressing the microglial marker Tmem119 as early as postnatal day 6 and become fully ramified throughout the brain by postnatal day 28 (Bennett et al., 2016; Nayak et al., 2014). Microglia originate from yolk-sac progenitor cells, and can be replenished by bone-marrow derived macrophages upon insult; both cell types can be subjected to microbial signals during early development (Erny et al., 2015; Khosravi et al., 2014). A central role for the microbiota in the development and maturation of the microglia has recently emerged (Erny et al., 2015; Matcovitch-Natan et al., 2016). In the absence of the microbiota, mice harbor microglia with significantly altered developmental states. These microglia display morphological characteristics and a gene expression profile that indicate an arrest in their developmental maturation, and subsequently are maintained in an immature status (Erny et al., 2015; Matcovitch-Natan et al., 2016). Notably, microglia derived from GF mice display limited responses towards viral infection and microbially-associated molecular patterns (MAMPs). Such defective responses can be rescued by administration of short-chain fatty-acids (SCFAs)(Erny et al., 2015). The blood-brain-barrier (BBB) forms during gestation and serves as a selective barrier between the brain and circulation. The importance of gut microbiome and microbial metabolites in the formation of the BBB has been exemplified in GF mice (Braniste et al., 2014). In the absence of gut microorganisms, the BBB is more permeable to macromolecules, compared to conventionally-raised animals, mediated by decreased expression of key tight-junction proteins in the brain endothelium. Furthermore, permeability decreased upon colonization of GF animals, or, alternatively, administration of the SCFA butyrate that is produced as a result of bacterial fermentation in the gut (Braniste et al., 2014). Correspondingly, the BBB in the sterile fetus is permeable, compared to the adult BBB (Møllgård and Saunders, 1986). Recently, a lymphatic vasculature of the brain that drains from the cerebrospinal fluid in the adjacent subarachnoid space and the intrastitial fluid, to the deep cervical lymph nodes was discovered (Aspelund et al., 2015). This network allows easy passage for various immune cells as well as macromolecules and metabolites in and out of the brain (Aspelund et al., 2015; Louveau et al., 2015). The BBB and the brain lymphatic vasculature serve as a gateway for various signals to the brain, such as circulating immune cells and soluble molecules (including hormones and neurotransmitters, both host and microbial in origin), and along with stimulation of the vagus nerve, represent mechanisms that facilitate direct and indirect transmission of microbial signals from the gut to the brain. The maternal gut and vaginal microbiome during pregnancy and its effect on offspring behavior The maternal microbiome is distinct and dynamically changing during pregnancy compared to other periods in female life (DiGiulio et al., 2015; Koren et al., 2012; MacIntyre et al., 2015; Nuriel-Ohayon et al., 2016; Romero et al., 2014a). With the progression of pregnancy, various taxa change in abundance with some becoming more dominant over others. For example, in the gut some Proteobacteria and Actinobacteria appear to increase in relative abundance in the third trimester, compared to the first (DiGiulio et al., 2015; Koren et al., 2012). The vaginal microbiome during pregnancy remains dominated by Lactobacillus species; however as pregnancy progresses the species composition varies between stable, common states termed community state types (CSTs). CSTs with higher microbial diversity in the vagina are associated with preterm birth (DiGiulio et al., 2015; Romero et al., 2014b). Some data suggest that the mammalian fetus is not necessarily sterile as commonly predicted. Studies advocate that the placenta and, at times, the amniotic fluid surrounding the fetus, harbor distinct bacterial populations (Aagaard et al., 2014; DiGiulio, 2012; Kuperman and Koren, 2016; Zheng et al., 2015a); however these findings remain controversial and require further investigation, as Lauder and colleagues (2016) demonstrated. Significant work is needed to validate and implement this information for therapeutic or diagnostic applications. Variation in maternal microbial populations have been suggested to modulate the microbiome, neurodevelopment, and behavior of the offspring (Jašarević et al., 2015a). Perinatal administration of antibiotics can affect offspring health and immune status in both humans and mouse models (Russell et al., 2013; Stensballe et al., 2013). Administration of nonabsorbable antibiotics to rodent dams resulted in shifts of both the maternal and offspring gut microbiomes and induced hypoactivity compared to controls (Degroote et al., 2016; Tochitani et al., 2016). Moreover, offspring exhibited anxiety-like behavior and deficits in locomotion (Tochitani et al., 2016). Similarly, offspring Wistar rats exhibited reduced social behavior and increased anxiety as a result of nonabsorbable antibiotics administration to dams early in gestation (Degroote et al., 2016). A recent report indicates that maternal diet may also change both the microbial population and behavior of offspring (Buffington et al., 2016), following clinical data showing an association between maternal obesity and an ASD diagnosis in children (Krakowiak et al., 2012). A maternal high-fat diet in mice was sufficient to render offspring that are less social and exhibited repetitive behavior, compared to controls fed normal chow. The social deficit in these mice could be reversed by administration of Lactobacillus reuteri, found to be missing in the gut microbiome of high-fat diet offspring (Buffington et al., 2016). Similarly, behavioral deficits induced by antibiotic treatment during pregnancy could be rescued by cross-fostering offspring with control dams (Degroote et al., 2016; Tochitani et al., 2016). These examples indicate that maternal microbial populations can impact behavioral outcomes in the offspring. Whether these changes are mediated indirectly through effects on maternal behavior, or directly alter fetal brain development, remains to be determined. While the fetal environment may or may not harbor a microbiome, the fetus is inarguably exposed to microbial products from the mother, such as secondary metabolites, fermentation products, LPS, and/or peptidoglycan (PG). MacPherson and colleagues (2016) elegantly demonstrated that microbial metabolites produced in the maternal gut during transient colonization of otherwise GF mice with an auxotrophic strain of Escherichia coli can reach the fetal compartment and induce a specific developmental program prenatally. Bacterial cell wall components can also affect offspring: PG can cross the placenta and reach the fetal brain, where it induces proliferation of neurons in the frontal cortex, via increased expression of FOXG1, a critical regulator of forebrain development and neurogenesis. Offspring exposed to PG prenatally exhibit decreased cognitive function (Humann et al., 2016). Exposure to other microbial products prenatally and neonatally has impacts on offspring behavior. Offspring to dams exposed to propionic acid or LPS, injected subcutaneously, exhibited anxiety-like behaviors (Foley et al., 2014). These effects were observed even when offspring were exposed neonatally, suggesting a direct effect on the offspring rather than on maternal behavior (Foley et al., 2014). The maternal microbiome, the microbiome transmitted to offspring, their metabolites, and other microbial products are important in driving a developmental program in a healthy trajectory, and when perturbed are sufficient to induce behavioral deficits in offspring. The maternal immune system comes in close interaction with both gut microorganisms and the fetus. Immune activation during gestation has potentially severe implications on offspring physiology, neuropathology and behavior, as well as the microbiome (reviewed by Estes and McAllister, 2016; Knuesel et al., 2014). Large-scale epidemiological studies have demonstrated that prenatal infections significantly increase the risk for schizophrenia in offspring (Khandaker et al., 2013), and data support involvement in ASD as well, although less conclusively (Gardener et al., 2011). Based on these findings, rodent models for maternal immune activation (MIA) have been developed, where prenatal administration of the Toll-like receptor ligands LPS or Poly(inosine:cytosine) as surrogates for infection induced detrimental effects on offspring neuropathology and behavior (Estes and McAllister, 2016). Furthermore, changes in offspring microbiome following MIA have been reported with implications on the metabolomic profile in the serum of these offspring (Hsiao et al., 2013). Intervention with the human commensal bacterium Bacteroides fragilis corrected many of the adverse effects induced by MIA (Hsiao et al., 2013). Specifically, B. fragilis treatment decreased intestinal barrier permeability and lowered the concentration of potentially pathogenic metabolites (Hsiao et al., 2013). Furthermore, a recent study found that MIA phenotypes depend on Th17 cells and the production of IL-17A (Choi et al., 2016); Interestingly, the development of this T helper cell was previously shown to depend on gut bacteria (Ivanov et al., 2009). MIA models demonstrate one possible axis through which the gut microbiome and immune system act in concert to shape offspring physiology, behavior, and neuropathology. Postnatal brain development Postnatal brain development is predominantly governed by synaptic development and plasticity, including the overproduction and elimination of synapses during the first decade of human life (Paolicelli et al., 2011; Zuchero and Barres, 2015). Although neurogenesis is highly limited postnatally and is confined to the subventricular zone of the lateral ventricle and the subgranular zone of the hippocampal dentate gyrus, glial cells continue proliferating, migrating and differentiating throughout postnatal development and partially throughout life. Glial progenitors proliferate in the subventricular zone of the forebrain and migrate to various regions of the brain, where they differentiate to oligodendrocytes and astrocytes (Menn et al., 2006). Oligodendrocyte progenitor cells extend processes to neighboring axons and differentiate to oligodendrocytes and myelinate these axons, a process that extends over the first 2–3 decades of life in the frontal lobes of the cerebral cortex and that is crucial to development of higher cognitive function in humans. Long-term antibiotic treatment of adult mice is sufficient to induce decreased neurogenesis in the hippocampus of adult mice, and results in deficits in the novel object recognition task (Möhle et al., 2016). Voluntary exercise and probiotic treatment are sufficient to rescue these phenotypes (Möhle et al., 2016). Reduced numbers of CD45+CD11b+Ly-6chiCCR2+ monocytes, but not microglia, were observed in the brains of antibiotic-treated animals. Remarkably, CCR2−/− knock-out animals, as well as Ly-6chi-monocyte depleted animals, show reduced hippocampal neurogenesis. Lastly, adoptive transfer of Ly-6chi monocytes into antibiotic-treated animals was sufficient to rescue the neurogenesis phenotype in the hippocampus, indicating that circulating monocytes play an important role in adult neurogenesis (Möhle et al., 2016). These reports suggest that neurogenesis, apoptosis, and synaptic pruning may be regulated by signals from the microbiome. However, more research is needed to mechanistically study the role of the microbiome in these processes in both animal models and humans. Adult neurogenesis can also be promoted by serotonin (Alenina and Klempin, 2015), and gut bacteria have been shown to play a role in serotonergic pathways both in the gut and in various regions of the brain (reviewed by O’Mahoni et al., (2015). During postnatal brain development, astrocytes and microglia are thought to facilitate pruning of weak neuronal synapses by complement activation and subsequent phagocytosis (Hong and Stevens, 2016). In the absence of microglia, for example, the adult brain harbors significantly more synapses, exemplifying that these glial cells are necessary for synaptic pruning (Paolicelli et al., 2011; Zhan et al., 2014). While complement components are secreted by multiple cells in the CNS, astrocytes and microglia are major producers of complement (Bahrini et al., 2015; Stephan et al., 2012). Microglia produce copious amounts of C1q, the first protein in the complement activation cascade, and express various complement receptors (Schafer et al., 2012; Stephan et al., 2012). Synaptic pruning, as well as other processes, shape neuronal connections after cell differentiation and migration, refining neuronal networks following major events in postnatal brain development. Proper conductance in neuronal axons is essential for information and signal relay, and myelination is a critical process in the development of a healthy brain that continues well into adolescence (Davison and Dobbing, 1966). Two reports demonstrated that the presence of an intact gut microbiome modulates myelination. In these studies, myelin-related transcripts were increased in the prefrontal cortex, but not other brain regions, as a result of antibiotic-treatment (Gacias et al., 2016) or in GF mice (Hoban et al., 2016). Interestingly, while antibiotic treatment was sufficient to induce elevated expression of myelin-related genes in non-obese-diabetic (NOD) mice and could be transferred by microbiome transplantation from these mice to C57Bl/6 mice (Gacias et al., 2016), colonization of GF animals with SPF microbiota did not rescue the myelin phenotype (Hoban et al., 2016). These observations suggest that early-life exposure to the microbiome is necessary for dynamic response to changes in the microbiome later in life. In addition, Oligodendrocyte development and differentiation relies on various signals, among them are the chemokine CXCL1 and its receptor CXCR2. Notably, CXCL1 expression was shown to be differentially induced in the brains of an ischemic stroke mouse model, in mice with different microbiotas, suggesting a potential role during homeostasis and development, although further investigation is needed (Benakis et al., 2016). The early life gut microbiome and its impact on brain development and behavior Under the assumption that the fetus is sterile of bacteria, the first direct encounter an infant has with the microbial world is during birth. Infants born via vaginal delivery are colonized by microbial populations that are closely related to maternal vaginal populations, dominated by Lactobacillus and Prevotella species (Dominguez-Bello et al., 2010). In contrast, infants born by cesarean-section (C-section) are exposed to and colonized by skin microbes such as Staphylococcus and Corynebacterium (Bäckhed et al., 2015; Dominguez-Bello et al., 2010). This initial exposure to such distinct microbial populations has various implications over the health and development of a newborn, with long-term consequences. In addition, prenatal stress changes the vaginal microbiome, and has been shown to subsequently remodel the gut microbiome and metabolome of the offspring(Jašarević et al., 2015b). It has been documented that children born by C-section are at a higher risk for autoimmune diseases (Sevelsted et al., 2015). However, some restoration towards vaginal-derived microbial status can be achieved by exposing newborns to vaginal microorganisms derived from their mothers (Dominguez-Bello et al., 2016). The infant microbiome is highly sensitive to various perturbations like changes in diet and antibiotic treatment (Bäckhed et al., 2015; Koenig et al., 2011; Yassour et al., 2016). Moreover, infants are exposed to vertical acquisitions of novel microorganisms through intimate interaction with parents and siblings, as well as exposure to new environments (Bäckhed et al., 2015; Bordenstein and Theis, 2015; Rosenberg and Zilber-Rosenberg, 2016). Bäckhed and colleagues (2015) have identified signature taxa for various stages during the first year of life. While the newborns’ gut is predominantly aerobic and inhabited by Bifidobacterium, Enterococcus, Escherichia/Shigella, Streptococcus, Bacteroides, and Rothia, gut bacterial populations are significantly closer to those of mothers, and more anaerobic, by the age of one. At this stage, children are already colonized with Clostredium, Ruminococcus, Veilonella, Roseburia, Akkermansia, Alistipes, Eubacterium, Faecalibacterium and Prevotella, in addition to other bacteria. The gut microbiome during the first three years of life is more amenable and prone to perturbations. Interestingly, as some bacteria can be transmitted from mother to newborn, probiotic administration to mothers during pregnancy can transfer specific species to newborns (Dotterud et al., 2015). Hence, this period of life is critical for the establishment of a healthy, stable microbiome (Lloyd-Price et al., 2016a; Yatsunenko et al., 2012). Disruption of the microbiome, by administration of antibiotics or drastic changes in diet, or conversely, augmentation with probiotic microorganisms, has profound effects on the microbial community and its trajectory through life. Perturbations of the microbial community may have significant impacts on the developing individual, with long lasting effects on metabolism, physiology and immune status, as has been suggested in animal studies (Blaser, 2016; Kuperman and Koren, 2016; Zeissig and Blumberg, 2014). Others have reported that antibiotic administration during the first year of life was correlated with depression and behavioral difficulties later in life (Slykerman et al., 2016). The administration of antibiotics pre- or post-natally changes the physiological status of the mother or its offspring in animal models, and subsequently may impact the developmental trajectory of the offspring’s brain, or alternatively modulate behavior via primary effects on maternal behavior. Partial depletion of an animal’s associated microbiota for a short period of time may not always impact behavior; a short-term treatment of rat neonates with vancomycin had no effect on anxiety- and depressive-like behaviors in adulthood. However, during adulthood these rats showed visceral hypersensitivity, indicating that gut microbes can impact nociception (Amaral et al., 2008; O’Mahony et al., 2014). In adult mice, a seven-day course of nonabsorbable antibiotics was sufficient to decrease anxiety-like behavior. Interestingly, this effect was short-lasting, and behavior normalized to baseline within two weeks (Bercik et al., 2011a), as the microbiome likely returned to its initial state. Long-term broad-spectrum antibiotic treatment from weaning through adulthood restructured the microbiome and subsequently modulated brain chemistry and behavior (Desbonnet et al., 2015). Correspondingly, GF animals display various behavioral and developmental phenotypes, compared to SPF animals (Arentsen et al., 2015; Bercik et al., 2011a; Desbonnet et al., 2014; Gareau et al., 2011; Heijtz et al., 2011; Luczynski et al., 2016b; Neufeld et al., 2011) (Table 1, Fig. 1). These observations indicate a close connection between the microbiome and behavior, and suggest possible pathways through which they interact. Probiotic administration augments the population with a specific microbe, either transiently or permanently, and can change the microbiome profile and function, as well as interact with the host. Probiotic bacteria have been demonstrated to alter, reverse, or prevent various conditions in mouse models and humans. In addition, experiments with reconstitution of the gut microbial population with healthy consortia by fecal microbiota transplantation (FMT) are underway (Borody and Khoruts, 2012; Hourigan and Oliva-Hemker, 2016). Beneficial bacteria were found to reduce responses to stress and anxiety, depressive-like behavior, promote social behavior, decrease repetitive behavior, and improve cognitive function and communication in animals (Ait-Belgnaoui et al., 2012, 2014; Bercik et al., 2011b; Bravo et al., 2011; Buffington et al., 2016; Gareau et al., 2011; Hsiao et al., 2013; Sudo et al., 2004; Sun et al., 2016). This concept has been also expanded to humans where healthy volunteers that consumed a fermented milk product (containing several different probiotic bacteria) showed different brain activity during an emotional faces attention task, as measured by fMRI, in brain regions that control processing of sensation and emotion (Tillisch et al., 2013). Mounting evidence suggests that the communication between the brain and gut microbial populations is bi-directional (Bailey et al., 2011; Carabotti et al., 2015; Moussaoui et al., 2014; Park et al., 2013). Using the maternal separation model in mice, De Palma et al. (2015) demonstrated profound differences in the gut microbiome in response to early-life stress resulted in an anxiety-like phenotype. Moreover, gene expression in the amygdala differs between GF and SPF animals (Stilling et al., 2015). A reciprocal effect by gut bacteria has been reported as well, where specific bacteria, or complete microbial assemblages, had effects on host stress- and depression-like behaviors (Bercik et al., 2011a; Gacias et al., 2016; Sudo et al., 2004). While it is yet unclear if these examples are driven by a direct gut-brain interaction, or mediated by other physiological factors induced by the disease state, these reports and others exemplify potential interactions between the microbiome, the gastrointestinal tract, and the brain. The adult “steady-state” microbiome In adulthood, the microbiome reaches a relative equilibrium in terms of bacterial abundance and diversity, and does not change significantly under stable environmental or health conditions. Known determinants that shape the microbiome are genetics (Goodrich et al., 2016), diet (Carmody et al., 2015; David et al., 2014), lifestyle (Allen et al., 2015; Kang et al., 2014) and geography (Rampelli et al., 2015; Yatsunenko et al., 2012). Health is defined rather ambiguously as “..the absence of any overt disease” (Aagaard et al., 2013; Lloyd-Price et al., 2016a). The healthy human microbiome was recently reviewed extensively by Huttenhower and colleagues (2016a). The human microbiome is niche-specific, with microbial diversity and abundance differing significantly from niche to niche. Each of these niches is colonized by specific microbial assemblages, with the phyla Bacteroidetes and Firmicutes dominating the intestine, Streptococcus sp. dominating the oral cavity, Corynebacterium, Propionibacterium, and Staphylococcus dominating the skin, and Lactobacillus dominating the vagina (Lloyd-Price et al., 2016a). Higher microbial diversity is correlated with health and functional redundancy (Lozupone et al., 2012; Moya and Ferrer, 2016). Importantly, the healthy microbiome is temporally stable, even when subjected to recurrent mild disturbances (Dethlefsen et al., 2008; Oh et al., 2016; Schloissnig et al., 2013). Decreased diversity, or lack of redundancy, in the microbiome has been reported in multiple diseases (Lloyd-Price et al., 2016a). The microbiome in neurodevelopmental and mood disorders Neurodevelopmental disorders are classically studied from a genetic perspective (Parikshak et al., 2015; de la Torre-Ubieta et al., 2016). However, gastrointestinal comorbidities and food allergies are common in neurodevelopmental disorders, suggesting a role for the gut microbiome (de Theije et al., 2014). Thus, an appreciation for a microbial role in these conditions has been gained through profiling bacterial populations in fecal samples of patients and controls. Recent reports support the notion that the microbiome, or its disruption, can contribute to the pathology of various neurologic disorders, using mouse models and intervention studies. Evidence in rodent models suggests a direct link between the gut microbiota and stress and anxiety (reviewed by (Foster and McVey Neufeld, 2013). These observations in animal models is supported by data in human subjects that associates the gut microbiome in IBD to stress disorders (Bonaz and Bernstein, 2013; Fond et al., 2014). Autism spectrum disorder (ASD) The gut microbiome of ASD children has been studied in multiple different cohorts using various methodologies, and a consensus between studies was rarely reported. It appears as though these small studies fail to generate a coherent picture, although differences in species richness and their diversity between ASD and controls have been repeatedly reported (De Angelis et al., 2013; Finegold et al., 2002, 2010; Kang et al., 2013; Parracho et al., 2005; Son et al., 2015; Williams et al., 2011, 2012). The gut microbiome of ASD patients have increased abundance and diversity of Clostridia species, and a general increase in non-spore forming anaerobes and microaerophilic bacteria, compared to neurotypical controls (De Angelis et al., 2013; Finegold et al., 2002, 2010; Parracho et al., 2005). Gastrointestinal comorbidities are significantly more prevalent in children with ASD compared to controls (Mannion et al., 2013; McElhanon et al., 2014). These comorbidities often coincide with differences in the gut microbiome of these children (Son et al., 2015; Williams et al., 2011, 2012). Interestingly, Sutterella was found in close association with the intestinal epithelium of ASD children presenting gastrointestinal symptoms, while it was absent in controls (Williams et al., 2012). Interestingly, Kang et al. (2013), reported the absence of specific probiotic members of the community such as Prevotella from the ASD gut bacterial population, suggesting that augmenting the microbiome with a specific microbe may be beneficial. These studies, and others, indicate a potential relationship between the microbiome (or specific microbes) and the brain in autism. Large-scale cross-center studies, using standardized methodologies, would help delineate what are the significant differences in the gut microbiome of ASD patients. Subsequent functional studies into the properties of the gut microbiome in ASD will shed light on mechanism by which the gut microbiome contributes to pathology and behavior. Understanding interactions between specific microbial species and the brain during development may help unravel the etiology of this enigmatic disorder. Schizophrenia To date, few studies of the schizophrenia microbiome exist. Higher incidence of lactic-acid bacteria in the oropharyngeal microbiome was reported, compared to controls (Castro-Nallar et al., 2015). Interestingly, a Lactobacillus specific phage was also discovered in high abundance (Yolken et al., 2015). Another study reported a blood-specific microbiome in schizophrenia patients, compared to controls, with higher alpha and beta diversity (Mangul et al., 2016; Severance et al., 2013). Further research in large populations is needed to show whether these differences are consistent throughout the population, further profiles microbiomes associated with this condition, and test the potential roles for gut microbes in the etiology of schizophrenia. Depression (Major Depressive Disorder; MDD) Evidence in rodent models suggest that the gut microbiome plays a role in depressive-like behaviors (Bravo et al., 2011; Desbonnet et al., 2010; Ferreira Mello et al., 2013). Gastrointestinal symptoms are associated with depression, with approximately 20% of patients reporting such symptoms (Mussell et al., 2008). One hypothesis claims that depression, or subsets of this disorder, is a microglial disorder, as the onset of depression often follows intense inflammatory episodes in the brain, or conversely, decline in microglial function (Yirmiya et al., 2015). Interestingly, minocycline, a tetracycline antibacterial agent known to inhibit the activation of microglia, can diminish depressive behaviors in rodents (Molina-Hernández et al., 2008; Zheng et al., 2015b) and humans (Miyaoka et al., 2012), and has been suggested as a potent antidepressant. In light of recent evidence on the role of the microbiome in microglia maturation and activation (Erny et al., 2015; Matcovitch-Natan et al., 2016), it is compelling to speculate that the microbiome impacts depression by influencing microglial maturation and activation. It should be noted, however, that it is yet to be determined whether the antidepressive effects of minocycline are due to its antimicrobial properties, inhibition of microglial activation, or a combination of the two. Recently, Zheng et al. (2016) have shown that the beta-diversity of the gut microbiome in MDD patients is significantly different from that of healthy controls, with significantly more Actinobacteria and less Bacteroidetes in MDD associated microbial populations. Further, the authors transplanted human samples from MDD and control samples to GF mice, and show that recipients of MDD samples exhibit depressive-like phenotypes, compared to controls (Zheng et al., 2016). These findings were reproduced by another group, where decreased bacterial richness and diversity were reported in depression, and depressive-like phenotypes could be transmitted by fecal transplantation into rats (Kelly et al., 2016). Mouse and human studies now show that the microbiome plays an active role in driving depressive-like behaviors, suggesting potential new avenues for therapeutic development. Neurodegeneration and the microbiome in old age Throughout aging, mammals undergo physiological changes that increase susceptibility to disease. Interestingly, the incidence of some gastrointestinal diseases increase with age (Britton and McLaughlin, 2013), and prevalence of diagnosed GI disorders is approximately 24% in people over 65 (Alameel et al., 2012). Notably, the enteric nervous system degenerates with age, starting in adulthood. Cholinergic nerves, as well as enteric glia cells, are lost in both the myenteric plexus and the submucosal plexus (Phillips and Powley, 2007). This degeneration is partially responsible for the increase in bowel motility symptoms prevalent in the elderly (O’Mahony et al., 2002). The microbiome also undergos profound remodeling in elderly populations (over 65 years old)(Biagi et al., 2011; O’Toole and Claesson, 2010; Salazar et al., 2014). Major shifts in bacterial taxa have been reported in the fecal microbiome of elderly people, compared to infants and young adults, and these shifts correlate with health status and frailty in elderly people (Claesson et al., 2011; van Tongeren et al., 2005). Notably, while in adulthood Firmicutes outnumber Bacteroidetes in the gut, it appears as though the ratio shifts in favor of Bacteroidetes in the elderly (Claesson et al., 2011; Mariat et al., 2009). Specifically, Bacteroides, Alistipes, Parabacteroides, and Proteobacteria (gamma-proteobacteria, specifically) are significantly more prevalent in the elderly compared to younger adults (Claesson et al., 2011; Mariat et al., 2009). Moreover, various studies show that the microbiome of elderly living in the community have a similar microbial diversity to that of younger adults than that of elderly people staying in short or long term care facilities (Claesson et al., 2012). Langlie et al. (2014) reported that similar differences occur in older mice, compared to a younger group, and is highly correlated with mouse frailty. A clinical study, testing the effects of Lactobacillus rhamnosus GG consumption in an elderly cohort, found that the probiotic restructured the bacterial population towards an anti-inflammatory phenotype that favors the survival and growth of beneficial microbes and elevated SCFA production (Eloe-Fadrosh et al., 2015). The microbiome in the elderly population is significantly different from that of younger adults, is less diverse and resilient, and can be modulated by environmental factors and interventions. Remarkably, the gut microbiome of centenarians differs significantly from that of other adults (Biagi et al., 2010, 2016). Biagi and colleagues found that a core gut microbiome, comprised of species in the Bacteroidaceae, Lachnospiraceae, and Ruminococcaceae families, is associated with the human host throughout life, with decreasing cumulative abundance. Specifically, Coprococcus, Roseburia, and Faecalibacterium, were found to be negatively correlated with age (Biagi et al., 2016). Other genera are positively correlated with age; among these are Oscillospira and Akkermansia (Biagi et al., 2016). Interestingly, certain subdominant bacterial species are enriched in centenarians, some of which are known to exert beneficial functions on their host, and may play a role in maintenance of health in old age (Biagi et al., 2016). Insights into the microbiome of centenarians may unravel specific microbes with beneficial and protective effect on the host. The microbiome in neurodegenerative diseases How initial interactions with gut microbes alter events later in life, such as during neurodegenerative diseases, is still unclear. A small number of studies to date have demonstrated different gut microbial populations in both human and animal models of neurodegenerative disease. Preliminary observations in the APP/PS1 mouse model of Alzheimer’s disease (AD) indicate that these animals harbor decreased Allobaculum and Akkermansia, with an increase in Rikenellaceae compared to wild-type controls (Harach et al., 2015). Concurrently, individuals afflicted with Parkinson’s disease (PD) display significantly different fecal and mucosal microbial populations (Hasegawa et al., 2015; Keshavarzian et al., 2015; Scheperjans et al., 2015). Prevotellaceae show decreased, while Lactobacilliaceae have increased, abundance compared to controls (Hasegawa et al., 2015; Scheperjans et al., 2015). Relative levels of Enterobacteraceae in feces is sufficient to discriminate between specific forms of PD; patients displaying the tremor-dominant form of PD had significantly lower Enterobacteraceae relative abundance than those with the more severe postural and gait instability (Scheperjans et al., 2015). In fact, intestinal biopsies of PD patients have indicated increased tissue-associated E. coli compared to healthy controls (Forsyth et al., 2011), further demonstrating the presence of an altered gut microbial community in individuals diagnosed with neurodegenerative diseases. However, this research correlating changes in the microbiome and neurodegeneration remain largely descriptive; how different microbial populations arise, and their physiological consequences, if any, remain unknown. Neuroinflammation is postulated to play a key role in the pathology of neurodegenerative diseases (Cappellano et al., 2013; Glass et al., 2010). Proinflammatory cytokines produced both in the brain and periphery modulate neuronal function and can initiate pathologic cell death (Koprich et al., 2008; McCoy and Tansey, 2008). Given the importance of microglia functions in both the prevention and promotion of neurodegenerative processes, it is tempting to speculate that the gut microbiota may influence these inflammatory diseases of the aging brain. As discussed above, bacterial fermentation products, namely SCFAs, can drive the maturation of microglia and are needed for maintenance of mature microglia (Erny et al., 2015). Interestingly, decreased SCFA concentrations in feces from PD patients, compared to controls, were recently reported (Unger et al., 2016). Passage of MAMPs from the intestine and into the brain may produce low levels of inflammation (Pal et al., 2015). Such persistent, proinflammatory signaling has been linked to the severity of neurodegenerative disease (Cappellano et al., 2013; Glass et al., 2010). Interestingly, a recent report suggests that a physiologic role of the amyloid protein Aβ, known to form pathogenic plaques in AD, is as an antimicrobial agent to eliminate bacterial infection in the brain (Kumar et al., 2016). Studies to understand the long-term consequences of neurophysiological changes by the microbiota are critical. While dysfunction in BBB integrity and maturation of microglia could have global effects, a recent study has revealed a specific example of neurodegenerative disease mediated by gut microbes. In a mouse model of sporadic uveitis, an inflammation of the middle layer of the eye, animals exhibit increased inflammation and loss of cells in the neuroretina leading to vision dysfunction. This inflammation is mediated by autoreactive T cells that recognize antigens present in the retina, a typically immune-privileged tissue. The activation of autoreactive T cells in these mice is dependent on the microbiota (Horai et al., 2015). In fact, not only are these T cells reactive towards retinal proteins, they are also capable of recognizing microbial antigens present in the gut (Horai et al., 2015). At this time, however, the microbial antigen driving this autoimmune neuroinflammation is unknown. Nonetheless, this observation suggests that “molecular mimicry” of host molecules present in the microbiota can trigger autoimmune responses that promote neurodegeneration. Other microbial molecules that mimic host structures have been suggested to play roles in promoting immune responses during AD and PD (Friedland, 2015; Hill and Lukiw, 2015). Therefore, one might consider the early developmental presence of certain gut microbes that produce these molecular mimics to potentially act as risk factors for specific immune and neurodegenerative diseases. However, no direct observation of this hypothesis has been reported to date, and the link between age-dependent influences to the microbiome and neuropathology are an active area of research. Perspective The microbiome plays a significant role in the well-being of its host. While much of the research on this topic to date has demonstrated that different bacterial populations are associated with certain clinical conditions, it is unclear for the most part whether these differences are causative, promote and/or enhance disease, or instead are a consequence of otherwise unrelated pathophysiology. Future research should tackle this challenging question in order to understand the intricate interaction between mammals (or any other host) and their associated microbial community. We must not continue exercises in simply cataloging bacterial populations. Rather, we must extend this foundational research approach to test the functional and ecological roles that a given microbial population plays, as well as decipher the physiological effects individual bacteria or consortia of bacteria have on their animal hosts. It is of importance to address questions of cause and effect: are changes in the microbiome underlying the pathophysiology or are they a result thereof? Are the effects on behavior direct, or a result of other fundamental physiological changes? Are there defined microbial features that are necessary and sufficient to support proper neurodevelopment and prevent neurodegeneration? The use of animal models is a great tool for studying basic processes in health and disease. However, we must use caution in extrapolating results to the human condition, and strive to use preclinical findings as one of several approaches to inform human health and disease. While research on the gut-brain axis is still in relative infancy, certain basic rules have begun to emerge. It appears as though specific neurological pathways evolved to respond to the effect of microbial population, while others are unaffected by microbiome “instruction” and subject to purely genomic or other environmental cues. Interaction with host-associated microbial communities, either directly via microbial metabolites or indirectly by the immune, metabolic or endocrine systems, can supply the nervous system with real-time information about the environment. These cues converge to control basic developmental processes in the brain such as barrier function, immune surveillance, and neurogenesis. The mechanistic understanding of how different microbial populations, beneficial or pathogenic, govern these and other functions related to health and disease holds promise in the diagnosis, treatment, and prevention of specific neuropathologies. Determining how a microbiome, changing with Westernization and other environmental factors, impacts a human population with growing rates of neurodevelopmental disorders and increasing life expectancy represents an urgent challenge to biomedical research, and to society. The authors apologize to colleagues whose work could not be included in this Review. We thank Drs. Hiutung Chu and Wei-li Wu, as well as Carly Stewart for critical reading of this manuscript. The authors are supported by the Meixner Postdoctoral Fellowship in Translational Research (to G.S.) and the Larry L. Hillblom Foundation Postdoctoral Fellowship (to T.R.S). Research in the Mazmanian laboratory is funded by grants from the National Institutes of Health (MH100556, DK078938, GM099535 and NS085910), the Department of Defense, the Heritage Medical Research Institute, and the Simons Foundation. Figure 1 Intersections of gut microorganisms and basic developmental processes Basic developmental processes driven directly or indirectly by gut microbes and their products. (A) Gut microorganisms relay messages to the brain via various direct and indirect mechanisms. (B) Basic neurodevelopmental processes are modulated as a result of colonization of GF animals or depletion of gut bacteria by antibiotics. Specifically, the following processes are modulates: blood-brain barrier (BBB) formation and integrity (Braniste et al., 2014), neurogenesis (Möhle et al., 2016; Ogbonnaya et al., 2015), microglia maturation and ramification (Erny et al., 2015; Matcovitch-Natan et al., 2016), myelination (Gacias et al., 2016; Hoban et al., 2016), and expression of neurotrophins (Bercik et al., 2011a, 2011b; Desbonnet et al., 2015), neurotransmitters (Bercik et al., 2011a; O’Mahony et al., 2015), and their respective receptors. Figure 2 Major events in mammalian brain development Developmental trajectories and key neurodevelopmental events in mice and humans (adapted from(Knuesel et al., 2014; Pressler and Auvin, 2013; Semple et al., 2013). E-embryonic age, P-postnatal age, GSW-gestational week. Bacterial taxa on the right panel are the dominant ones at each life stage (Bäckhed et al., 2015; Lloyd-Price et al., 2016b; Nuriel-Ohayon et al., 2016). Table 1 Perturbation of the microbiome and microbial products can affect behavioral outcomes in mouse models and humans Treatment / pertubation to microbiome Effects on behavior Known/Persumed Mechanism (by correlation) Reference Prenatal effects Abx Offspring exhibited anxiety-like behavior and hypoactivity, as well as decreased sociability Dysbiotic microbiome Tochitani et al., 2016; Degroote et al., 2016 High-fat diet Social deficit and repetitive behavior in offsprings Dysbiotic microbiome; deficient VTA synaptic plasticity and oxytocin Buffington et al., 2016 Peptidoglycan Offspring show decreased cognitive function TLR-2 mediated neuroprolifiration via FoxG1 induction in fetal cortex Humann et al., 2016 Prenatal stress Change in offspring microbiome, gut and brain metabolome Dysbiotic microbiome; altered free amino-acid levels in offspring brain Jašarević et al., 2015b Maternal immune activation by Poly(I:C) administration Change in offspring microbiome and metabolome, increase repetitive behavior, anxiety-like behavior, social deficit, communication deficit IL-6 and IL-17A mediated behavioral and cortical development abnormalities; Dysbiotic microbiome. Smith et al., 2007; Hsiao et al., 2013; Choi et al., 2016 Propionic acid Offspring exhibited anxiety-like behavior - Foley et al., 2014 Postnatal effects Perinatal Abx Visceral hypersensitivity Dysbiotic microbiome; decreased expression of various genes involved in pain preception in the lumbosacral region of the spine O’Mahony et al., 2014 Abx (short-term) Short-term anxiolytic effect Dysbiotic microbiome; increased BDNF levels in hippocampus, and decreased in the amygdala; phenotype independent of sympathetic and parasympathetic pathways Bercik et al., 2011a Abx (Long-term) Anxiolytic effect, cognitive deficits Dysbiotic microbiome; increased tryptophan and decreased kynurenine in serum; increased noradrenaline in hippocampus and increased L-DOPA in amygdala; decreased BDNF in the hippocampus and vasopressin expression in the hypothalamus Desbonnet et al., 2015 Deficits in memory formation Decreased BDNF and c-Fos expression in the CA1 region of the hippocampus Gareau et al., 2011 Colonization of GF Swiss Webster with BALB/c microbiome increased exploratory behavior, while the reciprocal colonization of GF BALB/c mice with Swiss Webster microbiome reduced exploration GF Swiss Webster mice comlonize with Swiss Webster microbiota have higher levels of BDNF in the hippocampus, but not in the amygdala, compared to GF mice colonized with microbiota from BALB/c mice Bercik et al., 2011a Germ-Free Animals Increased motor activity and reduced anxiety-like behavior compared to SPF GF mice, compared to SPF controls, show: increased turnover of noradrenaline, dopamine, and serotonin in the striatum ; decreased expression of NGFI-A in the frontal cortex, BDNF in the basolateral amygdala and CA1 region of hippocampus, and dopamine D1 receptor in the dendate gyrus; differences in gene expression in hippocampus, frontal cortex, and striatum; higher expression of synaptophysin and PSD-95 (increased synaptogenesis) Heijtz et al., 2011 Reduced anxiety-like behavior GF mice had decreased expression of NR2B in the amygdala; increased expression of BDNF and decreased expression of 5HT1A in the dentate gyrus Neufeld et al., 2011 Increased stress-induced HPA response GF animals, compared to SPF controls, show: increased levels of stress-induced acetylcholine and corticosterone (HPA-axis); decreased expression of NR-1 in the cortex, and of NR-2a in the cortex and hippocampus; BDNF levels were lower in the cortex and hippocampus Sudo et al., 2004 Reduced social behavior - Desbonnet et al., 2014 Increased social behavior Decreased expression of specific BDNF transcripts in the amygdala Arentsen et al., 2015 Clostridium butyricum administration restored cognitive function in mouse model for vascular dementia Restoration correlated with increased levels of the SCFA butyrate in feces and brains, and accompanied by activation of the BDNF-PI3K/Akt pathway in the hippocampus Liu et al., 2015 Clostridium butyricum protected from cerebral ischemia/reperfusion injury in diabetic mice Anti-apoptotic effects via Akt activation; restoration of bacterial diversity Sun et al., 2016 Lactobacillus farciminis prevented stress-induced intestinal permeability and neuroinflammation in mice Decreased intestinal permeability that mediated the following stress-induced phenotypes: increased CRF expression in PVN; increased expression of proinflammatory cytokines in blood and increased plasma levels of corticosterone Ait-Belgnaoui et al., 2012 Lactobacillus helveticus R0052 and Bifidobacterium longum R0175 (Probio’Stick®) protected from effects of water avoidence stress in mice and decreased intestinal barrier dysfunction Probiotic treatment attentuated plasma levels of HPA/ANS-related corticosterone, adrenaline, and noradrenaline, as a result of stress induction, as well as activation of nuclei in the PVN, amygdala, CA3 of the hippocampus, and the dentate gyrus. Additionally, The expression of various genes related to Ait-Belgnaoui et al., 2014 Probiotic Administration Lactobacillus rhamnosus (JB-1) reduced stress induced anxiety- and depression-like behaviors in mice Probiotic decreased levels of stress-induced plasma corticosterone; changed expression levels of GABAB1b, GABAAa1, and GABAAa2 thorughout the brain in a vagus-dependent manner Bravo et al., 2011 Lactobacillus rhamnosus (R0011) and Lactobacillus helveticus (R0052) (Lacidofil) reduced anxiety-like behaviors in mice Increased BDNF and c-Fos expression in the hippocampus Gareau et al., 2011 Lactobacillus reuteri corrected social deficit in mice Probiotic restored levels of oxytocin producing cells in the PVN and synaptic plasticity in the VTA Buffington et al., 2016 Monoassociation with Bifidobacterium infantis reversed increased stress-induced HPA response in GF mice Treatment of GF mice with probiotic or colonizing with SPF microbiome early in development normalized levels of stress-induced acetylcholine and corticosterone; normalized expression of NR1 in the cortex, and of NR2a in the cortex and hippocampus; normalized BDNF levels in the cortex and hippocampus Sudo et al., 2004 Bifidobacterium longum NCC3001 reduced colitis-induced anxiety-like behaviors in mice Probiotic decreased anxiety, but not pathology, in a vagus-dependant manner; excitability of enteric neurons incubated in probiotic-fermented media was lower than controls Bercik et al., 2011b Fermented milk product affected brain activity in regions processing emotion and sensation in healthy human subjects Probiotic reduced activity in response to an emotional faces attention task in the insula cortex and somatosensory cortex Tillisch et al., 2013 Bacteroides fragilis corrected anxiety-like and repetitive behaviors in mice Partial restoration of microbial community; restoration of intestinal barrier function Hsiao et al., 2013 Abx - Antibiotics; GF - Germ-free ; SPF - Specific pathogen free; VTA - ventral tegmental area; BDNF - brain-derived neurotropic ;HPAfactor; NR2B - N-methyl-D-aspartate receptor subunit 2B; 5HT1A - serotonine receptor 1A; SCFA - short-chain fatty-acid; HPA - hypothalamus-pituitary-adrenal axis; ANS - autonomous nervous system; PVN - paraventricular nucleus of the hypothalamus; This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. Aagaard K Petrosino J Keitel W Watson M Katancik J Garcia N Patel S Cutting M Madden T Hamilton H 2013 The Human Microbiome Project strategy for comprehensive sampling of the human microbiome and why it matters FASEB J 27 1012 1022 23165986 Aagaard K Ma J Antony KM Ganu R Petrosino J Versalovic J 2014 The placenta harbors a unique microbiome Sci Transl Med 6 237ra65 Ait-Belgnaoui A Durand H Cartier C Chaumaz G Eutamene H Ferrier L Houdeau E Fioramonti J Bueno L Theodorou V 2012 Prevention of gut leakiness by a probiotic treatment leads to attenuated HPA response to an acute psychological stress in rats Psychoneuroendocrinology 37 1885 1895 22541937 Ait-Belgnaoui A Colom A Braniste V Ramalho L Marrot A Cartier C Houdeau E Theodorou V Tompkins T 2014 Probiotic gut effect prevents the chronic psychological stress-induced brain activity abnormality in mice Neurogastroenterol Motil 26 510 520 24372793 Alenina N Klempin F 2015 The role of serotonin in adult hippocampal neurogenesis Behav Brain Res 277 49 57 25125239 Allen JM Berg Miller ME Pence BD Whitlock K Nehra V Gaskins HR White BA Fryer JD Woods JA 2015 Voluntary and forced exercise differentially alters the gut microbiome in C57BL/6J mice J Appl Physiol 118 1059 1066 25678701 Amaral FA Sachs D Costa VV Fagundes CT Cisalpino D Cunha TM Ferreira SH Cunha FQ Silva TA Nicoli JR 2008 Commensal microbiota is fundamental for the development of inflammatory pain Proc Natl Acad Sci U S A 105 2193 2197 18268332 Arentsen T Raith H Qian Y Forssberg H Diaz Heijtz R 2015 Host microbiota modulates development of social preference in mice Microb Ecol Health Dis 26 29719 26679775 Aspelund A Antila S Proulx ST Karlsen TV Karaman S Detmar M Wiig H Alitalo K 2015 A dural lymphatic vascular system that drains brain interstitial fluid and macromolecules J Exp Med 212 991 999 26077718 Bäckhed F Roswall J Peng Y Feng Q Jia H Kovatcheva-Datchary P Li Y Xia Y Xie H Zhong H 2015 Dynamics and Stabilization of the Human Gut Microbiome during the First Year of Life Cell Host Microbe 17 690 703 25974306 Bahrini I Song J-H Diez D Hanayama R 2015 Neuronal exosomes facilitate synaptic pruning by up-regulating complement factors in microglia Sci Rep 5 7989 25612542 Bailey MT Dowd SE Galley JD Hufnagle AR Allen RG Lyte M 2011 Exposure to a social stressor alters the structure of the intestinal microbiota: implications for stressor-induced immunomodulation Brain Behav Immun 25 397 407 21040780 Benakis C Brea D Caballero S Faraco G Moore J Murphy M Sita G Racchumi G Ling L Pamer EG 2016 Commensal microbiota affects ischemic stroke outcome by regulating intestinal [gamma][delta] T cells Nat Med 22 516 523 27019327 Bennett ML Bennett FC Liddelow SA Ajami B Zamanian JL Fernhoff NB Mulinyawe SB Bohlen CJ Adil A Tucker A 2016 New tools for studying microglia in the mouse and human CNS Proc Natl Acad Sci U S A 113 E1738 E1746 26884166 Bercik P Denou E Collins J Jackson W Lu J Jury J Deng Y Blennerhassett P Macri J McCoy KD 2011a The intestinal microbiota affect central levels of brain-derived neurotropic factor and behavior in mice Gastroenterology 141 599 609 609.e1 e3 21683077 Bercik P Park AJ Sinclair D Khoshdel A Lu J Huang X Deng Y Blennerhassett PA Fahnestock M Moine D 2011b The anxiolytic effect of Bifidobacterium longum NCC3001 involves vagal pathways for gut–brain communication Neurogastroenterology & Motility 23 1132 1139 21988661 Biagi E Nylund L Candela M Ostan R Bucci L Pini E Nikkïla J Monti D Satokari R Franceschi C 2010 Through ageing, and beyond: gut microbiota and inflammatory status in seniors and centenarians PLoS One 5 e10667 20498852 Biagi E Franceschi C Rampelli S Severgnini M Ostan R Turroni S Consolandi C Quercia S Scurti M Monti D 2016 Gut Microbiota and Extreme Longevity Curr Biol 26 1480 1485 27185560 Blaser MJ 2016 Antibiotic use and its consequences for the normal microbiome Science 352 544 545 27126037 Bonaz BL Bernstein CN 2013 Brain-gut interactions in inflammatory bowel disease Gastroenterology 144 36 49 23063970 Bordenstein SR Theis KR 2015 Host Biology in Light of the Microbiome: Ten Principles of Holobionts and Hologenomes PLoS Biol 13 e1002226 26284777 Borody TJ Khoruts A 2012 Fecal microbiota transplantation and emerging applications Nat Rev Gastroenterol Hepatol 9 88 96 Braniste V Al-Asmakh M Kowal C Anuar F Abbaspour A Tóth M Korecka A Bakocevic N Ng LG Guan NL 2014 The gut microbiota influences blood-brain barrier permeability in mice Sci Transl Med 6 263ra158 Bravo JA Forsythe P Chew MV Escaravage E Savignac HM Dinan TG Bienenstock J Cryan JF 2011 Ingestion of Lactobacillus strain regulates emotional behavior and central GABA receptor expression in a mouse via the vagus nerve Proc Natl Acad Sci U S A 108 16050 16055 21876150 Britton E McLaughlin JT 2013 Ageing and the gut Proc Nutr Soc 72 173 177 23146206 Buffington SA Di Prisco GV Auchtung TA Ajami NJ Petrosino JF Costa-Mattioli M 2016 Microbial Reconstitution Reverses Maternal Diet-Induced Social and Synaptic Deficits in Offspring Cell 165 1762 1775 27315483 Cappellano G Carecchio M Fleetwood T Magistrelli L Cantello R Dianzani U Comi C 2013 Immunity and inflammation in neurodegenerative diseases Am J Neurodegener Dis 2 89 107 23844334 Carabotti M Scirocco A Maselli MA Severi C 2015 The gut-brain axis: interactions between enteric microbiota, central and enteric nervous systems Ann Gastroenterol Hepatol 28 203 209 Carmody RN Gerber GK Luevano JM Jr Gatti DM Somes L Svenson KL Turnbaugh PJ 2015 Diet dominates host genotype in shaping the murine gut microbiota Cell Host Microbe 17 72 84 25532804 Castro-Nallar E Bendall ML Pérez-Losada M Sabuncyan S Severance EG Dickerson FB Schroeder JR Yolken RH Crandall KA 2015 Composition, taxonomy and functional diversity of the oropharynx microbiome in individuals with schizophrenia and controls PeerJ 3 e1140 26336637 Ceni C Unsain N Zeinieh MP Barker PA 2014 Neurotrophins in the Regulation of Cellular Survival and Death Neurotrophic Factors Lewin GR Carter BD Springer Berlin Heidelberg 193 221 Chang C-Y Ke D-S Chen J-Y 2009 Essential fatty acids and human brain Acta Neurol Taiwan 18 231 241 20329590 Choi GB Yim YS Wong H Kim S Kim H Kim SV Hoeffer CA Littman DR Huh JR 2016 The maternal interleukin-17a pathway in mice promotes autism-like phenotypes in offspring Science 351 933 939 26822608 Claesson MJ Cusack S O’Sullivan O Greene-Diniz R de Weerd H Flannery E Marchesi JR Falush D Dinan T Fitzgerald G 2011 Composition, variability, and temporal stability of the intestinal microbiota of the elderly Proc Natl Acad Sci U S A 108 Suppl 1 4586 4591 20571116 Clarke G Grenham S Scully P Fitzgerald P Moloney RD Shanahan F Dinan TG Cryan JF 2013 The microbiome-gut-brain axis during early life regulates the hippocampal serotonergic system in a sex-dependent manner Mol Psychiatry 18 666 673 22688187 David LA Maurice CF Carmody RN Gootenberg DB Button JE Wolfe BE Ling AV Devlin AS Varma Y Fischbach MA 2014 Diet rapidly and reproducibly alters the human gut microbiome Nature 505 559 563 24336217 Davison AN Dobbing J 1966 Myelination as a vulnerable period in brain development Br Med Bull 22 40 44 5321813 De Angelis M Piccolo M Vannini L Siragusa S De Giacomo A Serrazzanetti DI Cristofori F Guerzoni ME Gobbetti M Francavilla R 2013 Fecal microbiota and metabolome of children with autism and pervasive developmental disorder not otherwise specified PLoS One 8 e76993 24130822 Degroote S Hunting D Baccarelli AA Takser L 2016 Maternal gut and fetal brain connection: Increased anxiety and reduced social interactions in Wistar rat offspring following peri-conceptional antibiotic exposure Prog Neuropsychopharmacol Biol Psychiatry De Palma G Blennerhassett P Lu J Deng Y Park AJ Green W Denou E Silva MA Santacruz A Sanz Y 2015 Microbiota and host determinants of behavioural phenotype in maternally separated mice Nat Commun 6 7735 26218677 Desbonnet L Garrett L Clarke G Kiely B Cryan JF Dinan TG 2010 Effects of the probiotic Bifidobacterium infantis in the maternal separation model of depression Neuroscience 170 1179 1188 20696216 Desbonnet L Clarke G Shanahan F Dinan TG Cryan JF 2014 Microbiota is essential for social development in the mouse Mol Psychiatry 19 146 148 23689536 Desbonnet L Clarke G Traplin A O’Sullivan O Crispie F Moloney RD Cotter PD Dinan TG Cryan JF 2015 Gut microbiota depletion from early adolescence in mice: Implications for brain and behaviour Brain Behav Immun 48 165 173 25866195 Dethlefsen L Huse S Sogin ML Relman DA 2008 The pervasive effects of an antibiotic on the human gut microbiota, as revealed by deep 16S rRNA sequencing PLoS Biol 6 e280 19018661 DiGiulio DB 2012 Diversity of microbes in amniotic fluid Semin Fetal Neonatal Med 17 2 11 22137615 DiGiulio DB Callahan BJ McMurdie PJ Costello EK Lyell DJ Robaczewska A Sun CL Goltsman DSA Wong RJ Shaw G 2015 Temporal and spatial variation of the human microbiota during pregnancy Proc Natl Acad Sci U S A 112 11060 11065 26283357 Dominguez-Bello MG Costello EK Contreras M Magris M Hidalgo G Fierer N Knight R 2010 Delivery mode shapes the acquisition and structure of the initial microbiota across multiple body habitats in newborns Proc Natl Acad Sci U S A 107 11971 11975 20566857 Dominguez-Bello MG De Jesus-Laboy KM Shen N Cox LM Amir A Gonzalez A Bokulich NA Song SJ Hoashi M Rivera-Vinas JI 2016 Partial restoration of the microbiota of cesarean-born infants via vaginal microbial transfer Nat Med 22 250 253 26828196 Dotterud CK Avershina E Sekelja M Simpson MR Rudi K Storrø O Johnsen R Øien T 2015 Does Maternal Perinatal Probiotic Supplementation Alter the Intestinal Microbiota of Mother and Child? J Pediatr Gastroenterol Nutr 61 200 207 25782657 Erny D Hrabě de Angelis AL Jaitin D Wieghofer P Staszewski O David E Keren-Shaul H Mahlakoiv T Jakobshagen K Buch T 2015 Host microbiota constantly control maturation and function of microglia in the CNS Nat Neurosci 18 965 977 26030851 Estes ML McAllister AK 2016 Maternal immune activation: Implications for neuropsychiatric disorders Science 353 772 777 27540164 Ferreira Mello BS Monte AS McIntyre RS Soczynska JK Custódio CS Cordeiro RC Chaves JH Mendes Vasconcelos SM Nobre HV Júnior Florenço de Sousa FC 2013 Effects of doxycycline on depressive-like behavior in mice after lipopolysaccharide (LPS) administration J Psychiatr Res 47 1521 1529 23835040 Finegold SM Molitoris D Song Y Liu C Vaisanen M-L Bolte E McTeague M Sandler R Wexler H Marlowe EM 2002 Gastrointestinal microflora studies in late-onset autism Clin Infect Dis 35 S6 S16 12173102 Finegold SM Dowd SE Gontcharova V Liu C Henley KE Wolcott RD Youn E Summanen PH Granpeesheh D Dixon D 2010 Pyrosequencing study of fecal microflora of autistic and control children Anaerobe 16 444 453 20603222 Foley KA Ossenkopp KP Kavaliers M Macfabe DF 2014 Pre- and neonatal exposure to lipopolysaccharide or the enteric metabolite, propionic acid, alters development and behavior in adolescent rats in a sexually dimorphic manner PLoS One 9 e87072 24466331 Fond G Loundou A Hamdani N Boukouaci W Dargel A Oliveira J Roger M Tamouza R Leboyer M Boyer L 2014 Anxiety and depression comorbidities in irritable bowel syndrome (IBS): a systematic review and meta-analysis Eur Arch Psychiatry Clin Neurosci 264 651 660 24705634 Forsyth CB Shannon KM Kordower JH Voigt RM Shaikh M Jaglin JA Estes JD Dodiya HB Keshavarzian A 2011 Increased intestinal permeability correlates with sigmoid mucosa alpha-synuclein staining and endotoxin exposure markers in early Parkinson’s disease PLoS One 6 e28032 22145021 Foster JA McVey Neufeld K-A 2013 Gut–brain axis: how the microbiome influences anxiety and depression Trends Neurosci 36 305 312 23384445 Friedland RP 2015 Mechanisms of molecular mimicry involving the microbiota in neurodegeneration J Alzheimers Dis 45 349 362 25589730 Gacias M Gaspari S Santos P-MG Tamburini S Andrade M Zhang F Shen N Tolstikov V Kiebish MA Dupree JL 2016 Microbiota-driven transcriptional changes in prefrontal cortex override genetic differences in social behavior Elife 5 Gardener H Spiegelman D Buka SL 2011 Perinatal and neonatal risk factors for autism: a comprehensive meta-analysis Pediatrics 128 344 355 21746727 Gareau MG Wine E Rodrigues DM Cho JH Whary MT Philpott DJ Macqueen G Sherman PM 2011 Bacterial infection causes stress-induced memory dysfunction in mice Gut 60 307 317 20966022 Geschwind DH Rakic P 2013 Cortical evolution: judge the brain by its cover Neuron 80 633 647 24183016 Ginhoux F Greter M Leboeuf M Nandi S See P Gokhan S Mehler MF Conway SJ Ng LG Stanley ER 2010 Fate mapping analysis reveals that adult microglia derive from primitive macrophages Science 330 841 845 20966214 Glass CK Saijo K Winner B Marchetto MC Gage FH 2010 Mechanisms underlying inflammation in neurodegeneration Cell 140 918 934 20303880 Gomez de Agüero M Ganal-Vonarburg SC Fuhrer T Rupp S Uchimura Y Li H Steinert A Heikenwalder M Hapfelmeier S Sauer U 2016 The maternal microbiota drives early postnatal innate immune development Science 351 1296 1302 26989247 Goodrich JK Davenport ER Waters JL Clark AG Ley RE 2016 Cross-species comparisons of host genetic associations with the microbiome Science 352 532 535 27126034 Harach T Marungruang N Dutilleul N Cheatham V Mc Coy KD Neher JJ Jucker M Fåk F Lasser T 2015 Reduction of Alzheimer’s disease beta-amyloid pathology in the absence of gut microbiota Hasegawa S Goto S Tsuji H Okuno T Asahara T Nomoto K Shibata A Fujisawa Y Minato T Okamoto A 2015 Intestinal Dysbiosis and Lowered Serum Lipopolysaccharide-Binding Protein in Parkinson’s Disease PLoS One 10 e0142164 26539989 Heijtz RD Wang S Anuar F Qian Y Björkholm B Samuelsson A Hibberd ML Forssberg H Pettersson S 2011 Normal gut microbiota modulates brain development and behavior Proceedings of the National Academy of Sciences 108 3047 3052 Hill JM Lukiw WJ 2015 Microbial-generated amyloids and Alzheimer’s disease (AD) Front Aging Neurosci 7 9 25713531 Hoban AE Stilling RM Ryan FJ Shanahan F Dinan TG Claesson MJ Clarke G Cryan JF 2016 Regulation of prefrontal cortex myelination by the microbiota Transl Psychiatry 6 e774 27045844 Hong S Stevens B 2016 Microglia: Phagocytosing to Clear, Sculpt, and Eliminate Dev Cell 38 126 128 27459063 Horai R Zárate-Bladés CR Dillenburg-Pilla P Chen J Kielczewski JL Silver PB Jittayasothorn Y Chan CC Yamane H Honda K 2015 Microbiota-Dependent Activation of an Autoreactive T Cell Receptor Provokes Autoimmunity in an Immunologically Privileged Site Immunity 43 343 353 26287682 Hourigan SK Oliva-Hemker M 2016 Fecal microbiota transplantation in children: a brief review Pediatr Res 80 2 6 26982451 Hsiao EY McBride SW Hsien S Sharon G Hyde ER McCue T Codelli JA Chow J Reisman SE Petrosino JF 2013 Microbiota modulate behavioral and physiological abnormalities associated with neurodevelopmental disorders Cell 155 1451 1463 24315484 Humann J Mann B Gao G Moresco P Ramahi J Loh LN Farr A Hu Y Durick-Eder K Fillon SA 2016 Bacterial Peptidoglycan Traverses the Placenta to Induce Fetal Neuroproliferation and Aberrant Postnatal Behavior Cell Host Microbe 19 388 399 26962947 Ichim G Tauszig-Delamasure S Mehlen P 2012 Neurotrophins and cell death Exp Cell Res 318 1221 1228 22465479 Ivanov II Atarashi K Manel N Brodie EL Shima T Karaoz U Wei D Goldfarb KC Santee CA Lynch SV 2009 Induction of intestinal Th17 cells by segmented filamentous bacteria Cell 139 485 498 19836068 Jašarević E Rodgers AB Bale TL 2015a A novel role for maternal stress and microbial transmission in early life programming and neurodevelopment Neurobiol Stress 1 81 88 25530984 Jašarević E Howerton CL Howard CD Bale TL 2015b Alterations in the Vaginal Microbiome by Maternal Stress Are Associated With Metabolic Reprogramming of the Offspring Gut and Brain Endocrinology 156 3265 3276 26079804 Kang DW Park JG Ilhan ZE Wallstrom G Labaer J Adams JB Krajmalnik-Brown R 2013 Reduced incidence of Prevotella and other fermenters in intestinal microflora of autistic children PLoS One 8 e68322 23844187 Kang SS Jeraldo PR Kurti A Miller MEB Cook MD Whitlock K Goldenfeld N Woods JA White BA Chia N 2014 Diet and exercise orthogonally alter the gut microbiome and reveal independent associations with anxiety and cognition Mol Neurodegener 9 36 25217888 Kelly JR Borre Y O’Brien C Patterson E el Aidy S Deane J Kennedy PJ Beers S Scott K Moloney G 2016 Transferring the blues: Depression-associated gut microbiota induces neurobehavioural changes in the rat J Psychiatr Res Keshavarzian A Green SJ Engen PA Voigt RM Naqib A Forsyth CB Mutlu E Shannon KM 2015 Colonic bacterial composition in Parkinson’s disease Mov Disord 30 1351 1360 26179554 Khandaker GM Zimbron J Lewis G Jones PB 2013 Prenatal maternal infection, neurodevelopment and adult schizophrenia: a systematic review of population-based studies Psychol Med 43 239 257 22717193 Khosravi A Yáñez A Price JG Chow A Merad M Goodridge HS Mazmanian SK 2014 Gut microbiota promote hematopoiesis to control bacterial infection Cell Host Microbe 15 374 381 24629343 Knuesel I Chicha L Britschgi M Schobel SA Bodmer M Hellings JA Toovey S Prinssen EP 2014 Maternal immune activation and abnormal brain development across CNS disorders Nat Rev Neurol 10 643 660 25311587 Koenig JE Spor A Scalfone N Fricker AD Stombaugh J Knight R Angenent LT Ley RE 2011 Succession of microbial consortia in the developing infant gut microbiome Proc Natl Acad Sci U S A 108 Suppl 1 4578 4585 20668239 Koprich JB Reske-Nielsen C Mithal P Isacson O 2008 Neuroinflammation mediated by IL-1$\beta$ increases susceptibility of dopamine neurons to degeneration in an animal model of Parkinson’s disease J Neuroinflammation 5 1 12 18171484 Koren O Goodrich JK Cullender TC Spor A Laitinen K Bäckhed HK Gonzalez A Werner JJ Angenent LT Knight R 2012 Host remodeling of the gut microbiome and metabolic changes during pregnancy Cell 150 470 480 22863002 Krajmalnik-Brown R Lozupone C Kang D-W Adams JB 2015 Gut bacteria in children with autism spectrum disorders: challenges and promise of studying how a complex community influences a complex disease Microb Ecol Health Dis 26 26914 25769266 Krakowiak P Walker CK Bremer AA Baker AS Ozonoff S Hansen RL Hertz-Picciotto I 2012 Maternal metabolic conditions and risk for autism and other neurodevelopmental disorders Pediatrics 129 e1121 e1128 22492772 Kriegstein A Alvarez-Buylla A 2009 The glial nature of embryonic and adult neural stem cells Annu Rev Neurosci 32 149 184 19555289 Kumar DKV Choi SH Washicosky KJ Eimer WA Tucker S Ghofrani J Lefkowitz A McColl G Goldstein LE Tanzi RE 2016 Amyloid-β peptide protects against microbial infection in mouse and worm models of Alzheimer’s disease Sci Transl Med 8 340ra72 Kuperman AA Koren O 2016 Antibiotic use during pregnancy: how bad is it? BMC Med 14 91 27312712 Lauder AP Roche AM Sherrill-Mix S Bailey A Laughlin AL Bittinger K Leite R Elovitz MA Parry S Bushman FD 2016 Comparison of placenta samples with contamination controls does not provide evidence for a distinct placenta microbiota Microbiome 4 29 27338728 Lloyd-Price J Abu-Ali G Huttenhower C 2016a The healthy human microbiome Genome Med 8 51 27122046 Lloyd-Price J Abu-Ali G Huttenhower C 2016b The healthy human microbiome Genome Med 8 51 27122046 Louveau A Smirnov I Keyes TJ Eccles JD Rouhani SJ Peske JD Derecki NC Castle D Mandell JW Lee KS 2015 Structural and functional features of central nervous system lymphatic vessels Nature 523 337 341 26030524 Lozupone CA Stombaugh JI Gordon JI Jansson JK Knight R 2012 Diversity, stability and resilience of the human gut microbiota Nature 489 220 230 22972295 Luczynski P Whelan SO O’Sullivan C Clarke G Shanahan F Dinan TG Cryan JF 2016a Adult microbiota-deficient mice have distinct dendritic morphological changes: differential effects in the amygdala and hippocampus Eur J Neurosci Luczynski P McVey Neufeld K-A Oriach CS Clarke G Dinan TG Cryan JF 2016b Growing up in a Bubble: Using Germ-Free Animals to Assess the Influence of the Gut Microbiota on Brain and Behavior Int J Neuropsychopharmacol MacIntyre DA Chandiramani M Lee YS Kindinger L Smith A Angelopoulos N Lehne B Arulkumaran S Brown R Teoh TG 2015 The vaginal microbiome during pregnancy and the postpartum period in a European population Sci Rep 5 8988 25758319 Mangul S Olde Loohuis LM Ori A Jospin G Koslicki D Yang HT Wu T Boks MP Lomen-Hoerth C Wiedau-Pazos M 2016 Total RNA Sequencing reveals microbial communities in human blood and disease specific effects Mannion A Leader G Healy O 2013 An investigation of comorbid psychological disorders, sleep problems, gastrointestinal symptoms and epilepsy in children and adolescents with Autism Spectrum Disorder Res Autism Spectr Disord 7 35 42 Mariat D Firmesse O Levenez F Guimarăes V Sokol H Doré J Corthier G Furet J-P 2009 The Firmicutes/Bacteroidetes ratio of the human microbiota changes with age BMC Microbiol 9 123 19508720 Marín O Rubenstein JLR 2003 Cell migration in the forebrain Annu Rev Neurosci 26 441 483 12626695 Matcovitch-Natan O Winter DR Giladi A Vargas Aguilar S Spinrad A Sarrazin S Ben-Yehuda H David E Zelada González F Perrin P 2016 Microglia development follows a stepwise program to regulate brain homeostasis Science McCoy MK Tansey MG 2008 TNF signaling inhibition in the CNS: implications for normal brain function and neurodegenerative disease J Neuroinflammation 5 45 18925972 McElhanon BO McCracken C Karpen S Sharp WG 2014 Gastrointestinal symptoms in autism spectrum disorder: a meta-analysis Pediatrics 133 872 883 24777214 Menn B Garcia-Verdugo JM Yaschine C Gonzalez-Perez O Rowitch D Alvarez-Buylla A 2006 Origin of oligodendrocytes in the subventricular zone of the adult brain J Neurosci 26 7907 7918 16870736 Miyaoka T Wake R Furuya M Liaury K Ieda M Kawakami K Tsuchie K Taki M Ishihara K Araki T 2012 Minocycline as adjunctive therapy for patients with unipolar psychotic depression: an open-label study Prog Neuropsychopharmacol Biol Psychiatry 37 222 226 22349578 Möhle L Mattei D Heimesaat MM Bereswill S Fischer A Alutis M French T Hambardzumyan D Matzinger P Dunay IR 2016 Ly6Chi Monocytes Provide a Link between Antibiotic-Induced Changes in Gut Microbiota and Adult Hippocampal Neurogenesis Cell Rep 15 1945 1956 27210745 Molina-Hernández M Tellez-Alcántara NP Pérez-García J Olivera-Lopez JI Jaramillo-Jaimes MT 2008 Antidepressant-like actions of minocycline combined with several glutamate antagonists Prog Neuropsychopharmacol Biol Psychiatry 32 380 386 17933448 Møllgård K Saunders NR 1986 The development of the human blood-brain and blood-CSF barriers Neuropathol Appl Neurobiol Moussaoui N Braniste V Ait-Belgnaoui A Gabanou M Sekkal S Olier M Théodorou V Martin PGP Houdeau E 2014 Changes in intestinal glucocorticoid sensitivity in early life shape the risk of epithelial barrier defect in maternal-deprived rats PLoS One 9 e88382 24586321 Moya A Ferrer M 2016 Functional Redundancy-Induced Stability of Gut Microbiota Subjected to Disturbance Trends Microbiol 24 402 413 26996765 Mussell M Kroenke K Spitzer RL Williams JBW Herzog W Löwe B 2008 Gastrointestinal symptoms in primary care: prevalence and association with depression and anxiety J Psychosom Res 64 605 612 18501261 Nayak D Roth TL McGavern DB 2014 Microglia development and function Annu Rev Immunol 32 367 402 24471431 Neufeld KM Kang N Bienenstock J Foster JA 2011 Reduced anxiety-like behavior and central neurochemical change in germ-free mice Neurogastroenterol Motil 23 255 264 e119 21054680 Nuriel-Ohayon M Neuman H Koren O 2016 Microbial changes during pregnancy, birth and infancy Front Microbiol 7 26858696 Ogbonnaya ES Clarke G Shanahan F Dinan TG Cryan JF O’Leary OF 2015 Adult Hippocampal Neurogenesis Is Regulated by the Microbiome Biol Psychiatry 78 e7 e9 25700599 Oh J Byrd AL Park M Kong HH Segre JA NISC Comparative Sequencing Program 2016 Temporal Stability of the Human Skin Microbiome Cell 165 854 866 27153496 O’Mahony D O’Leary P Quigley EMM 2002 Aging and intestinal motility: a review of factors that affect intestinal motility in the aged Drugs Aging 19 515 527 12182688 O’Mahony SM Felice VD Nally K Savignac HM Claesson MJ Scully P Woznicki J Hyland NP Shanahan F Quigley EM 2014 Disturbance of the gut microbiota in early-life selectively affects visceral pain in adulthood without impacting cognitive or anxiety-related behaviors in male rats Neuroscience 277 885 901 25088912 O’Mahony SM Clarke G Borre YE Dinan TG Cryan JF 2015 Serotonin, tryptophan metabolism and the brain-gut-microbiome axis Behav Brain Res 277 32 48 25078296 Oppenheim RW 1989 The neurotrophic theory and naturally occurring motoneuron death Trends Neurosci 12 252 255 2475935 Pal GD Shaikh M Forsyth CB Ouyang B Keshavarzian A Shannon KM 2015 Abnormal lipopolysaccharide binding protein as marker of gastrointestinal inflammation in Parkinson disease Front Neurosci 9 306 26388718 Paolicelli RC Bolasco G Pagani F Maggi L Scianni M Panzanelli P Giustetto M Ferreira TA Guiducci E Dumas L 2011 Synaptic pruning by microglia is necessary for normal brain development Science 333 1456 1458 21778362 Parikshak NN Gandal MJ Geschwind DH 2015 Systems biology and gene networks in neurodevelopmental and neurodegenerative disorders Nat Rev Genet 16 441 458 26149713 Park H Poo M-M 2013 Neurotrophin regulation of neural circuit development and function Nat Rev Neurosci 14 7 23 23254191 Park AJ Collins J Blennerhassett PA Ghia JE Verdu EF Bercik P Collins SM 2013 Altered colonic function and microbiota profile in a mouse model of chronic depression Neurogastroenterol Motil 25 733 e575 23773726 Parracho HMRT Bingham MO Gibson GR McCartney AL 2005 Differences between the gut microflora of children with autistic spectrum disorders and that of healthy children J Med Microbiol 54 987 991 16157555 Pärtty A Kalliomäki M Wacklin P Salminen S Isolauri E 2015 A possible link between early probiotic intervention and the risk of neuropsychiatric disorders later in childhood: a randomized trial Pediatr Res 77 823 828 25760553 Phillips RJ Powley TL 2007 Innervation of the gastrointestinal tract: patterns of aging Auton Neurosci 136 1 19 17537681 Pressler R Auvin S 2013 Comparison of Brain Maturation among Species: An Example in Translational Research Suggesting the Possible Use of Bumetanide in Newborn Front Neurol 4 36 23596438 Rakic P 1988 Specification of cerebral cortical areas Science 241 170 176 3291116 Rampelli S Schnorr SL Consolandi C Turroni S Severgnini M Peano C Brigidi P Crittenden AN Henry AG Candela M 2015 Metagenome Sequencing of the Hadza Hunter-Gatherer Gut Microbiota Curr Biol 25 1682 1693 25981789 Romero R Hassan SS Gajer P Tarca AL Fadrosh DW Nikita L Galuppi M Lamont RF Chaemsaithong P Miranda J 2014a The composition and stability of the vaginal microbiota of normal pregnant women is different from that of non-pregnant women Microbiome 2 4 24484853 Romero R Hassan SS Gajer P Tarca AL Fadrosh DW Bieda J Chaemsaithong P Miranda J Chaiworapongsa T Ravel J 2014b The vaginal microbiota of pregnant women who subsequently have spontaneous preterm labor and delivery and those with a normal delivery at term Microbiome 2 18 24987521 Rosenberg E Zilber-Rosenberg I 2016 Microbes Drive Evolution of Animals and Plants: the Hologenome Concept MBio 7 Round JL Mazmanian SK 2009 The gut microbiota shapes intestinal immune responses during health and disease Nat Rev Immunol 9 313 323 19343057 Russell SL Gold MJ Willing BP Thorson L McNagny KM Finlay BB 2013 Perinatal antibiotic treatment affects murine microbiota, immune responses and allergic asthma Gut Microbes 4 158 164 23333861 Schafer DP Lehrman EK Kautzman AG Koyama R Mardinly AR Yamasaki R Ransohoff RM Greenberg ME Barres BA Stevens B 2012 Microglia sculpt postnatal neural circuits in an activity and complement-dependent manner Neuron 74 691 705 22632727 Scheperjans F Aho V Pereira PAB Koskinen K Paulin L Pekkonen E Haapaniemi E Kaakkola S Eerola-Rautio J Pohja M 2015 Gut microbiota are related to Parkinson’s disease and clinical phenotype Mov Disord 30 350 358 25476529 Schloissnig S Arumugam M Sunagawa S Mitreva M Tap J Zhu A Waller A Mende DR Kultima JR Martin J 2013 Genomic variation landscape of the human gut microbiome Nature 493 45 50 23222524 Semple BD Blomgren K Gimlin K Ferriero DM Noble-Haeusslein LJ 2013 Brain development in rodents and humans: Identifying benchmarks of maturation and vulnerability to injury across species Prog Neurobiol 106–107 1 16 Sevelsted A Stokholm J Bønnelykke K Bisgaard H 2015 Cesarean section and chronic immune disorders Pediatrics 135 e92 e98 25452656 Severance EG Gressitt KL Stallings CR Origoni AE Khushalani S Leweke FM Dickerson FB Yolken RH 2013 Discordant patterns of bacterial translocation markers and implications for innate immune imbalances in schizophrenia Schizophr Res 148 130 137 23746484 Severance EG Yolken RH Eaton WW 2014 Autoimmune diseases, gastrointestinal disorders and the microbiome in schizophrenia: more than a gut feeling Schizophr Res Slykerman RF Thompson J Waldie KE Murphy R Wall C Mitchell EA 2016 Antibiotics in the first year of life and subsequent neurocognitive outcomes Acta Paediatr Son JS Zheng LJ Rowehl LM Tian X Zhang Y Zhu W Litcher-Kelly L Gadow KD Gathungu G Robertson CE 2015 Comparison of Fecal Microbiota in Children with Autism Spectrum Disorders and Neurotypical Siblings in the Simons Simplex Collection PLoS One 10 e0137725 26427004 Stensballe LG Simonsen J Jensen SM Bønnelykke K Bisgaard H 2013 Use of antibiotics during pregnancy increases the risk of asthma in early childhood J Pediatr 162 832 838.e3 23140881 Stephan AH Barres BA Stevens B 2012 The complement system: an unexpected role in synaptic pruning during development and disease Annu Rev Neurosci 35 369 389 22715882 Stiles J Jernigan TL 2010 The basics of brain development Neuropsychol Rev 20 327 348 21042938 Stilling RM Ryan FJ Hoban AE Shanahan F Clarke G Claesson MJ Dinan TG Cryan JF 2015 Microbes & neurodevelopment--Absence of microbiota during early life increases activity-related transcriptional pathways in the amygdala Brain Behav Immun 50 209 220 26184083 Sudo N Chida Y Aiba Y Sonoda J Oyama N Yu X-N Kubo C Koga Y 2004 Postnatal microbial colonization programs the hypothalamic–pituitary–adrenal system for stress response in mice J Physiol 558 263 275 15133062 Sun J Wang F Ling Z Yu X Chen W Li H Jin J Pang M Zhang H Yu J 2016 Clostridium butyricum attenuates cerebral ischemia/reperfusion injury in diabetic mice via modulation of gut microbiota Brain Res 1642 180 188 27037183 de Theije CGM Bavelaar BM Lopes da Silva S Korte SM Olivier B Garssen J Kraneveld AD 2014 Food allergy and food-based therapies in neurodevelopmental disorders Pediatr Allergy Immunol 25 218 226 24236934 Tillisch K Labus J Kilpatrick L Jiang Z Stains J Ebrat B Guyonnet D Legrain-Raspaud S Trotin B Naliboff B 2013 Consumption of fermented milk product with probiotic modulates brain activity Gastroenterology 144 1394 1401 1401.e1 e4 23474283 Tochitani S Ikeno T Ito T Sakurai A Yamauchi T Matsuzaki H 2016 Administration of Non-Absorbable Antibiotics to Pregnant Mice to Perturb the Maternal Gut Microbiota Is Associated with Alterations in Offspring Behavior PLoS One 11 e0138293 26789865 van Tongeren SP Slaets JPJ Harmsen HJM Welling GW 2005 Fecal microbiota composition and frailty Appl Environ Microbiol 71 6438 6442 16204576 de la Torre-Ubieta L Won H Stein JL Geschwind DH 2016 Advancing the understanding of autism disease mechanisms through genetics Nat Med 22 345 361 27050589 Unger MM Spiegel J Dillmann K-U Grundmann D Philippeit H Bürmann J Faßbender K Schwiertz A Schäfer K-H 2016 Short chain fatty acids and gut microbiota differ between patients with Parkinson’s disease and age-matched controls Parkinsonism Relat Disord Varvel NH Grathwohl SA Baumann F Liebig C Bosch A Brawek B Thal DR Charo IF Heppner FL Aguzzi A 2012 Microglial repopulation model reveals a robust homeostatic process for replacing CNS myeloid cells Proc Natl Acad Sci U S A 109 18150 18155 23071306 Williams BL Hornig M Buie T Bauman ML Cho Paik M Wick I Bennett A Jabado O Hirschberg DL Lipkin WI 2011 Impaired carbohydrate digestion and transport and mucosal dysbiosis in the intestines of children with autism and gastrointestinal disturbances PLoS One 6 e24585 21949732 Williams BL Hornig M Parekh T Lipkin WI 2012 Application of novel PCR-based methods for detection, quantitation, and phylogenetic characterization of Sutterella species in intestinal biopsy samples from children with autism and gastrointestinal disturbances MBio 3 Yassour M Vatanen T Siljander H Hämäläinen A-M Härkönen T Ryhänen SJ Franzosa EA Vlamakis H Huttenhower C Gevers D 2016 Natural history of the infant gut microbiome and impact of antibiotic treatment on bacterial strain diversity and stability Sci Transl Med 8 343ra81 Yatsunenko T Rey FE Manary MJ Trehan I Dominguez-Bello MG Contreras M Magris M Hidalgo G Baldassano RN Anokhin AP 2012 Human gut microbiome viewed across age and geography Nature 486 222 227 22699611 Yirmiya R Rimmerman N Reshef R 2015 Depression as a microglial disease Trends Neurosci 38 637 658 26442697 Yolken RH Severance EG Sabunciyan S Gressitt KL Chen O Stallings C Origoni A Katsafanas E Schweinfurth LAB Savage CLG 2015 Metagenomic Sequencing Indicates That the Oropharyngeal Phageome of Individuals With Schizophrenia Differs From That of Controls Schizophr Bull 41 1153 1161 25666826 Zeisel SH 2004 Nutritional importance of choline for brain development J Am Coll Nutr 23 621S 626S 15640516 Zeissig S Blumberg RS 2014 Life at the beginning: perturbation of the microbiota by antibiotics in early life and its role in health and disease Nat Immunol 15 307 310 24646587 Zhan Y Paolicelli RC Sforazzini F Weinhard L Bolasco G Pagani F Vyssotski AL Bifone A Gozzi A Ragozzino D 2014 Deficient neuron-microglia signaling results in impaired functional brain connectivity and social behavior Nat Neurosci 17 400 406 24487234 Zheng J Xiao X Zhang Q Mao L Yu M Xu J 2015a The Placental Microbiome Varies in Association with Low Birth Weight in Full-Term Neonates Nutrients 7 6924 6937 26287241 Zheng L-S Kaneko N Sawamoto K 2015b Minocycline treatment ameliorates interferon-alpha-induced neurogenic defects and depression-like behaviors in mice Front Cell Neurosci 9 5 25674053 Zheng P Zeng B Zhou C Liu M Fang Z Xu X Zeng L Chen J Fan S Du X 2016 Gut microbiome remodeling induces depressive-like behaviors through a pathway mediated by the host’s metabolism Mol Psychiatry 21 786 796 27067014 Zuchero JB Barres BA 2015 Glia in mammalian development and disease Development 142 3805 3809 26577203

PMC005xxxxxx/PMC5127450.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 8708534 734 Arch Psychiatr Nurs Arch Psychiatr Nurs Archives of psychiatric nursing 0883-9417 1532-8228 27888976 5127450 10.1016/j.apnu.2016.07.004 NIHMS805252 Article The Variability of Nursing Attitudes Toward Mental Illness: An Integrative Review de Jacq Krystyna NP Norful Allison Andreno MSN, RN, ANP-BC Larson Elaine PhD, RN Columbia University School of Nursing, 617 West 168 Street, New York, NY 10032, USA Corresponding author: Krystyna de Jacq, kd2221@columbia.edu 29 7 2016 12 7 2016 12 2016 01 12 2017 30 6 788796 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. Mental illnesses are common worldwide, and nurses’ attitudes toward mental illness have an impact on the care they deliver. This integrative literature review focused on nurses’ attitudes toward mental illness. Four databases were searched between January 1, 1995 to October 31, 2015 selecting studies, which met the following inclusion criteria: 1) English language; and 2) Research in which the measured outcome was nurses’ attitudes toward mental illness. Fifteen studies conducted across 20 countries that 4,282 participants met the inclusion criteria. No study was conducted in the United States (U.S.). Studies reported that nurses had mixed attitudes toward mental illness, which were comparable to those of the general public. More negative attitudes were directed toward persons with schizophrenia. Results indicate the need for further research to determine whether attitudes among nurses in the U.S. differ from those reported from other countries and to examine potential gaps in nursing curriculum regarding mental illness. Mental illnesses are common worldwide and represent the fifth leading disorder globally (Whiteford et al., 2013). About 450 million people suffer from mental illnesses worldwide (World Health Organization, 2001). In the United States (U.S.) alone, over 43.7 million of adults, 18.6% of all the population, have a mental illness diagnosis (National Alliance for the Mentally Ill, 2013). Effective treatments exist, but only 39% of people with diagnosed mental illness receive treatment and among those who receive treatment, one in five terminate treatment prematurely (NIMH, 2001, Olfson et al., 2009). Various factors play a role in decision-making as it pertains to seeking help for mental illness. Those factors include financial concerns, poor self-perception, limited access and stigma (Mojtabai et al., 2011). Goffman (1963) defines social stigma as an attribute that is discredited by society. Hatzenbuehler, Phelan, & Link, 2013, suggested in a recent review that stigma related to mental illness causes health inequalities by preventing people from seeking help that they need. People with depression are more likely to suffer from physical health comorbidities and are reported to be twice as likely as non-depressed patients to have two or more physical illnesses (Smith et al., 2014). According to the Anxiety and Depression Association of America (ADAA), anxiety disorders cost the U.S. more than $42 billion per year, representing almost a third of total mental health spending (ADAA, 2010). People who suffer from anxiety disorders are three to five times more likely to visit primary care and gastroenterology than people without the disorder, resulting in increased health care costs (Hoffman, Dukes, & Wittchen, 2008). Delaying treatment for mental illness may result in negative consequences. The longer the duration of untreated illness, the worse the outcomes in psychosis, mood disorders and anxiety disorders (Dell’osso, Glick, Baldwin, & Altamura, 2012). Furthermore, after initiation of treatment, non-adherence and drop out rates may result in unfavorable outcomes (Barrett et al., 2008). A negative patient-provider relationship, or personal and professional characteristics of the providers, may compel the patient to leave treatment (Reneses, Munoz, & Lopez-Ibor, 2009). Hoge et al., (2014) performed a study at a U.S. Veterans Administration Hospital and reported that dissatisfaction with the provider was one of the reasons for patients to drop out of treatment. Furthermore, in a recent integrative review, Newman, D., O’Reilly, P., Lee, S. H., & Kennedy, C. (2015) underlined the importance of relationships between the providers, such as nurses, and the patients who were seeking help for mental health problems. In addition to the patient-provider relationship, the impact of provider stigma is emerging in the literature, and has been identified as the strongest barrier toward help seeking behavior of individuals with mental illness (Clement et al., 2015, Corrigan, 2004; Evans-Lacko, Brohan, Mojtabai, & Thornicroft, 2012; Hinshaw & Stier, 2008; Kim, Britt, Klocko, Riviere, & Adler, 2011). Newman et al., (2015) re-iterated the importance of stigma, affirming that negative nursing attitudes toward mental illness have a profound impact on the delivery of care. Similarly, McDonald et al. (2003) confirm that the nurses’ care of patients is negatively impacted if the patient has a mental illness. The investigators presented vignettes that represented three patients admitted to the emergency room with a possible myocardial infarction. 1) The patient was taking an antipsychotic medication; 2) The patient was taking alprazolam (Xanax), a medication used to treat anxiety disorder; and 3) The patient had no history of psychiatric treatment (control). A significant difference in symptom recognition was found. Only 31% of nurses who read the first vignette identified a possibility of myocardial infarction in a patient taking antipsychotic medications compared to 51% of nurses in the control group. Additionally, when patients were experiencing increased anxiety, 78.9% of nurses in the control group stated that they could be having a heart attack versus 45.5% only in the psychotic patient group. This study highlights a general tendency of nurses to stereotype patients with mental illness thereby responding differently to them (McDonald et al., 2003). Corrigan et al., (2014) found that providers’ attitudes were different toward patients with a diagnosis of mental illness than toward those without. Although the factors that influence attitudes regarding mental illness have been studied for many years (Ajzen, 2005; Ajzen & Fishbein, 1980; Fishbein, 2010; Fishbein & Ajzen, 1975; Fishbein, Ajzen, Albarracin, & Hornik, 2007), to our knowledge, there has been no integrative literature review exploring nursing attitudes toward patients with mental illness. Obtaining a clear understanding of nursing attitudes may, inform policy and be used to implement change to ensure optimal patient care. Aim The aim of this integrative review is to explore nurses’ attitudes toward patients with mental illness. Methodology Defining Mental Illness The Centers for Disease Control and Prevention (2013) defines mental illness as “disorders generally characterized by dysregulation of mood, thought, and/or behavior, as recognized by the Diagnostic and Statistical Manual, 4th edition, of the American Psychiatric Association.” People with mental illness have impaired thinking, and their feelings may affect their ability to function on a daily basis. For the purpose of this review, we used the terms mental illness, mental disorders, and psychological problems interchangeably, which included, but not limited to, mood and psychotic disorders, as well as anxiety. Given the change in mental illness criteria introduced by DSM IV in 1994, only studies that used DSM IV and DSM V were included (American Psychiatric Association, 1994, 2013). Literature Search The conduct of this integrative review was guided by the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) statement (Liberati, Altman, Tetzlaff, & Mulrow, 2009). We searched the following databases: Ovid MEDLINE, PsycINFO, CINAHL, and PubMed in September, 2015. The following Medical Subject Heading (MeSH) terms were searched: (‘mental illness’ OR ‘mental health’) AND (‘nurses’ OR ‘nurs*’) AND (‘stereotyp*’ OR ‘stigma’ OR ‘prejudice’ OR ‘discrimination’ OR ‘attitudes” OR ‘beliefs’). Data were initially extracted from the four databases by the first author who screened all articles’ titles and abstracts. Two authors independently assessed selected full text articles for eligibility, and the discrepancies were resolved by discussion. The inclusion criteria were studies published between January 1, 1995 and October 31, 2015 in English and included nurses as participants in which the measured outcome was nursing attitudes toward mental health and/or illness in patients. Personal accounts, editorials, and/or single case studies, studies not written in English, and studies that explored attitudes of other professionals were excluded. Quality Appraisal The methodological quality of the studies was assessed using Quality Assessment Tool for Observational Cohort and Cross-Sectional Studies (QATOCCS) from the National Institute of Health, National Heart, Lung, and Blood Institute. The QATOCCS was modified to fit the needs of cross-sectional studies, as many questions were relevant to cohort studies only. Two researchers appraised the quality of the studies and 100% consensus of each study’s quality was achieved. Studies were rated in tertiles: low quality (0 – 33%), moderate quality (34 – 66%), and high quality (67 – 100%). Results The initial database search yielded 2,615 articles, and 2,343 remained after duplicates were removed. Following title screening, 770 papers were identified as potentially eligible and 701 articles were excluded after title and abstract review, leaving 69 articles for full text screening. Fourteen articles met the inclusion criteria. A search of the reference lists of the 14 final articles yielded an additional five articles eligible for inclusion in the study. A full text review by two researchers was performed again and one of the five articles was included in the final review yielding 15 studies that met initial eligibility criteria. Quality Appraisal Two researchers reached consensus on the quality of each study. Twelve studies were determined to be of high quality (Arvaniti et al., 2009; Chambers et al., 2010; Foster et al., 2008; Hamdan-Mansour & Wardam, 2009; Hsiao et al., 2015; Linden & Kavanagh, 2012; Magliano et al., 2004; Munro & Baker, 2007; Nordt, 2006; Scheerder et al., 2011; Serafini et al., 2011; Sevigny et al., 1999). One study score within the moderate quality range (Kukulu & Ergun, 2007). Two studies received lower quality scores because some key methodological elements were not reported, including sampling, sample recruitment and size, and lack of information about study measures. One of these lacked sufficient methodological rigor to be included, leaving 14 studies remaining in the final synthesis of the review. A PRISMA flow diagram is presented in Figure 1. The studies were conducted across 20 countries. None of the studies were performed in the U.S. Eight of the studies were conducted in Europe (two of which included more than one country), four in Asia, and three in the Middle East. Of the 14 studies, six focused on attitudes toward schizophrenia and/or depression, while the remaining nine concentrated on mental illness in general. All of the studies had a cross-sectional design. Twelve studies included mental health nurses who worked with mentally ill inpatients or outpatients. Aydin, Yigit, Inandi, & Kirpinar, (2003) conducted a study in an outpatient, non-psychiatric setting. Arvaniti et al., (2009) and Scheerder et al., (2010) performed their studies on medical rather than psychiatric units. Study characteristics and key findings are presented in Table 1. Study Measures Numerous measures were utilized across countries. Three studies used the Attitudes Toward Acute Mental Illness Scale (ATAMH33) (Baker, Richards, & Campbell, 2005) and four studies used Community Attitudes towards Mental Illness (CAMI) (Taylor, Dear, & Hall, 1979; Taylor & Dear, 1981). The following measures were in at least one study: The Level of Contact Report (Holmes, Corrigan, Williams, & Canar, 1999), the Opinion about Mental Illness (Madianos, Madianou, Vlachonikolis, & Stefanis, 1987), the Authoritarianism Scale (Adorno, 1950), Social Distance (Arkar, 1991), Burden of Illness (Eker & Arkar, 1991), Jefferson Scale of Empathy-Health Profession version (Hojat, Gonnella, Nasca, Mangione, & et al., 2002), Attitudes of Mental Illness Questionnaire (Luty, Fekadu, Umoh, & Gallagher, 2006), Social Interaction Scale (Kelly, St Lawrence, Smith, & Hood, 1987), Social Acceptance Scale (Angermeyer & Matschinger, 1997), and Standardized Stigma Questionnaire (Haghighat, 2005). Kukulu and Ergun, (2007) utilized an adaptation of multiple instruments, but the researchers were unable to assess its validity because the instruments’ descriptions and psychometric testing were only being available in studies published in the Turkish language. Findings In these studies, attitudes toward mental illness were compared between psychiatric nurses and nurses working in non-psychiatric settings as well as between nurses and the general public. Finally as discussed below, four common themes emerged: 1) etiology of mental illness; 2) social restrictiveness and distance; 3) perceived dangerousness; 4) attitudes specific to schizophrenia and depression. Attitudes of psychiatric nurses compared to nurses working in other settings Nursing attitudes were examined first by comparing nurses that were working on psychiatric wards compared to non-psychiatric nurses working on a medical ward or outpatient clinics. However, no study compared directly psychiatric versus non-psychiatric nurses. Authors of three studies reported the attitudes of non-psychiatric nurses (Arvantini et al., 2009; Aydin et al., 2003; Scheerder et al., 2010). Arvantini et al., (2009) reported both positive and negative nursing attitudes toward mental illness. For example, 60.7% of nurses in this study agreed that mentally ill patients should be separated from patients without mental illness. On the contrary, 76% of psychiatric and non-psychiatric nurses in this study viewed mentally ill patients as not being dangerous. Aydin et al., (2003) reported that nurses endorsed social discrimination more than the doctors and showed low support for social integration. They also endorsed social restriction more than other professionals, such as doctors and medical students. However, nurses endorsed social care questions at a higher level than other groups. Negative nursing attitudes toward patients with schizophrenia and depression were also reported. This finding was consistent among studies that examined both psychiatric and non-psychiatric nurses. Scheerder et al. (2010) found that non-psychiatric nurses held mostly positive attitudes toward people with depression. Sixty percent of nurses considered depression as an illness and 81.9% of respondents agreed (n=1533) that depression was treatable. However, nurses’ attitudes were less positive compared to other mental health professionals, such as clinical social workers, psychologists, and counselors, which can be explained by lack of specialty training among nurses as compared to professionals in mental health. There was a variability of psychiatric nurses attitudes across studies. Three studies reported positive attitudes (Chambers et al., 2010; Linden & Kavanagh, 2012; Munro & Baker, 2007) and four studies exemplified negative attitudes (Hamdan-Mansour & Wardam, 2009; Hsiao et al., 2015; Magliano et al., 2004; Sevigny et al., 1999). The remaining studies were a combination of both positive and negative (Foster et al., 2008; Kukulu & Ergun, 2007; Nordt et al., 2006; Serafini et al., 2011). In a large European study, Chambers et al. (2010) assessed attitudes of 810 mental health nurses and reported that respondents rejected authoritarian attitudes as well as the desire for social distance toward people with mental illness and not only displayed benevolent attitudes, but also endorsed community integration. Linden and Kavanagh (2011) reported similar results. Munro and Baker (2007) reported that their respondents mostly agreed with positive statements, such as “psychiatric illness deserves at least as much attention as physical illness” (95.7% agreement) and disagreed with negative statements, such as “depression occurs in people with weak personality” (90% disagreed). It is important to mention, that even in studies that reported mostly positive attitudes, there were some negative attitudes, such as consideration that psychiatric drugs were used to control disruptive behavior (61.7% agreement), and that nurses perceived mentally ill patients with pessimism (semantic differential: pessimism – optimism). Authors of four studies reported that psychiatric nurses had mostly negative attitudes (Hamdan-Mansour & Wardam, 2009; Hsiao et al., 2015; Magliano et al., 2004; Sevigny et al, 1999). Majority of nurse respondents considered that psychiatric illness did not deserve as much attention as physical illness (94.6%, 87/92), 84.8% (78/92) considered that a person with mental illness had no control over her or his emotions, and 68.5% (63/92) agreed that depression was occurring in people with weak personality (Hamdan-Mansour & Wardam, 2009). Hsiao et al., (2015) found that psychiatric nurses had significantly more negative attitudes toward patients with schizophrenia than nurses who worked in community-based clinics, and that nurses had more negative attitudes toward people with schizophrenia than those with depression. Magliano et al., (2004) reported that 86% (163/190) of nurses considered people with schizophrenia as unpredictable, and 87% (165/190) considered that people were keeping away from patients with schizophrenia. Nurses also agreed that patients with schizophrenia should not have children (72%, 137/190), and that they should not get married (63%, 119/190), (Magliano et al., 2004). Even though most responses were negative, nurses also agreed with positive statements and considered that patients with schizophrenia should be allowed to vote (63%, 119/190), and that they were as able to work as other people (79%, 150/190), (Magliano et al., 2004). Sevigny et al., (1999) reported that nurses mostly held negative attitudes toward mentally ill people and generally more negative than physicians. Thirty eight percent of nurses considered a mental illness as any other illness (n=74) and 63% displayed authoritarian attitudes toward mentally ill patients. Nurses in Sevigny et al., (1999) also reported positive attitudes. Almost 60% of respondents disagreed that lack of discipline and will power was causing mental illness. Four studies reported mixed attitudes (Foster et al., 2008; Kukulu et al., 2007; Serafini et al., 2011; Nordt et al., 2006). The authors of all four studies reported results that showed negative and positive attitudes toward mental illness. Nordt et al., (2006) reported that nurses endorsed negative stereotypes of mentally ill people, but opposed restriction of civil rights of the mentally ill. Serafini et al., (2011) reported that while 75% of nurses believed that people with schizophrenia were unpredictable and 80% expressed a desire for social distance, 60% did not believe that people with schizophrenia were dangerous (n=50). Kukulu and Ergun, (2007) also confirmed the desire for social distance: while 56.7% of nurses said that they could work with a person with schizophrenia, 91.7% would not marry a person with that disorder (n=543). Foster et al., (2008) also reported mixed attitudes among their respondents: while 91.3% of nurses considered that people with a psychiatric history should be given jobs with responsibilities, 91.3% said that psychiatric medications were used to control disruptive behavior instead of being used to control the symptoms (n=23). Attitudes of nurses compared to the general public Three studies compared nurses’ attitudes toward mental illness with non-healthcare professionals such as family members and the general public, with mixed results (Magliano et al., 2004; Nordt et al., 2006; Scheerder et al., 2011). Magliano et al. (2004) reported that nurses (n=190) had more negative attitudes than the relatives (n= 709) of patients with mental illness. For example, 86% of nurses believed that patients with schizophrenia were unpredictable compared with only 65% of relatives having the same attitude. In addition, 72% of nurses compared to 32% of relatives considered that mentally ill patients should be punished for wrong behavior in the same manner as other people. In regards to personal civil rights, nurses and relatives had similar attitudes about whether those with schizophrenia should have children (29%) or have the right to vote (66%). Finally, while almost half of the relatives (44%) considered that mentally ill people could work as other people, 79% of nurses disagreed. In a second study, Nordt et al., (2006) compared five groups, including nurses and 253 members of the general population. The nurses and the general population agreed with negative stereotypes of the mentally ill at a similar level. However, while 54% of nurses opposed revocation of the Driver’s License, 65.7% of the general public endorsed that restriction. More members of the general public than nurses considered that the mentally ill people should not vote (19.6% vs. 2.8%), and while almost all nurses agreed to compulsory admission (98.2%), 67.5% of general public respondents endorsed this option. In the third study (Scheerder et al., 2010), community facilitators (clergy, police, youth workers, pharmacists, social workers and volunteers) were asked their opinions about depression and were compared with mental health professionals and nurses. While 77% of community facilitators considered that depression is a real disease, 60% of nurses endorsed that opinion. Both groups agreed that depression could be treated (83.4% of community facilitators vs. 81.9% of nurses). Specific themes Etiology of mental illness Seven studies reported nurses’ beliefs about the etiology of mental illness (Foster et al., 2008; Kukulu & Ergun, 2007; Magliano et al., 2004; Munro & Baker, 2007; Scheerder et al., 2011; Serafini et al., 2011; Sevigny et al, 1999). Nurses predominantly have the attitude that mental illness is a disease of a hereditary nature (range: 65%–93%). Additional attitudes about the etiology of mental illness included personal weakness, result of alcohol and/or drug use, and stress and family conflict. Most nurse respondents (59–90%) did not consider mental illness as emanating from a lack of will power (Munro & Baker, 2007; Sevigny et al., 1999). Social restrictiveness and distance Social restrictiveness in mental illness stigma literature measured the desire to restrict people with mental illness from roles in society. Social distance refers to the proximity that one desires between self and a mentally ill person in a social situation. Nine studies reported nurses’ attitudes toward social restrictions that should be imposed on the mentally ill as well as the social distance that the respondents preferred to maintain from this population (Arvaniti et al., 2009; Aydin et al., 2003; Chambers et al., 2010; Kukulu et al., 2007; Linden & Kavanagh, 2012; Magliano et al., 2004; Munro & Baker, 2007; Nordt et al., 2006; Sevigny et al., 2011. Attitudes toward social restrictions and distance were measured through questions that examined attitudes toward right to vote, revocation of one’s driver’s license, isolation of the mentally ill from the residential neighborhoods, mandatory abortion for women with diagnosed schizophrenia, and opposition to marrying people with mental disorders. Almost half of the nurses (46%, 311/676) in one study agreed that people who suffered from any mental health issues should have their driver’s license revoked (Nordt et al., 2006). The majority of respondents would oppose a marriage of a family member to a person with mental illness. Almost two-thirds (63%) of nurses in one study agreed that patients with schizophrenia should not marry at all (Magliano et al., 2004). Similarly, in another study, 100% of respondents agreed that they would not want their sister to marry someone with a mental disorder (Aydin et al., 2003). The majority of these respondents (76.2%, 32/42), also agreed that they would not rent their apartments to mentally ill people (Aydin et al., 2003). Perceived dangerousness Studies presented mixed attitudes and beliefs regarding the level of dangerousness, unpredictability, and emotional instability of mentally ill. Serafini et al., (2011) reported that 16 of 40 nurses (40%) considered patients with schizophrenia to be dangerous, while Munro and Baker (2007) reported that 85% of respondents did not. Kukulu and Ergun (2007) reported that over half of nurses (53%) agree they would be frightened if people with mental illness lived close by. Severely mentally ill people were perceived as unpredictable (from 75% to 86% agreement). Questions concerning the lack of control over emotions showed mixed opinions. Hamdan-Mansour and Wardam, (2009) reported that 84.8% of the 92 nurses agreed with the statement that: “mentally ill have no control over their emotions”, while Foster et al., (2008) reported the opposite with almost 70% of nurses disagreeing with the following statement: “mentally ill patients have no control over the emotions”. Schizophrenia and depression Three studies compared specific attitudes toward schizophrenia and depression (Aydin et al., 2003; Hsiao et al., 2015; Nordt et al., 2006). Attitudes were generally more positive toward patients with depression. The comparisons included discrimination toward housing, use of services, work and proximity in social settings, such as their comfort level working with someone who has a mental illness. Aydin et al., (2003) reported that more nurses would be disturbed if they had to shop at a market run by a person with schizophrenia (33.3%) rather than the depression (11.1%). While 38.1% of nurses would be disturbed to work with a person with schizophrenia, only 5.9% would feel that way working with a person with depression. However, in some social situation, discrimination toward people with depression or schizophrenia were at the same level: 100% of respondents would not want their sisters to marry either one, 76% would not go to a hairdresser with either disorder and 76% would not rent a house to any of them. Hsiao et al., (2015) and Nordt et al., (2006) findings supported that nurses had more negative attitudes toward patients with schizophrenia rather than major depression. Discussion The studies included in this review examined nurse attitudes toward mental illness across 20 countries. Globally, nurses tend to have mixed attitudes toward different aspects of mental illness. Evidence about the difference in attitudes of psychiatric nurses and non-psychiatric nurses was contradictory. However, one study determined that the higher the education level of the nurse, the more likely the nurse would have a more positive attitude about mental illness. This suggests that education regarding mental illness could potentially alleviate negative attitudes associated with mental illness among nurses. Furthermore, the mixed attitudes found in this review may be partially explained by different cultural beliefs. Among the eight studies conducted in one or more European countries, both positive and negative nursing attitudes were reported, both within and across countries (Arvaniti et al., 2009; Chambers et al., 2010; Linden & Kavanagh, 2012; Magliano et al., 2004; Munro & Baker, 2007; Nordt et al., 2006; Scheerder et al., 2011; Serafini et al., 2011). In contrast, the majority of studies conducted in Middle Eastern or Asian countries, reported more negative than positive nursing attitudes, suggesting that culture may play an influential role in nursing perception of mental illness. Another factor that might have contributed to the finding that nurses’ attitudes toward the mentally ill were quite mixed was the fact that various measurement tools were used. More than half of the studies (8/14) used different tools. Three tools alone were questionnaires adapted by researchers. This makes the comparisons of results across studies difficult. Finally, the results of this study were surprising in that professional nurses’ attitudes toward mental illness were comparable to attitudes among the general public rather than reflective of professional expertise (Al-Krenawi, Graham, Dean, & Eltaiba, 2004; Angermeyer & Dietrich, 2006; Ozmen et al., 2004; Schomerus et al., 2012; Tsang, Tam, Chan, & Cheung, 2003). One would anticipate that professional training would have an impact on attitudes toward these patients. The fact that nurses who worked on psychiatric units did not express more positive attitudes toward their patients as compared to nurses who worked in general medicine might be due to perception bias. These nurses often see patients readmitted for care with multiple psychiatric hospitalizations, which may influence their attitude toward mental illness capacity and prognosis. Linden and Kavanagh (2011) support this explanation, in that nurses from mental health community settings who worked with more stable patients endorsed more positive attitudes than those who worked on acute inpatient wards. If nurses have clear guidelines regarding how to approach patients with various mental illnesses, how to address their symptoms, and what therapeutic interventions are most effective, they may feel more empowered in their nursing roles, thus promoting a more positive outlook on mental illness. Further, management can be influential by providing explicit and overt support for culture change toward more supportive attitudes of patients diagnosed with mental illness. Limitations This review has some limitations. The English language limitation as well as the limited number of databases searched might have led to omission of relevant studies. Furthermore, we did not include the grey literature in this review. Conclusions and Future Research In summary, this review found that nursing attitudes toward people with mental illness varied, both within and across countries and mimicked attitudes similar to the general public. Since no studies were conducted in the U.S., there is a need to examine the attitudes of nurses toward those with mental illness and compare the U.S. to other countries. It is crucial to assess nurses’ attitudes toward mental illness and explore the factors associated with positive beliefs. A better understanding of mental illness and related nursing attitudes will help to inform delivery of care to those patients who suffer from mental illness. Funding Source This review was funded by the Jonas Center For Nursing and Veteran Healthcare and the National Institute of Nursing Research (NINR) Comparative and Cost-Effectiveness Research Training for Nurse Scientists, T32 NR014205. Figure 1 Flow Diagram Table 1 Study Characteristics Study Country Sample Measures Key Findings Arvaniti et al., 2009 Greece 130 nurses 76 physicians 140 other staff 239 medical students 10 medical wards and one psychiatric ward of a general hospital The Level of Contact Report (LCR) Opinion about Mental Illness (OMI) The Authoritarianism Scale (AS) (% Nurse Agreement) Social discrimination Mentally ill patients are dangerous (24%) Mentally ill patients should not marry (53%) Mentally ill patients should be separated from patients without mental illness (60.7%) Social restriction: Mentally ill patients should not vote (31%) Nurses endorsed more restrictive attitudes than physicians More knowledge about mental illness was associated with more positive attitudes Social integration: Nurses were more negative than physicians but less authoritarian than medical students Aydin et al., 2003 Turkey 40 nurses 40 academicians 40 physicians 40 hospital employees Medical clinics Schizophrenia and depression vignettes measures: Social distance Burden of illness Nurses exhibited more negative attitudes toward a person with schizophrenia than depression 100% of respondents said they would be disturbed by having a sister marrying a mentally ill person with schizophrenia and depression, but less bothered by working at the same place Chambers et al., 2010 Finland Italy Lithuania Portugal Ireland 810 nurses Psychiatric Hospitals (n=21) Community Attitudes toward the Mentally Ill (CAMI) Nurses exhibited positive attitudes toward mentally ill across all countries Most positive attitude toward mental illness (Portugal) Most negative attitude toward mental illness (Lithuania) Foster et al., 2008 Fiji 23 nurses 48 orderlies Psychiatric hospital Attitudes Toward Acute Mental Health Scale Etiology(% agreement) “Mental illnesses are caused by genetic factors” (65.2%) Attitudes “Psychiatric illness deserves as much attention as physical illness” (86.9%) “Mentally ill have no control over their emotions” (30.5%) “Manner in which you talk to patients affects their mental state” (91.3%) Hamdan-Mansour and Wardam., 2009 Jordan 92 nurses Acute and chronic mental health inpatient and outpatient facilities Attitudes Toward Acute Mental Health Scale Significant difference in attitudes between older and younger nurses Special training in psychiatric nursing led to more positive attitudes Higher level of education was associated with more positive attitudes. Nurse Agreement (%) “Psychiatric Illness deserves as much attention as physical illness” (5.4%) “Depression occurs in people with weak personality” (68.5%) “Mentally ill patients have no control over their emotions” (84.8%) “Mental illnesses are genetic in origin” (76.1%). Hsiao et al., 2015 Taiwan 180 nurses. Psychiatric hospitals (n=3) Jefferson Scale of Empathy-Health Profession version (JSE-HP) Attitudes of Mental Illness Questionnaire (AMIQ) More negative attitudes towards schizophrenia than depression (p <.001) The older the nurse, the more positive attitude (p <.01) The more experience the more positive attitudes (p <.001) There was a positive correlation between empathy and attitudes toward mental illness (p <.01) There were no gender difference (p =.84) Staff nurses endorsed more negative attitudes than nurse managers (p <.02) Nurses on acute psychiatric units endorsed more negative attitudes toward schizophrenia than nurses who worked in community-based outpatient clinics (p =.006) Kukulu and Ergun, 2007 Turkey 543 nurses Psychiatric wards of teaching hospitals Questionnaire reported in Turkish language Etiology(% Agreement) Schizophrenia present from birth (93.2%) Schizophrenia caused by social problems (51.4%) Social distance “People with schizophrenia should be free in society” (31.9%) Could work with people with schizophrenia (56.7%) Could marry a person with schizophrenia (8.3%) Have a neighbor with schizophrenia (42.9%) Rent home to a person with schizophrenia (63.2%) Linden and Kavanagh 2012 Ireland 121 nurses 66 student mental health nurses Inpatient and Community Setting (n=2) Community Attitudes toward Mental Illness Scale (CAMI) Social Interaction Scale (SIS) CAMI Nurses disagreed with social restrictiveness and authoritative attitudes toward mental illness Nurses agreed with integrating mentally ill into the community Nurses agreed with exhibiting benevolent attitudes toward those with mental illness Community mental health nurses showed more positive attitudes than those who worked in inpatient setting SIS Inpatient mental health nurses showed more socially restrictive attitudes than nurses in a community setting. Magliano et al., 2004 Italy 190 nurses 110 psychiatrists 709 patient relatives Mental health services (n=30) Opinions About Mental Illness Questionnaire Etiology of schizophrenia (% Nurse agreement) Heredity (74%); Stress (53%); Alcohol (42%), Drugs (48%); Family conflict (48%), Trauma (36%) Social functioning “Patients with mental illness should work as other people” (79%) “Patients with schizophrenia are unpredictable (86%) Civil rights Patients with schizophrenia should be responsible in court (72%) Patients with schizophrenia should vote (66%) Patients with schizophrenia should not get married (63%) Patients with schizophrenia should not have children (72%) Wife of patient with schizophrenia should be allowed to divorce upon diagnosis (50%) Munro and Baker 2007 England 141 nurses Acute mental health unit Attitude Toward Acute Mental Health Scale Positive Attitudes (% Nurse Agreement) “Psychiatric illness deserves at least as much attention as physical illness” (80%) “Depression occurs in people with a weak personality” (20%) Negative Attitudes Psychiatric drugs are used to control disruptive behavior” (67%) Neutral Attitudes “Mental illness is genetic in origin” (46.4%) “People are born vulnerable to mental illness” (45.7%) Nordt et al., 2006 Switzerland 684 nurses 204 psychiatrists 185 other professionals Psychiatric wards of hospitals (n=29) Outpatient clinics (n=3) 1737 members of the general public Computer Assisted Telephone Interview Social Acceptance Nurses and Psychiatrists reported similar negative attitudes Nurses endorsed higher social distance toward people with schizophrenia than toward people with depression. Social restrictiveness (% Nurse Agreement) Mentally ill people should have driver’s license revoked (46%) In favor of withdrawing the right to vote (2.8%) Mentally ill should abort when pregnant (9.8%) In favor of compulsory admission (98.2%) Scheerder et al., 2011 European Alliance Against Depression Belgium Estonia France Germany Hungary Ireland Italy Scotland Slovenia 887 nurses 334 nursing assistants 169 mental health professionals (physicians and mental health professionals) 968 community facilitators (clergy, social workers) from a training program and professional associations Adaptation of 3 tools Depression Attitude Questionnaire Defeat Depression Questionnaire Instruments of EAAD partner countries Mental health professionals had the least negative attitudes toward people with depression and the use of antidepressants. Nurses had more negative attitudes toward people with depression Positive Attitudes toward mental health (% agreement) “Depression is a real disease” (60% nurses; 95% physicians) “Depression can be treated” (81.9% nurses; 95.8% physicians) Negative attitudes toward treatment (% agreement) Antidepressants (AD) are addictive (67.8% nurses; 24.8% physicians) AD is effective in treating depression (47.9% nurses; 83.1% physicians) AD can change one’s personality (66% nurses; 28.6% physicians) Etiology of mental illness Heredity (In agreement) 56.5% nurses; 65.1% physicians Serafini et al., 2011 Italy 50 nurses 50 medical physicians 50 medical students 52 psychiatric outpatients from a university hospital Responses to questions about vignettes Standardized Stigmatization Questionnaire (SSQ), (Part 1) Vignettes (% Nurse Agreement) Positive attitudes Genetic basis of schizophrenia (80%) Negative attitudes Most people think that people with schizophrenia are unpredictable (75%) Most people want to keep their distance from people with schizophrenia (80%) Neutral attitudes Most people thought that people with schizophrenia are dangerous (50%) SSQ Significant response difference between medical doctors and nurses (p.038) Sevigny et al., 1999 China 74 nurses 26 physicians Psychiatric hospital Community Attitudes Toward the Mentally Ill Nurses endorsed mostly negative attitudes(% Nurse Agreement) “Mental illness is an illness as any other” (38%) “Most women who were once patients in a mental hospital can be trusted to take care of babies” (30%) “Mental patients need the same kind of control and discipline as a young child” (63%) “Anyone with a history of mental problems should be excluded from taking public office” (71%) “The mentally ill should not be given any responsibility” (78%) Nurses endorsed more negative attitudes than physicians This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. Anxiety and Depression Association of America 2010 Facts and Statistics Retrieved from http://www.adaa.org/about-adaa/press-room/facts-statistics Adorno TW 1950 The Authoritarian personality 1 New York Harper Ajzen I 2005 Attitudes, personality and behavior 2 Maidenhead, England Open University Press Ajzen I Fishbein M 1980 Understanding attitudes and predicting social behavior Englewood Cliffs, NJ Prentice-Hall Al-Krenawi A Graham JR Dean YZ Eltaiba N 2004 Cross-national study of attitudes towards seeking professional help: Jordan, United Arab Emirates (UAE) and Arabs in Israel International Journal of Social Psychiatry 50 2 102 114 10.1177/0020764004040957 15293428 American Psychiatric Association 1994 Diagnostic and statistical manual of mental disorders 4 Washington, DC American Psychiatric Association American Psychiatric Association 2013 Diagnostic and statistical manual of mental disorders 5 Arlington, VA American Psychiatric Publishing Angermeyer MC Dietrich S 2006 Public beliefs about and attitudes towards people with mental illness: A review of population studies Acta Psychiatrica Scandinavica 113 3 163 179 10.1111/j.1600-0447.2005.00699.x 16466402 Angermeyer MC Matschinger H 1997 Social distance towards the mentally ill: results of representative surveys in the Federal Republic of Germany Psychological Medicine 27 1 131 141 9122293 Arkar H 1991 The social refusing of mental health patient Journal of Psychiatry and Neurological 4 6 9 Arvaniti A Samakouri M Kalamara E Bochtsou V Bikos C Livaditis M 2009 Health service staff’s attitudes towards patients with mental illness Soc Psychiatry Psychiatr Epidemiol 44 8 658 665 10.1007/s00127-008-0481-3 19082905 Aydin N Yigit A Inandi T Kirpinar I 2003 Attitudes of hospital staff toward mentally ill patients in a teaching hospital, Turkey International Journal of Social Psychiatry 49 1 17 26 12793512 Baker JA Richards DA Campbell M 2005 Nursing attitudes towards acute mental health care: Development of a measurement tool Journal of Advanced Nursing 49 5 522 529 10.1111/j.1365-2648.2004.03325.x 15713184 Barrett MS Chua WJ Crits-Christoph P Gibbons MB Casiano D Thompson DON 2008 Early withdrawal from mental health treatment: implications for psychotherapy practice Psychotherapy (Chicago, Ill) 45 2 247 267 Centers for Disease Control and Prevention 2013 Mental Illness Atlanta, GA Chambers M Guise V Välimäki M Botelho MA Scott A Staniuliené V Zanotti R 2010 Nurses’ attitudes to mental illness: A comparison of a sample of nurses from five European countries International Journal of Nursing Studies 47 3 350 362 10.1016/j.ijnurstu.2009.08.008 19804882 Chiu-Yueh H Huei-Lan L Yun-Fang T 2015 Factors influencing mental health nurses’ attitudes towards people with mental illness Internation Journal of Mental Health Nursing 24 3 272 280 10.1111/inm.12129 Clement S Schauman O Graham T Maggioni F Evans-Lacko S Bezborodovs N …Thornicroft G 2015 What is the impact of mental health-related stigma on help-seeking? A systematic review of quantitative and qualitative studies Psychological Medicine 45 1 11 27 10.1017/S0033291714000129 24569086 Corrigan P 2004 How Stigma Interferes With Mental Health Care American Psychologist 59 7 614 625 10.1037/0003-066X.59.7.614 15491256 Corrigan P Mittal D Reaves CM Haynes TF Han X Morris S Sullivan G 2014 Mental health stigma and primary health care decisions Psychiatry Research 218 1 35 38 10.1016/j.psychres.2014.04.028 24774076 Dell’osso B Glick ID Baldwin DS Altamura AC 2012 Can Long-Term Outcomes Be Improved by Shortening the Duration of Untreated Illness in Psychiatric Disorders? A Conceptual Framework Psychopathology 46 1 14 21 http://dx.doi.org/10.1159/000338608 22890286 Edward K Munro I 2009 Nursing considerations for dual diagnosis in mental health International Journal of Nursing Practice 15 2 74 79 10.1111/j.1440-172X.2009.01731.x 19335524 Eker D Arkar H 1991 Experienced Turkish Nurses’ Attitudes towards mental illness and the predictor variables of their attitudes International Journal of Social Psychiatry 37 3 214 222 10.1177/002076409103700308 1743906 Evans-Lacko S Brohan E Mojtabai R Thornicroft G 2012 Association between public views of mental illness and self-stigma among individuals with mental illness in 14 European countries Psychological Medicine 42 8 1741 1752 http://dx.doi.org/10.1017/S0033291711002558 22085422 Fishbein M 2010 Predicting and changing behavior: the reasoned action approach New York Psychology Press Fishbein M Ajzen I 1975 Belief, attitude, intention, and behavior: An introduction to theory and research Reading, MA Addison-Wesley Fishbein M Ajzen I Albarracin D Hornik RC 2007 Prediction and change of health behavior: applying the reasoned action approach Mahwah, NJ L. Erlbaum Associates Foster K Usher K Baker JA Gadai S Ali S 2008 Mental health workers’ attitudes toward mental illness in Fiji Australian Journal of Advanced Nursing 25 3 72 79 Haghighat R 2005 The development of an instrument to measure stigmatization: Factor analysis and origin of stigmatization The European Journal of Psychiatry 19 3 10.4321/S0213-61632005000300002 Hamdan-Mansour AM Wardam LA 2009 Attitudes of Jordanian mental health nurses toward mental illness and patients with mental illness Issues in Mental Health Nursing 30 11 705 711 19874099 Hatzenbuehler ML Phelan JC Link BG 2013 Stigma as a fundamental cause of population health inequalities American Journal of Public Health 103 5 813 821 10.2105/AJPH.2012.301069 23488505 Hinshaw SP Stier A 2008 Stigma as related to mental disorders Annual Revue of Clinical Psychology 4 367 393 doi:0.1146/annurev.clinpsy.4.022007.141245 Hoffman DL Dukes EM Wittchen HU 2008 Human and economic burden of generalized anxiety disorder Depress Anxiety 25 1 72 90 10.1002/da.20257 17146763 Hoge CW Grossman SH Auchterlonie JL Riviere LA Milliken CS Wilk JE 2014 PTSD treatment for soldiers after combat deployment: Low utilization of mental health care and reasons for dropout Psychiatric Services 65 8 997 1004 10.1176/appi.ps.201300307 24788253 Hojat M Gonnella JS Nasca TJ Mangione S 2002 Physician empathy: Definition, components, measurement, and relationship to gender and specialty The American Journal of Psychiatry 159 9 1563 1569 12202278 Holmes EP Corrigan PW Williams P Canar J 1999 Changing attitudes about schizophrenia Schizophrenia Bulletin 25 3 447 456 10478780 Kelly JA St Lawrence JS Smith & Hood HV 1987 Medical students’ attitudes toward AIDS and homosexual patients Journal of medical education 62 7 549 556 3599050 Kim PY Britt TW Klocko RP Riviere LA Adler AB 2011 Stigma, negative attitudes about treatment, and utilization of mental health care among soldiers Military Psychology 23 1 65 81 10.1080/08995605.2011.534415 Kukulu K Ergun G 2007 Stigmatization by nurses against schizophrenia in Turkey: A questionnaire survey Journal of Psychiatric and Mental Health Nursing 14 3 302 309 http://dx.doi.org/10.1111/j.1365-2850.2007.01082.x 17430454 Liberati A Altman DG Tetzlaff J Mulrow C 2009 The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate healthcare interventions: explanation and elaboration British medical journal (Clinical research ed) 339 7 21 1 b2700 b2700 10.1136/bmj.b2700 Linden M Kavanagh R 2012 Attitudes of qualified vs. student mental health nurses towards an individual diagnosed with schizophrenia Journal of Advanced Nursing 68 6 1359 1368 10.1111/j.1365-2648.2011.05848.x 21981761 Luty J Fekadu D Umoh O Gallagher J 2006 Validation of a short instrument to measure stigmatised attitudes towards mental illness The Psychiatrist 30 7 257 260 Madianos MG Madianou D Vlachonikolis J Stefanis CN 1987 Attitudes towards mental illness in the Athens area: implications for community mental health intervention Acta Psychiatrica Scandinavica 75 2 158 165 3565059 Magliano L De Rosa C Fiorillo A Malangone C Guarneri M Marasco C … Working Group of the Italian National, S 2004 Beliefs of psychiatric nurses about schizophrenia: a comparison with patients’ relatives and psychiatrists International Journal of Social Psychiatry 50 4 319 330 15648745 McDonald DD Frakes M Apostolidis B Armstrong B Goldblatt S Bernardo D 2003 Effect of a psychiatric diagnosis on nursing care for nonpsychiatric problems Research in Nursing & Health 26 3 225 232 http://dx.doi.org/10.1002/nur.10080 12754730 Mojtabai R Olfson M Sampson NA Jin R Druss B Wang PS … Kessler RC 2011 Barriers to mental health treatment: results from the National Comorbidity Survey Replication Psychological Medicine 41 8 1751 1761 http://dx.doi.org/10.1017/S0033291710002291 21134315 Munro S Baker JA 2007 Surveying the attitudes of acute mental health nurses Journal of Psychiatric & Mental Health Nursing 14 2 196 202 10.1111/j.1365-2850.2007.01063.x 17352783 National Alliance for the Mentally Ill 2013 Mental Illness Facts and Numbers Washington, DC CDC Newman D O’Reilly P Lee SH Kennedy C 2015 Mental health service users’ experiences of mental health care: an integrative literature review Journal of Psychiatric & Mental Health Nursing 22 3 171 182 10.1111/jpm.12202 25707898 NIMH 2011 Any Mental Illness Among Adults Washington DC Nordt C Rössler W Lauber C 2006 Attitudes of mental health professionals toward people with schizophrenia and major depression Schizophrenia Bulletin 32 4 709 714 10.1093/schbul/sbj065 16510695 Olfson M Mojtabai R Sampson NA Hwang I Druss B Wang PS … Kessler RC 2009 Dropout from outpatient mental health care in the United States Psychiatr Serv 60 7 898 907 10.1176/appi.ps.60.7.898 19564219 Ozmen E Ogel K Aker T Sagduyu A Tamar D Boratav C 2004 Public attitudes to depression in urban Turkey Social Psychiatry and Psychiatric Epidemiology 39 12 1010 1016 10.1007/s00127-004-0843-4 15583910 Reneses B Munoz E Lopez-Ibor JJ 2009 Factors predicting drop-out in community mental health centres World Psychiatry 8 3 173 177 19812755 Scheerder G Van Audenhove C Arensman E Bernik B Giupponi G Horel AC … Hegerl U 2011 Community and health professionals’ attitude toward depression: A pilot study in nine EAAD countries International Journal of Social Psychiatry 57 4 387 401 10.1177/0020764009359742 20223779 Schomerus G Schwahn C Holzinger A Corrigan PW Grabe HJ Carta MG Angermeyer MC 2012 Evolution of public attitudes about mental illness: A systematic review and meta-analysis Acta Psychiatrica Scandinavica 125 6 440 452 10.1111/j.1600-0447.2012.01826.x 22242976 Serafini G Pompili M Haghighat R Pucci D Pastina M Lester D … Girardi P 2011 Stigmatization of schizophrenia as perceived by nurses, medical doctors, medical students and patients Journal of Psychiatric Mental Health Nursing 18 7 576 585 http://dx.doi.org/10.1111/j.1365-2850.2011.01706.x 21848591 Sevigny R Yang W Zhang P Marleau JD Yang Z Su L … Wang H 1999 Attitudes toward the mentally ill in a sample of professionals working in a psychiatric hospital in Beijing (China) International Journal of Social Psychiatry 45 1 41 55 10443248 Smith DJ Court H McLean G Martin D Langan Martin J Guthrie B … Mercer SW 2014 Depression and multimorbidity: A cross-sectional study of 1,751,841 patients in primary care The Journal of Clinical Psychiatry 75 11 1202 1208 quiz 1208 10.4088/jcp.14m09147 25470083 Taylor SM Dear MJ Hall GB 1979 Attitudes toward the mentally ill and reactions to mental health facilities Social Science & Medicine. Part D: Medical Geography 13 4 281 290 http://dx.doi.org/10.1016/0160-8002(79)90051-0 Taylor SM Dear MJ 1981 Scaling community attitudes toward the mentally ill Schizophrenia Bulletin 7 2 225 240 7280561 Tsang HWH Tam PKC Chan F Cheung WM 2003 Stigmatizing attitudes towards individuals with mental illness in Hong Kong: Implications for their recovery Journal of Community Psychology 31 4 383 396 10.1002/jcop.10055 Whiteford HA Degenhardt L Rehm J Baxter AJ Ferrari AJ Erskine HE … Vos T 2013 Global burden of disease attributable to mental and substance use disorders: Findings from the Global Burden of Disease Study 2010 The Lancet 382 9904 1575 1586 http://dx.doi.org/10.1016/S0140-6736(13)61611-6

PMC005xxxxxx/PMC5127453.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 0144042 6320 Otolaryngol Clin North Am Otolaryngol. Clin. North Am. Otolaryngologic clinics of North America 0030-6665 1557-8259 27888914 5127453 10.1016/j.otc.2016.08.004 NIHMS814038 Article Bacterial Pathogens and The Microbiome Vickery Thad W. BA 1 Ramakrishnan Vijay R. MD 2 1 University of Colorado School of Medicine, 13001 E 17th Pl, Aurora, CO 80045 2 Department of Otolaryngology-Head and Neck Surgery, University of Colorado, 12631 E 17th Ave, B205, Aurora, CO 80045 Corresponding Author: Vijay Ramakrishnan, MD Vijay.Ramakrishnan@UCDenver.edu 1 9 2016 2 2017 01 2 2018 50 1 2947 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. bacteria chronic rhinosinusitis microbiome culture-independent microbiology sinusitis 1. Introduction Chronic rhinosinusitis (CRS) continues to be one of the most prevalent healthcare problems in the United States. Despite the significant morbidity, loss of productivity, and healthcare costs associated with CRS, the underlying processes that lead to disease remain poorly understood. The non-specific clinical symptoms of nasal obstruction, rhinorrhea, facial pain, and anosmia may represent a common endpoint for various inflammatory mechanisms occurring in different anatomic areas. CRS is increasingly being appreciated as a clinical syndrome with a wide spectrum of overlapping disease physiology. For instance, CRS with nasal polyps (CRSwNP) often is characterized by eosinophilic inflammation and increased production of histamine, IL-5 and IL-13,1 whereas CRS without nasal polyps (CRSsNP) is often considered a predominantly neutrophilic disease characterized by high levels of IL-1, IL-6 and TNFα.2 In practice, however, there are CRSsNP patients with high levels of eosinophils and CRSwNP patients that exhibit robust neutrophilic infiltration within the sinonasal epithelium. Thus our classification of CRS in clinical practice is often not as simple as we would prefer. I. CRS Pathophysiology and Immune homeostasis CRS is characterized by persistent inflammation, a dysregulated immune response, and host-microbial interactions that together result in disruption of epithelial barrier function, poor wound healing, tissue remodeling, and clinical symptoms. Historically the significance of bacteria in acute and chronic rhinosinusitis has focused on the interactions between a single bacterial pathogen and its host. However, developing concepts in microbial ecology, laboratory methods in culture-independent microbiology, and bioinformatics, are furthering our capacity to study complex microbial communities as an entire functional unit. The nasal cavity and paranasal sinuses are the first tissues exposed to airborne environmental challenges, including pathogenic and non-pathogenic bacteria, viruses, fungi, allergens and toxins. The mucosal surface utilizes several immune mechanisms to promote homeostasis, which can broadly be divided into innate and adaptive immunity. Many host factors impact the functionality of the immune response that is thought to predispose individuals to the development of CRS.3 Innate vs. adaptive immunity Innate immunity classically refers to the non-specific defense mechanisms that are rapidly activated following exposure to antigenic material and confer immediate protection. Within the upper respiratory tract, this includes the physical barrier provided by the ciliated pseudostratified columnar respiratory epithelium that lines the sinonasal cavity. This resilient barrier contains interspersed goblet cells that secrete a viscoelastic mucus layer atop the epithelial surface composed of high molecular weight glycoprotein mucins and heavily glycosylated molecules. In conjunction with beating cilia, the enriched mucus layer promotes non-specific mucociliary clearance of microbes and irritant particles.4 Barrier dysfunction can contribute to CRS, and when coupled with defects in mucociliary clearance that promote bacterial colonization, bacterial invasion and further barrier disruption may occur.3,5 Classic genetic defects in ciliary function, such as cystic fibrosis and primary ciliary dyskinesia, are often used as examples but acquired ciliary dysfunction occurs in CRS as well.6 Poor barrier function and dysfunctional mucociliary clearance are host defects that to predispose individuals to pathogen colonization and the development of recurrent infection.7–9 Sinonasal epithelial cells secrete enzymes, opsonins, defensins, permeabolizing proteins and other endogenous antimicrobial products into the apical mucus layer. These host defense molecules are important for directly neutralizing microbes and recruiting inflammatory cells that modulate the immune response. Specifically, epithelial cells secrete enzymes such as lysozyme, peroxidases, and chitinases, the small cationic permeabolizing proteins such as the defensins and cathelicidins. Additionally, proteins such as lactoferrin, mucins, C-reactive protein and secretory leukocyte proteinase inhibitor (SLPI), collectively provide protection from bacteria, fungi, and viruses at the apical surface.9,1011 When pathogens do invade the sinonasal epithelium, circulating professional phagocytes, possessing pattern recognition receptors (PRRs) on their cell surface, recognize pathogen associated molecular patterns (PAMPs) and damage associated molecular patterns (DAMPs). When host epithelial cell PRRs bind pathogenic or damaged cell proteins, the acute inflammatory pathways are activated and the tissue becomes primed for the adaptive immune response.12,10 Toll-like receptors (TLRs) are a subset of PRRs that allow phagocytes to recognize bacterial motifs such as lipopolysacharide (via TLR-4) and CpG repeats (via TLR-9) associated with Gram-negative bacteria and bacterial DNA respectively. Current literature suggests that variable gene expression of TLRs and other cytokines may predispose some individuals to develop CRS.13 For example, patients with CRS demonstrate variable expression of TLR2 mRNA.14,15 TLR2 is expressed on sinonasal epithelial cells and recognizes several bacterial, fungal, and viral proteins. Binding of a ligand to TLR2 activates the acute inflammatory response and also primes the adaptive immune response by increasing expression of MHC-II co-stimulatory molecules required to present antigens to T-follicular helper cells. The resulting Th1 response is necessary for clearance of S. pneumonia infection, for example.16,17 Another study identified decreased TLR9 mRNA expression in CRSsNP;18 however, in depth study of TLR-mediated disease in CRS is lacking. In concert with the acute inflammatory response, cytotoxic natural killer (NK) cells are important to mounting an appropriate innate immune response. While NK cells are often understood to induce apoptosis in viral infected cells and tumor cells, they are becoming increasingly of interest in CRS. Patients with CRS have dysfunctional NK cells that demonstrate impaired degranulation and diminished release of TNFα and IFNγ.19 The adaptive immune system is stimulated when PMNs and macrophages are recruited to the site of infection to fully eradicate bacterial infection and establish long-lasting cell-mediated immunity.20 Antigen presenting cells such as monocytes, macrophages, B-lymphocytes and dendritic cells process complex protein antigens and present antigenic peptides on their surface that are then presented to T helper cells on the appropriate MHC. Binding of the antigen-MHC complex to the T-cell receptor activates the adaptive immune system and leads to longer lasting T-effector cells and antibody producing plasma cells.21 While much attention is given to defective innate immunity in the development of CRS, the adaptive immune system is crucial for developing an appropriate T-cell response, and much of the CRS literature supports polarized populations of T-follicular helper cells (either Th1 or Th2) in subsets of CRS. Th1 cells produce IFNγ, IL-2, and TNFα, which ultimately lead to a robust cell-mediated response and phagocyte-mediated inflammation. Th2 cells produce IL-4, IL-5, and IL-13 that promote a strong antibody response and the accumulation of eosinophils while inhibiting the phagocytes and the resulting potent inflammatory response.22 Furthermore, the T-follicular helper cell environment influences macrophage differentiation into M1 macrophages that are necessary for a mounting a vigorous inflammatory response against intracellular pathogens, or M2 macrophages that are associated with eosinophilic states.23 M2 macrophages are predominant in CRSwNP and may play a role in exacerbating polyposis because they are unable to phagocytose S. aureus.24 The inappropriate persistence of Th1 inflammation in CRSsNP and the ongoing recruitment of eosinophils and Th2 cells in CRSwNP are hallmarks of adaptive immune dysfunction characteristic in CRS. Surface mucosal niche Once thought to be a “sterile” environment in the healthy state, the paranasal sinuses are now widely appreciated to harbor rich and diverse populations of commensal bacteria. Commensals are most often defined as the community of microorganisms that colonize epithelial surfaces without causing harm to the host. Our understanding of “commensal” bacteria is evolving to recognize the shifting selective pressures in the microbial community that likely include other forms of symbiosis between host cells and neighboring bacteria including parasitism, mutualism and amensalism. Host defense mechanisms contribute to a unique mucosal environment that constantly applies selective pressure on epithelial microbes resulting in a “surface mucosal niche.” This niche varies between individuals and between different anatomic compartments.25, 26 There may be drivers of this niche beyond host mucosal immune function, including environmental exposures (eg, smoking) or medication usage (eg, antibiotics) [Figure 1]. Bacteria, either live or their by-products such as DNA, quorum-sensing molecules, or metabolized waste, interact with the immune system and can influence inflammatory processes. II. Role of bacteria in initiation and sustenance of inflammation Historically, many described CRS as a disease that resulted from a persistent or incompletely treated acute infection. Current understanding is that CRS is an inflammatory process rather than infectious, but that inflammatory mechanisms modulated by commensal and pathogenic bacteria contribute to disease formation, beyond the simplistic notion of a single microbial pathogen interacting with a host and causing disease. Certainly, chronic inflammatory diseases can result from direct bacterial invasion at mucosal surfaces, resulting in compromised barrier function, a coordinated innate and adaptive immune response, and an acute inflammatory response which may evolve into prolonged inflammation leading to tissue damage, remodeling and fibrosis.27 It is possible that in some patients, bacterial infection initiates the inflammatory process that never resolves; yet, in others it may be the case that a non-infectious inciting event initiates an inflammatory response that alters the native mucosal surface bacterial niche, which then propagates the disease once the original disease-causing event has concluded [Figure 2].28 Earlier studies of sinonasal microbiota utilized culture-based microbiology techniques, which have recently been usurped by nucleic acid-based molecular techniques that allow for more sensitive and less biased detection of microbes, as well as the ability to characterize entire communities of microbes within the same sample.29, 30 Without evidence to support a definitive role for bacterial “infection” as the etiology for CRS, future studies are needed to better understand the role of bacteria in host susceptibility to disease development. 2. The Paranasal Sinus Biome in Health and Disease I. The microbiome Host-microbe interactions are an established contributing factor in the formation of CRS, but evidence for the presence or absence of a single microbe resulting in disease is lacking. The shifting paradigm focuses on the unique composition of the entire bacterial community that colonizes the sinonasal mucosa also known as the microbiome. The microbiome is a potentially diverse community of microbiota existing in a delicate symbiotic relationship within a human microenvironment. These organisms possess great genetic potential to act as disease modifiers and contribute to the maintenance of health and formation of disease on all epithelial surfaces including upper and lower airway.31, 32, 33 In intestinal epithelia, early microbial colonization is essential for normal immunologic development and influences susceptibility to inflammatory and allergic diseases later in life.34 In neonates, for example, early lower airway colonization with pathogens such as S. pneumonia, H. influenza, and M. catarrhalis increases the risk of recurrent wheezing and asthma.35 Additionally, several groups have demonstrated the importance of the microbiome on host adaptive immune system through modulation of dendritic cells, Th17 and T-reg cells.36, 37, 38,39 Although much of bacterial microbiome research is still emerging, it is one of the most intriguing current topics of CRS research. There remains a paucity of microbiome research on fungi, viruses, and bacteriophages in CRS, although growing evidence suggests that these microbes may contribute to a larger “metaorganism” in which interactions between and among all microbes and their host shape human health. Early efforts to describe microbial roles in CRS resulted in the “fungal hypothesis” which suggested that CRS resulted from an overexuberant host response to Alternaria fungi.40 While this theory does not explain many of the defects observed in CRS, and current guidelines recommend against antifungal therapy for CRS patients, there is continued research interest in the fungal microbiome.41 Recent studies using sequencing of the fungal 18S ribosomal RNA gene demonstrate a rich and diverse population of commensal fungal taxa in middle meatal lavages from healthy and CRS patients. Patients with CRS were found to have quantitative increases in the total amount of fungal 18S ribosomal RNA when compared to controls.42 A 2014 prospective study by Cleland et al. identified 207 unique fungal genera in 23 CRS patients and 11 controls. Interestingly, fungal genera traditionally associated with CRS such as Alternaria and Aspergillus were found in very low abundance. This study also assessed post-operative changes in the fungal microbiome and found decreased richness at 6 and 12 weeks after surgery.43 No studies to date have thoroughly profiled viral or bacteriophage populations within the paranasal sinuses, although there is certainly interest. For instance, recent evidence suggests that upper airway rhinovirus infection can alter the nasopharyngeal microbiome.44 Further studies directed at characterizing the entire microbiome population in health are necessary in order to develop a more robust understanding of the role microbial diversity plays in sinonasal health as well as the generation of disease. II. Bacteria in health Since the advent of culture-independent molecular techniques, several groups have used various techniques to characterize the sinonasal bacterial community in healthy individuals. Efforts to identify a distinct microbial profile in health are inconclusive, and there is currently no consensus on which bacteria predominate in or define the healthy state. Although several studies have characterized the bacterial communities in the upper airway, cross-study comparisons are difficult due to several variations in sampling methods, bacterial primer selection, sequencing methods, and data analysis pipelines.45 Even so, there are several patterns that frequently emerge. These include an abundance of Propionibacterium acnes, S. epidermidis, S. aureus, and Corynebacterium spp in health.46, 47, 26, 33 Of note, Staphylococcus spp. including S. aureus and coagulase negative Staphylococci are present in healthy subjects, and can behave in either pathogenic or commensal fashion based on particular strain, bacterial gene expression, and surrounding microbial interactions. While native bacteria act to promote homeostasis within the sinonasal epithelia, disruption of this balance known as dysbiosis may allow commensal organisms to act as opportunistic pathogens in disease states.48 The Human Microbiome Project (HMP) consortium compiled 16S rRNA gene sequences collected from 18 different sites across the body in order to better associate the microbiome with human health.49,50 Further analysis of the HMP data demonstrates that different human subjects possess unique native bacterial communities with highly variable taxonomic composition at the same body sites. These bacterial community profiles strongly associated with life-history characteristics such as history of being breastfed as an infant, gender, and level of education. Interestingly, despite profound inter- and intra-person variability in the bacterial microbiome, community function remains constant and the community types from one body site is predictive of community types at other body sites within the same individual.50,51 The association of bacterial community composition with life history factors such as presence or absence of breastfeeding raises questions about the source of commensal microbes. In the gastrointestinal literature, there are several studies that demonstrate that diet shapes the microbiome composition, and different microbial profiles are associated with diseases such as inflammatory bowel disease (IBD).52 In addition to dietary factors, the upper airway microbiota are a known source of microbes that colonize the lung and stomach.53 Ultimately, the aggregate of bacterial taxa in the sinus niche need to be better studied at a “metagenome” level to begin deciphering community function, and how specific bacterial ecosystems are established. III. Microbiome changes in disease Several statistical indices are use to describe species diversity within a microbial community. Alpha diversity is the intra-community diversity as measured by the total number and composition of species. Beta diversity is an estimate of the diversity (number and composition) comparison between different habitats. Abundance refers to the total amount of bacterial DNA that corresponds to specific bacterial taxa found within a sample. Richness describes the total number of species identified within a sample, i.e. greater numbers of distinct species means higher richness. Evenness is a measure of how similar in number each species within a bacterial community is. Several studies have identified perturbations in the microbiome during CRS. In general, S. aureus and anaerobes including Prevotella, Fusobacterium, Bacteroides spp. and Peptostreptococcus spp. are consistently more abundant in CRS versus healthy controls.26,41,46,47,54 Interestingly, although the abundance of pathogenic bacteria is increased in CRS patients, the overall total quantity of bacteria does not seem to change compared with healthy individuals. This indicates that the sinus niche is filled by a given amount of bacteria, and a relative dominance of the niche by pathogens is associated with disease. In these studies, despite having a similar overall bacterial burden, microbial richness, evenness and diversity of bacterial colonies was greatly diminished in CRS.26,33,54 This observation supports the hypothesis that disturbances in the microbial community are a part of CRS. Just as ecologists associate the macroscopic biodiversity and biomass of rainforests and coral reef habitats with the overall health of the ecosystem, the microscopic biodiversity of bacteria within a host mucosal niche is considered a hallmark of health. The frequent observation that CRS is associated with decreased microbial diversity parallels many other disease states. For example, decreased diversity in the gut microbiome is associated with obesity, active inflammatory bowel disease and psoriatic arthritis.55–57 Although we lack data to suggest that promoting bacterial diversity within the sinonasal niche would prevent or improve CRS, this may be a relevant consideration moving forward since the mainstays of CRS therapy are long-term broad spectrum antibiotics and corticosteroids which carry the potential to alter the microbiome composition.58,59 IV. Are microbiome alterations a cause or by-product of disease? Recent cross-sectional population studies have detected differences in the microbiome between CRS patients and healthy individuals, but what does this really mean? One possibility is that community alterations in the microbiome lead to epithelial barrier and immune dysfunction resulting in disease. Alternatively, persistent inflammation at the mucosal surface may result in a prolonged immune response, and/or alterations in the local microenvironment that shift the microbial community. Moreover, therapies administered for the disease (e.g., antibiotic therapy and steroids) likely impact the microbiome and may confound these observations. The multitude of variables present that impact the mucosal niche makes establishing this causation difficult in the absence of animal models. In reality, any or all of these possibilities may occur [Figure 1]. The concept of perturbations in the microbial community resulting in disease, i.e. dysbiosis, has been explored in the gastrointestinal tract. The normal gut flora play an essential role in maintaining immune homeostasis and disruptions in these mutualistic relationships are believed to lead to several conditions ranging from antibiotic-associated diarrhea and inflammatory bowel disease to necrotizing enterocolitis and colorectal cancer.48 While we know that CRS is associated with shifts in the bacterial microbiome it remains unclear if this is an inciting factor in the development or propagation of disease or merely a consequence. Emerging data from cross-sectional analysis of specific patient populations suggests that host factors, such as a history of tobacco use, presence of asthma, or recent antibiotic use can impact the microbiota.30,46,60 Given the panoply of disease mechanisms involved, the prolonged utilization of medical therapies, and the lack of a universal mouse model for the disease, human studies of CRS will require large-scale, multilevel, longitudinal design in order to delineate patterns of microbial perturbations and their significance. Such studies are critical to determining the role of the microbiome in disease formation, chronicity, severity and prognosis, and response to therapies. 3. Clinical implications and treatment considerations of bacterial pathogens I. Pathogens often implicated in CRS Traditional culture-based study vs. molecular techniques Many culture-based studies of sinus specimens from CRS patients identify the most common pathogens associated with CRS as S. aureus, P. aeruginosa, and with specialized culture techniques, several species of anaerobes.61–64 The bacterial pathogens implicated in CRS are distinct from those most often identified in acute bacterial rhinosinusitis (ABRS). Just as CRS is thought to be multifactorial, incidence of ABRS may be higher in the setting of particular environmental exposures, allergies, smoking, ciliary impairment, sinonasal anatomic variations, and transient bacterial infections.65–67 However, the most often identified bacteria cultured in patients with ABRS are S. pneumoniae, H. influenzae, S. pyrogenes, M. catarrhalis, and S aureus.68 Patients with refractory CRS often demonstrate both cellular and humoral immune dysfunction, as described above,69 and in these cases the most commonly isolated pathogens are S. aureus and Pseudomonas aeruginosa.63,64 Patients with CRS have a higher incidence of antibiotic-resistance; this observation is especially notable in patients undergoing revision endoscopic sinus surgery (ESS) when compared to patients undergoing surgery for the first time. 70 True (ie, physiologic) antibiotic resistance in a dysbiotic community is difficult to assess, but may be even higher than that documented in culture studies of antibiotic susceptibility.71 In the wake of emerging molecular and imaging techniques, several groups have interrogated the functional role of biofilms in CRS.72,73,74 As the field embraces molecular identification techniques, it has become apparent that conventional culture methods are not on par with newer culture-independent techniques. Current data suggest that molecular techniques such as 16S gene sequencing demonstrate greater biodiversity, increased sensitivity, and the ability to identify anaerobic groups with greater specificity when compared to culture.46,47 In practice, Hauser et al. found clinical lab cultures to identify the dominant bacteria only 60% of the time.75 Presumably, identification of the dominant bacteria is the information sought by the clinician, as dysbiosis and decreased diversity in disease points to a single or few dominant organisms out of community balance as the source of pathology. Culture-directed antibiotics are thought to be beneficial if there are signs of active infection such as purulence, or if many antibiotics have already been administered, but the role of cultures is less clear in the absence of this history. As our understanding of CRS etiology evolves, and our capacity to examine the entire microbial community increases, the utility of cultures in CRS requires further evaluation. II. S. aureus and superantigens Staphylococcus aureus and coagulase negative staphylococcus are the most common putative pathogens identified in CRS in North America.68,76 Toxigenic strains of S. aureus are known to be potent disease modifiers capable of disrupting barrier function, invading epithelial cells, modulating immune cells, and promoting polyp formation.77 Staphylococcal superantigenic toxins (SAgs) are small molecules secreted during toxigenic S. aureus infection that crosslink antigen presenting cells and T-cells by simultaneously binding MHC-II and the T-cell receptor. In contrast to conventional antigens, this direct activation of CD4+ and CD8+ T-cells leads to a profound polyclonal immune response, and generation of a Th2 cytokine environment. This promotes eosinophil recruitment, the generation of polyclonal IgE, Treg inhibition, mast cell degranulation, and the activation of M2 macrophages, which collectively result in generation of nasal polyps.78–80 The incidence and awareness of methicillin-resistant S. aereus (MRSA) carriage in the anterior nares has increased in recent time and is known to contribute to surgical site infection and biofilm creation, slow mucosal healing after endoscopic sinus surgery, and result in greater overall morbidity.72 Although S. aureus is a major pathogen inCRS, most species remain methicillin-sensitive (MSSA). 81, 82, 46 Since the colonization of healthy individuals with toxigenic strains of Staphylococcus is surprisingly common, multiple host and microenvironment factors likely contribute to Staphylococcus-implicated effects in CRS. S. aureus, P. aeruginosa, H. influenza, S. pneumonia, M. catarrhalis and Stenotrophomonas are known to establish biofilms on the mucosal surface. S. aureus biofilms in particular have been suggested to promote Th2 cytokines independently of SAgs.83 Biofilm formation with any pathogen occurs in response to selective pressures within the mucosal niche, including antibiotic use, tobacco exposure, and loss of integrity of host epithelial immune barrier. The presence of biofilms is associated with increased disease severity and recalcitrance since antibiotics and host defenses do not efficiently penetrate the dense, 3-dimensional, polysaccharide matrix. Once biofilms are established they are difficult to clear, and serve as a protected reservoir for pathogens to evolve virulence factors and shed planktonic bacteria that can incite acute exacerbations of disease.84 III. P. aeruginosa Pseudomonas aeruginosa infection of the paranasal sinuses is most frequently discovered in immune compromised individuals, patients undergoing revision sinus surgery, or in the setting of ciliary dysfunction such as in cystic fibrosis.85, 86, 87 P. aeruginosa secretes several virulence factors including adhesins, secreted toxins (eg. exoenzyme S and exotoxin A), proteases, effector proteins (eg. elastase and alkaline protease), and pigments (eg. pyocyanin) which act to evade the immune system and disable host cells. Additionally, P. aeruginosa uses active quorum sensing to organize tenacious biofilms at the epithelial surface and further evade host defenses.88 IV. Anaerobes Although advances in molecular sequencing have highlighted the preponderance of anaerobes in the CRS microbiome, in clinical practice, the identification of specific anaerobes is often complicated by limitations of culture techniques. The expansion of anaerobes in subsets of CRS may also result from alterations in the microbial community and local microenvironment.89 30, 9091 Currently, anaerobes pose several challenges to the effective management of CRS. When identified by culture, this often influences antibiotic selection to include anaerobe coverage. In the case of anaerobes such as Bacteriodetes and Fusobacteria, these pathogens are known to share mobile genetic elements that confer antibiotic resistance and increased virulence with other bacteria in the community.92,93 Thus, these anaerobes, when expanded in CRS, may work synergistically with other opportunistic pathogens throughout a bacterial community leading to a pathogenic niche. V. Antibiotic resistance associated with pathogens in CRS study Antibiotic resistance is a natural phenomenon and is increasingly common in patients with refractory CRS.70 Bacteria may acquire antibiotic resistance through random mutation or through exchange of genetic material. As discussed previously, the sinonasal bacterial niche experiences constant selective pressure from neighboring microbes and the host immune defenses. When this bacterial community proliferates in the presence of antibiotics there is additional selective pressure that culls species with the defense mechanisms to withstand antimicrobials in the environment. There are several mobile genetic elements or “r genes” that confer resistance to particular antibiotics and may be passed vertically through generations of bacteria, or horizontally from one bacterium to a neighboring bacterium.94 Since antibiotics used to treat refractory CRS selectively eliminate the most susceptible members of the community, the potential of the microbiome to remain diverse and facilitate pathogen exclusion is reduced, theoretically allowing resistant bacteria to flourish and dominate. Innovative new strategies are needed to combat the rise of resistant superinfections in refractory CRS; perhaps an approach that includes promoting a diverse and healthy microbial community to exclude multi-resistant pathogens will be explored in the future. 4. Emerging research concepts in the microbiome I. Role in mucosal immune function While progress has been made to characterize the bacterial microbiome in health and disease, studies of host-microbe and microbe-microbe interactions occurring within the mucosal niche have only begun. The concept of “pathogen exclusion” refers to the ability of a microbial community to selectively inhibit the dominance of a single pathogen. This can occur actively through direct inhibition of pathogens by commensals within the niche. For example, S. aureus may be inhibited by neighboring microbes including S. epidermidis, Corynebacterium spp., P. aeruginosa, and fungal species.25,95 While the host innate immune defense utilizes several antimicrobial strategies to modulate surface microbes, the bacterial community itself possesses a much more dynamic capacity to shape microbes within the niche through the use of small molecule signals (quorum sensing), utilization of alternative nutrient sources, secreted antimicrobials, and transfer of mobile genetic elements over many generations. Expansion of pathogens may also be limited by passive mechanisms such as competition for limited nutrients and the accumulation of metabolic waste products from indigenous microbes that inhibit pathogenic strains. For example, in the mouse intestine, enterohemorrhagic strains of E. coli (EHEC) are known to compete with non-pathogenic strains for amino acids and other nutrients.96,97 II. Dysbiosis and community dynamics Dysbiosis is the concept that disruption of mutualistic relationships in the microbiome may compromise health and contribute to human disease. If the healthy microbiome exists to serve a function or multiple functions, exogenous stressors that induce a dysbiosis may then influence the functional ability of the core microbiome, creating a temporary susceptibility for transitioning to a diseased state [Figure 2]. It remains unclear whether dysbiosis is a primary trigger that leads to disease pathogenesis and if these community imbalances may be involved in determining severity or duration of disease.48 But, this is certainly an area of interest, as stability and resilience of the microbiome has a degree of interpersonal variance.50,98 III. Determinants of the human microbiome The microbiome appears to be established early in infancy and is necessary for the development of normal mucosal immunity. Several gestational and post-partum environmental factors are known to contribute to the development of a healthy intestinal microbiome including route of birth, gestational age at birth, breastfeeding, exposure to cigarette smoke or antibiotics, household milieu, and socioeconomic status.99 In addition to environmental factors, there is intriguing evidence from twin studies to suggest that microbiome determinants are heritable but that genetics may play only a partial role in microbiome composition. These studies demonstrate that individuals have significant co-variability in the species represented within their gut microbiome, but that specific alleles held between monozygotic and dizygotic twins underlie the heritability of certain bacterial phyla.55,100 Stability refers to the degree of random change with time, and the resilience of a microbial community is defined by its ability to return to baseline following a perturbation in the community. Studies examining the stability and resilience of the sinonasal microbiome are limited, but emerging evidence suggests that the healthy adult microbiome is relatively stable over time.25 Hauser et al. determined that patients undergoing ESS and receiving post-operative oral antibiotics demonstrated significant shift in the microbial composition two weeks post-operatively, but that after six weeks, the bacterial composition returned to the pre-intervention baseline in many subjects.98 The significance of stability and resilience of the sinonasal microbiome is unclear at this point but may be identified as an important factor in disease susceptibility or response to therapy. For instance, we often hear the clinical scenario of a patient with an upper respiratory infection that subsequently develops sinus disease. If the native microbiota are not a stable and resilient population to begin with, then this perturbation could lead to prolonged shifts in the microbiome that then predispose to the formation of sinus disease, akin to the “two-hit” hypothesis of carcinogenesis.101 IV. Restoring a healthy microbiome – Prebiotics and Probiotics With increasing evidence suggesting that dysbiosis participates in the onset or propagation of CRS, and data supporting the concept that microbial richness and diversity contribute to health, much interest has been devoted toward restoration of healthy microbial community function. The most obvious example of these efforts is with the use of probiotics and prebiotics. Probiotics are live bacteria or yeast that are introduced to a microbial community actively functioning to restore balance and/or functionality to the niche. While significant research has been devoted to oral probiotics for the restoration of gastrointestinal microbes, similar trials investigating oral or topical probiotics in CRS are lacking. To date there has been one randomized controlled trial examining the utility of probiotics in CRS, which did not find any benefit.102 Murine models have demonstrated proof-of-concept for the utility of introducing competing bacteria into the sinonasal milieu of infectious CRS using S. epidermidis as a “probiotic” niche-occupier in S.aureus-induced CRS in mice.103 Prebiotics are non-nutritive fibers that, when delivered into a bacterial community, serve as substrates for bacteria and are thought to support the growth of desirable microbes within the community. The uses of prebiotics for nasal health are widely marketed by “nutraceutical” companies. However, there have not been any randomized controlled data that supports the use of prebiotics in the treatment of CRS. Lastly, many of our antibiotics carry broad-spectrum activity and eradicate many species in addition to the target pathogen (i.e., collateral damage), which may actually be unfavorable [Figure 3]. The development of narrow-spectrum antibiotics to selectively eliminate individual taxa from a microbial niche is appealing. In order to rationally move probiotics, prebiotics, and narrow-specrum antibiotics into the rhinology clinic, we must develop a detailed description of the “core microbiome,” or the core microbial components necessary for an individual to maintain sinonasal health. Restoring community function in the healthy state is a primary microbiome goal, just as restoring mucociliary function has been a functional goal in the management of CRS.6 5. A general approach to antibiotic therapy for refractory CRS I. Current guidelines for antibiotics in CRS Although sinonasal saline irrigation and topical corticosteroids are the mainstays of medical therapy for CRS, current clinical practice guidelines do recommend the routine use of antibiotics.104, 105 If purulence is present on examination in a CRS patient, short-term culture directed therapy has been recommended.106, 107, 108 In CRSwNP patients that experience an acute exacerbation or persistent symptoms of CRS, a short course of systemic corticosteroid therapy remains the standard of care with the best early and long-term improvement in polyp scores,109 although doxycycline may have a role in the patients as well110 Current clinical practice guidelines recommend consideration of long-term macrolide therapy such as clarithromycin 250–500mg daily, roxithromycin 150mg daily, or azithromycin 500mg weekly for 3 months in patients with CRSsNP only, as patients with nasal polyps did not benefit from long-term macrolide therapy when compared with placebo.111, 112, 104 The routine use of antibiotics in CRS when acute infection is not present has been recently called into question, given the cost and potential harms of antibiotic use when the effectiveness of this therapy is not clear.113 II. Bacteria: good, bad, or just there? Increasing data suggest that diverse bacterial communities are important for maintaining human health, although much of our current CRS therapy is non-specific and potentially eradicates large groups of both commensal and pathogenic bacteria. Recent culture-independent study has demonstrated that bacteria are present in roughly similar quantities in both healthy and diseased states.29 Certainly, it is apparent that bacteria may function in a beneficial or detrimental manner to the host. However, it appears that much of the bacteria present are simply there, without a known role or function. It has also been noted that pathogens may be present in the healthy state, although in low abundance and housed within a richer and diverse population of other microbes.26 At this point in time it is challenging to determine if bacteria detected in disease always necessitates a therapeutic strategy for its eradication. Differences in microbial communities identified in recent CRS studies very well may be associations discovered due to confounding factors such as the extensive and prolonged therapies used for the disease, namely antibiotics. As a result, a more rational approach to judicious use of antibiotic therapy is needed in CRS. Research reported in this publication was supported by the National Institute On Deafness And Other Communication Disorders of the National Institutes of Health under Award Numbers K23DC014747 (V.R.R.) and T32DC01228003 (T.W.V). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. Figure 1 Potential factors influencing the sinus microbiome. Data from Turnbaugh PJ, Ley RE, Hamady M, et al. The human microbiome project. Nature 2007;449(7164):804–10. Figure 2 Significant perturbations of the healthy microbiome state can result in a degraded state that is susceptible to disease. Once degraded and/or diseased, a goal may be to restore the rich and diverse healthy state. Data from Lozupone CA, Stombaugh JI, Gordon JI, et al. Diversity, stability and resilience of the human gut microbiota. Nature 2012;489(7415):220–30. Figure 3 Antibiotic use, coupled with subject-specific resilience, may potentially result in a degraded microbiome. Synopsis Bacterial pathogens and microbiome alterations can contribute to the initiation and propagation of mucosal inflammation in chronic rhinosinusitis (CRS). In this article, we review the clinical and research implications of key pathogens, discuss the role of the microbiome, and connect bacteria to mechanisms of mucosal immunity relevant in CRS. Key Points The sinuses are not sterile. A population of bacteria are present in both health and disease, roughly in the same overall abundance, but qualitatively different in its makeup. As of yet, no single bacterial species or set of species has been definitively shown to be protective or causative in CRS. The overall function of the bacterial community may be most important, rather than the presence of absence of a single pathogen. Therapies used to treat CRS may induce microbiome alterations. Further research is indicated and required in this exciting field. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. Disclosure Statement: The authors have no financial conflicts of interest to disclose. References 1 Bachert C Gevaert P Holtappels G Johansson SG van Cauwenberge P Total and specific IgE in nasal polyps is related to local eosinophilic inflammation J Allergy Clin Immunol 2001 107 4 607 614 doi:10.1067/mai.2001.112374 11295647 2 Pawankar R Nonaka M Inflammatory mechanisms and remodeling in chronic rhinosinusitis and nasal polyps Curr Allergy Asthma Rep 2007 7 3 202 208 17448332 3 Kern RC Conley DB Walsh W Perspectives on the etiology of chronic rhinosinusitis: an immune barrier hypothesis am j rhinol 2008 22 6 549 559 doi:10.2500/ajr.2008.22.3228 18786300 4 Knowles MR Boucher RC Mucus clearance as a primary innate defense mechanism for mammalian airways J Clin Invest 2002 109 5 571 577 doi:10.1172/JCI15217 11877463 5 Peters AT Kato A Zhang N Evidence for altered activity of the IL-6 pathway in chronic rhinosinusitis with nasal polyps J Allergy Clin Immunol 2010 125 2 397 403.e10 doi:10.1016/j.jaci.2009.10.072 20159251 6 Gudis D Zhao K-Q Cohen NA Acquired cilia dysfunction in chronic rhinosinusitis am j rhinol allergy 2012 26 1 1 6 doi:10.2500/ajra.2012.26.3716 22391065 7 Hadfield PJ Rowe-Jones JM Mackay IS The prevalence of nasal polyps in adults with cystic fibrosis Clin Otolaryngol Allied Sci 2000 25 1 19 22 10764232 8 Bush A Chodhari R Collins N Primary ciliary dyskinesia: current state of the art Arch Dis Child 2007 92 12 1136 1140 doi:10.1136/adc.2006.096958 17634184 9 Ooi EH Psaltis AJ Witterick IJ Wormald P-J Innate immunity Otolaryngologic Clinics of North America 2010 43 3 473 87 vii doi:10.1016/j.otc.2010.02.020 20525504 10 Claeys S De Belder T Holtappels G Human beta-defensins and toll-like receptors in the upper airway Allergy 2003 58 8 748 753 12859553 11 Ramanathan M Lee W-K Spannhake EW Lane AP Th2 cytokines associated with chronic rhinosinusitis with polyps down-regulate the antimicrobial immune function of human sinonasal epithelial cells am j rhinol 2008 22 2 115 121 doi:10.2500/ajr.2008.22.3136 18416964 12 Vandermeer J Sha Q Lane AP Schleimer RP Innate immunity of the sinonasal cavity: expression of messenger RNA for complement cascade components and toll-like receptors Arch Otolaryngol Head Neck Surg 2004 130 12 1374 1380 doi:10.1001/archotol.130.12.1374 15611395 13 Kato A Schleimer RP Beyond inflammation: airway epithelial cells are at the interface of innate and adaptive immunity Curr Opin Immunol 2007 19 6 711 720 doi:10.1016/j.coi.2007.08.004 17928212 14 Dong Z Yang Z Wang C Expression of TLR2 and TLR4 messenger RNA in the epithelial cells of the nasal airway am j rhinol 2005 19 3 236 239 16011126 15 Detwiller KY Smith TL Alt JA Trune DR Mace JC Sautter NB Differential expression of innate immunity genes in chronic rhinosinusitis am j rhinol allergy 2014 28 5 374 377 doi:10.2500/ajra.2014.28.4082 25198021 16 Hertz CJ Kiertscher SM Godowski PJ Microbial lipopeptides stimulate dendritic cell maturation via Toll-like receptor 2 J Immunol 2001 166 4 2444 2450 doi:10.4049/jimmunol.166.4.2444 11160304 17 Schwandner R Dziarski R Wesche H Rothe M Kirschning CJ Peptidoglycan- and lipoteichoic acid-induced cell activation is mediated by toll-like receptor 2 J Biol Chem 1999 274 25 17406 17409 10364168 18 Ramanathan M Lee W-K Dubin MG Lin S Spannhake EW Lane AP Sinonasal epithelial cell expression of toll-like receptor 9 is decreased in chronic rhinosinusitis with polyps am j rhinol 2007 21 1 110 116 doi:10.2500/ajr.2007.21.2997 17283572 19 Kim JH Kim GE Cho GS Natural killer cells from patients with chronic rhinosinusitis have impaired effector functions PLoS ONE 2013 8 10 e77177 doi:10.1371/journal.pone.0077177 24204766 20 Blair C Naclerio RM Yu X Thompson K Sperling A Role of type 1 T helper cells in the resolution of acute Streptococcus pneumoniae sinusitis: a mouse model J Infect Dis 2005 192 7 1237 1244 doi:10.1086/444544 16136467 21 Hamilos DL Antigen presenting cells Immunol Res 1989 8 2 98 117 2659691 22 Broere F Apasov SG Sitkovsky MV Eden W Nijkamp FP Parnham JM A2 T cell subsets and T cell-mediated immunity Principles of Immunopharmacology: 3rd revised and extended edition 2011 15 27 Birkhäuser Basel Basel doi:10.1007/978-3-0346-0136-8_2 23 Martinez FO Helming L Gordon S Alternative activation of macrophages: an immunologic functional perspective Annu Rev Immunol 2009 27 451 483 doi:10.1146/annurev.immunol.021908.132532 19105661 24 Krysko O Holtappels G Zhang N Alternatively activated macrophages and impaired phagocytosis of S. aureus in chronic rhinosinusitis Allergy 2011 66 3 396 403 doi:10.1111/j.1398-9995.2010.02498.x 20973804 25 Yan M Pamp SJ Fukuyama J Nasal microenvironments and interspecific interactions influence nasal microbiota complexity and S. aureus carriage Cell Host & Microbe 2013 14 6 631 640 doi:10.1016/j.chom.2013.11.005 24331461 26 Ramakrishnan VR Feazel LM Gitomer SA Ir D Robertson CE Frank DN The microbiome of the middle meatus in healthy adults PLoS ONE 2013 8 12 e85507 doi:10.1371/journal.pone.0085507 24386477 27 Tabas I Glass CK Anti-Inflammatory Therapy in Chronic Disease: Challenges and Opportunities Science 2013 339 6116 166 172 doi:10.1126/science.1230720 23307734 28 Lozupone CA Stombaugh JI Gordon JI Jansson JK Knight R Diversity, stability and resilience of the human gut microbiota Nature 2012 489 7415 220 230 doi:10.1038/nature11550 22972295 29 Ramakrishnan VR Feazel LM Abrass LJ Frank DN Prevalence and abundance of Staphylococcus aureus in the middle meatus of patients with chronic rhinosinusitis, nasal polyps, and asthma Int Forum Allergy Rhinol 2013 3 4 267 271 doi:10.1002/alr.21101 23055295 30 Ramakrishnan VR Hauser LJ Feazel LM Ir D Robertson CE Frank DN Sinus microbiota varies among chronic rhinosinusitis phenotypes and predicts surgical outcome J Allergy Clin Immunol 2015 136 2 334 42.e1 doi:10.1016/j.jaci.2015.02.008 25819063 31 Turnbaugh PJ Ley RE Hamady M Fraser-Liggett CM Knight R Gordon JI The human microbiome project Nature 2007 449 7164 804 810 doi:10.1038/nature06244 17943116 32 Erb-Downward JR Thompson DL Han MK Analysis of the lung microbiome in the “healthy” smoker and in COPD PLoS ONE 2011 6 2 e16384 doi:10.1371/journal.pone.0016384 21364979 33 Abreu NA Nagalingam NA Song Y Sinus microbiome diversity depletion and Corynebacterium tuberculostearicum enrichment mediates rhinosinusitis Sci Transl Med 2012 4 151 151ra124 doi:10.1126/scitranslmed.3003783 34 McLoughlin RM Mills KHG Influence of gastrointestinal commensal bacteria on the immune responses that mediate allergy and asthma J Allergy Clin Immunol 2011 127 5 1097 107 quiz1108–9. doi:10.1016/j.jaci.2011.02.012 21420159 35 Bisgaard H Hermansen MN Buchvald F Childhood asthma after bacterial colonization of the airway in neonates N Engl J Med 2007 357 15 1487 1495 doi:10.1056/NEJMoa052632 17928596 36 Ivanov II Atarashi K Manel N Induction of intestinal Th17 cells by segmented filamentous bacteria Cell 2009 139 3 485 498 doi:10.1016/j.cell.2009.09.033 19836068 37 Ivanov II Frutos R de L Manel N Specific microbiota direct the differentiation of IL-17-producing T-helper cells in the mucosa of the small intestine Cell Host & Microbe 2008 4 4 337 349 doi:10.1016/j.chom.2008.09.009 18854238 38 Worbs T Bode U Yan S Oral tolerance originates in the intestinal immune system and relies on antigen carriage by dendritic cells J Exp Med 2006 203 3 519 527 doi:10.1084/jem.20052016 16533884 39 Atarashi K Tanoue T Shima T Induction of colonic regulatory T cells by indigenous Clostridium species Science 2011 331 6015 337 341 doi:10.1126/science.1198469 21205640 40 Sasama J Sherris DA Shin S-H Kephart GM Kern EB Ponikau JU New paradigm for the roles of fungi and eosinophils in chronic rhinosinusitis Current Opinion in Otolaryngology & Head and Neck Surgery 2005 13 1 2 8 15654207 41 Boase S Foreman A Cleland E The microbiome of chronic rhinosinusitis: culture, molecular diagnostics and biofilm detection BMC Infect Dis 2013 13 210 doi:10.1186/1471-2334-13-210 23656607 42 Aurora R Chatterjee D Hentzleman J Prasad G Sindwani R Sanford T Contrasting the microbiomes from healthy volunteers and patients with chronic rhinosinusitis JAMA Otolaryngol Head Neck Surg 2013 139 12 1328 1338 doi:10.1001/jamaoto.2013.5465 24177790 43 Cleland EJ Bassiouni A Bassioni A The fungal microbiome in chronic rhinosinusitis: richness, diversity, postoperative changes and patient outcomes Int Forum Allergy Rhinol 2014 4 4 259 265 doi:10.1002/alr.21297 24500871 44 Allen EK Koeppel AF Hendley JO Turner SD Winther B Sale MM Characterization of the nasopharyngeal microbiota in health and during rhinovirus challenge Microbiome 2014 2 22 doi:10.1186/2049-2618-2-22 25028608 45 Wilson MT Hamilos DL The Nasal and Sinus Microbiome in Health and Disease Curr Allergy Asthma Rep 2014 14 12 485 doi:10.1007/s11882-014-0485-x 25342392 46 Feazel LM Robertson CE Ramakrishnan VR Frank DN Microbiome complexity and Staphylococcus aureus in chronic rhinosinusitis The Laryngoscope 2012 122 2 467 472 doi:10.1002/lary.22398 22253013 47 Stephenson M-F Mfuna L Dowd SE Molecular characterization of the polymicrobial flora in chronic rhinosinusitis J Otolaryngol Head Neck Surg 2010 39 2 182 187 20211106 48 Frank DN Zhu W Sartor RB Li E Investigating the biological and clinical significance of human dysbioses Trends in Microbiology 2011 19 9 427 434 doi:10.1016/j.tim.2011.06.005 21775143 49 Human Microbiome Project Consortium A framework for human microbiome research Nature 2012 486 7402 215 221 doi:10.1038/nature11209 22699610 50 Human Microbiome Project Consortium Structure, function and diversity of the healthy human microbiome Nature 2012 486 7402 207 214 doi:10.1038/nature11234 22699609 51 Ding T Schloss PD Dynamics and associations of microbial community types across the human body Nature 2014 509 7500 357 360 doi:10.1038/nature13178 24739969 52 Reshef L Kovacs A Ofer A Pouch Inflammation Is Associated With a Decrease in Specific Bacterial Taxa Gastroenterology 2015 149 3 718 727 doi:10.1053/j.gastro.2015.05.041 26026389 53 Bassis CM Erb-Downward JR Dickson RP Analysis of the upper respiratory tract microbiotas as the source of the lung and gastric microbiotas in healthy individuals mBio 2015 6 2 e00037 doi:10.1128/mBio.00037-15 25736890 54 Choi EB Hong SW Kim DK Decreased diversity of nasal microbiota and their secreted extracellular vesicles in patients with chronic rhinosinusitis based on a metagenomic analysis Allergy 2014 69 4 517 526 doi:10.1111/all.12374 24611950 55 Turnbaugh PJ Hamady M Yatsunenko T A core gut microbiome in obese and lean twins Nature 2009 457 7228 480 484 doi:10.1038/nature07540 19043404 56 Ott SJ Musfeldt M Wenderoth DF Reduction in diversity of the colonic mucosa associated bacterial microflora in patients with active inflammatory bowel disease Gut 2004 53 5 685 693 doi:10.1136/gut.2003.025403 15082587 57 Scher JU Ubeda C Artacho A Decreased bacterial diversity characterizes the altered gut microbiota in patients with psoriatic arthritis, resembling dysbiosis in inflammatory bowel disease Arthritis Rheumatol 2015 67 1 128 139 doi:10.1002/art.38892 25319745 58 Huang EY Inoue T Leone VA Using corticosteroids to reshape the gut microbiome: implications for inflammatory bowel diseases Inflamm Bowel Dis 2015 21 5 963 972 doi:10.1097/MIB.0000000000000332 25738379 59 Modi SR Collins JJ Relman DA Antibiotics and the gut microbiota J Clin Invest 2014 124 10 4212 4218 doi:10.1172/JCI72333 25271726 60 Ramakrishnan VR Frank DN Impact of cigarette smoking on the middle meatus microbiome in health and chronic rhinosinusitis Int Forum Allergy Rhinol 2015 5 11 981 989 doi:10.1002/alr.21626 26272413 61 Brook I Frazier EH Correlation between microbiology and previous sinus surgery in patients with chronic maxillary sinusitis Ann Otol Rhinol Laryngol 2001 110 2 148 151 11219522 62 Finegold SM Flynn MJ Rose FV Bacteriologic findings associated with chronic bacterial maxillary sinusitis in adults Clin Infect Dis 2002 35 4 428 433 doi:10.1086/341899 12145727 63 Nadel DM Lanza DC Kennedy DW Endoscopically guided cultures in chronic sinusitis am j rhinol 1998 12 4 233 241 9740915 64 Cleland EJ Bassiouni A Wormald P-J The bacteriology of chronic rhinosinusitis and the pre-eminence of Staphylococcus aureus in revision patients Int Forum Allergy Rhinol 2013 3 8 642 646 doi:10.1002/alr.21159 23468020 65 Sande MA Gwaltney JM Acute community-acquired bacterial sinusitis: continuing challenges and current management Clin Infect Dis 2004 39 Suppl 3 S151 S158 doi:10.1086/421353 15546110 66 Skoner DP Complications of allergic rhinitis J Allergy Clin Immunol 2000 105 6 Pt 2 S605 S609 doi:10.1067/mai.2000.106150 10856165 67 Alho O-P Nasal airflow, mucociliary clearance, and sinus functioning during viral colds: effects of allergic rhinitis and susceptibility to recurrent sinusitis am j rhinol 2004 18 6 349 355 15706980 68 Fokkens WJ Lund VJ Mullol J European Position Paper on Rhinosinusitis and Nasal Polyps 2012 Rhinol Suppl 2012 23 1 298 69 Chee L Graham SM Carothers DG Ballas ZK Immune dysfunction in refractory sinusitis in a tertiary care setting The Laryngoscope 2001 111 2 233 235 doi:10.1097/00005537-200102000-00008 11210866 70 Kingdom TT Swain RE The microbiology and antimicrobial resistance patterns in chronic rhinosinusitis Am J Otolaryngol 2004 25 5 323 328 doi:10.1016/j.amjoto.2004.03.003 15334396 71 Estrela S Whiteley M Brown SP The demographic determinants of human microbiome health Trends in Microbiology 2015 23 3 134 141 doi:10.1016/j.tim.2014.11.005 25500524 72 Suh JD Ramakrishnan V Palmer JN Biofilms Otolaryngologic Clinics of North America 2010 43 3 521 30 viii doi:10.1016/j.otc.2010.02.010 20525507 73 Psaltis AJ Ha KR Beule AG Tan LW Wormald P-J Confocal scanning laser microscopy evidence of biofilms in patients with chronic rhinosinusitis The Laryngoscope 2007 117 7 1302 1306 doi:10.1097/MLG.0b013e31806009b0 17603329 74 Oncel S Pinar E Sener G Calli C Karagoz U Evaluation of bacterial biofilms in chronic rhinosinusitis J Otolaryngol Head Neck Surg 2010 39 1 52 55 20122345 75 Hauser LJ Feazel LM Ir D Sinus culture poorly predicts resident microbiota Int Forum Allergy Rhinol 2015 5 1 3 9 doi:10.1002/alr.21428 25278448 76 Larson DA Han JK Microbiology of sinusitis: does allergy or endoscopic sinus surgery affect the microbiologic flora? Current Opinion in Otolaryngology & Head and Neck Surgery 2011 19 3 199 203 doi:10.1097/MOO.0b013e328344f67a 21358332 77 Bachert C Zhang N Patou J van Zele T Gevaert P Role of staphylococcal superantigens in upper airway disease Curr Opin Allergy Clin Immunol 2008 8 1 34 38 doi:10.1097/ACI.0b013e3282f4178f 18188015 78 Pezato R Świerczyńska-Krępa M Niżankowska-Mogilnicka E Derycke L Bachert C Pérez-Novo CA Role of imbalance of eicosanoid pathways and staphylococcal superantigens in chronic rhinosinusitis Allergy 2012 67 11 1347 1356 doi:10.1111/all.12010 22978320 79 Pérez Novo CA Waeytens A Claeys C Cauwenberge PV Bachert C Staphylococcus aureus enterotoxin B regulates prostaglandin E2 synthesis, growth, and migration in nasal tissue fibroblasts J Infect Dis 2008 197 7 1036 1043 doi:10.1086/528989 18419541 80 Corriveau M-N Zhang N Holtappels G Van Roy N Bachert C Detection of Staphylococcus aureus in nasal tissue with peptide nucleic acid-fluorescence in situ hybridization am j rhinol allergy 2009 23 5 461 465 doi:10.2500/ajra.2009.23.3367 19807976 81 Parvez N Jinadatha C Fader R Universal MRSA nasal surveillance: characterization of outcomes at a tertiary care center and implications for infection control South Med J 2010 103 11 1084 1091 doi:10.1097/SMJ.0b013e3181f69235 20926991 82 Jervis-Bardy J Foreman A Field J Wormald P-J Impaired mucosal healing and infection associated with Staphylococcus aureus after endoscopic sinus surgery am j rhinol allergy 2009 23 5 549 552 doi:10.2500/ajra.2009.23.3366 19807992 83 Foreman A Holtappels G Psaltis AJ Adaptive immune responses in Staphylococcus aureus biofilm-associated chronic rhinosinusitis Allergy 2011 66 11 1449 1456 doi:10.1111/j.1398-9995.2011.02678.x 21834937 84 Lebeaux D Ghigo J-M Beloin C Biofilm-related infections: bridging the gap between clinical management and fundamental aspects of recalcitrance toward antibiotics Microbiol Mol Biol Rev 2014 78 3 510 543 doi:10.1128/MMBR.00013-14 25184564 85 Psaltis AJ Weitzel EK Ha KR Wormald P-J The effect of bacterial biofilms on post–sinus surgical outcomes am j rhinol 2008 22 1 1 6 doi:10.2500/ajr.2008.22.3119 18284851 86 Sabini P Josephson GD Reisacher WR Pincus R The role of endoscopic sinus surgery in patients with acquired immune deficiency syndrome Am J Otolaryngol 1998 19 6 351 356 9839907 87 Godoy JM Godoy AN Ribalta G Largo I Bacterial pattern in chronic sinusitis and cystic fibrosis Otolaryngol Head Neck Surg 2011 145 4 673 676 doi:10.1177/0194599811407279 21521896 88 Hansen SK Rau MH Johansen HK Evolution and diversification of Pseudomonas aeruginosa in the paranasal sinuses of cystic fibrosis children have implications for chronic lung infection ISME J 2012 6 1 31 45 doi:10.1038/ismej.2011.83 21716309 89 Carenfelt C Lundberg C The role of local gas composition in pathogenesis of maxillary sinus empyema Acta Otolaryngol 1978 85 1–2 116 121 343480 90 Jousimies-Somer HR Savolainen S Ylikoski JS Macroscopic purulence, leukocyte counts, and bacterial morphotypes in relation to culture findings for sinus secretions in acute maxillary sinusitis J Clin Microbiol 1988 26 10 1926 1933 3182987 91 Siqueira JF Rôças IN Community as the unit of pathogenicity: an emerging concept as to the microbial pathogenesis of apical periodontitis Oral Surg Oral Med Oral Pathol Oral Radiol Endod 2009 107 6 870 878 doi:10.1016/j.tripleo.2009.01.044 19464660 92 Coyne MJ Zitomersky NL McGuire AM Earl AM Comstock LE Evidence of extensive DNA transfer between bacteroidales species within the human gut mBio 2014 5 3 e01305 e01314 doi:10.1128/mBio.01305-14 24939888 93 McKay TL Ko J Bilalis Y DiRienzo JM Mobile genetic elements of Fusobacterium nucleatum Plasmid 1995 33 1 15 25 doi:10.1006/plas.1995.1003 7753905 94 Davies J Davies D Origins and evolution of antibiotic resistance Microbiol Mol Biol Rev 2010 74 3 417 433 doi:10.1128/MMBR.00016-10 20805405 95 Iwase T Uehara Y Shinji H Staphylococcus epidermidis Esp inhibits Staphylococcus aureus biofilm formation and nasal colonization Nature 2010 465 7296 346 349 doi:10.1038/nature09074 20485435 96 Leatham MP Banerjee S Autieri SM Mercado-Lubo R Conway T Cohen PS Precolonized human commensal Escherichia coli strains serve as a barrier to E. coli O157:H7 growth in the streptomycin-treated mouse intestine Infection and Immunity 2009 77 7 2876 2886 doi:10.1128/IAI.00059-09 19364832 97 Fabich AJ Jones SA Chowdhury FZ Comparison of carbon nutrition for pathogenic and commensal Escherichia coli strains in the mouse intestine Infection and Immunity 2008 76 3 1143 1152 doi:10.1128/IAI.01386-07 18180286 98 Hauser LJ Ir D Kingdom TT Robertson CE Frank DN Ramakrishnan VR Investigation of bacterial repopulation after sinus surgery and perioperative antibiotics Int Forum Allergy Rhinol 2016 6 1 34 40 doi:10.1002/alr.21630 26388320 99 Munyaka PM Khafipour E Ghia J-E External influence of early childhood establishment of gut microbiota and subsequent health implications Front Pediatr 2014 2 109 doi:10.3389/fped.2014.00109 25346925 100 Goodrich JK Waters JL Poole AC Human genetics shape the gut microbiome Cell 2014 159 4 789 799 doi:10.1016/j.cell.2014.09.053 25417156 101 Knudson AG Mutation and cancer: statistical study of retinoblastoma Proc Natl Acad Sci USA 1971 68 4 820 823 5279523 102 Mukerji SS Pynnonen MA Kim HM Singer A Tabor M Terrell JE. Probiotics as adjunctive treatment for chronic rhinosinusitis: a randomized controlled trial Otolaryngol Head Neck Surg 2009 140 2 202 208 doi:10.1016/j.otohns.2008.11.020 19201289 103 Cleland EJ Drilling A Bassiouni A James C Vreugde S Wormald P-J Probiotic manipulation of the chronic rhinosinusitis microbiome Int Forum Allergy Rhinol 2014 4 4 309 314 doi:10.1002/alr.21279 24415658 104 Rudmik L Soler ZM Medical Therapies for Adult Chronic Sinusitis: A Systematic Review JAMA 2015 314 9 926 939 doi:10.1001/jama.2015.7544 26325561 105 Rosenfeld RM Piccirillo JF Chandrasekhar SS Clinical practice guideline (update): adult sinusitis Otolaryngol Head Neck Surg 2015 152 2 Suppl S1 S39 doi:10.1177/0194599815572097 106 Dellamonica P Choutet P Lejeune JM Efficacy and tolerance of cefotiam hexetil in the super-infected chronic sinusitis. A randomized, double-blind study in comparison with cefixime Ann Otolaryngol Chir Cervicofac 1994 111 4 217 222 7726479 107 Legent F Bordure P Beauvillain C Berche P A double-blind comparison of ciprofloxacin and amoxycillin/clavulanic acid in the treatment of chronic sinusitis Chemotherapy 1994 40 Suppl 1 8 15 108 Huck W Reed BD Nielsen RW Cefaclor vs amoxicillin in the treatment of acute, recurrent, and chronic sinusitis Arch Fam Med 1993 2 5 497 503 8118565 109 Martinez-Devesa P Patiar S Oral steroids for nasal polyps Cochrane Database Syst Rev 2011 7 CD005232 doi:10.1002/14651858.CD005232.pub3 21735400 110 van Zele T Gevaert P Holtappels G Oral steroids and doxycycline: two different approaches to treat nasal polyps J Allergy Clin Immunol 2010 125 5 1069 1076.e4 doi:10.1016/j.jaci.2010.02.020 20451040 111 Wallwork B Coman W Mackay-Sim A Greiff L Cervin A A double-blind, randomized, placebo-controlled trial of macrolide in the treatment of chronic rhinosinusitis The Laryngoscope 2006 116 2 189 193 doi:10.1097/01.mlg.0000191560.53555.08 16467702 112 Videler WJ Badia L Harvey RJ Lack of efficacy of long-term, low-dose azithromycin in chronic rhinosinusitis: a randomized controlled trial Allergy 2011 66 11 1457 1468 doi:10.1111/j.1398-9995.2011.02693.x 21884529 113 Soler ZM Oyer SL Kern RC Antimicrobials and chronic rhinosinusitis with or without polyposis in adults: an evidenced-based review with recommendations Int Forum Allergy Rhinol 2013 3 1 31 47 doi:10.1002/alr.21064 22736403

PMC005xxxxxx/PMC5127595.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 9000850 2507 Curr Opin Pediatr Curr. Opin. Pediatr. Current opinion in pediatrics 1040-8703 1531-698X 27653703 5127595 10.1097/MOP.0000000000000421 NIHMS822049 Article Indoor allergen exposure and asthma outcomes Sheehan William J. ab Phipatanakul Wanda ab a Division of Allergy and Immunology, Boston Children’s Hospital, Boston, Massachusetts, USA b Harvard Medical School, Boston, Massachusetts, USA Correspondence to Wanda Phipatanakul, MD, MS, Division of Allergy and Immunology, Boston Children’s Hospital, 300 Longwood Avenue, Boston, MA 02115, USA. Tel: +1 17 355 6117; fax: +1 617 730 0248; wanda.phipatanakul@childrens.harvard.edu 29 10 2016 12 2016 01 12 2017 28 6 772777 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. Purpose of review The aim of the present review is to discuss updates on research regarding the relationship between indoor allergen exposure and childhood asthma with a focus on clinical effects, locations of exposure, and novel treatments. Recent findings Recent data continue to demonstrate that early life sensitization to indoor allergens is a predictor of asthma development later in life. Furthermore, avoidance of exposure to these allergens continues to be important especially given that the vast majority of children with asthma are sensitized to at least one indoor allergen. New research suggests that mouse allergen, more so than cockroach allergen, may be the most relevant urban allergen. Recent evidence reminds us that children are exposed to clinically important levels of indoor allergens in locations away from their home, such as schools and daycare centers. Exposure to increased levels of indoor mold in childhood has been associated with asthma development and exacerbation of current asthma; however, emerging evidence suggests that early exposure to higher fungal diversity may actually be protective for asthma development. Novel treatments have been developed that target TH2 pathways thus decreasing asthmatic responses to allergens. These therapies show promise for the treatment of severe allergic asthma refractory to avoidance strategies and standard therapies. Summary Understanding the relationship between indoor allergens and asthma outcomes is a constantly evolving study of timing, location, and amount of exposure. allergens asthma dust mites molds mouse INTRODUCTION Exposure to aeroallergens is an important factor in the pathogenesis and control of allergic diseases, including asthma. Indoor allergens are of particular importance and principally include house dust mites, pets such as dogs and cats, pests such as cockroach and rodents, and molds. The relative importance of these different allergens varies based on different environmental factors depending on geographic, climatic, socioeconomic, and housing conditions. The measurement of indoor allergen levels in dust and air samples has allowed for the determination of risk levels associated with the development of sensitization and symptomology. In the present review, we summarize recent data regarding indoor allergen exposures and the associations with asthma morbidity. This review will focus on the most common indoor allergens including allergens from dust mite, cat, dog, mouse, cockroach, and molds. A specific focus has been placed on intriguing recent findings including new methods to measure indoor allergens, inner-city exposures, and allergen exposures in schools or daycare centers. In addition, new treatments that effectively disrupt the association between indoor allergen exposure and asthma symptoms will be presented. This review will not discuss exposures to indoor air pollutants or microbial products that are important, but beyond the scope of this topic. OVERVIEW AND EPIDEMIOLOGY Studies indicate that more than 80% of school age children with asthma are sensitized to at least one indoor allergen and that allergic sensitization is a strong predictor of disease persistence in later life [1–3]. In fact, in a cohort of children hospitalized for asthma, 91% were found to be sensitized to at least one indoor allergen [4]. Bjerg et al. [5] have shown that while pollen sensitization is strongly associated with the development of rhinitis, indoor allergen sensitization was more associated with asthma. The timing of sensitization is also an important factor as a recent study demonstrated that aeroallergen sensitization at younger ages was associated with an increased risk of asthma in later childhood [6▪]. NEW WAYS TO MEASURE INDOOR ALLERGEN EXPOSURE AND TO IDENTIFY ALLERGEN SENSITIZATION Indoor allergens are typically measured in dust samples that are collected by either vacuuming settled dust or gathering airborne dust from filtered air within a room. Samples may be collected from multiple sites within a home. After collection, fine dust can then be extracted and tested for quantification of individual allergens. Typical dust sampling has relied on spot checks at a certain time point of vacuuming, which measures exposure levels in that exact location. Tovey et al. [7] recently developed a novel method to measure dust mite exposures in individuals over 24-h periods and found that exposures fluctuate over time and beds may not always be the main site of exposure as general considered. Although home sampling is feasible in the clinical research community, there have not been easy-to-use methods for patients or parents to measure the allergen levels in their homes by themselves. Winn et al. [8] evaluated the use of an in-home test kit that allowed parents of dust mite allergic children the ability to easily and quickly quantify the level of dust mites within the home. This study demonstrated that dust mite levels in the ‘testing’ homes were significantly reduced over time compared with control homes that only received education [8]. The authors speculate that immediate knowledge of dust mite levels motivated the parents to more thoroughly perform dust mite avoidance strategies. Future studies are needed to further evaluate the utility and role of patient-directed or parent-directed home allergen measurements in the care of asthma and allergies. As healthcare trends toward precision medicine where therapies are tailored toward an individual’s exact need, allergic sensitization testing is also heading in this direction. In keeping with this idea of more focused treatments, Chen et al. [9] demonstrated the clinical utility of latent class analysis to identify patterns of allergic sensitization among children. Analyzing these patterns of allergic sensitization revealed additional associations with asthma than the evaluation of a single allergen or total IgE [9]. Identifying patterns of polysensitization may be useful in treatment decisions and prognosis discussion. For example, it was shown that sensitization to cockroach actually reduced the symptom scores in participants with concomitant dust mite and pollen sensitization [10]. Future studies may continue this trend by viewing the whole sensitization profile in a subject instead of focusing on single allergen results. THE KEY INNER-CITY ALLERGEN: COCKROACH VERSUS MOUSE Indoor allergen exposure in inner-city areas has been of particular interest given that children living in urban areas have increased asthma severity, decreased asthma control, and greater healthcare use [11]. In the late 1990s, Rosenstreich et al. [12] found cockroach allergen to be highly detectable in inner-city homes. Furthermore, this study demonstrated that children with asthma who were sensitized and exposed to high levels of cockroach allergen had increased asthma morbidity [12]. This landmark study led to cockroach allergen being deemed the important indoor allergen in urban areas; however, mouse allergen has been gaining recognition in recent years [13]. Mouse allergen has been discovered to be prevalent in inner-city homes [14] with home exposure associated with increased asthma morbidity [15]. Outside the home, a study of urban schools found mouse allergen to be the primary school-based allergen, not cockroach allergen [16]. In support of this, Ahluwalia et al. [17] reported that mouse and not cockroach allergen was the major allergen of relevance in inner-city Baltimore. In a city with high levels of both mouse and cockroach allergens, mouse allergen was more strongly and consistently associated with worse asthma outcomes including acute care visits, decreased spirometry, and bronchodilator reversibility [17]. Supporting these results, a more recent study of urban Baltimore and Boston found that sensitization to mouse allergen, but not cockroach, was an independent risk factor for rhinitis in inner-city children with asthma [18]. Although allergen reduction tactics in the inner city often focus on mouse and cockroach allergen, recent results have led to the suggestion that community-based asthma intervention strategies should prioritize reducing mouse allergen exposure [17]. Admittedly, the predominance and clinical importance of mouse versus cockroach allergen in a particular urban area depends on a variety of factors including weather and building conditions. It is expected that future studies will continue to focus not only on identifying the important inner-city allergen, but also efforts to eliminate exposures and improve asthma outcomes. AWAY FROM HOME: EXPOSURE TO INDOOR ALLERGENS IN OTHER CHILDHOOD SETTINGS, SUCH AS SCHOOL AND DAYCARE CENTERS Early studies on indoor allergen exposures and asthma outcomes focused on the home environment. Although parents can do their best to reduce or eliminate allergen exposures in the home, children spend a large portion of the day in schools or daycare centers. In recent years, the exposure to indoor allergens has been studied in these locations outside the home. Cat and dog allergens seem to be particularly ubiquitous in society because of the easy transfer of these pet allergens. This has been confirmed by data demonstrating that clinically relevant levels of pet allergens are found in homes without these animals [19]. Furthermore, a recent study of teenagers found that a large majority of cat-sensitized individuals did not live in a home with a cat [20▪] suggesting that sensitization occurs because of exposures outside the home or due to allergens tracked into the home. For children, the main location for this possible outside exposure is the school setting. In support of this, Almqvist et al. [21] found that cat allergen was transported into schools and associated with worsening of asthma for students with cat allergy who were subjected to this indirect pet exposure. Indoor allergen exposures in schools are not limited to allergens being tracked in from home, but are sometimes because of allergens intrinsic to the school settings. Permaul et al. [16] noted mouse allergen to be present in high levels in inner-city schools due to infestation within the schools and not passive transfer from students’ homes [22]. For younger children, exposure to dust mites and molds in day care centers was associated with wheezing [23]. Likewise, researchers in Denmark found that high classroom dampness was associated with increased wheezing and decreased spirometry in exposed students [24]. Finally, mold exposure in school classrooms was shown to be significantly associated with current asthma symptoms and asthma symptoms that improved over holidays and weekends when the students were out of school [25▪▪]. This pattern of symptomatology is analogous to workplace exposures and health outcomes for adults [26]. These findings have prompted suggestions that governmental health policies should consider environmental interventions in schools to improve childhood asthma. It is expected that future research will focus on intervention tactics in schools to reduce allergen exposures and improve students’ respiratory symptoms. These interventions have been successful in homes [27] and now need to be trialed in the school setting. MOLD EXPOSURE AND ASTHMA Molds grow best in warm and moist environments within the home including basements, windows frames, and bathrooms. Aspergillus and Penicillium species are generally regarded as the predominant indoor molds, and Alternaria is important in both indoor and outdoor environments. Recent literature in this field has focused on the association of mold exposure in early childhood with the development of atopic conditions including asthma later in childhood. Sharpe et al. [28▪] recently discovered that reporting mildew odor in the home was associated with an increased risk of childhood asthma. Supporting this, a birth cohort study demonstrated that exposure to moisture damage and visible mold in early infancy were associated with asthma development [29]. More specific measurement of mold exposure yielded similar findings as Oluwole et al. [30] documented that mold levels in dust samples from both play area floors and bedroom mattresses were significantly associated with current asthma. In a meta-analysis, increased levels of Penicillium, Aspergillus, Cladosporium, and Alternaria species were associated with increased exacerbation of current asthma symptoms in children [31]. New research has now focused on the diversity of mold exposures and the associated clinical outcomes because there is growing evidence that exposure to a wide diversity of mold species may actually be protective. For example, Tischer et al. [32▪▪] found that exposure to higher fungal diversity shortly after birth was associated with a decreased risk of developing wheezing and aeroallergen sensitization in later childhood. Future work should continue to evaluate the role of fungal diversity on the development of atopic conditions. NEW TREATMENTS TO DISRUPT THE ALLERGEN–ASTHMA ASSOCIATION Allergen avoidance is considered the first line of treatment for patients with indoor allergen sensitivities and asthma. In general practice, allergen avoidance should be tailored based on an individual’s-specific allergen sensitivities and known environmental exposures. In research, most of the studies in which environmental controls have been effective implemented a multifaceted approach. For example, Morgan et al. [27] demonstrated that a comprehensive environmental intervention in the home was able to reduce levels of indoor allergens and resulted in improved asthma outcomes in children. Certainly, future studies will continue to focus on efficient and cost-effective strategies to reduce allergen exposures in homes and locations outside the home. After allergen avoidance, the mainstays of initial medication therapy include antihistamines, corticosteroids, and leukotriene receptor antagonists. In recent years, novel therapies targeting TH2 pathways have been successful in decreasing the asthmatic response to allergen triggers. These medications provide hope for the future for patients with allergies and asthma refractory to environmental avoidance strategies and standard medications. Most notably, Krug et al. [33▪▪] recently reported promising outcomes in individuals taking a DNAzyme that is able to cleave and inactivate GATA3, an important transcription factor of the TH2 pathway. In patients with allergic asthma, treatment with this DNAzyme was able to significantly attenuate both late and early asthmatic responses after allergen challenge [33▪▪]. Similar disruption of the allergen exposure and asthma pathway was seen after treatment with a monoclonal antibody against thymic stromal lymphopoietin, which is a cytokine thought to be important in initiating allergic inflammation [34]. Other new therapeutic approaches targeting specific cytokines in the TH2 inflammatory pathway have shown promising results for improving asthma outcomes [35–37]. Finally, recent advances in dust mite allergen immunotherapy have shown promising results. Virchow et al. [38] evaluated individuals with dust mite allergy-related asthma and found treatment with novel dust mite sublingual allergen immunotherapy, when compared with placebo, was shown to improve asthma control. A very important secondary finding in this study was that there were no differences in outcomes detected between polysensitized individuals and those monosensitized to dust mite, demonstrating the clinical efficacy of single allergen immunotherapy even in patients with sensitization to multiple allergens [38]. Zolkipli et al. [39▪] used dust mite allergen oral immunotherapy prophylactically in high-risk, but not yet sensitized, infants less than 12 months of age and found that this early life treatment reduced the development of sensitization to any allergen. The use of single allergen immunotherapy is often the concern regarding oral immunotherapy; however, the targeting of treatment with dust mite allergen makes sense given that this is known to be the most frequent indoor allergen associated with asthma worldwide. In fact, recent laboratory data demonstrated that house dust mite exposure selectively increased the proliferation of bronchial smooth muscle in patients with severe asthma [40]. It is expected that future research will continue to develop oral immunotherapy with a particular focus on dust mites. CONCLUSION Exposure to indoor aeroallergens is an important factor in the pathogenesis and control of childhood asthma. Although previous research focused on allergen exposures in homes, emerging research is shifting toward the importance of allergen exposures in schools and daycare centers. Furthermore, new techniques are being developed to more easily and efficiently measure levels of allergen exposure. In urban areas, mouse allergen has become an important exposure and may replace cockroach allergen as the most relevant indoor allergen exposure. Studies continue to demonstrate the association between mold exposure in early childhood and asthma development and symptoms. Although avoidance strategies are always the first line of defense in the treatment of allergies and asthma, novel therapeutics are being developed for patients with severe asthma triggered by allergens. Financial support and sponsorship This work was supported in part by grants K24AI106822, K23AI104780, U10HL109172, U10HL098102, and U01AI110397 from the National Institutes of Health. This work was also conducted in part by support from the American Lung Association/American Academy of Allergy, Asthma, and Immunology Respiratory Diseases Faculty Award, and Deborah Munroe Noonan Memorial Award. There was also support from Harvard Catalyst/The Harvard Clinical and Translational Science Center (NIH Award # 8UL1TR000170) and financial contributions from Harvard University and its affiliated academic healthcare centers. The content is solely the responsibility of the authors and does not necessarily represent the official views of Harvard Catalyst, Harvard University, and its affiliated academic healthcare centers, the National Center for Research Resources, or the National Institutes of Health CTSU PI (Nagler). KEY POINTS Indoor allergen sensitization is common in young children with asthma, and this sensitization is a predictor of asthma persistence later in life. Cockroach allergen has traditionally been considered the primary inner-city indoor allergen; however, recent data suggest that mouse allergen may be the most relevant urban allergen exposure. In addition to home exposures, children are exposed to a wide variety of clinically relevant allergens in schools and daycare centers. In general, exposure to mold in homes and schools has been associated with asthma development and exacerbation of current asthma; however, emerging evidence suggests that early exposure to higher fungal diversity may actually be protective for asthma development. In recent years, novel therapies targeting TH2 pathways that have been successful in decreasing asthmatic responses to allergens show promise for the treatment of severe allergic asthma refractory to avoidance strategies and standard therapies. Conflicts of interest There are no conflicts of interest. REFERENCES AND RECOMMENDED READING Papers of particular interest, published within the annual period of review, have been highlighted as: ▪ of special interest ▪▪ of outstanding interest 1 Sporik R Holgate ST Platts-Mills TA Cogswell JJ Exposure to house-dust mite allergen (Der p I) and the development of asthma in childhood. A prospective study New Engl J Med 1990 323 502 507 2377175 2 Illi S von Mutius E Lau S Perennial allergen sensitisation early in life and chronic asthma in children: a birth cohort study Lancet 2006 368 763 770 16935687 3 Eggleston PA Rosenstreich D Lynn H Relationship of indoor allergen exposure to skin test sensitivity in inner-city children with asthma J Allergy Clin Immunol 1998 102 4 Pt 1 563 570 9802363 4 Beck AF Huang B Kercsmar CM Allergen sensitization profiles in a population-based cohort of children hospitalized for asthma Ann Am Thorac Soc 2015 12 376 384 25594255 5 Bjerg A Ekerljung L Eriksson J Increase in pollen sensitization in Swedish adults and protective effect of keeping animals in childhood Clin Exp Allergy 2016 6▪ Rubner FJ Jackson DJ Evans MD Early life rhinovirus wheezing, allergic sensitization, and asthma risk at adolescence J Allergy Clin Immunol 2016 This study demonstrated that aeroallergen sensitization at younger ages was associated with an increased risk of asthma in later childhood 7 Tovey ER Liu-Brennan D Garden FL Time-based measurement of personal mite allergen bioaerosol exposure over 24 hour periods PloS One 2016 11 e0153414 27192200 8 Winn AK Salo PM Klein C Efficacy of an in-home test kit in reducing dust mite allergen levels: results of a randomized controlled pilot study J Asthma 2016 53 133 138 26308287 9 Chen Q Zhong X Acosta L Allergic sensitization patterns identified through latent class analysis among children with and without asthma Ann Allergy Asthma Immunol 2016 116 212 218 26945495 10 He W Jimenez F Martinez H Cockroach sensitization mitigates allergic rhinoconjunctivitis symptom severity in patients allergic to house dust mites and pollen J Allergy Clin Immunol 2015 136 658 666 26026342 11 Szefler SJ Gergen PJ Mitchell H Morgan W Achieving asthma control in the inner city: do the National Institutes of Health Asthma Guidelines really work? J Allergy Clin Immunol 2010 125 521 526 quiz 7–8 20226288 12 Rosenstreich DL Eggleston P Kattan M The role of cockroach allergy and exposure to cockroach allergen in causing morbidity among inner-city children with asthma New Engl J Med 1997 336 1356 1363 9134876 13 Ownby DR Will the real inner-city allergen please stand up? J Allergy Clin Immunol 2013 132 836 837 23978444 14 Phipatanakul W Eggleston PA Wright EC Wood RA Mouse allergen. I. The prevalence of mouse allergen in inner-city homes. The National Cooperative Inner-City Asthma Study J Allergy Clin Immunol 2000 106 1070 1074 11112888 15 Phipatanakul W Litonjua AA Platts-Mills TA Sensitization to mouse allergen and asthma and asthma morbidity among women in Boston J Allergy Clin Immunol 2007 120 954 956 17590423 16 Permaul P Hoffman E Fu C Allergens in urban schools and homes of children with asthma Pediatr Allergy Immunol 2012 23 543 549 22672325 17 Ahluwalia SK Peng RD Breysse PN Mouse allergen is the major allergen of public health relevance in Baltimore City J Allergy Clin Immunol 2013 132 830 835 e1 2 23810154 18 Sedaghat AR Matsui EC Baxi SN Mouse sensitivity is an independent risk factor for rhinitis in children with asthma J Allergy Clin Immunol Pract 2016 4 82 88 e1 26441149 19 Arbes SJ Jr Cohn RD Yin M Dog allergen (Can f 1) and cat allergen (Fel d 1) in US homes: results from the National Survey of Lead and Allergens in Housing J Allergy Clin Immunol 2004 114 111 117 15241352 20▪ Perzanowski MS Ronmark E James HR Relevance of specific IgE antibody titer to the prevalence, severity and persistence of asthma among 19-year-olds in Northern Sweden J Allergy Clin Immunol 2016 Article in Press. This study uncovered that a majority of cat-sensitized individuals did not live in a home with a cat indicating that sensitization to pets can occur without obvious exposure 21 Almqvist C Wickman M Perfetti L Worsening of asthma in children allergic to cats, after indirect exposure to cat at school Am J Respir Crit Care Med 2001 163 3 Pt 1 694 698 11254526 22 Permaul P Sheehan WJ Baxi SN Predictors of indoor exposure to mouse allergen in inner-city elementary schools Ann Allergy Asthma Immunol 2013 111 299 301 e1 24054369 23 Carreiro-Martins P Papoila AL Caires I Effect of indoor air quality of day care centers in children with different predisposition for asthma Pediatr Allergy Immunol 2016 27 299 306 26663443 24 Holst G Host A Doekes G Allergy and respiratory health effects of dampness and dampness-related agents in schools and homes: a cross-sectional study in Danish pupils Indoor Air 2015 25▪▪ Chen CH Chao HJ Chan CC Current asthma in schoolchildren is related to fungal spores in classrooms Chest 2014 146 123 134 This study in school children demonstrated that mold exposures within the school were associated with asthma symptoms that improved over holidays and weekends. These results support the theory that school exposures in children are analogous to workplace exposures in adults 24676386 26 Lim FL Hashim Z Than LT Asthma, airway symptoms and rhinitis in Office Workers in Malaysia: associations with house dust mite (HDM) allergy, cat allergy and levels of house dust mite allergens in office dust PloS One 2015 10 e0124905 25923543 27 Morgan WJ Crain EF Gruchalla RS Results of a home-based environmental intervention among urban children with asthma New Engl J Med 2004 351 1068 1080 15356304 28▪ Sharpe RA Thornton CR Tyrrell J Variable risk of atopic disease due to indoor fungal exposure in NHANES Clin Exp Allergy 2015 45 1566 1578 This study presented results from a birth cohort where exposure to visible mold and water damage in early infancy were associated with asthma development 25845975 29 Karvonen AM Hyvarinen A Korppi M Moisture damage and asthma: a birth cohort study Pediatrics 2015 135 e598 e606 25687143 30 Oluwole O Kirychuk SP Lawson JA Indoor mold levels and current asthma among school-aged children in Saskatchewan, Canada Indoor Air 2016 4 24 31 Sharpe RA Bearman N Thornton CR Indoor fungal diversity and asthma: a meta-analysis and systematic review of risk factors J Allergy Clin Immunol 2015 135 110 122 25159468 32▪▪ Tischer C Weikl F Probst AJ Urban dust microbiome: impact on later atopy and wheezing Environ Health Perspect 2016 This study found that exposure to higher fungal diversity shortly after birth was associated with a decreased risk of developing wheezing and aeroallergen sensitization. These findings support the theory that an increased diversity of exposure may actually be protective against atopy development 33▪▪ Krug N Hohlfeld JM Kirsten AM Allergen-induced asthmatic responses modified by a GATA3-specific DNAzyme New Engl J Med 2015 372 1987 1995 This study was a randomized, double-blind, placebo-controlled trial, evaluating the effect of treatment with a novel DNA enzyme (DNAzyme) that cleaves and inactivates GATA3, an important transcription factor of the TH2 pathway. The results demonstrated that treatment significantly attenuated both late and early asthmatic responses after allergen provocation 25981191 34 Gauvreau GM O’Byrne PM Boulet LP Effects of an anti-TSLP antibody on allergen-induced asthmatic responses New Engl J Med 2014 370 2102 2110 24846652 35 Wenzel S Ford L Pearlman D Dupilumab in persistent asthma with elevated eosinophil levels New Engl J Med 2013 368 2455 2466 23688323 36 Haldar P Brightling CE Hargadon B Mepolizumab and exacerbations of refractory eosinophilic asthma New Engl J Med 2009 360 973 984 19264686 37 Corren J Lemanske RF Hanania NA Lebrikizumab treatment in adults with asthma New Engl J Med 2011 365 1088 1098 21812663 38 Virchow JC Backer V Kuna P Efficacy of a house dust mite sublingual allergen immunotherapy tablet in adults with allergic asthma: a randomized clinical trial JAMA 2016 315 1715 1725 27115376 39▪ Zolkipli Z Roberts G Cornelius V Randomized controlled trial of primary prevention of atopy using house dust mite allergen oral immunotherapy in early childhood J Allergy Clin Immunol 2015 136 1541 1547 e1 11 High-risk, but not yet sensitized, infants treated with dust mite allergen oral immunotherapy had reduced development of sensitization to any allergen 26073754 40 Trian T Allard B Dupin I House dust mites induce proliferation of severe asthmatic smooth muscle cells via an epithelium-dependent pathway Am J Respir Crit Care Med 2015 191 538 546 25569771

PMC005xxxxxx/PMC5127630.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 101688624 45516 Sci Immunol Sci Immunol Science immunology 2470-9468 27917411 5127630 10.1126/sciimmunol.aaf7153 HHMIMS831435 Article PTPN22 inhibition resets defective human central B cell tolerance Schickel Jean-Nicolas 1 Kuhny Marcel 1 Baldo Alessia 1 Bannock Jason M. 1 Massad Christopher 1 Wang Haowei 1 Katz Nathan 1 Oe Tyler 1 Menard Laurence 1 Soulas-Sprauel Pauline 2 Strowig Till 1* Flavell Richard 13 Meffre Eric 1† 1 Department of Immunobiology, Yale University School of Medicine, New Haven, CT 06511, USA 2 CNRS UPR 3572, Laboratory of Immunopathology and Therapeutic Chemistry/Laboratory of Excellence Medalis, Molecular and Cellular Biology Institute (IBMC), Strasbourg, France 3 Howard Hughes Medical Institute, Chevy Chase, MD 20815–6789, USA † Corresponding author. eric.meffre@yale.edu * Present address: Helmholtz Centre for Infection Research, Braunschweig, Germany. 23 11 2016 22 7 2016 2016 22 7 2017 1 1 aaf7153This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. The 1858T protein tyrosine phosphatase nonreceptor type 22 (PTPN22 T) allele is one of the main risk factors associated with many autoimmune diseases and correlates with a defective removal of developing autoreactive B cells in humans. To determine whether inhibiting PTPN22 favors the elimination of autoreactive B cells, we first demonstrated that the PTPN22 T allele interfered with the establishment of central B cell tolerance using NOD-scid-common γ chain knockout (NSG) mice engrafted with human hematopoietic stem cells expressing this allele. In contrast, the inhibition of either PTPN22 enzymatic activity or its expression by RNA interference restored defective central B cell tolerance in this model. Thus, PTPN22 blockade may represent a therapeutic strategy for the prevention or treatment of autoimmunity. INTRODUCTION Rituximab, an anti-CD20 monoclonal antibody that eliminates B cells, has shown efficacy in type 1 diabetes (T1D), rheumatoid arthritis (RA), and multiple sclerosis (MS) and exposes a role for B cells in promoting autoimmunity (1–3). However, anti–B cell therapy does not reset early B cell tolerance checkpoints defective in T1D likely because these impaired autoreactive B cell counterselection steps may be primary to the development of this autoimmune disease (4). Indeed, asymptomatic individuals carrying the PTPN22 T allele display elevated frequencies of autoreactive B cells in their blood similar to those in patients with T1D, RA, or systemic lupus erythematosus (SLE) (5). This PTPN22 variant harbors a single nucleotide change (cytidine to thymidine) at residue 1858, which results in a single amino acid substitution from arginine to tryptophan at position 620 (620W) of the PTPN22/LYP protein and has been associated with an increased risk for the development of many autoimmune diseases including RA, T1D, and SLE (6). In contrast, the rare loss-of-function 263Q PTPN22 variant was reported to confer protection against SLE and RA, suggesting that decreased PTPN22 phosphatase activity inhibits autoimmunity (7, 8). Here, we aimed to develop an alternative efficient therapy for autoimmune diseases by targeting the intrinsic genetic defects responsible for impaired central B cell tolerance. We therefore assessed whether 620W PTPN22 expression is sufficient to induce defects in central B cell tolerance and whether they could be corrected after inhibiting PTPN22 function. RESULTS Central B cell tolerance is defective in humanized mouse engrafted with hematopoietic stem cells carrying PTPN22 T allele(s) To further study the impact of PTPN22 variants on central B cell tolerance, we engrafted NOD-scid-common γ chain knockout (NSG) immunodeficient mice with CD34+ hematopoietic stem cells (HSCs) isolated from human fetuses that carry (or do not carry) PTPN22 T allele(s) (9–11) (Fig. 1A and table S1). Humanized NSG mice displayed high frequencies of CD45+ human cells detected by flow cytometry about 3 months after engraftment with HSCs, regardless of the presence of PTPN22 T allele(s) (Fig. 1B). Ratios between human CD19+ B andCD3+ T lymphocytes were also similar in NSG mice transplanted with PTPN22 C/C, PTPN22 C/T, or PTPN22 T/T HSCs, demonstrating that the PTPN22 T allele does not affect either B or T cell development (Fig. 1B). Pooled immunoglobulin heavy-chain (IgH) sequence analyses from new emigrant B cells of PTPN22 C/T or T/T NSG mice revealed no consistent differences in IgH variable (VH), diversity (D), or joining (J) gene usage, compared to PTPN22 C/C NSG mice (fig. S1, A to C). However, in contrast to new emigrant B cells of PTPN22 C/C NSG mice, the presence of a PTPN22 T allele favored the usage of different D reading frames encoding hydrophobic residues known to favor self-reactivity and which correlated with an abnormal central B cell tolerance checkpoint (12–14) (fig. S1D). The analyses of antibody reactivity revealed that frequencies of polyreactive clones in splenic CD19+CD27−CD10+IgMhiCD21lo new emigrant B cells from NSG mice transplanted with PTPN22 C/C HSCs isolated from seven distinct fetuses were low and similar to those of new emigrant B cells isolated from the blood of PTPN22 C/C healthy donors (Fig. 1C, fig. S2A, and tables S2 to S8). The low frequencies of new emigrant B cells reactive to human epithelial type 2 (HEp-2) cells and the virtual absence of antinuclear clones in this B cell compartment reveal that central B cell tolerance is established normally in humanized mice in the absence of the PTPN22 T allele (Fig. 1D and fig. S2, B and C). In contrast, we found that new emigrant B cells isolated from the spleen of NSG mice engrafted with PTPN22 C/T or T/T HSCs contained many autoreactive clones expressing polyreactive and HEp-2–reactive antibodies with similar frequencies to those observed in healthy donors carrying PTPN22 T allele(s) (5) (Fig. 1, C and D, fig. S2, A and B, and tables S9 to S11). Indirect immunofluorescence assays with HEp-2 cell–coated slides revealed that the proportions of antinuclear new emigrant B cell in NSG mice engrafted with PTPN22 C/T or T/T HSCs were increased but failed to reach significance (fig. S2C). We conclude that the presence of the PTPN22 T allele in HSCs results in defective central B cell tolerance and the release of large numbers of autoreactive B cells from the bone marrow. 620W PTPN22 overexpression interferes with central B cell tolerance To determine whether B cell–intrinsic expression of 620W PTPN22 phosphatases is sufficient to interfere with the removal of developing autoreactive immature B cells in the bone marrow, we transduced PTPN22 C/C HSCs with lentiviruses expressing green fluorescent protein (GFP) and the 620W PTPN22 autoimmunity-favoring variant, the common 620R or the 263Q loss-of-function PTPN22 enzyme (Fig. 2A). Human CD19+ B cells developed in NSG mice engrafted with HSCs, whether transduced or not with the different lentiviruses, revealed that lentiviral transduction did not alter HSC engraftment or B cell development (fig. S3A). Quantification of lentivirus-encoded PTPN22 expression in GFP-positive sorted B cells revealed that all three PTPN22 variants were at least 20 times more abundant than endogenous PTPN22 proteins in GFP-negative B cell counterparts (fig. S3B). The presence of 620W PTPN22 altered the counterselection of developing autoreactive B cells because GFP-positive new emigrant B cells expressing this variant contained many autoreactive clones that produce polyreactive antibodies (Fig. 2B, fig. S4A, and tables S12 to S17). High proportions of HEp-2–reactive and antinuclear GFP-positive new emigrant B cells corroborated this defective central B cell tolerance checkpoint (Fig. 2C and fig. S4, B to D). In contrast, GFP-negative B cell counterparts that developed in the same NSG mice rarely expressed polyreactive antibodies and displayed low frequencies of HEp-2–reactive and antinuclear clones, revealing that these B cells were properly selected in the absence of 620W PTPN22 expression (Fig. 2, B and C, and fig. S4, A to D). In addition, GFP-positive new emigrant B cells expressing either 620R PTPN22 or the loss-of-function 263Q PTPN22 variant displayed normal proportions of polyreactive, HEp-2–reactive, and antinuclear clones, demonstrating normal central B cell tolerance (Fig. 2, B and C, fig. S4, A to D, and tables S18 to S21). Regardless of how the 620W amino acid replacement alters PTPN22 function, our data demonstrate that B cell–intrinsic 620W PTPN22 expression is sufficient to interfere with the removal of developing autoreactive B cells and the establishment of human central B cell tolerance. Inhibition of PTPN22 enzymatic activity resets central B cell tolerance Because the rare loss-of-function 263Q PTPN22 variant was reported to confer protection against SLE and RA, we hypothesize that decreased PTPN22 phosphatase activity may inhibit autoimmunity (7, 8). LTV-1 is a compound that was identified to selectively inhibit human PTPN22 enzymatic activity (15). To assess the impact of the inhibition of 620W PTPN22 enzymatic activity on central B cell tolerance, we injected PTPN22 C/T or T/T engrafted NSG mice about 3 months after transplant with 0.75 mg of the LTV-1 compound twice daily for a week and determined the frequency of autoreactive new emigrant B cells (Fig. 3A). We found that LTV-1 treatment substantially reduced the frequencies of polyreactive new emigrant B cells in PTPN22 C/T or T/T transplanted mice, similar to those in NSG mice engrafted with HSCs that did not carry the PTPN22 T allele (Fig. 3B and tables S22 to S25). In addition, PTPN22 inhibition by LTV-1 also normalized the frequencies of HEp-2–reactive new emigrant B cells in PTPN22 C/T or T/T engrafted mice (Fig. 3C), and antinuclear clone frequencies remained very low (fig. S5). Central B cell tolerance was also restored by injecting five times less LTV-1 (0.15 mg per injection), although polyreactivity frequencies were slightly higher than those in mice treated with 0.75 mg per injection (Fig. 3B). These data therefore suggest a wide range of effective PTPN22 inhibition by LTV-1 (Fig. 3, B and C). Hence, inhibition of 620W PTPN22 enzymatic activity resets central B cell tolerance that is normally impaired by the presence of the PTPN22 T allele. Inhibition of PTPN22 expression during B cell development resets central B cell tolerance Although central B cell tolerance appears to be mainly regulated by B cell–intrinsic pathways involving B cell receptor (BCR) and, potentially, Toll-like receptor signaling (16), this checkpoint might be restored via B cell–extrinsic pathways normalized by 620W PTPN22 inhibition. In addition, the LTV-1 PTPN22 inhibitor may also nonspecifically alter the function of other phosphatases. To determine whether specific B cell–intrinsic PTPN22 blockade is responsible for the correction of central tolerance, we developed a strategy to inhibit the expression of PTPN22 in developing B cells using RNA interference (17). We engrafted NSG mice with PTPN22 C/T or T/T HSCs transduced with a GFP-tagged lentivirus expressing PTPN22-specific short hairpin RNA (shRNA) (Fig. 4A). We identified two PTPN22-specific shRNA (shRNA #1 and shRNA #3) that could inhibit about 80% of PTPN22 expression detected by Western blot using human RAMOS B cell line and chose shRNA #1 for all further experiments (fig. S6A). A high proportion of GFP-positive human B cells expressing PTPN22 shRNA #1 developed in NSG mice, revealing that transduced HSCs retained engraftment and B cell development capacities (Fig. 4B). In addition, GFP expression correlated with more than 90% decrease of PTPN22 expression in developing B cells (Fig. 4C). Blocking PTPN22 expression in GFP-positive PTPN22 C/T or T/T new emigrant B cells reduced the production of polyreactive and HEp-2–reactive clones compared with GFP-negative counterparts that often expressed autoreactive antibodies (Fig. 4, D and E, fig. S6, B and C, and tables S26 to S31). In addition, we have previously shown using control shRNA lentiviruses that HSC transduction per se does not interfere with the counterselection of autoreactive B cells (17). Together, these data demonstrate that the inhibition of PTPN22 expression in developing B cells can induce efficient removal of autoreactive clones and therefore restore central B cell tolerance that is otherwise impaired when the 620W PTPN22 variant is expressed. DISCUSSION Here, we reported that the PTPN22 T allele is responsible for the production of autoreactive B cells that escape central tolerance. These results are in agreement with the presence of PTPN22 T allele(s) correlating with impaired early B cell tolerance checkpoints (5). These observations may also explain why the PTPN22 T allele confers high risk of developing many autoimmune diseases as it induces central B cell tolerance defects observed in patients with T1D, RA, and SLE (4, 18, 19). The impact of the PTPN22 T allele was also investigated by generating knockin mice expressing the analogous mutation R619W in the murine Ptpn22 ortholog (20). The presence of R619W variants altered T and B cell selection and expansion, leading to systemic autoimmunity (20). In addition, B cell–restricted expression of the R619W variant was sufficient to promote autoimmunity, further demonstrating the prevalent tolerogenic role of the 620R PTPN22 variant (20). Together, elevated frequencies of autoreactive B cells resulting from the presence of the 619W or 620W PTPN22 variant in mice and humans, respectively, may increase the probability to present self-antigens and initiate autoimmunity. Central B cell tolerance could be reset in PTPN22 C/T or T/T subjects by inhibiting PTPN22 enzymatic activity or expression. In line with these observations, decreased PTPN22 phosphatase activity associated with the loss-of-function 263Q PTPN22 variant has already been suggested to protect against autoimmunity, likely by preventing the production of autoreactive B cells in the bone marrow (7, 8). How does PTPN22 inhibition favor the elimination of autoreactive B cells? The presence of the 620W PTPN22 variant decreases both T cell receptor (TCR) and BCR signaling and will therefore affect the selection of the repertoire of these receptors during early T and B cell development, respectively (21–23). Indeed, decreased BCR signaling results in impairment to induce tolerance mechanisms in immature B cells that bind self-antigens and allows some autoreactive clones to escape central B cell tolerance (16, 24). Because PTPN22 inhibition increases TCR signaling (15), it is likely that PTPN22 inhibition will also enhance BCR signaling and will thereby reestablish proper threshold for the removal of developing autoreactive B cells in the bone marrow. PTPN22 inhibition also likely resets TCR signaling thresholds altered by 620W PTPN22 variants (21, 22) and will therefore modify the TCR repertoire of both T effector and regulatory T cells selected in the thymus of PTPN22 T carriers. In conclusion, PTPN22 is a major regulator of human central B cell tolerance; its inhibition can normalize the elimination of developing autoreactive B cells and may thereby thwart the development of autoimmunity. MATERIALS AND METHODS Human progenitor cell isolation and injection in NSG mice Human CD34+ cells were purified from fetal liver samples by density gradient centrifugation followed by positive immunomagnetic selection with anti-human CD34 microbeads (Miltenyi Biotech). Newborn NSG mice (within first 3 days of life) were sublethally irradiated (x-ray irradiation with X-RAD 320 irradiator at 180 cGy), and 100,000 to 150,000 CD34+ cells in 20 µl of phosphate-buffered saline were injected into the liver with a 22-gauge needle (Hamilton Company).Mice were used for experiments 10 to 12 weeks after transplantation. NSG mice treated with the PTPN22 inhibitor were injected with the PTPN22 inhibitor 0.75 or 0.15 mg ip twice daily for a week. All animals were treated, and experiments were conducted in accordance with the Yale institutional reviewed guidelines on treatment of experimental animals. PTPN22 overexpression and silencing and CD34+ HSC transduction The pTRIP-Ubi-GFP lentiviral vector was used for overexpression of PTPN22 variants and shRNA delivery. Vector constructions have been previously described (17, 25). The following sequences were used for human PTPN22 targeting: shRNA #1, 5′-CTAGTGCTCTTGGTGTATATT-3′; shRNA #2, 5′-CTGTTGCCAACATCCTCTA-3′; shRNA #3, 5′-AAGAATCCACCTGACTTCC-3′. Lentiviral particles were produced by transient transfection of 293T cells, as previously described (26). Viruses were then used to transduce CD34+ HSCs in the presence of protamine sulfate (Sigma). Single-cell sorting B cells were enriched from splenocytes using magnetic separation with CD19 microbeads (Miltenyi Biotech) and stained with CD19–Pacific Blue, CD10-phycoerythrin-Cy7, CD21-allophycocyanin, and IgM-biotin (all from BioLegend) before purification. Single CD19+CD10+CD21loGFP− or CD19+CD10+CD21loGFP+ new emigrant B cells were sorted on a FACSAria (BD Biosciences) into 96-well polymerase chain reaction (PCR) plates and were immediately frozen on dry ice. cDNA synthesis, Ig gene amplification, antibody production, and antibody purification RNA from single cells was reverse-transcribed in the original 96-well plate in 12.5-µl reactions containing 100 U of Superscript II RT (Gibco BRL) for 45 min at 42°C. Reverse transcription PCRs, primer sequences, cloning strategy, expression vectors, and antibody expression and purification were as previously described (27). Enzyme-linked immunosorbent assays and immunofluorescence assays Antibody reactivity analysis was performed as previously described with the highly polyreactive ED38 antibody as positive control for HEp-2 reactivity and polyreactivity (28). Antibodies were considered polyreactive when they recognized all three distinct antigens: double-stranded DNA (dsDNA), insulin, and lipopolysaccharide (LPS). For indirect immunofluorescence assays, HEp-2 cell–coated slides (Bion Enterprises Ltd.) were incubated in a moist chamber at room temperature with purified recombinant antibodies at 50 to 100 µg/ml, according to the manufacturer’s instructions. Fluorescein isothiocyanate–conjugated goat anti-human IgG was used as detection reagent. Flow cytometry The following monoclonal antibodies against human antigens were used: anti-CD10 (HI10a), anti-CD19 (HIB19), anti-CD27 (O323), anti-CD45 (HI30), anti-CD21 (B-ly4), and anti-IgM (G20-127) (the first four antibodies were from BioLegend and the latter two were from BD Biosciences). Cells were acquired with an LSR II (BD Biosciences) and analyzed with FlowJo software. Representative gating strategies are shown in fig. S7. Immunoblot Total cell lysates were separated by SDS–polyacrylamide gel electrophoresis, transferred to polyvinylidene difluoride membranes, probed with mouse anti-PTPN22 (Invitrogen), and detected by chemiluminescence (Amersham ECL Prime Western Blotting Detection Reagent) using a GBox documentation system (Syngene). For quantification, blots were stripped with stripping buffer (Pierce) and reprobed with a mouse anti–β-actin antibody (Sigma-Aldrich). Unprocessed immunoblot images are shown in fig. S8. Statistical analysis Statistical analysis was performed using GraphPad Prism (version 5.0; GraphPad). Data are reported as means ± SD. Differences between groups of research subjects were analyzed for statistical significance with unpaired two-tailed Student’s t tests. A P value of ≤0.05 was considered significant. Supplementary Material Supplemental Material We thank M. Kalp for advice on the PTPN22 inhibitor and L. Devine and C. Wang for cell sorting. Funding: This work was supported by a grant provided by AbbVie. J.-N.S. and E.M. are authors on U. S. Patent #47162-5220-P1-US.605388 “Compositions and Methods for Inhibiting PTPN22.” Fig. 1 Defective central B cell tolerance in humanized mouse engrafted with HSCs carrying PTPN22 T allele(s) (A) Schematic diagram depicting the generation of humanized mice. CD34+ HSCs that carry (or do not carry) PTPN22 T allele(s) were injected into the liver of 3-day-old recipient NSG mice. (B) Representative flow cytometry analysis of the frequency of human (h) CD45+, CD3+, and CD19+ cells in the blood of the indicated recipient mice. The summary of blood engraftment from NSG mice transplanted with PTPN22 C/C, C/T, or T/T HSCs is represented. Each dot represents an individual mouse, and the bars indicate mean values. The frequencies of polyreactive (C) and HEp-2–reactive (D) new emigrant B cells from different types of humanized mice transplanted with indicated HSCs were determined and compared with those of healthy donors that carry (or do not carry) PTPN22 T allele(s). Dotted lines show positive control. For each B cell fraction, the frequency of reactive (solid area) and nonreactive (open area) clones is summarized in pie charts, with the total number of clones tested indicated in the center. In summarized reactivity panels on the right, each diamond represents an individual, and each dot represents a mouse. Averages are shown with a bar, and statistically significant differences are indicated. OD, optical density. Fig. 2 620W PTPN22 overexpression interferes with central B cell tolerance (A) Humanized mice were generated with CD34+ HSCs transduced with lentiviruses, allowing the expression of different variants of PTPN22 before being injected into the liver of 3-day-old recipient NSG mice. LTR, long terminal repeat. The frequencies of polyreactive (B) and HEp-2–reactive (C) new emigrant B cells from sorted GFP-positive fractions expressing 620W PTPN22, 620R PTPN22, or 263Q PTPN22 were determined and compared with those of GFP-negative new emigrant B cells. Dotted lines show positive control. For each B cell fraction, the frequency of reactive (solid area) and nonreactive (open area) clones is summarized in pie charts, with the total number of clones tested indicated in the center. In summarized reactivity panels on the right, each symbol represents a mouse overexpressing 620W PTPN22 (green dots), 620R PTPN22 (green squares), or 263Q PTPN22 (green triangles). Averages are shown with a bar, and statistically significant differences are indicated. Fig. 3 Inhibition of PTPN22 enzymatic activity resets central B cell tolerance (A) Schematic diagram depicting the PTPN22 inhibitor treatment strategy. NSG mice generated with CD34+ HSCs carrying PTPN22 T allele(s) were injected twice daily with 0.75 or 0.15 mg of the PTPN22 inhibitor for 1 week. The frequencies of polyreactive (B) and HEp-2–reactive (C) new emigrant B cells from NSG mice carrying PTPN22 T allele(s) and treated with the PTPN22 inhibitor were determined and compared with those of nontreated NSG mice. Dotted lines show positive control. For each B cell fraction, the frequency of reactive (solid area) and nonreactive (open area) clones is summarized in pie charts, with the total number of clones tested indicated in the center. In summarized reactivity panels on the right, each dot represents an untreated mouse and full and half-filled diamonds represent mice treated with either 0.75 or 0.15 mg of the LTV-1 PTPN22 inhibitor, respectively. Averages are shown with a bar, and statistically significant differences are indicated. Fig. 4 Inhibition of PTPN22 expression during B cell development resets central B cell tolerance (A) CD34+ HSCs carrying PTPN22 T allele(s) were transduced with lentiviruses, allowing the expression of PTPN22 shRNA before injection into the liver of 3-day-old NSG mice. (B) Representative flow cytometry analysis of CD19+ cells isolated from the spleen of NSG mouse engrafted with PTPN22 C/T HSCs transduced with a GFP-tagged lentivirus expressing PTPN22-specific shRNA. CD19+ B cells were stained with anti-hCD19, anti-IgM, and anti-hCD10 antibodies. The frequencies of GFP-negative and GFP-positive shRNA+ new emigrant B cells are shown. (C) PTPN22 protein expression in GFP-negative and GFP-positive shRNA+ hCD19+ cells isolated from the spleen of NSG mice; β-actin is used for normalization of protein expression. Percentage of knockdown is indicated. (D and E) B cell–intrinsic PTPN22 expression is required for central B cell tolerance. The frequencies of polyreactive (D) and HEp-2–reactive (E) new emigrant B cells from sorted GFP-positive fractions expressing PTPN22 shRNA were determined and compared with those of GFP-negative new emigrant B cells. Dotted lines show positive control. For each B cell fraction, the frequency of reactive (solid area) and nonreactive (open area) clones is summarized in pie charts, with the total number of clones tested indicated in the center. In summarized reactivity panels on the right, each symbol represents a mouse. The average is shown with a bar, and statistically significant differences are indicated. SUPPLEMENTARY MATERIALS immunology.sciencemag.org/cgi/content/full/1/1/aaf7153/DC1 Fig. S1. New emigrant B cells isolated from NSG mice engrafted with PTPN22 C/T or T/T HSCs display normal IgH repertoire. Fig. S2. Defective central B cell tolerance in humanized mouse engrafted with HSCs carrying PTPN22 T allele(s). Fig. S3. Overexpression of PTPN22 variants in NSG mice. Fig. S4. 620W PTPN22 overexpression interferes with the central B cell tolerance checkpoint. Fig. S5. Frequencies of antinuclear new emigrant B cells in PTPN22 C/T or T/T NSG mice treated with the LTV-1 PTPN22 inhibitor. Fig. S6. Inhibition of PTPN22 expression during B cell development resets central B cell tolerance. Fig. S7. Representative gating strategies. Fig. S8. Unprocessed immunoblot images. Table S1. Fetal donor characteristics. Table S2. Repertoire and reactivity of antibodies from new emigrant B cells of mouse #1. Table S3. Repertoire and reactivity of antibodies from new emigrant B cells of mouse #2. Table S4. Repertoire and reactivity of antibodies from new emigrant B cells of mouse #3. Table S5. Repertoire and reactivity of antibodies from new emigrant B cells of mouse #4. Table S6. Repertoire and reactivity of antibodies from new emigrant B cells of mouse #5. Table S7. Repertoire and reactivity of antibodies from new emigrant B cells of mouse #6. Table S8. Repertoire and reactivity of antibodies from new emigrant B cells of mouse #7. Table S9. Repertoire and reactivity of antibodies from new emigrant B cells of mouse #8. Table S10. Repertoire and reactivity of antibodies from new emigrant B cells of mouse #9. Table S11. Repertoire and reactivity of antibodies from new emigrant B cells of mouse #10. Table S12. Repertoire and reactivity of antibodies from GFP-negative new emigrant B cells of mouse #11. Table S13. Repertoire and reactivity of antibodies from GFP-positive new emigrant B cells expressing 620W PTPN22 from mouse #11. Table S14. Repertoire and reactivity of antibodies from GFP-negative new emigrant B cells of mouse #12. Table S15. Repertoire and reactivity of antibodies from GFP-positive new emigrant B cells expressing 620W PTPN22 from mouse #12. Table S16. Repertoire and reactivity of antibodies from GFP-negative new emigrant B cells of mouse #13. Table S17. Repertoire and reactivity of antibodies from GFP-positive new emigrant B cells expressing 620W PTPN22 from mouse #13. Table S18. Repertoire and reactivity of antibodies GFP-positive new emigrant B cells expressing R620 PTPN22 from mouse #14. Table S19. Repertoire and reactivity of antibodies from GFP-positive new emigrant B cells expressing R620 PTPN22 from mouse #15. Table S20. Repertoire and reactivity of antibodies from GFP-positive new emigrant B cells expressing 263Q PTPN22 from mouse #16. Table S21. Repertoire and reactivity of antibodies from GFP-positive new emigrant B cells expressing 263Q PTPN22 from mouse #17. Table S22. Repertoire and reactivity of antibodies from new emigrant B cells of mouse #18 treated with 0.75 mg of the LTV-1 PTPN22 inhibitor. Table S23. Repertoire and reactivity of antibodies from new emigrant B cells of mouse #19 treated with 0.75 mg of the LTV-1 PTPN22 inhibitor. Table S24. Repertoire and reactivity of antibodies from new emigrant B cells of mouse #20 treated with 0.15 mg of the LTV-1 PTPN22 inhibitor. Table S25. Repertoire and reactivity of antibodies from new emigrant B cells of mouse #21 treated with 0.15 mg of the LTV-1 PTPN22 inhibitor. Table S26. Repertoire and reactivity of antibodies from GFP-negative new emigrant B cells of mouse #21. Table S27. Repertoire and reactivity of antibodies from GFP-positive new emigrant B cells expressing PTPN22 shRNA from mouse #21. Table S28. Repertoire and reactivity of antibodies from GFP-negative new emigrant B cells of mouse #22. Table S29. Repertoire and reactivity of antibodies from GFP-positive new emigrant B cells expressing PTPN22 shRNA from mouse #22. Table S30. Repertoire and reactivity of antibodies from GFP-negative new emigrant B cells of mouse #23. Table S31. Repertoire and reactivity of antibodies from GFP-positive new emigrant B cells expressing PTPN22 shRNA from mouse #23. Author contributions: E.M. started the collaboration with R.F. and T.S. that shared protocols for the generation of engrafted NSG mice. P.S.-S. built and provided the pTRIP lentiviral vector. J.-N.S. and E.M. designed the experiments. J.-N.S., M.K., A.B., J.M.B., C.M., H.W., N.K., T.O., and L.M. performed the experiments and were responsible for the statistical analysis. J.-N.S. and E.M. wrote the manuscript. All authors reviewed the manuscript and provided scientific input. Competing interests: The other authors declare that they have no competing interests. REFERENCES AND NOTES 1 Pescovitz MD Greenbaum CJ Krause-Steinrauf H Becker DJ Gitelman SE Goland R Gottlieb PA Marks JB McGee PF Moran AM Raskin P Rodriguez H Schatz DA Wherrett D Wilson DM Lachin JM Skyler JS Type 1 Diabetes TrialNet Anti-CD20 Study Group Rituximab, B-lymphocyte depletion, and preservation of β-cell function N. Engl. J. Med 2009 361 2143 2152 19940299 2 Edwards JCW Szczepański L Szechiński J Filipowicz-Sosnowska A Emery P Close DR Stevens RM Shaw T Efficacy of B-cell–targeted therapy with rituximab in patients with rheumatoid arthritis N. Engl. J. Med 2004 350 2572 2581 15201414 3 Hauser SL Waubant E Arnold DL Vollmer T Antel J Fox RJ Bar-Or A Panzara M Sarkar N Agarwal S Langer-Gould A Smith CH HERMES Trial Group B-cell depletion with rituximab in relapsing–remitting multiple sclerosis N. Engl. J. Med 2008 358 676 688 18272891 4 Chamberlain N Massad C Oe T Cantaert T Herold KC Meffre E Type 1 Diabetes TrialNet Pathway to Prevention Study Group Rituximab does not reset defective early B cell tolerance checkpoints J. Clin. Invest 2016 126 282 287 26642366 5 Menard L Saadoun D Isnardi I Ng Y-S Meyers G Massad C Price C Abraham C Motaghedi R Buckner JH Gregersen PK Meffre E The PTPN22 allele encoding an R620W variant interferes with the removal of developing autoreactive B cells in humans J. Clin. Invest 2011 121 3635 3644 21804190 6 Bottini N Peterson EJ Tyrosine phosphatase PTPN22: Multifunctional regulator of immune signaling, development, and disease Annu. Rev. Immunol 2014 32 83 119 24364806 7 Orrú V Tsai SJ Rueda B Fiorillo E Stanford SM Dasgupta J Hartiala J Zhao L Ortego-Centeno N D’Alfonso S Italian Collaborative Group Arnett FC Wu H Gonzalez-Gay MA Tsao BP Pons-Estel B Alarcon-Riquelme ME He Y Zhang Z-Y Allayee H Chen XS Martin J Bottini N A loss-of-function variant of PTPN22 is associated with reduced risk of systemic lupus erythematosus Hum. Mol. Genet 2009 18 569 579 18981062 8 Rodríguez-Rodríguez L Taib WRW Topless R Steer S González-Escribano MF Balsa A Pascual-Salcedo D González-Gay MA Raya E Fernandez-Gutierrez B González-Álvaro I Bottini N Witte T Viken MK Coenen MJH van Riel PLCM Franke B den Heijer M Radstake TRDJ Wordsworth P Lie BA Merriman TR Martín J The PTPN22 R263Q polymorphism is a risk factor for rheumatoid arthritis in Caucasian case–control samples Arthritis Rheum 2011 63 365 372 21279993 9 Shultz LD Lyons BL Burzenski LM Gott B Chen X Chaleff S Kotb M Gillies SD King M Mangada J Greiner DL Handgretinger R Human lymphoid and myeloid cell development in NOD/LtSz-scid IL2Rγnull mice engrafted with mobilized human hematopoietic stem cells J. Immunol 2005 174 6477 6489 15879151 10 Rongvaux A Willinger T Martinek J Strowig T Gearty SV Teichmann LL Saito Y Marches F Halene S Palucka AK Manz MG Flavell RA Development and function of human innate immune cells in a humanized mouse model Nat. Biotechnol 2014 32 364 372 24633240 11 Kalscheuer H Danzl N Onoe T Faust T Winchester R Goland R Greenberg E Spitzer TR Savage DG Tahara H Choi G Yang Y-G Sykes M A model for personalized in vivo analysis of human immune responsiveness Sci. Transl. Med 2012 4 125ra30 12 Corbett SJ Tomlinson IM Sonnhammer ELL Buck D Winter G Sequence of the human immunoglobulin diversity (D) segment locus: A systematic analysis provides no evidence for the use of DIR segments, inverted D segments, “minor” D segments or D-D recombination J. Mol. Biol 1997 270 587 597 9245589 13 Ng Y-S Wardemann H Chelnis J Cunningham-Rundles C Meffre E Bruton’s tyrosine kinase is essential for human B cell tolerance J. Exp. Med 2004 200 927 934 15466623 14 Meyers G Ng Y-S Bannock JM Lavoie A Walter JE Notarangelo LD Kilic SS Aksu G Debré M Rieux-Laucat F Conley ME Cunningham-Rundles C Durandy A Meffre E Activation-induced cytidine deaminase (AID) is required for B-cell tolerance in humans Proc. Natl. Acad. Sci. U.S.A 2011 108 11554 11559 21700883 15 Vang T Liu WH Delacroix L Wu S Vasile S Dahl R Yang L Musumeci L Francis D Landskron J Tasken K Tremblay ML Lie BA Page R Mustelin T Rahmouni S Rickert RC Tautz L LYP inhibits T-cell activation when dissociated from CSK Nat. Chem. Biol 2012 8 437 446 22426112 16 Meffre E The establishment of early B cell tolerance in humans: Lessons from primary immunodeficiency diseases Ann. N. Y. Acad. Sci 2011 1246 1 10 22236425 17 Cantaert T Schickel J-N Bannock JM Ng Y-S Massad C Oe T Wu R Lavoie A Walter JE Notarangelo LD Al-Herz W Kilic SS Ochs HD Nonoyama S Durandy A Meffre E Activation-induced cytidine deaminase expression in human B cell precursors is essential for central B cell tolerance Immunity 2015 43 884 895 26546282 18 Samuels J Ng Y-S Coupillaud C Paget D Meffre E Impaired early B cell tolerance in patients with rheumatoid arthritis J. Exp. Med 2005 201 1659 1667 15897279 19 Yurasov S Wardemann H Hammersen J Tsuiji M Meffre E Pascual V Nussenzweig MC Defective B cell tolerance checkpoints in systemic lupus erythematosus J. Exp. Med 2005 201 703 711 15738055 20 Dai X James RG Habib T Singh S Jackson S Khim S Moon RT Liggitt D Wolf-Yadlin A Buckner JH Rawlings DJ A disease-associated PTPN22 variant promotes systemic autoimmunity in murine models J. Clin. Invest 2013 123 2024 2036 23619366 21 Rieck M Arechiga A Onengut-Gumuscu S Greenbaum C Concannon P Buckner JH Genetic variation in PTPN22 corresponds to altered function of T and B lymphocytes J. Immunol 2007 179 4704 4710 17878369 22 Salmond RJ Brownlie RJ Morrison VL Zamoyska R The tyrosine phosphatase PTPN22 discriminates weak self peptides from strong agonist TCR signals Nat. Immunol 2014 15 875 883 25108421 23 Arechiga AF Habib T He Y Zhang X Zhang Z-Y Funk A Buckner JH Cutting edge: The PTPN22 allelic variant associated with autoimmunity impairs B cell signaling J. Immunol 2009 182 3343 3347 19265110 24 Grimaldi CM Hicks R Diamond B B cell selection and susceptibility to autoimmunity J. Immunol 2005 174 1775 1781 15699102 25 Ruer-Laventie J Simoni L Schickel J-N Soley A Duval M Knapp A-M Marcellin L Lamon D Korganow A-S Martin T Pasquali J-L Soulas-Sprauel P Overexpression of Fkbp11, a feature of lupus B cells, leads to B cell tolerance breakdown and initiates plasma cell differentiation Immun. Inflamm. Dis 2015 3 265 279 26417441 26 Schickel J-N Pasquali J-L Soley A Knapp A-M Decossas M Kern A Fauny J-D Marcellin L Korganow A-S Martin T Soulas-Sprauel P Carabin deficiency in B cells increases BCR-TLR9 costimulation-induced autoimmunity EMBO Mol. Med 2012 4 1261 1275 23109291 27 Tiller T Meffre E Yurasov S Tsuiji M Nussenzweig MC Wardemann H Efficient generation of monoclonal antibodies from single human B cells by single cell RT-PCR and expression vector cloning J. Immunol. Methods 2008 329 112 124 17996249 28 Wardemann H Yurasov S Schaefer A Young JW Meffre E Nussenzweig MC Predominant autoantibody production by early human B cell precursors Science 2003 301 1374 1377 12920303

PMC005xxxxxx/PMC5127716.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 101275111 33888 Annu Rev Pathol Annu Rev Pathol Annual review of pathology 1553-4006 1553-4014 24050625 5127716 10.1146/annurev-pathol-012513-104653 NIHMS831025 Article The Intracellular Life of Cryptococcus neoformans Coelho Carolina 12 Bocca Anamelia L. 3 Casadevall Arturo arturo.casadevall@einstein.yu.edu 1 1 Department of Microbiology and Immunology, Albert Einstein College of Medicine, Yeshiva University, Bronx, New York 10461 2 Center for Neuroscience and Cell Biology and Institute of Microbiology, Faculty of Medicine, University of Coimbra, 3004-504 Coimbra, Portugal 3 Department of Cellular Biology, Institute of Biological Sciences, University of Brasilia, Brasilia 70910900, Brazil 21 11 2016 16 9 2013 2014 29 11 2016 9 219238 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. Cryptococcus neoformans is a fungal pathogen with worldwide distribution. Serological studies of human populations show a high prevalence of human infection, which rarely progresses to disease in immunocompetent hosts. However, decreased host immunity places individuals at high risk for cryptococcal disease. The disease can result from acute infection or reactivation of latent infection, in which yeasts within granulomas and host macrophages emerge to cause disease. In this review, we summarize what is known about the cellular recognition, ingestion, and killing of C. neoformans and discuss the unique and remarkable features of its intracellular life, including the proposed mechanisms for fungal persistence and killing in phagocytic cells. granuloma Trojan horse disease tolerance fungal immunity intracellular pathogen C. neoformans killing Introduction to The Biology of Cryptococcus Neoformans Environmental Organism and Treatment of Disease Cryptococcus neoformans was first described in 1894 by Otto Busse, when the organism was recovered from a lesion in a woman's tibia (1). The pathogenic yeast can be found worldwide in several environmental niches and has been isolated from soil, trees, and animals, in particular from avian guano (1, 2). Exposure to C. neoformans does not usually lead to overt disease, and epidemiological data led to the accepted view that establishment of an asymptomatic latent state may be the most common outcome of infection (3–5). Even from the early clinical cases described, an association between cryptococcosis and immunosuppression was already inferred (39, 40). In fact, in immunosuppressed patients, reactivation of infection is frequently fatal. Patients develop pneumonia and meningoencephalitis, and brain involvement predicts high mortality and morbidity, even with aggressive antifungal drug therapy (6). Immunity to Cryptococcosis Serological studies show that 80% of children in urban environments have been infected with C. neoformans, without any discernible clinical manifestations (4, 7). Primary infection most likely occurs via inhalation of spores or desiccated yeast cells from environmental sources. The physical characteristics of these infectious particles, such as size and capacity to become airborne, allow deposition in the lungs (8). There, the yeast particles encounter an alveolar macrophage or dendritic cell and trigger an immune response, culminating in sterilization or, most likely, restriction of infection within a granuloma. The resulting granulomas are usually well circumscribed, self-limited, and benign (Figure 1) and are composed mainly of mature mononuclear phagocytes, histiocytes, and giant multinucleated cells enveloping the yeast cells (9). Efficient control of C. neoformans requires a delicate balance of both Th1- and Th2-type responses (10–12). Depletion of cytokines by genetic disruption or antibody neutralization has confirmed that a Th1-type response is essential to control infection; these studies are summarized in Table 1. In fact, mouse strains show differential susceptibilities that correlate with a stronger Th1 versus Th2 skewing (13) and with the presence of complement cascade member C5 (14). Depletion of Th1-type cytokines, such as interferon-γ (IFN-γ) and interleukin (IL)-12, consistently results in decreased mouse survival (15, 16), whereas loss of hallmark Th2-type cytokines increases mouse survival (17). In these models, Th1 or Th2 cytokine bias is reflected in both granuloma composition and control of fungal burden (18). Although a predominantly Th1-type response results in mouse survival, too strong of a Th1-type polarization cannot prevent brain dissemination (19–22) and associated mortality, and the Th2 component is required for the most efficient immune response. Although an impressive body of work has been carried out to characterize cytokine dependence, an understanding of immunity to cryptococcosis is still incomplete. For example, lack of the Th1 major cytokine tumor necrosis factor α (TNF-α) did not influence mouse survival, but administration of TNF-α was beneficial (23). As another example, Th17 immunity was crucial for Candida albicans mucosal immunity (24) but appears to play a lesser role in cryptococcal disease: In models of cryptococcosis, deletion of Th17-type responses did not influence the outcome of primary infection or the efficiency of vaccination (25). Macrophages are crucial for control of cryptococcosis, as evidenced by the observation that depletion of host macrophages and dendritic cells results in dramatically reduced survival after C. neoformans challenge (26, 27). Two studies of the effects of macrophage depletion on lung fungal burden produced contradictory results (27, 28); however, both studies demonstrated that mouse macrophages require a particular activation profile to become fungicidal (28). Macrophages with a mixed classical and alternative activation phenotype are seen during experimental models of cryptococcosis (19). Although they are less studied, other types of innate immune cells are found in granulomas and may play a role in defense against cryptococcosis (29). The presence of either excess eosinophils or excess neutrophils is associated with poor control of infection in mice (30, 31), whereas eosinophils might have a beneficial role in rats (32). An extensive body of literature shows that induced or passively administered antibodies can mediate significant protection from cryptococcosis (33). However, the role of humoral immunity in the cryptococcosis model is not adequately explained by classical mechanisms of antibody-mediated immunity, which has led to the discovery of novel immunoregulatory functions of antibodies (33). Various investigators have addressed the immunological mechanism for effective immunization against C. neoformans challenge (3, 27, 34, 35). For example, immunization with capsular mannoproteins was able to prolong mouse survival (34). An alternative approach was to design an IFN-γ-producing C. neoformans (IFN-γ is a strong Th1-type cytokine) (25, 36–38). This strategy resulted in complete protection from a posterior challenge, accompanied by a Th1-biased lung cytokine pattern, classical activation of macrophages, and increased production of nitric oxide (NO) (37), and demonstrated how appropriate manipulation of the host immune system, in particular macrophage activation, can be an effective therapeutic option. At this time, there is a reasonable consensus that defense against cryptococcosis depends on an appropriate collaboration of Th1 cells with macrophages. Evidence That Intracellular Residence Contributes to Virulence and Immune Escape Evidence from Pathological Studies C. neoformans lesions in autopsies (9, 39–41) and experimental models (42) show fungal cells inside granulomas, known as cryptococcomas (Figure 1b,c). Cryptococcal granulomas are less inflammatory than Mycobacterium tuberculosis granulomas, suggesting a dormant and controlled infection. In well-organized granulomas the yeast is localized within the cytosol of giant cells or macrophages, but in the absence of granulomas yeasts are both intracellular and extracellular (Figure 1f) (41). Neutrophilic infiltrates are not common in human cryptococcal lesions, whereas CD4+ T cells are found in immunocompetent patients. In rats (42), mice (3, 43), and rabbits (44), C. neoformans can be found associated with lung macrophages, in some cases for months, without obvious clinical manifestations. In mice, C. neoformans is rapidly ingested by phagocytes, and in one model of experimental infection, there was a fluctuation in intracellular and extracellular residence during the first 24 h (43). At day 7, a shift occurred toward the intracellular lifestyle, coincident with formation of granulomas. At day 28, most yeast cells were found within multinucleated giant cells, as illustrated in Figure 1e. In this model, the budding index was higher for intracellular than for extracellular C. neoformans, sparking the hypothesis that intracellular residency is favorable for C. neoformans growth. Hence, both early infection and long-term persistence find C. neoformans cells associated with host macrophages, supporting the importance of intracellular residence within them (43). Evidence from Animal Models Animal models have found evidence consistent with the view that fungal residency within macrophages contains the infection while allowing the fungus to persist in tissue. Rats are more resistant than mice are to C. neoformans infection, but the two rodent systems have provided complementary information. Rats' superior resistance to cryptococcosis is associated with a more effective macrophage fungicidal capacity, an effect attributed to increased production by macrophages of lysozyme and reactive oxygen species (ROS) (28, 42). Similar to the situation in mice, C. neoformans resistance in rats is associated with a strong Th1 response balanced with an adequate Th2 component (18). Rats that control infection develop mature granulomas containing eosinophils, whereas rats with an excessive Th1 response develop more inflammatory granulomas with central necrosis and caseation. Early in the course of rat infection, extracellular C. neoformans is prominent, but after granuloma formation the percentage of intracellular fungi increases, with a concomitant reduction in fungal burden (45). Further evidence that macrophages are required for both control and persistence of disease came from the observation that macrophage depletion can prevent yeast dissemination into the mouse brain (46, 47). This result is consistent with the notion that fungal dissemination to the brain involves the transport of viable yeast cells inside host macrophages. The idea that C. neoformans has a favorable niche within murine macrophages was directly investigated by constructing a yeast strain that could survive only within acidic environments. During the course of infection, an acidic environment is found solely in the phagosome. This strain, although confined to the phagocytic compartment, was still virulent in natural killer– and T cell–depleted mice, indicating that yeast virulence occurs from the intracellular compartment (47). In the same immunosuppressed mice, depletion of alveolar macrophages delayed mouse death, supporting the concept that the macrophages are a niche for intracellular survival of C. neoformans (47). The Intracellular Life Cycle of Cryptococcus Neoformans Fungal Entry and Recognition Fungal cell wall components, such as α-glucans, β-glucans, and chitin, are recognized by pattern recognition receptors (PRRs) present in immune cells, triggering cellular activation and, in the case of phagocytic receptors, ingestion of the fungal particle. However, the capsule is highly antiphagocytic, and without opsonins there is no significant ingestion of yeast cells in vitro. Because acapsular C. neoformans is readily ingested through complement receptors and/or β-glucan receptors (48), it has been hypothesized that the large polysaccharide capsule conceals most fungal PRR ligands, thereby decreasing phagocytosis by host cells (Figure 2) (49). In fact, for efficient phagocytosis in vitro (Figure 3), opsonization with antibody or complement is necessary, after which phagocytosis proceeds through a complex interplay of Fc receptors, complement receptors (50), and Dectin-1 (51). Despite the capsule's antiphagocytic properties in vitro, C. neoformans ingestion occurs readily in vivo. The opsonin or the receptor responsible for in vivo ingestion has not been definitively identified. The complement system is the most likely candidate because complement-deficient animals have greater susceptibility to cryptococcosis (14, 52). C. neoformans spores are acapsular, and thus their surfaces expose more β-glucans than do the surfaces of the yeasts; therefore, when spores are the infectious particles, Dectin-1 and other β-glucan PRRs might be readily activated (8) and mediate rapid ingestion of C. neoformans. Cell wall β-glucans can be recognized by Dectin-1, Toll-like receptor 2 (TLR2), Nodlike receptors, and several scavenger receptors. In addition, CD36 and scavenger receptor F1 (SCARF1) are responsible for immune cell binding of C. neoformans in the mouse lung (53). Recognition of the yeast particle is not limited to the immune cell extracellular membrane but continues within the phagolysosome, and even the host cytosol is monitored for the presence of fungal components. In C. albicans infection, Dectin-1 and complement receptors accumulated at sites of phagocytosis but dissociated from the phagosome shortly after internalization, while mannose receptors fused into nascent phagosomes, displaying a coordinated cooperation (54). In contrast, in Aspergillus fumigatus, Dectin-1 remained within the phagosome and was capable of interacting with β-glucans within the acidic compartment (55). Activation of Dectin-1 by β-glucans in vitro led to enhanced macrophage fungicidal activity, presumably because Dectin-1 mediated inflammasome activation and proinflammatory cytokine production (Figure 4a), which can trigger a more effective antifungal response. Therefore, disguise of β-glucans by the C. neoformans capsule may impair maximal macrophage activation. Thus, defects in recognition of C. neoformans by Dectin-1 might explain why mice deficient in Dectin-1 do not have increased susceptibility to C. neoformans infection (56). This hypothesis has been proven in C. albicans, where Dectin-1 dependency is fungal strain dependent due to differences in cell wall composition (57). Other receptors have been shown to be crucial for C. neoformans recognition. Both TLR2- (58) and mannose-deficient mice (59) have decreased immunity to cryptococcal challenge, and the TLR9 receptor is important because of cytosolic detection of fungal DNA (60, 61). In summary, mannose receptor, complement receptors, CD36, SCARF1, TLR2, and TLR9 are all crucial receptors for C. neoformans recognition in the lung, and cross talk between multiple PRRs is necessary for maximal immune response. Cells other than immune cells might also recognize the presence of C. neoformans, and IL-8 secretion by epithelial cells has been detected (62). Within the lungs, despite extensive adhesion to the epithelium, very little invasion of epithelial cells by C. neoformans occurs (63). However, the yeast is commonly found within lung capillaries and can cross the blood-brain and endothelial barriers, which leads to the conclusions that the yeast is able to cross host tissues (64) and that epithelial cells play a role in the pathogenesis of C. neoformans. Phagosome Maturation C. neoformans has not been shown to interfere with phagosomal maturation. A phagosome containing C. neoformans is able to acidify (65), and this acidification is beneficial for fungal replication (65–67). The existing characterization of the C. neoformans phagosome shows that lysosomal fusion occurs (Figure 5) and phagosomes quickly acquire an array of phagosomal markers (Figure 6) (65, 66, 68). Autophagic markers colocalize to the C. neoformans phagosome (65), but the yeast has not been found within an autophagic compartment. Autophagy mediators may perform functions in this phagosome distinct from their canonical functions; such hypothetical activities would explain why depletion of Atg2, Atg5, or Atg9 decreases uptake and/or replication of C. neoformans (65, 69) and why depletion of Atg5 affects survival after C. albicans but not C. neoformans challenge. Evidence for Host Cytotoxicity Despite normal phagosome maturation, macrophage phagosomes become leaky after C. neoformans infection, as measured by light and electron microscopy (70). Leakiness of the phagosome would have a myriad of consequences: loss of acidity, leakage of macrophage-damaging phagosomal enzymes, easy fungal access to cytoplasmic nutrients, and release of strong immunomodulatory capsular components into the cytosol (Figure 4b). At this time, it is not clear whether the leakiness of phagosomes reflects a loss of phagosomal integrity due to macrophage damage, a direct effect of the fungus, or a combination of both. It is also hard to reconcile the fact that the yeast prefers an acidic phagosome with the notion of fungi residency within a leaky, nonacidic phagosome. Reports of fungal damage to host macrophages are scarce. Lipid peroxidation was observed in rat alveolar macrophages exposed to C. neoformans, which presumably occurs as a result of excessive ROS production by the macrophage (71) and not due to direct fungal toxicity. In vivo, cells that have ingested C. neoformans display features of affected lysosomes; they are known as hueco cells, after the Spanish word for hole, given their perforated appearance in electron microscopy preparations (43). Capsulated, but not acapsular, C. neoformans can trigger apoptosis in macrophages (72), and this observation has been replicated for isolated capsular components (73). Phagocytosis can stimulate proliferation of macrophage cells (74, 75), yet in prolonged C. neoformans infection, ingestion of yeast cells specifically inhibited cyclin D1 expression (75) and decreased macrophage mitosis, indicating cell cycle arrest (76). Similarly, the presence of extracellular yeast triggered aneuploidy and cell cycle impairment in macrophages (72). The realization that fungi, like bacteria such as Mycobacterium tuberculosis (77), can manipulate the host cell cycle to their advantage is an exciting development in fungal pathogenesis. However, the type of macrophage adaptations necessary to support the observed long-term residence of fungal pathogens has not been elucidated. Killing of Cryptococcus neoformans Human macrophages restrict C. neoformans growth for up to 24 h after infection (78), a finding indicative of damage to the fungus. Within the phagosome, the yeast is exposed simultaneously to low pH, ROS, reactive nitrogen species, and nutrient starvation (79). These challenges are counteracted by equally powerful mechanisms on the yeast side. Upon ingestion, the yeast upregulates gene expression of oxidative stress enzymes (80), starvation responses, and the autophagic machinery (81). These collaborate with the antioxidant properties of fungal melanin and the capsule to efficiently protect the fungus from host attack. In a model of NADPH oxidase–null mice, cryptococcal infection is contained and the fungal load in both brain and lung is decreased (82), suggesting that inflammatory ROS are prejudicial to the host rather than to the fungus. One antimicrobial molecule proven to be inhibitory to C. neoformans in acidic conditions is NO (83). The enzyme that produces NO, iNOS (inducible nitric oxide synthase), is present in C. neoformans granulomas in the lung (37, 42, 84), and NO has a protective role in the cryptococcosis mouse model (85, 86). The understanding of these complex effects is hampered by the difficulty in separating the direct fungicidal and indirect immunoregulatory effects of NO, but because C. neoformans with defective nitrosative defenses is only slightly less virulent than is wild-type C. neoformans (87), immunoregulation seems to be the predominant effect of NO. T cells and natural killer cells exert direct antifungal activity, at least in vitro (88, 89), through an unknown mechanism. Neutrophils and dendritic cells can kill opsonized fungi through oxidative and nonoxidative mechanisms (90, 91), and the myeloperoxidase system contributes significantly to antifungal activity against C. neoformans, given that myeloperoxidase-knockout mice have dramatically decreased survival after cryptococcal infection (92). Nonoxidative mechanisms include Cathepsin-B-induced structural changes and rupture of the fungal cell wall (93) in dendritic cells, whereas neutrophils have been reported to use both oxidative burst and nonoxidative molecules such as calprotectin and defensins (91). In macrophages, microbicidal activity depends on macrophage activation, in which Th1-type responses result in the upregulation of ROS, reactive nitrogen species, proteases, and lipid mediators (94), all of which would render macrophages more effective in pathogen killing. Such Th1 stimulation can also decrease phagosomal hydrolase activity to increase major histocompatibility complex presentation and stimulation of adaptive immunity (95). However, in the case of C. neoformans infection, even IFN-γ stimulation of macrophages failed to elicit efficient killing in vitro (78). Therefore, the contribution of macrophages' oxidative and nonoxidative defenses to fungal control remains unknown. Nonlytic Exocytosis Upon phagocytosis, C. neoformans can undergo morphological changes, such as capsular enlargement, that aid its survival within, and even its escape from, host phagocytes (96). Some of these changes include fungal giant cell (titan cell) formation (97, 98), cell-to-cell spread (99), and nonlytic exocytosis (NLE) (100, 101). The presence of mechanisms to flee from phagosomes or traverse to an adjacent cell is compelling evidence of the yeast's adaptation to an intracellular lifestyle. NLE occurs after phagosomal maturation and requires fungal viability (100–102). Curiously, phagosomal permeability always precedes NLE, whereas actin flashes around the phagosome seem to counteract fungal escape (Figure 7) (103). Interference with host cytoskeletal machinery decreases NLE (104), and yeast cells have been found to interact with host cytoskeletal Rac1, a small GTP-binding Rho family protein, to penetrate the blood-brain barrier (105), indicating that the host cytoskeleton can be subverted to promote fungal escape. The most surprising feature of NLE is how little macrophage damage ensues immediately afterward, with the exception of giant vacuole formation in the cytoplasm of the host cell (100). NLE appears to be tightly modulated by macrophage permissiveness. Macrophages activated by Th2 cytokines in vitro showed an increase in intracellular proliferation and a decrease in extrusion rate when compared with nonstimulated macrophages (106). Th2 cytokines enhance iron uptake and storage by macrophages (107), which may transform the phagosome into a more hospitable environment for the yeast. As mentioned above, acidification of the phagosome is beneficial for C. neoformans (65), and blockage of acidification increases NLE rates (102, 104). These results could be interpreted to suggest the curious hypothesis that a less favorable intracellular niche leads to increased fungal escape via NLE. Trojan Horse Hypothesis for Extrapulmonary Dissemination The Trojan horse hypothesis posits that a pathogen gains entry into the blood-brain barrier through dissemination within immune cells (108). In this scenario, the macrophage functions as a Trojan horse, carrying the fungus throughout the body and contributing to dissemination and the breaching of epithelial and endothelial barriers. For C. neoformans, a Trojan horse mechanism for dissemination is supported by the observation that depletion of alveolar macrophages prevents brain dissemination (47). Similarly, injection of ex vivo infected macrophages into mice resulted in increased brain fungal burden (46). However, alternative mechanisms of penetration into the brain are possible, such as active penetration of endothelial cells, by either a transcellular or paracellular mechanism (64, 109), and C. neoformans proteins that contribute to differential lung/brain infection ratios have been identified in a mutant screen (110, 111). For example, phospholipase B mutants have reduced virulence and invasion of the brain (112). Phospholipase B was found to interact with host cytoskeletal Rac1 to promote brain invasion (105), supporting the idea that C. neoformans may use transcellular mechanisms in addition to the Trojan-horse mechanism. Cryptococcus Neoformans is An Intracellular Pathogen Establishment of a latent intracellular residency is a very common outcome after phagocyte– fungal cell interactions (see sidebar, The Amoeba-Macrophage Connection). Although C. neoformans is not an obligate intracellular pathogen, intracellular residency is an environment where C. neoformans can persist and even travel, if we attribute brain invasion to migratory infected macrophages. However, it remains unclear why latency, and not eradication of infection, is such a common outcome. One explanation postulated the damage-response framework (113), which was further developed with the tolerance hypothesis (114). According to the tolerance hypothesis, resistance mechanisms minimize pathogen burden, whereas tolerance mechanisms maximize host function without affecting microbe burden. Consequently, brain, lungs, and heart are the most susceptible organs to immune damage (114). Two of these organs are major targets of C. neoformans. In chronic infection models, the yeast spreads to spleen and liver early in infection but is later cleared (42), consistent with a lower risk to these organs of immune damage (due in part to their greater regenerative capacities). Thus, control of infection by intracellular latency, but not clearance, might be a tolerance mechanism to minimize brain and lung damage. Within this postulate, intracellular residency is a tolerance mechanism that would minimize both direct fungal damage to the host and exposure of fungi to the immune response (which would trigger immunopathology), allowing maximal host function (114). These considerations raise the question of why C. neoformans has particular tropism for the lung and brain, but not the heart, for which we cannot formulate a credible explanation. When cryptococcal pathogenesis is viewed in the context of the tolerance hypothesis, it appears that fungal intracellular residence is an outcome that presents advantages to both organisms. Conclusions and Unresolved Questions There is still much to be discovered regarding the survival of C. neoformans within macrophages and its capacity for lung intracellular residence in pathogenesis. Most cases of cryptococcosis are initiated by lung pathology, a finding that provides evidence consistent with a pulmonary reservoir for latency. However, animal studies show that dissemination to the brain occurs shortly after pulmonary infection, which suggests that the brain could also be a reservoir for the yeast. If so, how does the yeast establish latency within the immunoprivileged brain, and are there particular mechanisms of fungal control within the brain? Within the lung, yeast control is achieved through the formation of specialized granulomas. Granulomas originate from immune cell cooperation, including macrophages and granulomas generated in vitro that have already been used as a C. albicans infection model (115). That macrophage granulomas and giant cells possess cellular and molecular characteristics distinct from those of macrophages (116) could explain the observed reduced fungicidal capacity of macrophages in vitro. An alternative explanation could be that immune cells must cooperate, meaning that macrophages would have to acquire a microbicidal molecule from other immune cells. Microbial ligands can activate innate immunity in the absence of adequate adaptive immunity (117). Our results (51) have shown an increase in fungicidal activity due to β-glucan stimulation. Given that protection could be elicited with proper innate cell stimulation, without the need for CD4+ T cells, we suggest that microbial ligands might have therapeutic value, in particular for immunocompromised patients in whom proper T cell stimulation is not possible. In conclusion, C. neoformans is capable of surviving within mammalian hosts, contained within the intracellular environment of macrophages. The intracellular residency might reflect the most advantageous equilibrium for the host and the pathogen duo and seems to have evolved serendipitously from an ancient relationship with amoebae. Understanding the features of intracellular life can help to prevent C. neoformans–associated deaths. The authors acknowledge Julie M. Wolf for help in obtaining the TEM images and for invaluable critical reading of the manuscript. We also acknowledge all the personnel at the Analytical Imaging Facility, National Cancer Institute support grant P30CA013330, for their technical assistance on the electron microscopy images. This work was supported by NIH grants HL059842-3, A1033774, A1052733, and AI033142 to A.C. and PhD grant SFRH/BD/33471/2008 by Fundação Ciência e Tecnologia to C.C. Figure 1 Histopathology of Cryptococcus neoformans lung infection. Photomicrographs of lung tissue from Balb/c mice infected with C. neoformans (blue arrowheads), stained with hematoxylin and eosin. (a) Initial infection, showing diffuse pneumonitis and infiltration of immune cells and yeast into the alveolar space (200×). (b) Typical granuloma formation 5 days postinfection (200×). (c) Typical granuloma formation 15 days postinfection (100×). (d) Magnification of panel c, showing the presence of histiocytes (red arrowheads) (400×). (e) At later stages of infection, giant cells (yellow arrowhead) contain C. neoformans (400×). (f) C. neoformans replicating within the alveolar space, visualized by periodic acid–Schiff stain (400×). Figure 2 Schematic of recognition of Cryptococcus neoformans by immune cells. Recognition of C. neoformans by immune cells depends on several receptors and extensive cross talk between those receptors. Recognition of capsular components was determined in isolation and likely also occurs for the whole capsule. Most of these receptors are not opsonic, meaning they cannot mediate ingestion. The in vivo opsonins are thought to be serum components iC3b and C5, such that the yeast is ingested via cooperation between complement receptors, FcRs, and possibly Dectin-1. Abbreviations: FcR, Fc receptor; MR, mannose receptor; TLR, Toll-like receptor. Figure 3 Scanning electron micrographs showing Cryptococcus neoformans and macrophage interaction in vitro. Bone marrow–derived macrophages were infected with antibody-opsonized C. neoformans, and macrophage membranes are shown interacting with yeast cells. (a) Yeast cells are recognized when macrophage membranes probe the extracellular environment around them. (b) Capsulated yeast cells are ingested as the macrophage membrane engulfs them. (c) Ingestion is finalized when the membrane closes upon the yeast cell; a neighboring extracellular yeast is also shown. Panel a courtesy of Sabriya Stukes; panels b and c acquired with the help of Julie M. Wolf. Figure 4 Schematic of immune signaling cascades triggered by Cryptococcus neoformans recognition. (a) Dectin-1 signaling pathway. Dectin-1 can induce both Syk-dependent and Raf (Syk-independent) pathways. Dectin-1 can activate macrophages through the Syk pathway, triggering phagocytosis; following phagocytosis, Dectin-1 activation, coupled to ROS production, contributes to inflammasome activation or fungal killing and activates the transcription factor NF-κB through CARD9, triggering inflammatory cytokine production. The Raf-1 (Syk-independent) pathway enhances NF-κB and inflammatory cytokines. (b) Inflammasome pathway. The Syk-dependent pathway requires combination of two signals. The first signal, which can be mediated by TLR activation, together with a second signal, such as ROS production and/or lysosomal damage, induces the oligomerization of the NLRP3 complex, activation of caspase 1, and production of IL-1β. Abbreviations: ASC, apoptosis-associated speck-like protein containing a C-terminal CARD; Bcl10, B cell leukemia/lymphoma 10; CARD9, caspase recruitment domain–containing protein 9; CLR, C-type lectin receptor; IL, interleukin; MALT-1, mucosa-associated lymphoid tissue 1; NF-κB, nuclear factor κ-light-chain enhancer of activated B cells; NLRP3, Nod-like receptor family, pyrin domain–containing 3; PLCγ2, phospholipase Cγ2; ROS, reactive oxygen species; Syk, spleen tyrosine kinase; TLR, Toll-like receptor; TNF-α, tumor necrosis factor α Figure 5 Transmission electron micrographs showing Cryptococcus neoformans and macrophage interaction in vitro. Blue arrowheads indicate possible lysosomal fusion events. (a) Macrophage with ingested C. neoformans. (b) Magnification of panel a, highlighting macrophage organelles, particularly lysosomes, in proximity with the phagosome. (c) C. neoformans budding within a phagosome. (d) Magnification of panel c, displaying C. neoformans organelles. Abbreviations: L, lysosome; M, mitochondrion; Nu, nucleus. Figure 6 Phagocytic events upon Cryptococcus neoformans ingestion. To date, no manipulation of the phagocytic compartment by C. neoformans has been described. The interplay between macrophage fungicidal mechanisms and C. neoformans results in host damage, mainly to the phagosomal compartment and to the regulation of the host cell cycle. Abbreviations: ROS, reactive oxygen species; RNS, reactive nitrogen species. Figure 7 Possible outcomes for Cryptococcus neoformans infection of murine macrophages. The interaction between C. neoformans and host macrophages can result in different outcomes, and the frequency with which they occur influences the course of infection. Table 1 Role of immune components in mouse model of cryptococcosis Type Immune polarization Immune component Infection route Outcome when removeda Reference(s) Recognition/binding — CD36 i.v. Decreased survival 53 Mannose receptor i.n. Decreased survival 59 C5 i.v. Decreased survivalb 14 C3 i.v. Decreased survival 52 TLR2 i.p. Decreased survival 58 TLR4 i.p. No or limited effect 58, 123 TLR9 i.n. Decreased survival 61 Dectin-1 i.v. No effect 56 Immune cells — Neutrophils i.t. Increased survival 30 Eosinophils i.t., i.n. Increased survivalc 29, 31 Macrophages i.v. Decreased survival 26 Macrophages/dendritic cells i.t, i.v. Decreased survival 27 B cells i.v. No effectd 124 CD4+ T cells i.t. Decreased survivalc 11, 125 CD8+ T cells i.v. Decreased survival 11, 126 Cytokines Th1 IL-12 i.v. Decreased survival 17 Th1 IFN-γ i.v., i.t Decreased survival 15, 126 Th1 IL-18 i.n., i.t. Decreased survival 61, 127 Th1 TNF-α i.t. No effect 23 Th2 IL-13 i.n. Increased survival 128 Th2 IL-4 i.v. Increased survival 17 Th1/Th17 IL-23 i.p, i.v Decreased survival 129 Th17 IL-17 i.n. Important for early response 25 Th2 uPA i.t. Decreased survival 130 Antiinflammatory IL-10 i.v. Decreased survival 131 Inflammatory IL-6 i.v. Earlier death 131 Antiinflammatory TGF-β i.n. Mixed resultsc 132 Effector molecules — Nitric oxide synthase i.t Decreased survival rate 86 NADPH oxidase i.n. Increased survivalc 82 Myeloperoxidase i.n., i.v. Decreased survival 92 Signaling molecules — MyD88 i.p. Earlier death 58 Abbreviations: IFN, interferon; IL, interleukin; i.n., intranasal; i.p., intraperitoneal; i.t., intratracheal; i.v., intravenous; TGF, transforming growth factor; TLR, Toll-like receptor; TNF, tumor necrosis factor; uPA, urokinase-type plasminogen activator. a Compared with wild-type mice. b Inferred from correlating mouse strain susceptibility with presence of C5. c No survival study performed, but conclusion is supported by pathology and cytokine profile. d According to Reference 133, B cells play a role in regulating immunity and establishing protection. The Amoeba-Macrophage Connection Cryptococcus neoformans is a soil organism that has no requirement for mammalian pathogenesis in its life cycle. Why would C. neoformans develop such a sophisticated intracellular pathogenic strategy? Studies of the interaction of C. neoformans with amoebae suggested how this strategy might have evolved. Amoebae are predators on C. neoformans in soil, which can be replicated in a laboratory setting (118, 119). Analysis of the interactions of C. neoformans with Acanthamoeba castellanii revealed remarkable similarities to the response elicited by interaction with mammalian macrophages; similar virulence factors are required for pathogenesis in both hosts (120). Subsequent studies have established that other phenomena associated with the interaction of macrophages, such as capsule growth and NLE, can be replicated in C. neoformans–amoeba interactions (121, 122). On the basis of these observations, the capacity of C. neoformans to survive in macrophages and cause disease in mammals was proposed to be the result of selection by such biotic factors as amoebae in the environment (113). According to this synthesis, environmental pressures selected for traits that were needed to survive phagocytic predators and that incidentally also conferred the capacity for mammalian virulence (113). Future Issues What is the mechanism of control of the primary C. neoformans infection? What immune effector mechanism, lost during immunosuppression or loss of CD4+ T cells, is responsible for control of latent C. neoformans infection? Are the mechanisms responsible for control of a primary infection the same as those that will control an established disseminated infection? Is it possible to prevent C. neoformans from crossing the blood-brain barrier? Disclosure Statement: The authors are not aware of any affiliations, memberships, funding, or financial holdings that might be perceived as affecting the objectivity of this review. Literature Cited 1 Barnett JA 2010 A history of research on yeasts 14: medical yeasts part 2, Cryptococcus neoformans Yeast 27 11 875 904 20641025 2 Lin X Heitman J 2006 The biology of the Cryptococcus neoformans species complex Annu Rev Microbiol 60 69 105 16704346 3 Lindell DM Ballinger MN McDonald RA Toews GB Huffnagle GB 2006 Immunologic homeostasis during infection: coexistence of strong pulmonary cell–mediated immunity to secondary Cryptococcus neoformans infection while the primary infection still persists at low levels in the lungs J Immunol 177 7 4652 61 16982904 4 Goldman DL Khine H Abadi J Lindenberg DJ Pirofski La 2001 Serologic evidence for Cryptococcus neoformans infection in early childhood Pediatrics 107 5 e66 11331716 5 Garcia-Hermoso D Janbon G Dromer F 1999 Epidemiological evidence for dormant Cryptococcus neoformans infection J Clin Microbiol 37 10 3204 9 10488178 6 Longo D Fauci A Kasper D Hauser S Jameson J Loscalzo J 2011 Harrison's Principles of Internal Medicine 1, 2 New York McGraw-Hill 18th 7 Chen LC Goldman DL Doering TL Pirofski La Casadevall A 1999 Antibody response to Cryptococcus neoformans proteins in rodents and humans Infect Immun 67 5 2218 24 10225877 8 Giles SS Dagenais TRT Botts MR Keller NP Hull CM 2009 Elucidating the pathogenesis of spores from the human fungal pathogen Cryptococcus neoformans Infect Immun 77 8 3491 500 19451235 9 Shibuya K Hirata A Omuta J Sugamata M Katori S 2005 Granuloma and cryptococcosis J Infect Chemother 11 3 115 22 15990974 10 Lindell DM Ballinger MN McDonald RA Toews GB Huffnagle GB 2006 Diversity of the T-cell response to pulmonary Cryptococcus neoformans infection Infect Immun 74 8 4538 48 16861640 11 Huffnagle GB Yates JL Lipscomb MF 1991 Immunity to a pulmonary Cryptococcus neoformans infection requires both CD4+ and CD8+ T cells J Exp Med 173 4 793 800 1672543 12 Lim TS Murphy JW 1980 Transfer of immunity to cryptococcosis by T-enriched splenic lymphocytes from Cryptococcus neoformans-sensitized mice Infect Immun 30 1 5 11 7002791 13 Zaragoza O Alvarez M Telzak A Rivera J Casadevall A 2007 The relative susceptibility of mouse strains to pulmonary Cryptococcus neoformans infection is associated with pleiotropic differences in the immune response Infect Immun 75 6 2729 39 17371865 14 Rhodes JC 1985 Contribution of complement component C5 to the pathogenesis of experimental murine cryptococcosis Sabouraudia 23 3 225 34 4023888 15 Hoag KA Lipscomb MF Izzo AA Street NE 1997 IL-12 and IFN-γ are required for initiating the protective Th1 response to pulmonary cryptococcosis in resistant C.B-17 mice Am J Respir Cell Mol Biol 17 6 733 39 9409560 16 Chen GH McDonald RA Wells JC Huffnagle GB Lukacs NW Toews GB 2005 The γinterferon receptor is required for the protective pulmonary inflammatory response to Cryptococcus neoformans Infect Immun 73 3 1788 96 15731080 17 Decken K Köhler G Palmer-Lehmann K Wunderlin A Mattner F 1998 Interleukin-12 is essential for a protective Th1 response in mice infected with Cryptococcus neoformans Infect Immun 66 10 4994 5000 9746609 18 Kobayashi M Ito M Sano K Koyama M 2001 Granulomatous and cytokine responses to pulmonary Cryptococcus neoformans in two strains of rats Mycopathologia 151 3 121 30 11678586 19 Arora S Olszewski MA Tsang TM McDonald RA Toews GB Huffnagle GB 2011 Effect of cytokine interplay on macrophage polarization during chronic pulmonary infection with Cryptococcus neoformans Infect Immun 79 5 1915 26 21383052 20 Jain AV Zhang Y Fields WB McNamara DA Choe MY 2009 Th2 but not Th1 immune bias results in altered lung functions in a murine model of pulmonary Cryptococcus neoformans infection Infect Immun 77 12 5389 99 19752036 21 Zhang Y Wang F Tompkins KC McNamara A Jain AV 2009 Robust Th1 and Th17 immunity supports pulmonary clearance but cannot prevent systemic dissemination of highly virulent Cryptococcus neoformans H99 Am J Pathol 175 6 2489 500 19893050 22 Stenzel W Müller U Köhler G Heppner FL Blessing M 2009 IL-4/IL-13-dependent alternative activation of macrophages but not microglial cells is associated with uncontrolled cerebral cryptococcosis Am J Pathol 174 2 486 96 19147811 23 Kawakami K Qifeng X Tohyama M Qureshi MH Saito A 1996 Contribution of tumour necrosis factor-alpha (TNF-α) in host defence mechanism against Cryptococcus neoformans Clin Exp Immunol 106 3 468 74 8973614 24 Conti HR Gaffen SL 2010 Host responses to Candida albicans: Th17 cells and mucosal candidiasis Microbes Infect 12 7 518 27 20381638 25 Wozniak KL Young ML Wormley FL 2011 Protective immunity against experimental pulmonary cryptococcosis in T cell–depleted mice Clin Vaccine Immunol 18 5 717 23 21450975 26 Monga DP 1981 Role of macrophages in resistance of mice to experimental cryptococcosis Infect Immun 32 3 975 78 6265377 27 Osterholzer JJ Milam JE Chen GH Toews GB Huffnagle GB Olszewski MA 2009 Role of dendritic cells and alveolar macrophages in regulating early host defense against pulmonary infection with Cryptococcus neoformans Infect Immun 77 9 3749 58 19564388 28 Shao X Mednick A Alvarez M van Rooijen N Casadevall A Goldman DL 2005 An innate immune system cell is a major determinant of species-related susceptibility differences to fungal pneumonia J Immunol 175 5 3244 51 16116215 29 Feldmesser M Casadevall A Kress Y Spira G Orlofsky A 1997 Eosinophil–Cryptococcus neoformans interactions in vivo and in vitro Infect Immun 65 5 1899 907 9125578 30 Mednick AJ Feldmesser M Rivera J Casadevall A 2003 Neutropenia alters lung cytokine production in mice and reduces their susceptibility to pulmonary cryptococcosis Eur J Immunol 33 6 1744 53 12778493 31 Piehler D Stenzel W Grahnert A Held J Richter L 2011 Eosinophils contribute to IL-4 production and shape the T-helper cytokine profile and inflammatory response in pulmonary cryptococcosis Am J Pathol 179 2 733 44 21699881 32 Garro AP Chiapello LS Baronetti JL Masih DT 2010 Rat eosinophils stimulate the expansion of Cryptococcus neoformans-specific CD4+ and CD8+ T cells with a T-helper 1 profile Immunology 132 2 174 87 21039463 33 Casadevall A 2012 Amoeba provide insight into the origin of virulence in pathogenic fungi Adv Exp Med Biol 710 1 10 22127880 34 Mansour MK Yauch LE Rottman JB Levitz SM 2004 Protective efficacy of antigenic fractions in mouse models of cryptococcosis Infect Immun 72 3 1746 54 14977983 35 Kawakami K 2004 Regulation by innate immune T lymphocytes in the host defense against pulmonary infection with Cryptococcus neoformans Jpn J Infect Dis 57 4 137 45 15329444 36 Hardison SE Wozniak KL Kolls JK Wormley FL 2010 Interleukin-17 is not required for classical macrophage activation in a pulmonary mouse model of Cryptococcus neoformans infection Infect Immun 78 12 5341 51 20921149 37 Hardison SE Herrera G Young ML Hole CR Wozniak KL Wormley FL 2012 Protective immunity against pulmonary cryptococcosis is associated with STAT1-mediated classical macrophage activation J Immunol 189 8 4060 68 22984078 38 Hardison SE Ravi S Wozniak KL Young ML Olszewski MA Wormley FL 2010 Pulmonary infection with an interferon-γ-producing Cryptococcus neoformans strain results in classical macrophage activation and protection Am J Pathol 176 2 774 85 20056835 39 Baker RD Haugen RK 1955 Tissue changes and tissue diagnosis in cryptococcosis; a study of 26 cases Am J Clin Pathol 25 1 14 24 14349908 40 Schwartz DA 1988 Characterization of the biological activity of Cryptococcus infections in surgical pathology. The Budding Index and Carminophilic Index Ann Clin Lab Sci 18 5 388 97 2460017 41 Shibuya K Coulson WE Wollman JS Wakayama M Ando T 2001 Histopathology of cryptococcosis and other fungal infections in patients with acquired immunodeficiency syndrome Int J Infect Dis 5 2 78 85 11468102 42 Goldman DL Lee SC Mednick AJ Montella L Casadevall A 2000 Persistent Cryptococcus neoformans pulmonary infection in the rat is associated with intracellular parasitism, decreased inducible nitric oxide synthase expression, and altered antibody responsiveness to cryptococcal polysaccharide Infect Immun 68 2 832 38 10639453 43 Feldmesser M Kress Y Novikoff P Casadevall A 2000 Cryptococcus neoformans is a facultative intracellular pathogen in murine pulmonary infection Infect Immun 68 7 4225 37 10858240 44 Nessa K Gross NT Jarstrand C Johansson A Camner P 1997 In vivo interaction between alveolar macrophages and Cryptococcus neoformans Mycopathologia 139 1 1 7 9511230 45 Goldman D Lee SC Casadevall A 1994 Pathogenesis of pulmonary Cryptococcus neoformans infection in the rat Infect Immun 62 11 4755 61 7927751 46 Charlier C Nielsen K Daou S Brigitte M Chrétien F Dromer F 2009 Evidence of a role for monocytes in dissemination and brain invasion by Cryptococcus neoformans Infect Immun 77 1 120 27 18936186 47 Kechichian TB Shea J Del Poeta M 2007 Depletion of alveolar macrophages decreases the dissemination of a glucosylceramide-deficient mutant of Cryptococcus neoformans in immunodeficient mice Infect Immun 75 10 4792 98 17664261 48 Cross CE Bancroft GJ 1995 Ingestion of acapsular Cryptococcus neoformans occurs via mannose and β-glucan receptors, resulting in cytokine production and increased phagocytosis of the encapsulated form Infect Immun 63 7 2604 11 7790075 49 Giles SS Zaas AK Reidy MF Perfect JR Wright JR 2007 Cryptococcus neoformans is resistant to surfactant protein A mediated host defense mechanisms PLoS ONE 2 12 e1370 18159253 50 Taborda CP Casadevall A 2002 CR3 (CD11b/CD18) and CR4 (CD11c/CD18) are involved in complement-independent antibody-mediated phagocytosis of Cryptococcus neoformans Immunity 16 6 791 802 12121661 51 Bocca AL Coelho C Casadevall A 2012 Dectin-1 pathway activation is important to Cryptococcus neoformans killing J Immunol 188 Meet. Abstr. Suppl 67.10 http://www.jimmunol.org/cgi/content/meeting_abstract/188/1_MeetingAbstracts/67.10 52 Shapiro S Beenhouwer DO Feldmesser M Taborda C Carroll MC 2002 Immunoglobulin G monoclonal antibodies to Cryptococcus neoformans protect mice deficient in complement component C3 Infect Immun 70 5 2598 604 11953401 53 Means TK Mylonakis E Tampakakis E Colvin RA Seung E 2009 Evolutionarily conserved recognition and innate immunity to fungal pathogens by the scavenger receptors SCARF1 and CD36 J Exp Med 206 3 637 53 19237602 54 Heinsbroek SEM Taylor PR Martinez FO Martinez-Pomares L Brown GD Gordon S 2008 Stage-specific sampling by pattern recognition receptors during Candida albicans phagocytosis PLoS Pathog 4 11 e1000218 19043561 55 Faro-Trindade I Willment JA Kerrigan AM Redelinghuys P Hadebe S 2012 Characterisation of innate fungal recognition in the lung PLoS ONE 7 4 e35675 22536422 56 Nakamura K Kinjo T Saijo S Miyazato A Adachi Y 2007 Dectin-1 is not required for the host defense to Cryptococcus neoformans Microbiol Immunol 51 11 1115 19 18037789 57 Marakalala MJ Vautier S Potrykus J Walker LA Shepardson KM 2013 Differential adaptation of Candida albicans in vivo modulates immune recognition by Dectin-1 PLoS Pathog 9 4 e1003315 23637604 58 Biondo C Midiri A Messina L Tomasello F Garufi G 2005 MyD88 and TLR2, but not TLR4, are required for host defense against Cryptococcus neoformans Eur J Immunol 35 3 870 78 15714580 59 Dan JM Kelly RM Lee CK Levitz SM 2008 Role of the mannose receptor in a murine model of Cryptococcus neoformans infection Infect Immun 76 6 2362 67 18391001 60 Nakamura K Miyazato A Xiao G Hatta M Inden K 2008 Deoxynucleic acids from Cryptococcus neoformans activate myeloid dendritic cells via a TLR9-dependent pathway J Immunol 180 6 4067 74 18322216 61 Wang JP Lee CK Akalin A Finberg RW Levitz SM 2011 Contributions of the MyD88-dependent receptors IL-18R, IL-1R, and TLR9 to host defenses following pulmonary challenge with Cryptococcus neoformans PLoS ONE 6 10 e26232 22039448 62 Guillot L Carroll SF Badawy M Qureshi ST 2008 Cryptococcus neoformans induces IL-8 secretion and CXCL1 expression by human bronchial epithelial cells Respir Res 9 1 9 18211687 63 Ganendren R Carter E Sorrell T Widmer F Wright L 2006 Phospholipase B activity enhances adhesion of Cryptococcus neoformans to a human lung epithelial cell line Microbes Infect 8 4 1006 15 16487740 64 Sabiiti W May RC 2012 Capsule independent uptake of the fungal pathogen Cryptococcus neoformans into brain microvascular endothelial cells PLoS ONE 7 4 e35455 22530025 65 Qin QM Luo J Lin X Pei J Li L 2011 Functional analysis of host factors that mediate the intracellular lifestyle of Cryptococcus neoformans PLoS Pathog 7 6 e1002078 21698225 66 Artavanis-Tsakonas K Love JC Ploegh HL Vyas JM 2006 Recruitment of CD63 to Cryptococcus neoformans phagosomes requires acidification Proc Natl Acad Sci USA 103 43 15945 50 17043215 67 Levitz SM Harrison TS Tabuni A Liu X 1997 Chloroquine induces human mononuclear phagocytes to inhibit and kill Cryptococcus neoformans by a mechanism independent of iron deprivation J Clin Investig 100 6 1640 46 9294133 68 Wozniak KL Levitz SM 2008 Cryptococcus neoformans enters the endolysosomal pathway of dendritic cells and is killed by lysosomal components Infect Immun 76 10 4764 71 18678670 69 Nicola AM Albuquerque P Martinez LR Dal-Rosso RA Saylor C 2012 Macrophage autophagy in immunity to Cryptococcus neoformans and Candida albicans Infect Immun 80 9 3065 76 22710871 70 Tucker SC Casadevall A 2002 Replication of Cryptococcus neoformans in macrophages is accompanied by phagosomal permeabilization and accumulation of vesicles containing polysaccharide in the cytoplasm Proc Natl Acad Sci USA 99 5 3165 70 11880650 71 Gross NT Hultenby K Mengarelli S Camner P Jarstrand C 2000 Lipid peroxidation by alveolar macrophages challenged with Cryptococcus neoformans, Candida albicans or Aspergillus fumigatus Med Mycol 38 6 443 49 11204882 72 Ben-Abdallah M Sturny-Leclère A Avé P Louise A Moyrand F 2012 Fungal-induced cell cycle impairment, chromosome instability and apoptosis via differential activation of NF-κB PLoS Pathog 8 3 e1002555 22396644 73 Villena SN Pinheiro RO Pinheiro CS Nunes MP Takiya CM 2008 Capsular polysaccharides galactoxylomannan and glucuronoxylomannan from Cryptococcus neoformans induce macrophage apoptosis mediated by Fas ligand Cell Microbiol 10 6 1274 85 18284419 74 Luo Y Tucker SC Casadevall A 2005 Fc- and complement-receptor activation stimulates cell cycle progression of macrophage cells from G1 to S J Immunol 174 11 7226 33 15905568 75 Luo Y Casadevall A 2010 Intracellular cryptococci suppress Fc-mediated cyclin D1 elevation Commun Integr Biol 3 4 390 91 20798836 76 Coelho C Tesfa L Zhang J Rivera J Gonçalves T Casadevall A 2012 Analysis of cell cycle and replication of mouse macrophages after in vivo and in vitro Cryptococcus neoformans infection using laser scanning cytometry Infect Immun 80 4 1467 78 22252872 77 Thi EP Lambertz U Reiner NE 2012 Sleeping with the enemy: how intracellular pathogens cope with a macrophage lifestyle PLoS Pathog 8 3 e1002551 22457616 78 Levitz SM Farrell TP 1990 Growth inhibition of Cryptococcus neoformans by cultured human monocytes: role of the capsule, opsonins, the culture surface, and cytokines Infect Immun 58 5 1201 9 2182538 79 Shoham S Levitz SM 2005 The immune response to fungal infections Br J Haematol 129 5 569 82 15916679 80 Derengowski LS Paes HC Albuquerque P Tavares AHFP Fernandes L 2013 The transcriptional response of Cryptococcus neoformans to ingestion by Acanthamoeba castellanii and macrophages provides insights into the evolutionary adaptation into the mammalian host Eukaryot Cell 12 5 761 74 23524994 81 Fan W Kraus PR Boily MJ Heitman J 2005 Cryptococcus neoformans gene expression during murine macrophage infection Eukaryot Cell 4 8 1420 33 16087747 82 Snelgrove RJ Edwards L Williams AE Rae AJ Hussell T 2006 In the absence of reactive oxygen species, T cells default to a Th1 phenotype and mediate protection against pulmonary Cryptococcus neoformans infection J Immunol 177 8 5509 16 17015737 83 Alspaugh JA Granger DL 1991 Inhibition of Cryptococcus neoformans replication by nitrogen oxides supports the role of these molecules as effectors of macrophage-mediated cytostasis Infect Immun 59 7 2291 96 2050398 84 Goldman D Cho Y Zhao M Casadevall A Lee SC 1996 Expression of inducible nitric oxide synthase in rat pulmonary Cryptococcus neoformans granulomas Am J Pathol 148 4 1275 82 8644868 85 Rivera J Mukherjee J Weiss LM Casadevall A 2002 Antibody efficacy in murine pulmonary Cryptococcus neoformans infection: a role for nitric oxide J Immunol 168 7 3419 27 11907100 86 Aguirre KM Gibson GW 2000 Differing requirement for inducible nitric oxide synthase activity in clearance of primary and secondary Cryptococcus neoformans infection Med Mycol 38 5 343 53 11092381 87 de Jesús-Berríos M Liu L Nussbaum JC Cox GM Stamler JS Heitman J 2003 Enzymes that counteract nitrosative stress promote fungal virulence Curr Biol 13 22 1963 68 14614821 88 Murphy JW Hidore MR Wong SC 1993 Direct interactions of human lymphocytes with the yeast-like organism, Cryptococcus neoformans J Clin Investig 91 4 1553 66 8473499 89 Levitz SM Dupont MP Smail EH 1994 Direct activity of human T lymphocytes and natural killer cells against Cryptococcus neoformans Infect Immun 62 1 194 202 8262627 90 Kelly RM Chen J Yauch LE Levitz SM 2005 Opsonic requirements for dendritic cell-mediated responses to Cryptococcus neoformans Infect Immun 73 1 592 98 15618199 91 Mambula SS Simons ER Hastey R Selsted ME Levitz SM 2000 Human neutrophil-mediated nonoxidative antifungal activity against Cryptococcus neoformans Infect Immun 68 11 6257 64 11035733 92 Aratani Y 2006 Contribution of the myeloperoxidase-dependent oxidative system to host defence against Cryptococcus neoformans J Med Microbiol 55 9 1291 99 16914663 93 Hole CR Bui H Wormley FL Wozniak KL 2012 Mechanisms of dendritic cell lysosomal killing of Cryptococcus Sci Rep 2 739 23074646 94 Laskin DL Sunil VR Gardner CR Laskin JD 2011 Macrophages and tissue injury: agents of defense or destruction? Annu Rev Pharmacol Toxicol 51 267 88 20887196 95 Trost M English L Lemieux S Courcelles M Desjardins M Thibault P 2009 The phagosomal proteome in interferon-γ-activated macrophages Immunity 30 1 143 54 19144319 96 Bliska JB Casadevall A 2009 Intracellular pathogenic bacteria and fungi—a case of convergent evolution? Nat Rev Microbiol 7 165 71 19098923 97 Zaragoza O García-Rodas R Nosanchuk JD Cuenca-Estrella M Rodríguez-Tudela JL Casadevall A 2010 Fungal cell gigantism during mammalian infection PLoS Pathog 6 6 e1000945 20585557 98 Okagaki LH Strain AK Nielsen JN Charlier C Baltes NJ 2010 Cryptococcal cell morphology affects host cell interactions and pathogenicity PLoS Pathog 6 6 e1000953 20585559 99 Alvarez M Burn T Luo Y Pirofski La Casadevall A 2009 The outcome of Cryptococcus neoformans intracellular pathogenesis in human monocytes BMC Microbiol 9 51 19265539 100 Alvarez M Casadevall A 2006 Phagosome extrusion and host-cell survival after Cryptococcus neoformans phagocytosis by macrophages Curr Biol 16 21 2161 65 17084702 101 Ma H Croudace JE Lammas DA May RC 2006 Expulsion of live pathogenic yeast by macrophages Curr Biol 16 21 2156 60 17084701 102 Nicola AM Robertson EJ Albuquerque P Derengowski LDS Casadevall A 2011 Nonlytic exocytosis of Cryptococcus neoformans from macrophages occurs in vivo and is influenced by phagosomal pH mBio 2 4 e00167 11 21828219 103 Johnston SA May RC 2010 The human fungal pathogen Cryptococcus neoformans escapes macrophages by a phagosome emptying mechanism that is inhibited by Arp2/3 complex–mediated actin polymerisation PLoS Pathog 6 8 e1001041 20714349 104 Carnell M Zech T Calaminus SD Ura S Hagedorn M 2011 Actin polymerization driven by WASH causes V-ATPase retrieval and vesicle neutralization before exocytosis J Cell Biol 193 5 831 39 21606208 105 Maruvada R Zhu L Pearce D Zheng Y Perfect J 2012 Cryptococcus neoformans phospholipase B1 activates host cell Rac1 for traversal across the blood-brain barrier Cell Microbiol 14 10 1544 53 22646320 106 Voelz K Lammas DA May RC 2009 Cytokine signaling regulates the outcome of intracellular macrophage parasitism by Cryptococcus neoformans Infect Immun 77 8 3450 57 19487474 107 Weiss G Bogdan C Hentze MW 1997 Pathways for the regulation of macrophage iron metabolism by the anti-inflammatory cytokines IL-4 and IL-13 J Immunol 158 1 420 25 8977218 108 Kim WK Corey S Alvarez X Williams K 2003 Monocyte/macrophage traffic in HIV and SIV encephalitis J Leukoc Biol 74 5 650 56 12960230 109 Chang YC Stins MF McCaffery MJ Miller GF Pare DR 2004 Cryptococcal yeast cells invade the central nervous system via transcellular penetration of the blood-brain barrier Infect Immun 72 9 4985 95 15321990 110 He X Lyons DM Toffaletti DL Wang F Qiu Y 2012 Virulence factors identified by Cryptococcus neoformans mutant screen differentially modulate lung immune responses and brain dissemination Am J Pathol 181 4 1356 66 22846723 111 Rao PK Singh CR Jagannath C Li Q 2009 A systems biology approach to study the phagosomal proteome modulated by mycobacterial infections Int J Clin Exp Med 2 3 233 47 19918316 112 Cox GM McDade HC Chen SC Tucker SC Gottfredsson M 2001 Extracellular phospholipase activity is a virulence factor for Cryptococcus neoformans Mol Microbiol 39 1 166 75 11123698 113 Casadevall A Pirofski La 2003 The damage-response framework of microbial pathogenesis Nat Rev Microbiol 1 1 17 24 15040176 114 Medzhitov R Schneider DS Soares MP 2012 Disease tolerance as a defense strategy Science 335 6071 936 41 22363001 115 Alvarez-Rueda N Albassier M Allain S Deknuydt F Altare F Le Pape P 2012 First human model of in vitro Candida albicans persistence within granuloma for the reliable study of host-fungi interactions PLoS ONE 7 6 e40185 22768252 116 Vignery A 2008 Macrophage fusion: molecular mechanisms Methods Mol Biol 475 149 61 18979243 117 Evans SE Tuvim MJ Fox CJ Sachdev N Gibiansky L Dickey BF 2011 Inhaled innate immune ligands to prevent pneumonia Br J Pharmacol 163 1 195 206 21250981 118 Bunting LA Neilson JB Bulmer GS 1979 Cryptococcus neoformans: gastronomic delight of a soil ameba Sabouraudia 17 3 225 32 394365 119 Ruiz A Neilson JB Bulmer GS 1982 Control of Cryptococcus neoformans in nature by biotic factors Sabouraudia 20 1 21 29 6801787 120 Steenbergen JN Casadevall A 2003 The origin and maintenance of virulence for the human pathogenic fungus Cryptococcus neoformans Microbes Infect 5 7 667 75 12787743 121 Chrisman CJ Alvarez M Casadevall A 2010 Phagocytosis of Cryptococcus neoformans by, and nonlytic exocytosis from, Acanthamoeba castellanii Appl Environ Microbiol 76 18 6056 62 20675457 122 Chrisman CJ Albuquerque P Guimaräes AJ Nieves E Casadevall A 2011 Phospholipids trigger Cryptococcus neoformans capsular enlargement during interactions with amoebae and macrophages PLoS Pathog 7 5 e1002047 21637814 123 Nakamura K Miyagi K Koguchi Y Kinjo Y Uezu K 2006 Limited contribution of Toll-like receptor 2 and 4 to the host response to a fungal infectious pathogen, Cryptococcus neoformans FEMS Immunol Med Microbiol 47 1 148 54 16706798 124 Monga DP Kumar R Mohapatra LN 1979 Experimental cryptococcosis in normal and B-cell-deficient mice Infect Immun 23 1 1 3 570553 125 Hill JO Aguirre KM 1994 CD4+ T cell–dependent acquired state of immunity that protects the brain against Cryptococcus neoformans J Immunol 152 5 2344 50 7907637 126 Yuan RR Casadevall A Oh J Scharff MD 1997 T cells cooperate with passive antibody to modify Cryptococcus neoformans infection in mice Proc Natl Acad Sci USA 94 6 2483 88 9122221 127 Kawakami K Koguchi Y Qureshi MH Kinjo Y Yara S 2000 Reduced host resistance and Th1 response to Cryptococcus neoformans in interleukin-18 deficient mice FEMS Microbiol Lett 186 1 121 26 10779723 128 Müller U Stenzel W Köhler G Werner C Polte T 2007 IL-13 induces disease-promoting type 2 cytokines, alternatively activated macrophages and allergic inflammation during pulmonary infection of mice with Cryptococcus neoformans J Immunol 179 8 5367 77 17911623 129 Kleinschek MA Müller U Brodie SJ Stenzel W Köhler G 2006 IL-23 enhances the inflammatory cell response in Cryptococcus neoformans infection and induces a cytokine pattern distinct from IL-12 J Immunol 176 2 1098 106 16393998 130 Gyetko MR Chen GH McDonald RA Goodman R Huffnagle GB 1996 Urokinase is required for the pulmonary inflammatory response to Cryptococcus neoformans. A murine transgenic model J Clin Investig 97 8 1818 26 8621764 131 Beenhouwer DO Shapiro S Feldmesser M Casadevall A Scharff MD 2001 Both Th1 and Th2 cytokines affect the ability of monoclonal antibodies to protect mice against Cryptococcus neoformans Infect Immun 69 10 6445 55 11553589 132 Williams AE Humphreys IR Cornere M Edwards L Rae A Hussell T 2005 TGF-β prevents eosinophilic lung disease but impairs pathogen clearance Microbes Infect 7 3 365 74 15784186 133 Casadevall A Pirofski La 2012 Immunoglobulins in defense, pathogenesis, and therapy of fungal diseases Cell Host Microbe 11 5 447 56 22607798

PMC005xxxxxx/PMC5127727.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 7900832 6672 Prog Lipid Res Prog. Lipid Res. Progress in lipid research 0163-7827 1873-2194 27528189 5127727 10.1016/j.plipres.2016.08.003 NIHMS812940 Article LIPID SOMERSAULTS: UNCOVERING THE MECHANISMS OF PROTEIN-MEDIATED LIPID FLIPPING Pomorski Thomas Günther 12* Menon Anant K. 3* 1 Faculty of Chemistry and Biochemistry, Molecular Biochemistry, Ruhr University Bochum, Universitätstrasse 150, D-44780 Bochum, Germany 2 Centre for Membrane Pumps in Cells and Disease - PUMPKIN, Department of Plant and Environmental Sciences, University of Copenhagen, Frederiksberg C, Denmark 3 Department of Biochemistry, Weill Cornell Medical College, 1300 York Avenue, New York, NY 10065, USA * Corresponding authors (tgp@plen.ku.dk; akm2003@med.cornell.edu) 26 8 2016 12 8 2016 10 2016 01 10 2017 64 6984 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. Membrane lipids diffuse rapidly in the plane of the membrane but their ability to flip spontaneously across a membrane bilayer is hampered by a significant energy barrier. Thus spontaneous flip-flop of polar lipids across membranes is very slow, even though it must occur rapidly to support diverse aspects of cellular life. Here we discuss the mechanisms by which rapid flip-flop occurs, and what role lipid flipping plays in membrane homeostasis and cell growth. We focus on conceptual aspects, highlighting mechanistic insights from biochemical and in silico experiments, and the recent, ground-breaking identification of a number of lipid scramblases. flippase floppase membrane asymmetry phosphatidylserine photoreceptor scramblase 1. Introduction The defining feature of a biological membrane is its bilayer structure, a sandwich of two monolayers of phospholipids visible as a trilaminar track by thin-section electron microscopy. The two monolayers are coupled, a fact that is easily revealed by noting that the membrane bends if the number of lipids on one side exceeds that on the other. This is akin to the bending of a bimetallic strip in which the coupled metals have different coefficients of thermal expansion. Gorter and Grendel provided the first experimental evidence for bilayer organization in 1925 when they compared the area occupied by lipids extracted from red blood cells with the predicted area of the cell membrane using surface chemistry approaches [1] [2]. Sheetz and Singer proposed the bilayer couple hypothesis in 1974 [3]. Bilayer membranes form spontaneously when phospholipids are dispersed in water, a consequence of the hydrophobic effect discussed by Tanford [4]. The bilayers formed in this way are usually nested, one within the other, to form a multilamellar structure like the layered skin of an onion. These structures can be dispersed into individual unilamellar vesicles by sonication, or by extrusion through filters after freezing and thawing. The resulting vesicles can be quite small, <50 nanometer in diameter, in which case there are more lipids in the outer leaflet than in the inner leaflet. For large unilamellar vesicles, >150 nanometer in diameter, the number of lipids in each monolayer is almost identical. Phospholipids within an individual leaflet of the bilayer are very dynamic. They undergo a number of intramolecular motions that can be so great as to sometimes bring the methyl end of their acyl chains into the vicinity of the glycerol moiety. They also exhibit rapid rotational and lateral diffusion (Figure 1). Thus, a lipid spins around its own axis, normal to the plane of the membrane, with a characteristic time of ~1 nanosecond, and can diffuse laterally to occupy the position of a neighboring lipid (a displacement on the order of ~1 nanometer) within ~100 nanoseconds. In contrast to these fast movements, phospholipid exchange between the two leaflets of the bilayer occurs only slowly. Thus, reorientation of a phospholipid across the ~3 nanometer thickness of a membrane has a characteristic time of ~100 hours. The low frequency with which phospholipids flip spontaneously across pure lipid bilayers was first reported in 1971 by Kornberg and McConnell [6]. They reconstituted trace quantities of spin-labeled phospholipid analogues into synthetic vesicles, chemically reduced all labeled lipids in the outer leaflet with ascorbate to generate asymmetric vesicles in which the non-reduced lipid probes were located only in the luminal leaflet, and then monitored the translocation of the probes from the inner to the outer leaflet again using ascorbate. They reported a frequency of translocation of ~10−5 seconds−1 at 30°C. By carrying out the measurement at different temperatures they could deduce the activation energy E of the translocation process and also estimate the prefactor A in the Arrhenius rate equation (rate = A•exp (−E/kBT), where kB is the Boltzmann constant and T is the absolute temperature): E~20 kcal mol−1, A~109 seconds−1. The prefactor is typical for reactions that occur in liquids whereas the huge activation barrier, roughly equivalent to the energy derived from hydrolysis of 3 ATP molecules to ADP under standard conditions, is readily attributed to the fact that polar, ionic, or zwitterionic phospholipid headgroups have to traverse the highly hydrophobic interior of the bilayer as the lipid reorients from one leaflet to the other. An initially asymmetric bilayer is therefore relatively stable, with decay of its asymmetry occurring only very slowly over a time frame of 100 hours. While transbilayer movement of polar lipids only rarely occurs in synthetic systems, fast flip-flop is crucial for cellular life. Thus, constitutive flip-flop of phospholipids is necessary to meet the demands of cell growth, while regulated flip-flop events are needed to sculpt the cellular responses to physiological challenges. Transbilayer translocation of lipids is also needed for membrane homeostasis, to control transbilayer lipid asymmetry in membranes, and to coordinate with protein machinery in generating intracellular transport vesicles. In almost all cases, these transport events are mediated or regulated by proteins without which flip-flop rates would be too slow to match physiological demands. Historically, the role of these proteins was defined through activity measurements that distinguished two types of translocation events: those that required ATP to move lipids vectorially across a membrane and those that were ATP-independent. Although these activity measurements were first reported in the late 1970s and 1980s, the identity of the proteins themselves remained mysterious. Only recently have members of both protein categories been identified and their activities verified by reconstitutions into phospholipid vesicles. The discovery of the molecular identity of the transporters has come hand in hand with fresh insights into the mechanisms of transbilayer translocation of lipids. This has been supplemented by informative in silico approaches using molecular dynamics methods to understand lipid translocation. Our objective in this review is to structure and highlight this explosion of new information. 2. Parameters influencing spontaneous lipid flip-flop Although flip-flop of typical membrane lipids is slow, not all lipids flip-flop slowly. Lipids with a simple hydroxyl headgroup (ceramide, diacylglycerol, and cholesterol) have a very high spontaneous rate of flipping (t1/2 ~seconds/minutes), and glucosylceramide with an uncharged yet polar headgroup has a faster translocation rate than zwitterionic phosphatidylcholine (PC) or phosphatidylethanolamine (PE) (t1/2 ~10 hours, versus ~100 hours for PE) [7–11]. Anionic lipids such as phosphatidic acid (PA) or phosphatidylglycerol (PG) flip-flop slowly at neutral pH, but at pH~5 when they are fully protonated and uncharged they exchange between the leaflets of the bilayer very rapidly (t1/2 ~seconds/minutes) [12]. Indeed, elimination of headgroup polarity by synthetic chelators can promote the fast flipping of a variety of phospholipids [13, 14]. These points are readily illustrated by a simple visual experiment with giant unilamellar vesicles (GUVs) (Figure 2). The GUVs are placed under hyperosmotic conditions so that they become flaccid and adopt a relaxed prolate ellipsoid shape. On adding phospholipid, the GUVs undergo a characteristic shape change to accommodate the excess lipid in the outer leaflet of the membrane (Figure 2A). This happens because the leaflets are coupled and because the added phospholipid does not exchange into the inner leaflet on the time-scale of the experiment (several minutes). If the added lipid happens to be one that flip-flops rapidly, for example a lipid such as ceramide, the shape change eventually reverses as flip-flop normalizes the number of lipids in the two leaflets (Figure 2B). Studies of model membranes as well as molecular dynamics simulations (see for example [15, 16]) have shown that lipid flip-flop is also affected by the physical properties of the bilayer. An essential factor is lipid packing. Interestingly, molecular packing defects can enhance the flip-flop rate [17–19]. Such defects can occur at the border of coexisting liquid-ordered and liquid-disordered phases in the bilayer, and they become especially pronounced at the main phase transition temperature for membranes composed of a single lipid species [17, 19]. Thus, 7-nitrobenz-2-oxa-1,3-diazol-4-yl (NBD)-labeled PE equilibrates extremely rapidly (t1/2 ~6 min) across a dipalmitoyl-PC bilayer at 41°C (the main phase transition temperature for DPPC) compared with an equilibration half-time of ~80 hours at 23°C in egg PC membranes [11]. At temperatures above the solid-liquid phase transition, flip-flop of short-chain phospholipids in human erythrocytes and PC membranes is reduced by cholesterol [19, 20]. Cholesterol suppresses packing defects by condensing the phospholipid arrangement and increasing the thickness of the hydrophobic core. Perturbations of the bilayer structure can be triggered by imperfect matching between the transmembrane domains of proteins and the boundary lipid phase. This led to the proposal that the interplay between transmembrane domains and non-bilayer favoring lipids might be sufficient to allow fast flip-flop in the ER [21–23]. In line with these ideas, recent studies demonstrated that low-complexity synthetic transmembrane peptides are able to increase the rate of flipping of a variety of phospholipids in synthetic vesicles [24, 25]. However, a large number of peptide molecules were needed per vesicle to see this effect. For example, reconstitution of 100 peptides/vesicle corresponding to the transmembrane protein EDEM1 could improve the rate of NBD-PC flipping to 20 lipids per second [25]. However, this contrasts with the rapid flipping seen when vesicles contain only one or a few copies of an authentic scramblase (see below). In summary, the nature of the lipid headgroup and the structure of the membrane itself are key parameters that dictate the rate of spontaneous flip-flop. 3. Transbilayer lipid movement in biological membranes: an overview To facilitate the energetically unfavorable movement of a lipid’s polar head group through the hydrophobic membrane interior, biological membranes are equipped with specific membrane proteins. These proteins are classified as flippases/floppases or scramblases, depending on whether they mediate ATP-dependent transport against a concentration gradient or ATP-independent bidirectional transport (Figure 3). Translocation mediated by scramblases can occur at rates faster than 10,000 seconds−1 (see e.g. [26, 27]), while based on the highest ATP hydrolysis rates reported so far, ATP-driven transporters would reach turnover numbers no greater than ~100 seconds−1 [28] ATP-driven lipid transporters belong to the family of P4-ATPases or ABC transporters. Scramblases come in two flavors: they are either constitutively active or regulated by physiological stimuli such as a rise in intracellular Ca2+. Here we preview these different categories of transporters before exploring each in detail in subsequent sections. Constitutively active phospholipid scramblases Biological membranes are not made from scratch but rather assembled organically by a process in which lipids and proteins are synthesized and integrated into preexisting membranes. Biogenic membranes, such as the endoplasmic reticulum (ER) and bacterial cytoplasmic membrane (bCM), are capable of such self-synthesis. Lipid synthesis is asymmetric such that newly formed lipids are deposited in the cytoplasmic leaflet of the ER or bCM [29]. For bilayer formation, new lipids must be moved to the opposite leaflet on a physiologically relevant time frame. This is accomplished by an ATP-independent, constitutively active phospholipid scramblase that has been extensively characterized but whose molecular identity is surprisingly yet unknown (see below, Section 4.1). Nevertheless, two other constitutively active scramblases have been identified: MprF, a bifunctional bacterial protein that functions as a resistance factor to mitigate the cytolytic effects of cationic antimicrobial peptides and, quite unexpectedly, G protein-coupled receptors (GPCRs). We discuss these topics in Sections 4.2 and 4.3. Constitutively active glycolipid scramblases for glycoconjugate biosynthesis Both the ER and bCM host a variety of glycosylation pathways, for example protein N-glycosylation, O-mannosylation, C-mannosylation and GPI anchoring in eukaryotes and cell wall and cell envelope synthesis in bacteria. Each of these pathways requires the critical participation of glycolipids that carry the building blocks for glycoconjugate assembly. Glycolipid synthesis invariably begins on the cytoplasmic face of the ER or bCM, but the final product is used on the luminal/periplasmic side necessitating transbilayer lipid movement. The scramblases needed for moving these lipids across the membrane are not known, except in the case of cell wall and cell envelope synthesis where scramblase candidates have been identified. These topics are discussed in Sections 5.1–5.3. Control of transbilayer lipid asymmetry by ATP-driven lipid transporters and regulated scramblases The plasma membrane is strongly asymmetric with respect to the transbilayer distribution of lipids. This has been known since the 1970s through the work of Bretscher, as well as van Deenen and colleagues [30, 31]. For example, choline-containing lipids and glycolipids are predominantly located at the exoplasmic leaflet whereas phosphoinositides and aminophospholipids are located in the cytoplasmic leaflet. This asymmetric arrangement of lipids is a consequence of multiple factors, including biophysical membrane properties that limit the ability of a lipid to cross the bilayer spontaneously, retentive mechanisms that trap lipids in one leaflet of the bilayer, as well as the presence of plasma membrane-localized ATP-dependent lipid transporters (presented in Section 6) that selectively catalyze the inward (flip) and outward (flop) vectorial movement of lipids. The function of the transporters at a minimum is to correct any loss of asymmetry that results from spontaneous flip-flop and/or deposition of new lipids through membrane trafficking and non-vesicular pathways. The extent to which they are responsible for generating lipid asymmetry in the first place is not yet clear [32]. Plasma membrane lipid asymmetry is lost in very specific physiological contexts, notably in cells undergoing apoptosis and in activated blood platelets. This is accomplished by regulated lipid scrambling, triggered by caspases (in apoptotic cells) and Ca2+ (in activated platelets). Proteins required for these scrambling events have very recently been identified and are discussed in Section 7. 4. Constitutive scrambling of phospholipids 4.1. Phospholipid scrambling in biogenic membranes: growth of the membrane bilayer Biogenic membranes such as the ER and bacterial cytoplasmic membrane are ‘self-synthesizing’ membranes capable of synthesizing and integrating their protein and lipid components. Much is known about how proteins are assembled into these membranes, but integration of membrane lipids remains poorly understood. The key issue is the problem of lipid scrambling: phospholipids are synthesized on the cytoplasmic face of biogenic membranes but must be scrambled across the bilayer to populate the opposing leaflet. This is necessary for uniform expansion of the bilayer and must occur rapidly, on a timescale relevant to cell growth. For a bacterial cell that doubles every 30 minutes, ~5000 phospholipids must be scrambled every second. Reports published in the late 1970s [34] and late 1980s [35, 36] demonstrated phospholipid scrambling in ER microsomes and bacterial cytoplasmic membrane vesicles, and more recent publications describe the reconstitution of this activity in lipid vesicles [37, 38]. The ER scrambling activity displays a relatively low specificity as non-natural structural isomers of glycerophospholipids (for example lipids with an sn-2,3-diacylglycerol moiety instead of the -1,2-diacylglycerol found in eukaryotes [39] as well as ceramide-based lipids such as sphingomyelin and glucosylceramide are translocated equally well within the limited time-resolution of the activity assays [10, 11]. However, some lipids such as the isoprenoid-based glycolipids involved in the biosynthesis of protein N-glycans are not translocated by the phospholipid scramblase [40]. The reconstitution studies show unambiguously that not all proteins can scramble lipids. For example, detergent extracts of biogenic membranes (crude preparation of membrane proteins) can be fractionated by velocity gradient centrifugation, and only certain fractions show activity when reconstituted into vesicles. Thus, proteins that sediment operationally at ~4S have phospholipid scramblase activity whereas those that sediment more rapidly do not [38, 41]. Protein modification studies suggest that there are at least two proteins that contribute to overall scramblase activity in the ER [42] but their identifications await further investigations. 4.2. A bacterial defense mechanism requires phospholipid scrambling: the role of MprF The bacterial cytoplasmic membrane is a target for membrane disrupting peptides (bacteriocins and defensins) that are attracted to the negative surface charge of the membrane created by PG. To protect against the action of these so-called cationic antimicrobial peptides a number of bacteria modulate their surface charge by modifying PG with lysine or alanine. The polytopic membrane protein MprF catalyzes this modification by transferring an amino acid from the corresponding aminoacyl-tRNAs to PG [43]. The reaction occurs on the cytoplasmic face of the membrane and MprF is proposed to then translocate the aminoacyl-PG to the external surface. The two functions of MprF are separable and can be assigned to different domains of the protein [44]. The synthase domain was recently crystallized [45], while the enigmatic scramblase domain (referred to as flippase in the literature and in Figure 4) remains to be studied in detail. MprF is a large polytopic membrane protein [46] and its scramblase function is associated with the N-terminal 6–8 transmembrane segments that contain charged residues as well as a proline that are critical for scramblase function. While there seems to be little doubt that MprF is indeed a dual function protein, only its synthase function is well defined; its proposed lipid scrambling function remains to be tested using purified protein. We refer to MprF as a scramblase as no obvious metabolic energy source is associated with its transport function. However, this is a point that awaits clarification through reconstitution studies with purified protein. 4.3. Rhodopsin-mediated phospholipid scrambling The outer segments of photoreceptor cells in the retina are stacked with discs that contain the light-sensing protein rhodopsin (Figure 5). It has been known since the early 1990s that disc membranes are able to scramble phospholipids: thus, when labeled phospholipids are added to isolated discs, they equilibrate across the disc membrane on a time-scale of a few minutes [47–49]. Recent studies identified rhodopsin as the scramblase responsible for this phenomenon [27, 50, 51]. When reconstituted into lipid vesicles, rhodopsin is capable of scrambling all common phospholipids extremely rapidly, at a rate >10,000 phospholipids per rhodopsin per second. Rhodopsin’s scramblase activity was found to be constitutive, independent of its light-sensing function. Thus, both the apo-protein opsin and retinal-containing rhodopsin are able to scramble lipids. The necessity for constitutive lipid scrambling in discs is not fully understood, but may be related to the function of the ABC transporter ABCA4 which is found in disc membranes [51]. ABCA4 is proposed to function as importer in moving N-retinylidene PE from the luminal side to the cytoplasmic face of discs. N-retinylidene PE transport helps to reduce synthesis of the bis-retinoid A2E, a risk factor associated with age-related macular degeneration, but a by-product is that it causes lipid accumulation on the cytoplasmic side of discs at the expense of the luminal side. Rhodopsin’s scramblase activity would correct this transbilayer lipid imbalance, which would otherwise distort the disc membrane. Furthermore, by reducing bilayer stress, rhodopsin-mediated lipid scrambling would permit ABCA4 to function. Discs also possess a P-type ATPase, ATP8A2, and its potential function in discs is discussed in Section 8.3. 5. Constitutive scrambling for glycoconjugate biosynthesis 5.1. Scrambling of dolichol-based glycolipids: key steps in protein glycosylation in the ER The majority of proteins that enter the secretory pathway are N-glycosylated. Glycosylation occurs when the protein emerges into the ER lumen and is recognized by oligosaccharyltransferase (OST), as illustrated in Figure 6. The glycolipid Glc3Man9GlcNAc2-PP-dolichol (G3M9-DLO) provides the oligosaccharide that is needed for OST-mediated glycosylation. G3M9-DLO is synthesized in the ER through a multi-step pathway in which sugars are added sequentially to dolichyl phosphate, an isoprenoid lipid with a very long hydrocarbon chain [52]. Sugar addition occurs in two stages and on different sides of the ER [52, 53]. The first 7 steps of G3M9-DLO synthesis convert dolichyl-P to Man5GlcNAc2-PP-dolichol (M5-DLO) on the cytoplasmic face of the ER. Then, M5-DLO is flipped into the ER lumen and extended in 7 further steps to G3M9-DLO. The sugar donors for these luminal reactions are mannose-P-dolichol (MPD) and glucose-P-dolichol (GPD) that are synthesized on the cytoplasmic face of the ER and must be flipped to the luminal side. Although flipping of these glycolipids is ATP-independent and bidirectional, their consumption on the luminal side of the ER promotes a flow of lipid from the cytoplasmic to the luminal side. The scramblases that move M5-DLO, MPD, and GPD across the ER membrane have not been identified. Genetic approaches assigned M5-DLO and MPD scramblase activity to the ER membrane proteins Rft1 and MPDU1, respectively, but disappointingly neither assignment held up in biochemical tests. For example, reconstitution-based assays were used to show that Rft1 could be separated from M5-DLO flippase activity by a number of chromatographic methods [54], and the absence of MPDU1 did not prevent the transbilayer movement of a MPD analog into the ER of cells where the plasma membrane had been permeabilized by a pore-forming toxin to allow access of the analog to the cytosol [55]. Thus, most likely Rft1 and MPDU1 are accessory proteins that assist the actual scramblases, or are proteins somehow involved in maintaining the lipid substrate in transport-competent form [52, 54, 56, 57]. One explanation for the failure of genetic approaches to identify the ER glycolipid scramblases is incomplete coverage in the screens reported to date: this is not surprising as, e.g., the first two enzymes of the DLO biosynthetic pathway were not revealed in the original screen that identified most other DLO glycosyltransferases. An alternative explanation is that the scramblases are dual function proteins whose transport functions are obscured by another function. Although the identities of the three ‘glycosylation scramblases’ are not known, their activities have been measured in ER microsomes as well as in vesicles reconstituted with ER membrane proteins [40, 41, 54, 58–61]. These studies reveal that the M5-DLO and MPD scramblase activities can be resolved by conventional chromatography (for example, ion exchange on DE-52 and lectin affinity with Con A-Sepharose), that they are distinct from the ER activity that flips glycerophospholipids, and that they have exquisite specificities. Thus, Manβ-P-dolichol, the natural isomer of MPD, is scrambled >100-fold more rapidly than non-natural Manα-P-dolichol in vesicles reconstituted with ER membrane proteins from yeast or rat liver. Likewise, ‘biosynthetic’ M5-DLO is scrambled more rapidly than a structural isomer corresponding to ‘processed’ M5-DLO. 5.2. Glycolipid scrambling and bacterial cell wall assembly The bacterial cell wall consists primarily of peptidoglycan, a sugar polymer cross-linked by short peptides. The cell wall provides protection against osmotic stresses and its synthesis is a target of many antibiotics. A key intermediate in cell wall biosynthesis is Lipid II, which provides the glycopeptide building blocks for cell wall assembly (Figure 7). In Lipid II, the glycopeptide is linked via a diphosphate bridge to an undecaprenol lipid. Lipid II is synthesized on the cytoplasmic face of the bacterial inner membrane and then translocated to the periplasmic face for cell wall construction. The identity of the scramblase required for Lipid II translocation is controversial [62, 63]: at least three candidates have been proposed, FtsW [64, 65], MurJ [66, 67], and AmJ [68]. FtsW belongs to the SEDS (shape, elongation, division, and sporulation) family whereas MurJ is a member of the multidrug/oligosaccharidyl-lipid/polysaccharide (MOP) exporter superfamily. AmJ is functionally redundant with MurJ, yet the two proteins share no sequence similarity. Reconstitution based assays using purified protein show unambiguously that FtsW can scramble a fluorescent derivative of Lipid II, whereas MurJ cannot. However, cell-based assays show that MurJ is required for Lipid II flipping whereas depletion of SEDS proteins, including FtsW, has no effect on Lipid II translocation. Both approaches have their limitations [69]. For example, the lack of activity of the purified MurJ protein may simply be the result of its lability during purification and reconstitution. Also, the role of MurJ revealed in vivo may be indirect: while the accumulation of Lipid II in MurJ deficient cells is consistent with MurJ being a Lipid II transporter, this phenotype in itself does not provide proof that transport is directly due to MurJ. Conversely, while FtsW shows Lipid II scramblase activity on reconstitution into vesicles, FtsW-deficient cells unexpectedly accumulate downstream products of the Lipid II biosynthetic pathway. Recent work described a Bacillus subtilis strain lacking all 10 members of the MOP superfamily that are known to be expressed in these cells. The decuple mutant was viable and grew at rates similar to that of wild-type cells, but turned out to express the protein AmJ, which is unrelated to the MOP superfamily and necessary for viability in the absence of MOP family members. AmJ may provide essential Lipid II transport activity when MOP superfamily members, including MurJ, are not available. These data only fuel the controversy surrounding the true identity of the Lipid II scramblase. Much more work will be needed to clarify this issue and it may turn out that these proteins act redundantly at some level. An excellent recent review elaborates on these points, making the case for and against each of the flippase candidates [62]. 5.3. Synthesis of the bacterial cell envelope The cell envelope of Gram-negative bacteria is dominated by lipopolysaccharide (LPS), a complex glycolipid composed minimally of Lipid A (endotoxin) capped by a core oligosaccharide. In many bacteria, LPS is extended by O-antigen, a polysaccharide that consists of repeating units of 2–6 sugar residues. The composition and length of O-antigen repeats can be quite diverse and are typical of individual bacterial strains, enabling strains to be serotyped. The O-antigen repeat unit is synthesized as a glycolipid, undecaprenol-PP-(O-antigen unit), on the cytoplasmic side of the bacterial inner membrane and must be translocated to the periplasmic side to be polymerized into undecaprenol-PP-(O-antigen unit)n, followed by transfer of the O-antigen polymer to Lipid A-core oligosaccharide. Translocation of the undecaprenol-PP-(O-antigen unit) is likely to be facilitated by Wzx proteins [70]. These proteins are polytopic membrane proteins (12 transmembrane spans) encoded by genes that are found within gene clusters associated with O-antigen biosynthesis, and they are closely related to the multidrug and toxin extrusion (MATE) family of inner membrane efflux proteins. Homology to the MATE family suggests that Wzx proteins may use an ion gradient to power the movement of the undecaprenol-PP-(O-antigen unit) [70]. However, in the only example where the lipid translocation role of a Wzx protein was explicitly examined using everted bacterial membrane vesicles, no energy requirement was found [71]. Thus, it remains to be determined whether the Wzx proteins are directly responsible for lipid movement, and if so, whether they function as translocases or scramblases. 6. Control of transbilayer lipid asymmetry by ATP-driven lipid transporters ATP-driven transporters are primary active pumps responsible for a net transfer of specific lipids to one side of a membrane. Current genetic and biochemical evidence indicates that these energy-dependent proteins are primarily members of the P4 subfamily of P-type ATPases (P4-ATPases) and the ATP-binding cassette (ABC) family of transporters. Several excellent reviews have recently treated different aspects of both families of pumps [72–75]. Here we highlight the recent progress in the characterization of their lipid specificities. 6.1. P4-ATPases – inward lipid translocases with different substrate specificities P4-ATPases catalyze the translocation of phospholipids from the exoplasmic to the cytosolic leaflet of membranes. They belong to the family of P-type ATPases, whose members include the ion transporting Na+/K+-ATPase and the Ca2+-ATPase. These proteins are multispan transmembrane pumps that couple ATP hydrolysis to the transport of substrates against their concentration gradient. In doing so, they cycle between a number of conformational states, auto-phosphorylating and dephosphorylating a conserved aspartate residue within a conserved signature sequence (hence the designation ‘P-type’). The lipid-translocating sub-family of P-type ATPases is found exclusively in eukaryotes and classified as type IV, therefore these transporters are referred to as P4-ATPases. Most of these pumps are known to associate with an accessory subunit known as Cdc50 proteins resulting in a heterodimeric complex. This association is required for both proper localization and activity of the pump [76–78] but seems not to affect its substrate specificity [79]. The number of CDC50 isoforms appears to differ between organisms. While only one isoform has been described in Caenorhabditis elegans [80], three isoforms are present in yeast, humans, and the unicellular parasite Leishmania, and up to five exist in plants. Some P4-ATPases interact specifically with only one β-subunit isoform [81–84] while others are more promiscuous and can interact with several isoforms [79, 85–87]. The molecular basis for this difference is unknown. Although the activity of translocases was measured in the 1980s in pioneering work done with red blood cells [88, 89], their molecular identities were only revealed more recently in Baker’s yeast (Saccharomyces cerevisiae). In total, S. cerevisiae harbors five members of this family, namely Neo1p (Neomycin resistant 1), Drs2p (Deficient for Ribosomal Subunit 2), Dnf1p (Drs2p/Neo1p family), Dnf2p, and Dnf3p. Studies on four of these P4-ATPases revealed striking differences in their cellular locations and lipid specificities. Drs2p and Dnf3p are mainly confined to intracellular membranes of the late secretory and endocytic pathways and primarily transport the aminophospholipids phosphatidylserine (PS) and PE [90–92]. By contrast, Dnf1p and Dnf2p are characterized as plasma membrane translocases with relatively broad phospholipid specificities, including PC, lysophospholipids and synthetic alkylphospholipids [93–95] [96, 97]. In mammals, three subunits of the CDC50 family (CDC50A, B and C) and at least 14 P4-ATPases, designated ATP8A1 through ATP11C, have been identified. Among them, ATP8A1, ATP8A2, ATP8B3, ATP11A, and ATP11C have, so far, been connected to aminophospholipid translocation. ATP8A1 is dependent on PS and PE for ATPase activity [98, 99], is able to translocate fluorescent PS upon expression in yeast [100], and regulates PS asymmetry in the late secretory pathway [101]. Likewise, ATP8A2, a P4-ATPase highly expressed in the brain, testes, and retina, exhibits PS-dependent ATPase activity and the ability to translocate fluorescent PS, and to some extend PE, in proteoliposomes [28, 84, 102]. In the spermatozoa of mice, ATP8B3 is necessary for PS asymmetry and fertilization [103]. ATP11C, a P4-ATPase important for B cell and erythrocyte development, was found to play a crucial role in PS translocation [104, 105]. During apoptosis, ATP11C undergoes caspase-mediated cleavage and is consequently inactivated, thereby contributing to PS exposure on the cell surface [106, 107]. ATP11C is crucial for PS flipping in CHO-K1 cells and a lack of the functional ATP11C protein is responsible for the defect in PS uptake in UPS-1 cells [108]. Notably, not all mammalian P4- ATPases are aminophospholipid specific translocators. ATP8B1, initially characterized as a translocase for aminophospholipids [109] and cardiolipin [110], was found to translocate PC rather than PS upon overexpression in cell lines with low endogenous phospholipid translocase activity [111]. In the same setup, ATP8B2 and ATP10A were shown to specifically transport PC [111, 112]. Using heterologous expression in yeast, ATP8B5 from mouse testes was found to transport PC and PE [92]. Moreover, previous studies on mammalian cells uncovered a role of the human P4-ATPase subunit CDC50A (TMEM30a) in the uptake of the inflammatory lipid PAF (platelet-activating factor, a short-chain PC) and synthetic alkylphospholipids, implying the presence of yet unidentified P4-ATPase(s) that in complex with CDC50A facilitate(s) the uptake of alkylphospholipids [113]. Recent studies in the plant Arabidopsis thaliana further substantiate the notion that members of the P4-ATPase family differ in their substrate specificities. The Arabidopsis genome encodes 12 P4-ATPases termed ALA1 to ALA12 (for AminophosphoLipid ATPase) [114]. The lipid specificity of some of these plant ALA proteins has been characterized. ALA2 flips PS across the lipid bilayer in the endosomal system. Golgi-localized ALA3 facilitates transport of PS, PE and PC but not of the lysoPC analogue miltefosine [79, 86]. ALA10 localizes to the plasma membrane and internalizes a broad range of exogenous phospholipids, including lysoPC [115]. Strikingly, many P4-ATPases localizing to subcellular membranes seem to display a narrow substrate specificity only recognizing one (typically PS) or few lipid species as substrate. By contrast, several P4-ATPases at the plasma membrane of eukaryotic cells transport a broad range of exogenous phospholipids, including lysolipids and derivatives, suggesting functions for these P4-ATPases in lipid scavenging [94, 95, 115]. In fact, the cellular processes known to be directly or indirectly affected by P4-ATPases have expanded over the last years to include the regulation of membrane traffic, cytoskeletal dynamics, cell division, lipid metabolism, and lipid signaling (reviewed in [73, 75, 116]). 6.2. ABC transporters – a family of exporters and importers ABC transporters comprise a superfamily of membrane proteins that actively transport chemically diverse substrates across biological membranes. They typically consist of four core domains: two transmembrane domains surrounding a single central cavity, presumably along the pathway of substrate transport, and two ATP-binding cassettes (ABCs), which supply the energy for substrate transport through the binding and hydrolysis of ATP [117, 118]. In eukaryotes, ABC transporters exist as either full transporters, in which all four domains are present within a single polypeptide chain, or half-transporters, in which an ABC and transmembrane domain reside within a polypeptide chain that assembles as homo- or heterodimers. In prokaryotes, the domains can exist as individual subunits or together in various combinations to generate an active transporter. Humans have almost 50 different ABC transporters that are grouped into seven sub-families (A–G). The members of these subfamilies are not exclusively expressed at the plasma membrane but also localizes to intracellular organelles such as peroxisomes (ABCD1–3) [119], lamellar bodies (ABCA3, ABCA12) [120–122], lysosomes (ABCA2, ABCA5) [123, 124], and endosomes (ABCG1, ABCG2) [125, 126]. Several members of these subfamilies have been implicated in the transport of lipids or lipid-related compounds. Their substrate specificities have mostly been inferred from disease-associated phenotypes, analysis of knock-out mice, and cell-based studies. Progress in purification and reconstitution of some members has provided directed evidence for their capability to directly catalyze phospholipid transport. Purified ABCB1 reconstituted into proteoliposomes flips a variety of short-chain fluorescent phospholipids and sphingolipids [127], including simple glycolipids, but is unable to restore transport of PC into the bile of Mdr2 (Abcb4)-knockout mice [128]. Thus, direct evidence for the transport of endogenous lipids is still lacking. The glutathione dependent multidrug transporter ABCC1 transports fluorescent PC after reconstitution in proteoliposomes [129] and may help to maintain the outward orientation of natural choline phospholipids in the plasma membrane [130–132]. Reconstituted ABCA1 actively transports fluorescent PC, PS, and SM with a preference for PC, whereas ABCA7 preferentially transports fluorescent PS [133]. In yeast, Yor1p and Pdr5 have been implicated in the transport of PE [93, 134] while Ybt1 translocates PC across the vacuolar membrane as part of choline recycling [135]. Thus, similar to P4-ATPases discussed in the previous section, individual ABC transporters differ in their substrate specificities. Notably, several plant ABC transporters have been implicated in the transport of cuticular lipids comprising C16 and C18 hydroxy and epoxy fatty acids [136–140]. Although it has been possible to demonstrate lipid transport by purified ABC transporters after reconstitution into phospholipid vesicles, such experiments remain challenging since the amplitude of transport is very small. This is because even the smallest movement of phospholipids from the outer leaflet to the inner leaflet of the vesicles exerts bilayer stress (the bilayer couple hypothesis [3] that prevents further transport. One example where this is not a problem is the case of PglK, a bacterial ABC transporter that transports a specific lipid that is not a component of the bulk membrane. This example is discussed below. Apart from outwardly-directed transporters, eukaryotes also express ABC transporters that transport lipids towards the cytosolic leaflet of cellular membranes. Such an inward-directed lipid translocase activity has been demonstrated for the mammalian ABC transporter ABCA4 flipping the Schiff base adduct of retinal and PE known as N-retinylidene-PE from the lumen to the cytosolic leaflet of photoreceptor disc membranes and proteoliposomes [133, 141]. Notably, ABCC7 expression has been correlated with an increased uptake of the signaling lipids sphingosine-1-phosphate and lysophosphatidic acid, implying that other mammalian members might act as inward lipid transporters as well [142]. In the yeast Candida albicans, a subfamily member (Cdr3p) has been identified that exhibits an inward-directed phospholipid translocase activity. Some yeast ABC transporters (Aus1p and Pdr11p in S. cerevisiae; Aus1p in Candida glabrata) facilitate exogenous sterol uptake [143–148]. Three Arabidopsis proteins, TGD1-3, are proposed to form an ABC transporter complex in the inner envelope of the chloroplast that imports PA into the inside of the chloroplast for the synthesis of thylakoid lipids [149, 150]. Finally, ABC transporters are also widely expressed in prokaryotes, and some bacterial ABC transporters function as lipid translocases. MsbA is one of these proteins being an essential inner membrane transporter in Gram-negative Escherichia coli genetically linked to the export of the Lipid A core of lipopolysaccharides to the bacterial outer membrane. Depletion of cellular MsbA or the presence of a conditionally inactivating mutation in this protein results in loss of lipid A and phospholipid transport from the cytoplasmic to the outer membrane in bacterial cells, suggesting a general lipid translocase function for MsbA [151, 152]. Reconstitution studies using purified MsbA showed that it can transport a variety of fluorescent glycerophospholipid and even simple glycosphingolipids, albeit with low activity [153], while other data suggest that this is not likely to be the case [22, 154]. Another ABC protein, PglK, is required for the translocation of isoprenoid-linked oligosaccharides [155–157]. This function was recently confirmed by reconstituting PglK-mediated flipping of isoprenoid-linked oligosaccharide intermediates from the outer leaflet to the lumen of proteoliposomes [158]; comments in [159] and [160]. 7. Loss of plasma membrane asymmetry and surface exposure of PS: role of regulated scramblases It has long been recognized that PS, normally sequestered in the cytoplasmic leaflet of the plasma membrane by inward ATP-driven transporters, becomes exposed at the cell surface in response to specific stimuli [161]. Thus blood cells, especially platelets, expose PS when activated and this is important in providing a nucleating platform for the maturation of coagulation factors and the production of blood clots. Likewise, cells undergoing programmed cell death expose PS and this helps to target them for removal through the phagocytic action of macrophages. In a specific and quite dramatic example, PS is exposed at the tips of photoreceptor cells in the retina in a diurnal manner and this contributes a signal for the consumption of the tips by neighboring retinal pigment epithelial cells [162]. Recent discoveries have implicated membrane proteins belonging to the TMEM16 and Xkr families in PS exposure [163–165]. The former are specifically associated with Ca2+-dependent PS exposure related to blood clotting whereas the latter play a role in apoptosis. Many cells expose PS when treated with a Ca2+ ionophore in the presence of extracellular Ca2+. It is generally assumed that ionophore treatment prompts an increase in intracellular Ca2+ that, in turn, activates a plasma membrane scramblase. The predicted high rate of scrambling (>10,000 phospholipids per scramblase per second) counters the slow corrective effects of P-type ATPases (1–100 phospholipids per ATPase per second) in restoring PS asymmetry, and thus PS can be detected at the cell surface by fluorescent probes such as fluorescein-labeled Annexin V [165]. Scramblase action is not specific to PS; all common phospholipids are translocated between the two leaflets of the membrane, resulting in loss of lipid asymmetry of which a byproduct is PS exposure. Mouse Ba/F3 cells expose PS in response to ionophore even under low Ca2+ conditions [165]. A subline of these cells was selected by repeated fluorescence-activated cell sorting and shown to be highly sensitive to Ca2+ ionophore-elicited PS exposure. To identify the protein(s) responsible for hyperactive scrambling in these cells, a cDNA library was generated, fractionated into clones >2.5 kb and 1.0–2.5 kb, and introduced into the parental Ba/F3 cell line. Further rounds of selection revealed the identities of the genes responsible for enhanced PS exposure. The use of cDNA clones >2.5 kb identified TMEM16F [165], whereas clones of 1.0–2.5kb identified Xkr8 [163]. Both proteins belong to evolutionarily conserved protein families and subsequent work has implicated other members of each family in scrambling [163, 164, 166]. Scott syndrome is a bleeding disorder caused by the inability of activated blood platelets to expose PS, and this disease has figured prominently in the search for the Ca2+-activated scramblase [161]. Scott syndrome patients express a severely truncated non-functional TMEM16F protein, a result of a mutation in the TMEM16F gene that causes a frame shift leading to premature termination [165]. In contrast, the Ba/F3 subline that is hyperactive in PS exposure expresses TMEM16F with a constitutively activating D409G mutation. While these correlations make it clear that TMEM16F plays a critical role in lipid scrambling, it is possible that the protein functions indirectly. Indeed, as TMEM16 belongs to a family of ion channels, one hypothesis is that it is involved in providing Ca2+ in a localized way to support the activity of the ‘real scramblase’. This idea is unlikely to be correct as direct evidence that TMEM16 proteins have Ca2+-activated scramblase activity has come from reconstitution studies. Purified fungal TMEM16 homologues (afTMEm16 and nhTMEM16) were reconstituted into large unilamellar vesicles and shown to scramble phospholipids [167, 168]. Scrambling was sensitive to Ca2+ as the rate of scrambling in the presence of Ca2+ was at least an order of magnitude faster than in its absence. Some TMEM16 proteins do not scramble lipids and this provides an opportunity to understand what makes a TMEM16 protein a scramblase. Analysis of chimeras between the inactive TMEM16A protein, a Ca2+-activated chloride channel with no scramblase activity, and TMEM16F identified a sequence in 16F that, when transplanted into 16A, allowed the chimeric protein to promote Ca2+-dependent PS exposure in cells [169]. While a definitive conclusion will only be possible after reconstitution and testing of purified protein, these results nevertheless strongly suggest that TMEM16F is indeed a scramblase. The other protein family implicated in lipid scrambling is less well understood. Of the 8 members of the mouse Xkr family, three are needed to facilitate PS exposure in cells undergoing programmed cell death [166]. It is unclear whether these proteins are scramblases themselves, but as they are polytopic membrane proteins located at the plasma membrane they are well positioned to either carry out or regulate scrambling. Xkr proteins need to be activated in order to play a role in PS exposure, and, unlike the Ca2+ trigger needed for TMEM16 proteins, the trigger here appears to be caspase-mediated cleavage of the C-terminal tail. 8. Lipid transporters and membrane lipid asymmetry Differential distribution of phospholipids across the bilayer is a feature that defines the many different membranes in a eukaryotic cell. A remarkable assortment of selective and non-selective lipid transporters with overlapping sub-cellular distributions and substrate specificities appears to help in controlling the lipid arrangement in the various membrane systems (Figure 3). Thus, their activities need to be differentially regulated. In addition, lipid arrangement also depends on the membrane lipid composition and retentive mechanisms that trap lipids on one side of the bilayer. Here we will highlight three examples illustrating how cells might combine these principles to control the texture of their membranes. 8.1. Eukaryotic plasma membranes: lipid asymmetry based on constrained flip-flop In the plasma membrane, the late Golgi, and endosomal compartments ATP-driven vectorial transporters are responsible for a net transfer of specific lipids to one side of a membrane and thereby regulate their transbilayer lipid arrangement. A prime example is the specific transport of PS and PE towards the cytosolic leaflet of these membranes by members of the P4-ATPase family. This selective transport is held responsible for generating and maintaining an asymmetric lipid distribution with PS and PE largely confined to the cytosolic leaflet. Some P4-ATPases display a PC translocase activity implying that certain cell types might restrict PC to the cytoplasmic leaflet as well. The clearance of PC and other phospholipids from the cell surface would lead to an enrichment of sphingolipids in the exoplasmic leaflet. As sphingolipids have saturated acyl chains and therefore pack at a higher density than glycerophospholipids, their enrichment in the exoplasmic leaflet of the plasma membrane would support the barrier function of this organelle. Sphingolipids are primarily synthesized in the luminal leaflet of Golgi and delivered to the cell surface by vesicular transport, which explains their asymmetric distribution across the plasma membrane [170]. However, it is still unclear what permits the accumulation of the majority of PC in the outer plasma membrane leaflet of most mammalian cells. One possibility is that ABC proteins move phospholipids unspecifically to the outer leaflet while P4-ATPases restore specific lipids, e.g. PS, PE, to the inner leaflet [171] [172] [127] [129] [173]. Alternatively, it was proposed on theoretical grounds that inward transport of aminophospholipids, together with passive fluxes, would be sufficient to accumulate choline containing lipids in the outer leaflet [174] [175]. Discrimination between these models awaits evaluation of the substrate specificity and transport efficiency of the different ATP-dependent lipid transporters and the characterization of their regulation in the living cell. Despite these questions, it is clear that the plasma membrane does not support constitutive, protein-mediated lipid scrambling. Non-selective ATP-independent scramblases are either retained in the ER or silenced during their trafficking to the plasma membrane, possibly by the increasing sterol content from the ER (around 5 mol%) to the plasma membrane (more than 40 mol% sterol), leading to a thicker, more organized bilayer that is not compatible with scrambling. It is interesting to speculate that scramblases destined for the plasma membrane actually contribute to lipid scrambling in the ER while they are in transit through the ER [27]. 8.2. Early secretory organelles: lipid asymmetry in presence of scramblases In early secretory organelles, scramblases are important for the proper assembly of the membrane. They allow lipids to equilibrate rapidly between the two bilayer leaflets, and thereby in principle promote a symmetric lipid distribution across the bilayer. However, studies with a genetically-encoded fluorescent PS sensor indicate that PS is not detectable at the ER membrane, consistent with it residing in the luminal leaflet of the ER, an asymmetric arrangement opposite to that of the plasma membrane [176]. In line with this notion, phospholipid monolayers around lipid droplets, which are thought to arise from the cytoplasmic leaflet of the ER, contain hardly any PS [177]. This suggests the presence of retentive mechanisms that trap PS on the luminal side of the ER. Such retentive mechanisms could include calcium-mediated interactions of PS with luminal proteins [178]. In addition, recent work uncovered lipid-transfer proteins that mediate non-vesicular transport of PS from the cytosolic leaflet of the ER to the cytosolic surface of the plasma membrane, in exchange for phosphatidylinositol-4-phosphate [179–181]. Such lipid-exchange cycles might contribute in the control of PS asymmetry in early secretory organelles and the plasma membrane. 8.3. Photoreceptor discs: lipid asymmetry based on asymmetric charge distribution Photoreceptor disc membranes are another example of a subcellular membrane with an asymmetric lipid arrangement. These membranes are asymmetric with respect to PS (65–80% in the cytoplasmic leaflet), and roughly symmetric in their transbilayer distribution of PC and PE [47–49]. This asymmetry was predicted by the transbilayer coupling model of Hubbell [182], i.e., the combination of phospholipid scrambling and the asymmetric charge distribution (positive on the cytoplasmic face of discs), created by the large number of oriented rhodopsin molecules in the disc membrane, explains why negatively charged PS is mainly located on the cytoplasmic side at steady state whereas zwitterionic PC and PE are symmetrically distributed across the disc membranes. Interestingly, the transbilayer asymmetry of PS changes reversibly in response to light, presumably because of associated changes in transbilayer charge asymmetry [49]. Given the transbilayer coupling model [182] it is not necessary to invoke the activity of the aminophospholipid specific transporter ATP8A2 to explain PS asymmetry. Indeed the function of ATP8A2 in discs is enigmatic as its activity would needlessly deplete ATP. We speculate that Atp8a2 may be silent in disc membranes because of unrelieved auto-inhibition, or because of regulation by phosphorylation as shown for other lipid-translocating P-type ATPases [183, 184]. Atp8a2 may instead play a functional role in the secretory pathway [101, 185] by regulating protein traffic to the disc membrane. 9. General mechanisms of handling lipids by lipid transporters How would proteins facilitate lipid flip-flop? A simple look at the Arrhenius rate equation presented in section 1, suggests that if the prefactor stays the same then the energy barrier must decrease from ~20 kcal mol−1 to ~7 kcal mol−1 in order to accelerate the rate of flipping to that measured for known scramblases. While it is also possible that the prefactor may change, if the frequency with which lipids access the flippase is somehow enhanced, it is typical to consider mechanisms that result in reduction of the energy barrier or those that short-circut the bilayer by fusing the two leaflets. Thus, the usual depiction of phospholipid movement across a bilayer shows a path where the two leaflets of the bilayer are connected. Evidence for such a path is provided in studies of pore-forming amphipathic peptides, such as magainin, which are predicted to form toroidal pores in the membrane [186, 187]. Such pores would necessarily connect the two leaflets leading to lipid mixing/scrambling. In silico studies promote the importance of water [188] (Figure 8). Thus, the presence of a transbilayer region involving water, or at least polar molecules, would promote lipid movement as the lipid headgroup would be able to interact with this region specifically, thereby minimizing the energy cost of transfer. A lipid transporter might provide such a pathway via a hydrophilic membrane-facing groove. The translocation mechanism is imagined to resemble the swiping of a card through a card reader (Figure 9). In this “card reader” model, the polar headgroup of the phospholipid (the magnetic strip on the card) is protected during passage across the hydrophobic interior of the membrane until it emerges on the other side; the acyl chains of the lipid remain in the hydrophobic milieu of the membrane during this process. Thus, the groove of the flippase/card-reader provides a low energy path for the lipid headgroup by sequestering it from the unfavorable hydrophobic environment of the membrane interior. Evidence for such potential grooves has been provided for several lipid transporters. For example, structural homology modeling and molecular dynamics simulations on the mammalian P4-ATPase ATP8A2 provided an indication for a groove formed by the transmembrane segments M1, M2, M4, and M6, which likely contains water-filled pockets [190]. The cytoplasmic end of this groove coincides with a bound PE molecule present in several crystal structures of the sarcoplasmic reticulum Ca2+-ATPase [191], in line with a transport pathway from the exoplasmic side to a cytoplasmically facing exit site situated approximately at the location of the PE molecule in SERCA. Based on mutational studies on the yeast P4-ATPases Dnf1p and Drs2p, a transport pathway along potential grooves formed by the transmembrane segments M1, M3, and M4, or alternatively M3, M4, and M5 has been proposed [192]. Furthermore, current models for ABC-type translocases propose an alternating opening of lateral entry and exit pathways for the transported phospholipid substrate to the central cavity formed by the transmembrane helices. Coarse-grained modelling has revealed lipid binding to the ABC transporter MsbA within the grooves formed between transmembrane helices [193]. Similarly, scramblases appear to display such potential grooves. The crystal structure of a Ca2+-bound TMEM16 scramblase (nhTMEM16) was recently reported [168]. The structure is that of a dimer, the native state in which TMEM16 proteins are isolated from cells. Each monomer has a remarkable groove in its transmembrane domain on the side of the protein opposite to that of the dimer interface. The groove is ~1 nm wide and of polar character, and could thereby accommodate the headgroup of a transiting phospholipid. While considerable work needs to be done to verify that this is indeed the translocation pathway, this structural feature of nhTMEM16 provides the first hint of how a scramblase might function, reinforcing models that have been discussed for many years. Apart from utilizing a hydrophilic membrane facing groove, some scramblases might perturb the lipid bilayer thereby facilitating lipid flip-flop. Such a mechanism might apply to rhodopsin-mediated lipid scrambling. Although it was originally suggested that a possible lipid translocation pathway might be provided by the central water-lined core of the protein [50], this was ruled out since access to the core is occluded by a loop (the E2-loop) on the exoplasmic side, and since the core itself is blocked by retinal in rhodopsin and metarhodopsin II, both of which are active as scramblases [27]. A suitable environment for lipid translocation could be provided instead by specific residues on the surface of the rhodopsin transmembrane (TM) helical bundle or by membrane packing defects that are specifically generated in the vicinity of rhodopsin. As other rhodopsin-like GPCRs, e.g. the A2A adenosine receptor, β1-, and β2-adrenergic receptors also scramble lipids [27, 50]; a structural element common amongst these proteins likely contributes to the scramblase activity. GPCRs contain an intracellular amphipathic helix 8 that immediately follows the last transmembrane helix. As this helix runs parallel to the membrane and often contains palmitoylated cysteine, it has the potential to disturb bilayer structure on one side of the membrane; propagation of this disturbance across the bilayer may contribute to bidirectional scrambling. 10. Concluding remarks and future perspectives As is clear from this overview, phospholipid translocation across cellular membranes is facilitated by a remarkable assortment of lipid translocases and scramblases. Some progress has been made on the in vitro reconstitution of purified P4-ATPases, ABC transporters, and scramblases, thereby demonstrating that these proteins indeed participate directly in lipid translocation. Yet, analysis of natural lipid translocation by these enzymes remains challenging due to the hydrophobic nature of the lipid substrate. Nevertheless, such studies are important to determine their transport kinetic properties and substrate specificities, and thus to understand their physiological role and interplay in the various cellular membrane systems. Additional biophysical approaches will be required to test the impact of the lipid composition – especially cholesterol content – on lipid translocation by translocases and scramblases. The answer to this question cannot be easily obtained by experiments on living cells, nor can it be obtained via ensemble reconstitution experiments because of the compositional heterogeneity of reconstitutions [196]. The key lies in single vesicle experiments where the lipid content is precisely known. At the same time, further biophysical studies on model membrane systems are needed to unravel how the physical properties of the bilayer itself dictate the rate of spontaneous flip-flop. Techniques such as small-angle neutron scattering and sum-frequency vibrational spectroscopy that do not require lipids labeled with bulky reporter groups are promising but have yielded contradictory results in a few instances. For example, using sum-frequency vibrational spectroscopy, fast flip-flop rates (t1/2 <10 min) were measured for DPPC and DMPC in supported fluid bilayers attached directly to silica [197–199]. In comparison, flip-flop was predictably slow in bilayers formed on polymer supports [200], suggesting that the roughness of the silica surface induces imperfections within the bilayer which promote fast flipping [201]. Clarification of these issues is warranted in future studies. Promising steps have also been taken to dissect the inner workings of lipid-translocating proteins. This includes the improvement of protein expression and purification approaches and crystallization attempts [202, 203]. Yet, the lipid-translocating proteins identified so far belong to large protein superfamilies, and many family members remain to be characterized. At the same time, there is now the probability that individual proteins have multiple functions depending on their cellular context, as has been demonstrated for the G protein-coupled receptor rhodopsin. Dissecting how all these enzymes are coordinately regulated within the cell will be a challenge for years to come. Notably, mitochondria and chloroplasts derive many of their membrane lipids from the ER through a non-vesicular pathway that exploits regions of close membrane contact between the two organelles [204, 205]: ER-derived lipids arriving at the surface of these organelles must flip across the outer and inner membranes to gain access to the organelle interior [206–208]. Identification of the mitochondrial and plastid translocases/scramblase(s) will be important in future studies. Similarly, the phospho- and glycolipid scramblases of biogenic membranes have not yet been identified. Although some bacterial proteins such as MprF appear to be able to scramble phospholipids, none of these proteins are needed for viability, and, as a group, they are unlikely to shoulder the burden of scrambling phospholipids to sustain growth. It has been almost a decade since the last substantial progress was made on lipid flipping in biogenic membranes. The spate of recent discoveries in the scramblase field may provide new impetus to go after these fundamentally important transporters. The authors are particularly grateful to the members of their research teams for stimulating discussions and comments on the manuscript; we would like to warmly thank Ida Louise Jørgensen for her commitment to improving and eliminating errors in the manuscript and Kalpana Pandey for generating Figure 9. We thank Natascha Ruiz for comments on the section relating to peptidoglycan synthesis. TGP acknowledges support from the Research Centre ‘bioSYNergy’ funded by the University of Copenhagen’s Excellence Programme, the Danish National Research Foundation through the PUMPKIN Center of Excellence (DNRF85), and the Danish Council for Independent Research |Natural Sciences (DFF1323-00297). AKM acknowledges support from the National Institutes of Health (grants GM106717, EY024207 and NS093457). Abbreviations ABC transporter ATP-binding cassette transporter bCM bacterial cytoplasmic membrane DLO dolichol-linked oligosaccharide ER endoplasmic reticulum GPCR G protein-coupled receptor GPD glucose-P-dolichol GPI glycosylphosphatidylinositol LPS lipopolysaccharide MPD mannose-P-dolichol NBD 7-nitrobenz-2-oxa-1,3-diazol-4-yl OST oligosaccharyltransferase PA phosphatidic acid PC phosphatidylcholine PE phosphatidylethanolamine PG phosphatidylglycerol PI phosphatidylinositol PS phosphatidylserine Figure 1 Phospholipid motions in a membrane Phospholipid bilayers are two-dimensional fluids. Individual lipid molecules have a cross-sectional area of ~0.7 nm2. In each monolayer of the membrane bilayer they can rotate very rapidly around their head-to-tail axis with a characteristic time of 10−9 seconds, and diffuse laterally within the plane of a membrane leaflet with a translational diffusion coefficient of ~10−8 cm2 seconds−1, i.e. the time taken for a phospholipid to move ~1 nm to replace a neighboring phospholipid is ~100 nanoseconds. In contrast, spontaneous exchange of phospholipids between leaflets (flip-flop) is slow, taking typically ~100 hours. The energy barrier that must be overcome in order to move the phospholipid headgroup through the hydrophobic interior of the membrane is >20 kcal mol−1. Adapted from Mouritsen ‘Life - As a Matter of Fat’ [5]. Figure 2 Shape change in GUVs on expanding the outer monolayer of the membrane A, Lysophosphatidylcholine (16:0) was added to a prolate GUV and the sample was observed by differential interference contrast microscopy. A time-lapse sequence is shown, starting at the left and ending at the right. As the phospholipid does not exchange between the leaflets of the bilayer on the time-scale of this experiment (~6 minutes), and because the two leaflets of the membrane are coupled, the GUV undergoes a predicted shape change to minimize bilayer stress caused by the excess lipid in one leaflet [3]. B, The same experiment as in panel A, except that C6 ceramide (d18:1/6:0) was added to a prolate GUV. The shape change induced by excess ceramide in the outer leaflet is evident. However, as C6 ceramide flip-flops rapidly, the number of lipids in the two leaflets of the GUV membrane eventually normalizes to restore the original shape of the GUV. Images courtesy of Patricia Pipaluk Mia Mathiassen. Figure 3 Lipid transporters and membrane lipid asymmetry The endoplasmic reticulum (ER) harbors constitutive scramblases that facilitate rapid flip-flop of lipids and allow them to equilibrate between the two membrane leaflets independently of ATP. This system is unable to accumulate a given lipid in one leaflet. Thus, retentive mechanisms are required to trap lipids (e.g. PS) on the luminal side of the ER; also, for example, consumption of glycolipid biosynthetic intermediates such as DLOs on the luminal side of the ER drives scrambling from the cytoplasmic to the luminal leaflet. In the plasma membrane (PM) of eukaryotic cells, flip-flop of phospholipids is constrained by the absence/silencing of constitutive scramblases. Thus, ATP-dependent flippases (P4-ATPase family members) and floppases (ABC transporters) can maintain an asymmetric phospholipid distribution by moving specific lipids towards or away from the cytosolic leaflet. Cellular activation triggered by cytosolic calcium, caspases or other stimuli can collapse the lipid asymmetry by the transient activity of ATP-independent scramblases. Note that the term “flippase” is sometimes used to designate an enzyme that catalyses lipid flip-flop in both directions [33]. PC, phosphatidylcholine; PS, phosphatidylserine; DLOs, lipid-linked oligosaccharides. Figure 4 MprF-mediated bacterial CAMP resistance MprF is a bifunctional protein. Its synthase domain (S) transfers lysine from lysyl-tRNA to phosphatidylglycerol (PG) to synthesize lysyl-PG, whereas its flippase domain (F) transfers lysyl-PG across the inner membrane to the exoplasmic/periplasmic side. Negatively charged PG attracts cationic anti-microbial peptides (CAMPs), whereas lysyl-PG being neutral does not. Figure 5 Lipid transporters in photoreceptor discs ABCA4 is an ABC transporter specific for PE and N-retinylidene-PE (NRPE); ATP8A2 is a P4-ATPase specific for PS and PE – however, it is not clear whether it is active in discs (see text); rhodopsin (Rho) is a scramblase that translocates common phospholipids (PL) in an ATP-independent manner. Arrows show the direction of lipid transport. ABCA4 is unusual amongst mammalian ABC transporters because it is the only one reported thus far that functions as an importer or flippase. Figure redrawn from [51]. Figure 6 Glycolipid scrambling is necessary for protein N-glycosylation in the ER G3M9-DLO, the oligosaccharide donor for protein N-glycosylation, is synthesized in the ER in a multi-step, topologically split pathway. The first 7 steps convert dolichyl-P (dol-P) to M5-DLO on the cytoplasmic face of the ER. Then, M5-DLO is flipped into the ER lumen and extended in 7 further steps to G3M9-DLO. The sugar donors for these luminal reactions are MPD and GPD that are synthesized on the cytoplasmic face of the ER and must be flipped to the luminal side. In addition to its role in N-glycan biosynthesis, MPD is required in the ER lumen for GPI anchor biosynthesis, O-mannosylation, and C-mannosylation. Oligosaccharyltransferase (OST) transfers the oligosaccharide from G3M9-DLO to Asn residues within glycosylation sequences in translocating proteins as they emerge from the translocon into the ER lumen. The dolichyl-PP product of the OST reaction is recycled. The multistep synthesis and transfer of the oligosaccharide require at least 40 gene products. Figure 7 Bacterial cell wall (peptidoglycan) assembly The peptidoglycan building block is assembled on the lipid undecaprenyl phosphate on the cytoplasmic side of the bacterial inner membrane (IM). The enzyme MraY uses UDP-N-acetylmuramic acid-L-Ala-γ-D-Glu-A2pm-D-Ala-D-Ala (UDP-MurNAc-pentapeptide) to synthesize Lipid I. MurG then catalyzes the transfer of GlcNAc from UDP-GlcNAc to Lipid I to generate Lipid II. Lipid II is flipped across the inner membrane (depicted here as a bidirectional process, although this is not fully established) where transglycosylases (TG) polymerize the GlcNAc-MurNac-pentapeptide units into glycan chains attached to undecaprenol by a pyrophosphate linkage. These chains are crosslinked to pre-existing peptidoglycan by transpeptidases while the terminal D-Ala residues in each unit are removed by carboxypeptidases. Figure redrawn from Ref. [62]. Figure 8 In silico analysis of pore-mediated lipid flip-flop Molecular dynamics simulation showing that the appearance of a water pore facilitates the spontaneous migration of lipids across a phospholipid membrane. (A) 0 picoseconds, (C) 118.9 nanoseconds, (D) 122.4 nanoseconds, (E) 152.7 nanoseconds. Lipids (except for the flip-flopped one) are not shown; water is shown in red and white, acyl chains of the flip-flopped lipid are shown in yellow, and its choline and phosphate groups are shown in orange and green, respectively. Adapted and reprinted by permission from [189], American Chemical Society, copyright 2007. Figure 9 Model for substrate flipping by TMEM16 and ATP8A2 (A) Credit card swiping model. The magnetic strip on the card (= polar headgroup of the phospholipid being transported) is protected from the lipid environment (by passage through the groove in the card reader) as it transits the hydrophobic interior of the membrane. See text for details and for a discussion. Adapted from a figure drawn by Adam Steinberg and published in Ref.[194]. Panel (B) and (C) show alternate views of the proposed conduit groove in nhTMEM16 (PDBID 4WIS) and homology modeled bovine ATP8a2 (Uniprot ID: C7EXK4, cytoplasmic domains are omitted), respectively. Both proteins are shown in surface representation with dark salmon color. The grooves are highlighted in blue with a hypothetical lipid placed inside shown as green CPK representation. The ATP8A2 homology model was generated based on the sarcoplasmic-endoplasmic reticulum Ca2+ ATPase I structure (PDBID: 3B9B) and in vacuo energy minimized using GROMOS96 implementation in Swiss-Pdb Viewer [195]. All models were generated with Pymol (DeLano Scientific, San Carlos, California). This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. 1 Gorter E Grendel F ON BIMOLECULAR LAYERS OF LIPOIDS ON THE CHROMOCYTES OF THE BLOOD The Journal of Experimental Medicine 1925 41 439 43 19868999 2 Zwaal RFA Demel RA Roelofsen B van Deenen LLM The lipid bilayer concept of cell membranes Trends in Biochemical Sciences 1976 1 112 4 3 Sheetz MP Singer SJ Biological membranes as bilayer couples. A molecular mechanism of drug-erythrocyte interactions Proc Natl Acad Sci USA 1974 71 4457 61 4530994 4 Tanford C The hydrophobic effect: formation of micelles and biological membranes New York John Wiley & Sons 1973 5 Mouritsen OG Life - As a Matter of Fat Berlin Heidelberg GmbH & Co K.: Springer-Verlag 2005 6 Kornberg RD McConnell HM Inside-outside transitions of phospholipids in vesicle membranes Biochemistry 1971 10 1111 20 4324203 7 Ganong BR Bell RM Transmembrane movement of phosphatidylglycerol and diacylglycerol sulfhydryl analogs Biochemistry 1984 23 4977 83 6498172 8 Lopez-Montero I Rodriguez N Cribier S Pohl A Velez M Devaux PF Rapid transbilayer movement of ceramides in phospholipid vesicles and in human erythrocytes J Biol Chem 2005 280 25811 9 15883154 9 Bai J Pagano RE Measurement of Spontaneous Transfer and Transbilayer Movement of BODIPY-Labeled Lipids in Lipid Vesicles Biochemistry 1997 36 8840 8 9220970 10 Buton X Hervé P Kubelt J Tannert A Burger KNJ Fellmann P Transbilayer Movement of Monohexosylsphingolipids in Endoplasmic Reticulum and Golgi Membranes† Biochemistry 2002 41 13106 15 12390039 11 Chalat M Menon I Turan Z Menon AK Reconstitution of Glucosylceramide Flip-Flop across Endoplasmic Reticulum: IMPLICATIONS FOR MECHANISM OF GLYCOSPHINGOLIPID BIOSYNTHESIS The Journal of Biological Chemistry 2012 287 15523 32 22427661 12 Cullis PR Hope MJ Bally MB Madden TD Mayer LD Fenske DB Influence of pH gradients on the transbilayer transport of drugs, lipids, peptides and metal ions into large unilamellar vesicles Biochimica et Biophysica Acta (BBA) - Reviews on Biomembranes 1997 1331 187 211 9325441 13 Boon JM Smith BD Facilitated phosphatidylcholine flip-flop across erythrocyte membranes using low molecular weight synthetic translocases J Am Chem Soc 2001 123 6221 6 11427044 14 Boon JM Smith BD Synthetic membrane transporters Current Opinion in Chemical Biology 2002 6 749 56 12470727 15 Ogushi F Ishitsuka R Kobayashi T Sugita Y Rapid flip-flop motions of diacylglycerol and ceramide in phospholipid bilayers Chemical Physics Letters 2012 522 96 102 16 Bennett WFD Tieleman DP Molecular simulation of rapid translocation of cholesterol, diacylglycerol, and ceramide in model raft and nonraft membranes Journal of Lipid Research 2012 53 421 9 22246847 17 De Kruijff B Van Zoelen EJJ Effect of the phase transition on the transbilayer movement of dimyristoyl phosphatidylcholine in unilamellar vesicles Biochimica et Biophysica Acta (BBA) - Biomembranes 1978 511 105 15 667054 18 Wimley WC Thompson TE Exchange and flip-flop of dimyristoyl phosphatidylcholine in liquid-crystalline, gel and two-component, two-phase large unilamellar vesicles Biochemistry 1990 29 1296 303 2322564 19 John K Schreiber S Kubelt J Herrmann A Müller P Transbilayer movement of phospholipids at the main phase transition of lipid membranes: implications for rapid flip-flop in biological membranes Biophysical Journal 2002 83 3315 23 12496099 20 Morrot G Herve P Zachowski A Fellmann P Devaux PF Aminophospholipid translocase of human erythrocytes: phospholipid substrate specificity and effect of cholesterol Biochemistry 1989 28 3456 62 2742848 21 Kol MA de Kroon AIPM Killian JA de Kruijff B Transbilayer Movement of Phospholipids in Biogenic Membranes Biochemistry 2004 43 2673 81 15005602 22 Kol MA van DA de Kroon AI de KB Translocation of phospholipids is facilitated by a subset of membrane-spanning proteins of the bacterial cytoplasmic membrane J Biol Chem 2003 278 24586 93 12714595 23 Kol MA van Laak AN Rijkers DT Killian JA de Kroon AI de KB Phospholipid flop induced by transmembrane peptides in model membranes is modulated by lipid composition Biochemistry 2003 42 231 7 12515559 24 Langer M Sah R Veser A Gütlich M Langosch D Structural Properties of Model Phosphatidylcholine Flippases Chemistry & Biology 2013 20 63 72 23352140 25 Nakao H Ikeda K Ishihama Y Nakano M Membrane-Spanning Sequences in Endoplasmic Reticulum Proteins Promote Phospholipid Flip-Flop Biophysical Journal 2016 110 2689 97 27332127 26 Marx U Lassmann G Holzhutter HG Wustner D Muller P Hohlig A Rapid flip-flop of phospholipids in endoplasmic reticulum membranes studied by a stopped-flow approach Biophys J 2000 78 2628 40 10777759 27 Goren MA Morizumi T Menon I Joseph JS Dittman JS Cherezov V Constitutive phospholipid scramblase activity of a G protein-coupled receptor Nat Commun 2014 5 28 Coleman JA Vestergaard AL Molday RS Vilsen B Andersen JP Critical role of a transmembrane lysine in aminophospholipid transport by mammalian photoreceptor P4-ATPase ATP8A2 Proc Natl Acad Sci U S A 2012 109 1449 54 22307598 29 Chauhan N Farine L Pandey K Menon AK Bütikofer P Lipid topogenesis - 35 years on Biochemica et Biophysica (BBA) Molecular and Cell Biology of Lipids Forthcoming 2016 30 Bretscher MS Asymmetrical lipid bilayer structure for biological membranes Nat New Biol 1972 236 11 2 4502419 31 Verkleij AJ Zwaal RFA Roelofsen B Comfurius P Kastelijn D van Deenen LLM The asymmetric distribution of phospholipids in the human red cell membrane. A combined study using phospholipases and freeze-etch electron microscopy Biochim Biophys Acta 1973 323 178 93 4356540 32 Mioka T Fujimura-Kamada K Tanaka K Asymmetric distribution of phosphatidylserine is generated in the absence of phospholipid flippases in Saccharomyces cerevisiae Microbiology Open 2014 3 803 21 25220349 33 Sanyal S Menon AK Flipping lipids: why an whats the reason for? ACS Chem Biol 2009 34 Rothman JE Kennedy EP Rapid transmembrane movement of newly synthesized phospholipids during membrane assembly Proc Natl Acad Sci USA 1977 74 1821 5 405668 35 Bishop WR Bell RM Assembly of the endoplasmic reticulum phospholipid bilayer: the phosphatidylcholine transporter Cell 1985 42 51 60 3893747 36 Backer JM Dawidowicz EA Reconstitution of a phospholipid flippase from rat liver microsomes Nature 1987 327 341 3 3587349 37 Hrafnsdóttir S Menon AK Reconstitution and Partial Characterization of Phospholipid Flippase Activity from Detergent Extracts of the Bacillus subtilis Cell Membrane Journal of Bacteriology 2000 182 4198 206 10894727 38 Menon AK Watkins WE Hrafnsdottir S Specific proteins are required to translocate phosphatidylcholine bidirectionally across the endoplasmic reticulum Curr Biol 2000 10 241 52 10712902 39 Vishwakarma RA Vehring S Mehta A Sinha A Pomorski T Herrmann A New fluorescent probes reveal that flippase-mediated flip-flop of phosphatidylinositol across the endoplasmic reticulum membrane does not depend on the stereochemistry of the lipid Org Biomol Chem 2005 3 1275 83 15785818 40 Sanyal S Menon AK Specific transbilayer translocation of dolichol-linked oligosaccharides by an endoplasmic reticulum flippase Proc Natl Acad Sci USA 2009 106 767 72 19129492 41 Sanyal S Frank CG Menon AK Distinct flippases translocate glycerophospholipids and oligosaccharide diphosphate dolichols across the endoplasmic reticulum Biochemistry 2008 47 7937 46 18597486 42 Chang Q-l Gummadi SN Menon AK Chemical Modification Identifies Two Populations of Glycerophospholipid Flippase in Rat Liver ER† Biochemistry 2004 43 10710 8 15311932 43 Ernst CM Peschel A Broad-spectrum antimicrobial peptide resistance by MprF-mediated aminoacylation and flipping of phospholipids Molecular Microbiology 2011 80 290 9 21306448 44 Ernst CM Staubitz P Mishra NN Yang S-J Hornig G Kalbacher H The Bacterial Defensin Resistance Protein MprF Consists of Separable Domains for Lipid Lysinylation and Antimicrobial Peptide Repulsion PLoS Pathogens 2009 5 e1000660 19915718 45 Hebecker S Krausze J Hasenkampf T Schneider J Groenewold M Reichelt J Structures of two bacterial resistance factors mediating tRNA-dependent aminoacylation of phosphatidylglycerol with lysine or alanine Proceedings of the National Academy of Sciences 2015 112 10691 6 46 Ernst CM Kuhn S Slavetinsky CJ Krismer B Heilbronner S Gekeler C The Lipid-Modifying Multiple Peptide Resistance Factor Is an Oligomer Consisting of Distinct Interacting Synthase and Flippase Subunits mBio 2015 6 47 Wu G Hubbell WL Phospholipid asymmetry and transmembrane diffusion in photoreceptor disk membranes Biochemistry 1993 32 879 88 8380712 48 Hessel E Herrmann A Müller P Schnetkamp PPM Hofmann K-P The transbilayer distribution of phospholipids in disc membranes is a dynamic equilibrium European Journal of Biochemistry 2000 267 1473 83 10691986 49 Hessel E Müller P Herrmann A Hofmann KP Light-induced Reorganization of Phospholipids in Rod Disc Membranes Journal of Biological Chemistry 2001 276 2538 43 11062249 50 Menon I Huber T Sanyal S Banerjee S Barré P Canis S Opsin is a phospholipid flippase Current biology: CB 2011 21 149 53 21236677 51 Ernst OP Menon AK Phospholipid scrambling by rhodopsin Photochemical & Photobiological Sciences 2015 52 Schenk B Fernandez F Waechter CJ The ins(ide) and outs(ide) of dolichyl phosphate biosynthesis and recycling in the endoplasmic reticulum Glycobiology 2001 11 61R 70R 53 Hirschberg CB Snider MD Topography of Glycosylation in The Rough Endoplasmic Reticulum and Golgi Apparatus Annual Review of Biochemistry 1987 56 63 87 54 Frank CG Sanyal S Rush JS Waechter CJ Menon AK Does Rft1 flip an N-glycan lipid precursor? Nature 2008 454 E3 E4 18668045 55 Anand M Rush JS Ray S Doucey M-A Weik J Ware FE Requirement of the Lec35 Gene for All Known Classes of Monosaccharide-P-Dolichol-dependent Glycosyl transferase Reactions in Mammals Molecular Biology of the Cell 2001 12 487 501 11179430 56 Jelk J Gao N Serricchio M Signorell A Schmidt RS Bangs JD Glycoprotein Biosynthesis in a Eukaryote Lacking the Membrane Protein Rft1 The Journal of Biological Chemistry 2013 288 20616 23 23720757 57 Rush JS Gao N Lehrman MA Matveev S Waechter CJ Suppression of Rft1 Expression Does Not Impair the Transbilayer Movement of Man(5)GlcNAc(2)-P-P-Dolichol in Sealed Microsomes from Yeast The Journal of Biological Chemistry 2009 284 19835 42 19494107 58 Rush JS van Leyen K Ouerfelli O Wolucka B Waechter CJ Transbilayer movement of Glc-P-dolichol and its function as a glucosyl donor: protein-mediated transport of a water-soluble analog into sealed ER vesicles from pig brain Glycobiology 1998 8 1195 205 9858641 59 Rush JS Waechter CJ Transmembrane movement of a water-soluble analogue of mannosylphosphoryldolichol is mediated by an endoplasmic reticulum protein The Journal of Cell Biology 1995 130 529 36 7622555 60 Sanyal S Menon AK Stereoselective transbilayer translocation of mannosyl phosphoryl dolichol by an endoplasmic reticulum flippase Proceedings of the National Academy of Sciences of the United States of America 2010 107 11289 94 20534553 61 Rush JS Waechter CJ Functional Reconstitution into Proteoliposomes and Partial Purification of a Rat Liver ER Transport System for a Water-Soluble Analogue of Mannosylphosphoryldolichol Biochemistry 2004 43 7643 52 15182207 62 Ruiz N Lipid Flippases for Bacterial Peptidoglycan Biosynthesis Lipid insights 2015 8 21 31 26792999 63 Scheffers D-J Tol MB LipidII: Just Another Brick in the Wall? PLoS Pathogens 2015 11 e1005213 26679002 64 Mohammadi T Sijbrandi R Lutters M Verheul J Martin NI den Blaauwen T Specificity of the Transport of Lipid II by FtsW in Escherichia coli The Journal of Biological Chemistry 2014 289 14707 18 24711460 65 Mohammadi T van Dam V Sijbrandi R Vernet T Zapun A Bouhss A Identification of FtsW as a transporter of lipid-linked cell wall precursors across the membrane The EMBO Journal 2011 30 1425 32 21386816 66 Ruiz N Bioinformatics identification of MurJ (MviN) as the peptidoglycan lipid II flippase in Escherichia coli Proceedings of the National Academy of Sciences 2008 105 15553 7 67 Sham L-T Butler EK Lebar MD Kahne D Bernhardt TG Ruiz N MurJ is the flippase of lipid-linked precursors for peptidoglycan biogenesis Science 2014 345 220 2 25013077 68 Meeske AJ Sham L-T Kimsey H Koo B-M Gross CA Bernhardt TG MurJ and a novel lipid II flippase are required for cell wall biogenesis in Bacillus subtilis Proceedings of the National Academy of Sciences of the United States of America 2015 112 6437 42 25918422 69 Young KD A flipping cell wall ferry Science 2014 345 139 40 25013047 70 Islam ST Lam JS Synthesis of bacterial polysaccharides via the Wzx/Wzy-dependent pathway Canadian Journal of Microbiology 2014 60 697 716 25358682 71 Rick PD Barr K Sankaran K Kajimura J Rush JS Waechter CJ Evidence That the wzxE Gene of Escherichia coli K-12 Encodes a Protein Involved in the Transbilayer Movement of a Trisaccharide-Lipid Intermediate in the Assembly of Enterobacterial Common Antigen Journal of Biological Chemistry 2003 278 16534 42 12621029 72 Coleman JA Quazi F Molday RS Mammalian P4-ATPases and ABC transporters and their role in phospholipid transport Biochimica Et Biophysica Acta 2013 1831 555 74 23103747 73 Hankins HM Baldridge RD Xu P Graham TR Role of Flippases, Scramblases and Transfer Proteins in Phosphatidylserine Subcellular Distribution Traffic 2015 16 35 47 25284293 74 López-Marqués RL Poulsen LR Bailly A Geisler M Pomorski TG Palmgren MG Structure and mechanism of ATP-dependent phospholipid transporters Biochimica et Biophysica Acta (BBA) - General Subjects 2015 1850 461 75 24746984 75 Panatala R Hennrich H Holthuis JCM Inner workings and biological impact of phospholipid flippases Journal of Cell Science 2015 76 Saito K Fujimura-Kamada K Furuta N Kato U Umeda M Tanaka K Cdc50p, a protein required for polarized growth, associates with the Drs2p P-type ATPase implicated in phospholipid translocation in Saccharomyces cerevisiae Mol Biol Cell 2004 15 3418 32 15090616 77 Bryde S Hennrich H Verhulst PM Devaux PF Lenoir G Holthuis JC CDC50 proteins are critical components of the human class-1 P4-ATPase transport machinery J Biol Chem 2010 285 40562 72 20961850 78 Lenoir G Williamson P Puts CF Holthuis JC Cdc50p Plays a Vital Role in the ATPase Reaction Cycle of the Putative Aminophospholipid Transporter DRS2P J Biol Chem 2009 79 López-Marqués RL Poulsen LR Hanisch S Meffert K Buch-Pedersen MJ Jakobsen MK Intracellular targeting signals and lipid specificity determinants of the ALA/ALIS P4-ATPase complex reside in the catalytic ALA α-subunit Mol Biol Cell 2010 21 791 801 20053675 80 Chen B Jiang Y Zeng S Yan J Li X Zhang Y Endocytic sorting and recycling require membrane phosphatidylserine asymmetry maintained by TAT-1/CHAT-1 PLoSGenet 2010 6 e1001235 81 Puts CF Panatala R Hennrich H Tsareva A Williamson P Holthuis JCM Mapping Functional Interactions in a Heterodimeric Phospholipid Pump The Journal of Biological Chemistry 2012 287 30529 40 22791719 82 Pérez-Victoria FJ Sánchez-Cañete MP Castanys S Gamarro F Phospholipid translocation and miltefosine potency require both L. donovani miltefosine transporter and the new protein LdRos3 in Leishmania parasites J Biol Chem 2006 281 23766 75 16785229 83 Kato U Emoto K Fredriksson C Nakamura H Ohta A Kobayashi T A novel membrane protein, Ros3p, is required for phospholipid translocation across the plasma membrane in Saccharomyces cerevisiae J Biol Chem 2002 277 37855 62 12133835 84 Coleman JA Molday RS Critical role of the β-subunit CDC50A in the stable expression, assembly, subcellular localization, and lipid transport activity of the P4-ATPase ATP8A2 J Biol Chem 2011 286 17205 16 21454556 85 van der Velden LM Wichers CGK van Breevoort AED Coleman JA Molday RS Berger R Heteromeric interactions required for abundance and subcellular localization of human CDC50 proteins and class 1 P4-ATPases J Biol Chem 2010 285 40088 96 20947505 86 Poulsen LR López-Marqués RL McDowell SC Okkeri J Licht D Schulz A The Arabidopsis P4-ATPase ALA3 localizes to the Golgi and requires a β-subunit to function in lipid translocation and secretory vesicle formation Plant Cell 2008 20 658 76 18344284 87 Costa Sara R Marek M Axelsen Kristian B Theorin L Pomorski Thomas G López-Marqués Rosa L Role of post-translational modifications at the β-subunit ectodomain in complex association with a promiscuous plant P4-ATPase Biochemical Journal 2016 473 1605 15 27048590 88 Seigneuret M Devaux PF ATP-dependent asymmetric distribution of spin-labeled phospholipids in the erythrocyte membrane: relation to shape changes Proc Natl Acad Sci USA 1984 81 3751 5 6587389 89 Daleke DL Huestis WH Erythrocyte morphology reflects the transbilayer distribution of incorporated phospholipids The Journal of Cell Biology 1989 108 1375 85 2925790 90 Alder-Baerens N Lisman Q Luong L Pomorski T Holthuis JC Loss of P4 ATPases Drs2p and Dnf3p disrupts aminophospholipid transport and asymmetry in yeast post-Golgi secretory vesicles Mol Biol Cell 2006 17 1632 42 16452632 91 Zhou X Graham TR Reconstitution of phospholipid translocase activity with purified Drs2p, a type-IV P-type ATPase from budding yeast Proc Natl Acad Sci USA 2009 106 16586 91 19805341 92 Xu P Okkeri J Hanisch S Hu R-Y Xu Q Pomorski TG Identification of a novel mouse P4-ATPase family member highly expressed during spermatogenesis J Cell Sci 2009 122 2866 76 19657017 93 Pomorski T Lombardi R Riezman H Devaux PF van Meer G Holthuis JCM Drs2p-related P-type ATPases Dnf1p and Dnf2p are required for phospholipid translocation across the yeast plasma membrane and serve a role in endocytosis Mol Biol Cell 2003 14 1240 54 12631737 94 Riekhof WR Voelker DR Uptake and utilization of lyso-phosphatidylethanolamine by Saccharomyces cerevisiae J Biol Chem 2006 281 36588 96 17015438 95 Riekhof WR Wu J Gijón MA Zarini S Murphy RC Voelker DR Lysophosphatidylcholine metabolism in Saccharomyces cerevisiae: the role of P-type ATPases in transport and a broad specificity acyltransferase in acylation J Biol Chem 2007 282 36853 61 17951629 96 Riekhof WR Wu J Jones JL Voelker DR Identification and characterization of the major lysophosphatidylethanolamine acyltransferase in Saccharomyces cerevisiae J Biol Chem 2007 282 28344 52 17652094 97 Baldridge RD Xu P Graham TR Type IV P-type ATPases Distinguish Mono- versus Diacyl Phosphatidylserine Using a Cytofacial Exit Gate in the Membrane Domain J Biol Chem 2013 288 19516 27 23709217 98 Ding J Wu Z Crider BP Ma Y Li X Slaughter C Identification and functional expression of four isoforms of ATPase II, the putative aminophospholipid translocase. Effect of isoform variation on the ATPase activity and phospholipid specificity J Biol Chem 2000 275 23378 86 10801890 99 Paterson JK Renkema K Burden L Halleck MS Schlegel RA Williamson P Lipid specific activation of the murine P4-ATPase Atp8a1 (ATPase II) Biochemistry 2006 45 5367 76 16618126 100 Soupene E Kemaladewi DU Kuypers FA ATP8A1 activity and phosphatidylserine transbilayer movement J Receptor Ligand Channel Res 2008 1 1 10 20224745 101 Lee S Uchida Y Wang J Matsudaira T Nakagawa T Kishimoto T Transport through recycling endosomes requires EHD1 recruitment by a phosphatidylserine translocase The EMBO Journal 2015 34 669 88 25595798 102 Coleman JA Kwok MC Molday RS Localization, purification and functional reconstitution of the P4-ATPASE, ATP8A2, a phosphatidylserine flippase in photoreceptor disc membranes J Biol Chem 2009 103 Wang L Beserra C Garbers DL A novel aminophospholipid transporter exclusively expressed in spermatozoa is required for membrane lipid asymmetry and normal fertilization Dev Biol 2004 267 203 15 14975727 104 Yabas M Coupland LA Cromer D Winterberg M Teoh NC D’Rozario J Mice Deficient in the Putative Phospholipid Flippase ATP11C Exhibit Altered Erythrocyte Shape, Anemia, and Reduced Erythrocyte Life Span The Journal of Biological Chemistry 2014 289 19531 7 24898253 105 Yabas M Jing W Shafik S Bröer S Enders A ATP11C Facilitates Phospholipid Translocation across the Plasma Membrane of All Leukocytes PLoS ONE 2016 11 e0146774 26799398 106 Segawa K Kurata S Yanagihashi Y Brummelkamp TR Matsuda F Nagata S Caspase-mediated cleavage of phospholipid flippase for apoptotic phosphatidylserine exposure Science 2014 344 1164 8 24904167 107 Segawa K Kurata S Nagata S Human Type IV P-type ATPases That Work as Plasma Membrane Phospholipid Flippases and Their Regulation by Caspase and Calcium Journal of Biological Chemistry 2016 291 762 72 26567335 108 Takada N Takatsu H Miyano R Nakayama K Shin H-W ATP11C mutation is responsible for the defect in phosphatidylserine uptake in UPS-1 cells Journal of Lipid Research 2015 56 2151 7 26420878 109 Paulusma CC Folmer DE Ho-Mok KS de Waart DR Hilarius PM Verhoeven AJ ATP8B1 requires an accessory protein for endoplasmic reticulum exit and plasma membrane lipid flippase activity Hepatology 2008 47 268 78 17948906 110 Ray NB Durairaj L Chen BB McVerry BJ Ryan AJ Donahoe M Dynamic regulation of cardiolipin by the lipid pump, ATP8b1, determines the severity of lung injury in experimental bacterial pneumonia Nature medicine 2010 16 1120 7 111 Takatsu H Tanaka G Segawa K Suzuki J Nagata S Nakayama K Phospholipid Flippase Activities and Substrate Specificities of Human Type IV P-type ATPases Localized to the Plasma Membrane Journal of Biological Chemistry 2014 112 Naito T Takatsu H Miyano R Takada N Nakayama K Shin H-W Phospholipid Flippase ATP10A Translocates Phosphatidylcholine and Is Involved in Plasma Membrane Dynamics Journal of Biological Chemistry 2015 290 15004 17 25947375 113 Chen R Brady E McIntyre TM Human TMEM30a Promotes Uptake of Antitumor and Bioactive Choline Phospholipids into Mammalian Cells The Journal of Immunology 2011 186 3215 25 21289302 114 Axelsen KB Palmgren MG Evolution of substrate specificities in the P-type ATPase superfamily J Mol Evol 1998 46 84 101 9419228 115 Poulsen LR Lopez-Marques RL Pedas PR McDowell SC Brown E Kunze R A phospholipid uptake system in the model plant Arabidopsis thaliana Nat Commun 2015 6 116 Lopez-Marques R Theorin L Palmgren M Pomorski T P4-ATPases: lipid flippases in cell membranes Pflugers Arch - Eur J Physiol 2014 466 1227 40 24077738 117 Higgins CF Gottesman MM Is the multidrug transporter a flippase? Trends Biochem Sci 1992 17 18 21 1374941 118 Linton KJ Higgins CF The Escherichia coli ATP-binding cassette (ABC) proteins Molecular Microbiology 1998 28 5 13 9593292 119 Kemp S Theodoulou FL Wanders RJA Mammalian peroxisomal ABC transporters: from endogenous substrates to pathology and clinical significance British Journal of Pharmacology 2011 164 1753 66 21488864 120 Shulenin S Nogee LM Annilo T Wert SE Whitsett JA Dean M ABCA3 Gene Mutations in Newborns with Fatal Surfactant Deficiency New England Journal of Medicine 2004 350 1296 303 15044640 121 Sakai K Akiyama M Sugiyama-Nakagiri Y McMillan JR Sawamura D Shimizu H Localization of ABCA12 from Golgi apparatus to lamellar granules in human upper epidermal keratinocytes Experimental Dermatology 2007 16 920 6 17927575 122 Whitsett JA Wert SE Weaver TE Alveolar Surfactant Homeostasis and the Pathogenesis of Pulmonary Disease Annual review of medicine 2010 61 105 19 123 Ile KE Davis W Jr Boyd JT Soulika AM Tew KD Identification of a novel first exon of the human ABCA2 transporter gene encoding a unique N-terminus Biochimica et Biophysica Acta (BBA) - Gene Structure and Expression 2004 1678 22 32 15093135 124 Kubo Y Sekiya S Ohigashi M Takenaka C Tamura K Nada S ABCA5 Resides in Lysosomes, and ABCA5 Knockout Mice Develop Lysosomal Disease-Like Symptoms Molecular and Cellular Biology 2005 25 4138 49 15870284 125 Tarling EJ Edwards PA ATP binding cassette transporter G1 (ABCG1) is an intracellular sterol transporter Proceedings of the National Academy of Sciences 2011 108 19719 24 126 Tarr PT Edwards PA ABCG1 and ABCG4 are coexpressed in neurons and astrocytes of the CNS and regulate cholesterol homeostasis through SREBP-2 Journal of Lipid Research 2008 49 169 82 17916878 127 Romsicki Y Sharom FJ Phospholipid flippase activity of the reconstituted P-glycoprotein multidrug transporter Biochemistry 2001 40 6937 47 11389609 128 Smit JJM Schinkel AH Oude Elferink RPJ Groen AK Wagenaar E van Deemter L Homozygous disruption of the murine MDR2 P-glycoprotein gene leads to a complete absence of phospholipid from bile and to liver disease Cell 1993 75 451 62 8106172 129 Huang Z Chang X Riordan JR Huang Y Fluorescent modified phosphatidylcholine floppase activity of reconstituted multidrug resistance-associated protein MRP1 Biochim Biophys Acta 2004 1660 155 63 14757231 130 Kamp D Haest CW Evidence for a role of the multidrug resistance protein (MRP) in the outward translocation of NBD-phospholipids in the erythrocyte membrane Biochimica Et Biophysica Acta 1998 1372 91 101 9651491 131 Raggers RJ van Helvoort A Evers R van Meer G The human multidrug resistance protein MRP1 translocates sphingolipid analogs across the plasma membrane J Cell Sci 1999 112 Pt 3 415 22 9885294 132 Dekkers DW Comfurius P van Gool RG Bevers EM Zwaal RF Multidrug resistance protein 1 regulates lipid asymmetry in erythrocyte membranes Biochemical Journal 2000 350 531 5 10947968 133 Quazi F Molday RS Differential Phospholipid Substrates and Directional Transport by ATP Binding Cassette Proteins ABCA1, ABCA7, and ABCA4 and Disease-causing Mutants J Biol Chem 2013 134 Decottignies A Grant AM Nichols JW de Wet H McIntosh DB Goffeau A ATPase and multidrug transport activities of the overexpressed yeast ABC protein Yor1p J Biol Chem 1998 273 12612 22 9575223 135 Gulshan K Moye-Rowley WS Vacuolar import of phosphatidylcholine requires the ATP-binding cassette transporter Ybt1 Traffic 2011 12 1257 68 21649806 136 Pighin JA Zheng H Balakshin LJ Goodman IP Western TL Jetter R Plant Cuticular Lipid Export Requires an ABC Transporter Science 2004 306 702 4 15499022 137 Bird D Beisson F Brigham A Shin J Greer S Jetter R Characterization of Arabidopsis ABCG11/WBC11, an ATP binding cassette (ABC) transporter that is required for cuticular lipid secretion† The Plant Journal 2007 52 485 98 17727615 138 Luo B Xue X-Y Hu W-L Wang L-J Chen X-Y An ABC Transporter Gene of Arabidopsis thaliana, AtWBC11, is Involved in Cuticle Development and Prevention of Organ Fusion Plant and Cell Physiology 2007 48 1790 802 17989085 139 Panikashvili D Savaldi-Goldstein S Mandel T Yifhar T Franke RB Höfer R The Arabidopsis DESPERADO/AtWBC11 Transporter Is Required for Cutin and Wax Secretion Plant Physiology 2007 145 1345 60 17951461 140 Ukitsu H Kuromori T Toyooka K Goto Y Matsuoka K Sakuradani E Cytological and Biochemical Analysis of COF1, an Arabidopsis Mutant of an ABC Transporter Gene Plant and Cell Physiology 2007 48 1524 33 17971336 141 Quazi F Lenevich S Molday RS ABCA4 is an N-retinylidene-phosphatidylethanolamine and phosphatidylethanolamine importer Nat Commun 2012 3 925 22735453 142 Boujaoude LC Bradshaw-Wilder C Mao C Cohn J Ogretmen B Hannun YA Cystic Fibrosis Transmembrane Regulator Regulates Uptake of Sphingoid Base Phosphates and Lysophosphatidic Acid: MODULATION OF CELLULAR ACTIVITY OF SPHINGOSINE 1-PHOSPHATE Journal of Biological Chemistry 2001 276 35258 64 11443135 143 Lorenz RT Parks LW Involvement of heme components in sterol metabolism of Saccharomyces cerevisiae Lipids 1991 26 598 603 1779707 144 Wilcox LJ Balderes DA Wharton B Tinkelenberg AH Rao G Sturley SL Transcriptional Profiling Identifies Two Members of the ATP-binding Cassette Transporter Superfamily Required for Sterol Uptake in Yeast Journal of Biological Chemistry 2002 277 32466 72 12077145 145 Li Y Prinz WA ATP-binding Cassette (ABC) Transporters Mediate Nonvesicular, Raft-modulated Sterol Movement from the Plasma Membrane to the Endoplasmic Reticulum Journal of Biological Chemistry 2004 279 45226 34 15316012 146 Reiner S Micolod D Zellnig G Schneiter R A Genomewide Screen Reveals a Role of Mitochondria in Anaerobic Uptake of Sterols in Yeast Molecular Biology of the Cell 2006 17 90 103 16251356 147 Nagi M Tanabe K Ueno K Nakayama H Aoyama T Chibana H The Candida glabrata sterol scavenging mechanism, mediated by the ATP-binding cassette transporter Aus1p, is regulated by iron limitation Molecular Microbiology 2013 88 371 81 23448689 148 Marek M Silvestro D Fredslund MD Andersen TG Pomorski TG Serum albumin promotes ATP-binding cassette transporter-dependent sterol uptake in yeast FEMS Yeast Research 2014 14 1223 33 25331273 149 Wang Z Anderson NS Benning C The Phosphatidic Acid Binding Site of the Arabidopsis Trigalactosyldiacylglycerol 4 (TGD4) Protein Required for Lipid Import into Chloroplasts The Journal of Biological Chemistry 2013 288 4763 71 23297418 150 Roston RL Gao J Murcha MW Whelan J Benning C TGD1, -2, and -3 Proteins Involved in Lipid Trafficking Form ATP-binding Cassette (ABC) Transporter with Multiple Substrate-binding Proteins The Journal of Biological Chemistry 2012 287 21406 15 22544736 151 Zhou Z White KA Polissi A Georgopoulos C Raetz CRH Function of Escherichia coli MsbA, an essential ABC family transporter, in lipid A and phospholipid biosynthesis J Biol Chem 1998 273 12466 75 9575204 152 Doerrler WT Reedy MC Raetz CR An Escherichia coli mutant defective in lipid export J Biol Chem 2001 276 11461 4 11278265 153 Eckford PD Sharom FJ The reconstituted Escherichia coli MsbA protein displays lipid flippase activity Biochemical Journal 2010 429 195 203 20412049 154 Tefsen B Geurtsen J Beckers F Tommassen J de Cock H Lipopolysaccharide Transport to the Bacterial Outer Membrane in Spheroplasts Journal of Biological Chemistry 2005 280 4504 9 15576375 155 Wacker M Linton D Hitchen PG Nita-Lazar M Haslam SM North SJ N-Linked Glycosylation in Campylobacter jejuni and Its Functional Transfer into E. coli Science 2002 298 1790 3 12459590 156 Young NM Brisson J-R Kelly J Watson DC Tessier L Lanthier PH Structure of the N-Linked Glycan Present on Multiple Glycoproteins in the Gram-negative Bacterium, Campylobacter jejuni Journal of Biological Chemistry 2002 277 42530 9 12186869 157 Alaimo C Catrein I Morf L Marolda CL Callewaert N Valvano MA Two distinct but interchangeable mechanisms for flipping of lipid-linked oligosaccharides EMBO J 2006 25 967 76 16498400 158 Perez C Gerber S Boilevin J Bucher M Darbre T Aebi M Structure and mechanism of an active lipid-linked oligosaccharide flippase Nature 2015 524 433 8 26266984 159 Verchere A Menon AK Structural biology: Lipid gymnastics Nature 2015 524 420 2 26266983 160 Lehrman MA Flipping a Lipid-Linked Oligosaccharide? You Must Whip It! Trends in Biochemical Sciences 2015 40 715 7 26476576 161 Bevers EM Williamson PL Phospholipid scramblase: An update FEBS Letters 2010 584 2724 30 20302864 162 Ruggiero L Connor MP Chen J Langen R Finnemann SC Diurnal, localized exposure of phosphatidylserine by rod outer segment tips in wild-type but not Itgb5−/− or Mfge8−/− mouse retina Proceedings of the National Academy of Sciences 2012 109 8145 8 163 Suzuki J Denning DP Imanishi E Horvitz HR Nagata S Xk-Related Protein 8 and CED-8 Promote Phosphatidylserine Exposure in Apoptotic Cells Science 2013 341 403 6 23845944 164 Suzuki J Fujii T Imao T Ishihara K Kuba H Nagata S Calcium-dependent phospholipid scramblase activity of TMEM16 protein family members J Biol Chem 2013 288 13305 16 23532839 165 Suzuki J Umeda M Sims PJ Nagata S Calcium-dependent phospholipid scrambling by TMEM16F Nature 2010 468 834 8 21107324 166 Suzuki J Imanishi E Nagata S Exposure of phosphatidylserine by Xk-related protein family members during apoptosis Journal of Biological Chemistry 2014 167 Malvezzi M Chalat MN Janjusevic R Picollo A Terashima H Menon AK Ca(2+)-dependent phospholipid scrambling by a reconstituted TMEM16 ion channel Nature communications 2013 4 2367 168 Brunner JD Lim NK Schenck S Duerst A Dutzler R X-ray structure of a calcium-activated TMEM16 lipid scramblase Nature 2014 516 207 12 25383531 169 Yu K Whitlock JM Lee K Ortlund EA Yuan Cui Y Hartzell HC Identification of a lipid scrambling domain in ANO6/TMEM16F eLife 2015 4 e06901 26057829 170 van Meer G Voelker DR Feigenson GW Membrane lipids: where they are and how they behave Nat Rev Mol Cell Biol 2008 9 112 24 18216768 171 Kälin N Fernandes J Hrafnsdóttir S van Meer G Natural phosphatidylcholine is actively translocated across the plasma membrane to the surface of mammalian cells J Biol Chem 2004 279 33228 36 15175345 172 Dekkers DWC Comfurius P Schroit AJ Bevers EM Zwaal RFA Transbilayer movement of NBD-labeled phospholipids in red blood cell membranes: outward-directed transport by the multidrug resistance protein 1 (MRP1) Biochemistry 1998 37 14833 7 9778357 173 van Helvoort A Smith AJ Sprong H Fritzsche I Schinkel AH Borst P MDR1 P-glycoprotein is a lipid translocase of broad specificity, while MDR3 P-glycoprotein specifically translocates phosphatidylcholine Cell 1996 87 507 17 8898203 174 Heinrich R Brumen M Jaeger A Müller P Herrmann A Modelling of Phospholipid Translocation in the Erythrocyte Membrane: A Combined Kinetic and Thermodynamic Approach Journal of Theoretical Biology 1997 185 295 312 9156082 175 Frickenhaus S Heinrich R Kinetic and thermodynamic aspects of lipid translocation in biological membranes Biophysical Journal 1999 76 1293 309 10049313 176 Fairn GD Schieber NL Ariotti N Murphy S Kuerschner L Webb RI High-resolution mapping reveals topologically distinct cellular pools of phosphatidylserine The Journal of Cell Biology 2011 194 257 75 21788369 177 Tauchi-Sato K Ozeki S Houjou T Taguchi R Fujimoto T The Surface of Lipid Droplets Is a Phospholipid Monolayer with a Unique Fatty Acid Composition Journal of Biological Chemistry 2002 277 44507 12 12221100 178 Holthuis JCM Menon AK Lipid landscapes and pipelines in membrane homeostasis Nature 2014 510 48 57 24899304 179 Chung J Torta F Masai K Lucast L Czapla H Tanner LB PI4P/phosphatidylserine countertransport at ORP5- and ORP8-mediated ER–plasma membrane contacts Science 2015 349 428 32 26206935 180 Maeda K Anand K Chiapparino A Kumar A Poletto M Kaksonen M Interactome map uncovers phosphatidylserine transport by oxysterol-binding proteins Nature 2013 501 257 61 23934110 181 Moser von Filseck J Čopič A Delfosse V Vanni S Jackson CL Bourguet W Phosphatidylserine transport by ORP/Osh proteins is driven by phosphatidylinositol 4-phosphate Science 2015 349 432 6 26206936 182 Hubbell WL Transbilayer coupling mechanism for the formation of lipid asymmetry in biological membranes. Application to the photoreceptor disc membrane Biophysical Journal 1990 57 99 108 2297564 183 Zhou X Sebastian TT Graham TR Auto-inhibition of Drs2p, a Yeast Phospholipid Flippase, by Its Carboxyl-terminal Tail The Journal of Biological Chemistry 2013 288 31807 15 24045945 184 Roelants FM Baltz AG Trott AE Fereres S Thorner J A protein kinase network regulates the function of aminophospholipid flippases Proc Natl Acad Sci USA 2010 107 34 9 19966303 185 Coleman JA Zhu X Djajadi HR Molday LL Smith RS Libby RT Phospholipid flippase ATP8A2 is required for normal visual and auditory function and photoreceptor and spiral ganglion cell survival Journal of Cell Science 2014 127 1138 49 24413176 186 Matsuzaki K Murase O Fujii N Miyajima K An Antimicrobial Peptide, Magainin 2, Induced Rapid Flip-Flop of Phospholipids Coupled with Pore Formation and Peptide Translocation Biochemistry 1996 35 11361 8 8784191 187 Illya G Deserno M Coarse-Grained Simulation Studies of Peptide-Induced Pore Formation Biophysical Journal 2008 95 4163 73 18641080 188 Gurtovenko AA Anwar J Vattulainen I Defect-Mediated Trafficking across Cell Membranes: Insights from in Silico Modeling Chemical Reviews 2010 110 6077 103 20690701 189 Gurtovenko AA Vattulainen I Lipid transmembrane asymmetry and intrinsic membrane potential: two sides of the same coin J Am Chem Soc 2007 129 5358 9 17417854 190 Vestergaard AL Coleman JA Lemmin T Mikkelsen SA Molday LL Vilsen B Critical roles of isoleucine-364 and adjacent residues in a hydrophobic gate control of phospholipid transport by the mammalian P4-ATPase ATP8A2 Proceedings of the National Academy of Sciences of the United States of America 2014 111 E1334 E43 24706822 191 Obara K Miyashita N Xu C Toyoshima I Sugita Y Inesi G Structural role of countertransport revealed in Ca2+ pump crystal structure in the absence of Ca2+ Proc Natl Acad Sci USA 2005 102 14489 96 16150713 192 Baldridge RD Graham TR Identification of residues defining phospholipid flippase substrate specificity of type IV P-type ATPases Proc Natl Acad Sci USA 2012 109 E290 E8 22308393 193 Ward AB Guvench O Hills RD Jr Coarse grain lipid-protein molecular interactions and diffusion with MsbA flippase Proteins 2012 80 2178 90 22552999 194 Pomorski T Menon AK Lipid flippases and their biological functions Cell MolLife Sci 2006 63 2908 21 195 Guex N Peitsch MC SWISS-MODEL and the Swiss-Pdb Viewer: An environment for comparative protein modeling ELECTROPHORESIS 1997 18 2714 23 9504803 196 Larsen J Hatzakis NS Stamou D Observation of Inhomogeneity in the Lipid Composition of Individual Nanoscale Liposomes Journal of the American Chemical Society 2011 133 10685 7 21688773 197 Liu J Conboy JC 1,2-Diacyl-Phosphatidylcholine Flip-Flop Measured Directly by Sum-Frequency Vibrational Spectroscopy Biophysical Journal 2005 89 2522 32 16085770 198 Anglin TC Conboy JC Lateral Pressure Dependence of the Phospholipid Transmembrane Diffusion Rate in Planar-Supported Lipid Bilayers Biophysical Journal 2008 95 186 93 18339755 199 Anglin TC Cooper MP Li H Chandler K Conboy JC Free Energy and Entropy of Activation for Phospholipid Flip-Flop in Planar Supported Lipid Bilayers The Journal of Physical Chemistry B 2010 114 1903 14 20073520 200 Kiessling V Crane JM Tamm LK Transbilayer Effects of Raft-Like Lipid Domains in Asymmetric Planar Bilayers Measured by Single Molecule Tracking Biophysical Journal 2006 91 3313 26 16905614 201 Khan A Ducker WA Mao M Flip-Flop in Adsorbed Bilayers The Journal of Physical Chemistry B 2006 110 23365 72 17107187 202 Jacquot A Montigny C Hennrich H Barry R le Maire M Jaxel C Phosphatidylserine stimulation of Drs2p.Cdc50p lipid translocase dephosphorylation is controlled by phosphatidylinositol-4-phosphate J Biol Chem 2012 287 13249 61 22351780 203 Groen A Romero MR Kunne C Hoosdally SJ Dixon PH Wooding C Complementary functions of the flippase ATP8B1 and the floppase ABCB4 in maintaining canalicular membrane integrity Gastroenterology 2011 141 1927 37 21820390 204 Helle SCJ Kanfer G Kolar K Lang A Michel AH Kornmann B Organization and function of membrane contact sites Biochimica et Biophysica Acta (BBA) - Molecular Cell Research 2013 1833 2526 41 23380708 205 Wang Z Benning C Chloroplast lipid synthesis and lipid trafficking through ER–plastid membrane contact sites Biochemical Society Transactions 2012 40 457 63 22435830 206 Janssen MJFW Koorengevel MC de Kruijff B de Kroon AIPM Transbilayer movement of phosphatidylcholine in the mitochondrial outer membrane of Saccharomyces cerevisiae is rapid and bidirectional Biochimica et Biophysica Acta (BBA) - Biomembranes 1999 1421 64 76 10561472 207 Gallet PF Zachowski A Julien R Fellmann P Devaux PF Maftah A Transbilayer movement and distribution of spin-labelled phospholipids in the inner mitochondrial membrane Biochimica et Biophysica Acta (BBA) - Biomembranes 1999 1418 61 70 10209211 208 Rajasekharan A Gummadi SN Flip-Flop of Phospholipids in Proteoliposomes Reconstituted from Detergent Extract of Chloroplast Membranes: Kinetics and Phospholipid Specificity PLoS ONE 2011 6 e28401 22174798

PMC005xxxxxx/PMC5127731.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 7809375 3455 Dev Neurosci Dev. Neurosci. Developmental neuroscience 0378-5866 1421-9859 27622292 5127731 10.1159/000448585 NIHMS808399 Article Hypothermia and rewarming activate a macroglial unfolded protein response independent of hypoxic-ischemic brain injury in neonatal piglets Lee Jennifer K. MD a Wang Bing MD, PhD a Reyes Michael BA a Armstrong Jillian S. BS a Kulikowicz Ewa MS a Santos Polan T. MD a Lee Jeong-Hoo BS a Koehler Raymond C. PhD a Martin Lee J. PhD b a Department of Anesthesiology and Critical Care Medicine, Johns Hopkins University, Baltimore, Maryland b Department of Pathology, Division of Neuropathology, Johns Hopkins University, Baltimore, Maryland Corresponding Author: Jennifer K. Lee, MD, Department of Anesthesiology and Critical Care Medicine, Johns Hopkins Hospital, Charlotte R. Bloomberg Children's Center, 1800 Orleans Street, Room 6321, Baltimore, MD 21287, Phone: 410-955-6412, FAX: 410-502-5312, jklee@jhmi.edu 15 9 2016 14 9 2016 2016 14 9 2017 38 4 277294 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. Therapeutic hypothermia provides incomplete neuroprotection after hypoxia-ischemia (HI)-induced brain injury in neonates. We previously showed that cortical neuron and white matter apoptosis are promoted by hypothermia and early rewarming in a piglet model of HI. The unfolded protein response (UPR) may be one of the potential mediators of this cell death. Here, neonatal piglets underwent HI or sham surgery followed by 29 hours of normothermia, 2 hours of normothermia+27 hours of hypothermia or 18 hours of hypothermia+rewarming. Piglets recovered for 29 hours. Immunohistochemistry for endoplasmic reticulum to nucleus signaling-1 protein (ERN1), a marker of UPR activation, was used to determine the ratios of ERN1+ macroglia and neurons in the motor subcortical white matter and cerebral cortex. The ERN1+ macroglia were immunophenotyped as oligodendrocytes and astrocytes by immunofluorescent co-labeling. Temperature (p=0.046) and HI (p<0.001) independently affected the ratio of ERN1+ macroglia. In sham piglets, sustained hypothermia (p=0.011) and rewarming (p=0.004) increased the ERN1+ macroglia ratio above that in normothermia. HI prior to hypothermia diminished the UPR. Ratios of ERN1+ macroglia correlated to white matter apoptotic profile counts in shams (r=0.472; p=0.026), thereby associating UPR activation with white matter apoptosis during hypothermia and rewarming. Accordingly, macroglial cell counts decreased in shams that received sustained hypothermia (p=0.009) or rewarming (p=0.007) compared to those in normothermic shams. HI prior to hypothermia neutralized the macroglial cell loss. Neither HI nor temperature affected ERN1+ neuron ratios. In summary, delayed hypothermia and rewarming activate the macroglial UPR, which is associated with white matter apoptosis. HI may decrease the macroglial endoplasmic reticulum stress response after hypothermia and rewarming. neonate brain injury ischemia endoplasmic reticulum hypothermia apoptosis INTRODUCTION Therapeutic hypothermia reduces the risk of death and neurologic disability in neonatal hypoxic-ischemic encephalopathy (HIE). [1] However, approximately half of survivors have persistent neurodevelopmental disabilities despite receiving hypothermia. [1, 2] We theorize that adverse effects from delaying the induction of hypothermia [3] and rewarming [4, 5] may be partially responsible for these neurologic impairments. We previously demonstrated that rewarming increases cortical neuronal apoptosis and caspase-3 cleavage after hypoxia-ischemia (HI) in a piglet model. [4] Hypothermia and rewarming also promote glial apoptosis in the white matter, an effect that may be independent of HI in some brain regions. [5] Cortical and white matter injuries on MRI predict neurodevelopmental and motor impairments in survivors of HIE. [6, 7] Thus, there is need to define the mechanisms through which apoptotic cascades are activated after HI, hypothermia, and rewarming; to clarify targets for adjuvant therapies; and to minimize off-target effects from hypothermia and rewarming. HI, hypothermia, and rewarming are cellular stressors that alter cellular metabolism, protein synthesis, quality control, and survival. [4, 5, 8-10] These events can be upstream signals for cell death. Endoplasmic reticulum (ER) stress is involved in the cellular heat-shock response [11] and may also mediate cellular injury during hypothermia and rewarming. ER stress activates the canonical unfolded protein response (UPR) that slows protein accumulation in the ER lumen and upregulates gene transcription of ER-resident chaperones and enzymes to relieve the ER stress. [12, 13] UPR activation is generally considered to be neuroprotective. [14, 15] However, in some adult white matter diseases, such as multiple sclerosis and Charcot-Marie-Tooth, [16] the UPR becomes maladaptive and induces apoptosis in myelinating glia. [17] Information is scant on the UPR's role in ischemic brain injury. ER stress from ischemia and reperfusion activates the UPR, and it remains activated for 3 hours after HI in neonatal rats. [12] ER stress with UPR dysfunction also has been reported in adult rodent models of cerebral ischemia. [18] Whether the UPR persists for a longer duration after HI and during hypothermia and rewarming must be determined if we are to clarify the UPR's potential as a clinical biomarker or therapeutic target. Therefore, we used a piglet model of HI to test the hypotheses that 1) HI with normothermic recovery will activate the UPR whereas overnight hypothermia will suppress the UPR after HI; 2) rewarming will activate the UPR; and 3) UPR activation will be associated with apoptosis. We tested these hypotheses in white matter macroglial cells and cortical neurons and used the same rewarming rate of 0.5°C/hour that is used in clinical practice. [3] MATERIALS AND METHODS Animal preparation All procedures were approved by the Johns Hopkins University Animal Care and Use Committee, and all protocols abided by the United States Public Health Service Policy on the Humane Care and Use of Laboratory Animals and the Guide for the Care and Use of Laboratory Animals. Animal care was in compliance with the National Institutes of Health Guidelines and ensured comfort throughout the experiments. Male piglets (3–5 days old, 1–2.5 kg) were randomized to receive HI injury or sham surgery. Piglets in the sham group received the same anesthetic, duration of anesthesia, and surgery as those in the HI group. The median age in each experimental group was 3–4 days. An age-matched, male, naïve group of piglets that did not receive anesthesia, surgery, or HI was an additional control group. We anesthetized and intubated the piglets with oxygen, nitrous oxide, and isoflurane as previously described. [4] After inserting femoral arterial and venous catheters, which takes approximately 10–15 minutes, we discontinued the isoflurane, administered 70% nitrous oxide with 30% oxygen, and gave an intravenous (IV) bolus of fentanyl 20 μg/kg followed by a continuous infusion of 10–20 μg/kg/h. We administered vecuronium at 0.2 mg/kg/h IV to all piglets to prevent shivering during hypothermia and to ensure that all treatment groups received the same anesthetic regimen. We previously showed that this protocol does not alter the level of apoptosis above normal, baseline, developmental programmed cell death. [4, 5] Low-dose phenylephrine or dopamine infusions were initiated, when necessary, to maintain the piglets’ mean arterial blood pressure at ≥45 mmHg and above the lower limit of autoregulation. [19] The piglets also received 5% dextrose in 0.45% saline at 10 ml/hour IV. HI injury We have previously reported the HI injury protocol. [4, 5, 19-21] Briefly, the inspired oxygen concentration is reduced to 10% for 45 minutes. A brief reoxygenation period with 21% oxygen for 5 minutes is used to increase the likelihood of cardiac resuscitation. Then, we clamp the endotracheal tube to produce 7–8 minutes of asphyxia. Piglets are resuscitated with chest compressions, 100 μg/kg epinephrine, and 50% oxygen. Piglets without return of spontaneous circulation (ROSC) after 3 minutes of compressions are excluded. Once ROSC is attained, we reduce the oxygen concentration to 30% for the rest of the experiment. The resultant metabolic acidosis is corrected with sodium bicarbonate, and hypocalcemia is treated with calcium chloride. Piglets evaluated by histology received 7 minutes of asphyxia, and piglets evaluated by immunoblotting received 7–8 minutes of asphyxia (11 pigs with 7 minutes of asphyxia and 7 pigs with 8 minutes of asphyxia). Temperature Piglets in the HI or sham groups were randomized to one of three temperature treatments as previously described [4, 5]: 1) normothermia (38.5–39.5°C), 2) whole body hypothermia (34.0°C), or 3) hypothermia followed by rewarming (0.5°C/hour) until the goal sample size in each group was reached. (Fig 1) This rewarming rate is the same as that used clinically. [3] In piglets assigned to receive hypothermia, we initiated cooling with ice packs and a cooling blanket at 2 hours after ROSC (or the time equivalent in shams) to mimic clinical delays in hypothermia induction. In the rewarming groups, whole body temperature was increased beginning at 20 hours (after 18 h of hypothermia) at a goal rate of 0.5°C/hour until normothermia (38.5°C) was reached. All piglets were euthanized at 29 hours. Histology Pentobarbital (50 mg/kg, IV) was administered to achieve a deep level of anesthesia, and then piglets were transcardially perfused with cold phosphate-buffered saline (PBS) and 4% paraformaldehyde for brain tissue fixation. Paraffin embedded forebrains were cut into 10-μm sections. Anterior-posterior levels in coronal sections were matched according to anatomic regions. We used immunoperoxidase histochemistry on paraffin sections as previously described [5] to identify neurons and macroglia that were positive for endoplasmic reticulum to nucleus signaling 1 (ERN1) in the cerebral cortex and subcortical white matter of the motor gyrus at the anterior striatal and posterior hippocampal levels. ERN1 is a marker of UPR activation also known as inositol requiring enzyme-1-α. [22] Mouse monoclonal IgG anti-ERN1 (1:200, Abgent, San Diego, CA) was used. We tested several ERN1 antibodies in piglet tissue for comparison to a positive control recombinant ERN1 protein (Origene, Rockville, MD). We chose the ERN1 antibody that had selective specificity in piglet tissue at the molecular weight that corresponded to the recombinant ERN1 protein; this antibody has been well characterized and used for immunocytochemistry. [23] In addition, we verified that our ERN1 antibody did not cross-react with a recombinant ERN2 protein (ABM, Inc., Richmond, BC, Canada) by western blot. To distinguish the morphology of microglial cells from that of oligodendrocytes and astrocytes, additional brain sections were incubated with the primary antibody anti-ionized calcium-binding adapter molecule 1 (Iba-1; 1:10, Wako, Richmond, VA). We developed the sections using 3,3′-diaminobenzidine substrate followed by a cresyl violet counterstain. Negative control tissue sections were exposed to secondary antibody without primary antibody. A single investigator (JKL) who was blinded to treatment group manually counted ERN1+ and ERN1-negative macroglia in six random microscope fields in the center of the subcortical white matter of the motor gyrus at 1000X magnification with oil immersion. A second investigator (BW) who was also blinded to treatment group manually counted ERN1+ and ERN1-negative neurons on one side of the motor gyrus, from the sulcus between the cingulate and motor gyri to the tip of the motor gyrus, in cortical layers 2 and 3 and the beginning of layer 4 at 400X magnification. (Fig 2) Macroglia were identified by size (usually <10 μm in diameter) and distinguished from neurons by both size and neuronal morphologic criteria as previously reported. [5] Criteria for neuron identification included a large cell body (typically 8–20 μm in diameter), multipolar or triangular morphology, open non-condensed nucleus with chromatin strands and often a nucleolus. ERN1+ macroglia profiles were defined by intracellular brown staining within the nucleus and cytoplasm. ERN1+ neurons were identified by brown immunoreactivity seen as discrete foci within the nucleus. (Fig 3A–C). Microglia were distinguished from astrocytes and oligodendrocytes morphologically and immunophenotypically; microglia were not included in the ERN1 cell counts. Microglia, as shown by Iba-1 immunoreactivity, have irregular and elongated nuclear morphology and soma with branch points and processes (Fig 3D). [24] We did not encounter activated, large, transformed microglia with a macrophage-like morphology in cortex or white matter. Therefore, putative astrocytes and oligodendrocytes were counted, but cells with a microglial morphology were not counted. An experimental neuropathologist (LJM) verified the identification and ERN1 classification of the macroglia and neurons. To phenotype the ERN1+ macroglia in the white matter, we used immunofluorescence to co-label cells with ERN1, glial fibrillary acidic protein (GFAP), and 4',6-diamidino-2-phenylindole (DAPI) or with ERN1, oligodendrocyte transcription factor lineage protein-2 (olig2), and DAPI to identify astrocytes and oligodendrocytes, respectively, with intranuclear ERN1 staining. We used mouse monoclonal IgG anti-ERN1 primary antibody (1:50; Abgent) and rabbit polyclonal IgG anti-GFAP primary antibody (1:200, Dako, Carpinteria, CA) to immunophenotype ERN1+ astrocytes or mouse monoclonal IgG anti-ERN1 primary antibody (1:25; Abgent) and rabbit polyclonal IgG anti-olig2 primary antibody (1:20, GeneTex, Inc., Irvine, CA) to immunophenotype ERN1+ oligodendrocytes. We used the secondary antibodies goat polyclonal IgG anti-mouse Alexa Fluor 488 conjugate (1:50, Thermo Scientific, Rockford, IL) and goat polyclonal IgG anti-rabbit Alexa Fluor 555 conjugate (1:50, Thermo Scientific, Rockford, IL) as well as VectaShield hard set mounting medium with DAPI (Vector Laboratories, Burlingame, CA). Negative control tissue sections were incubated with fluorescent secondary antibody without primary antibody. We identified ERN1+ astrocytes by co-localization of GFAP (red) and intranuclear ERN1 (green; Fig 4A–D) under 1000X magnification with oil immersion using SPOT software (v5.1, Sterling Heights, MI). ERN1+ oligodendrocytes were identified by co-localization of olig2 (red) and intranuclear ERN1 (green; Fig 4E–H). We used brightfield microscopy with immunoperoxidase to count ERN1+ cells (rather than immunofluorescence) to discern the cellular morphology and count oligodendrocytes and astrocytes with an oil objective lens at 1000X for optimal resolution when visualizing the fine granules of immunoreactivity in the nucleus. The observer needed to fine focus in the Z-axis focal plane to fully appreciate the ERN1 immunoreactivity. Standard epifluorescence microscopy of immunofluorescence in brain sections under oil immersion at 1000X has lower resolution. We also evaluated the relationship between ERN1 immunoreactivity and apoptosis in the subcortical white matter. We previously reported white matter apoptotic profile counts using hematoxylin & eosin (H&E) stain and verified with terminal deoxynucleotidyl transferase dUTP nick end labeling in the piglets of the current study. The apoptosis occurred primarily among macroglia, including oligodendrocytes. [5] Apoptotic profiles were identified as cells with a few (<4) crescent-shaped or spherical clumps of chromatin, cytoplasmic condensation, cell shrinkage, eosinophilic cytoplasm, and an intact cytoplasmic membrane. [25] We used the H&E apoptotic profile counts from our prior study [5] to conduct paired comparisons of ERN1 immunoreactivity and apoptosis within pigs. Because apoptotic profiles are best visualized using our morphologic criteria by H&E stain [4, 5] and we evaluated ERN1 using immunohistochemistry with 3,3′-diaminobenzidine substrate, we could not directly correlate apoptotic profiles counts to ERN1+ macroglia counts in the same microscope fields. Apoptotic profiles were instead compared to ERN1+ cell counts in the same region of interest – the subcortical white matter of the motor gyrus – at the striatal and hippocampal anatomic levels. Immunoblotting We examined ERN1, protein kinase RNA-like endoplasmic reticulum kinase (PERK), and heat shock protein 70 (HSP70) levels by immunoblotting. We cut fresh piglet brains into 1-cm slabs and obtained white matter punch samples with a dermal micropunch. Briefly, frozen lysates from the subcortical white matter of the sensorimotor cortex were homogenized in RIPA lysis buffer (Cell Signaling Technology, Danvers, MA) plus phosphatase inhibitor (Roche Applied Science, Penzberg, Germany), protease inhibitor cocktail (Invitrogen, Grand Island, NY), and reducing agent (DTT 50 mM, Sigma Aldrich) with a weight-to-volume ratio of 0.05 g to 300 μL. We determined the supernatant protein concentration using the Pierce BCA protein assay kit (Thermo Scientific, Carlsbad, CA). Protein samples (40 μg) were treated with loading buffer 4X (LDS sample buffer, Invitrogen), separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis, and used for western blotting as described. [4, 5] Blots were probed with monoclonal mouse anti-ERN1 IgG (1:500, Abgent), monoclonal rabbit anti-PERK IgG (1:1000, Cell Signaling Technology), and monoclonal mouse anti-HSP70 IgG (1:1000, Abcam, Cambridge MA). Immunoreactivity was detected by enhanced chemiluminescence (Thermo Scientific, Carlsbad, CA) to quantify the optical density. Band intensities were quantified using ImageLab v. 5.0 software (Bio-Rad Laboratories) with the Coomassie Brilliant Blue stain (Bio-Rad Laboratories, Philadelphia, PA) serving as the protein loading control. We normalized the ERN1, HSP70, and PERK optical densities to that of the Coomassie Brilliant Blue stain for analysis. In some experiments, we loaded positive controls with recombinant ERN1 protein (OriGene, Rockville, MD) and 293T cell lysate (ProSci, Poway, CA; positive controls for PERK and HSP70 antibodies). We ran brain homogenate from one piglet per group (naïve, sham+normothermia, sham+hypothermia, sham+rewarming, HI+normothermia, HI+hypothermia, and HI+ rewarming) on a single gel in four separate experiments. Sample size calculations Because we did not have a priori knowledge about the magnitude of change and variability in ERN1+ cell ratios that would be induced by hypothermia and rewarming after HI, we conducted a sample size estimate after four experiments had been completed in each of the HI+hypothermia and HI+rewarming groups. Preliminary data of ERN1+ macroglial ratios indicated a difference in means of 2.4% between the groups with a within-group standard deviation of 1.1%. To reject the null hypothesis that the ratio of ERN1+ macroglia would differ by less than 2.4% between HI piglets that receive sustained hypothermia and those that receive hypothermia with rewarming at a power of 0.8 with an alpha level of 0.05, we calculated that we would need 4 piglets/group. We chose sample sizes of 6–8 piglets/group to permit some error in our initial estimates of variability. Statistical analysis The mean total macroglial cell count was calculated from the sums of ERN1+ and ERN1-negative macroglia in 6 random microscope fields. The mean ratio of ERN1+ macroglia/total macroglia from 6 microscope fields was used for analysis. Neurons were analyzed as the total neuronal count (sum of ERN1+ and ERN1-negative neurons) and the ratio of ERN1+ neurons/total neurons. Significance was assumed at p<0.05. To examine the effects of the anesthetic on the UPR, we used t-tests for parametric data and Mann Whitney rank sum tests for non-parametric data to compare the ratios of ERN1+ macroglia and neurons between naïve unanesthetized and normothermic sham piglets. To examine the effects of HI and temperature on the ERN1+ cell ratios and cell counts, we transformed non-parametric data with a log(x+1) function to generate parametric data when necessary. Then, a 2-way ANOVA was conducted with HI condition as factor 1 (sham or HI) and temperature as factor 2 (normothermia, hypothermia, or rewarming). We conducted post-hoc pairwise comparisons using the Holm Sidak method to examine the effects of HI injury when data were stratified by temperature and also to evaluate the effects of temperature when data were stratified by HI condition. We previously reported that hypothermia and rewarming promote macroglial apoptosis in the subcortical white matter of the motor gyrus in sham and HI piglets.[5] To evaluate the relationship between the UPR and apoptosis, we used Spearman correlations to conduct pairwise comparisons within pigs that had both ERN1+ ratios and apoptotic profile counts in macroglia [5] in the subcortical white matter of the motor gyrus. To analyze the immunoblot data, the densities of ERN1/Coomassie, PERK/Coomassie, and HSP70/Coomassie were normalized to naïve/Coomassie and then compared by Friedman two-way analysis of ranks with the four independent gels blocked as a between-subject factor. Finally, we generated a summary table (table 3) using results from the current study and our previously published work in white matter apoptosis [5]. We conducted a 2-way ANOVA with HI condition as factor 1 (sham or HI) and temperature as factor 2 (normothermia, hypothermia, or rewarming). We then used post-hoc pairwise comparisons using the Holm Sidak method to compare the ERN1+ macroglia ratios, apoptotic profile counts from our prior study [5], and total macroglial cell counts in the subcortical white matter of the motor gyrus with normothermia as the control. Hypothermic or rewarmed sham piglets were compared to normothermic shams, and hypothermic or rewarmed HI piglets were compared to the normothermic HI piglets. RESULTS Mortality Forty-four piglets underwent HI injury. One HI+normothermia piglet could not be resuscitated after asphyxia, and three HI+hypothermia piglets died from hypotension during anesthesia. Thus, 91% of the HI piglets completed the protocol with 29 hours of recovery under anesthesia. Forty-three piglets had sham surgery. One sham+normothermia and two sham+rewarming piglets died from hypotension during anesthesia. Thirteen naïve piglets that did not receive anesthesia or surgery were prepared as additional controls. In total, 93 piglets including those from our prior two studies [4, 5] were used in the final analysis. (Fig 1) Physiology We previously reported physiologic and blood gas data in the piglets, including temperature, blood pressure, oxyhemoglobin saturation, pH, PaCO2, hemoglobin, and electrolyte levels, in our previous reports on cortical neuron [4] and white matter apoptosis. [5] In addition, mean glucose levels were 186 mg/dL (SD: 95) in rewarmed and 148 mg/dL (SD: 71) in hypothermic shams. We summarize select parameters during the HI protocol here to describe the physiologic severity of the HI injury. During hypoxia, the mean oxyhemoglobin saturations were 25% (SD: 7) in HI+normothermia (n=6), 30% (SD: 9) in HI+hypothermia (n=8), and 29% (SD: 7) in HI+rewarming (n=8) groups. The oxyhemoglobin saturations decreased during asphyxia to 7% (SD: 11) in HI+normothermia, 7% (SD: 3) in HI+hypothermia, and 4% (SD: 3) in HI+rewarming piglets. Hypothermic piglets had core temperatures of approximately 34°C and were successfully rewarmed to normothermia without exceeding the goal temperature. Anesthetic minimally affects the UPR We first evaluated the effect of the anesthetic on the UPR. The ratios of ERN1+ macroglia and ERN1+ neurons were similar between naïve unanesthetized piglets and anesthetized normothermic sham piglets at the striatal (p=0.694 for macroglia; p=0.867 for neurons) and hippocampal levels (p=0.452 for macroglia; p=0.950 for neurons; Fig 5). Therefore, the anesthetic regimen and surgical preparation had minimal effect on the UPR. The macroglial UPR is affected by HI, hypothermia, and rewarming and is associated with apoptosis HI (p<0.001) and temperature (p=0.046) independently and interactively (p<0.001) affected the ERN1+ macroglia ratio at the striatal level. The ERN1+ cells in the subcortical white matter were immunophenotyped as oligodendrocytes and astrocytes (Fig 4). When data were stratified by HI injury, normothermic (p=0.015) and rewarmed (p=0.014) HI piglets had greater ERN1+ macroglia ratios than did HI piglets that remained hypothermic (Fig 6A). Among shams, sustained hypothermia (p=0.011) and rewarming (p=0.004) each increased the macroglial ERN1+ ratio above that in normothermia. When data were stratified by temperature, hypothermic shams had greater ERN1+ macroglial ratios than did hypothermic HI piglets (p<0.001). Rewarmed shams also had more ERN1+ macroglia than did rewarmed HI piglets (p=0.004). Absolute cell counts and ERN1+ cell ratios are presented in Tables 1 and 2. HI (p=0.002), but not temperature (p=0.349), affected the total macroglial cell counts (ERN1+ plus ERN1-negative) with an interactive effect (p<0.001) in white matter at the striatal level (Fig 6B). When data were stratified by injury, HI piglets that remained normothermic had fewer macroglial cells than did HI piglets that were rewarmed (p=0.018). Normothermic shams had more macroglia than did hypothermic (p=0.009) and rewarmed (p=0.007) shams. Among the hypothermic groups, HI piglets had more macroglia than did shams (p=0.011). Similarly, HI rewarmed piglets had more macroglia than sham rewarmed piglets (p<0.001). At the posterior hippocampal anatomic level, only temperature affected the ERN1+ macroglia ratio (p=0.009); HI did not have an effect (p=0.244, Fig 6C). When we controlled for HI, rewarming increased the ERN1+ macroglia ratio more than normothermia (p=0.007). In comparisons stratified by injury, rewarmed shams had more ERN1+ macroglia than did normothermic shams (p=0.024). One rewarmed sham piglet did not have ERN1 macroglia counts because the subcortical white matter had tissue damage artifact at the hippocampal anatomic level. (The cortical tissue was not damaged and neurons were counted in this piglet.) The number of macroglia in white matter was not affected by HI (p=0.516) or temperature (p=0.392; Fig 6D) at the hippocampal level. We previously reported that hypothermia and rewarming promote apoptosis in the subcortical white matter of the motor gyrus and other white matter regions in sham and HI piglets. [5] Among piglets that had both ERN1+ macroglia and apoptotic profile counts at the striatal level, [5] the ratio of ERN1+ macroglia correlated to the number of apoptotic profiles in the motor subcortical white matter in shams that received normothermia, hypothermia, or rewarming (p=0.026; Fig 7A). ERN1+ and apoptotic macroglial counts were not correlated in HI piglets after normothermia, hypothermia, or rewarming (p=0.956; Fig 7B). At the hippocampal level, the ratio of ERN1+ macroglia did not correlate with the number of apoptotic profiles in the subcortical white matter of sham (n=21; r=0.109; p=0.633) or HI piglets (n=22; r=0.269; p=0.222). HI, hypothermia, and rewarming decrease neuron counts without affecting the neuronal UPR Absolute cell counts and ratios for ERN1+ neurons are presented in Tables 1 and 2. Neither HI (p=0.104) nor temperature (p=0.738) affected the ratio of ERN1+ neurons in the motor cortex at the striatal level (Fig 8A). Both HI (p=0.019) and temperature (p=0.002) affected the number of neurons without an interaction (p=0.120; Fig 8B). When we controlled for HI, rewarming decreased the number of neurons relative to normothermia (p=0.001). Post-hoc comparisons showed that rewarmed shams had fewer neurons than did normothermic (p<0.001) and hypothermic (p=0.045) shams. Moreover, rewarmed shams had lower neuronal cell counts than did rewarmed HI piglets (p=0.003). At the hippocampal level, HI (p=0.251) and temperature (p=0.070) did not influence the ratio of ERN1+ neurons in the motor cortex (Fig 8C). Temperature (p=0.005), but not HI (p=0.798), affected the neuronal count (Fig 8D) without an interaction (p=0.703). When we controlled for HI, both hypothermia (p=0.006) and rewarming (p=0.022) decreased the number of neurons below that observed during normothermia. In post-hoc comparisons, normothermic HI piglets had more neurons than did hypothermic HI piglets (p=0.043). Neuron counts could not be completed in one HI+normothermia and three HI+rewarming piglets because of cortical damage to the mounted tissue; the entire motor cortex must be intact to quantify ERN1+ and ERN1-negative neurons. (These pigs had intact subcortical white matter, however, and thus had macroglial cell counts.) Immunoblotting showed similar levels of ERN1, PERK, and HSP70 levels among groups Immunoreactivity on Western blots for ERN1 (p=0.345), PERK (p=0.297), and HSP70 (p=0.836) was similar among normothermic sham, hypothermic sham, rewarmed sham, normothermic HI, hypothermic HI, and rewarmed HI piglets (Fig 9). DISCUSSION This study identifies potential deleterious, cell type-specific effects from delayed hypothermia and rewarming on the newborn brain. We determined in a neonatal piglet model that hypothermia and rewarming cause macroglial ER stress that is associated with white matter apoptosis independent of HI. These findings have critical and clinically relevant implications for the use of hypothermia in the developing brain. First, temperature and HI affected the UPR in oligodendrocytes and astrocytes both independently and interactively. Both sustained hypothermia and hypothermia with rewarming increased the ERN1+ macroglia ratio above that of normothermia in shams, whereas HI prior to hypothermia diminished the macroglial UPR. Second, macroglial UPR activation correlated with white matter apoptosis as measured in our prior study [5] in shams, thereby associating UPR activation with white matter apoptosis after hypothermia and rewarming. This apoptosis was robust and decreased oligodendrocyte and astrocyte cell counts in shams. However, the associations between macroglial UPR activation and apoptosis with hypothermia and rewarming were not observed after HI. Third, hypothermia and rewarming induced cortical neuron loss in the motor gyrus, but this change was not related to ERN1 upregulation. Pre-hypothermia HI blocked the macroglial cell loss after hypothermia and rewarming but did not influence the cortical neuron sensitivity to hypothermia and rewarming. Altogether, these findings lead us to conclude that in the neonatal brain: 1) hypothermia and rewarming activate the UPR in oligodendrocytes and astrocytes and promote white matter apoptosis independent of HI; 2) hypothermia and rewarming induce cortical neuron loss through mechanisms not clearly associated with the UPR; and 3) HI may interrupt the macroglial ER stress response and subsequent cell death with promotion of apoptosis by mechanisms separate from the hypothermia-induced UPR. (Table 3) The potential cytotoxic effects of hypothermia and rewarming on oligodendrocytes, astrocytes, and neurons in uninjured or minimally injured developing brain deserve further study given increasing interest in treating mild HIE with hypothermia and for designing clinical studies. [26] Therapeutic hypothermia initiated during the latent phase soon after HI [27] is neuroprotective in part by attenuating secondary energy failure. [28] Disturbed cerebral metabolism with mitochondrial dysfunction occurs early after HI [29] with secondary energy failure at approximately 24 hours in a piglet HI model of hypoxia with carotid occlusion; hypothermia for 12 hours mitigates the energy failure through 60 hours. [30] In our piglet HI model of whole-body hypoxia-asphyxia, mitochdondrial failure occurs by 12 hours with normothermia. [31] We therefore expect that the 18-29 hours of hypothermia in our protocols targeted the latent phase after HI prior to secondary deterioration. [27] We did not extend hypothermia to 72 hours, the duration used in the clinical treatment of HIE, due to neurotoxic effects of the anesthetic (unpublished data) and limited resources. Hypothermia confers neuroprotection in our piglet HI model by preserving viable neurons and reducing NMDA receptor activation, nitric oxide synthase relocalization, oxidative stress, and cellular necrosis. [32-35] However, hypothermia may promote apoptosis in cortex after HI [4] and in white matter with or without HI [5], which may reflect an off-target effect of hypothermia and rewarming. Based on the similar ERN1+ macroglia and neuron ratios between sham normothermic and naïve unanesthetized piglets, our 29 hour anesthetic minimally affected the UPR in brain. We hypothesized that HI would induce ER stress with UPR activation, that post-HI hypothermia would suppress the UPR, and that cellular stress from rewarming would reactivate the UPR. We formulated these hypotheses based on our prior observations in HI piglets [31] and assumptions that hypothermia would decrease protein translation and rewarming would reinitiate translation. [36] Morphological and biochemical evidence for ER stress has been identified in piglet striatum using a variation of the model used here. [31] Electron microscopy has shown ER damage with 3-6 h after HI and prominent release of ER luminal protein by 3 h after HI. [31] Biochemical and morphological evidence for ER stress in brain has been found in neonatal rodent HI models with 3 hours of recovery. [12] ER stress from HI activates the protein kinase activity of PERK, an ER-resident protein that phosphorylates the eukaryotic initiation factor-2 α-subunit (eIF2-α), to slow translation of mRNA to protein [12, 37] and activate the UPR. [38] Ischemia also impairs heat shock protein translation in immature brain [39]: heat shock proteins facilitate de novo protein folding and refolding of misfolded proteins. Impaired heat shock protein function after HI would cause protein misfolding and compound the ER stress. These perturbations could influence neural cell viability. A recent study of HI in neonatal mice found an association between ER stress and a newly described form of cortical neuron degeneration called macrozeiosis. [40] This neurodegeneration was morphologically distinct from necrosis and classical apoptosis and had a characteristic signature typified by cytoplasmic shedding. [40] In our piglet study, hypothermia and rewarming induced cortical neuron loss. While the mechanism for this neuronal elimination is not known, it was distinct from the ER stress-associated induction of apoptosis in white matter macroglia. This study starkly emphasizes the need for histological verification of the cell types in which the UPR is occurring. We designed our study to examine the independent effects of temperature and HI and the interaction between the two factors. Our experimental design included normothermia with or without prior HI; sustained hypothermia with or without prior HI; and hypothermia followed by rewarming with or without prior HI. Each group provided relevant information about the UPR and cell death. Piglets that remained hypothermic after HI had lower ERN1+ macroglia ratios than did HI normothermic piglets. Our histologic measurements focused on oligodendrocytes and astrocytes. This finding suggests that post-HI hypothermia suppressed the UPR through 29 hours of recovery. The ability of hypothermia to mitigate ER stress in piglet brain and presumably protein misfolding after HI may be related to decreased cell metabolism, protein translation, [41] and oxidative stress [42] and modulation of the inflammatory response. [43] HI piglets treated with hypothermia have diminished NMDA receptor activation, nitric oxide synthase activation, and oxidized protein burden. [32] In contrast to our observations after HI, hypothermia increased the ERN1+ macroglia ratio above that of normothermia in sham piglets. Thus, hypothermia itself appears to induce ER stress with UPR activation in oligodendrocytes and astrocytes, and prior HI may interrupt the hypothermia-induced ER stress. It was previously shown that hypoxic preconditioning decreases caspase-12 activation in hippocampal slice cultures exposed to oxygen-glucose deprivation [44] and that caspase-12 is linked to neuronal cell death in rodents with ER stress after ischemia. [13] However, the role of capsase-12 in human neurodegeneration is uncertain because this gene has acquired polymorphisms during evolution that cause it to encode for a truncated protein without caspase activity. [45] Rewarming also promoted macroglial UPR activation. Rewarmed HI piglets had higher ERN1+ macroglia ratios than did HI piglets that remained hypothermic, and rewarming activated the UPR above that observed during normothermia in shams. Therefore, the ER stress response could provide a potential target for adjuvant therapies to reduce the risk of neurologic injury from rewarming in HIE. [46-48] Because all piglets were studied at 29 hours of recovery after HI or sham surgery, the rewarmed groups received shorter duration of hypothermia than those that remained hypothermic. The observed UPR activation in rewarmed HI piglets could also reflect shorter duration of hypothermia. Nonetheless, other studies corroborate the potential cytotoxic effects of rewarming. For example, faster rewarming rates from hypothermic cardiopulmonary bypass are also associated with stroke and higher plasma glial fibrillary acidic protein levels, a marker of astrocyte injury. [49] Rewarming reinitiates protein translation through dephosphorylation of eIF2-α in mammalian cells recovering from cold shock. [36] Upregulated protein translation with the accumulation of misfolded proteins in addition to cellular stressors from hypothermia with rewarming, such as inflammation [50] and oxidative stress, [35, 51, 52] may activate the UPR. [53, 54] Furthermore, the ratio of ERN1+ macroglia correlated with apoptosis in the subcortical white matter of sham piglets. [5] Congruent with our findings that hypothermia and rewarming activate the UPR, both sustained hypothermia and hypothermia with rewarming reduced the macroglial cell count below that of normothermia in shams. The increasing interest in therapeutic hypothermia for mild HIE and the routine use of hypothermic cardiopulmonary bypass for neonatal and pediatric cardiac surgery warrants investigation into the off-targets effects of hypothermia and rewarming. Hypothermia and rewarming may have deleterious effects in the uninjured or minimally injured developing white matter, including the UPR activation, apoptosis, and oligodendrocyte and astrocyte loss observed in our sham piglets. HI prior to hypothermia prevented macroglial cell loss from sustained hypothermia and hypothermia with rewarming. Perhaps HI enabled macroglia to tolerate the proteomic stress of hypothermia, or HI interrupted or impaired the mechanisms required to fully execute the UPR with apoptosis. For example in HI piglet striatum, the ER shows evidence for pathological damage by 3 h [31] so the UPR might not be operative. Hypothermia and rewarming without HI, however, resulted in a robust UPR with apoptosis in white matter macroglia. The relationships between HI, hypothermia, rewarming, and ERN-related UPR were limited to oligodendrocytes and astrocytes. The ratio of ERN1+ neurons was not associated with HI or temperature at 29 hours of recovery, thereby demonstrating differential activation of the UPR among specific cell types and anatomic regions. The UPR's importance in neurologic injury after HI may be focused on white matter injury and might serve as a target for adjuvant therapies designed to decrease white matter injury in HIE. White matter injury remains a prominent component of neurologic injury in HIE despite the use of therapeutic hypothermia, [46, 47] and white matter injuries predict neurodevelopmental outcomes in survivors of HIE. [7] The difference in cell-specific UPR activation may be due to high protein synthesis in myelinating glia [55] that could impart vulnerability to temperature-related ER stress. The UPR is associated with apoptosis in myelinating glia in multiple sclerosis and Charcot-Marie-Tooth. [16, 17] Astrocytes are also highly metabolically active cells that are activated by hypothermia and oxygen-glucose deprivation. [56] They are involved in neurotransmitter and ion clearance from the interstitial space, [57, 58] neuroinflammation, [59] brain development, and cell differentiation, proliferation, and plasticity. [60] It is also possible that the neuronal UPR subsided by 29 hours of recovery. Sustained hypothermia and hypothermia with rewarming decreased the number of neurons in the motor cortex below that of normothermia. The cytotoxic effects of hypothermia and rewarming in neurons were independent of HI and not related to robust ERN1-related UPR activation within neurons. This again highlights the potential deleterious effects of hypothermia and rewarming to the developing brain in the absence of moderate or severe HI and in uninjured brain regions remote from areas of moderate or severe injury. The risks of cooling neonates for mild HIE must be further studied. Hypothermia after HI also reduced the number of cortical neurons relative to that observed with normothermia after HI. We observed intracortical gyrus differences in the magnitude of cortical neuron loss and white matter changes at striatal and hippocampal levels, but these differences could be due to differences in cortical connectivity that can be found even within the same gyrus. [61] Additional experiments are necessary to identify the mechanisms responsible for this rapid cortical neuron loss. Our Western blotting experiments did not show evidence for UPR activation in white matter extracts, but our histologic data revealed strong evidence for UPR activation. Western blotting may be limited by the use of a crude homogenate in which sources of protein include all glial cells, neurons, axons, and vascular and blood-borne cells. In contrast, histologic evaluation permits high resolution identification and precise counting of macroglia with intranuclear ERN1+ foci. It is also possible that the magnitudes of between-group differences in ERN1, PERK, and HSP70 protein expression were too small to be detected in a crude homogenate assay. We previously demonstrated increased levels of cortical cleaved caspase-3 after hypothermia and rewarming following HI in our model. [4] Our study had some limitations. Although we conducted a power analysis to determine our sample sizes, the number of piglets per experimental group was small, and we used only males. The animals remained under general anesthesia for 29 hours for comfort, a practice that is not used clinically. Nonetheless, the fact that the ratios of ERN1+ macroglia and neurons were similar between unanesthetized naïve and anesthetized sham normothermic piglets indicated that the anesthetic minimally affected the UPR. Another limitation is that we examined only one time point and only one histologic marker for UPR after HI or sham surgery; the neuronal UPR response may have subsided by 29 hours, and we cannot define a timeline for macroglial UPR activation during hypothermia and rewarming. Moreover, clinical hypothermia for HIE is provided for more than 24 hours. Therefore, the UPR needs to be examined during longer durations of hypothermia to define the UPR's potential as a clinical biomarker or target for adjuvant therapies. Extending hypothermia beyond 72 hours, however, may not be beneficial and could affect neuroinflammation [62] and exacerbate cortical neuron loss. Although we cannot determine whether white matter macroglial UPR activation will result in long-term neurologic consequences, in clinical MRI studies white matter injury is evident 2 weeks after therapeutic hypothermia,[46, 47] and is associated with neurodevelopmental impairments at 2 years of age.[63] The possible effect of HI in disengaging the UPR during hypothermia and rewarming is difficult to apply clinically, and different degrees of hypoxia may affect subsequent responses to cellular stresses differently. HI did not activate the UPR above that observed after sham surgery during 29 hours normothermic recovery. The UPR may have subsided at this relatively late time point as other studies demonstrate UPR activation in rodents 3 hours after HI. [12] Microglia may also be involved in the UPR, and we did not evaluate ERN1 upregulation in microglia. Glucose was not well controlled, and hyperglycemia can contribute to UPR activation. [64] CONCLUSION We identified that hypothermia and rewarming induce ER stress with UPR activation in oligodendrocytes and astrocytes of the developing brain. Our finding that macroglial UPR activation was associated with apoptosis and a decrease in macroglial cell count in shams demonstrated that hypothermia and rewarming are neurotoxic independent of HI. Hypothermia and rewarming also caused cortical neuron loss independent of HI and unassociated with UPR activation. Pre-hypothermia HI interrupts the UPR during hypothermia and rewarming. Modulating the UPR in macroglial cells and the maladaptive responses of cortical neurons during hypothermia and rewarming could improve neuroprotection in the white matter and cerebral cortex during clinical situations that warrant the therapeutic use of hypothermia. Acknowledgements Dr. Lee was supported by grants from the National Institutes of Health (NIH; K08 NS080984 [NINDS]; R21HD072845 [NICHD]), a Johns Hopkins University Clinician Scientist Award, and an American Heart Association Grant-in-Aid. Drs. Koehler and Martin were supported by a grant from the NIH (R01 NS060703). Dr. Martin was also supported by a grant from NIH NINDS (R01 NS079348). Dr. Lee has research funding from Medtronic for a clinical study. We are grateful to Claire Levine, MS, ELS, for her editorial assistance and to Dawn Spicer for her assistance with the immunohistochemistry. Abbreviations DAPI 4',6-diamidino-2-phenylindole ER endoplasmic reticulum ERN1 endoplasmic reticulum to nucleus signaling-1 protein GFAP macroglial fibrillary acidic protein HSP70 heat shock protein 70 HI hypoxia-ischemia Iba-1 ionized calcium-binding adaptor molecule 1 IQR interquartile range Olig2 oligodendrocyte transcription factor lineage protein-2 PERK protein kinase RNA-like endoplasmic reticulum kinase SD standard deviation UPR unfolded protein response Fig 1 Study design with randomization of piglets to hypoxic-ischemic injury or sham surgery as well as to normothermia (normoT), delayed hypothermia (hypoT), or hypothermia with rewarming. The number of piglets per group are listed. Piglet groups reported in our prior studies are denoted [4] a and [5] b. Fig 2 Representative macrophotographs showing the anatomic regions of interest at the striatal (A) and hippocampal (B) anatomic levels. Neuronal and macroglial profiles were counted in the cortex (i) and subcortical white matter (ii) of the motor gyrus. Photographs were taken at 10X magnification; grayscale images of hematoxylin & eosin stained sections are shown. Fig 3 Representative images show immunohistochemistry for endoplasmic reticulum to nucleus signaling 1 (ERN1) with 3,3’-diaminobenzidine substrate and cresyl violet (A, B, D); hematoxylin & eosin stain of neurons and macroglia (C); and immunohistochemistry for Iba-1-positive microglia with 3,3’-diaminobenzidine substrate and cresyl violet (panel D inset). (A) The subcortical white matter of the motor gyrus in a hypothermic sham piglet is shown. The arrows denote examples of ERN1+ macroglia, and the arrowheads identify examples of ERN1-negative macroglia. (B) Examples of ERN1+ neurons (arrows) and ERN1-negative neurons (arrowheads) in the motor cortex of a hypothermic sham piglet. (C) Representative images of neurons and macroglia in the motor cortex of a normothermic sham piglet illustrate the differential morphology between cell types. The arrows identify examples of neurons, distinguished by the multipolar or triangular neuronal morphology, large cell body (characteristically 8-20 μm in diameter), and open non-condensed nucleus with chromatin strands and often a nucleolus. The arrowheads show examples of the smaller macroglial cells, which include astrocytes and oligodendrocytes. Some macroglial cells are located in “satellite” positions next to the larger neurons (panel inset). (D) Representative images of microglia, macroglia, and neurons in the subcortical white matter of the motor gyrus in a sham hypothermic piglet. The arrow identifies a microglial cell, distinguished by the elongated nucleus with dark chromatin. Microglia were not included in the ERN1 macroglia counts. Arrowheads denote macroglia that were counted for ERN1 immunoreactivity. The double arrowheads indicate a neuron. The inset panel shows immunohistochemistry with ionized calcium-binding adaptor molecule 1 antibody (marker for microglia), 3,3’-diaminobenzidine substrate, and cresyl violet. The arrow shows a representative microglial cell with nuclear morphology that can be differentiated from macroglia (arrowheads). Photos for panels A and B were taken at 1000X with oil immersion; the scale bars are 20 μm. The photo for panel C was taken at 400X magnification, and the scale bar is 50 μm. The inset photo for panel C was taken at 600X magnification with oil immersion, and the scale bar is 10 μm. The photos for panel D was taken at 1000X with oil immersion; the scale bar is 10 μm. Fig. 4 Immunofluorescent co-staining in the subcortical white matter of the motor gyrus. Panels A–D show glial fibrillary acidic protein (GFAP; red), endoplasmic reticulum to nucleus signaling 1 (ERN1; green) and 4',6-diamidino-2-phenylindole (DAPI; blue) in a hypothermic sham piglet. Panels E–H show co-staining for oligodendrocyte transcription factor (Olig2; red), ERN1 (green), and DAPI (blue) in a rewarmed sham piglet. (D) The single arrows denote GFAP-positive (red) astrocytes with intranuclear (blue) ERN1 (green) staining in this merged image. The double arrowheads identify a neuron that is distinguished by its large size and multipolar morphology. The double arrows show a GFAP-negative macroglial cell with intranuclear ERN1 staining. We therefore conducted additional immunostaining to identify the ERN1+, GFAP-negative cells. (H) The arrows identify olig2+ (red) oligodendrocytes with intranuclear (blue) ERN1 (green) staining in this merged image. The single arrowhead denotes an olig2+ oligodendrocyte that is ERN1-negative. The double arrowheads show an ERN1+ macroglial cell that is olig2-negative. Photos were taken at 1000X with oil immersion; the scale bars are 20 μm. Fig 5 The anesthetic had minimal effect on the unfolded protein response (UPR). The ratios of ERN1+ macroglia were similar in naïve unanesthetized and sham normothermic piglets at the (A) striatal (p=0.694) and (B) hippocampal levels (p=0.452). The ratios of ERN1+ neurons were also similar in naïve and sham normothermic piglets at the (C) striatal (p=0.867) and (D) hippocampal levels (p=0.950). Data were analyzed by t-tests or Mann Whitney rank sum tests. Box plots with whiskers (5th-95th percentiles) are shown. Each circle represents one piglet. Fig 6 ERN1+ macroglia and total macroglial cell counts in the subcortical white matter of the motor gyrus at the striatal (A, B) and hippocampal (C, D) levels. (A) Both hypoxia-ischemia (HI; p<0.001) and temperature (p=0.046) independently affected the ratio of ERN1+ macroglia, and there was an interaction between HI and temperature (p<0.001). In post-hoc pairwise comparisons, ERN1 expression differed among piglets that received normothermia (normoT), sustained hypothermia (hypoT), or rewarming (rewarm): *p<0.02 vs. HI+hypoT; **p<0.02 vs. sham+normoT; †p<0.001 vs. HI+hypoT; ††p=0.004 vs. HI+rewarm. (B) HI (p=0.002), but not temperature (p=0.349), affected the macroglial cell count. There was an interactive effect between HI and temperature on the number of macroglia (p<0.001). Post-hoc pairwise comparisons showed differences in macroglial counts: *p=0.018 vs. HI+normoT; **p<0.010 vs. sham+normoT; †p=0.011 vs. HI+hypoT; ††p<0.001 vs. HI+rewarm. (C) Temperature (p=0.009), but not hypoxia-ischemia (HI; p=0.244), affected the ratio of ERN1+ macroglia. When we controlled for HI, rewarming increased the ratio of ERN1+ macroglia above that during normothermia (normoT; p=0.007). In post-hoc pairwise comparisons, rewarmed shams had more ERN1+ macroglia than did normothermic shams (*p=0.024 vs. sham+normoT). (D) The macroglial cell count was not affected by HI (p=0.516) or temperature (p=0.392). Data were analyzed by 2-way ANOVA with Holm Sidak post-hoc pairwise comparisons. Bar graphs show means and SD. Each circle represents one pig. Fig 7 Correlation between ERN1+ macroglia and apoptotic profiles in the subcortical white matter. We previously reported the macroglial apoptotic profile counts of the pigs in the current study. [5] (A) The ratio of ERN1+ macroglia correlated with the number of apoptotic profiles in sham (A) but not HI (B) piglets that received normothermia, hypothermia, or rewarming. Data were analyzed by Spearman correlations. Each circle represents one piglet. Fig 8 ERN1+ neurons and neuronal cell counts in the motor cortex at the striatal (A, B) and hippocampal (C, D) levels. (A) HI (p=0.104) and temperature (p=0.738) did not affect the ratio of ERN1+ neurons. (B) Both HI (p=0.019) and temperature (p=0.002) independently affected the neuron count. When we controlled for HI, rewarming (rewarm) decreased the number of neurons below that observed during normothermia (normoT; p<0.001). Post-hoc pairwise comparisons showed differences in neuronal counts: *p<0.05 vs. sham+rewarm; **p=0.003 vs. HI+rewarm. (C) Neither HI (p=0.251) nor temperature (p=0.070) affected the ratio of ERN1+ neurons. (D) Temperature (p=0.005), but not HI (p=0.798), affected the number of neurons. When we controlled for HI, both sustained hypothermia (hypoT) and hypothermia with rewarming decreased the neuron count when compared to that with normothermia (p=0.006 for hypoT vs. normoT; p=0.022 for rewarm vs. normoT). In post-hoc pairwise comparisons, hypothermia decreased the number of neurons compared to that with normothermia in HI piglets (*p=0.043 vs. HI+normoT). Data were analyzed by 2-way ANOVA with Holm Sidak post-hoc pairwise comparisons. Bar graphs show means and SD. Each circle represents one pig. Fig 9 Immunoreactivity of PERK, HSP70, and ERN1 in the subcortical white matter of the sensorimotor cortex. (A) Representative Western blots of naïve, sham+normothermia (NormoT), sham+hypothermia (HypoT), sham+rewarming (RW), hypoxia-ischemia (HI)+NormoT, HI+HypoT, and HI+RW piglets. (B–D) The relative densities normalized to naïve piglets were similar between groups for PERK (p=0.297), HSP70 (p=0.836), and ERN1 (p=0.345). Data were analyzed by a Friedman two-way analysis of ranks in which the four independent gels were blocked as a between-subject factor. Box plots with 5th-95th percentiles are shown. Each circle represents one piglet. Table 1 Absolute counts of endoplasmic reticulum to nucleus signaling 1-positive (ERN1+) and total cells. Glial counts are the average of six microscope fields under 1000X magnification with oil immersion. Neuron counts are from one side of the entire motor gyrus.* Anatomic location Normothermia Hypothermia Hypothermia + rewarming Sham HI Sham HI Sham HI Glia: striatal level ERN1 + 4 (2, 6) 7 (5, 7) 16 (11, 22) 1 (0, 1) 26 (10, 28) 4 (3, 8) Total 85 (79, 96) 74 (73, 80) 69 (52, 75) 84 (80, 86) 63 (57, 71) 104 (91, 108) Glia: hippocampal level ERN1+ 8 (4, 22) 9 (3, 18) 21 (13, 23) 13 (3, 24) 21 (14, 33) 26 (16, 33) Total 96 (86, 104) 69 (68, 86) 69 (57, 77) 79 (72, 91) 66 (64, 72) 83 (73, 91) Neurons: striatal level ERN1+ 267 (86, 422) 321 (65, 733) 528 (375, 819) 88 (66, 266) 459 (149, 554) 164 (46, 455) Total 2732 (2388, 2973) 2780 (2440, 3049) 2000 (1956, 2138) 2417 (2163, 2573) 1439 (1274, 1498) 2395 (1553, 3150) Neurons: hippocampal level ERN1+ 257 (44, 441) 301 (267, 405) 701 (505, 939) 121 (85, 424) 251 (84, 284) 289 (92, 316) Total 2273 (1816, 2871) 2387 (2098, 2458) 1918 (1680, 2023) 1517 (1368, 1878) 1847 (1576, 1978) 1861 (1783, 2037) Abbreviations: HI, hypoxia-ischemia; IQR, interquartile range. * Data are shown as medians (IQR). Table 2 Endoplasmic reticulum to nucleus signaling 1-positive (ERN1+) ratios in glia and neurons* Anatomic location Normothermia Hypothermia Hypothermia + rewarming Sham HI Sham HI Sham HI Glia: striatal level 0.05 (0.03, 0.08) 0.09 (0.06, 0.11) 0.23 (0.17, 0.31) 0.01 (0.003, 0.02) 0.35 (0.17, 0.51) 0.04 (0.03, 0.10) Glia: hippocampal level 0.08 (0.04, 0.22) 0.09 (0.03, 0.18) 0.21 (0.13, 0.23) 0.13 (0.03, 0.24) 0.21 (0.14, 0.33) 0.26 (0.16, 0.33) Neurons: striatal level 0.09 (0.03, 0.19) 0.14 (0.03, 0.29) 0.24 (0.16, 0.40) 0.03 (0.02, 0.13) 0.39 (0.11, 0.46) 0.08 (0.02, 0.25) Neurons: hippocampal level 0.11 (0.02, 0.28) 0.13 (0.12, 0.23) 0.36 (0.31, 0.56) 0.06 (0.05, 0.33) 0.12 (0.06, 0.15) 0.16 (0.05, 0.16) Abbreviations: HI, hypoxia-ischemia; IQR, interquartile range. * Data are shown as the median ratio of ERN1+ cells (ERN1+/total cells; IQR). Table 3 Summary of treatment effects on the unfolded protein response, apoptosis, and glial cell loss in the subcortical white matter 29 hours after hypoxia-ischemia or sham surgery.* Treatment ERN1+ glia ratio in white matter Apoptosis in white matter glia White matter total glial cell count Striatal anatomic level     HI + hypothermia − + NS     HI + rewarming NS + +     Hypothermia alone + + −     Rewarming alone + + − Hippocampal anatomic level     HI + hypothermia NS NS NS     HI + rewarming NS NS NS     Hypothermia alone + + NS     Rewarming alone + + NS * This summary is extrapolated from the current study and our past study on white matter apoptosis (Wang and Armstrong, et al., 2016). The rewarming rate is 0.5°C/hour. Data were analyzed by two-way ANOVA, and the results of post-hoc pairwise comparisons using the Holm Sidak method are shown. Hypothermic and rewarmed sham groups were compared to the normothermic sham group, and hypothermic and rewarmed HI groups were compared to the normothermic HI group. Abbreviations: HI, hypoxia-ischemia; +, increase in parameter compared to the normothermic group at p<0.05; −, decrease in parameter compared to the normothermic group at p<0.05; NS, no significant difference from the normothermic group. None of the authors has a conflict of interest associated with this research. References 1 Azzopardi D Strohm B Marlow N Effects of hypothermia for perinatal asphyxia on childhood outcomes. N Engl J Med 2014 371 140 149 25006720 2 Shankaran S Pappas A McDonald SA Childhood outcomes after hypothermia for neonatal encephalopathy. N Engl J Med 2012 366 2085 2092 22646631 3 Shankaran S Laptook AR Ehrenkranz RA Whole-body hypothermia for neonates with hypoxic-ischemic encephalopathy. N Engl J Med 2005 353 1574 1584 16221780 4 Wang B Armstrong JS Lee JH Rewarming from therapeutic hypothermia induces cortical neuron apoptosis in a swine model of neonatal hypoxic-ischemic encephalopathy. J Cereb Blood Flow Metab 2015 5 Wang B Armstrong JS Reyes M White matter apoptosis is increased by delayed hypothermia and rewarming in a neonatal piglet model of hypoxic ischemic encephalopathy. Neuroscience 2016 In press 6 van Schie PE Schijns J Becher JG Barkhof F van Weissenbruch MM Vermeulen RJ Long-term motor and behavioral outcome after perinatal hypoxic-ischemic encephalopathy. Eur J Paediatr Neurol 2015 19 354 359 25683783 7 Cavalleri F Lugli L Pugliese M Prognostic value of diffusion-weighted imaging summation scores or apparent diffusion coefficient maps in newborns with hypoxic-ischemic encephalopathy. Pediatr Radiol 2014 44 1141 1154 24715056 8 Sun H Tang Y Li L Guan X Wang D Effects of local hypothermia on neuronal cell apoptosis after intracerebral hemorrhage in rats. J Nutr Health Aging 2015 19 291 298 25732214 9 Zhu X Zelmer A Kapfhammer JP Wellmann S Cold-inducible RBM3 inhibits PERK phosphorylation through cooperation with NF90 to protect cells from endoplasmic reticulum stress. FASEB J 2015 10 Prieto Robles C Argibay Lago A Hernanz Del Rio A Rodriguez Castano IM Fernandez-Rodriguez D Optimization of glucose metabolism in patients undergoing therapeutic hypothermia after sudden cardiac arrest. Rev Enferm 2015 38 46 50 11 Yu A Li P Tang T Wang J Chen Y Liu L Roles of Hsp70s in stress responses of microorganisms, plants, and animals. Biomed Res Int 2015 2015 510319 26649306 12 Badiola N Penas C Minano-Molina A Induction of ER stress in response to oxygen-glucose deprivation of cortical cultures involves the activation of the PERK and IRE-1 pathways and of caspase-12. Cell Death Dis 2011 2 e149 21525936 13 Nakka VP Gusain A Raghubir R Endoplasmic reticulum stress plays critical role in brain damage after cerebral ischemia/reperfusion in rats. Neurotox Res 2010 17 189 202 19763736 14 Smith HL Li W Cheetham ME Molecular chaperones and neuronal proteostasis. Semin Cell Dev Biol 2015 40 142 152 25770416 15 Rzechorzek NM Connick P Patani R Selvaraj BT Chandran S Hypothermic preconditioning of human cortical neurons requires proteostatic priming. EBioMedicine 2015 2 528 535 26287272 16 Lin W Popko B Endoplasmic reticulum stress in disorders of myelinating cells. Nat Neurosci 2009 12 379 385 19287390 17 Gow A Wrabetz L CHOP and the endoplasmic reticulum stress response in myelinating glia. Curr Opin Neurobiol 2009 19 505 510 19744850 18 Kumar R Krause GS Yoshida H Mori K DeGracia DJ Dysfunction of the unfolded protein response during global brain ischemia and reperfusion. J Cereb Blood Flow Metab 2003 23 462 471 12679723 19 Larson AC Jamrogowicz JL Kulikowicz E Cerebrovascular autoregulation after rewarming from hypothermia in a neonatal swine model of asphyxic brain injury. J Appl Physiol (1985) 2013 115 1433 1442 24009008 20 Lee JK Brady KM Mytar JO Cerebral blood flow and cerebrovascular autoregulation in a swine model of pediatric cardiac arrest and hypothermia. Crit Care Med 2011 39 2337 2345 21705904 21 Lee JK Yang ZJ Wang B Noninvasive autoregulation monitoring in a swine model of pediatric cardiac arrest. Anesth Analg 2012 114 825 836 22314692 22 Chonghaile TN Gupta S John M Szegezdi E Logue SE Samali A BCL-2 modulates the unfolded protein response by enhancing splicing of X-box binding protein-1. Biochem Biophys Res Commun 2015 466 40 45 26319553 23 Hyrskyluoto A Bruelle C Lundh SH Ubiquitin-specific protease-14 reduces cellular aggregates and protects against mutant huntingtin-induced cell degeneration: Involvement of the proteasome and ER stress-activated kinase IRE1alpha. Hum Mol Genet 2014 23 5928 5939 24951540 24 Lafrenaye AD Todani M Walker SA Povlishock JT Microglia processes associate with diffusely injured axons following mild traumatic brain injury in the micro pig. J Neuroinflammation 2015 12 186 015-0405-6 26438203 25 Martin LJ Al-Abdulla NA Brambrink AM Kirsch JR Sieber FE Portera-Cailliau C Neurodegeneration in excitotoxicity, global cerebral ischemia, and target deprivation: A perspective on the contributions of apoptosis and necrosis. Brain Res Bull 1998 46 281 309 9671259 26 Tagin M Zhu C Gunn AJ Beneficence and nonmaleficence in treating neonatal hypoxicischemic brain injury. Dev Neurosci 2015 37 305 310 25720376 27 Gunn AJ Gluckman PD Head cooling for neonatal encephalopathy: The state of the art. Clin Obstet Gynecol 2007 50 636 651 17762415 28 Thornton C Hagberg H Role of mitochondria in apoptotic and necroptotic cell death in the developing brain. Clin Chim Acta 2015 451 35 38 25661091 29 Bainbridge A Tachtsidis I Faulkner SD Brain mitochondrial oxidative metabolism during and after cerebral hypoxia-ischemia studied by simultaneous phosphorus magnetic-resonance and broadband near-infrared spectroscopy. Neuroimage 2014 102 Pt 1 173 183 23959202 30 Thoresen M Penrice J Lorek A Mild hypothermia after severe transient hypoxiaischemia ameliorates delayed cerebral energy failure in the newborn piglet. Pediatr Res 1995 37 667 670 7603788 31 Martin LJ Brambrink AM Price AC Neuronal death in newborn striatum after hypoxiaischemia is necrosis and evolves with oxidative stress. Neurobiol Dis 2000 7 169 191 10860783 32 Mueller-Burke D Koehler RC Martin LJ Rapid NMDA receptor phosphorylation and oxidative stress precede striatal neurodegeneration after hypoxic ischemia in newborn piglets and are attenuated with hypothermia. Int J Dev Neurosci 2008 26 67 76 17950559 33 Agnew DM Koehler RC Guerguerian AM Hypothermia for 24 hours after asphyxic cardiac arrest in piglets provides striatal neuroprotection that is sustained 10 days after rewarming. Pediatr Res 2003 54 253 262 12736390 34 Ni X Yang ZJ Wang B Early antioxidant treatment and delayed hypothermia after hypoxia-ischemia have no additive neuroprotection in newborn pigs. Anesth Analg 2012 115 627 637 22745113 35 Zhu J Wang B Lee JH Additive neuroprotection of a 20-HETE inhibitor with delayed therapeutic hypothermia after hypoxia-ischemia in neonatal piglets. Dev Neurosci 2015 37 376 389 25721266 36 Hofmann S Cherkasova V Bankhead P Bukau B Stoecklin G Translation suppression promotes stress granule formation and cell survival in response to cold shock. Mol Biol Cell 2012 23 3786 3800 22875991 37 Harding HP Zhang Y Ron D Protein translation and folding are coupled by an endoplasmicreticulum-resident kinase. Nature 1999 397 271 274 9930704 38 Lu M Lawrence DA Marsters S Opposing unfolded-protein-response signals converge on death receptor 5 to control apoptosis. Science 2014 345 98 101 24994655 39 Sun X Crawford R Liu C Luo T Hu B Development-dependent regulation of molecular chaperones after hypoxia-ischemia. Neurobiol Dis 2015 82 123 131 26070787 40 Chavez-Valdez R Flock DL Martin LJ Northington FJ Endoplasmic reticulum pathology and stress response in neurons precede programmed necrosis after neonatal hypoxia-ischemia. Int J Dev Neurosci 2016 48 58 70 26643212 41 Knight JR Bastide A Roobol A Eukaryotic elongation factor 2 kinase regulates the cold stress response by slowing translation elongation. Biochem J 2015 465 227 238 25353634 42 Zhao H Chen Y Effects of mild hypothermia therapy on the levels of glutathione in rabbit blood and cerebrospinal fluid after cardiopulmonary resuscitation. Iran J Basic Med Sci 2015 18 194 198 25810895 43 Yuan X Ghosh N McFadden B Hypothermia modulates cytokine responses after neonatal rat hypoxic-ischemic injury and reduces brain damage. ASN Neuro 2014 6 10.1177/1759091414558418. Print 2014 44 Bickler PE Clark JP Gabatto P Brosnan H Hypoxic preconditioning and cell death from oxygen/glucose deprivation co-opt a subset of the unfolded protein response in hippocampal neurons. Neuroscience 2015 310 306 321 26404874 45 Fischer H Koenig U Eckhart L Tschachler E Human caspase 12 has acquired deleterious mutations. Biochem Biophys Res Commun 2002 293 722 726 12054529 46 Tekes A Poretti A Scheurkogel MM Apparent diffusion coefficient scalars correlate with near-infrared spectroscopy markers of cerebrovascular autoregulation in neonates cooled for perinatal hypoxic-ischemic injury. AJNR Am J Neuroradiol 2015 36 188 193 25169927 47 Howlett JA Northington FJ Gilmore MM Cerebrovascular autoregulation and neurologic injury in neonatal hypoxic-ischemic encephalopathy. Pediatr Res 2013 74 525 535 23942555 48 Burton VJ Gerner G Cristofalo E A pilot cohort study of cerebral autoregulation and 2-year neurodevelopmental outcomes in neonates with hypoxic-ischemic encephalopathy who received therapeutic hypothermia. BMC Neurol 2015 15 209 015-0464-4 26486728 49 Hori D Everett AD Lee JK Rewarming rate during cardiopulmonary bypass is associated with release of glial fibrillary acidic protein. Ann Thorac Surg 2015 100 1353 1358 26163357 50 Bisschops LL Hoedemaekers CW Mollnes TE van der Hoeven JG Rewarming after hypothermia after cardiac arrest shifts the inflammatory balance. Crit Care Med 2011 51 Dohi K Miyamoto K Fukuda K Status of systemic oxidative stress during therapeutic hypothermia in patients with post-cardiac arrest syndrome. Oxid Med Cell Longev 2013 2013 562429 24066191 52 Alva N Palomeque J Carbonell T Oxidative stress and antioxidant activity in hypothermia and rewarming: Can RONS modulate the beneficial effects of therapeutic hypothermia? Oxid Med Cell Longev 2013 2013 957054 24363826 53 Guerra R Vera-Aguilar E Uribe-Ramirez M Exposure to inhaled particulate matter activates early markers of oxidative stress, inflammation and unfolded protein response in rat striatum. Toxicol Lett 2013 222 146 154 23892126 54 Llorente IL Burgin TC Perez-Rodriguez D Martinez-Villayandre B Perez-Garcia CC Fernandez-Lopez A Unfolded protein response to global ischemia following 48 h of reperfusion in the rat brain: The effect of age and meloxicam. J Neurochem 2013 127 701 710 23763503 55 Millson GC Hunter GD Protein synthesis in normal and scrapie mouse brain. J Neurochem 1968 15 447 453 5691263 56 Agematsu K Korotcova L Morton PD Gallo V Jonas RA Ishibashi N Hypoxia diminishes the protective function of white-matter astrocytes in the developing brain. J Thorac Cardiovasc Surg 2016 151 265 272 e3 26412317 57 Oberheim NA Takano T Han X Uniquely hominid features of adult human astrocytes. J Neurosci 2009 29 3276 3287 19279265 58 Araque A Carmignoto G Haydon PG Oliet SH Robitaille R Volterra A Gliotransmitters travel in time and space. Neuron 2014 81 728 739 24559669 59 Lu XM Shear DA Deng-Bryant Y Comprehensive evaluation of neuroprotection achieved by extended selective brain cooling therapy in a rat model of penetrating ballistic-like brain injury. Ther Hypothermia Temp Manag 2015 60 Ohnishi M Urasaki T Ochiai H Selective enhancement of wnt4 expression by cyclic AMP-associated cooperation between rat central astrocytes and microglia. Biochem Biophys Res Commun 2015 467 367 372 26431871 61 Jelsing J Hay-Schmidt A Dyrby T Hemmingsen R Uylings HB Pakkenberg B The prefrontal cortex in the gottingen minipig brain defined by neural projection criteria and cytoarchitecture. Brain Res Bull 2006 70 322 336 17027768 62 Davidson JO Yuill CA Zhang FG Wassink G Bennet L Gunn AJ Extending the duration of hypothermia does not further improve white matter protection after ischemia in term-equivalent fetal sheep. Sci Rep 2016 6 25178 27121655 63 Martinez-Biarge M Bregant T Wusthoff CJ White matter and cortical injury in hypoxicischemic encephalopathy: Antecedent factors and 2-year outcome. J Pediatr 2012 161 799 807 22682614 64 Yang Q Gao H Dong R Wu YQ Sequential changes of endoplasmic reticulum stress and apoptosis in myocardial fibrosis of diabetes mellitus-induced rats. Mol Med Rep 2016 13 5037 5044 27121167

PMC005xxxxxx/PMC5127734.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 7807270 22115 Environ Int Environ Int Environment international 0160-4120 1873-6750 27745946 5127734 10.1016/j.envint.2016.10.007 NIHMS823586 Article Predictors of urinary flame retardant concentration among pregnant women Hoffman Kate 1 Lorenzo Amelia 1 Butt Craig 1 Adair Linda 2 Herring Amy H. 2 Stapleton Heather M. 1 Daniels Julie 2 1 Nicholas School of the Environment, Duke University, Durham, North Carolina, USA 2 Gillings School of Global Public Health, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, USA Address Correspondence To: Kate Hoffman, Ph.D., Nicholas School of the Environment, Duke University, A220 LSRC Box 90328, Durham, NC 27708 USA, phone: (919) 684-6952, fax: (919) 684-8741, kate.hoffman@duke.edu 20 10 2016 13 10 2016 1 2017 01 1 2018 98 96101 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. Background Organophosphate compounds are commonly used in residential furniture, electronics, and baby products as flame retardants and are also used in other consumer products as plasticizers. Although the levels of exposure biomarkers are generally higher among children and decrease with age, relatively little is known about the individual characteristics associated with higher levels of exposure. Here, we investigate urinary metabolites of several organophosphate flame retardants (PFRs) in a cohort of pregnant women to evaluate patterns of exposure. Methods Pregnant North Carolina women (n=349) provided information on their individual characteristics (e.g. age and body mass index (BMI)) as a part of the Pregnancy Infection and Nutrition Study (2002–2005). Women also provided second trimester urine samples in which six PFR metabolites were measured using mass spectrometry methods. Results PFR metabolites were detected in every urine sample, with BDCIPP, DHPH, ip-PPP and BCIPHIPP detected in >80% of samples. Geometric mean concentrations were higher than what has been reported previously for similarly-timed cohorts. Women with higher pre-pregnancy BMI tended to have higher levels of urinary metabolites. For example, those classified as obese at the start of pregnancy had ip-PPP levels that were 1.52 times as high as normal weight range women (95% confidence interval: 1.23, 1.89). Women without previous children also tended to have higher urinary levels of DPHP, but lower levels of ip-PPP. In addition, we saw strong evidence of seasonal trends in metabolite concentrations (e.g. higher DPHP, BDCIPP, and BCIPHIPP in summer, and evidence of increasing ip-PPP between 2002 and 2005). Conclusions Our results indicate ubiquitous exposure to PFRs among NC women in the early 2000s. Additionally, our work suggests that individual characteristics are related to exposure and that temporal variation, both seasonal and annual, may exist. organophosphate flame retardants (PFRs) pregnancy exposure 1. Introduction Flame retardant chemicals have been added to a variety of household products to meet flammability standards for decades. Until the mid-2000s, polybrominated diphenyl ethers (PBDEs) accounted for a large proportion of flame retardants used in household products including polyurethane foam and electronics; however, regulatory action and concern over the persistence, bioaccumulation, and toxicity of PBDEs led to an increased use of alternative flame retardants (Stapleton et al. 2012b; van der Veen and de Boer 2012). Organophosphate flame retardants (PFRs) are now among the most commonly used PBDE alternatives in industries that manufacture residential furniture, electronics (e.g. TVs) and baby products (e.g. nursing pillows). They are commonly added to flame retardant mixtures, such as Firemaster® 550 (FM550), and to other consumer products as plasticizers (Ballesteros-Gomez et al. 2014; Fang et al. 2013; Patisaul et al. 2013; Stapleton et al. 2008; Stapleton et al. 2009; Stapleton et al. 2011). PFRs have been detected with high frequency in recent studies of home, office, and automobile dust, demonstrating that they leach from products and suggesting ubiquitous exposure [e.g. (Brandsma et al. 2013; Brommer and Harrad 2015; Cao et al. 2014; Carignan et al. 2013; Cristale et al. 2016; Hoffman et al. 2015b; Stapleton et al. 2008; Stapleton et al. 2009; Stapleton et al. 2011)]. Additionally, an accumulating body of research indicates that the vast majority of U.S. adults (>90%) have detectable levels of PFR metabolites in their urine, and similar detection frequencies have been reported in Canadian, European, Asian and Australian populations (e.g. (Butt et al. 2014 and 2016; Cequier et al. 2015; Dodson et al. 2014; Hoffman et al. 2014; Hoffman et al. 2015a; Hoffman et al. 2015b; Kosarac et al. 2016; Meeker et al. 2013a; Van den Eede et al. 2015; Su et al. 2015)). Although data suggest that metabolite levels vary by age, with younger individuals shown to have higher exposures (e.g. Butt et al. 2014 and 2016; Hoffman et al. 2015b; Van den Eede et al. 2015), the individual characteristic and behaviors associated with higher levels of exposure are not well understood. In our present work we investigate the levels of exposure in a large pregnancy cohort, and additionally assess factors associated with higher levels of PFR metabolites in urine samples. We focus on widely used PFRs and six metabolites (Figure 1). Identifying factors contributing to higher levels of exposure to these compounds is particularly important because certain PFRs can disrupt normal endocrine function (Liu et al. 2012; Wang et al. 2013; Meeker et al. 2013a) and 2013b), are carcinogenic (Faust and August 2011; Gold et al. 1978), neurotoxic (Dishaw et al. 2011), reproductive toxicants (Meeker et al. 2013a and 2013b; Liu et al. 2013; Farhat et al. 2013), and potentially adipogenic (Patisaul et al. 2013; Pillai et al. 2014). In addition, recent data suggests that PFRs may have similar or greater toxicity than their PBDE predecessors, particularly with respect to neurodevelopmental outcomes (Behl et al. 2015; Behl. et al 2016). 2. Methods 2.1 Study Population The Pregnancy Infection and Nutrition (PIN) Study enrolled a cohort of central North Carolina women in early pregnancy and conducted follow-up through delivery (PIN 2012). PIN women were recruited from the University of North Carolina prenatal care clinic, and delivered their infants at University of North Carolina hospitals between 2001 and 2006 (n = 2009; PIN phase 3). This analysis is part of a larger project investigating the impacts of exposure to environmental chemicals on children’s growth. This sample is limited to 349 mothers recruited during the final four years of the cohort study, whose children had growth measurements collected at multiple time points (infants born 2002–2005). Self-administered questionnaires, telephone interviews, and home visits were used to collect pregnancy and postpartum health and lifestyle information throughout pregnancy and after the child’s birth (PIN 2012). All study protocols were approved by the institutional review board at the University of North Carolina at Chapel Hill and all mothers provided informed consent prior to completing any study activities. 2.2 Urine Collection and Analysis During the late-second or early-third trimester, PIN women collected a spot urine sample in a standard urine collection cup. The time and date of collection was recorded, and urine samples were aliquoted into polyethylene storage tubes and frozen at −80° C until analysis. Urine samples were extracted using enzyme deconjugation and solid phase extraction (SPE) techniques as previously described (Van den Eede et al. 2013) but adapted for 5 ml of urine (Butt et al. 2016). In brief, samples were thawed, 5 ml of urine were aliquoted into a clean glass test tube, the internal standard mixture was spiked (10 ng of d10-BDCIPP, 8.8 ng of d10-DPHP; 25 ng of d12-TCEP) and samples vortexed. After pH adjustment with sodium acetate (1.75 ml of 1 M sodium acetate, pH 5), the enzyme solution was added (250 µl of 1000 units/ml µ-glucuronidase, 33 units/ml sulfatase in 0.2 M sodium acetate buffer), and the samples were vortexed and incubated overnight in a 37°C water bath. Samples were extracted and cleaned using SPE with a StrataX-AW (60 mg, 3 ml) column, and were reconstituted in 500 ul of 1:1 water:methanol, as previously described (Butt et al. 2016). Internal standard recovery was quantified by spiking with 13C2-DPHP. Extracts were analyzed using electrospray ionization (ESI) liquid chromatography tandem mass spectrometry (LC-MS/MS) with a Phenomenex Luna C18 column on an Agilent 1100 series LC and an Agilent 6410B tandem mass spectrometer as previously described (Butt et al. 2014 and 2016). Data were acquired under multiple reaction monitoring conditions using optimized parameters. Analyte responses were normalized to internal standard responses. BCIPP and BDCIPP were normalized using d10-BDCIPP, DPHP, ip-PPP and tb-PPP were normalized using d10-DPHP and BCIPHIPP was normalized using d12-TCEP. The mean recovery of the mass-labelled standards in the urine samples (n=349) was 97% (standard error (SE) = 2.1%) for d10-DPHP, 98% (SE=3.0%) for d10-BDCIPP and 34% (SE=1.0%) for d12-TCEP. The low d12-TCEP recovery is partially due to quantification inaccuracies resulting from matrix suppression since the d12-TCEP recovery was 55–73% in the blank samples (clean water only). Analyte values were blank corrected using the mean laboratory blank levels. Method detection limits (MDLs) were calculated as three times the standard deviation of the laboratory blanks, normalized to the average urine volume (3 ml). Sample were assessed in three batches and MDLs were calculated separately for each batch (MDLs: 136–333 pg/ml for BCIPP, 127–243 pg/ml for DPHP, 60–197 pg/ml for BDCIPP, 37–177 pg/ml for ip-PPP, 213–846 pg/ml for tb-PPP and 3–33 pg/ml for BCIPHIPP. Specific gravity (SG) was measured in each urine sample prior to analysis using a digital handheld refractometer (Atago). Relative method accuracy was assessed by measuring PFR metabolites in SRM 3673 (n=3). Specific gravity-normalized concentrations were 1.56 ng/ml (SE=0.09) for BDCIPP, 0.44 ng/ml (0.02) for BCIPHIPP, 0.65 ng/ml (0.08) for DPHP and 6.0 ng/ml (0.30) for ip-PPP. These values are similar to those previously reported by our lab for SRM 3673 with the exception of ip-PPP, which was 0.6-times those of the current study (Hammel et al. 2016). BCPP and tb-PPP were not detected in the SRM. To investigate the impacts of differences in urine dilution on results, we conducted analyses of urinary metabolites using raw PFR metabolite measures as well as using SG-corrected concentrations (Boeniger et al. 1993). Corrected and uncorrected concentrations were very highly correlated [Spearman correlations (rs) >0.82 for all metabolites] and results were very similar using both methods. Here we present only the results obtained with the specific gravity corrected concentrations to facilitate comparison with previous studies. 2.3 Statistical Analysis Preliminary analyses indicated that urinary PFR metabolite levels were not normally distributed and were positively skewed (i.e. skewed right). Accordingly, we used non-parametric analyses or log10-transformed metabolite concentrations in statistical analyses. We calculated descriptive statistics for each PFR metabolite and conducted additional analyses for those that were detected in >70% of urine samples. For these metabolites, samples with concentrations below the method limit of detection (MDL) were replaced with the MDL/2 prior to adjustment for specific gravity (see below). Spearman correlations (rs) were used to assess relationships between urinary PFR metabolites. We used linear regression models with log10-transformed metabolite levels as the outcome to assess maternal predictors of PFR levels. Predictive analyses were conducted only for metabolites detected in at least 80% of the samples. Beta coefficients from these models were exponentiated (10β) for interpretation and represent the multiplicative change relative to the reference category for categorical variables, and the multiplicative change for a one unit increase for continuous variables. In addition to univariate models, we conducted multivariate regression analyses including all variables of interest [age (≤25, 26–30, 31–35 and ≥36), race (white and non-white), education (≤15and ≥16, pre-pregnancy BMI (underweight, normal range, overweight and obese), parity(0 and≥1), gestational duration at the time of sample collection (24–26, 27–28 and 29–30), date of sample collection (continuous measure scaled to years) and season of collection (December–February, March–May, June–August, September–November)] to obtain the independent effect of each variable. Unadjusted and adjusted models produced very similar results. Here, we present adjusted analyses; however, unadjusted results are shown in Supplemental Table 1. All statistical analyses were conducted in SAS (version 9.4; Cary, NC) and statistical significance was set at alpha=0.05. 3. Results and Discussion Women averaged 29.6 years of age at the time of enrollment and were highly educated, with nearly 70% having a college education (Table 1). Women included in the present study (Table 1) were more likely to be white, have higher educational attainment, and be older than mothers in the larger PIN cohort (Daniels et al. 2010; PIN 2012). Nearly half of the participants were primiparous (47.6%); and the majority (55.6%) had a BMI within the normal range at the start of their pregnancy. Urine samples were collected between 24 and 30 weeks gestation, and the average collection time was gestational week 27. 3.1 PRF Metabolite Levels BDCIPP, DPHP, ip-PPP and BCIPHIPP were detected frequently in urine samples (92.8%, 83.7%, 99.4% and 98.3%, respectively), and concentrations varied considerably between women. Among these compounds, concentration ranged from non-detectable to approximately 100 ng/mL for all analytes (Table 2). BCIPP and tb-PPP were detected less frequently (48.7% and 2.0% detect, respectively). Correlations between DPHP and other PFR metabolites were generally small, but statistically significant (rs from 0.11 for ip-PPP to 0.31 for BDCIPP, with p<0.05 for both). BDCIPP was additionally correlated with BCIPHIPP (rs=0.21, p=0.001) but not ip-PPP, and BCIPHIPP was not correlated with ip-PPP among women in our cohort. The full correlation matrix is shown in Supplemental Table 2. While some overlap in patterns of use or exposure pathways for the parent PFRs is probable, the small magnitude of correlation suggests potential for different sources of exposure or differences in their toxicokinetics (or pharmacokinetics). Although PRFs are commonly considered replacements for the PentaBDE mixture, which was phased out in the U.S. at the approximate time of our sample collection, our results indicate that exposures were ubiquitous in the early 2000s, suggesting that PFR use was already common. This finding is consistent with previous work identifying PFRs in National Institute of Standards and Technology Standards Reference Material dust samples collected in the 1990s (e.g. SRM 2585; Van den Eede et al. 2011). Urinary DPHP and BDCIPP levels in the present study were similar to those measured in a 2011–2012 cohort of pregnant women from North Carolina (Hoffman et al. 2014). This was surprising since the present cohort was sampled approximately 6–8 years earlier than the 2011–2012 cohort, and PFR chemical usage is thought to have been increased since the phase-out of Penta BDE(Stapleton et al. 2012b; van der Veen and de Boer 2012). In addition, DPHP and BDCIPP levels in the present cohort are approximately 1 order of magnitude greater than those from a Massachusetts cohort studied during a similar time frame (2002–2007) (Meeker et al. 2013a); however, the Massachusetts cohort was exclusively male which may explain observed differences. Data have previously shown that women have higher levels of DPHP than men (e.g. Hoffman et al. 2015). TPHP was recently detected in nail polish, which has been offered as a possible explanation for this pattern (Mendelsohn et al. 2016); however, we are unaware whether TPHP was used in nail polish in the early 2000s. In addition BDCIPP levels were also higher in our cohort compared to the men in Meeker et. al. 2013a. Metabolic differences could also explain this pattern. Data from Hays et al. (2015) demonstrate that urine flaw rates differ between males and females (higher in males over 12 years of age), a factor which may impact urinary PFR metabolite concentrations. The ip-PPP levels were approximately 3–6 times higher than recent cohorts from New Jersey and California (Butt et al. 2014 and 2016). Interestingly, the women in the PIN cohort were also found to have higher levels of PBDEs in their breast milk than women in other similarly timed U.S. cohorts (Daniels et al. 2010). 3.2 Predictors of PFR levels We found little evidence of association between PFR metabolites and maternal age at the start of pregnancy after adjusting for other factors. Only associations between maternal age and BCIPHPP remained suggestive of an inverse association after adjustment; though confidence intervals were imprecise (Table 3). Although past research has shown that metabolite concentrations decrease with age (Van den Eede et al. 2015 and Hoffman et al. 2015b), the age range of participants in our cohort was relatively narrow, potential limiting our ability to detect real difference occurring with age. Women experiencing their first pregnancy had lower levels of ip-PPP (10β=0.83; 95% CI: 0.71, 0.97; p=0.02), but significantly higher levels of urinary DPHP (10β=1.27; 95% CI: 1.04–1.55; p=0.02). Information on differences in consumer patterns between primiparous women and those with previous children could be helpful in identifying drivers of these associations. Neither race nor education were strongly associated with urinary levels. However, the cohort was primarily white and well-educated women, reducing our power to fully investigate patterns by these demographics. Still, we did observe higher levels of ip-PPP among women with less education (25% higher for women with less than a 4 year college degree; 10β=1.25; 95% CI: 1.01, 1.56; p=0.04). Educational attainment, a marker of socioeconomic position (SEP), has previously been associated with biomarkers of PBDE exposure in children (e.g. Stapleton et al. 2012a). Compared to women with a normal pre-pregnancy BMI, those who were overweight or obese prior to pregnancy had higher levels of urinary BDCIPP, DPHP and ip-PPP. For example, obese women had urinary ip-PPP levels 1.52 times those of women with pre-pregnancy BMIs in the normal range (95% CI: 1.23, 1.89; p=0.0002). Our previous research indicates that rats exposed to FM550 in early-life gain weight more readily, suggesting that FM550’s components may be obesogenic (Patisaul et al. 2013). Additional work with FM550 suggests that the obesogenic potential may be driven by PFRs present in FM550 (e.g. TPHP and ip-TPHP), which are ligands for the peroxisome proliferator-activated receptor gamma which is a critical nuclear receptor in adipocyte differentiation and lipid storage (Belcher et al. 2014; Fang et al. 2015; Pillai et al. 2014). However, it is also possible that PFR metabolism or excretion is intrinsically associated with BMI. For example, previous work from Hays et al. (2015) indicated that failing to account for urine dilution could induce correlation between categories of BMI and urinary BPA metabolites. Although we have conducted analyses both with and without correction for dilution (i.e. specific gravity) and observed similar associations with both methods, other factors linking BMI and PFR metabolism or excretion could be at play. Alternatively, differences in behavior and activity patterns that are associated with BMI may explain differences in PFR exposure. Based on our study design we are unable to distinguish which factors may be causal in associations between urinary metabolites and BMI; however, this is an important consideration for future research. The week of gestation during which the urine sample was collected tended to be inversely associated with BDCIPP and DPHP; however, associations were imprecisely estimated and not statistically significant. If real, differences in kidney function and metabolism during pregnancy may explain these patterns. These results are an important consideration for epidemiologic studies investigating the consequences of prenatal exposure to PFRs with a single urine sample during pregnancy and suggest that gestational timing of sample collection may be an important factor driving measured concentrations. Season of collection was the strongest predictor of urinary PFR metabolite concentrations. Compared to samples collected in the winter, samples collected in the summer had significantly higher concentrations of BDCIPP, DHPH and BCIPHIPP. For example, BDCIPP concentrations in samples collected in the summer were approximately 3 times higher in the summer (10β = 3.97; 95% CI: 2.96, 5.32; p<0.0001). Figure 2 depicts individual concentration measures as well as the average month temperature in our study area. This pattern suggests that behavior or exposure varies with temperature. Past research has shown that indoor dust samples collected in China had lower PFR concentrations in the summer months (Cao et al. 2014). This could mean that PFRs are partition into air more readily in warmer summer months, potentially increasing inhalation exposure which is increasingly recognized as an important exposure route for PFRs (e.g. Xu et al. 2016). However, although we did not directly evaluate indoor temperatures, they are likely to be relatively stable over time in North Carolina (central heating and air conditioning are exceedingly common in the study area) which suggests that exposure in other environments (e.g. outdoors or in vehicles) could be a substantial contributor to total the body burden. For example, PFRs are commonly detected in dust samples collected in cars (e.g. Harrada, 2016). It is possible that exposure inside cars in the summer is higher due to higher temperatures. In contrast, concentrations of ip-PPP were lower in the spring and summer than in the winter. Cumulatively, these results indicate that season of collection could be an important confounder in future epidemiology studies using spot urine samples as a proxy for longer-term PFR exposure. Despite a relatively narrow time frame of sample collection, we also observed statistical evidence of increases in urinary ip-PPP concentration; levels of ip-PPP increased by 18% per year (10β = 1.18; 95% CI: 1.08, 1.28; p=0.0003). Our results represent the first large-scale assessment of individual factors related to the levels of urinary PFR metabolites during pregnancy. These results should be interpreted in the context of several limitations. First, our cohort was relatively homogeneous, with the majority of women reporting white race and having high educational attainment, potentially limiting the generalizability of these results to other populations. However, the cohort’s homogeneity may also be an asset for planned future research investigating health impacts of exposure, as it could reduce the potential for confounding. Another limitation is that these results rely on a single spot urine sample. Although we have previously shown that measures of BDCIPP and DPHP in urine to be relatively stable over the course of pregnancy (Hoffman et al. 2014), and moderately to highly stable over a week (Hoffman et al. 2015b), we were unable to evaluate the potential for differences over time in this cohort. In addition, previous research has shown that urinary metabolite concentrations may vary throughout the day (Cequier et al. 2015; Hoffman et al. 2015b). While we were not able to evaluate diurnal variation because the vast majority of urine samples were collected in the same time window (>95% of samples were collected between 0700 and 1200 hours), the temporal standardization served to control such variation for this comparison. Because samples were mainly collected in the early morning, they are likely to represent exposure over the previous night, which we expect many women would have spent in their homes. 4. Conclusions Although PFRs are commonly thought to be a replacement for the Penta-BDE mixture, which was phased out of used as flame retardants in the U.S. in the mid-2000s, our results indicate that exposure to PFRs was wide-spread by 2002. We observed strong seasonal trends in metabolite level suggesting season of collection may be an important factor to consider in future epidemiologic investigations. In addition, our results suggest that levels vary by BMI, parity, and education. Additional data are needed to identify the mechanisms explaining observed associations and to determine whether the levels of exposure that we observed are associated with any adverse help impacts among pregnant women of their children. Supplementary Material This research was supported in part by grants from the National Institute of Environmental Health Sciences (R21 ES023904and P30ES10126) and the U.S. Environmental Protection Agency (RD832736). The work of KH was funded in part by a training grant from the National Institute of Environmental Health Sciences (T32 ES007018) Abbreviations BCIPHIPP 1-hydroxy-2-propyl bis(1-chloro-2-propyl) phosphate BMI body mass index BCIPP bis(1-chloro-2-propyl) phosphate BDCIPP bis(1,3-dichloro-2-propyl) phosphate CI confidence interval DPHP diphenyl phosphate FM550 Firemaster® 550 GM geometric mean ip-PPP isopropyl-phenyl phenyl phosphate MDL method limit of detection PFRs organophosphate flame retardants PBDEs polybrominated diphenyl ethers PIN Pregnancy Infection and Nutrition Study tb-PPP tert-butyl phenyl phenyl phosphate Figure 1 Chemical structures of urinary PFR metabolites monitored. TPHP metabolite = DPHP; Isopropyl-phenyl diphenyl phosphate metabolite = ip-PPP; Tertbutyl-phenyl diphenyl phosphate metabolite= tb-PPP; TDCIPP metabolite = BDCIPP; and tris(1-chloro-2-isopropyl) phosphate (TCIPP metabolites) = BCIPP and BCIPHIPP. Figure 2 Individual Urinary BDCIPP (ng/mL) concentrations plotted by date of sample collection (green dots), overlaid with the average month temperature in Chapel Hill, North Carolina (Degrees Celsius; grey line). Table 1 Selected characteristics of 349 pregnant North Carolina women (2002–2005). N % Total 349 100.0 Age ≤25 76 21.8 26–30 126 36.1 31–35 107 30.7 ≥36 40 11.5 Race white 278 79.7 non-white 71 20.3 Education (years) ≤15 106 30.4 ≥16 243 69.6 Parity 0 166 47.6 ≥1 183 52.4 Pre-pregnancy BMI Underweight 46 13.2 Normal range 194 55.6 Overweight 42 12.0 Obese 67 19.2 Gestational weeks at urine sample collection 24–26 weeks 72 20.6 27–28 weeks 139 39.8 29–30 weeks 138 39.5 Season of urine sample collection Winter (Dec – Feb) 81 23.2 Spring (Mar – May) 90 25.8 Summer (Jun – Aug) 94 26.9 Fall (Sep – Nov) 84 24.1 Table 2 Detection frequency, geometric mean and distribution information (ng/mL) for urinary PFR metabolites (N=349). Sample were assessed in three batches and MDLs were calculated separately for each batch (MDL ranges: 136–333 pg/ml for BCIPP, 127–243 pg/ml for DPHP, 60–197 pg/ml for BDCIPP, 37–177 pg/ml for ip-PPP, 213–846 pg/ml for tb-PPP and 3–33 pg/ml for BCIPHIPP. Metabolite % Detect GMa 25th %ile 50th %ile 75th %ile Maximum BCIPP 48.7 -- -- 0.7 1.1 6.1 BDCIPP 92.8 1.8 0.8 1.9 3.6 140 DPHP 83.7 1.4 0.8 1.3 2.7 112 ip-PPP 99.4 6.8 4.2 7.1 10.9 69 tb-PPP 2.0 -- -- -- -- 8.6 BCIPHIPP 98.3 0.5 0.2 0.4 0.8 98 a GM: geometric mean Table 3 Multiplicative change (10β) in PFR metabolite concentration by maternal characteristic (simultaneously adjusted for all included factors). Predictor BDCIPP DPHP ip-PPP BCIPHIPP 10β (95% CI) p 10β (95% CI) p 10β (95% CI) p 10β (95% CI) p Age ≤25 Reference -- Reference -- Reference -- Reference -- 26–30 1.06 (0.78, 1.45) 0.71 1.12 (0.83, 1.52) 0.44 0.99 (0.79, 1.26) 0.96 0.92 (0.63, 1.34) 0.66 31–35 0.93 (0.66, 1.30) 0.66 0.97 (0.70, 1.35) 0.87 0.84 (0.65, 1.09) 0.20 0.74 (0.49, 1.12) 0.15 ≥36 1.04 (0.70, 1.55) 0.84 0.99 (0.68, 1.46) 0.97 1.10 (0.81, 1.49) 0.53 0.67 (0.41, 1.09) 0.11 Race white Reference -- Reference -- Reference -- Reference -- non-white 1.09 (0.84, 1.42) 0.52 0.91 (0.70, 1.17) 0.45 0.88 (0.72, 1.07) 0.21 0.89 (0.64, 1.22) 0.46 Education (years) ≤15 1.03 (0.77, 1.38) 0.83 1.07 (0.81, 1.41) 0.65 1.25 (1.01, 1.56) 0.04 0.79 (0.56, 1.13) 0.20 ≥16 Reference -- Reference -- Reference -- Reference -- Parity 0 0.93 (0.75, 1.14) 0.46 1.27 (1.04, 1.55) 0.02 0.83 (0.71, 0.97) 0.02 1.22 (0.95, 1.57) 0.12 ≥1 Reference -- Reference -- Reference -- Reference -- Pre-pregnancy BMI Underweight 1.01 (0.74, 1.37) 0.95 1.12 (0.83, 1.51) 0.45 1.09 (0.86, 1.37) 0.48 0.73 (0.50, 1.07) 0.10 Normal range Reference -- Reference -- Reference -- Reference -- Overweight 1.09 (0.79, 1.50) 0.59 1.21 (0.89, 1.65) 0.22 1.36 (1.07, 1.73) 0.01 0.9 (0.61, 1.33) 0.60 Obese 1.17 (0.88, 1.57) 0.28 1.22 (0.92, 1.61) 0.17 1.52 (1.23, 1.89) 0.0002 1.14 (0.80, 1.62) 0.46 Gestational weeks at urine sample collection 24–26 weeks Reference -- Reference -- Reference -- Reference -- 27–28 weeks 0.84 (0.64, 1.11) 0.22 0.86 (0.66, 1.13) 0.28 0.89 (0.72, 1.10) 0.30 0.87 (0.62, 1.23) 0.43 29–30 weeks 0.85 (0.65, 1.13) 0.26 0.85 (0.65, 1.12) 0.25 1.07 (0.87, 1.32) 0.52 1.10 (0.78, 1.54) 0.60 Date (years) 1.08 (0.96, 1.21) 0.20 0.96 (0.84, 1.08) 0.54 1.18 (1.08, 1.28) 0.0003 1.08 (0.92, 1.23) 0.33 Season of Sample Collection Winter (Dec – Feb) Reference -- Reference -- Reference -- Reference -- Spring (Mar – May) 2.73 (2.05, 3.65) <.0001 1.05 (0.80, 1.39) 0.73 0.67 (0.54, 0.83) 0.0003 1.43 (1.01, 2.03) 0.04 Summer (Jun – Aug) 3.97 (2.96, 5.32) <.0001 1.62 (1.22, 2.15) 0.0008 0.75 (0.60, 0.93) 0.01 1.96 (1.37, 2.79) 0.0002 Fall (Sep – Nov) 1.73 (1.29, 2.32) 0.0003 1.15 (0.86, 1.53) 0.34 1.01 (0.81, 1.27) 0.90 1.40 (0.98, 2.01) 0.06 Highlights PFR metabolites were detected in all urine samples provided by pregnant women. Geometric mean concentrations were higher than for similarly-timed cohorts. Women with higher pre-pregnancy BMI had higher levels of urinary metabolites. PFR metabolite concentrations in urine vary seasonally. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain Conflict Disclosure: The authors have no actual or potential conflicting relationships relevant to this article to disclose. References Ballesteros-Gomez A Brandsma SH de Boer J Leonards PE Analysis of two alternative organophosphorus flame retardants in electronic and plastic consumer products: Resorcinol bis-(diphenylphosphate) (PBDPP) and bisphenol a bis (diphenylphosphate) (BPA-BDPP) Chemosphere 2014 116 10 14 24556545 Behl M Rice JR Smith MV Co CA Bridge M Hsieh JH Freedman JH Boyd WA Comparative Toxicity of Organophosphate Flame Retardants and Polybrominated Diphenyl Ethers to C. elegans Toxicol Sci 2016 [Epub ahead of print] Behl M Hsieh JH Shafer TJ Mundy WR Rice JR Boyd WA Freedman JH Hunter ES 3rd Jarema KA Padilla S Tice RR Use of alternative assays to identify and prioritize organophosphorus flame retardants for potential developmental and neurotoxicity Neurotoxicol Teratol 2015 52 Pt B 181 193 26386178 Belcher SM Cookman CJ Patisaul HB Stapleton HM In vitro assessment of human nuclear hormone receptor activity and cytotoxicity of the flame retardant mixture FM 550 and its triarylphosphate and brominated components Toxicol Lett 2014 228 93 102 24786373 Boeniger MF Lowry LK Rosenberg J Interpretation of urine results used to assess chemical exposure with emphasis on creatinine adjustments: A review Am Ind Hyg Assoc J 1993 54 615 627 8237794 Brandsma SH Sellstrom U de Wit CA de Boer J Leonards PE Dust measurement of two organophosphorus flame retardants, resorcinol bis(diphenylphosphate) (RBDPP) and bisphenol a bis(diphenylphosphate) (BPA-BDPP), used as alternatives for BDE-209 Environ Sci Technol 2013 47 14434 14441 24261654 Brommer S Harrad S Sources and human exposure implications of concentrations of organophosphate flame retardants in dust from uk cars, classrooms, living rooms, and offices Environ Int 2015 83 202 207 26232632 Butt CM Congleton J Hoffman K Fang M Stapleton HM Metabolites of organophosphate flame retardants and 2-ethylhexyl tetrabromobenzoate in urine from paired mothers and toddlers Environ Sci Technol 2014 48 10432 10438 25090580 Butt CM Hoffman K Chen A Lorenzo A Congleton J Stapleton HM Regional comparison of organophosphate flame retardant (PFR) urinary metabolites and tetrabromobenzoic acid (TBBA) in mother-toddler pairs from California and New Jersey Environ Int 2016 pii: S0160-4120(16)30247-1 Cao Z Xu F Covaci A Wu M Yu G Wang B Differences in the seasonal variation of brominated and phosphorus flame retardants in office dust Environ Int 2014 65 100 106 24480750 Carignan CC McClean MD Cooper EM Watkins DJ Fraser AJ Heiger-Bernays W Predictors of tris(1,3-dichloro-2-propyl) phosphate metabolite in the urine of office workers Environ Int 2013 55 56 61 23523854 Cequier E Sakhi AK Marce RM Becher G Thomsen C Human exposure pathways to organophosphate triesters — a biomonitoring study of mother-child pairs Environ Int 2015 75 159 165 25461425 Cristale J Hurtado A Gomez-Canela C Lacorte S Occurrence and sources of brominated and organophosphorus flame retardants in dust from different indoor environments in Barcelona, Spain Environ Res 2016 149 66 76 27179204 Daniels JL Pan IJ Jones R Anderson S Patterson DG Jr Needham LL Individual characteristics associated with PBDE levels in U.S. human milk samples Environ Health Perspect 2010 118 155 160 20056574 Dishaw LV Powers CM Ryde IT Roberts SC Seidler FJ Slotkin TA Is the pentabde replacement, tris (1,3-dichloro-2-propyl) phosphate (TDCPP), a developmental neurotoxicant? Studies in pc12 cells Toxicology and applied pharmacology 2011 256 281 289 21255595 Dodson RE Van den Eede N Covaci A Perovich LJ Brody JG Rudel RA Urinary biomonitoring of phosphate flame retardants: Levels in California adults and recommendations for future studies Environ Sci Technol 2014 48 13625 13633 25388620 Fang M Webster TF Gooden D Cooper EM McClean MD Carignan C Investigating a novel flame retardant known as V6: Measurements in baby products, house dust, and car dust Environ Sci Technol 2013 47 4449 4454 23565680 Fang ML Webster TF Ferguson PL Stapleton HM Characterizing the peroxisome proliferator-activated receptor (ppar gamma) ligand binding potential of several major flame retardants, their metabolites, and chemical mixtures in house dust Environmental Health Perspectives 2015 123 166 172 25314719 Farhat A Crump D Chiu S Williams KL Letcher RJ Gauthier LT In ovo effects of two organophosphate flame retardants—TCPP and TDCPP—on pipping success, development, mrna expression, and thyroid hormone levels in chicken embryos Toxicol Sci 2013 134 92 102 23629516 Farhat A Buick JK Williams A Yauk CL O’Brien JM Crump D Tris(1,3-dichloro-2-propyl) phosphate perturbs the expression of genes involved in immune response and lipid and steroid metabolism in chicken embryos Toxicology and Applied Pharmacology 2014 275 104 112 24407104 Faust JB August LM Evidence on the carcinogenicity of tris(1,3-dichloro-2-propyl) phosphate 2011 Sacramento, ca Reproductive and cancer hazard assessment branch, office of environmental health hazard assessment, california environmental protection agency Gold MD Blum A Ames BN Another flame retardant, tris-(1,3-dichloro-2-propyl)-phosphate, and its expected metabolites are mutagens Science 1978 200 785 787 347576 Hammel SC Hoffman K Webster TF Anderson KA Stapleton HM Measuring personal exposure to organophosphate flame retardants using silicone wristbands and hand wipes Environ Sci Technol 2016 50 4483 4491 26975559 Harrada S Brommera S Mueller JF Concentrations of organophosphate flame retardants in dust from cars, homes, and offices: An international comparison Emerging Contaminants 2016 2 2 66 72 Hayes SM Aylward LL Blount BC Variation in urine flow rates according to demographic characteristics and body mass index in NHANES; Potential confounding of associations between health outcomes and urinary biomarker concentrations Environ Health Perspect 123 4 293 300 Hoffman K Daniels JL Stapleton HM Urinary metabolites of organophosphate flame retardants and their variability in pregnant women Environ Int 2014 63 169 172 24316320 Hoffman K Butt CM Chen A Limkakeng AT Jr Stapleton HM High exposure to organophosphate flame retardants in infants: Associations with baby products Environ Sci Technol 2015a 49 14554 14559 26551726 Hoffman K Garantziotis S Birnbaum LS Stapleton HM Monitoring indoor exposure to organophosphate flame retardants: Hand wipes and house dust Environ Health Perspect 2015b 123 160 165 25343780 Kosarac I Kubwabo C Foster WG Quantitative determination of nine urinary metabolites of organophosphate flame retardants using solid phase extraction and ultra performance liquid chromatography coupled to tandem mass spectrometry (UPLC-MS/MS) J Chromatogr B 2016 1014 24 30 Liu X Ji K Choi K Endocrine disruption potentials of organophosphate flame retardants and related mechanisms in h295r and mvln cell lines and in zebrafish Aquat Toxicol 2012 114–115 173 181 Liu X Ji K Jo A Moon HB Choi K Effects of tdcpp or tpp on gene transcriptions and hormones of hpg axis, and their consequences on reproduction in adult zebrafish (danio rerio) Aquat Toxicol 2013 134–135 104 111 Meeker JD Cooper EM Stapleton HM Hauser R Exploratory analysis of urinary metabolites of phosphorus-containing flame retardants in relation to markers of male reproductive health Endocrine Disruptors 2013a Meeker JD Stapleton HM House dust concentrations of organophosphate flame retardants in relation to hormone levels and semen quality parameters Environ Health Perspect 2010 118 3 318 323 20194068 Mendelsohn E Hagopian A Hoffman K Butt CM Lorenzo AM Congleton J Webster TF Stapleton HM Nail Polish as a Source of Exposure to Triphenyl Phosphate Environ Int 2016 86 Patisaul HB Roberts SC Mabrey N McCaffrey KA Gear RB Braun J Accumulation and endocrine disrupting effects of the flame retardant mixture Firemaster® 550 in rats: An exploratory assessment J Biochem Mol Toxicol 2013 27 124 136 23139171 Pillai HK Fang M Beglov D Kozakov D Vajda S Stapleton HM Ligand binding and activation of ppargamma by Firemaster® 550: Effects on adipogenesis and osteogenesis in vitro Environ Health Perspect 2014 122 1225 1232 25062436 PIN Pin — pregnancy, infection, and nutrition study 2012 Available: http://www.cpc.unc.edu/projects/pin Stapleton HM Allen JG Kelly SM Konstantinov A Klosterhaus S Watkins D Alternate and new brominated flame retardants detected in U.S. house dust Environ Sci Technol 2008 42 6910 6916 18853808 Stapleton HM Klosterhaus S Eagle S Fuh J Meeker JD Blum A Detection of organophosphate flame retardants in furniture foam and U.S. house dust Environ Sci Technol 2009 43 7490 7495 19848166 Stapleton HM Klosterhaus S Keller A Ferguson PL van Bergen S Cooper E Identification of flame retardants in polyurethane foam collected from baby products Environ Sci Technol 2011 45 5323 5331 21591615 Stapleton HM Eagle S Sjodin A Webster TF Serum PBDEs in a north carolina toddler cohort: Associations with handwipes, house dust, and socioeconomic variables Environ Health Perspect 2012a 120 1049 1054 22763040 Stapleton HM Sharma S Getzinger G Ferguson PL Gabriel M Webster TF Novel and high volume use flame retardants in us couches reflective of the 2005 pentaBDE phase out Environ Sci Technol 2012b 46 13432 13439 23186002 Su G Letcher RJ Yu H Determination of organophosphate diesters in urine samples by a high-sensitivity method based on ultra high pressure liquid chromatography-triple quadrupole-mass spectrometry J Chromatogr A 2015 1426 154 160 26686560 Van den Eede N Dirtu AC Neels H Covaci A Analytical developments and preliminary assessment of human exposure to organophosphate flame retardants from indoor dust Environ Int 37 454 461 21176966 Van den Eede N Neels H Jorens PG Covaci A Analysis of organophosphate flame retardant diester metabolites in human urine by liquid chromatography electrospray ionisation tandem mass spectrometry J Chromatogr A 2013 1303 48 53 23849782 Van den Eede N Heffernan AL Aylward LL Hobson P Neels H Mueller JF Age as a determinant of phosphate flame retardant exposure of the australian population and identification of novel urinary PFR metabolites Environ Int 2015 74 1 8 25277340 Van der Veen I de Boer J Phosphorus flame retardants: Properties, production, environmental occurrence, toxicity and analysis Chemosphere 2012 88 1119 1153 22537891 Wang Q Liang K Liu J Yang L Guo Y Liu C Exposure of zebrafish embryos/larvae to tdcpp alters concentrations of thyroid hormones and transcriptions of genes involved in the hypothalamic-pituitary-thyroid axis Aquat Toxicol 2013 126 207 213 23220413 Xu F Giovanoulis G van Waes S Padilla-Sanchez JA Papadopoulou E Comprehensive Study of Human External Exposure to Organophosphate Flame Retardants via Air, Dust, and Hand Wipes: The Importance of Sampling and Assessment Strategy Environ Sci Technol 2016 50 7752 7760 27350238

PMC005xxxxxx/PMC5127740.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 1246405 2835 Cell Immunol Cell. Immunol. Cellular immunology 0008-8749 1090-2163 27622385 5127740 10.1016/j.cellimm.2016.09.004 NIHMS815976 Article PI3K signaling in Leishmania infections Kima Peter E. Department of Microbiology and Cell Science, University of Florida For Correspondence: Peter E. Kima, Microbiology and Cell Science, University of Florida, Building 981, Box 110700, Gainesville FL 32669, pkima@ufl.edu 13 9 2016 7 9 2016 11 2016 01 11 2017 309 1922 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. PI3K signaling plays a role in the host response to Leishmania infections. At the cellular level PI3K signaling is engaged by the parasite to control several cellular processes, which ensures parasite persistence. At the systemic level, there is evidence that recruitment of regulatory cells into lesions is impaired in the absence of robust PI3K signaling. In this mini-review the more recent studies that investigated the roles of PI3K signaling in Leishmania infections are discussed. 1. Overview of Phosphatidyl-Inositol 3-Kinase (PI3K) signaling Amongst other head groups of phospholipids in membranes, the inositol head group plays an important role in signal transduction. Phosphorylation at positions 3 to 5 of the inositol ring transforms it into substrates to which lipid binding proteins are recruited and phosphorylated. These phosphorylated proteins in turn phosphorylate downstream molecules thus initiating a signal transduction cascade. Three classes of enzymes called phosphatidyl-inositol (PtdIns) 3 kinases phosphorylate the 3 position of the inositol ring (summarized in Table 1). These enzymes differ in their structure, their preferred substrate and the stimuli that activate them. The substrate of the Class III enzymes is an unphosphorylated inositol ring, which upon phosphorylation at the 3 position becomes PtdIns -3-phosphate (also known as PI3P). The substrates for the prototypic Class III enzyme, vps34, are primarily found on internal membranes that delimit organelles, which explains why they play important roles in vesicular trafficking (1). The Class II enzymes phosphorylate PtdIns - 4-phosphate (PtdIns(4)P) converting it to PtdIns 3,4 phosphate (PtdIns(3,4)P2) (2). These enzymes are the least well understood of the 3 classes. The substrates for Class I enzymes are PtdIns 4,5-phosphate (also known as PIP2) that are found predominantly on the plasma membrane; PIP2 are phosphorylated to PtdIns 3,4,5 phosphate (PIP3) (3). The Class I enzymes are further divided into 2 subclasses that differ in their overall structure and in the stimuli that activate them. The Class Ia enzymes function as heterodimers that include a catalytic subunit and a regulatory subunit. The catalytic subunit of 110kDa is encoded by 3 genes termed pik3ca (gene product is PI3Kα), pik3cb (gene product is PI3Kβ), and pik3cd (gene product is PI3Kδ). The regulatory or adaptor subunit of 85kDa is encoded by 4 genes Pik3r1, PIk3r2, Pik3r3 (gene products p85α, p85β and p85γ; alternative transcription initiation from Pik3r gives rise p55α, p50α and p55γ). The Class Ia kinases are recruited via the Src homology domains on the regulatory subunit to receptor tyrosine kinases. These tyrosine kinases may be associated with growth hormone receptors or antigen receptors. The Class 1b kinases are composed of a single member, PI3Kγ, which is homologous to the p110 catalytic subunit of the Class 1a enzymes. Unlike the Class Ia enzymes, Class Ib enzymes associate with adaptors different from the p85 subunit and are activated primarily by G-protein coupled receptors. Only the contributions of the Class I PI3kinases have been evaluated in Leishmania infections. Phosphorylation of the 3 position of the inositol ring is also dependent on the action of phosphatases. The phosphatase and tensin homologue deleted on chromosome 10 (PTEN) dephosphorylates PtdIns at the 3 position by converting PI(3,4,5)P3 to PI(4,5)P2 (4). Removal of the 3-phosphate on the inositol ring down modulates PI3K mediated signaling. Another relevant phosphatase is the SH2 domain-containing inositol 5-phosphatase type 2 (SHIP2) that converts PI(3,4,5)P3 to PI(3,4,)P2 (5). The role of PTEN has been evaluated in Leishmania infections as well. All three PtdIns substrates that are phosphorylated at position 3 of the inositol ring PI3P, PI(3,4,)P2 and PI(3,4,5)P3 recruit proteins with distinct lipid binding domains. Molecules with a pleckstrin homology domain such as the Ser/Thr kinase protein kinase B (PKB)/Akt and Tyr protein kinases BTK, bind to PI(3,4,5)P3 and PI(3,4,)P2 (3), while molecules with a PX homology domain or FYVE domain such as p40-phox or the early endosomal antigen 1 (EEA1), respectively bind to PI3P (Figure 1). Differential recruitment of these lipid binding proteins determines the signal transduction scheme and consequently the outcome of the activation of PI3K signaling. 2. PI3K signaling in Leishmania infections In Leishmania infections the contributions of PI3K signaling to the host response has been evaluated by assessing the role of molecules at critical points of the signal transduction scheme. Some studies have evaluated the role of individual PI3K catalytic subunits or the regulatory subunits in Leishmania infections. More recent studies that have focused the p110δ and p110γ catalytic subunits have shown that impaired function of either of these enzymes results in limited Leishmania infections. Studies of the role of molecules at the next level of PI3K signaling have focused primarily on the role of PKB/Akt; activation of this molecule mediates parasite persistence by promoting cell survival, production of leishmanicidal responses such as NO and secretion of anti-inflammatory cytokines. Beyond PKB/Akt signaling some studies have evaluated the role of the target of rapamycin (mTOR) and glycogen synthase kinase. Finally, the role of the phosphatase PTEN has also been evaluated in Leishmania infections. In light of the recent review by Lambertz et al (2012) (6), this review will primarily focus on newer studies. 2.1 Role of the catalytic subunits of PI3K enzymes The laboratories of Uzonna and Satoskar have each independently and together, explored the role of either p110δ or p110γ in Leishmania infections. They have exploited the availability of mice deficient in either of these enzymes and also the availability of a specific inhibitor to p110γ to explore how these molecules control the progress of experimental infections by several Leishmania species. The P110δ catalytic subunit is expressed more abundantly by hematopeoitic cells including B cells and T cells. A number of studies from Vanhaesebroeck’s laboratory had shown that p110δ plays essential roles in B cell and T cell development and in their activation via the antigen receptor (7, 8); moreover they showed that in mice in which p110δ function was compromised by insertion of an inactivating knock-in mutation in the p110δ (referred to as p110δ(D910A)) the function of CD4+CD25+Foxp3+ regulatory T cells (T-regs) was altered. In light of a large number of studies that had established an essential role for CD4+ T cell subsets in resistance or susceptibility in the L. major experimental Leishmania model, Uzonna and colleagues (9,10) assessed the progress of L. major infections in the p110δ(D910A) mice. Whether on ‘resistance’ or ‘susceptible’ backgrounds, L major infections exhibited reduced lesion size and accelerated parasite clearance. In additional experiments that included adoptive transfer of enriched T-regs from wild-type mice into p110δ(D910A) mice they showed that increased resistance was due to impaired expansion and effector functions of Tregs. In a follow-up study, Khadem et al (2014) (11) showed that infection of p110δ(D910A) mice with L. donovani parasites, the causative agents of visceral leishmaniasis, also resulted in a limited infection course that presented as reduced splenomegaly and hepatomegaly and limited parasitemia in these organs. They also showed that in L. donovani infections as well, Treg expansion was impaired, which was interpreted to mean that loss of Treg function was the major determinant of resistance in the D910A mice infected with L. donovani. Another PI3K subunit, the p110γ, which is the lone Class Ib PI3K catalytic has also been evaluated for its role in the progress of experimental Leishmania infections. The studies of Satoskar and colleagues were feasible because of the availability of mice in which the p110γ gene had been deleted (p110γ−/− mouse) and the availability of a specific inhibitor of p110γ, AS-605240 (12,13). The course of infection with L. mexicana, a causative agent of cutaneous leishmaniasis, was found to be suppressed in the p110γ−/− mice as compared to C57BL/6 wild type controls. Treatment of infected mice with AS-605240 had a comparable effect to the p110γ−/− mice. The authors showed that limited infections in these mice was due in part to limited phagocytic uptake of Leishmania parasites by macrophages and neutrophils where loss of PI3K activity impairs phagocytosis. The limited infection course was also due to impaired secretion of TH2 cytokines that otherwise promote susceptibility to leishmaniasis. Moreover, they showed that blockade of p110γ−/− mediated signaling reduced the recruitment of CD4+ FoxP3+ Tregs to the infection lesion site. Taken together, these studies established an important role for PI3K signaling in Leishmania infections. It is plausible to propose that these molecules are great targets for the control of leishmaniasis. In light of the fact that small molecules such as AS-605240 have been and continue to be developed that specifically target the different PI3K subunits, it can be anticipated that studies of inhibitors to PI3K molecules will lead to the control of these infections. 2.2. Role of PKB/AKT in leishmaniais Amongst the PI3K membrane proximal transducers of signaling, PKB/Akt has received the greatest attention. PKB/Akt contains a pleckstrin homology domain through which it binds to PI(3,4,5)P3 or PI(3,4) on membranes. There, PKB/Akt is phosphorylated at Thr308 by the phosphoinositide-dependent kinase 1 (PDK1) and also by several kinases including the mammalian target of rapamycin (mTORC2) at Ser473. PKB/Akt has numerous downstream targets, which upon their phosphorylation, they promote several processes including metabolism, proliferation and cell survival. As stated earlier, Lambertz et al (2012) (6) presented a comprehensive review of studies that demonstrated the critical of PKB/Akt signaling in Leishmania infections. More recent studies include a report by Calegari-Silva et al (2015) (14) who evaluated the role of PKB/Akt in the inhibition of nitric oxide release in L. amazonensis infections. Specifically, they considered the role of PKB/Akt in the activation of NFκB, which has been shown to control the transcription of the iNOS gene. Activation of NFκB can either result in the translocation of p50 and p65 subunits into the nucleus where they positively regulate relevant gene expression or translocation of p50 homodimers instead, which results in suppression of gene expression. They showed that even upon incubation of cells with lipopolysaccharide (LPS) which is a potent activator of NO, the activation of PKB/Akt by L. amazonensis infection, resulted in the translocation into the nucleus of the inhibitory p50 subunits that form p50 homodimers in the nucleus. This led to the suppression of iNOS transcription and consequently NO wasn’t produced in response to LPS in L. amazonensis infection. In another study, Vázquez-López et al (2015)(15) showed the effects of activation of PKB/Akt by Leishmania infection wasn’t limited only to macrophage infections. They showed that in dendritic cells as well, L. mexicana parasites activate PKB/Akt to promote cell survival in response otherwise potent apoptotic inducers, while suppressing the MAP Kinase p38 and JNK. The role of PKB/Akt in the production of or response to cytokines in Leishmania infections is complex. Recently, a study by Muhkerjee et al (2013) (16) reported on an intriguing role for PKB/Akt in the resistance of L. donovani parasites to antimony. They showed that antimony resistance by L. donovani (SbRLD) had a two-stage mechanism of resistance. In the first stage, SbRLD parasites induce the production of IL-10. Interestingly, production of IL-10 was attributable to the presence of a unique glycan with N-acetylgalactosamine as a terminal sugar expressed exclusively by resistant parasites. Moreover, this production of IL-10 wasn’t dependent on PKB/Akt activation. To fully achieve resistance to antimony these parasites induced the increased synthesis of the multi-drug resistance gene (MDR-1), which presumably is necessary for drug efflux from infected cells. They determined that the upregulation of MDR-1 was PI3K/Akt dependent and was largely due to IL-10 that acted in an autocrine fashion. The regulation of IL-10 production by Leishmania is complex, which might explain why several studies have arrived at different conclusions. In support for a role of PI3K/PKB/Akt in the regulation of IL-10 production in Leishmania infections, Nandan et al (2012) (17) evaluated L. donovani infections of primary human monocytes as well human and mouse macrophage cell lines. They showed that infection with L. donovani activated PI3K signaling as evidenced by the appearance of phosphorylated PKB/Akt. Activated PKB/Akt resulted in the phosphorylation of glycogen synthase kinase β (GSKβ), which relieves its suppression of CREB, an IL-10 transcription factor. IL-10 production in infected cells was inhibited by agents that suppress PI3K/PKB/Akt signaling. This confirmed the observations of Ohtani et al (2008) (18) that LPS induced IL-10 production was suppressed by inhibitors of PI3K signaling that prevent phosphorylation of GSKβ, which had a comparable effect as lithium chloride (LiCl) and other direct inhibitors of GSKβ function. Ruhland and Kima, (2009) (19) had also shown that PKB/Akt activation results in suppression of the production of pro-inflammatory cytokines such as IL-12 while promoting the production of anti-inflammatory cytokines such as IL-10. 2.3. Mammalian target of rapamycin (mTOR) in Leishmania infections mTOR is composed of two complexes that exhibit important differences. There is consensus that activation of the mTORC1-mediated signaling pathway is PKB/Akt dependent. Once activated, this complex in turn phosphorylates ribosomal protein S6 kinase (pS6k), eukaryotic initiation factor 4E (eIF4E) and eukaryotic initiation factor binding protein 1(4EBP1), which promote protein translation and cell growth (20). The activation scheme of mTORC2 is less certain; however, this complex apparently plays a role in the full activation of PKB/Akt in light of evidence that it phosphorylates PKB/Akt at Ser473, which promotes cell survival and other functions ascribed to PKB/Akt. Interestingly, mTORC2 is insensitive to rapamycin. Consistent with the fact that mTORC1 can be placed downstream in the PI3K/PKB/Akt signaling pathway, a few reports have provided evidence of mTORC1’s role in cytokine production during Leishmania infections (18, 21). These studies showed that Leishmania infection mediates suppression of IL-12 production in large part through activation of mTORC1, which also activates IL-10 production. Evidence for a more intriguing effect role of mTORC1 in Leishmania infections was presented by Jaramillo et al (2011) (22). They showed that Leishmania inhibit host cell protein synthesis by directly targeting mTORC1. Specifically, parasite expressed GP63 cleaves mTORC1, which results in its inactivation and limited phosphorylation of S6k, eIF4E and 4EBP1. This conclusion was arrived at using mutant parasites in which the GP63 gene had been deleted as well as mice in which the 4EBP1/2 genes were deleted. This later observation suggests that Leishmania control host cell processes not only by modulating the activation of signaling components but also by targeting key intermediates directly for degradation. 3. Outstanding unresolved observations Although it is known that p110δ are activated by receptor tyrosine kinases, while p110γ are activated by GPCRs, it is not known which parasite molecules on promastigotes or amastigotes forms initiate these responses. The preferential sensitivity of Tregs in Leishmania infections to loss of PI3K signaling is not explained Most of the studies that have evaluated PI3K signaling in Leishmania infections have shown that activation of this pathway is sustained well after (days) parasite entry. There isn’t presently a mechanistic explanation for how this is achieved during a Leishmania infection. PKB/AKT activation is known to occur at the plasma membrane where PI3K enzymes catalyze the production of PI(3,4)P and PI(3,4,5)P following the activation of a receptor tyrosine kinase or a G-protein coupled receptor as discussed above. Sustained PKB/AKT activation in Leishmania infections suggests that there is an intracellular membrane that may have components of the plasma membrane or complimentary components that are conducive to PKB/AKT activation. In their study Calegari-Silva et al (2015) (14) observed a reduction in infection in cells in which PKB/AKT levels were suppressed by siRNA treatment. This suggests that Leishmania survival in macrophages, even in the absence of any additional activation stimulus is dependent on activation of PKB/AKT. This suggests that modulation of PKB/AKT levels could be a productive strategy to control these infections. Resolution of these questions will enhance our understanding of the role of this critical signaling pathway PK is funded by NIH grant # 5R21AI115218-02 Figure 1 Phosphoinositide metabolism and PI3K signal transduction scheme. After phagocytic receptor engagement phosphoinositides are phosphorylated by several enzymes. Of interest in this discussion are the enzymes that phosphorylate the 3 position of the inositol ring. The class I and class II kinases promote the production of PtdIns(3,4)P and PtdIns (3,4,5). PKB/Akt is recruited to and binds to PtdIns (3,4)P and PtdIns(3,4,5)P at which point it is phosphorylated by PDK1 and mTORC2. Phosphorylated PKB/Akt has numerous direct targets, some of which are shown. Other downstream targets that PKB/Akt are shown including mTORC1. Table 1 PI3Kinases Enzyme Gene product recruitment Substrate-product Class I Class Ia 110kDa via src homology domain on regulatory subunit to hormone or antigen receptors PtdIns(4,5)P – PtdIns(3,4,5)P  Catalytic subunits  Pik3ca  Pik3cb  Pik3cd Regulatory subunits 85kDa  Pik3r1 P85α  Pik3r2 P85β  Pik3r (Alternative Transcription Initiation) P85γ P55α P50α P55γ Class Ib  Catalytic subunit Pi3kcg (PI3kγ) P110 Activated by G-protein coupled receptors PtdIns(4,5)P –PtgIns (3,4,5)P Adaptor subunit Pik3R5 P101 Pik3R6 P84 Class II Pi3kC2α 190Kda ? PtdIns and PtdIns(4)P – Ptdins(3,4)P Pi3kC2β 190kDa Pi3kC2γ 170kDa Class III Pik3R4 100kDa Recruited to endocytic pathway proteins PtdIns – PtdIns(3)P – Hvps34 (catalytic subunit) 150kDa References: Fry, M. J. (2001), Hawkins et al, (2006) Highlights Disruption of PI3K catalytic subunits results in suppressed Leishmania infections. PKB/Akt activation by Leishmania leads inhibition of NO production. PKB/Akt activation is sustained in Leishmania infections, which implies a non-conventional activation scheme. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. 1 Backer JM The regulation and function of Class III PI3Ks: novel roles for Vps34 Biochem J 410 2008 1 17 18215151 2 Falasca M Maffucci T Role of class II phosphoinositide 3-kinase in cell signalling Biochem Soc Trans 35 2007 211 4 17371240 3 Vanhaesebroeck B Guillermet-Guibert J Graupera M Bilanges B The emerging mechanisms of isoform-specific PI3K signalling Nat Rev Mol Cell Biol 11 2010 329 41 20379207 4 Pulido R PTEN: a yin-yang master regulator protein in health and disease Methods 77–78 2015 3 10 5 Viernes DR Choi LB Kerr WG Chisholm JD Discovery and development of small molecule SHIP phosphatase modulator Med Res Rev 34 2014 795 824 24302498 6 Lambertz U Silverman JM Nandan D McMaster WR Clos J Foster LJ Reiner NE Secreted virulence factors and immune evasion in visceral leishmaniasis J Leukoc Biol 91 2012 887 99 22442494 7 Okkenhaug K Vanhaesebroeck B PI3K in lymphocyte development, differentiation and activation Nat Rev Immunol 4 2003 317 30 8 Okkenhaug K Ali K Vanhaesebroeck B Antigen receptor signalling: a distinctive role for the p110delta isoform of PI3K Trends Immunol 28 2007 80 87 17208518 9 Liu D Zhang T Marshall AJ Okkenhaug K Vanhaesebroeck B Uzonna JE The p110delta isoform of phosphatidylinositol 3-kinase controls susceptibility to Leishmania major by regulating expansion and tissue homing of regulatory T cells J Immunol 183 2009 1921 33 19596993 10 Liu D Uzonna JE The p110 delta isoform of phosphatidylinositol 3-kinase controls the quality of secondary anti-Leishmania immunity by regulating expansion and effector function of memory T cell subsets J Immunol 184 2010 3098 105 20154209 11 Khadem F Mou Z Liu D Varikuti S Satoskar A Uzonna JE Deficiency of p110δ isoform of the phosphoinositide 3 kinase leads to enhanced resistance to Leishmania donovani PLoS Negl Trop Dis 8 2014 e2951 24945303 12 Cummings HE Barbi J Reville P Oghumu S Zorko N Sarkar A Keiser TL Lu B Rückle T Varikuti S Lezama-Davila C Wewers MD Whitacre C Radzioch D Rommel C Seveau S Satoskar AR Critical role for phosphoinositide 3-kinase gamma in parasite invasion and disease progression of cutaneous leishmaniasis Proc Natl Acad Sci U S A 109 2012 1251 1256 22232690 13 Oghumu S Satoskar AR PI3K-γ inhibitors in the therapeutic intervention of diseases caused by obligate intracellular pathogens Commun Integr Biol 6 2013 e23360 23749323 14 Calegari-Silva TC Vivarini ÁC Miqueline M Dos Santos GR Teixeira KL Saliba AM Nunes de Carvalho S de Carvalho L Lopes UG The human parasite Leishmania amazonensis downregulates iNOS expression via NF-κB p50/p50 homodimer: role of the PI3K/Akt pathway Open Biol 2015 150118 26400473 15 Vázquez-López R Argueta-Donohué J Wilkins-Rodríguez A Escalona-Montaño A Aguirre-García M Gutiérrez-Kobeh L Leishmania mexicana amastigotes inhibit p38 and JNK and activate PI3K/AKT: role in the inhibition of apoptosis of dendritic cells Parasite Immunol 37 2015 579 89 26352010 16 Mukherjee B Mukhopadhyay R Bannerjee B Chowdhury S Mukherjee S Naskar K Allam US Chakravortty D Sundar S Dujardin JC Roy S Antimony-resistant but not antimony-sensitive Leishmania donovani up-regulates host IL-10 to overexpress multidrug-resistant protein 1 Proc Natl Acad Sci U S A 110 2013 E575 82 23341611 17 Nandan D Camargo de Oliveira C Moeenrezakhanlou A Lopez M Silverman JM Subek J Reiner NE Myeloid cell IL-10 production in response to Leishmania involves inactivation of glycogen synthase kinase-3β downstream of phosphatidylinositol-3 kinase J Immunol 188 2012 367 378 22140263 18 Ohtani M Nagai S Kondo S Mizuno S Nakamura K Tanabe M Takeuchi T Matsuda S Koyasu S Mammalian target of rapamycin and glycogen synthase kinase 3 differentially regulate lipopolysaccharide-induced interleukin-12 production in dendritic cells Blood 112 2008 635 43 18492954 19 Ruhland A Kima PE Activation of PI3K/Akt signaling has a dominant negative effect on IL-12 production by macrophages infected with Leishmania amazonensis promastigotes Exp Parasitol 122 2009 28 36 19186178 20 Cheekatla SS Aggarwal A Naik S mTOR signaling pathway regulates the IL-12/IL-10 axis in Leishmania donovani infection Med Microbiol Immunol 201 2012 37 46 21567173 21 Dibble CC Cantley LC Regulation of mTORC1 by PI3K signaling Trends Cell Biol 25 2015 545 55 26159692 22 Jaramillo M Gomez MA Larsson O Shio MT Topisirovic I Contreras I Luxenburg R Rosenfeld A Colina R McMaster RW Olivier M Costa-Mattioli M Sonenberg N Leishmania repression of host translation through mTOR cleavage is required for parasite survival and infection Cell Host Microbe 9 2011 331 41 21501832 23 Fry MJ Phosphoinositide 3-kinase signalling in breast cancer: how big a role might it play? Breast Cancer Res 3 2001 304 312 11597319 24 Hawkins PT Anderson KE Davidson K Stephens LR Signalling through Class I PI3Ks in mammalian cells Biochem Soc Trans 34 2006 647 662 17052169

PMC005xxxxxx/PMC5127741.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 9809671 21092 Nat Neurosci Nat. Neurosci. Nature neuroscience 1097-6256 1546-1726 27749828 5127741 10.1038/nn.4417 NIHMS817258 Article Organization of long-range inputs and outputs of frontal cortex for top-down control Zhang Siyu § Xu Min § Chang Wei-Cheng Ma Chenyan Do Johnny Phong Hoang Jeong Daniel Lei Tiffany Fan Jiang Lan Dan Yang * Division of Neurobiology, Department of Molecular and Cell Biology, Helen Wills Neuroscience Institute, Howard Hughes Medical Institute, University of California, Berkeley, CA 94720 * To whom correspondence should be addressed: ydan@berkeley.edu § These authors contributed equally to this work. 1 10 2016 17 10 2016 12 2016 17 4 2017 19 12 17331742 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. Long-range projections from the frontal cortex are known to modulate sensory processing in multiple modalities. Although the mouse has become an increasingly important animal model for studying the circuit basis of behavior, the functional organization of its frontal cortical long-range connectivity remains poorly characterized. Here we used virus-assisted circuit mapping to identify the brain networks for top-down modulation of visual, somatosensory, and auditory processing. The visual cortex is reciprocally connected to the anterior cingulate area, whereas the somatosensory and auditory cortices are connected to the primary and secondary motor cortices. Anterograde and retrograde tracing identified the cortical and subcortical structures belonging to each network. Furthermore, using novel viral techniques to target subpopulations of frontal neurons projecting to the visual cortex versus the superior colliculus, we identified two distinct subnetworks within the visual network. These findings provide an anatomical foundation for understanding the brain mechanisms underlying top-down control of behavior. Long-range projections from the frontal cortex to sensory areas can powerfully modulate sensory processing, which may underlie sensorimotor integration and top-down attentional modulation. In mouse somatosensory cortex, active touch of an object by whiskers evokes large calcium signals in layer 5 pyramidal neuron dendrites, which depend on inputs from the vibrissa motor cortex1. The projection from the vibrissa motor cortex also disinhibits pyramidal neurons by activating vasoactive intestinal peptide-positive (VIP+) interneurons in the somatosensory cortex2. In the auditory cortex, inputs from the motor cortex suppress the auditory responses, through feedforward inhibition mediated by parvalbumin-positive interneurons3,4. In the visual cortex, activating the projection from the cingulate cortex can strongly enhance visual responses, which also depends on disinhibition mediated by local VIP+ neurons5. Thus, top-down modulation of sensory responses by projections from frontal cortical areas is widespread across sensory modalities. To understand how top-down modulation is implemented during behavior, an important step is to delineate the brain networks organized by long-range axonal projections to and from the frontal cortex. Although frontal projections to individual sensory cortical areas have been studied extensively, the brain networks encompassing both cortical and subcortical structures have not been mapped systematically and compared across modalities. Furthermore, for each modality, the corresponding frontal area projects to multiple targets in addition to the sensory cortex3,6–13. While in some cases similar signals are broadcast to multiple downstream targets through axon collaterals, in other cases different signals are transmitted by subgroups of neurons projecting to distinct postsynaptic targets. Examining the relationship between different output pathways is crucial for understanding how the frontal cortex coordinates activity of multiple brain areas to optimize behavior. In addition to divergent outputs, each frontal region also receives convergent inputs from various sources6,7,10,13–15. Mapping these inputs is essential for understanding the neural mechanisms that regulate the frontal cortical activity and shape the top-down signals. In this study, we used several virus-assisted circuit mapping techniques to characterize the long-range inputs and outputs of the frontal cortical regions connected to the visual, somatosensory, and auditory cortices of the mouse. We found largely separate brain networks for visual vs. somatosensory/auditory modalities. Furthermore, within the visual network, we delineated two distinct subnetworks for top-down control by mapping the inputs and outputs of two subsets of frontal neurons that project to the visual cortex vs. the superior colliculus. RESULTS To label the long-range inputs and outputs of each cortical area with fluorescent proteins, we injected various viral vectors into the mouse brain, as detailed below. After histological sectioning and fluorescence imaging, each brain sample was aligned to the Allen Mouse Brain Atlas to facilitate 3D whole-brain visualization and quantitative analyses (Fig. 1, also see Methods). The labeled neurons and axons were detected, and their locations were registered in the reference atlas (Table 1 and Supplementary Table 1). To facilitate data visualization at multiple levels of detail, we also used interactive sunburst diagrams (adapted from Allen Mouse Brain Atlas, http://www.brain-map.org/api/examples/examples/sunburst/) to represent the distribution of labeled inputs and outputs in all brain structures (http://top-down-network.org/). The brain structures are arranged hierarchically from inner to outer circles in the diagram, and the size of each sector represents the percentage of labeling in the corresponding structure. The numerical values can be read out by pointing the cursor, and each region of interest can be expanded with a mouse click. Identification of frontal regions for each sensory modality To identify the frontal regions directly innervating each sensory cortical area, we used rabies virus (RV)-mediated transsynaptic retrograde tracing, which labels monosynaptic inputs to selected starter cells with high specificity16,17. Avian-specific retroviral receptor (TVA), enhanced green fluorescent protein (EGFP), and rabies glycoprotein (RG) were expressed by injecting two Cre-inducible adeno-associated virus (AAV) vectors (AAV2-EF1α-DIO-TVA-EGFP and AAV2-CAG-DIO-RG) into the primary visual (VIS), somatosensory (SS), or auditory (AUD) cortex of CaMKIIα-Cre mice (Fig. 2a). Two weeks later, we injected a modified RV expressing tdTomato (RV-ΔG-tdTomato+EnvA), which only infects cells expressing TVA and requires RG to spread retrogradely to presynaptic cells. Figure 2b (upper panel) shows examples of starter cells (expressing both tdTomato and EGFP) in each injected sensory area. Across brain samples, the starter cells were distributed over large portions of VIS, SS and AUD (Supplementary Fig. 1), with similar laminar distributions among these areas (Supplementary Fig. 2). Transsynaptically labeled presynaptic neurons (expressing tdTomato only) were found in multiple cortical and subcortical regions (Supplementary Fig. 3, Supplementary Movie 1, Supplementary Table 1, and interactive sunburst diagram at http://top-down-network.org/). Within the frontal cortex, neurons transsynaptically labeled from VIS were found primarily in the anterior cingulate area (ACA) and the medial portion of secondary motor cortex (MOs), whereas those presynaptic to SS and AUD were mainly located in the primary motor cortex (MOp) and the lateral portion of MOs (Fig. 2b, lower panel, c,d and Supplementary Fig. 4). In all three modalities, inputs from the frontal cortex arose primarily from layers 2/3 and 5, consistent with previous studies3,5,8. We next mapped the brain regions receiving direct axonal projections from each sensory area by injecting AAV expressing mCherry (AAV2-CaMKIIα-mCherry) into VIS, SS, or AUD of wild-type mice (Fig. 2e–h, Supplementary Fig. 5, Supplementary Movie 2, Supplementary Table 1 and sunburst diagram at http://top-down-network.org/). Within the frontal cortex, labeled axons from VIS were found mostly in ACA and medial MOs18, and those from SS or AUD were distributed primarily in MOp and lateral MOs (Fig. 2f, lower panel, g,h). The spatial correspondence between the axonal projections of each sensory area (Fig. 2g,h) and its retrogradely labeled presynaptic neurons (Fig. 2c,d) indicates strong reciprocity of long-range corticocortical connections10. These anterograde and retrograde tracing experiments also indicate a clear segregation between the frontal areas connected to the visual (ACA and medial MOs) and the somatosensory and auditory cortices (MOp and lateral MOs). Other long-range connections of ACA and MO In addition to sensory cortices, the frontal areas also project to multiple other brain structures. To label these projections, we injected AAV expressing mCherry (AAV2-CaMKIIα-mCherry) into ACA or MO (Fig. 3a–d, Supplementary Fig. 6, Supplementary Movie 3 and Supplementary Table 1). Among the sensory cortical areas, ACA projects extensively to VIS and sparsely to SS and AUD, whereas MO projects extensively to SS, moderately to AUD and VIS, consistent with retrograde tracing from these sensory areas (Fig. 2b–d). Among other cortical areas, ACA projects extensively, and MO moderately, to the posterior parietal cortex (PTLp) and the retrosplenial area (RSP) (Fig. 3c1–c3). While PTLp makes reciprocal connections with VIS, SS, and AUD, RSP is densely connected only with VIS (Supplementary Fig. 3,5), suggesting that it belongs mainly to the visual network13. Within the frontal cortex, the prelimbic/infralimbic area (PL/ILA) receives more input from ACA, and the orbital area (ORB) receives more input from MO (Fig. 3c6). In the thalamus, the projections from ACA and MO were segregated mainly along the dorsal-ventral axis. For example, ACA projects more to the lateral posterior/lateral dorsal (LP/LD) thalamic nuclei, whereas MO projects more to posterior (PO), ventral posterior (VP), and ventral anterior-lateral/ventral medial (VAL/VM) complex10. Both ACA and MO project to the mediodorsal (MD) nucleus, which is densely connected to the prefrontal cortex19. Other major subcortical targets of the frontal regions include the striatum (STR) and superior colliculus (SC, Fig. 3d), with partially overlapping projections from ACA and MO (Fig. 3c1,c5). To map the inputs to these frontal areas, we injected the AAV and RV vectors for transsynaptic retrograde tracing (same as in Fig. 2a) into ACA or MO of CaMKIIα-Cre mice (Fig. 3e–h, Supplementary Fig. 6, Supplementary Movie 4 and Supplementary Table 1). Among the sensory cortices, MO injection led to dense labeling in SS but little labeling in VIS and AUD, whereas ACA injection caused the densest labeling in VIS (Fig. 3h), consistent with anterograde tracing (Fig. 2f–g). Among other cortical areas, we found extensive inputs from PTLp and RSP to ACA but only sparse inputs to MO (Fig. 3g,h), further attesting to their membership in the visual network (Fig. 4). Within the frontal cortex, PL/ILA provides more input to ACA than to MO, and ORB provides similar inputs to both regions. The striking similarity between the cortical distributions of inputs to (Fig. 3g,h) and projections from (Fig. 3c,d) each frontal region again demonstrates the reciprocity of corticocortical connections10. In the thalamus, the distributions of inputs to ACA and MO also largely mirrored the distributions of their axonal projections (Fig. 3c,d). ACA receives more inputs from LP/LD, whereas MO receives more inputs from PO, VP, and VAL/VM (Fig. 3g,h). On the other hand, MD projects to both ACA and MO, as expected for these prefrontal cortical areas19. Since LP/LD is reciprocally connected to both ACA and VIS (Fig. 3c,d,g,h and Supplementary Fig. 3,5), it forms an integral part of the visual network (Fig. 4a); PO, VP, and VAL/VM are connected to both MO and SS, suggesting that they are part of the somatosensory network (Fig. 4b). Among other subcortical regions, the pallidum projects to both ACA and MO (Fig. 3g5,h), consistent with the known cholinergic and non-cholinergic projections from the basal forebrain to the entire cortex10,20. Besides excitatory neurons, several subtypes of inhibitory interneurons have been implicated in long-range corticocortical interactions2–5. To determine whether the long-range axonal projections directly innervate these interneuron subtypes in each cortical area, we injected the AAV and RV vectors for transsynaptic tracing into VIS, RSP, PTLp and ACA of PV-, SOM- and VIP-Cre mice21,22 (Supplementary Table 1 and sunburst diagram at http://top-down-network.org/). Similar to excitatory neurons, all three subtypes of inhibitory neurons in VIS, PTLp and RSP receive monosynaptic inputs from ACA (Supplementary Fig. 7 and Fig. 2b–d), and all of them in ACA receive monosynaptic inputs from VIS, PTLp and RSP (Supplementary Fig. 8 and Fig. 3f–h). Thus, both top-down and bottom-up corticocortical projections directly recruit inhibitory interneurons in their target areas. Outputs of VIS- and SC-projecting ACA neurons Different projections from each cortical area in some cases originate from different subsets of neurons23,24, but in other cases may reflect axon collaterals of the same neurons. Distinguishing these possibilities is crucial for understanding how the frontal cortex coordinates the activity of different brain areas for top-down executive control, but this issue has not been addressed systematically in previous efforts mapping mesoscale mouse brain connectivity. In particular, we wondered what other brain regions are also innervated by the frontal cortical neurons projecting to the sensory cortex. We focused this analysis on the ACA neurons projecting to VIS (ACA→VIS neurons). To label these neurons and their axons, we injected the AAV vector expressing TVA (AAV2-EF1α-DIO-TVA-EGFP) into ACA of CaMKIIα-Cre mice, but the RV vector (RV-ΔG-tdTomato+EnvA) into VIS two weeks after AAV injection (Fig. 5a, upper panel). This allowed RV to enter the TVA-expressing ACA axon terminals in VIS, be transported retrogradely to the ACA neurons, and label all their axon collaterals with tdTomato (Fig. 5b, upper row). To enhance the visibility of labeled thin axons, we performed immunostaining for tdTomato to convert the fluorescence signal into nickel-enhanced diaminobenzidine (DAB) signal. In addition to the dense projection to VIS, we found that the ACA→VIS neurons also project extensively to PTLp and moderately to RSP (Fig. 5c1,c2,e). This suggests that similar modulatory signals are broadcast to VIS, PTLp and RSP, pointing to a tightly coordinated ACA – PTLp – RSP – VIS network for visual processing25 (Fig. 7, Supplementary Movie 5 and Supplementary Table 1). In contrast, the projection from the ACA→VIS neurons to PL/ILA was much sparser than that from the entire ACA population (Fig. 5c5). The thalamus also receives very little inputs from ACA→VIS neurons (Fig. 5c2,c3,e), suggesting that it is not strongly and directly influenced by the modulatory signals sent to VIS. Furthermore, while the SC receives a sizable projection from ACA (Fig. 3d), we found few labeled axons from the ACA→VIS neurons, suggesting that the SC projection originates from a separate ACA neuron population. This is consistent with the finding based on retrograde tracing from VIS and SC in primates26. Targeting the MO→SS neurons using the same technique revealed a similar level of selective axonal projections (Supplementary Fig. 9). The SC is also known to be important in top-down attentional modulation27–31. We thus examined the outputs of the SC-projecting ACA (ACA→SC) neurons by injecting the AAV vector expressing TVA into ACA and RV vector into SC (Fig. 5a, lower panel). Unlike the ACA→VIS neurons, which were distributed in both layers 2/3 and 5 (Fig. 5b, upper panel), the ACA→SC neurons were found primarily in layer 5 (Fig. 5b, lower panel)5. These spatial distributions are similar to those of intratelencephalic and pyramidal tract neurons in the motor cortex, which form non-overlapping populations of projection neurons with distinct roles in motor control23,24,32. We found little projection from the ACA→SC neurons to VIS (Fig. 5d1,e). In addition, the ACA→VIS but not ACA→SC neurons project to the contralateral ACA through the corpus callosum (Supplementary Movie 5), further supporting the correspondence between ACA→VIS/ACA→SC and intratelencephalic/pyramidal tract neurons. Compared to the ACA→VIS neurons, the ACA→SC neurons project much less to PTLp but more to PL/ILA (Figures 5d2,d5,e). Thus, the two subpopulations of ACA neurons show distinct cortical projection patterns. For the thalamic nuclei innervated by ACA, the inputs from ACA→SC neurons were much more extensive than those from ACA→VIS neurons (Fig. 5c–e). In contrast, the striatum receives extensive projections from both ACA→VIS and ACA→SC neurons, and their spatial distributions largely overlap. This is reminiscent of the striatal projections from both intratelencephalic and pyramidal tract neurons in the ipsilateral cortex23,24,32. To assess whether ACA→VIS and ACA→SC neurons form synapses in the identified areas, we expressed membrane-bound GFP (mGFP, for labeling axons) and synaptophysin-mRuby (SYP-mRuby, for labeling putative presynaptic sites) in these neurons33,34. CAV-FLExloxP-Flp was injected into VIS or SC and AAV-hSyn1-FLExFRT-mGFP-2A-synaptophysin-mRuby was injected into ACA of CaMK2α-Cre mice for Flp-dependent expression of mGFP and SYP-mRuby (Supplementary fig. 10). We found synaptophysin-mRuby labeling in all the major cortical and subcortical areas identified above (Fig. 5e), indicating synaptic innervation of those areas. Inputs to VIS- and SC-Projecting ACA Neurons The top-down signals from ACA→VIS and ACA→SC neurons to their distinct postsynaptic targets are determined by their respective inputs. To map the monosynaptic inputs to each subpopulation, we injected AAV expressing the trans-cellular tracer protein wheat germ agglutinin (WGA) fused with Cre recombinase (AAV2-EF1α-mCherry-IRES-WGA-Cre) into VIS or SC of wild-type mice, and AAV vectors with Cre-inducible expression of TVA and RG (same as in Fig. 2a) into their ACA. Four weeks after these AAV injections, RV expressing EGFP was injected into ACA (Fig. 6a). This viral strategy ensured that TVA and RG were expressed only in the ACA→VIS or ACA→SC neurons retrogradely labeled with Cre recombinase35, thus restricting RV labeling to their presynaptic inputs (Supplementary Fig. 11–14). We found that ACA→VIS but not ACA→SC neurons receive extensive monosynaptic inputs from VIS (Fig. 6b–e). Such selective innervation of sensory cortex-projecting neurons was also found in the somatosensory → motor cortex connection (Supplementary Fig. 15)8, resulting in a recurrent loop between the sensory cortex and a subset of frontal cortical neurons. Among other cortical areas, inputs from PL and ILA were much more extensive for ACA→SC neurons, whereas those from PTLp were much more extensive for ACA→VIS neurons (Fig. 6c2,c5,d2,d5,e), matching the distributions of axonal projections of the two ACA subpopulations (Fig. 5c2,c5,d2,d5,e). Together, these findings suggest that the ACA→VIS neurons have enhanced reciprocal connections with the posterior sensory and association areas, whereas the ACA→SC neurons are preferentially connected to the medial prefrontal cortex (Fig. 7, Supplementary Movie 6 and Supplementary Table 1). Thus, within the visual network, there are two subnetworks involving separate populations of ACA neurons. From the thalamus, the ACA→VIS neurons receive more inputs than ACA→SC neurons (Fig. 6e), opposite to the relative strengths of their projections to the thalamus (Fig. 5e). Thus the corticothalamic connections are much less reciprocal than the corticocortical connections (Fig. 7). Finally, we found inputs from the pallidum to ACA→VIS neurons but not to ACA→SC neurons. DISCUSSION Using a variety of virus-assisted circuit tracing techniques, we have mapped the long-range inputs and outputs of the frontal cortical regions that are directly connected to the visual, somatosensory, and auditory cortices. Both anterograde and retrograde tracing from the sensory areas indicate a clear spatial segregation between the frontal neurons connected to the visual cortex (ACA) vs. somatosensory and auditory cortices (MO) (Fig. 2). Anterograde and retrograde tracing from ACA and MO allowed us to delineate separate brain networks associated with different sensory modalities (Figs. 3,4). Furthermore, within the visual network, we identified two distinct subnetworks, involving subpopulations of ACA neurons that project to the visual cortex vs. the superior colliculus (Figs. 5–7). Our anterograde tracing approach is similar to that used for generating the Allen Mouse Brain Connectivity Atlas10, but we focused on the brain networks for top-down modulation of sensory processing. In addition, we have complemented anterograde tracing of axonal projections with RV-mediated retrograde tracing of input neurons. The results of these different tracing strategies are highly consistent with each other. The distributions of thalamic inputs to ACA and MO mapped with RV tracing (Fig. 3h) are also broadly consistent with a recent mapping study using anterograde tracing from the thalamus15. The visual and somatosensory/auditory networks we have identified using viral tracers generally correspond to the medial and somatic sensorimotor networks mapped in the Mouse Connectome Project using non-viral tracers13. However, there are some noticeable differences. While Zingg et al. (2014) grouped the visual and auditory areas into the same medial network and the somatosensory cortex in a separate network, we found that MOp and lateral MOs are connected to both AUD and SS whereas ACA and medial MOs are connected mainly to VIS (Fig. 2). This difference may reflect different emphases of the two studies; while Zingg et al. (2014) performed cluster analysis of all corticocortical connections, our study focused on the frontal-sensory cortex connections. In addition, preferential labeling of different neuronal subtypes by viral vs. non-viral tracers may also contribute to the difference between the two studies. Although the somatosensory and auditory networks overlap spatially in the frontal cortex, they are largely separate in other brain regions. For example, in the thalamus AUD receives extensive inputs from MG but SS mainly from VP, VAL and VM. Even in the frontal cortex, the neurons connected to AUD are likely to be distinct from those connected to SS at the level of single cells. Of course, it is also important to note that in addition to their distinct connections, the visual, somatosensory, and auditory networks also receive shared inputs and project to common targets in both the cortex (e.g., PTLp and ORB) and thalamus (e.g., MD), allowing cross talk between the different sensory modalities. While ACA→VIS neurons make dense reciprocal connections with sensory and association areas, ACA→SC neurons are preferentially connected to PL and ILA (Figs. 5,6). These medial prefrontal areas are known to be important for the control of actions7,36, and the SC is crucial for controlling saccadic eye movement37,38. Thus the two subnetworks may be specialized in different functions, one for sensory perception (RSP&PTLp ↔ ACA ↔ VIS ↔ RSP&PTLp, the “perception subnetwork”) and the other for action control (RSP&PL/ILA ↔ ACA → SC, the “action subnetwork”). Interestingly, we found inputs from the pallidum to ACA→VIS but not ACA→SC neurons (Fig. 6e). Cholinergic and non-cholinergic pallidal neurons project widely to the cortex and play important roles in brain-state-dependent modulation of sensory processing39–43. Their selective innervation of ACA→VIS neurons should allow preferential regulation of the perception subnetwork in a brain state-dependent manner. In contrast, RSP provides dense inputs to and thus may regulate the activity of both subnetworks. In both frontal and sensory cortices, intratelencephalic (IT) neurons provide extensive inputs to, but receive little innervation from, the pyramidal tract (PT) neurons24,32,44–46, suggesting a non-reciprocal connection from ACA→VIS to ACA→SC neurons. IT and PT neurons also differ in other cellular properties, e.g. with greater hyperpolarization-activated current (Ih) and faster action potentials in PT than IT neurons24. Interestingly, ACA→SC neurons provide much more projections to the thalamus (Fig. 5e), whereas ACA→VIS neurons receive more thalamic inputs (Fig. 6e). This points to a largely unidirectional thalamocortical loop for the interaction between ACA→VIS and ACA→SC neurons and between the two subnetworks (Fig. 7, blue lines). Note that the mouse LP is densely connected to VIS (Supplementary Fig. 3,5) and thought to be functionally analogous to the primate pulvinar10, which powerfully controls the responses in visual cortex47. The ACA→SC → LP projection (Fig. 5d2) may thus provide an additional pathway for top-down modulation of visual cortical processing. In the somatosensory network, the MO → PO → SS pathway (Fig. 3d and Supplementary Fig. 3d,f) may serve a similar function. However, deep-layer MO→AUD neurons were found to project to thalamus and brainstem motor regions3, suggesting a different organization of the auditory network. The long-range projections from ACA to both VIS and SC suggest a strong analogy between the mouse ACA and the primate frontal eye field (FEF)13. Optogenetic activation of ACA markedly enhances visual performance of the mouse and neuronal responses in VIS5, similar to the effect of FEF stimulation on attentional modulation in primates9,48,49. In the rat, a frontal orienting field has also been identified, whose activation can bias the orientating response, and presumably attention, to the contralateral side50. The current study indicates that the mouse ACA is a point of convergence between the visual perception and action subnetworks. Such anatomical characterization provides a blueprint for future physiological investigation of how each subnetwork contributes to top-down attentional modulation and behavioral control. METHODS Virus preparation AAV preparation followed previously reported protocol51. To construct AAV-EF1α-DIO-TVA-EGFP, TVA and EGFP linked by the 2A ‘self-cleaving’ peptides or rabies glycoprotein was cloned into pAAV-MCS in an antisense direction flanked by a pair of canonical loxP sites and a pair of lox2272 sites. TVA was subcloned from the AAV-TRE-HTG plasmid from L. Luo16. The AAV-CAG-DIO-Glycoprotein and AAV-CAG-DIO-TVA-mCherry vector was from Addgene (Plasmid #48333 and #48332)52. AAV particles (AAV2/2) were produced by co-transfection of packaging plasmids into HEK293T cells, and cell lysates were fractionated by iodixanol gradient ultracentrifugation. Viral particles were further purified from the crude fraction by heparin affinity column, desalted and concentrated with centrifugal filter (100K). The genomic titer of AAV2/2-EF1α-DIO-TVA-EGFP (4.4 × 1013 gc/mL), AAV2/2-CAG-DIO-TVA-mCherry (3.1 × 1012 gc/mL) and AAV2/2-CAG-DIO-Glycoprotein (8.7 × 1012 gc/mL) was determined by quantitative PCR. AAV2/2-CaMKIIα-mCherry (6.6 × 1012 gc/mL), AAV2/2-EF1α-mCherry-IRES-WGA-Cre (2.7 × 1012 gc/mL) and AAV2/2-EF1α-DIO-EYFP (4.2 × 1012 gc/mL) were purchased from the UNC Vector Core (Chapel Hill, NC). CAV-FLExloxP-Flp (5.0 × 1012 gc/mL) and AAV-DJ-hSyn1-FLExFRT-mGFP-2A-synaptophysin-mRuby (2.9 × 1013 gc/mL) were obtained from Stanford University (kind gift from Dr. Liqun Luo). Glycoprotein-deleted (ΔG) and EnvA-pseudotyped rabies virus (RV-ΔG-tdTomato+EnvA) was used for retrograde monosynaptic tracing from sensory and frontal cortical pyramidal neurons53,54. RV-ΔG-tdTomato (1.5 × 109 IU/mL) and RV-ΔG-EGFP (5 × 108 IU/mL) were amplified in B7GG cells and pseudotyped using BHK-EnvA cells in a manner similar to that previously described by Osakada and Callaway (2013)55. EnvA pseudotyped rabies virus was titered using HEK293-TVA cells. RV-ΔG-tdTomato was a gift from B. Lim. B7GG cells, BHK-EnvA cells and HEK293-TVA cells were gifts from E. Callaway. Animals and surgery All experimental procedures were approved by the Animal Care and Use Committee at the University of California, Berkeley. Experiments were performed on wild-type (C57), CaMKIIα-Cre (Jackson lab stock #005359), PV-Cre (#008069), SOM-Cre (#013044) and VIP-Cre (#010908) mice. Mice (>P60) were anesthetized with isoflurane (5% induction and 1.5% maintenance) and placed on a stereotaxic frame. Temperature was kept at 37 °C throughout the procedure using a heating pad. After asepsis, the skin was incised to expose the skull and the overlying connective tissue was removed. A craniotomy (~0.5 mm diameter) was made above the injection site. Viruses were loaded in a sharp micropipette mounted on a Nanoject II attached to a micromanipulator and then injected at a speed of 60 nL per min. For retrograde monosynaptic tracing, TVA receptor and rabies glycoprotein, which are required for virus infection and trans-synaptic spread, respectively, were expressed in Cre-positive neurons by co-injection of AAV2/2-EF1α-DIO-TVA-EGFP / AAV2/2-CAG-DIO-TVA-mCherry and AAV2/2-CAG-DIO-Glycoprotein (200–500 nL) into VIS (Bregma -3 mm, lateral 2.5 mm, depth 0.5 mm), SS (Bregma -1 mm, lateral 3 mm, depth 0.8 mm), AUD (Bregma -2.5 mm, lateral 4 mm and depth 0.5 mm), RSP (Bregma -1.8 mm, lateral 0.3 mm, depth 0.5 mm), PTLp (Bregma -1.8 mm, lateral 1.2 mm, depth 0.5 mm), ACA (Bregma +0.5 mm, lateral 0.3 mm, depth 0.9 mm) or MO (Bregma +1.5 mm, lateral 1.2 mm, depth 0.8 mm) of CaMKIIα-Cre mice and into VIS, RSP, PTLp and ACA of PV-Cre, SOM-Cre and VIP-Cre mice. Two weeks later, RV-ΔG-tdTomato+EnvA or RV-ΔG-EGFP+EnvA (500 nL) was injected into the same site as AAV injection. The histology experiments were performed 7 days after rabies virus injection. For anterograde tracing from cortical projection neurons, AAV2/2-CaMKIIα-mCherry (200–500 nL) was injected into VIS, SS, AUD, ACA or MO of wild type mice. For anterograde tracing from subgroups of frontal cortical projection neurons, AAV2/2-EF1α-DIO-TVA-EGFP (4.4 × 1011 gc/mL, 500 nL) was injected into ACA or MO of CaMKIIα-Cre mice. Two weeks later, RV-ΔG-tdTomato+EnvA (500 nL) was injected into VIS or SC in ACA injected mice and SS in MO injected mice. For axon arborization experiments, CAV-FLExloxP-Flp was injected into VIS or SC, and AAV-DJ-hSyn1-FLExFRT-mGFP-2A-synaptophysin-mRuby was injected into Cg of CaMKIIα-Cre mice. The histology experiments were performed 7–8 weeks after the injection. For retrograde monosynaptic tracing from subgroups of frontal cortical projection neurons, AAV2/2-EF1α-DIO-TVA-EGFP and AAV2/2-CAG-DIO-Glycoprotein (500 nL) were co-injected into ACA or MO of wildtype mice. At the same time AAV2/2-EF1α-mCherry-IRES-WGA-Cre (500 nL) was injected into VIS or SC in ACA injected mice and SS in MO injected mice. Four weeks later, RV-ΔG-EGFP+EnvA (500 nL) was injected into ACA or MO. For the control experiment on the distribution of WGA-Cre-labeled ACA→VIS and ACA→SC neurons, AAV2/2-EF1α-DIO-EYFP (500 nL) was injected into ACA, and AAV2/2-EF1α-mCherry-IRES-WGA-Cre (500 nL) were injected into VIS or SC (Supplementary Fig. 11). For the control experiment testing the specificity of WGA-Cre-mediated labeling of ACA→VIS and ACA→SC neurons, AAV2/2-EF1α-DIO-EYFP (500 nL) was injected into ACA, and AAV2/2-EF1α-mCherry-IRES-WGA-Cre (500 nL) and red retrobeads (150 nL) were injected into VIS or SC (Supplementary Fig. 12–13). For the control experiment measuring the spatial extent of RV-labeling in ACA without rabies glycoprotein18, AAV2/2-EF1α-DIO-TVA-EGFP (500 nL) was injected into ACA, and AAV2/2-EF1α-mCherry-IRES-WGA-Cre (500 nL) was injected into VIS or SC of wildtype mice. Four weeks later, RV-ΔG-EGFP+EnvA (500 nL) was injected into ACA (Supplementary Fig. 14). Histology The mice were deeply anesthetized with isoflurane and immediately perfused with chilled 0.1 M PBS followed by 4% paraformaldehyde (wt/vol) in PBS. The brain was removed and post-fixed overnight at 4 °C. After fixation, the brain was placed in 30% sucrose (wt/vol) in PBS solution for 1–2 d at 4 °C. After embedding and freezing, the brain was sectioned into 50 μm coronal slices using a cryostat. For fluorescence images, brain slices were washed with PBS for 0.5 hr and mounted with VECTASHIELD mounting medium with DAPI. For immunohistochemistry for tdTomato, brain slices were washed with PBS for 0.5 hr, quenched with 3% H2O2 for 0.5 hr, permeabilized using PBST (0.3% Triton X-100 in PBS) for 0.5 hr and then incubated with blocking solution (2% normal goat serum in PBST) for 1 hr followed by primary antibody incubation overnight at 4 ºC using anti-mCherry rat monoclonal antibody (M11217, Life Technologies; 1:1000). The next day, slices were washed three times with PBS, incubated with biotinylated secondary antibody (biotin-goat anti-rat IgG, 629540, Life Technologies; 1:1000) for 2 hrs and then incubated with VECTASTAIN→ ABC Reagent (PK-6100, Vector labs) overnight at 4 ºC. The next day, the slices were washed three times with PBS, incubated with DAB peroxidase substrate (SK-4100, Vector labs) for ~10 mins, washed with PBS again and then mounted with VECTASHIELD mounting medium. One out of every three sections were imaged using 20×/0.75 objective in a high-throughput slide scanner (Nanozoomer-2.0RS, Hamamatsu) for further processing. We also imaged selected example slices under a confocal microscope (Zeiss, LSM 710). 3D reconstruction and quantification A software package was developed in Matlab to analyze the digitized brain images. The analysis software consists of three modules: image registration, signal detection, and quantification/visualization. Registration module The registration module is a reference point-based image alignment software used to align images of brain sections to the Allen Mouse Brain Atlas for further quantification and 3D reconstruction. First, we manually selected a set of reference points in both the atlas and the brain image. The module then applied several geometric transformations (translation, rotation and scaling) of the brain section to optimize the match of the reference points between the brain image and the atlas. Since histological sectioning can sometimes cause tissue compression, we allowed the scaling factors along the dorsal-ventral and medial-lateral axes to be optimized independently. Following the transformation, the match between the image and the atlas was inspected, and further adjustments were made manually if necessary. Detection module The detection module has two independent sub-modules designed for counting RV-labeled cells and detecting axons, respectively. The cell counting module records the position of manually identified tdTomato-labeled neurons in each digitized brain section image. For axon detection, the ridge detection method was used (http://en.wikipedia.org/wiki/Ridge_detection). The following steps were taken to maximize the detection accuracy: (1) Image ridges were computed across multiple scales to extract all possible axon-like signals from each image. In the resulting binary ‘ridge image’, the number of pixels occupied by each detected axon depends on the length but not the thickness of the axon. In addition to valid axons, the ridge image also contains many noise pixels. (2) To remove the noise pixels due to the general background in the fluorescence image, we set a threshold based on the intensity distribution of the original image, and use this as a mask to remove the noise pixels in the ridge image obtained from step (1). (3) To remove the discrete noise pixels with fluorescence intensities higher than the general background (thus not removed by step 2), we first identified pixels that are spatially contiguous in the ridge image (after a spatial convolution with a Gaussian kernel), computed the size of each contiguous region, and removed the regions (of the original ridge image) below a threshold size. Steps 2 and 3 were repeated until satisfactory detection results were achieved. (4) The results were then visually inspected and the remaining noise pixels, which were mostly artifacts introduced during brain tissue processing, were rejected manually. Quantification/visualization module After detection and registration, signals were quantified across the whole brain and projected to the 3D reference atlas for better visualization. The 3D viewer plug-in of the ImageJ software was used to animate the final 3D model. The atlas, 3D reference mouse brain, quantification ontology, and layouts for sunburst plot were obtained from the open online resource of Allen Institute for Brain Science, licensed under the Apache License (Version 2.0). The input from each region was quantified by dividing the number of labeled neurons in that region by the total number of labeled neurons detected in the whole brain (excluding the injection site). The output (axon projection) to each region was quantified as the number of pixels occupied by detected axons in the cleaned ridge image (see Detection module above) divided by the total number of axon-occupied pixels in the entire brain (excluding the injection site and locations with known major fiber tracks). In addition, the density of labeled neurons and axons (number of neurons/length of axons divided by volume) in each structure was computed (Supplementary Table 1). To assess the data variability, we have computed the correlation coefficients (CCs) between individual brain samples for both input and output distributions. The mean CC between individual samples of the same group was 0.90±0.02 (s.e.m.). For the same brain sample, the CC between the whole-brain distributions of axons detected by two different observers was 0.96, and the CC between the whole-brain distributions of RV-labeled cells detected by two observers was >0.99. Code availability The atlas, 3D reference mouse brain, quantification ontology, and layouts for sunburst plot are freely available in the open online resource of Allen Institute for Brain Science, licensed under the Apache License (Version 2.0). The other computer codes used to generate the findings of this study are available from the corresponding author upon request. Statistical analyses A supplementary methods checklist is available summarizing statistical tests and results. Data collection and analysis were not performed blind to the conditions of the experiments. Data randomization was not applicable to our study, and no statistical methods were used to predetermine sample sizes but our sample sizes are similar to those reported in previous publications16–17. Data availability The data that support the findings of this study are available from the corresponding author upon request. Supplementary Material supp_figure supp_table vid1 Movie S1. Whole-Brain Distributions of RV-Labeled Neurons Presynaptic to VIS, SS and AUD. vid2 Movie S2. Whole-Brain Distributions of Axons from VIS, SS and AUD. vid3 Movie S3. Whole-Brain Distributions of Axons from ACA and MO. vid4 Movie S4. Whole-Brain Distributions of RV-Labeled Neurons Presynaptic to ACA and MO. vid5 Movie S5. Whole-Brain Distributions of Axons from ACA→VIS and ACA→SC Neurons. vid6 Movie S6. Whole-Brain Distributions of RV-Labeled Neurons Presynaptic to ACA→VIS and ACA→SC Neurons. We thank B. Kim, S. Zhu and P. Kim for technical help. We thank L. Luo, K. Beier, B. Lim and E. Callaway for viral vectors and cell lines. We thank T. Kamigaki for helpful discussion. This work was supported by NIH R01 EY018861. Figure 1 Steps for data processing. (a) Mapping of raw image (upper panel) onto corresponding coronal section of Allen Mouse Brain Atlas (lower panel). The registration module applied several geometric transformations (translation, rotation and scaling) of the raw image to optimize the match of reference points (yellow crosses) between raw image and atlas. (b) Fluorescence signals (right) detected from raw image (left). Retrogradely labeled neurons were identified manually (red crosses). Anterogradely labeled axons were detected using a ridge detection method (white lines, see Methods). (c) Detected signals were projected to Allen Mouse Brain Atlas and quantified as both the percentage and density of labeling in each brain region (Supplementary table 1). Figure 2 Mapping connections between sensory and frontal cortices. (a) Viral vectors and injection procedure for RV-mediated transsynaptic retrograde tracing from sensory cortices. (b) Upper panel, injection sites in VIS, SS and AUD (scale bar, 1 mm). Inset, enlarged view of region in white box; Starter cells, yellow; scale bar, 40 μm. Lower panel, fluorescence images of ACA and MO (yellow box in coronal diagram) showing RV-labeled input neurons (red) to each sensory area (scale bar, 200 μm). Inset, enlarged view of region in white box (scale bar, 40 μm). Green, EGFP; red, tdTomato; blue, DAPI. (c) Percentages of input neurons in ACA, MOs, MOp retrogradely labeled from VIS (blue, n = 3 mice), SS (red, n = 3), AUD (green, n = 3). Each circle represents one mouse. Error bar, ±s.e.m. (d) Summary of RV-labeled neurons in all samples of each group (scale bar, 1 mm). Dots, detected neurons. White masks, injection sites excluded from analysis. (e) Viral vector and injection procedure for tracing the axonal projections from each sensory area. (f) Upper panel, injection sites in VIS, SS, AUD (scale bar, 1 mm). Lower panel, fluorescence images of ACA and MO (yellow box in coronal diagram) showing axons from each sensory area (scale bar, 200 μm). Red, mCherry; blue, DAPI. (g) Percentages of labeled axons in ACA, MOs, MOp from VIS (n = 3), SS (n = 3), AUD (n = 3). Each circle represents one mouse. Error bar, ±s.e.m. (h) Summary of axons detected in all samples of each group (scale bar, 1 mm). White masks, injection sites excluded from analysis. Figure 3 Whole-brain distributions of axonal projections and input neurons of ACA and MO. (a) Injection procedure for tracing projections. (b) Fluorescence images of ACA and MO (red box in coronal diagram) at injection sites (scale, 200 μm). Red, mCherry; blue, DAPI. (c) Axons detected in all samples of each group (MO, yellow; ACA, magenta; Overlap, white. Scale, 1 mm). White masks, injection sites excluded from analysis. (d) Percentages of labeled axons in cortical and subcortical structures (MO, n = 3; ACA, n = 3). Included are cortical areas with >1% labeling and thalamic structures with >0.8% labeling. Error bar, ±s.e.m. (e) Injection procedure for RV-mediated retrograde tracing. (f) Fluorescence images of ACA and MO (red box in coronal diagram) at injection sites (scale, 200 μm). Inset, enlarged view of region in white box showing starter cells (yellow; scale, 10 μm). Green, EGFP; red, tdTomato; blue, DAPI. (g) RV-labeled neurons detected in all samples of each group (MO, yellow; ACA, magenta. Scale, 1 mm). White masks, injection sites excluded from analysis. (h) Percentages of retrogradely labeled neurons in selected cortical and subcortical regions (MO, n = 3; ACA, n = 3). Included are areas with >2% (cortical) or >1% (thalamic) labeling. Error bar, ±s.e.m. Figure 4 Schematic diagram of visual and somatosensory networks. Shown are major connections in each network. Figure 5 Whole-brain distributions of axonal projections from ACA→VIS and ACA→SC neurons. (a) Viral vectors and injection procedure. (b) Upper panel: left, bright field image of ACA and MO showing RV-labeled neurons from VIS (scale, 200 μm). Inset, enlarged view of region in red box (scale, 20 μm). Immunostaining for tdTomato was performed to convert fluorescence signal (tdTomato expressed by RV) into nickel-enhanced DAB signal; right, images of VIS and PTLp (red box in coronal diagram), showing RV-labeled axons of ACA→VIS neurons (scale, 100 μm). Lower panel, similar to upper panel, but the ACA neurons were labeled by RV injection into SC (left), and axons from ACA→SC neurons are concentrated in PL/ILA and LP (right). (c) Axons detected in all ACA→VIS samples (scale, 1 mm). White masks, injection sites excluded from analysis. (d) Similar to (c), for axons of ACA→SC neurons (scale, 1 mm). (e) Percentages of labeled axons in selected cortical and subcortical structures (ACA, n = 3; ACA→VIS, n = 3; ACA→SC, n = 3). Error bar, ±s.e.m. Data for ACA axons are the same as in Fig. 2d, shown here for comparison. Figure 6 Whole-brain distributions of inputs to ACA→VIS and ACA→SC neurons. (a) Viral vectors and injection procedure. (b) Left, fluorescence images of AAV-mCherry-IRES-WGA-Cre injection site in VIS (upper) or SC (lower) (scale, 1 mm). Middle, injection site of other AAVs and RV in ACA (scale, 1 mm). Inset, enlarged view of region in white box showing AAV/RV infected neurons (green; scale, 20 μm). Right, retrogradely labeled neurons (green) in VIS and PL/ILA (scale, 200 μm). Inset, enlarged view of region in white box (scale, 20 μm). Green, EGFP; red, mCherry; blue, DAPI. (c) RV-labeled neurons detected in all ACA→VIS samples (scale, 1 mm). White masks, injection sites excluded from analysis. (d) Similar to (c), for inputs to ACA→SC neurons (scale, 1 mm). (e) Percentages of retrogradely labeled neurons in selected cortical and subcortical brain structures (ACA, n = 3 mice; ACA→VIS, n = 4; ACA→SC, n = 4). Error bar, ±s.e.m. Data for ACA inputs are the same as in Fig. 2h. Figure 7 Schematic diagram of visual subnetworks. Shown are major inputs and outputs of ACA→VIS and ACA→SC neurons. The putative unidirectional connection from ACA→VIS to ACA→SC neurons (dashed blue line) was based on previous literature. Table 1 Abbreviations of Anatomical Structures Abbreviation Definition Cortical areas ACA anterior cingulate area AI agranular insular area AUD auditory areas ECT ectorhinal area ILA infralimbic area MO somatomotor areas (MOs + MOp) MOp primary motor area MOs secondary motor area ORB orbital area PL prelimbic area PTLp posterior parietal association areas RSP retrosplenial area SS somatosensory areas TEa temporal association area VIS visual areas Thalamic nuclei AD anterodorsal nucleus AM anteromedial nucleus AV anteroventral nucleus of thalamus LD lateral dorsal nucleus of thalamus LGd dorsal part of the lateral geniculate complex LGv ventral part of the lateral geniculate complex LP lateral posterior nucleus of the thalamus MD mediodorsal nucleus of thalamus MG medial geniculate complex RT reticular nucleus of the thalamus PF parafascicular nucleus PO posterior complex of the thalamus VAL ventral anterior-lateral complex of the thalamus VM ventral medial nucleus of the thalamus VP ventral posterior complex of the thalamus Other structures PAL pallidum SC superior colliculus STR striatum AUTHOR CONTRIBUTIONS S.Z. and Y.D. conceived and designed the experiments. S.Z. performed and organized all the experiments. M.X. wrote the software for data analyses and analysed the data. W.-C.C., C.M. and J.D. prepared AAV and RV vectors for rabies virus based retrograde tracing. W.-C.C. also performed some viral injections. T.L. and D.J. performed the detection for labelled cells and axons in some brain samples. J.L.F. and D.J. performed the brain tissue sectioning with cryostat. S.Z., M.X. and Y.D. wrote the manuscript. COMPETING FINANCIAL INTERESTS The authors declare no competing financial interests. 1 Xu NL Nonlinear dendritic integration of sensory and motor input during an active sensing task Nature 492 247 251 2012 23143335 2 Lee S Kruglikov I Huang ZJ Fishell G Rudy B A disinhibitory circuit mediates motor integration in the somatosensory cortex Nat Neurosci 16 1662 1670 2013 24097044 3 Nelson A A circuit for motor cortical modulation of auditory cortical activity J Neurosci 33 14342 14353 2013 24005287 4 Schneider DM Nelson A Mooney R A synaptic and circuit basis for corollary discharge in the auditory cortex Nature 513 189 194 2014 25162524 5 Zhang S Selective attention. Long-range and local circuits for top-down modulation of visual cortex processing Science 345 660 665 2014 25104383 6 Desimone R Duncan J Neural mechanisms of selective visual attention Annu Rev Neurosci 18 193 222 1995 7605061 7 Fuster JM The prefrontal cortex : anatomy, physiology, and neuropsychology of the frontal lobe Lippincott-Raven Philadelphia 1997 8 Mao T Long-range neuronal circuits underlying the interaction between sensory and motor cortex Neuron 72 111 123 2011 21982373 9 Noudoost B Chang MH Steinmetz NA Moore T Top-down control of visual attention Curr Opin Neurobiol 20 183 190 2010 20303256 10 Oh SW A mesoscale connectome of the mouse brain Nature 508 207 214 2014 24695228 11 Petersen CC Cortical control of whisker movement Annu Rev Neurosci 37 183 203 2014 24821429 12 Stanton GB Bruce CJ Goldberg ME Topography of projections to posterior cortical areas from the macaque frontal eye fields J Comp Neurol 353 291 305 1995 7745137 13 Zingg B Neural networks of the mouse neocortex Cell 156 1096 1111 2014 24581503 14 Hooks BM Organization of cortical and thalamic input to pyramidal neurons in mouse motor cortex J Neurosci 33 748 760 2013 23303952 15 Hunnicutt BJ A comprehensive thalamocortical projection map at the mesoscopic level Nat Neurosci 17 1276 1285 2014 25086607 16 Miyamichi K Cortical representations of olfactory input by trans-synaptic tracing Nature 472 191 196 2011 21179085 17 Wickersham IR Monosynaptic restriction of transsynaptic tracing from single, genetically targeted neurons Neuron 53 639 647 2007 17329205 18 Wang Q Sporns O Burkhalter A Network analysis of corticocortical connections reveals ventral and dorsal processing streams in mouse visual cortex J Neurosci 32 4386 4399 2012 22457489 19 Jones EG The thalamus 2 Cambridge University Press 2007 20 Wu H Williams J Nathans J Complete morphologies of basal forebrain cholinergic neurons in the mouse eLife 3 e02444 2014 24894464 21 Fishell G Rudy B Mechanisms of inhibition within the telencephalon: “where the wild things are” Annu Rev Neurosci 34 535 567 2011 21469958 22 Xu X Roby KD Callaway EM Immunochemical characterization of inhibitory mouse cortical neurons: three chemically distinct classes of inhibitory cells J Comp Neurol 518 389 404 2010 19950390 23 Lodato S Shetty AS Arlotta P Cerebral cortex assembly: generating and reprogramming projection neuron diversity Trends Neurosci 38 117 125 2015 25529141 24 Shepherd GM Corticostriatal connectivity and its role in disease Nat Rev Neurosci 14 278 291 2013 23511908 25 Makino H Komiyama T Learning enhances the relative impact of top-down processing in the visual cortex Nat Neurosci 18 1116 1122 2015 26167904 26 Pouget P Visual and motor connectivity and the distribution of calcium-binding proteins in macaque frontal eye field: implications for saccade target selection Front Neuroanat 3 2 2009 19506705 27 Goldberg ME Wurtz RH Activity of superior colliculus in behaving monkey. II Effect of attention on neuronal responses J Neurophysiol 35 560 574 1972 4624740 28 Horwitz GD Newsome WT Separate signals for target selection and movement specification in the superior colliculus Science 284 1158 1161 1999 10325224 29 Ignashchenkova A Dicke PW Haarmeier T Thier P Neuron-specific contribution of the superior colliculus to overt and covert shifts of attention Nat Neurosci 7 56 64 2004 14699418 30 Kustov AA Robinson DL Shared neural control of attentional shifts and eye movements Nature 384 74 77 1996 8900281 31 Zenon A Krauzlis RJ Attention deficits without cortical neuronal deficits Nature 489 434 437 2012 22972195 32 Li N Chen TW Guo ZV Gerfen CR Svoboda K A motor cortex circuit for motor planning and movement Nature 519 51 56 2015 25731172 33 Beier KT Circuit Architecture of VTA Dopamine Neurons Revealed by Systematic Input-Output Mapping Cell 162 622 634 2015 26232228 34 Lerner TN Intact-Brain Analyses Reveal Distinct Information Carried by SNc Dopamine Subcircuits Cell 162 635 647 2015 26232229 35 Gradinaru V Molecular and cellular approaches for diversifying and extending optogenetics Cell 141 154 165 2010 20303157 36 Euston DR Gruber AJ McNaughton BL The role of medial prefrontal cortex in memory and decision making Neuron 76 1057 1070 2012 23259943 37 Robinson DA Eye movements evoked by collicular stimulation in the alert monkey Vision Res 12 1795 1808 1972 4627952 38 Wang L Liu M Segraves MA Cang J Visual Experience Is Required for the Development of Eye Movement Maps in the Mouse Superior Colliculus J Neurosci 35 12281 12286 2015 26338338 39 Fu Y A cortical circuit for gain control by behavioral state Cell 156 1139 1152 2014 24630718 40 Goard M Dan Y Basal forebrain activation enhances cortical coding of natural scenes Nat Neurosci 12 1444 1449 2009 19801988 41 Metherate R Ashe JH Nucleus basalis stimulation facilitates thalamocortical synaptic transmission in the rat auditory cortex Synapse 14 132 143 1993 8392756 42 Pinto L Fast modulation of visual perception by basal forebrain cholinergic neurons Nat Neurosci 16 1857 1863 2013 24162654 43 Nelson A Mooney R The Basal Forebrain and Motor Cortex Provide Convergent yet Distinct Movement-Related Inputs to the Auditory Cortex Neuron 90 635 648 2016 27112494 44 Brown SP Hestrin S Intracortical circuits of pyramidal neurons reflect their long-range axonal targets Nature 457 1133 1136 2009 19151698 45 Kiritani T Wickersham IR Seung HS Shepherd GM Hierarchical connectivity and connection-specific dynamics in the corticospinal-corticostriatal microcircuit in mouse motor cortex J Neurosci 32 4992 5001 2012 22492054 46 Morishima M Kawaguchi Y Recurrent connection patterns of corticostriatal pyramidal cells in frontal cortex J Neurosci 26 4394 4405 2006 16624959 47 Purushothaman G Marion R Li K Casagrande VA Gating and control of primary visual cortex by pulvinar Nat Neurosci 15 905 912 2012 22561455 48 Moore T Armstrong KM Selective gating of visual signals by microstimulation of frontal cortex Nature 421 370 373 2003 12540901 49 Noudoost B Moore T Control of visual cortical signals by prefrontal dopamine Nature 474 372 375 2011 21572439 50 Erlich JC Bialek M Brody CD A cortical substrate for memory-guided orienting in the rat Neuron 72 330 343 2011 22017991 51 Maheshri N Koerber JT Kaspar BK Schaffer DV Directed evolution of adeno-associated virus yields enhanced gene delivery vectors Nat Biotechnol 24 198 204 2006 16429148 52 Weissbourd B Presynaptic partners of dorsal raphe serotonergic and GABAergic neurons Neuron 83 645 662 2014 25102560 53 Lim BK Huang KW Grueter BA Rothwell PE Malenka RC Anhedonia requires MC4R-mediated synaptic adaptations in nucleus accumbens Nature 487 183 189 2012 22785313 54 Wall NR Wickersham IR Cetin A De La Parra M Callaway EM Monosynaptic circuit tracing in vivo through Cre-dependent targeting and complementation of modified rabies virus Proc Natl Acad Sci U S A 107 21848 21853 2010 21115815 55 Osakada F Callaway EM Design and generation of recombinant rabies virus vectors Nat Protoc 8 1583 1601 2013 23887178

PMC005xxxxxx/PMC5127761.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 8710219 1493 AIDS AIDS AIDS (London, England) 0269-9370 1473-5571 27677161 5127761 10.1097/QAD.0000000000001286 NIHMS820697 Article Topical gentian violet compared to nystatin oral suspension for the treatment of oropharyngeal candidiasis in HIV-1 Infected participants Mukherjee Pranab K Chen Huichao Patton Lauren L Evans Scott Lee Anthony Kumwenda Johnstone Hakim James Masheto Gaerolwe Sawe Frederick Pho Mai T Freedberg Kenneth A Shiboski Caroline H Ghannoum Mahmoud A * Salata Robert A * and the Oral HIV/AIDS Research Alliance (OHARA)/AIDS Clinical Trials Group (ACTG) 5265 Team Corresponding Author: Robert A. Salata, MD, Division of Infectious Diseases & HIV Medicine, University Hospitals Case Medical Center, 11100 Euclid Ave., Cleveland, OH 44106, Tele: (216) 844-3293, FAX: (216) 844-3145, robert.salata@uhhospitals.org * Contributed equally to the manuscript. 6 10 2016 2 1 2017 02 1 2018 31 1 8188 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. Objective Compare the safety and efficacy of topical gentian violet (GV) to that of nystatin oral suspension (NYS) for the treatment of oropharyngeal candidiasis (OC) in HIV-1 infected adults in resource-limited settings. Design Multicenter, open-label, evaluator-blinded, randomized clinical trial at 8 international sites, within the AIDS Clinical Trials Group. Study participants and Intervention Adult HIV-infected participants with OC, stratified by CD4 cell counts and antiretroviral therapy status at study entry, were randomized to receive either GV (0.00165%, BID) or NYS (500,000 units, QID) for 14 days. Main outcome measure(s) Cure or improvement after 14 days of treatment. Signs and symptoms of OC were evaluated in an evaluator-blinded manner. Results The study was closed early per DSMB after enrolling 221 participants (target = 494). Among the 182 participants eligible for efficacy analysis, 63 (68.5%) in the GV arm had cure or improvement of OC versus 61 (67.8%) in the NYS arm, resulting in a non-sizable difference of 0.007 (95% CI: -0.129, 0.143). There was no sizable difference in cure rates between the two arms (-0.0007; 95% CI: -0.146, 0.131). No GV-related adverse events were noted. No sizable differences were identified in tolerance, adherence, quality of life, or acceptability of study drugs. In GV arm, 61% and 39% of participants reported “no” and “mild-to-moderate” staining, respectively. Cost for medication procurement was significantly lower for GV versus NYS [Median $2.51 and $19.42, respectively, P = 0.01]. Conclusions Efficacy of GV was not statistically different than NYS, was well-tolerated, and its procurement cost was substantially less than NYS. Oral candidiasis oral complications in HIV infection gentian violet nystatin resource-limited setting cost per treatment Background Oral candidiasis (OC) is among the most common opportunistic infections observed in HIV-infected individuals and often presents as the initial manifestation of disease [1-4]. Left untreated, OC can contribute to morbidity including esophageal disease and weight loss [5]. Recent studies conducted by the Oral HIV AIDS Research Alliance (OHARA) of the AIDS Clinical Trials Group (ACTG) demonstrated that OC is strongly associated with tuberculosis in HIV-infected participants, independent of CD4 count [6]. Although the incidence of OC has declined in resource-rich countries following the introduction of highly active antiretroviral therapy (HAART) [7, 8], the prevalence of OC in HIV-infected participants in resource-limited settings (RLS) remains substantial, reaching as high as 33% [9-14]. Nystatin oral suspension (NYS) is the main antifungal used to treat OC in HIV-infected participants in RLS [15-18]. To identify a low cost alternative to current treatment, our group conducted preclinical studies and found GV possesses potent anti-Candida activity [19, 20]. GV has been recommended by the World Health Organization (WHO) as a topical treatment for OC in HIV-infected participants, at a concentration of 1% [21]. In an earlier study, our group reported that oral rinsing with GV at low concentrations (0.00165%, 16.5 μg/mL) was safe and well-tolerated, with no mucosal staining [22]. In the present study, we compared the safety and efficacy of topical GV to that of NYS in the treatment of OC in HIV-infected adult participants from RLS. Methods Design A5265 was a multi-center phase III, randomized, open-label, evaluator-blinded clinical trial conducted in eight non-U.S. Clinical Research Sites (CRSs) of the ACTG (detailed clinical protocol is provided as a Supplemental File). Evaluation of signs and symptoms of OC was performed by a trained clinician blinded to the treatment assignment. Study Sites The sites participating in the protocol A5265 were: (1) University of Natal, Durban, South Africa; YRG CARE Medical Center VHS, Chennai, India; (2) Joint Clinical Research Center, Kampala, Uganda; (3) Walter Reed Project, Kericho, Kenya; (4) AMPATH at MOI University, Eldoret, Kenya; (5) Botswana Harvard Partnership/Princess Marina Hospital, Gaborone, Botswana; (6) Botswana Harvard Partnership/Scottish Livingston Hospital, Molepolole, Botswana; (7) Johns Hopkins Research Project/College of Medicine, Blantyre, Malawi; and (8) UZ/UCSF HIV Prevention Trials Unit, Harare, Zimbabwe. At each site, the study protocol and consent form were approved as appropriate by their local institutional review board (IRB). Study Population The eligibility criteria for enrollment were as follows: HIV-infected adults (≥ 18 years of age) presenting with pseudomembranous candidiasis (PC); positive identification of Candida in oral swabs at screening; negative pregnancy test performed within 48 hours prior to study entry for female study volunteers of reproductive age, a Karnofsky performance score ≥60 prior to entry [23], with ability to comply with all study procedures and follow-up after informed consent. Participants were excluded for the following reasons: proven or presumptive esophageal candidiasis at study entry; use of any investigational drug within 30 days prior to study entry; concurrent vaginal candidiasis within 21 days prior to study entry; use of inhaled or systemic corticosteroids or antifungals within 14 days prior to study entry; use of antifungal agents or other oral topical treatments during the study period; allergy/sensitivity or any hypersensitivity to GV or NYS; active illicit drug or alcohol use or dependence that would interfere with adherence to study requirements; serious illness (HIV-1 associated or not) indicating a high likelihood of death in the 30 days after study entry; predictors of early mortality (e.g. anemia, low CD4 count); previous or current history of porphyria; presence of oral warts during the screening period or at study entry; or use of full or partial dentures at study entry. Participants who failed treatment were referred for care outside of the study but continued to be followed on study; a zole therapy was most often initiated with study drug failure. Randomization and Stratification Participants were randomized in a 1:1 ratio to two arms (GV or nystatin), and stratified according to the following criteria: (a) screening CD4+ cell count > 200 cells/mm3 or ≤ _200 cells/mm3, (b) taking ART at the time of study entry or not taking ART at the time of study entry (and not planning to initiate ART during the study-defined 14-day treatment period). Study Treatment At randomization, participants were stratified by CD4+ T-cell counts (> 200 cells/mm3 or ≤ 200 cells/mm3) and ART use at the time of study entry (Yes or No) to receive topical GV (0.00165%, 5 mL swish and gargle for 2 minutes and expectorate twice daily) or NYS oral suspension (5 mL of 100,000 units/mL swish for 2 minutes and swallow, four times daily) for 14 days. Participants were instructed not to eat or drink for 30 minutes before and after the administration of study drugs. Oral Examination and Scoring of Lesions The oral cavity was categorized into 6 sites (Supplementary Figure 1). Severity was scored from 0 to 3 (0:no lesions, 1: scattered non-confluent lesions <2 mm in diameter, 2:multiple non-confluent lesions >2 mm in diameter, and 3: extensive confluent lesions). A composite severity score (CSS, range: 0 to 18) was obtained after adding the scores from all 6 sites. All examiners received standardized training in establishing a clinical diagnosis of PC using a published case definition [24]. A primary physician asisted by a registered nurse practitioner led the assessment at each study site. Endpoints The primary endpoint was a single binary outcome, defined as cure (CSS = 0) or improvement (a decrease in severity of lesions) at the end of treatment. Therapy was considered to have failed if participants had no improvement or worsening of their scores, or if signs or symptoms worsened after ≥ 7 consecutive days of therapy. Secondary endpoints included: OC symptoms (discomfort and pain), yeast colony counts, adverse events (AEs), tolerance, adherence (medication diaries and bottle counts), self-reported quality of life (QOL), acceptability of study drugs, and cost per treatment regimen. Discomfort and pain were rated on a 4-point scale (from no pain=0, mild pain=1, moderate pain=2, and severe pain=3). Cost per treatment course was calculated using a microcosting, ingredients-based approach [25]. Self-reported resource utilization including outpatient clinic visits and inpatient hospital stays at study entry and at the week 6 visit was compared between each arm. Costs for NYS also included materials and labor costs for drug preparation if acquired in concentrated form. Drug acquisition and labor costs were obtained directly from study sites and converted to 2014 U.S. Dollars (USD). Sample Size Calculations and Data Analyses This study was sized to estimate the difference in clinical efficacy rates between GV and NYS for the treatment of OC with appropriate precision (width of 95% CI < 0.2). Repeated CIs (RCIs) were used to control type I error. One interim analysis was conducted by utilizing a 99.7% CI and the final analysis used a 95.1% CI (constructed based on the Lan-DeMets error-spending function corresponding to the O'Brien-Fleming boundary). Assuming response rates of GV and NYS of 50% (resulting in the greatest variation and hence the widest CI), an intention-to-treat population (ITT) analysis required 494 participants after adjusting for 10% noncompliance. The primary analysis was conducted on a modified ITT population, which was defined as participants who present with a clinical diagnosis of OC with a positive yeast culture for Candida spp. at baseline. Cases without oral exam records at week 2 were considered clinical failures. A comparison of clinical efficacy rates was performed using a 95.1% CI for the difference (reported as 95% to account for 95% simultaneous coverage probability given one interim analysis) between two proportions, with four subgroup analyses conducted based on two stratification factors (screening CD4+ count and ART use at the time of study entry). In addition, a comparison of cure rates was performed per the Data Safety Monitoring Board (DSMB) recommendation of using “cure” as the primary endpoint given the concern of staining with GV. The CIs were based on either asymptotic method when event counts were > 12 or the formula provided in Fleiss [26] where the CI was adjusted by 0.5*(1/n1 + 1/n2). These comparisons were also conducted based on observed data as sensitivity analyses. Study Accrual and Closure A5265 was opened to accrual on June 7, 2011, and the first subject enrolled on August 2, 2011. As of October 19, 2012, a total of 221 participants (targeted sample size=494) were enrolled. The study was halted due to the DSMB concern regarding the excessive mortality unrelated to study medications in the recruited population and subsequently the study was closed because OHARA funding was not renewed. Results Baseline Characteristics Summaries of selected baseline characteristics are shown in Table 1. A total of 221 participants were enrolled into the study (Figure 1); 181 participants completed the study protocol and 40 participants discontinued the study prior to study completion (see Table 2 for reasons of study discontinuation). Among the enrolled participants, 106 (48%) were in the age group 30-39 years, and 128 (58%) were females. Moreover, 203 (92%) participants were black non-Hispanic, and 18 (8%) were Asian or Pacific Islanders. At the time of study entry, 166 (75%) participants were not taking ART and not planning to initiate ART until the study-defined 14-day treatment period was complete. The CD4 count in 175 participants (79%) was between 0-200 cells/mm3 (Median: 53 cells/mm3) and the median of baseline viral RNA titer log10[HIV RNA (cp/ml)] was 5.30. In patients with CD4 counts below 200, the median and IQR results were similar between the two treatment arms: 37 [13, 69] for GV and 43 [10, 92] for NYS; In patients with CD4 counts greater than 200, the median [IQR] were also similar between the two arms: 356 [299, 512] for GV and 371 [265, 573] for NYS. Safety Monitoring There were 18 participants (10 in GV group and 8 in NYS group) who experienced any signs or symptoms of grade 3 or higher AEs and 12 participants (6 in GV and 6 in NYS) who experienced new laboratory toxicity events of grade 3 or higher. In the GV arm, 1 case was reported to be probably not related to GV, while all other new signs, symptoms, laboratory events, and AEs, were not related to GV. In the NYS arm, there were 2 reported cases probably not related, 1 case possibly-related, and the rest (3) were not related to the study drug. There were 21 deaths (12 in GV and9 in NYS), and the primary causes of death for 18 of these cases were: HIV infection or HIV-related diagnosis (15), non-HIV diagnosis (2), toxicity (1) secondary to herbal medication (Table 3). None of the deaths were considered related to study medications. Efficacy Analysis Population for Efficacy Analysis Of the 221 participants enrolled, a total of 182 (GV: 92; NYS: 90) had positive Candida culture results at baseline, and were eligible for the modified ITT efficacy analysis (Figure 1). Of these participants, 170 (GV: 87; NYS: 83) had oral exams at week 2. Cases without oral exam records at week 2 were considered as clinical failures, which included 5 participants in the GV arm and 7 in the NYS arm. For the ‘observed data’ analysis, 19 (GV: 10; NYS: 9) of the 221 (GV: 110; NYS: 111) enrolled participants did not have oral exams at week 2, thus excluded, which led to a total of 202 (GV: 100; NYS: 102) participants in the analysis (Figure 1). Evaluation of Clinical Efficacy We first evaluated the difference in clinical efficacy rates based on modified ITT analysis. Among the 182 participants eligible for analysis, 63 (68.5%) in the GV arm (n=92) had cure or improvement of OC compared to 61 (67.8%) in the NYS arm (n=90), resulting in a non-sizable difference (GV-NYS) of 0.007 (95% CI: -0.129, 0.143). Table 1 showed that the majority of participants were not on ART. Comparison of the two arms using a “cure” rate also showed no sizable difference (-0.0007, 95% CI: -0.146, 0.131). Sensitivity analyses based on observed data showed no sizable difference either (Table 4). Due to the sparse stratum cells presented in Table 4, no adjustment was made for the stratification in the efficacy analyses or the additional analyses below. Additional Outcome Measures by Treatment The analyses below were based on observed data Staining Evaluation of differences in staining of the oral mucosa in the GV arm revealed 61% of participants reported no staining; 28% reported mild staining (mainly the tongue); and 11% reported moderate staining of the oral cavity. Notably, no participant reported severe staining. There were no instances where GV was discontinued due to staining. No staining was observed in the NYS arm. Evaluation of symptoms Pain and discomfort associated with OC were evaluated at study entry, end-of-treatment, and at clinical relapse. At study entry, a total of 217 observations (106 in the GV arm; 111 in the NYS arm) were available for evaluation. At the end-of-treatment, a total of 204 observations were available to evaluate the symptoms using extended Mantel-Haenszel test between arms. The tests did not show a sizable treatment effect. A total of 8 participants had clinical relapse of OC between weeks 2 and 8 (5 in GV; 3 in NYS). All 5 participants reported no pain in the GV arm while 2 reported mild and 1 reported severe pain in the NYS arm. One participant from each arm reported severe discomfort at relapse; others had no discomfort. Quantitative yeast colony counts The number of colony forming units (CFUs) per milliliter was evaluated at entry, end-of-treatment, at week 6, and at clinical relapse (between weeks 2 and 8). The difference in the means of log-transformed CFUs at entry (GV – NYS, 95% CI), end-of-treatment, and at week 6 was 0.25 (-0.40, 0.89), 1.05 (0.32, 1.78), and 1.15 (-0.88, 1.17), respectively (Supplemental Table 1). At clinical relapse, 3 observations were available for each arm without marked difference in mean log CFU. Assessment of tolerance, adherence, quality of life, acceptability, and cost comparison There was no significant difference between the GV and NYS groups in regard to drug tolerance, adherence, self-reported QOL, and acceptability (p = 0.17, 0.83, 0.88, and 0.36, respectively). The estimated ratio of adherence between GV and NYS was 1.01 with 95% CI (0.93, 1.10), and the ratio of acceptability between the two groups was 1.03 with 95% CI (0.97, 1.10). Comparison of the cost per treatment course of GV and NYS by study CRS showed that the GV medication (requiring pharmacy preparation via reconstitution and/or dilution), procurement costs were sizably lower than treatment with NYS (no pharmacy preparation needed) [Median (Q1, Q3) = $2.51 (0.95, 3.11) and $19.42 (6.46, 36.28), respectively, p = 0.01; Supplementary Figure 2 and Supplementary Table 2]. Discussion The current study was the largest evaluator-blinded, randomized clinical trial that compared the efficacy of GV to NYS in the treatment of OC in HIV-infected adults in RLS. Our results showed that both treatments resulted in high clinical efficacy rates with no sizable differences between the two groups. The efficacy rates for both GV and NYS were higher than previous studies comparing the efficacy of these two agents [27, 28]. Our secondary analyses revealed no sizable differences in tolerance, adherence, QOL, or acceptability of study drugs. Oral NYS is commonly used in the treatment of OC in HIV-infected participants [29, 30]; however, its use is limited by bitter taste, gastrointestinal side effects, four times per day dosing, and a relatively high cost, which contribute to reduced drug adherence and lower efficacy. While there is substantial variation in procurement prices for medications across sites [31], our study showed that GV (despite required pharmacy preparation) was consistently associated with a lower cost per treatment dose compared to NYS (no pharmacy preparation needed). While labor needs may constrain healthcare in RLS, the need for pharmacy preparation for GV is counterbalanced by the substantially lower cost of GV. For example, the labor cost of preparing GV would need to be more than seven times the current estimates before total cost per GV treatment regimen reached that of NYS. In addition, the widespread availability of GV in RLS, its long-term physical and chemical stability, and the ease of preparation makes GV a viable and attractive treatment option in HIV-infected participants. Nyst et al. [27] conducted a randomized open-label study to compare the efficacy of GV versus NYS and oral ketoconazole for the treatment of OC in AIDS participants in the Democratic Republic of Congo. There was similar efficacy among all three drugs. However, this study was limited by the small sample size and the fact that the GV concentration used was 300-fold higher than the concentration used in the current study (0.5% vs. 0.00165%, respectively), and was associated with staining of teeth and gums, oral irritation and small superficial ulcers. In a separate study, Hodgson et al. [28] compared the efficacy and tolerability of 1% and 0.00165% (as used in this study) GV and NYS mouth rinses for oral PC in pediatric HIV-infected participants in Malawi. The lower concentration of GV (0.00165%) was as effective as a 1% GV solution with fewer side effects. Gentian violethas a long history of use and has been recommended by the WHO as a topical treatment for OC in HIV-infected participants, at a concentration of 1% [21]. However, since 1% GV strongly stains the oral cavity, its use has been associated with a stigmatizing effect. The current trial convincingly demonstrated that use of a lower concentration of GV does not lead to severe staining or discontinuation of GV treatment and that any staining noted was transient and generally mild. These results may provide an impetus for wider acceptance of GV as a treatment for OC, especially in RLS. Deaths reported among enrolled participants in the current study that served as the basis for the DSMB halt of the study are reflective of the study population and not associated with study medications. In this regard, mortality rates of 8 to 40% have been reported in HIV/AIDS participants in the era of ART in RLS [32-36]. The median CD4 cell count of 53 cells/mm3 for study participants in our study may help explain the deaths in this patient population. Limitations of the study include the inability to fully accrue the target enrollment, leading to a wider CI (0.27 using the estimated efficacy rates) when compared to the original design (0.20 assuming 50% response rates). Additionally, this study was limited to sub-Saharan Africa and India, and the generalization of results to other geographic sites needs to be confirmed in additional clinical trials. In conclusion, this is the largest randomized clinical trial evaluating the use of gentian violet for treatment of oral candidiasis in HIV-infected adults in resource-limited setting. While the efficacy of gentian violet was not statistically different thannystatin, it was well-tolerated, and drug procurement was substantially less costly than nystatin for treatment of HIV-associated oral candidiasis. Supplementary Material Supplemental Data File _.doc_.tif_ pdf_etc.__1 Supplementary Figure 1. Map of the oral cavity. (1) Left lower and upper labial mucosa and buccal mucosa, (2) right lower and upper labial mucosa and buccal mucosa, (3) hard palate, (4) soft palate, (5) tongue (dorsum, lateral, and ventral), (6) floor of mouth. Participating sites were encouraged to use the oral cavity diagram to assess ulcerations and staining. Both the oral examiner and clinical staff were asked to use this diagram during the oral exam and oral treatment interview, respectively. Supplementary Figure 2. Cost per treatment course of nystatin or gentian violet. Cost of treatment with nystatin or gentian violet were compared using non-parametric test (Mann Whitney, two-tailed) and a P-value < .05 was considered significant. Supplementary Table 1: Quantitative fungal counts in gentian violet (GV) and nystatin (NYS) arms, evaluated in swabs collected at different time points during the study Supplementary Table 2. Cost per treatment course (U.S. dollars) for nystatin (NYS) and gentian violet (GV). Supplemental Data File _.doc_.tif_ pdf_etc.__2 The authors would like to express their sincere appreciation to participants in the A5265 study, and members of the A5265 protocol team. Study A5265 was supported by the Oral HIV AIDS Research Alliance (OHARA, BRS-ACURE-S-11-000049-110229), funded by the National Institute of Dental and Craniofacial Research (NIDCR), the AIDS Clinical Trials Group (award number UM1AI068634, UM1AI068636 and UM1AI106701) from the National Institute of Allergy and Infectious Diseases, and supported by the CWRU/UH Center for AIDS Research (CFAR, NIH grant number P30 AI036219). The content is solely the responsibility of the authors and does not necessarily represent the official views of NIDCR, NIAID, or the NIH. The NCT number is NCT01427738. The other members of the protocol team besides the co-authors: Trinh Ly, M.D. (DAIDS Medical Officer); Isaac Rodriguez-Chavez, Ph.D. (NIDCR Clinical Science Representative); Lynette Purdue, Pharm.D. (Pharmacist); Suria Yesmin, B.S. (Clinical Trials Specialist); Deborah Greenspan, B.D.S., D. Sc., Cissy Kityo, M.Sc., Johnstone Kumwenda, FRCP, Lauren Patton, D.D.S., Srikanth Tripathy, M.D., M.B.B.S. (Investigators); Flavia Miiro, M.Sc. (CSS Representative); Linda Wieclaw, B.S. (Data Manager); Jenifer Baer, R.N. (Field Representative); Christina Blanchard-Horan, Ph.D. (International Program Specialist); Laura Hovind, M.S., Aimee Willett, B.A. (Laboratory Data Managers). The role(s) of the authors in the conducted study are outlined below: Author Role on Study 1. Pranab K Mukherjee Writing, data analyses, conception, design 2. Huichao Chen Data analyses and design 3. Lauren L Patton Conception and design 4. Scott Evans Data analyses and design 5. Anthony Lee Data analyses and design 6. Johnstone Kumwenda Performance (enrollment of patients) 7. James Hakim Performance (enrollment of patients) 8. Gaerolwe Masheto Performance (enrollment of patients) 9. Frederick Sawe Performance (enrollment of patients) 10. Mai T Pho Design and data analyses 11. Kenneth A Freedberg Design and data analyses 12. Caroline H Shiboski Conception and design 13. Mahmoud A Ghannoum Conception, design and performance 14. Robert A Salata Conception, design and performance Figure 1 CONSORT diagram for the study Table 1 Demographics and Baseline Characteristics Characteristic Gentian Violet (N=110) Nystatin (N=111) Total (N=221) Age (years) 18-19 0 (0%) 1 (1%) 1 (0%) 20-29 17 (15%) 30 (27%) 47 (21%) 30-39 54 (49%) 52 (47%) 106 (48%) 40-49 21 (19%) 17 (15%) 38 (17%) 50-59 16 (15%) 9 (8%) 25 (11%) Over 60 2 (2%) 2 (2%) 4 (2%) Gender Male 48 (44%) 45 (41%) 93 (42%) Female 62 (56%) 66 (59%) 128 (58%) Race/Ethnicity Black Non-Hispanic 97 (88%) 106 (95%) 203 (92%) Asian, Pacific Islander 13 (12%) 5 (5%) 18 (8%) Antiretroviral Therapy (ART) Usage at Entry On ART 27 (25%) 28 (25%) 55 (25%) Not on ART 83 (75%) 83 (75%) 166 (75%) CD4 Count (cells/mm3) 0-200 87 (79%) 88 (79%) 175 (79%) > 200 23 (21%) 23 (21%) 46 (21%) Baseline Log10 [HIV RNA (cp/ml)] N 110 110 220 # missing 0 1 1 Mean (SD) 4.89 (1.31) 5.03 (1.21) 4.96 (1.26) Min, Max 1.59, 6.42 1.59, 7.00 1.59, 7.00 Median 5.21 5.40 5.30 Q1, Q3 4.76, 5.72 4.78, 5.71 4.77, 5.72 Table 2 Reasons for study discontinuation in the gentian violet (GV) and nystatin (NYS) arms Off-Study Reason GV N (%*) NYS N (%*) Total N (%*) Completed protocol 91 (82.7%) 90 (81.1%) 181 (81.9%) Death 12 (10.9%) 9 (8.1%) 21 (9.5%) Severe debilitation, unable to continue 0 (0.0%) 1 (0.9%) 1 (0.5%) Subject/parent not able to get to clinic 5 (4.5%) 2 (1.8%) 7 (3.2%) Subject/parent withdrew consent prior to study completion 0 (0.0%) 2 (1.8%) 2 (0.9%) Subject/parent not willing to adhere to study requirements 1 (0.9%) 2 (1.8%) 3 (1.4%) ACTU unable to contact subject/parent 1 (0.9%) 5 (4.5%) 6 (2.7%) * Percentages were compared to the number of enrolled participants in each group (GV = 110, NYS = 111, Total = 221) ACTU = AIDS Clinical Trials Unit Table 3 Reasons for death in gentian violet (GV) and nystatin (NYS) treatment arms Primary Cause of Death Category GV (N, %) NYS (N, %) Total HIV infection or HIV-related diagnosis 8 (67%) 7 (78%) 15 (71%) Non-HIV diagnosis 2 (17%) 0 (0%) 2 (10%) Toxicity 0 (0%) 1 (11%) 1 (5%) No information available 2 (17%) 1 (11%) 3 (14%) Total 12 9 21 None of the deaths were considered related to study medications Table 4 Comparison of gentian violet (GV) and nystatin (NYS) in clinical efficacy and cure rates Variable GV NYS Diff. with 95% CI‡ Event / Sample size Clinical efficacy in the mITT 63/92 61/90 0.007 (-0.129, 0.143)  On ART*; CD4§ ≤ 200 11/15 11/17 0.086 (-0.297, 0.469)  On ART; CD4 > 200 4/5 2/2 -0.200 (-0.902, 0.502)  Not on ART; CD4 ≤ 200 38/61 41/63 -0.028 (-0.198, 0.142)  Not on ART; CD4 > 200 10/11 7/8 0.034 (-0.360, 0.429) Cure in the mITT 31/92 31/90 -0.007 (-0.146, 0.131)  On ART; CD4 ≤ 200 4/15 2/17 0.149 (-0.186, 0.484)  On ART; CD4 > 200 5/5 2/2 0.000 (-0.350, 0.350)  Not on ART; CD4 ≤ 200 16/61 21/63 -0.071 (-0.232, 0.090)  Not on ART; CD4 > 200 6/11 6/8 -0.205 (-0.735, 0.326) Clinical efficacy based on observed data 76/100 73/102 0.044 (-0.077, 0.166)  On ART; CD4 ≤ 200 12/16 12/18 0.083 (-0.281, 0.448)  On ART; CD4 > 200 9/9 7/10 0.300 (-0.091, 0.691)  Not on ART; CD4 ≤ 200 43/63 44/63 -0.016 (-0.178, 0.146)  Not on ART; CD4 > 200 12/12 10/11 0.091 (-0.167, 0.349) Cure based on observed data 37/100 41/102 -0.032 (-0.167, 0.103)  On ART; CD4 ≤ 200 4/16 2/18 0.139 (-0.178, 0.456)  On ART; CD4 > 200 9/9 7/10 0.300 (-0.091, 0.691)  Not on ART; CD4 ≤ 200 17/63 23/63 -0.095 (-0.258, 0.067)  Not on ART; CD4 > 200 7/12 9/11 -0.235 (-0.684, 0.214) ‡ Asymptotic 95% CI was used when event counts were > 12; Otherwise, formula by Fleiss [26] was used. * Antiretroviral Therapy (ART) Usage at Entry § Screening CD4 Count (cells/μL) 1 Costa CR Cohen AJ Fernandes OF Miranda KC Passos XS Souza LK Asymptomatic oral carriage of Candida species in HIV-infected patients in the highly active antiretroviral therapy era Rev Inst Med Trop Sao Paulo 2006 48 257 261 17086312 2 Greenspan D Treatment of oral candidiasis in HIV infection Oral Surg Oral Med Oral Pathol 1994 78 211 215 7936591 3 Patel M Shackleton JT Coogan MM Effect of antifungal treatment on the prevalence of yeasts in HIV-infected subjects J Med Microbiol 2006 55 1279 1284 16914661 4 Patton LL Phelan JA Ramos-Gomez FJ Nittayananta W Shiboski CH Mbuguye TL Prevalence and classification of HIV-associated oral lesions Oral Dis 2002 8 98 109 12164670 5 Pienaar ED Young T Holmes H Interventions for the prevention and management of oropharyngeal candidiasis associated with HIV infection in adults and children Cochrane Database Syst Rev 2010 11 CD003940 6 Shiboski CH Chen H Ghannoum MA Komarow L Evans S Mukherjee PK Role of oral candidiasis in TB and HIV co-infection: AIDS Clinical Trial Group Protocol A5253 Int J Tuberc Lung Dis 2014 18 682 688 24903939 7 Patton LL McKaig R Strauss R Rogers D Eron JJ Jr Changing prevalence of oral manifestations of human immuno-deficiency virus in the era of protease inhibitor therapy Oral Surg Oral Med Oral Pathol Oral Radiol Endod 2000 89 299 304 10710453 8 Yang YL Lo HJ Hung CC Li Y Effect of prolonged HAART on oral colonization with Candida and candidiasis BMC Infect Dis 2006 6 8 16423306 9 Arendorf TM Bredekamp B Cloete CA Sauer G Oral manifestations of HIV infection in 600 South African patients J Oral Pathol Med 1998 27 176 179 9563573 10 Arendorf TM Bredekamp B Cloete C Stephen LX Oral soft-tissue manifestations as presenting symptom/sign of HIV infection SADJ 1999 54 602 604 16892567 11 Blignaut E Raubenheimer E van Heerden WF Senekal R Dreyer MJ Oral yeast flora of a Kalahari population J Dent Assoc S Afr 1995 50 601 604 9461886 12 Kamiru HN Naidoo S Oral HIV lesions and oral health behaviour of HIV-positive patients attending the Queen Elizabeth II Hospital, Maseru, Lesotho SADJ 2002 57 479 482 12674869 13 Holmes CB Wood R Badri M Zilber S Wang B Maartens G CD4 decline and incidence of opportunistic infections in Cape Town, South Africa: implications for prophylaxis and treatment J Acquir Immune Defic Syndr 2006 42 464 469 16810113 14 Shiboski CH HIV-related oral disease epidemiology among women: year 2000 update Oral Dis 2002 8 44 48 12164659 15 Nebavi F Arnavielhe S Le Guennec R Menan E Kacou A Combe P Oropharyngeal candidiasis in AIDS patients from Abidjan (Ivory Coast): antifungal susceptibilities and multilocus enzyme electrophoresis analysis of Candida albicans isolates Pathol Biol (Paris) 1998 46 307 314 9769890 16 Ravera M Reggiori A Agliata AM Rocco RP Evaluating diagnosis and treatment of oral and esophageal candidiasis in Ugandan AIDS patients Emerg Infect Dis 1999 5 274 277 10221882 17 Vazquez JA Therapeutic options for the management of oropharyngeal and esophageal candidiasis in HIV/AIDS patients HIV Clin Trials 2000 1 47 59 11590489 18 Blignaut E Botes ME Nieman HL The treatment of oral candidiasis in a cohort of South African HIV/AIDS patients SADJ 1999 54 605 608 16892568 19 Traboulsi RS Mukherjee PK Ghannoum MA In vitro activity of inexpensive topical alternatives against Candida species isolated from the oral cavity of HIV-infected patients Int J Antimicrob Agents 2008 31 272 276 18242063 20 Traboulsi RS Mukherjee PK Chandra J Salata RA Jurevic R Ghannoum MA Gentian violet exhibits activity against biofilms formed by oral Candida isolates obtained from HIV-infected patients Antimicrob Agents Chemother 2011 55 3043 3045 21444708 21 World Health Organization Clinical Management of HIV/AIDS at District and PHC Levels (SEA/AIDS/101) New Delhi World Health Organization Regional Office for South-East Asia 1998 22 Jurevic RJ Traboulsi RS Mukherjee PK Salata RA Ghannoum MA Identification of gentian violet concentration that does not stain oral mucosa, possesses anti-candidal activity and is well tolerated Eur J Clin Microbiol Infect Dis 2011 30 629 633 21210170 23 Wenzel T Pindur G Morsdorf S Giacchi J Influence of HIV-infection on the Karnofsky score and general social functioning in patients with hemophilia Haemostasis 1998 28 106 110 10087436 24 Shiboski CH Patton LL Webster-Cyriaque JY Greenspan D Traboulsi RS Ghannoum M The Oral HIV/AIDS Research Alliance: updated case definitions of oral disease endpoints J Oral Pathol Med 2009 38 481 488 19594839 25 Drummond MF Sculpher MJ Torrance GW O'Brien BJ Stoddart GL Methods for the Economic Evaluation of Health Care Programmes 3rd New York Oxford University Press 2005 26 Fleiss JL Levin B Cho Paik M Statistical Methods for Rates and Proportions Third John Wiley & Sons, Inc 2003 27 Nyst MJ Perriens JH Kimputu L Lumbila M Nelson AM Piot P Gentian violet, ketoconazole and nystatin in oropharyngeal and esophageal candidiasis in Zairian AIDS patients Ann Soc Belg Med Trop 1992 72 45 52 1567268 28 Hodgson TA Dube Q Lewsey JD Zinga B Mhango T Tembo P The efficacy and tolerability of 1% and 0.00165% gentian violet and nystatin mouth rinses for paediatric oral pseudomembranous candidosis in Malawi: a three armed RCT 8th Biennial Congress of the European Association of Oral Medicine Zagreb, Croatia 2006 7 29 Garcia-Cuesta C Sarrion-Perez MG Bagan JV Current treatment of oral candidiasis: A literature review J Clin Exp Dent 2014 6 e576 582 25674329 30 Patton LL Bonito AJ Shugars DA A systematic review of the effectiveness of antifungal drugs for the prevention and treatment of oropharyngeal candidiasis in HIV-positive patients Oral Surg Oral Med Oral Pathol Oral Radiol Endod 2001 92 170 179 11505264 31 Martinson N Mohapi L Bakos D Gray GE McIntyre JA Holmes CB Costs of providing care for HIV-infected adults in an urban HIV clinic in Soweto, South Africa J Acquir Immune Defic Syndr 2009 50 327 330 19194308 32 Floyd S Marston M Baisley K Wringe A Herbst K Chihana M The effect of antiretroviral therapy provision on all-cause, AIDS and non-AIDS mortality at the population level--a comparative analysis of data from four settings in Southern and East Africa Trop Med Int Health 2012 17 e84 93 22943383 33 Lawn SD Harries AD Anglaret X Myer L Wood R Early mortality among adults accessing antiretroviral treatment programmes in sub-Saharan Africa AIDS 2008 22 1897 1908 18784453 34 Ogoina D Obiako RO Muktar HM Adeiza M Babadoko A Hassan A Morbidity and mortality patterns of hospitalised adult HIV/AIDS patients in the era of highly active antiretroviral therapy: A 4-year retrospective review from Zaria, northern Nigeria AIDS Res Treat 2012 55 1707 1718 35 Walker AS Prendergast AJ Mugyenyi P Munderi P Hakim J Kekitiinwa A Mortality in the year following antiretroviral therapy initiation in HIV-infected adults and children in Uganda and Zimbabwe Clin Infect Dis 2012 36 Kasamba I Baisley K Mayanja BN Maher D Grosskurth H The impact of antiretroviral treatment on mortality trends of HIV-positive adults in rural Uganda: a longitudinal population-based study, 1999-2009 Trop Med Int Health 2012 17 e66 73 22943381

PMC005xxxxxx/PMC5127764.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 8510246 21001 Ann Behav Med Ann Behav Med Annals of behavioral medicine : a publication of the Society of Behavioral Medicine 0883-6612 1532-4796 27333898 5127764 10.1007/s12160-016-9812-x NIHMS797885 Article The Daily Relationship Between Aspects of Food Insecurity and Medication Adherence Among People Living with HIV with Recent Experiences of Hunger Pellowski Jennifer A. PhD 1 Kalichman Seth C. PhD 2 Cherry Sabrina MPH 2 Conway-Washington Christopher BA 2 Cherry Chauncey Dr.PH 2 Grebler Tamar BA 2 Krug Larissa BA 2 1 Department of Psychiatry and Human Behavior, Alpert Medical School of Brown University, Providence, RI 2 Department of Psychology, University of Connecticut, Storrs, CT 26 6 2016 12 2016 01 12 2017 50 6 844853 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. Background Limited access to resources can significantly impact health behaviors. Previous research on food insecurity and HIV has focused on establishing the relationship between lacking access to nutritious food and antiretroviral (ARV) medication non-adherence in a variety of social contexts. Purpose This study aims to determine if several aspects of food insecurity co-occur with missed doses of medication on a daily basis among a sample of people living with HIV who have recently experienced hunger. Methods The current study utilized a prospective, observational design to test the daily relationship between food insecurity and medication non-adherence. Participants were followed for 45 days and completed daily assessments of food insecurity and alcohol use via interactive text message surveys and electronic medication adherence monitoring using the Wisepill. Results Fifty-nine men and women living with HIV contributed a total of 2,655 days of data. Results showed that severe food insecurity (i.e. hunger), but not less severe food insecurity (i.e. worrying about having food), significantly predicted missed doses of medication on a daily level. Daily alcohol use moderated this relationship in an unexpected way; when individuals were hungry and drank alcohol on a given day they were less likely to miss a dose of medication. Conclusions Among people living with HIV with recent experiences of hunger, this study demonstrates that there is a daily relationship between hunger and nonadherence to antiretroviral therapy. Future research is needed to test interventions designed to directly address the daily relationship between food insecurity and medication non-adherence. Food insecurity HIV/AIDS adherence alcohol use multi-level modeling HIV infections are largely concentrated among those with low socioeconomic status (SES) [1, 2]. Individuals living in poverty face a myriad of social disadvantages, including overcrowding, restricted access to nutritious food, and violence, all of which have direct implications for health and well-being [3–5]. People living in poverty with HIV have limited access to the resources necessary to “survive and thrive” [6]. For optimal adherence to medications, people living with HIV require tangible resources including food, shelter, money to purchase their medications, and transportation. Additionally, individuals must also have cognitive resources to cope with their disease and adequately carry out the necessary tasks associated with medication adherence There are also competing demands for these resources. For example, a person living with co-morbid HIV infection and a substance use disorder will have only limited funds and must decide between buying food and substances [7]. Food insecurity can be construed as a marker of socioeconomic marginalization. Food insecurity is “the limited availability of nutritionally adequate or safe food, or the inability to procure food in socially acceptable ways” [8]. Among people living with HIV, food insecurity is particularly detrimental because it can lead to malnutrition, which exacerbates the adverse health outcomes of immune suppression resulting from HIV infection. Both malnutrition and HIV impact immune system functioning and when these two conditions co-occur the impact on the immune system is compounded and synergistic [9]. In addition to the detrimental impacts that food insecurity has on the physical health of people living with HIV, food insecurity has also been associated with worse health behaviors, such as lower antiretroviral (ARV) medication adherence and higher levels of sexual risk behavior [10, 11]. In a systematic review of studies reporting on the relationship between food insecurity and ARV medication adherence, nine out of thirteen studies that adjusted for other markers of poverty found statistically significant relationships such that greater food insecurity predicted lower ARV adherence [11]. In one of the few longitudinal studies in this literature, Weiser, Yuan et al., (2013) followed 284 unstably housed participants for a median of 22 months in San Francisco, CA. Weiser, et al. [12] found that food insecurity was associated with higher odds of medication non-adherence, having incomplete HIV RNA viral suppression, and having a CD4 T-cell count of less than 200. From their review, Singer, Weiser and McCoy [11] concluded that across multiple settings in resource poor and resource rich countries, food insecurity is an important barrier to ARV adherence. Several explanations for this relationship have been offered including purposeful non-adherence due to prescriptions with food requirements, side effects, competing resource demands, and the exacerbation of hunger during the initiation phase of ARVs [13, 14]. Food insecurity is fairly common among substance users, however, it is unknown if there is an interaction between alcohol use and food insecurity. Among alcohol users in Atlanta, Kalichman et al. [15] found that 43% of participants experienced food insecurity. Alcohol use has been related to medication adherence among food insecure populations. In South Africa, Morojele, Kekvaletswe, & Nkosi [16] found that alcohol use was independently associated with ARV adherence when controlling for socioeconomic status, having a stable living situation and food insecurity. Alcohol use, particularly alcohol addiction, and food insecurity, are competing demands for limited resources that can have direct implications on ARV adherence. Alcohol use is an important factor to consider when observing the daily relationship between aspects of food insecurity and medication adherence. The relationship between food insecurity and ARV non-adherence is robust in the literature and previous studies have extensively controlled for other indicators of SES. Previous work has also established the causal relationship between food insecurity and ARV non-adherence [17,18]. However, it is unknown if days when a person lacks food are the same days when they miss their medications. The temporal relationship between food insecurity and non-adherence is assumed in interventions that provide food to improve adherence [19], but whether or not these two events occur on the same day is unknown. A true test of whether medication non-adherence occurs on days with limited access to food requires data collection of both food insecurity and ARV adherence at the day level in a prospective study design. The purpose of the current study was to test the daily relationship between several aspects of food insecurity and medication non-adherence among people living with HIV with recent experiences of hunger. Additionally, to better understand the interaction between alcohol use and food insecurity on medication adherence, daily alcohol use was tested as a moderator of the relationship between food insecurity and medication adherence, using alcohol use dichotomously (did not drink vs. drank alcohol) to determine if simply drinking alcohol on a given day interacts with aspects of food insecurity and continuously to determine if quantity of alcohol consumed interacts with food insecurity. Methods The current study employed an observational cohort design that included 45-days of daily text message surveys and daily electronic medication monitoring. Participants also completed a computer-assisted interview and provided their CD4 T-cell count and HIV RNA viral load retrieved from their medical records. Participants were compensated for all completed study activities and the University Institutional Review Board approved the study protocol. Population and Screening Study participants consisted of men and women living with HIV/AIDS in and around Atlanta, GA, an area with a substantial HIV epidemic [20]. Atlanta has the seventh highest number of new HIV diagnoses for large metropolitan areas [21]. Notices announcing the study were posted in infectious disease clinics and interested persons called the study site and were screened for the study. Eligibility criteria included (a) being 18 or older, (b) name-matching proof of positive HIV status and photo identification, (c) currently taking antiretroviral medications (ARVs), (d) self-reported missing at least one dose of medication in the past month, (e) self-reported being hungry in the past month but unable to eat because he/she could not afford food, (f) self-reported consumption of at least one alcoholic beverage within the past month, and (g) willingness to use a Wisepill device for medication management. If the participant was deemed eligible, the screener gave the participant an enrollment appointment. Enrollment Session Participants were consented into the study and then completed an audio-computer assisted self-interview (ACASI) to gather basic demographic information including gender, race/ethnicity, education, and age. After the participants completed the ACASI, the assessor instructed them on how to complete the daily text message surveys and each participant completed a practice survey in the office. A project cell phone was provided to each participant. Finally, the assessor explained electronic medication monitoring (see below). The participants were instructed to use an electronic device to hold their ARV medication for the entirety of the study and to refill it themselves as needed. Daily Data Collection We provided participants with a study cell phone to deliver daily interactive text-message assessments. Participants received the same text message survey every day for 45 days and the questions referred to the previous day’s activities. This type of daily measurement has been used to assess a wide variety of health behaviors and cognitions including sexual risk behaviors, depression, coping skills, and alcohol consumption [22–25]. In addition, we monitored daily ARV adherence using electronic devises for assessing adherence contemporaneous with food insecurity at the day level. A time frame of 45 days was chosen for two reasons. In order to estimate the association of day-level events, the number of assessments (days) within the individual had to be large rather than focusing on a large individual sample size. Additionally, individuals living in poverty who often rely on government or state assistance financial resources (i.e. food stamps, Social Security benefits), receive these payments only once or twice a month. It is possible that severity of food insecurity differ on the days leading up to payments versus days directly following payments. In order to capture these potential fluctuations, the time period in which participants were observed had to be greater than one month. Daily aspects of food insecurity We assessed several aspects of food insecurity on a daily basis. Participants answered three questions adapted from the USDA Household Food Security Scale [26]. These items are answered with a yes or a no response and are as follows: “I worried about my food running out yesterday,” “I ate less than I needed to yesterday,” and “I was hungry, but could not eat because I couldn’t afford food yesterday.” Responses were dichotomous, 1=Yes, 0=No. These particular aspects were selected because we felt they might have the highest variability on a daily basis, however, they do not capture other important aspects of food insecurity, such as quality of food. Daily alcohol use To assess alcohol, participants were asked “How many alcohol drinks did you have yesterday? If you did not drink say 0.” These responses were used in two ways: 1) dichotomously to indicate 0 = ‘Did not drink’ or 1 = ‘Drank alcohol’ to determine whether or not drinking on a given day impacts adherence and 2) continuously to capture whether or not quantity of alcohol matters. Medication adherence Daily medication adherence data was collected using the Wisepill device [27–29]. This device functions as a medication pillbox with sensors for opening and closing. When a participant opens the Wisepill device to take their medication, a date and time stamp is created which is then sent to a central server via general packet radio service (GPRS). This information is then accessible via a secure, internet-based interface. These data were downloaded from the website and compared to each participant’s dosing schedule. For the individual level analyses this information was used to calculate an aggregate adherence percentage for the 45-day period. For the day level analyses this information was coded dichotomously for missed doses. For example, if a participant opened their Wisepill device on a day only once for a medication that is prescribed twice daily the medication was coded as a missed day. Wisepill signal lapses of 48 hours or greater were investigated by a phone call to the participant to determine whether the lapse was due to a technical failure (battery failure, loss of cellular signal) or if it was due to a behavioral cause (pocket dose, missed dose) [27]. An adjusted adherence that factored in reported pocket doses was calculated [28]. Chart Abstracted HIV RNA Viral Load and CD4 T Cell Count During the study, participants were asked to obtain a copy of their latest HIV RNA viral load and CD4 T cell counts from their health care provider and these records could be no older than 3 months. These records had to be marked with a clinic stamp to verify its authenticity. For consistency across viral load chart values, we defined undetectable viral load as <50 copies/mL. Financial Assessment Following the 45-day observation period, participants completed a financial assessment with a trained staff member. Participants were asked about where the source of any received financial resources, the amount of money or money equivalent (i.e. food stamps) received, and when they received that payment for the past 45 days. Participants were also asked about any jobs (full time/part time/side work), short-term disability, Social Security, unemployment assistance, food stamps, and any other types of monetary assistance that they received. Statistical Analyses To characterize the sample, means and rates were calculated using demographic information collected in the ACASI, the chart abstracted lab reports, and the financial assessment. Rates were also calculated for the day-level behaviors from the text message surveys and the Wisepill device. To analyze the relationship between aspects of food insecurity and medication adherence, two-levels of analysis were used. At the individual level, behaviors were collapsed across the 45-day period. Bivariate and multivariate regressions were conducted to determine if there was a relationship between aspects of food insecurity and medication adherence. Next, multilevel modeling was conducted using event records as a statistical unit of analysis (or “level 1 unit”), nested within participants (or “level 2 unit). All of these analyses were conducted using R version 3.1.3. [30] and the library “lme4: Linear mixed-effects models using Eigen and S4” (version 1.7) [31]. Prior to testing the daily relationship between aspects of food insecurity and missed doses of medication, the impact of time was examined to determine whether there was any indication of reactivity to the daily measurement. If habituation or sensitization effects occur over the course of the study, we would expect to observe a significant upward or downward linear trend in the rates at which missed doses of medication and aspects of food insecurity occurred. To test the daily relationships between aspects of food insecurity, alcohol use, and missed doses of medication, the data was restructured. Because the food insecurity and alcohol use questions refer to the prior days’ experiences whereas the Wisepill data is reported for the current day, the data was restructured so that all of the data for the aspects of food insecurity, alcohol use and adherence referred to the same day. Fixed and random effects were used such that the intercepts and slopes of the models were modeled first as random effects in logistic regressions. If the residual variances for the slopes are determined to not vary across level 2 units, they were converted to fixed effects [32], with the exception of time variables in order to retain the growth model component of the model and the nested nature of the data. B estimates and standard errors are reported for model estimates. Missing data days were included in these analyses and we assume them to be missing at random (MAR). Missing data were handled using multiple imputation by chained equations using all variables in the dataset to estimate missing data using the R package “mice: Multivariate Imputation by Chained Equations” version 2.22 [33, 34]. Predictive mean matching, a semi-parametric imputation method, was used in order to preserve any non-linear relationships present in the data [33]. Results For the current study, 578 individuals were screened for eligibility and a total of 508 were deemed not eligible for the study (See Figure 1). The majority of participants (N=260) screened out because they did not meet multiple eligibility criteria. Of the three major criteria for entry into the study, 112 individuals did not report severe food insecurity in the previous month (i.e. hunger and unable to buy more food), 49 individuals did not report drinking alcohol in the past month, and 39 individuals were not currently taking antiretroviral medications at the time of screening. There were 70 individuals who met all eligibility criteria for study entry and were scheduled for an in-office enrollment appointment. Of these individuals, a total of 59 individuals attended their appointment and enrolled into the study; 57 individuals completed follow-up assessments. Demographic Characteristics The majority of the 59 participants enrolled in the current study were male (N=40; 67.8%; See Table 1), identified as African American/Black (96.6%), and were middle-aged (M= 48, SD = 6.6). In terms of education, 35.6% of participants completed high school or their GED and an additional 44.1% completed some education after high school. On average, participants had an income of $1,472.66 for the 45 days. Most (76.3%) received a type of disability payment. Additionally, 61% of participants received food stamps during the course of the study. More than half (61%) of participants were on ARV regimens that required taking their prescriptions with food. The sample was fairly healthy; the majority of participants had undetectable HIV RNA viral loads (81.1%). Participants also had relatively high CD4 T-cell counts, with only six having a CD4 T-cell count of 200 or less, the cut-off for an AIDS diagnosis. Day-Level Behavior Of the 2,655 days of collected data, there were 530 days (20.0%) where doses of medication were missed. Participants reported one or more aspects of food insecurity on 718 days (27.0%). Of these days, worrying about food running out occurred on 597 days (22.5%), eating less than needed because there wasn’t enough food occurred on 533 days (20.1%), and being hungry but not able to afford more food occurred on 298 days (11.2%). Participants reported drinking alcohol on 566 days (21.3%) with an average of 2.27 drinks (SD = 1.29) when they were drinking. The quantity of alcohol drinks was skewed (skewness = 2.7, SE = 0.050) with responses ranging from 0 to 12 drinks consumed in a day. However, given the meaningfulness of the minimum score (i.e. 0), this variable was not transformed [33]. Aspects of food insecurity and missed doses of medication occurred on at least one day for a large percentage of participants. Over the course of the study, 74.6% of participants reported at least one day of any of the three food insecurity indicators. Nearly all participants had at least one day where they missed a dose of medication (91.5%). A large proportion of participants had at least one day where they drank alcohol (83.1%). Of the 2655 days of text message surveys, 170 days were missing due to uncompleted surveys (6.4%). Additionally, 16 surveys were partially completed (0.6%). Aspects of Food Insecurity and Medication Adherence Across the 45 Days To replicate past findings, analyses were first conducted across the 45-day observational period using the individual as the case; medication adherence and aspects of food insecurity were collapsed across days. On average, participants were 83.2% adherent to the medication in their Wisepill device (SD =13.62; Range 50% – 100%). In a bivariate regression, all aspects of food insecurity combined predicted medication adherence such that greater food insecurity was associated with poorer medication adherence (β = −0.312, p = 0.016; Table 2). This relationship remained significant even after controlling for education and income over the 45 days (β = −0.345, p = 0.011). Bivariate regressions were also conducted using each individual indicator of food insecurity. Worrying about running out of food predicted medication adherence, such that greater worry predicted lower medication adherence (β = −0.309, p = 0.017). Eating less because there was not enough food and being hungry did not predict medication adherence in individual bivariate regressions (β = −0.189, p = 0.152, β = −0.216, p = 0.100, respectively). Aspects of Food Insecurity and Medication Adherence on a Daily Level Test of reactivity to the daily measurement Day in the study predicted missed doses such that the longer participants were in the study, the more likely they were on a given day to miss doses of medication (B = 0.017, SE = 0.006, p = 0.008). The same model was also run with all aspects of food insecurity combined. Day in the study predicted food insecurity such that the longer participants were in the study, the less likely they were to report at least one of the aspects of food insecurity on any given day (B = −0.029, SE = 0.010, p = 0.004). This analysis was also replicated for each indicator of food insecurity; “worried about not having enough to eat” and “eating less because there was not enough food” both showed significant time trends. However, day in the study did not significantly predict the question about hunger. All subsequent analyses control for the effect of day. Figure 2 shows the time trends of missed doses of medication and food insecurity. In addition, days that participants received financial payments and income did not impact adherence. Daily relationship between aspects of food insecurity and missed medication doses A multilevel model was constructed to test the daily relationship between the aspects of food insecurity measured in this study combined and missed doses of medication. Results showed that all the aspects of food insecurity as a composite did not predict missed doses of medication on a day-to-day basis (B = −0.053, SE = 0.236, p=0.82). After controlling for income over the 45 days and education, this relationship was still non-significant (B = −0.090, SE = 0.245, p=0.71). Multilevel models were also conducted using each individual indicator of food insecurity. Worrying about food running out and eating less because there wasn’t enough food were not significant predictors of daily missed doses (worry: B = −0.198, SE = 0.321, p=0.54; ate less: B = 0.019, SE = 0.242, p=0.94). However, severe food insecurity as indexed by the hunger item did significantly predict daily missed doses of medication such that daily reported hunger was associated with a greater likelihood of missing medications on a daily level (B = 0.519, SE = 0.254, p=0.046; See Model 1 in Table 3). This relationship held even after controlling for income over the 45 days and education (B = 0.570, SE = 0.285, p = 0.050; See Model 2 in Table 3). Alcohol use as a moderator Whether or not a participant was drinking on a given day was tested as a moderator of the daily relationship between hunger and missing a dose of medication (See Table 4). There were 85 days in which participants drank and were hungry and 18 participants contributed at least one of those days. Among the entire sample, drinking on a given day predicted missing a dose of medication on a daily level (B = 0.731, SE = 0.319, p = 0.022). When accounting for daily alcohol use, hunger was no longer a significant predictor of missing a dose of medication (B = 0.673, SE = 0.474, p = 0.156). There was a significant interaction between drinking and experiencing hunger (B = −1.919, SE = 0.575, p<0.001). This interaction, however, was in an unexpected direction; those who drank alcohol and experienced hunger on a given day had the lowest likelihood of missing a dose on that day. Those that did not drink alcohol but did experience hunger on a given day were the most likely to miss a dose of medication on that day. This analysis was also conducted using the alcohol measure continuously to test whether quantity of alcohol use on a given day would interact with hunger, however, this interaction was not significant (B = 0.100, SE = 0.706, p = 0.887). Discussion The current study replicates findings from both resource poor and resource rich settings that demonstrate that food insecurity is related to medication adherence, even when controlling for other markers of socioeconomic status, such as income and education. Additionally, this study extends the literature by looking at this relationship on a daily level. Although worrying about not having enough to eat and eating less did not predict adherence, hunger was directly related to missing medication on a daily basis. This day-level finding supports a direct temporal relationship between food and medication adherence, such that having tangible resources impacts health behaviors [8]. This also lends credence to the potential explanations for the relationship between food insecurity and ARV adherence. Food requirements of prescriptions and avoiding side effects associated with taking medications on an empty stomach are potential explanations of the link between day level food insecurity and day level medication non-adherence. Although these explanations were not specifically tested in this study, the current findings support future research to tease apart these mechanisms. The results of this study support the assumption that hunger and medication adherence are directly related, due to circumstances that occur at the daily level. However, worrying about having enough to eat and eating less were not associated with non-adherence on a daily basis but was associated with non-adherence across time. Less severe food insecurity may, therefore, serve as a marker for social disadvantage that predicts non-adherence in general with severe food insecurity predicting non-adherence on the same day. Additionally, worrying about having enough food to eat taps into the psychological consequences of the stress of poverty. Worrying about having enough to eat may impact medication adherence in a cumulative way rather than on a day-by-day basis. Future research should replicate these findings in other locations, including rural areas of both resource rich and resource poor countries to confirm this daily relationship. There was potential reactivity to the daily measurement both for missed doses of medication and food insecurity defined as indicating yes to any of the three daily food insecurity questions. Participants were more likely to miss doses of medications the longer that they were in the study. A possible explanation for this finding is that using the Wisepill device may have served as a reminder for participants to take their medications at the beginning of the study, but as the study progressed, the Wisepill device became less of a reminder. Reactivity to using the Wisepill has also been found in other studies with HIV positive adults in China and Uganda [27, 28]. Future research may take these factors into account and perhaps lag the time between starting to use the Wisepill device and starting the daily food insecurity measurement to reduce the potential effect of reactivity on adherence. There was also an interaction between hunger and drinking alcohol, but in a direction opposite of what was predicted. Those who drank alcohol and were hungry on a given day were the least likely to miss a dose of medication on that day. This interaction was not observed for quantity of alcohol use. One possible explanation for the direction of the interaction observed is that participants who were hungry but still drinking may have had less severe hunger than those that were not drinking. Differences in hunger severity may have masked the expected relationship, where greater alcohol use and food insecurity would predict non-adherence. Also possible is the role of alcohol use in coping with extreme poverty. Although maladaptive, drinking may reduce the stress of poverty, even if diverting resources away from food. Thus, stress reduction could account for the paradoxically better adherence among drinkers experiencing food insecurity. Finally, we cannot dismiss the possibility that the interaction between drinking and hunger could have been a statistical artifact in the data particularly because the interaction disappears when using the drinking variable continuously. Thus, our study should be replicated with more sensitive measurements for hunger before concluding that this counterintuitive interaction is reliable. There were limitations to this study that should be taken into account. First, the eligibility criteria was specifically chosen so that there would be enough events of each variable of interest (i.e. aspects of food insecurity, alcohol use, and non-adherence) within the daily study, however, this may have limited the generalizability of these findings because only 10% of individuals screened were enrolled into the study. Individuals screened out may have been using alcohol and experiencing hunger but reported being adherent to their medications. It would be important to understand the facilitators for these individuals to maintain their adherence during economic hardships and while using substances, such as high resiliency, greater social support, or lower levels of depression, however, this study was limited in its design to elucidate this. Additionally, days of hunger and non-adherence may both be impacted by a third, confounding variable, such as daily housing instability. This type of severe disruption to daily life should be measured and incorporated into future daily analyses. Another factor to take into consideration in the interpretation of these results is the skew of the quantity of alcohol consumed variable. The decision not to transform these data was due to the meaningful nature of “0” in this variable. To transform a variable a constant must be added to minimum value of the variable changing the interpretation of the results [35]. However, this has implications for the possible results of the interaction. The skew of the data may have contributed to the non-significance of this interaction. The reactivity seen in the daily data may have also contributed to the non-significance. Another limitation of the study is that it was conducted in one large southeastern U.S. city. It is possible that the results do not generalize to other cities within the U.S. or suburban or rural areas of the U.S. or elsewhere. Additionally, almost the entire sample for this study identified as African-American (96%). There may be racial factors that play a role in the relationship between aspects of food insecurity and medication adherence such as race-based medical mistrust [36], racial prejudice/discrimination [37], and racial stigma [38, 39]. Finally, only 3 aspects of food insecurity were used to explore its daily relationship with medication non-adherence. This limited number of items was used to reduce the burden of answering these questions on a daily basis and these particular items were chosen because we felt that they would change the most over time. These items, however, do not address quality of food or the procurement of food in socially unacceptable ways. Although eliminating these aspects led to a less nuanced picture of daily food insecurity, the items that were used in the daily assessment provide range of food insecurity, albeit a smaller one than if the full assessment was used. In order to improve medication adherence among people living with HIV, interventions on multiple levels are required. Structural changes are necessary to make food more available to those living in poverty. Subsidized grocery stores, expansion of food stamp programs, and increasing the availability of quality food in food deserts are all needed steps to increase access to food. Providers can also screen for food insecurity prior to prescribing medications that require food to collaboratively decide on appropriate regimens with their patients. Behavioral and psycho-education based interventions can address individual decision-making around medication adherence and also skills-building for how to navigate the social services system as well as how to seek out potential resources such as food pantries. This research was supported by the National Institute of Alcohol Abuse and Alcoholism grant R01 AA021471 (PI: Seth Kalichman), and the National Institute of Mental Health grants T32 MH074387 (PI: Seth Kalichman) and T32 MH078788 (PI: Larry Brown). This manuscript contains content submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy for Jennifer Pellowski to the University of Connecticut. The authors thank Dr. Tania B. Huedo-Medina for her statistical assistance. Figure 1 Screening and enrollment into the 45-day observational study Figure 2 Time Trends of Daily Day with Study Payments Marked (Days 15, 29, 43) Note: Each line represents the total number of events that occurred each day across participants Table 1 Demographic characteristics of participants enrolled in current study (N=59) N % Gender  Male 40 67.8  Female 12 20.3  Transgender 7 11.9 Race/Ethnicity  White 1 1.7  African American/Black 57 96.6  Hispanic/Latino 1 1.7 Education  Less than High School 12 20.3  High school/GED 21 35.6  More than High School 26 44.1 Types of Income Received Over 45 days  Job (Part Time/Full Time) 9 15.3  Any Disability (Short-time/SSDI) 45 76.3  Unemployment 1 1.7  Food Stamps 36 61.0  Other 5 8.5 Medication Food Requirement  Regimen Requires Food 36 61  Regimen Does Not Require Food 23 39 Viral Loada  Detectable 10 18.9  Undetectable 43 81.1 CD4 (T-Cell) Counta  200 or less 6 11.5  Greater than 200 46 88.5 M SD Income Over 45 Days $1,472.66 1323.93 Age Range (30–59) 48 6.62 Baseline Visual Analog Scale 84.54 18.62 a Note: There was missing viral load data from 6 participants and missing CD4 data from 7 participants Table 2 Bivariate and multivariate regressions predicting ART medication adherence across the 45-day observational period Predictors Bivariate Regression Multivariate Regression B SE Beta B SE Beta Any Food Indicator −4.25 1.714 −0.312* −4.698 1.786 −0.345*  Worry −4.211 1.716 −0.309*  Ate Less −2.573 1.772 −0.189  Hungry −2.943 1.762 −0.216^ Income over 45 days −1.293 1.841 −0.095 −2.178 1.82 −0.16 Education 0.398 1.804 0.029 1.781 1.789 0.133 ^ p<0.10, * p<0.05 Table 3 Fixed effects and random effects estimates for multilevel model of daily hunger predicting daily missed doses of medication with imputed data Model 1 Model 2 Fixed Effects Parameter B SE B SE Intercept −2.312*** 0.257 −2.442*** 0.265 Level 1 (Daily Level Data) Day 0.016* 0.006 0.017** 0.007 Hunger 0.519* 0.254 0.54* 0.274 Level 2 (Individual Level Data) Income Over 45 days 0.138 0.207 Education −0.168 0.198 Random Effects Intercept (σ^2) 2.238 1.510 2.195 1.481 Level 1 (Daily Level Data) Day 0.0004 0.021 0.0004 0.020 Day* Hungry 0.0002 0.016 0.0002 0.015 * p<0.05, ** p<0.01, *** p<0.001 Table 4 Fixed effects and random effects estimates for multilevel model with drinking (Yes/No) as a moderator Fixed Effects Parameter B SE Intercept −2.627*** 0.288 Level 1 (Daily Level Data) Day 0.19* 0.008 Hunger 0.673 0.474 Drinking (Yes/No) 0.731* 0.319 Drinking* Hunger −1.919*** 0.575 Day* Hunger 0.010 0.016 Day* Drinking −0.003 0.012 Level 2 (Individual Level Data) Income Over 45 days 0.137 0.197 Education −0.165 0.194 Random Effects Intercept (σ^2) 2.274 1.508 Level 1 (Daily Level Data) Day 0.001 0.025 Drinking* Hunger 0.298 0.546 * p<0.05, *** p<0.001 Authors’ Statement of Conflict of Interest and Adherence to Ethical Standards Authors Pellowski, Kalichman, S. Cherry, Conway-Washington, C. Cherry, Grebler, and Krug declare that they have no conflicts of interest. All participants were treated according to the APA guidelines for treatment of human participants and with the Helsinki Declaration of 1975. 1 Centers for Disease Control and Prevention Characteristics associated with HIV infection among heterosexuals in urban areas with high AIDS prevalence – 24 cities, United States, 2006–2007 MMWR Morb Mortal Wkly Rep 2011 60 1045 1049 21832975 2 Pellowski JA Kalichman SC Matthews KA Adler N A pandemic of the poor: social disadvantage and the U.S. HIV epidemic Am Psychol 2013 68 197 209 23688088 3 Lynch JW Smith GD Kaplan GA House JS Income inequality and mortality: importance to health of individual income, psychosocial environment, or material conditions BMJ 2000 320 1200 1204 10784551 4 Reif S Whetten K Lowe K Ostermann J Association of unmet needs for support services with medication use and adherence among HIV-infected individuals in the southeastern United States AIDS Care 2006 18 277 283 16809104 5 Young AS Sullivan G Bogart LM Koegel P Kanouse DE Needs for services reported by adults with severe mental illness and HIV Psychiatr Serv 2005 56 99 101 15637201 6 Johnson BT Redding CA DiClemente RJ A network-individual-resource model for HIV prevention AIDS Behav 2010 14 204 221 20862606 7 Hobfoll SE Conservation of resources: a new attempt at conceptualizing stress Am Psychol 1989 44 513 524 2648906 8 Weiser SD Hatcher A Frongillo EA Food insecurity is associated with greater acute care utilization among HIV-infected homeless and marginally housed individuals in San Francisco J Gen Intern Med 2013 28 91 98 22903407 9 Anabwani G Navario P Nutrition and HIV/AIDS in sub-Saharan Africa: An overview Nutrition 2005 21 96 99 15661483 10 Anema A Vogenthaler N Frongillo EA Kadiyala S Weiser SD Food insecurity and HIV/AIDS: current knowledge gaps, and research priorities Curr HIV/AIDS Rep 2009 6 224 231 19849966 11 Singer AW Weiser SD McCoy SI Does food insecurity undermine adherence to antiretroviral therapy? A systematic review AIDS Behav 2015 19 1510 1526 25096896 12 Weiser SD Yuan C Guzman D Food insecurity and HIV clinical outcomes in a longitudinal study of urban homeless and marginally housed HIV-infected individuals AIDS 2013 27 2953 2958 23921612 13 Kalichman SC Washington C Grebler T Medication adherence and health outcomes of people living with HIV who are food insecure and prescribed antiretrovirals that should be taken with food Infect Dis Ther 2015 4 79 91 14 Young S Wheeler AC McCoy SI Weiser SD A review of the role of food insecurity in adherence to care and treatment among adult and pediatric populations living with HIV and AIDS AIDS Behav 2014 18 Suppl 5 S505 515 23842717 15 Kalichman SC Grebler T Amaral CM Food insecurity and antiretroviral adherence among HIV positive adults who drink alcohol J Behav Med 2014 37 1009 1018 24022091 16 Morojele NK Kekwaletswe CT Nkosi S Associations between alcohol use, other psychosocial factors, structural factors and antiretroviral therapy (ART) adherence among South African ART recipients AIDS Behav 2014 18 519 524 23934270 17 Cantrell RA Sinkala M Megazinni K A pilot study of food supplementation to improve adherence to antiretroviral therapy among food insecure adults in Lusaka, Zambia J Acquir Immune Defic Syndr 2008 49 2 190 5 18769349 18 Tirivayi N Koethe JR Groot W Clinic-based food assistance is associated with increased medication adherence among HIV-infected adults on long-term antiretroviral therapy in Zambia J AIDS Clin Res 2012 3 171 23227443 19 de Pee S Grede N Mehra D Bloem MW The enabling effect of food assistance in improving adherence and/or treatment completion for antiretroviral therapy and tuberculosis treatment: a literature review AIDS Behav 2014 18 Suppl 5 S531 41 24619602 20 Georgia Department of Public Health HIV/AIDS Surveillance Georgia 2012 Available at: https://dph.georgia.gov/data-fact-sheet-summaries Accessibility verified August 17, 2015 21 Centers for Disease Control and Prevention Diagnoses of HIV infection in the United States and dependent areas 2013 Available at: http://www.cdc.gov/hiv/library/reports/surveillance/2013/surveillance_Report_vol_25.html Accessibility verified August 17, 2015 22 Affleck G Zautra A Tennen H Armeli S Multilevel daily process designs for consulting and clinical psychology: a preface for the perplexed J Consult Clin Psychol 1999 67 746 754 10535241 23 Barta WD Portnoy DB Kiene SM Tennen H Abu-Hasaballah K Ferrer R A daily process investigation of alcohol-involved sexual risk behavior among economically disadvantaged problem drinkers living with HIV/AIDS AIDS Behav 2008 12 729 740 18071894 24 Parsons JT Rosof E Mustanski B The temporal relationship between alcohol consumption and HIV-medication adherence: a multilevel model of direct and moderating effects Health Psychol 2008 27 628 637 18823189 25 Tennen H Affleck G Armeli S Carney MA A daily process approach to coping: linking theory, research, and practice Am Psychol 2000 55 626 636 10892205 26 Bickel D Nord M Price C Hamilton W Cook J United States Department of Agriculture (USDA) Guide to measuring household food security Available at: http://www.fns.usda.gov/sites/default/files/FSGuide_0.pdf Accessibility verified March 25, 2016 27 Bachman DeSilva M Gifford AL Keyi X Feasibility and acceptability of a real-time adherence device among HIV-positive IDU patients in China AIDS Res Treat 2013 957862 23956851 28 Haberer JE Kahane J Kigozi I Real-time adherence monitoring for HIV antiretroviral therapy AIDS Behav 2010 14 1340 1346 20809380 29 Pellowski JA Kalichman SC White D Amaral CM Hoyt G Kalichman MO Real-time medication adherence monitoring intervention: test of concept in people living with HIV infection J Assoc Nurses AIDS Care 2014 25 646 651 25043931 30 R Core Team R: A language and environment for statistical computing R Foundation for Statistical Computing Vienna, Austria Available at: http://www.r-project.org Accessibility verified August 17, 2015 31 Bates D Maechler M Bolker B Package ‘lme4: Linear mixed-effects models using Eigen and S4’ available through the Comprehensive R Archive Network. http://cran.r-project.org/web/packages/lme4/index.html Accessibility verified August 17, 2015 32 Snijders T Bosker R Multilevel analysis: An introduction to basic and advanced multilevel modeling London, England Sage 1999 33 van Buuren S Groothuis-Oudshoorn K mice: Multivariate Imputation by Chained Equations in R J Stat Softw 2011 45 1 67 34 van Buuren S Groothuis-Oudshoorn K mice: Multivariate Imputation by Chained Equations available through the Comprehensive R Archive Network. http://cran.r-project.org/web/packages/mice/index.html Accessibility verified August 17, 2015 35 Osborne J Notes on the use of data transformations Practical Assessment, Research, & Evaluation 2002 8 6 1 Retrieved from http://PAREonline.net/getvn.asp?v=8&n=6 36 Eaton LA Driffin DD Kegler C The role of stigma and medical mistrust in the routine health care engagement of black men who have sex with men Am J Public Health 2015 105 e75 e82 37 Bogart LM Wagner GJ Galvan FH Klein DJ Longitudinal relationships between antiretroviral treatment adherence and discrimination due to HIV-serostatus, race, and sexual orientation among African-American men with HIV Ann Behav Med 2010 40 184 190 20552416 38 Earnshaw VA Bogart LM Dovidio JF Williams DR Stigma and racial/ethnic HIV disparities: moving toward resilience Am Psychol 2013 68 225 236 23688090 39 Earnshaw VA Smith LR Chaudoir SR Amico KR Copenhaver MM HIV stigma mechanisms and well-being among PLWH: a test of the HIV stigma framework AIDS Behav 2013 17 1785 1795 23456594

PMC005xxxxxx/PMC5127766.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 0110674 8015 Virology Virology Virology 0042-6822 1096-0341 27792904 5127766 10.1016/j.virol.2016.10.015 NIHMS827317 Article Epstein-Barr Virus Latent Membrane Protein 2A (LMP2A) Enhances IL-10 production through the Activation of Bruton’s Tyrosine Kinase and STAT3 Incrocci Ryan a Barse Levi b Stone Amanda a Vagvala Sai a Montesano Michael a Subramaniam Vijay b Swanson-Mungerson Michelle a* a Department of Microbiology and Immunology, Chicago College of Osteopathic Medicine, Midwestern University, 555 31stStreet, Downers Grove, IL 60515 b Department of Biomedical Sciences, College of Health Sciences, Midwestern University, 555 31stStreet, Downers Grove, IL 60515 * Corresponding author. Michelle Swanson-Mungerson, Department of Microbiology and Immunology, Chicago College of Osteopathic Medicine, Midwestern University, 555 31st Street, Downers Grove, IL 60515 (p) +630-515-6129; (f) +630-515-7245; mswans@midwestern.edu 4 11 2016 25 10 2016 1 2017 01 1 2018 500 96102 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. Previous data demonstrate that Epstein-Barr Virus Latent Membrane Protein 2A (LMP2A) enhances IL-10 to promote the survival of LMP2A-expressing B cell lymphomas. Since STAT3 is an important regulator of IL-10 production, we hypothesized that LMP2A activates a signal transduction cascade that increases STAT3 phosphorylation to enhance IL-10. Using LMP2A-negative and –positive B cell lines, the data indicate that LMP2A requires the early signaling molecules of the Syk/RAS/PI3K pathway to increase IL-10. Additional studies indicate that the PI3K-regulated kinase, BTK, is responsible for phosphorylating STAT3, which ultimately mediates the LMP2A-dependent increase in IL-10. These data are the first to show that LMP2A signaling results in STAT3 phosphorylation in B cells through a PI3K/BTK-dependent pathway. With the use of BTK and STAT3 inhibitors to treat B cell lymphomas in clinical trials, these findings highlight the possibility of using new pharmaceutical approaches to treat EBV-associated lymphomas that express LMP2A. B cell Epstein-Barr Virus Latent Membrane Protein 2A Interleukin-10 Bruton’s Tyrosine Kinase (BTK) Signal Transducer and Activator of Transcription 3 (STAT3) Introduction Epstein-Barr virus (EBV) is a member of the gamma herpesvirus family that infects B lymphocytes in more than 90% of the world population (Kang and Kieff, 2015; Kempkes and Robertson, 2015). While most infections are asymptomatic or result in infectious mononucleosis (Thorley-Lawson et al., 2013), EBV is associated with the development of multiple autoimmune diseases and lymphomas of the immune system (Ascherio and Munger, 2015; Thorley-Lawson and Gross, 2004). One mechanism by which Epstein-Barr virus could contribute to these diseases is by influencing B cell function and survival. After initial infection, EBV transitions to a latent state in which few viral genes are expressed. There are multiple latency gene patterns identified in either normal latency and/or EBV-associated pathology (Price and Luftig, 2015). The EBV latency protein Latent membrane protein 2A (LMP2A), which contains 12 transmembrane domains with a long amino terminal domain, is expressed in multiple programs of EBV latency (Babcock et al., 1998; Babcock, Hochberg, and Thorley-Lawson, 2000; Babcock, Miyashita-Lin, and Thorley-Lawson, 2001; Bell et al., 2006; Decker, Klaman, and Thorley-Lawson, 1996; Hochberg et al., 2004; Niedobitek et al., 1997), suggesting the importance of this protein in normal latency and EBV-associated diseases. LMP2A acts as a B cell receptor (BCR) mimic to increase the survival of latently-infected B cells (Mancao et al., 2005; Mancao and Hammerschmidt, 2007; Portis and Longnecker, 2004). Previous studies indicate that LMP2A constitutively activates many of the kinases and signal transduction molecules used by the BCR, including Syk, Ras, PI3K, BTK, and AKT (Fruehling and Longnecker, 1997; Merchant and Longnecker, 2001; Portis and Longnecker, 2004) to promote B cell survival (Merchant and Longnecker, 2001; Portis and Longnecker, 2004). Additional studies demonstrate that LMP2A signaling in B cells directly results in an increase in anti-apoptotic factors, such as BCL-2 and BCL-xL (Bultema, Longnecker, and Swanson-Mungerson, 2009; Portis and Longnecker, 2004; Swanson-Mungerson, Bultema, and Longnecker, 2010). More recently, it has become appreciated that LMP2A indirectly promotes B cell survival by increasing the production of pro-survival cytokines, such as IL-10 (Incrocci, McCormack, and Swanson-Mungerson, 2013). Due to the redundant expression of LMP2A throughout many phases of the EBV life cycle, targeting its pro-survival abilities in EBV-associated tumors may be of therapeutic benefit. Pharmacological therapies to treat tumors typically induce the death of cells that are rapidly proliferating or by blocking signal transduction pathways that directly increase tumor cell survival (Dominguez-Brauer et al., 2015; Pistritto et al., 2016). However, an alternative approach may be to block the production of pro-survival factors, such as IL-10. Inhibiting the LMP2A-dependent increase in IL-10 that promotes tumor survival may provide a potential novel approach to enhance current chemotherapeutic strategies for EBV-associated lymphomas. Therefore, we sought to identify the signals required for LMP2A to increase IL-10 production in B cell lymphomas. Our findings indicate for the first time, that LMP2A activates BTK to phosphorylate STAT3 in B cell tumors, which mediates the LMP2A-dependent increase in IL-10. Due to the identification of new therapeutics that target BTK and STAT3 in clinical trials, these findings have important implications for innovative treatments of LMP2A-expressing B cell tumors. Materials and Methods Cell lines All B cell lines used in this study have been described previously (Ikeda and Longnecker, 2007). Briefly, the BJAB B cell lymphoma line was transduced with either the vector backbone alone or the vector backbone with LMP2A. Transduced cells were selected using hygromycin and gentamycin and LMP2A expression was identified in all selected cells by immunofluorescence and found to be similar in levels when compared to lymphoblastoid cell lines (Incrocci, McCormack, and Swanson-Mungerson, 2013). Independent clones were isolated and maintained in cRPMI media supplemented with hygromycin (0.4 ug/ml) (EMD Millipore) and gentamycin (2 ug/ml) (Sigma Aldrich) at 37°C/5% CO2. The lymphoblastoid cell lines LCL3 (LMP2A-positive) and ES1 (LMP2A-negative) were generously provided by Richard Longnecker (Northwestern University-Chicago, Illinois) and were maintained in cRPMI at 37° C/5% CO2. Analysis of IL-10 production 5x104 LMP2A-negative or LMP2A-positive B cell lines described above were grown in a 96-well plate in the absence or presence of an optimized concentration of the following pharmacological inhibitors: Syk (R788-EMD Millipore, 5 uM), Ras (Manumycin A-EMD Millipore, 0.5 uM), PI3K (Wortmannin-EMD Millipore, 10 uM), STAT3 (Stattic-EMD Millipore, up to 1.75 uM), or BTK (Ibrutinib-Selleckchem, up to 10 uM). All inhibitors were initially diluted in DMSO and final dilutions were reached in cRPMI. The cells that were not exposed to inhibitor were exposed to an equivalent amount of DMSO to control for the potential effects of DMSO alone in all experiments. After 24 hours, 20 ul of supernatants were isolated for use in an LDH assay (Sigma-Aldrich, St. Louis MO) to analyze inhibitor toxicity and 100 ul of supernatants were isolated for analysis using the Human IL-10 Ready, Set, Go® ELISA kit (Ebioscience). All of the inhibitors used in the study did not induce toxicity at 24 hours as determined by LDH assay (data not shown), confirming that any differences seen in IL-10 levels was not due to toxicity of the assay. Western Blot analysis for STAT3 All B cell lines (10 x 106) were incubated for 24 hours in the absence or presence of Syk (5 uM), Ras (Manumycin A- 0.5 uM), Wortmannin (10 uM), Ibrutinib (10 uM) or Stattic (1.75 uM) and then lysed in RIPA buffer (Pierce, Rockford IL) containing 25 mM Tris–HCl (pH 7.6), 150 mM NaCl, 1% NP-40, 1% sodium deoxycholate and 0.1% SDS, along with HaltTM Protease and Phosphatase inhibitors (Pierce, Rockford IL). Protein concentrations from cell lysates were quantified using a BCA assay (Pierce, Rockford IL) and equal amounts of protein were analyzed by Western blot. Antibodies against phosphorylated- and total-STAT3 (Cell Signaling Technology, Massachusetts) or GAPDH (Biolegend, San Diego, CA) were diluted 1:1000 in blocking buffer containing milk, followed by washes in TBST. A secondary anti-rabbit IgG-HRP conjugated Ab (Pierce, Rockford IL) diluted in blocking buffer (1:5000) was added to the blot followed by ECL imaging (GE Healthcare, Buckinghamshire UK) using a Bio-Rad imager. Image analysis and quantification was done using ImageJ software (National Institutes of Health). The ratio of phosphorylated STAT3 to Total STAT3 was determined by dividing the value of the band for phosphorylated STAT3 according to ImageJ software by the value of the band for total STAT3 under different culture conditions. Statistics Each experiment was performed at least three times. All ELISA experiments were initially analyzed by a two-way analysis of variance (ANOVA), followed by a Bonferonni post hoc test to compare individual groups using Prism Graphpad software. All experiments that demonstrated a p<0.05 by ANOVA and only findings that reached p<0.05 by Bonferroni comparisons were considered significant. Results LMP2A is a BCR mimic that directly promotes B cell survival through a RAS/PI3K dependent pathway (Portis and Longnecker, 2004) and indirectly by inducing IL-10 production in both LMP2A-expressing primary murine transgenic B cells and in B cell lymphomas (Incrocci, McCormack, and Swanson-Mungerson, 2013). An additional study at the time indicated that the Syk inhibitor, R406, blocked the enhancement of IL-10 production in EBV-positive tumor cells from post-transplant lymphoproliferative disease (PTLD) patients (Hatton et al., 2011). These studies suggested that LMP2A uses Syk to increase IL-10. However, PTLD cells express numerous EBV latency proteins and therefore we sought to confirm that LMP2A uses Syk to increase IL-10 by using multiple B cell lines that only express LMP2A. Two independently-derived B cell lines that express LMP2A (LMP2A1.1 and LMP2A1.2) and two LMP2A-negative control cell lines (Vector.1 and Vector.2) were exposed to the Syk inhibitor (R788, Fostamatinib) for 24 hours and IL-10 production was analyzed using ELISA. As shown in Figure 1A, the addition of the Syk inhibitor significantly decreased the LMP2A-dependent increase in IL-10 production. Additionally, the Syk inhibitor did not significantly affect IL-10 production by LMP2A-negative B cell lines, indicating that this effect is specific for LMP2A. Since Syk activation leads to Ras stimulation during BCR signaling (Beitz et al., 1999; Mocsai, Ruland, and Tybulewicz, 2010), we next tested if LMP2A required Ras to increase IL-10 production. As shown in Figure 1B, the addition of the Ras inhibitor, Manumycin A, also decreased the LMP2A-dependent increase in IL-10 production, without significantly affecting IL-10 production in the LMP2A-negative cell lines. One downstream target of Ras activation is p38K phosphorylation and signaling (Shin et al., 2005). Since p38K controls IL-10 regulation in monocytes, macrophages, and dendritic cells (Chi et al., 2006; Foey et al., 1998; Jarnicki et al., 2008; Kim et al., 2005), we initially tested if LMP2A activates p38K in the LMP2A-expressing B cells used in this study. However, Western blot analysis of protein lysates from LMP2A-negative and –positive B cell lines indicate that LMP2A does not increase p38K phosphorylation (data not shown). Therefore, the Ras-mediated increase in IL-10 production must be via an alternative pathway downstream of Ras. Ras leads to the activation and signaling via PI3K in B cells. Our previous work using the PI3K inhibitor, Ly294002, suggested that LMP2A utilizes PI3K to increase IL-10 production (Incrocci, McCormack, and Swanson-Mungerson, 2013). However, due to the identification that Ly294002 can also affect NF-kB activation (Avni, Glucksam, and Zor, 2012), we wanted to confirm our previous findings through the use of the more PI3K-specific inhibitor Wortmannin (Avni, Glucksam, and Zor, 2012). Therefore, LMP2A-expressing and non-expressing B cell lines were incubated in the absence or presence of Wortmannin for 24 hours and IL-10 levels were once again analyzed by ELISA. As shown in Figure 1C, Wortmannin decreased the LMP2A-dependent increase in IL-10 production, confirming that LMP2A requires PI3K activation to increase IL-10 levels. Our previous findings also indicate that LMP2A increases the levels of IL-10 RNA transcripts (Incrocci, McCormack, and Swanson-Mungerson, 2013) and therefore, we hypothesized that PI3K activation may result in the activation of a transcription factor that regulates IL-10 RNA levels. Since Signal Transducer and Activator of Transcription 3 (STAT3) increases IL-10 transcription (Gabrysova et al., 2014) and is regulated by PI3K (Hart et al., 2011), we hypothesized that the LMP2A-dependent activation of PI3K would result in STAT3 phosphorylation, which is a pre-requisite for STAT3 induction as a transcription factor (Harrison, 2012). To determine if LMP2A induces STAT3 activation, we analyzed the levels of STAT3 phosphorylation at tyrosine residue 705, which is critical for STAT3 dimerization and activation (Sellier et al., 2013), using Western blot analysis. As shown in Figure 2A, LMP2A-expressing cells demonstrate higher levels of STAT3 phosphorylation than non-LMP2A-expressing B cell lines. We also analyzed if LMP2A increased the phosphorylation of the regulatory site of STAT3 (serine 727), but we were unable to detect phosphorylation of this residue in any of the cell lines tested (data not shown), suggesting that LMP2A does not induce the phosphorylation of serine 727 to influence IL-10 production. If LMP2A-mediated activation of the Syk/Ras/PI3K pathway results in STAT3 phosphorylation and the subsequent regulation of IL-10 as suggested in Figure 1, then the addition of inhibitors to these signal transduction molecules should block STAT3 phosphorylation. As shown in Figure 2B, the addition of inhibitors to Syk, Ras, and PI3K blocked the LMP2A-mediated increase in STAT3 phosphorylation, but did not consistently affect the basal levels of STAT3 phosphorylation in non-LMP2A-expressing cells. Most importantly, we originally hypothesized that the PI3K-mediated activation of STAT3 was responsible for the LMP2A-mediated increase in IL-10 production. If this hypothesis is correct, then the addition of a STAT3 inhibitor should block the LMP2A-dependent increase in IL-10. When LMP2A-expressing cells were incubated in the presence of the STAT3 inhibitor, Stattic (Schust et al., 2006), the levels of STAT3 phosphorylation were decreased (Figure 2B) and the amount of IL-10 were significantly decreased in a concentration-dependent manner in only the LMP2A-expressing cells (Figure 2C), suggesting that LMP2A requires STAT3 phosphorylation to increase IL-10 production. Recent findings indicate that the link between PI3K and STAT3 phosphorylation is mediated by TEC family kinases (Vogt and Hart, 2011). LMP2A increases the activation of the TEC family kinase, Bruton’s tyrosine kinase (BTK) (Mano, 1999; Merchant and Longnecker, 2001), suggesting that this kinase may be the link connecting PI3K activation and STAT3 phosphorylation in LMP2A-expressing cell lines. If LMP2A activates BTK to enhance STAT3 phosphorylation, then a BTK inhibitor would block the LMP2A-mediated increase in STAT3 phosphorylation. To test this possibility, LMP2A-negative and –positive B cell lines were exposed to the BTK inhibitor, Ibrutinib, for 24 hours and STAT3 phosphorylation was analyzed by Western blot analysis. As shown in Figure 3A, the addition of Ibrutinib blocked the LMP2A-dependent phosphorylation of STAT3, without decreasing STAT3 phosphorylation in non-LMP2A-expressing cells. Furthermore, since Ibrutinib decreased STAT3 phosphorylation in LMP2A-expressing cells, we would expect that the addition of Ibrutinib would significantly inhibit the LMP2A-dependent increase in IL-10 production. As shown in Figure 3B, the addition of Ibrutinib decreased IL-10 production in LMP2A-expressing cells in a concentration-dependent manner, further confirming that LMP2A activates the PI3K/BTK/STAT3 pathway to increase IL-10 production in LMP2A-expressing B cells. LMP2A is not expressed in isolation and therefore, it is important to determine if LMP2A increases STAT3 phosphorylation in the presence of other latency proteins. Therefore, to determine if LMP2A increases STAT3 phosphorylation in the context of latent infection, we tested whether lymphoblastoid cell lines (LCLs) that expressed LMP2A demonstrate an increase in STAT3 phosphorylation. As shown in Figure 4A, STAT3 phosphorylation was higher in the LMP2A-positive LCL (LCL3) when compared to the LMP2A-negative lymphoblastoid cell line ES1. Furthermore, based on our studies using B cell lymphoma lines, we confirmed that the LMP2A-dependent increase in STAT3 phosphorylation was BTK-dependent. As shown in Figure 4A, the addition of Ibrutinib blocked the LMP2A-dependent increase in STAT3 phosphorylation in the LMP2A-positive LCL (LCL3), but did not show a significant effect on the LMP2A-negative LCL (ES1). These data suggest that in context of EBV latent infection, LMP2A utilizes BTK to increase the phosphorylation of STAT3. Taken together, we propose that LMP2A activates the PI3K/BTK/STAT3 pathway to enhance STAT3 phosphorylation, which may have significant impact on the survival and drug-resistance of EBV-associated tumors. Discussion These data are the first to show that LMP2A signaling results in the phosphorylation of STAT3 through BTK signaling in B cell lymphomas. These findings are exciting in light of the fact that Ibrutinib has received breakthrough designation from the FDA to expedite clinical trials (Zucca and Bertoni, 2013) and is being used in clinical trials alone or in combination to treat B cell tumors, such as chronic myeloid leukemia and acute myeloid leukemia (Maddocks and Blum, 2014). One advantage of using BTK as a drug target for lymphoma treatment, is due to its relatively limited expression in different cell types (Mohamed et al., 2009) and therefore has exhibited limited toxicity thus far (Wang et al., 2015). Since PI3K and STAT3 are used by many different cell types (Cantley, 2002; Levy and Lee, 2002), drugs that target these signal transduction molecules may have greater side effects (Do, Mace, and Rexwinkle, 2016; Ogura et al., 2015). Therefore, our results suggest that Ibrutinib may be an asset in the treatment of EBV-associated lymphomas with minimal side effects, which has not yet been investigated. One potential EBV-associated lymphoma that may be amenable to new treatments that target BTK and STAT3 is Hodgkin’s Lymphoma (HL). EBV is detected in 40–50% of HL cases (Khan and Coates, 1994; Kuppers et al., 2002), while greater than 80–99% of HL cases are positive for EBV in AIDS patients (Bibas and Antinori, 2009). The malignant cell in HL, the Hodgkin-Reed Sternberg (HRS) cell is infected with EBV in EBV-associated HL and expresses LMP2A. These EBV-infected HL cells overexpress STAT3, when compared to EBV-negative HL (Garcia et al., 2003), suggesting that STAT3 contributes to HRS cell development and survival in EBV-associated HL. Therefore, our findings indicate that inhibition of the LMP2A-mediated increase in STAT3 by targeting BTK may provide a novel approach to current therapies. In EBV-positive HRS cells, both LMP2A and another latency protein, LMP1, is expressed. Others have demonstrated that LMP1 can also induce STAT3 activation through Jak3-mediated tyrosine phosphorylation (Wang et al., 2010) and Protein Kinase C-δ-dependent serine phosphorylation (Kung, Meckes, and Raab-Traub, 2011). Therefore, it is interesting that EBV evolved to generate multiple pathways that result in STAT3 activation in this particular tumor and points to the critical requirement for the activation of STAT3 in these tumors. Our findings with both Ibrutinib and Stattic indicate that targeting HRS cells with these inhibitors may provide therapeutic value in the course of treatment of EBV-associated B cell tumors. Interestingly, multiple herpesviruses induce activation of STAT3 (Punjabi et al., 2007; Sen et al., 2012). For example, Kaposi’s sarcoma-associated herpesvirus (KSHV), which is a member of the gamma herpesvirus family, like EBV, also constitutively activates STAT3 during latent infection (Punjabi et al., 2007). However, in the case of KSHV, the activation of STAT3 is through the production of a soluble factor and is not directly due to signaling by KSHV latency proteins (Punjabi et al., 2007). Taken together, it appears that multiple oncogenic herpesvirus family members have evolved to utilize STAT3 activation for host cell survival. However, they achieve this goal via different mechanisms. Our data consistently demonstrate the LMP2A significantly increases IL-10 production (Figures 1–3). Even though an autocrine loop of IL-10 production and IL-10 receptor signaling on LMP2A-expressing B cells could result in increased STAT3 phosphorylation, we believe that the increase in STAT3 phosphorylation is directly due to LMP2A for the following reasons. First, IL-10 receptor signaling induces STAT3 phosphorylation through Jak kinase activation and not BTK activation (Murray, 2006; Murray, 2007). If the increase in STAT3 phosphorylation was due to increased IL-10 receptor signaling, the BTK inhibitor would not block the increased STAT3 phosphorylation observed in LMP2A-expressing cells. Additionally, IL-10 receptor signaling decreases p38K activation (Kontoyiannis et al., 2001) and Western blot analysis did not demonstrate any change in p38K phosphorylation in LMP2A-expressing cells (data not shown). Taken together, the findings indicate that LMP2A induces the phosphorylation of STAT3 through PI3K/BTK activation. Our previous studies demonstrate the significance of LMP2A in IL-10 production for B cell survival (Incrocci, McCormack, and Swanson-Mungerson, 2013). Additionally, it is possible that the LMP2A-mediated increase may also influence the anti-tumor response induced by cytotoxic T cells (CTLs), since IL-10 dampens cell-mediated immunity (Groux et al., 1998; Klinker and Lundy, 2012). A recent study confirms that LMP2A increases IL-10 production, but the change in IL-10 was insufficient to influence CTL activity in their study (Rancan et al., 2015). It is possible that the influence of IL-10 on CTL function may be dependent on the program of EBV latency and the interplay of latency proteins. For example, LCLs express all of the latency proteins, while Hodgkin’s lymphoma express LMP1 and LMP2A. Therefore, additional studies that assess the LMP2A-dependent increase in IL-10 on CTL activity are warranted. Conclusions Our findings indicate, for the first time, that LMP2A induces the phosphorylation of STAT3 through the activation of PI3K and BTK to enhance IL-10 production (Figure 4B). In light of the use of both BTK inhibitors in clinical trials, these findings potentially highlight novel pharmaceutical approaches to treat EBV-associated lymphomas that express LMP2A. This research was funded by NIH grant 1R15CA149690-01 (M.S.M), the Biomedical Sciences Program in the College of Health Sciences (L.B., V.S.), and the Kenneth A. Suarez Summer Research Program at Midwestern University (A.S., M.M., and S.V.). Fig 1 LMP2A uses Syk, Ras, and PI3K to increase IL-10 production. LMP2A-negative (Vector.1 and Vector.2) and LMP2A-positive (LMP2A1.1 and LMP2A1.2) B cell lines were incubated in the absence or presence of the (A) Syk inhibitor, R788 (Fostamatinib), (B) Ras inhibitor, Manumycin A, or (C) PI3K inhibitor, Wortmannin for 24 hours and supernatants were isolated for analysis using an IL-10 ELISA. Data are representative of 3 independent experiments with similar results. * indicates p<0.05 when compared to LMP2A-negative B cell lines, ** indicates p<0.05 when compared to the same cell line that was incubated in the absence of inhibitor Fig 2 LMP2A uses PI3K to activate STAT3 to increase IL-10 production. (A) Protein from LMP2A-negative or –positive B cell lines (10 x 106 cells) was isolated and analyzed for phosphorylated (Y705)-STAT3, total STAT3, or GAPDH by Western blot analysis. (B) Protein from LMP2A-negative or –positive B cell lines (10 x 106 cells) was isolated after a 24 hour incubation in the absence or presence of R788 (5 uM), Manumycin A (0.5 uM), Wortmannin (10 uM), or Stattic (1.75 uM). Phosphorylated (Y705)-STAT3, total STAT3, or GAPDH were analyzed by Western Blot analysis. (C) LMP2A-negative and –positive B cell lines were incubated in the absence or presence of increasing concentrations of the STAT3 inhibitor, Stattic, for 24 hours and supernatants were isolated for analysis using an IL-10 ELISA. Data are representative of 3 independent experiments with similar results. For Western blot analysis, the ratio of phosphorylated (Y705)-STAT3/total STAT3 is given underneath each respective band. * indicates p<0.05 when compared to LMP2A-negative B cell lines, ** indicates p<0.05 when compared to the same cell line that was incubated in the absence of inhibitor Fig 3 LMP2A activates BTK to phosphorylate STAT3 and increase IL-10 production. (A) Protein from LMP2A-negative or –positive B cell lines (10 x 106 cells) was isolated after a 24 hour incubation in the absence or presence of the BTK inhibitor, Ibrutinib (10 uM). Total and phosphorylated-STAT3 were analyzed by Western Blot analysis. (B) LMP2A-negative and –positive B cell lines were incubated in the absence or presence of increasing concentrations of Ibrutinib for 24 hours and supernatants were isolated for analysis using an IL-10 ELISA. Data in (A–B) are representative of 3 independent experiments with similar results. For Western blot analysis, the ratio of phosphorylated (Y705)-STAT3/total STAT3 is given underneath each respective band. * indicates p<0.05 when compared to LMP2A-negative B cell lines, ** indicates p<0.05 when compared to the same cell line that was incubated in the absence of inhibitor Fig 4 LMP2A increases STAT3 phosphorylation in the context of latent EBV infection. (A) Protein from LMP2A-negative (ES1) or –positive (LCL3) lymphoblastoid cell lines (10 x 106 cells) was isolated after a 24 hour incubation in the absence or presence of Ibrutinib (1.75 uM) or Stattic (1.75 uM). Phosphorylated (Y705)-STAT3, total STAT3, or GAPDH were analyzed by Western Blot analysis. Data are representative of three independent experiments with similar results. For Western blot analysis, the ration of phosphorylated (Y705)-STAT3/total STAT3 is given underneath each respective band. (B) Model of the mechanism by which LMP2A increases IL-10 production. LMP2A activates Syk to increase Ras and subsequent PI3K activation. PI3K activation results in the stimulation of BTK to phosphorylate STAT3 to increase the amounts of IL-10 RNA in LMP2A-expressing cells. Highlights LMP2A stimulates the PI3K-dependent activation of BTK to phosphorylate STAT3. The addition of the STAT3 inhibitor, Stattic, or the BTK inhibitor, Ibrutinib, blocks the LMP2A-mediated increase in STAT3 phosphorylation and IL-10 production in B cell lymphomas. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. Ascherio A Munger KL 2015 EBV and Autoimmunity Curr Top Microbiol Immunol 390 Pt 1 365 85 26424654 Avni D Glucksam Y Zor T 2012 The phosphatidylinositol 3-kinase (PI3K) inhibitor LY294002 modulates cytokine expression in macrophages via p50 nuclear factor kappaB inhibition, in a PI3K-independent mechanism Biochem Pharmacol 83 1 106 14 22005520 Babcock GJ Decker LL Volk M Thorley-Lawson DA 1998 EBV persistence in memory B cells in vivo Immunity 9 3 395 404 9768759 Babcock GJ Hochberg D Thorley-Lawson AD 2000 The expression pattern of Epstein-Barr virus latent genes in vivo is dependent upon the differentiation stage of the infected B cell Immunity 13 4 497 506 11070168 Babcock GJ Miyashita-Lin EM Thorley-Lawson DA 2001 Detection of EBV infection at the single-cell level. Precise quantitation of virus-infected cells in vivo Methods Mol Biol 174 103 10 11357631 Beitz LO Fruman DA Kurosaki T Cantley LC Scharenberg AM 1999 SYK is upstream of phosphoinositide 3-kinase in B cell receptor signaling J Biol Chem 274 46 32662 6 10551821 Bell AI Groves K Kelly GL Croom-Carter D Hui E Chan AT Rickinson AB 2006 Analysis of Epstein-Barr virus latent gene expression in endemic Burkitt’s lymphoma and nasopharyngeal carcinoma tumour cells by using quantitative real-time PCR assays J Gen Virol 87 Pt 10 2885 90 16963746 Bibas M Antinori A 2009 EBV and HIV-Related Lymphoma Mediterr J Hematol Infect Dis 1 2 e2009032 21416008 Bultema R Longnecker R Swanson-Mungerson M 2009 Epstein-Barr virus LMP2A accelerates MYC-induced lymphomagenesis Oncogene 28 11 1471 6 19182823 Cantley LC 2002 The phosphoinositide 3-kinase pathway Science 296 5573 1655 7 12040186 Chi H Barry SP Roth RJ Wu JJ Jones EA Bennett AM Flavell RA 2006 Dynamic regulation of pro- and anti-inflammatory cytokines by MAPK phosphatase 1 (MKP-1) in innate immune responses Proc Natl Acad Sci U S A 103 7 2274 9 16461893 Decker LL Klaman LD Thorley-Lawson DA 1996 Detection of the latent form of Epstein-Barr virus DNA in the peripheral blood of healthy individuals J Virol 70 5 3286 9 8627812 Do B Mace M Rexwinkle A 2016 Idelalisib for treatment of B-cell malignancies Am J Health Syst Pharm 73 8 547 55 26933132 Dominguez-Brauer C Thu KL Mason JM Blaser H Bray MR Mak TW 2015 Targeting Mitosis in Cancer: Emerging Strategies Mol Cell 60 4 524 36 26590712 Foey AD Parry SL Williams LM Feldmann M Foxwell BM Brennan FM 1998 Regulation of monocyte IL-10 synthesis by endogenous IL-1 and TNF-alpha: role of the p38 and p42/44 mitogen-activated protein kinases J Immunol 160 2 920 8 9551930 Fruehling S Longnecker R 1997 The immunoreceptor tyrosine-based activation motif of Epstein-Barr virus LMP2A is essential for blocking BCR-mediated signal transduction Virology 235 2 241 51 9281504 Gabrysova L Howes A Saraiva M O’Garra A 2014 The regulation of IL-10 expression Curr Top Microbiol Immunol 380 157 90 25004818 Garcia JF Camacho FI Morente M Fraga M Montalban C Alvaro T Bellas C Castano A Diez A Flores T Martin C Martinez MA Mazorra F Menarguez J Mestre MJ Mollejo M Saez AI Sanchez L Piris MA 2003 Hodgkin and Reed-Sternberg cells harbor alterations in the major tumor suppressor pathways and cell-cycle checkpoints: analyses using tissue microarrays Blood 101 2 681 9 12393683 Groux H Bigler M de Vries JE Roncarolo MG 1998 Inhibitory and stimulatory effects of IL-10 on human CD8+ T cells J Immunol 160 7 3188 93 9531274 Harrison DA 2012 The Jak/STAT pathway Cold Spring Harb Perspect Biol 4 3 Hart JR Liao L Yates JR 3rd Vogt PK 2011 Essential role of Stat3 in PI3K-induced oncogenic transformation Proc Natl Acad Sci U S A 108 32 13247 52 21788516 Hatton O Phillips LK Vaysberg M Hurwich J Krams SM Martinez OM 2011 Syk activation of phosphatidylinositol 3-kinase/Akt prevents HtrA2-dependent loss of X-linked inhibitor of apoptosis protein (XIAP) to promote survival of Epstein-Barr virus+ (EBV+) B cell lymphomas J Biol Chem 286 43 37368 78 21908615 Hochberg D Middeldorp JM Catalina M Sullivan JL Luzuriaga K Thorley-Lawson DA 2004 Demonstration of the Burkitt’s lymphoma Epstein-Barr virus phenotype in dividing latently infected memory cells in vivo Proc Natl Acad Sci U S A 101 1 239 44 14688409 Ikeda M Longnecker R 2007 Cholesterol is critical for Epstein-Barr virus latent membrane protein 2A trafficking and protein stability Virology 360 2 461 8 17150237 Incrocci R McCormack M Swanson-Mungerson M 2013 Epstein-Barr virus LMP2A increases IL-10 production in mitogen-stimulated primary B-cells and B-cell lymphomas J Gen Virol 94 Pt 5 1127 33 23303827 Jarnicki AG Conroy H Brereton C Donnelly G Toomey D Walsh K Sweeney C Leavy O Fletcher J Lavelle EC Dunne P Mills KH 2008 Attenuating regulatory T cell induction by TLR agonists through inhibition of p38 MAPK signaling in dendritic cells enhances their efficacy as vaccine adjuvants and cancer immunotherapeutics J Immunol 180 6 3797 806 18322186 Kang MS Kieff E 2015 Epstein-Barr virus latent genes Exp Mol Med 47 e131 25613728 Kempkes B Robertson ES 2015 Epstein-Barr virus latency: current and future perspectives Curr Opin Virol 14 138 44 26453799 Khan G Coates PJ 1994 The role of Epstein-Barr virus in the pathogenesis of Hodgkin’s disease J Pathol 174 3 141 9 7823246 Kim L Del Rio L Butcher BA Mogensen TH Paludan SR Flavell RA Denkers EY 2005 p38 MAPK autophosphorylation drives macrophage IL-12 production during intracellular infection J Immunol 174 7 4178 84 15778378 Klinker MW Lundy SK 2012 Multiple mechanisms of immune suppression by B lymphocytes Mol Med 18 123 37 22033729 Kontoyiannis D Kotlyarov A Carballo E Alexopoulou L Blackshear PJ Gaestel M Davis R Flavell R Kollias G 2001 Interleukin-10 targets p38 MAPK to modulate ARE-dependent TNF mRNA translation and limit intestinal pathology EMBO J 20 14 3760 70 11447117 Kung CP Meckes DG Jr Raab-Traub N 2011 Epstein-Barr virus LMP1 activates EGFR, STAT3, and ERK through effects on PKCdelta J Virol 85 9 4399 408 21307189 Kuppers R Schwering I Brauninger A Rajewsky K Hansmann ML 2002 Biology of Hodgkin’s lymphoma Ann Oncol 13 Suppl 1 11 8 12078890 Levy DE Lee CK 2002 What does Stat3 do? J Clin Invest 109 9 1143 8 11994402 Maddocks K Blum KA 2014 Ibrutinib in B-cell Lymphomas Curr Treat Options Oncol 15 2 226 37 24481980 Mancao C Altmann M Jungnickel B Hammerschmidt W 2005 Rescue of “crippled” germinal center B cells from apoptosis by Epstein-Barr virus Blood 106 13 4339 44 16076866 Mancao C Hammerschmidt W 2007 Epstein-Barr virus latent membrane protein 2A is a B-cell receptor mimic and essential for B-cell survival Blood 110 10 3715 21 17682125 Mano H 1999 Tec family of protein-tyrosine kinases: an overview of their structure and function Cytokine Growth Factor Rev 10 3–4 267 80 10647781 Merchant M Longnecker R 2001 LMP2A survival and developmental signals are transmitted through Btk-dependent and Btk-independent pathways Virology 291 1 46 54 11878875 Mocsai A Ruland J Tybulewicz VL 2010 The SYK tyrosine kinase: a crucial player in diverse biological functions Nat Rev Immunol 10 6 387 402 20467426 Mohamed AJ Yu L Backesjo CM Vargas L Faryal R Aints A Christensson B Berglof A Vihinen M Nore BF Smith CI 2009 Bruton’s tyrosine kinase (Btk): function, regulation, and transformation with special emphasis on the PH domain Immunol Rev 228 1 58 73 19290921 Murray PJ 2006 Understanding and exploiting the endogenous interleukin-10/STAT3-mediated anti-inflammatory response Curr Opin Pharmacol 6 4 379 86 16713356 Murray PJ 2007 The JAK-STAT signaling pathway: input and output integration J Immunol 178 5 2623 9 17312100 Niedobitek G Kremmer E Herbst H Whitehead L Dawson CW Niedobitek E von Ostau C Rooney N Grasser FA Young LS 1997 Immunohistochemical detection of the Epstein-Barr virus-encoded latent membrane protein 2A in Hodgkin’s disease and infectious mononucleosis Blood 90 4 1664 72 9269787 Ogura M Uchida T Terui Y Hayakawa F Kobayashi Y Taniwaki M Takamatsu Y Naoe T Tobinai K Munakata W Yamauchi T Kageyama A Yuasa M Motoyama M Tsunoda T Hatake K 2015 Phase I study of OPB-51602, an oral inhibitor of signal transducer and activator of transcription 3, in patients with relapsed/refractory hematological malignancies Cancer Sci 106 7 896 901 25912076 Pistritto G Trisciuoglio D Ceci C Garufi A D’Orazi G 2016 Apoptosis as anticancer mechanism: function and dysfunction of its modulators and targeted therapeutic strategies Aging (Albany NY) 8 4 603 19 27019364 Portis T Longnecker R 2004 Epstein-Barr virus (EBV) LMP2A mediates B-lymphocyte survival through constitutive activation of the Ras/PI3K/Akt pathway Oncogene 23 53 8619 28 15361852 Price AM Luftig MA 2015 To be or not IIb: a multi-step process for Epstein-Barr virus latency establishment and consequences for B cell tumorigenesis PLoS Pathog 11 3 e1004656 25790223 Punjabi AS Carroll PA Chen L Lagunoff M 2007 Persistent activation of STAT3 by latent Kaposi’s sarcoma-associated herpesvirus infection of endothelial cells J Virol 81 5 2449 58 17151100 Rancan C Schirrmann L Huls C Zeidler R Moosmann A 2015 Latent Membrane Protein LMP2A Impairs Recognition of EBV-Infected Cells by CD8+ T Cells PLoS Pathog 11 6 e1004906 26067064 Schust J Sperl B Hollis A Mayer TU Berg T 2006 Stattic: a small-molecule inhibitor of STAT3 activation and dimerization Chem Biol 13 11 1235 42 17114005 Sellier H Rebillard A Guette C Barre B Coqueret O 2013 How should we define STAT3 as an oncogene and as a potential target for therapy? JAKSTAT 2 3 e24716 24069560 Sen N Che X Rajamani J Zerboni L Sung P Ptacek J Arvin AM 2012 Signal transducer and activator of transcription 3 (STAT3) and survivin induction by varicella-zoster virus promote replication and skin pathogenesis Proc Natl Acad Sci U S A 109 2 600 5 22190485 Shin I Kim S Song H Kim HR Moon A 2005 H-Ras-specific activation of Rac-MKK3/6-p38 pathway: its critical role in invasion and migration of breast epithelial cells J Biol Chem 280 15 14675 83 15677464 Swanson-Mungerson M Bultema R Longnecker R 2010 Epstein-Barr virus LMP2A imposes sensitivity to apoptosis J Gen Virol 91 Pt 9 2197 202 20484564 Thorley-Lawson DA Gross A 2004 Persistence of the Epstein-Barr virus and the origins of associated lymphomas N Engl J Med 350 13 1328 37 15044644 Thorley-Lawson DA Hawkins JB Tracy SI Shapiro M 2013 The pathogenesis of Epstein-Barr virus persistent infection Curr Opin Virol 3 3 227 32 23683686 Vogt PK Hart JR 2011 PI3K and STAT3: a new alliance Cancer Discov 1 6 481 6 22348200 Wang ML Blum KA Martin P Goy A Auer R Kahl BS Jurczak W Advani RH Romaguera JE Williams ME Barrientos JC Chmielowska E Radford J Stilgenbauer S Dreyling M Jedrzejczak WW Johnson P Spurgeon SE Zhang L Baher L Cheng M Lee D Beaupre DM Rule S 2015 Long-term follow-up of MCL patients treated with single-agent ibrutinib: updated safety and efficacy results Blood 126 6 739 45 26059948 Wang Z Luo F Li L Yang L Hu D Ma X Lu Z Sun L Cao Y 2010 STAT3 activation induced by Epstein-Barr virus latent membrane protein1 causes vascular endothelial growth factor expression and cellular invasiveness via JAK3 And ERK signaling Eur J Cancer 46 16 2996 3006 20709526 Zucca E Bertoni F 2013 Toward new treatments for mantle-cell lymphoma? N Engl J Med 369 6 571 2 23924008

PMC005xxxxxx/PMC5127771.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 9602153 20344 Subst Use Misuse Subst Use Misuse Substance use & misuse 1082-6084 1532-2491 27682897 5127771 10.1080/10826084.2016.1222622 NIHMS823407 Article Client Acceptability for Integrating Antiretroviral Therapy in Methadone Maintenance Therapy Clinics in Sichuan, China Lin Chunqing a Li Li a Cao Xiaobin b a Semel Institute for Neuroscience and Human Behavior, University of California at Los Angeles, Los Angeles, California, USA b National Center for AIDS Prevention and Control, Chinese Center for Disease Control and Prevention, Beijing, China CONTACT: Xiaobin Cao, caoxiaobin@chinaaids.cn, National Center for AIDS/STD Prevention and Control, Chinese Center for Disease Control and Prevention, 155 Changbai Road, Changping District, Beijing 102206, China 18 10 2016 28 9 2016 2 1 2017 02 7 2017 52 1 119126 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. Background Using methadone maintenance therapy (MMT) clinics to deliver antiretroviral therapy (ART) has proven to be effective for promoting treatment initiation and adherence in drug users living with HIV. Objectives The objective of this study was to investigate the HIV-positive client acceptability of integrated ART services and to identify the reasons for and factors associated with service acceptability. Methods A total of 86 HIV-positive MMT clients were recruited from 12 MMT clinics in Sichuan Province, China. They participated in a cross-sectional survey that queried their willingness to receive seven different types of MMT-based ART services. The reasons for their willingness/unwillingness to accept these services were documented. The association between service acceptability and background characteristics was examined. Results The most accepted integrated services were ART-related counseling (75.6%) and referral (73.2%). Concerns regarding the provider’s lack of ART expertise and confidentiality issues were common barriers for the acceptance of MMT-based ART services. A trust relationship with MMT providers was a reason for service acceptance. Service acceptability was associated with a poorer perceived health status. Conclusions/Importance ART-related services, based on the client perspective, can be delivered at MMT clinics. However, service provider training and the protection of confidentiality must be strengthened for the effective implementation of integrated service delivery. Acceptability antiretroviral therapy China methadone maintenance therapy service integration The provision of optimal care to HIV-infected drug users has long been a major challenge (Vlahov & Celentano, 2006). Compared to individuals who do not use drugs, HIV-positive drug users are less likely to receive antiretroviral therapy (ART) (Gruskin, Ferguson, Alfven, Rugg, & Peersman, 2013; Zhang et al., 2011). For drug users who receive ART, support is particularly important because these individuals are at particular risk for non-adherence (Azar et al., 2015; Jiamsakul et al., 2014). A lack of adherence subsequently leads to drug resistance and suboptimal treatment outcomes (Sarang, Rhodes, & Sheon, 2013; Wolfe, Carrieri, & Shepard, 2010). Methadone maintenance therapy (MMT) clinics provide a promising infrastructure to promote ART for drug users living with HIV because the clients receive their daily methadone dose at the clinic and establish a trust relationship with the MMT providers (Berg, Litwin, Li, Heo, & Arnsten, 2011; Tran et al., 2012). Literature has documented that MMT contributes to a more rapid initiation of ART among HIV-infected drug users (Uhlmann et al., 2010; Zhao et al., 2015) and a decreased rate of ART discontinuation (Reddon et al., 2014). A randomized controlled trial used this strategy to provide directly observed ART in MMT clinics and reported that this strategy was more efficacious than self-administered ART for improving adherence and reducing the HIV viral load among MMT clients (Berg et al., 2011). Due to the benefits of MMT on ART, researchers have called for an initiative to integrate ART-related services into MMT clinics to maximize HIV treatment outcomes for HIV-positive drug users (Bachireddy et al., 2014; Lambdin, Mbwambo, Josiah, & Bruce, 2015). To implement this integrated service in real-world healthcare settings, the vital first step is to understand the acceptability of MMT-based ART services and preferences among its target users. The objective of the study was to investigate the acceptability of MMT-based ART services among HIV-positive MMT clients. This study was conducted in China. Since 2004, China has developed a nation-wide network of 758 community-based MMT clinics in 28 provinces, which have cumulatively treated more than 384,500 clients (Wu, Sullivan, Wang, Rotheram-Borus, & Detels, 2007; Yin et al., 2015). However, MMT-based ART services remain largely unavailable in the country (Zhao et al., 2015), and little is known about the clients’ level of acceptance of different types of integrated ART services. In this study, we interviewed HIV-positive MMT clients regarding acceptability and their reasons for accepting/rejecting different types of MMT-based ART services. Factors associated with the acceptability of the services were also identified. Methods Study setting and participants The study was conducted in Sichuan Province, China, from January to September 2013. Twelve MMT clinics with the highest HIV caseload were included in the study. The study participant eligibility criteria were as follows: (1) 20 years or older; (2) an HIV-positive diagnosis confirmed by Western blot; and (3) a current client of one of the 12 selected MMT clinics. Current ART status was not a criterion for inclusion or exclusion of study participants. Procedure At the time of the study, a total of 143 MMT clients met the eligibility criteria in the 12 selected MMT clinics, and the MMT service providers informed all clients of the study. The study information was communicated verbally and with a printed flyer. Approximately 60% (N = 90) of the clients demonstrated an interest in participating and were referred to our study recruiters. Our study recruiters met with the potential participants in a private room and explained the study purpose and procedure, the nature of confidentiality, the potential benefits and risks and their right to refuse to participate or withdraw without any penalty. Written informed consent was obtained prior to data collection. Ethics approval was obtained from the Institutional Review Boards of University of California, Los Angeles and the National Center for AIDS/STD Control and Prevention, Chinese Center for Disease Control and Prevention. Data collection The participants completed a brief survey, which took approximately 20 minutes to finish. The survey was conducted in a private room of the MMT clinic. The survey was administered using the Computer Assisted Personal Interview (CAPI) method, with trained interviewers reading questions to the participants and entering their responses directly on laptop computers. The participants received 30 yuan ($4.80 USD) for their time and effort. Measures The participants were asked about the acceptability of receiving seven types of services from MMT providers. The seven types of ART-related services, which were predesigned by the research team and a local expert panel based on previous research (Achmad et al., 2009; Berg et al., 2011; Wolfe et al., 2010) and the potential for incorporation into local MMT clinics, included the following: (1) directly observed ART dosage once per day; (2) phone call reminders to take ART medication; (3) text-message reminders to take ART medication; (4) ART-related counseling; (5) ART-related referrals; (6) peer support groups in the MMT clinic to improve ARV treatment adherence; and (7) involvement of family members in support of ART adherence. The responses were either 1 = Yes or 0 = No. A general acceptability score was generated by summing all positive answers (1 = Yes). For each type of service, the reasons for willingness/unwillingness of acceptance were probed using open-ended questions, and the responses were recorded verbatim. The background characteristics collected in this study included age, gender, marital status, years of education, personal income in the previous 30 days, local residence, date of MMT admission, years since HIV diagnosis, current ART status, HIV-serostatus disclosure, and perception of health status. The duration of MMT use (in years) was computed by subtracting the reported date of admission to the MMT clinic from the date of assessment. The HIV-serostatus disclosure was a single composite measure (Lee et al., 2010) based on the extent to which the participant has disclosed his/her serostatus to various groups of people, such as sexual partners, children, other family members/relatives, friends, neighbors, health care workers, community/village leaders, people in the community/village, coworkers, and others. The response categories included 0 = none of them, 1 = some of them, and 2 = all of them. Based on the 10 items, a summative composite scale was developed, with a range of 0 to 20 (Cronbach’s alpha = 0.84). The participants’ perception of their health status was measured using the Medical Outcomes Study-HIV Health Survey (MOS-HIV) general health perception subscale (Ichikawa & Natpratan, 2004). The participants were asked to evaluate their health in general (from 1 = very good to 5 = very poor), and evaluate the following: (1) whether they were somewhat ill, (2) whether they were as healthy as anybody they know, (3) whether their health was excellent, and (4) whether they had been feeling bad lately. Each of the above four items ranged from 1 = definitely true to 5 = definitely false. After reversing some of the items, all responses were summed to generate an overall perceived health score, with higher scores indicating a better perceived health status (Cronbach’s alpha = 0.86). Data analysis The statistical analyses of the quantitative data were performed using the SAS 9.4 (SAS Institute Inc., Cary, NC, USA) statistical software package. We first descriptively analyzed the frequency distribution of the demographics, and drug use-related factors were summarized. Second, the number and percentage of participants who indicated a willingness to accept each type of service were calculated to determine the acceptability of the services. A grounded theory approach was used to examine patterns and emergent themes across the responses to the open-ended questions regarding the reasons for acceptance/rejection of services (Corbin & Strauss, 1990). Third, we conducted an exploratory factor analysis to examine the structure of the seven types of services and categorized the different types of services into three dimensions (Suhr, 2005). Rotated factor loading of 0.40 was used as a cut-off point for the inclusion of a service in a certain dimension. Pearson’s correlations (r) were calculated to investigate the relationships between background characteristics and the acceptability of each dimension of service. A multiple linear regression analysis was performed to examine the correlation between the general acceptability score and the background characteristics. Results Study sample A total of 90 participants were recruited; four did not answer the acceptability-related questions and were excluded from the analysis. Of the 86 participants who completed the questionnaire, 72.1% were male. The mean age was 41.4 years, and approximately half (N = 46, 53.5%) of the participants were between 40 and 49 years of age. The average income in the past 30 days was 585.1 yuan (approximately $94 USD), and approximately one-third (N = 28, 33.7%) of the participants did not have any income in the last 30 days. The participants had used MMT for an average of 4.3 years, and they had been diagnosed as HIV-positive for an average of 6.3 years. Less than one-third (29.1%) of the participants were currently receiving ART (Table 1). Acceptability of MMT-based ART services Figure 1 shows the percentage of acceptance of each type of service. The most accepted integrated services were ART-related counseling services (75.6%), followed by referral services (73.3%). The acceptability of the other five types of services ranged from 53% to 60%. The clients who were currently receiving ART were more likely to accept directly observed ART, counseling, referral, and support groups and less likely to accept reminder services. However, the differences did not reach statistical significance. The reasons for willingness/unwillingness to accept each type of service are summarized below. Directly observed ART dosage The acceptability of directly observed ART varied. Some participants were open to keeping their ART medication in the MMT clinic and having MMT providers supervise them taking their medication, although other participants did not accept this option because their MMT visit schedule would conflict with their ART dosing habits. I am okay with it. I have to come here for methadone anyway. It is convenient to take a dose of ART drugs at the same time (male, 41 years, not on ART). It depends on if other people are doing the same (female, 40 years, not on ART). I do not want to leave my medicine in the MMT clinic. If someday I would not be able to come to the clinic, I would miss a dose. It is convenient for me to keep the medicine with me so that I can take a dose before I go to bed (male, 46 years, on ART). I prefer to just take the medicine at home. I usually just keep my medicine in a very visible place at home so that I will not forget. I am concerned that if I took the medicine here, other people would see it and find out my HIV status (female, 45 years, not on ART). Phone call reminders for ART dosing A phone call reminder would be unfeasible for those who do not personally own a phone. Concerns regarding confidentiality issues when delivering phone call reminders were common. The participants suggested that the providers should be meticulous when phrasing the reminder messages, such that no terms that would potentially disclose their HIV serostatus were mentioned. Please be careful when making those phone calls, do not mention HIV or AIDS or that type of thing because I am afraid people around me would hear it (female, 54 years, not on ART). I do not think I will forget to take medicine. A reminder would be helpful, but the problem is that I do not personally own a cellphone. I use my parents’ landline, so it is not really convenient when they are around (female, 27 years, on ART). Text message reminders for ART dosing Although slightly more participants indicated willingness to accept a phone call reminder than a text message reminder (55.8% vs. 53.5%), several participants expressed their preference for a text message reminder over a phone call reminder because a text message reminder was perceived as more private and more convenient. The participants also suggested frequencies for delivering the reminders. I am more comfortable receiving text messages because I feel like text message reminders are more private than phone calls (male, 40 years, not on ART). It is not necessary to send a reminder every day. Once every few days would be good enough (male, 50 years, not on ART). I prefer text messages over phone calls. Sometimes it is inconvenient for me to pick up the phone (male, 47 years, not on ART). ART-related counseling Counseling services were generally well-accepted, mainly due to the established trust relationship between the clients and the MMT providers. However, a number of participants noted that the MMT providers had inadequate knowledge regarding AIDS treatment, which they learned from their previous encounters. This belief has, to a large extent, reduced their willingness to seek ART-related counseling from MMT providers. It would be great if counseling services were provided here because I have a very close relationship with the doctors here (male, 33 years, on ART). Although the doctors here are nice to us and really care about us, I do not really ask them ART-related questions. I think it is not their specialty. I prefer to ask the CDC doctor; they know more and explain things better (male, 46 years, on ART). I have consulted them before, but they did not really know the answer. I once asked them if I could take the medication at a different time, if I need to adjust the dosage, and how to manage my severe side-effects, and they had no idea. I think it is a waste of time. I should just consult an AIDS specialist (male 46 years, on ART). I do not feel comfortable consulting the doctor here because other clients would find out. If there were a private room I would be willing to (undergo consultation). After all, I would like to know more about AIDS treatment, and it would be more convenient for us to consult the doctors here (female, 27 years, on ART). ART-related referral services The participants indicated that referral services would be helpful for them because the MMT providers were perceived as resourceful. However, the additional cost of other healthcare services was cited as a cause of concern for some of the participants. Referral services would be good. Presumably, they know which hospital provides better treatment (male, 46 years, on ART). I do not need referral services because I do not have money to receive other hospital services; they are always expensive. The last time I went to the infectious disease hospital, the pre-ART physical checkup cost me a fortune. I spent so much money and still did not get the treatment because my liver function was not good enough. They suggested that my liver should be treated first, but I do not have the money to do so. Therefore, I decided not to get ART because my disease is not curable anyway; you can only control it. In addition, I know ART has severe side effects (male, 47 years, not on ART). Peer support group for ART adherence Although some participants had positive experiences attending peer support groups in the past, others perceived such services as unnecessary. The disclosure of their HIV-positive serostatus was reported to be the major concern that reduced the participants’ willingness to participate in peer group activities. I am willing to participate in any type of group activities. I did so before (male, 42, on ART). I do not want to because people gossip. I do not want other people to know (female, 54 years, not on ART). I am pretty compliant with the doctor’s instructions already. I do not think I need help from other people (male, 41 years, on ART). Involvement of family members for ART adherence The trust relationship with MMT providers augmented the participants’ willingness to invite their family members to events organized by the MMT clinic. Nonetheless, several participants mentioned autonomy and a reluctance to add a greater psychological burden to their family members. At the same time, the involvement of family members was not a practical option for those who have not disclosed their serostatus. I would be willing to invite my families to the clinic. They trust the doctors here very much. They really believe in whatever they say (male, 41 years, not on ART). I do not want to discuss this (HIV) with my families. My father is old. I do not want him to worry because it is not good for his own health. Additionally, he would not be able to make decisions for me (male, 47 years, not on ART). Some of my family members, for example my in-laws, do not know my status yet. This is personal and nothing to be proud of (male, 30 years, not on ART). Factors associated with service acceptability Factor analysis of the seven types of services indicated a three-factor structure. Examination of factor loading and content suggested that the seven types of services represented three construct domains. Domain 1 consisted of directly observed ART, peer support groups, and the involvement of family members, and this domain was referred to as social support-related services. Domain 2, categorized as reminder services, included phone call reminders and text message reminders. Domain 3 included counseling and referral services and was called referred to as services. The participants who were currently receiving ART, those who have been diagnosed with HIV infection for a longer period of time, and those who perceived themselves as having poorer health conditions were more likely to accept social support-related services. No factors were found to be significantly correlated with reminder services. Self-perceived health status was negatively associated with the acceptability of clinical services (Table 2). The general acceptability score was significantly correlated only with health perception (β = −0.16, p = .0216) after controlling for other covariates. Discussion This study demonstrated a median to high level of acceptability of MMT-based ART services among HIV-positive MMT clients, which indicates that an MMT clinic-based ART program is a feasible strategy for promoting ART initiation and supporting ART adherence in China. The acceptability is credited to the trust relationship with MMT providers established through daily contact (Li, Wu, Cao, & Zhang, 2012). However, the integration of ART into MMT programs warrants strategic planning. The level of acceptability varied across the different types of services. It is critical to build a patient-centered model of care, such that the services provided are tailored to the specific needs of different groups of clients (Duncombe et al., 2015). The clients with poor health conditions should be the first to receive the integrated services because they exhibit a higher level of service acceptability than those who are relatively healthy. When planning the integrated services, MMT stakeholders should consider adopting the most accepted service components as the first step. Clinical services, including ART-related counseling and referral, demonstrated the highest acceptability across all service types in the study. However, the MMT service provider’s lack of expertise in this area was a prominent concern. This was also echoed in a focus group with MMT service providers (Lin, Cao, & Li, 2014). Providing integrated MMT and ART care is challenging because service providers usually do not have expertise in both addiction medicine and HIV care, which are two separate areas. The development of an MMT-ART provider network with formal governance and management is a more viable strategy to provide comprehensive and seamless care (Haire et al., 2012). Concerns regarding confidentiality should be addressed when planning the implementation of integrated services. In particular, for the delivery of phone calls and text message reminder services, the provider should be cautious when choosing the appropriate wording and avoid using phrases that would disclose the HIV serostatus of the client. Other types of services, particularly those involving friends or family members, may only be feasible for those who have fully disclosed their HIV-serostatus. Directly observed dosage has been proven to improve treatment adherence and decrease HIV viral replication (Pearson et al., 2006). Nonetheless, the lack of confidentiality and conflict with self-dosing habits poses a major concern, which reduces the feasibility of delivering this service to MMT clinics. Similar findings have also been reported in other studies (Saberi, Caswell, Jamison, Estes, & Tulsky, 2012; Wohl et al., 2003). Instead of imposing this type of service on the client, the provider should change tactics by informing the clients of the importance of treatment adherence, providing the clients with specific medication instructions, introducing mobile applications to support adherence, and guiding the client to establish a medication-taking routine. This study was limited by a high refusal rate and the corresponding self-selection bias. Furthermore, we could not predict the actual use if the ART-related services were provided. Due to the small sample size, this study could not identify the differences that may exist across the 12 separate MMT clinics. However, this study provided insights into the strategic planning of MMT-ART service integration. Stakeholders need to consider the provision of multiple strategies instead of following a single algorithm, such that the HIV-positive clients can choose the most suitable combination of services based on personal and relationship-based preferences. Enhanced in-service training for MMT providers and confidentiality protection are necessary to heighten the acceptability and optimize the integration of service delivery. We would like to thank the project team members in the Sichuan Provincial Center for Disease Control and Prevention for their contributions to this study. Funding This study was supported by the National Institute on Drug Abuse (NIDA) Grant R01DA033130 and National Institute of Mental Health (NIMH) Grant K01MH102147. Figure 1 Percentage of acceptance of each type of service by ART status. Table 1 Sample description (N = 86). N % Female 24 27.9  Age (mean ± SD) 41.4 5.7  Younger than 40 years 33 38.4  40–49 years 46 53.5  More than 50 years 7 8.1 Married 36 41.9 Years of education (mean ± SD) 8.9 2.5  Equal to or less than 6 years 21 24.4  7–9 years 40 46.5  More than 10 years 25 29.1 Income in the past 30 days (mean ± SD) 585.1 1182.6  Zero 28 33.7  Less than 500 yuan1 33 39.8  More than 500 yuan1 22 26.5 Years in MMT (mean ± SD) 4.3 2.8  Less than 3 years 31 36.1  3–6 years 27 31.4  More than 6 years 28 32.6 Years since HIV diagnosis (mean ± SD) 6.3 3.7  Less than 5 years 29 34.1  5–10 years 38 44.7  More than 10 years 18 21.2 Currently on ART 25 29.1 Disclosure2 (mean ± SD) 7.8 4.7 Health perception3 (mean ± SD) 12.6 4.6 1 1 yuan = 0.16 USD = 0.12 EUR (as 2013) 2 Disclosure: the extent to which the participant has disclosed his/her HIV serostatus to various groups of people (sexual partners, children, other family members/relatives, friends, neighbors, health care workers, community/village leaders, people in the community/village, coworkers, others) (Lee et al., 2010). 3 Health perception: perception of health status, measured using Medical Outcomes Study-HIV Health Survey (MOS-HIV) general health perception subscale (Ichikawa & Natpratan, 2004). Table 2 Pearson’s correlation coeffcient and corresponding p-values with service acceptabilitya. Social support Reminder services Clinical services Female 0.13 0.12 0.17 0.2427 0.2641 0.1120 Age −0.09 −0.16 −0.08 0.3935 0.1518 0.4899 Married −0.06 0.17 0.04 0.6003 0.1265 0.6917 Years of education 0.16 0.13 0.18 0.1338 0.2286 0.0976 Income in the past 30 days (yuan) 0.14 0.20 0.17 0.1849 0.0635 0.1232 Years in MMT 0.17 0.08 0.06 0.1130 0.4610 0.5587 Years since HIV diagnosis 0.22 0.02 0.00 0.0398* 0.8255 0.9775 Currently on ART 0.25 −0.12 0.13 0.0182* 0.2872 0.2498 Disclosure 0.17 0.08 0.08 0.1088 0.4404 0.4803 Health perception −0.27 −0.15 −0.26 0.0110* 0.1777 0.0143* * p < .05 a Service acceptability: the acceptability to receive ART-related services provided by MMT providers (the participants could answer 1 = yes or 0 = no to each of the seven different types of services). The seven types of services were categorized into three domains: social support, reminder services, and clinical services. Social support-related services consisted of (1) directly observed ART, (2) peer support groups, and (3) involvement of family members. Reminder services included (1) phone call reminders and (2) text message reminders. Clinical services included (1) counseling and (2) referral services. Declaration of interest The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the article. Achmad YM Istiqomah AN Iskandar S Wisaksana R van Crevel R Hidayat T 2009 Integration of methadone maintenance treatment and HIV care forinjecting drug users: A cohort study in Bandung, Indonesia Indonesian Society of Internal Medicine 41 Suppl 1 23 27 Azar P Wood E Nguyen P Luma M Montaner J Kerr T Milloy MJ 2015 Drug use patterns associated with risk of non-adherence to antiretroviral therapy among HIV-positive illicit drug users in a Canadian setting: A longitudinal analysis BMC Infectious Diseases 15 193 10.1186/s12879-015-0913-0 25927573 Bachireddy C Soule MC Izenberg JM Dvoryak S Dumchev K Altice FL 2014 Integration of health services improves multiple healthcare outcomes among HIV-infected people who inject drugs in Ukraine Drug and Alcohol Dependence 134 106 114 10.1016/j.drugalcdep.2013.09.020 24128379 Berg KM Litwin A Li X Heo M Arnsten JH 2011 Directly observed antiretroviral therapy improves adherence and viral load in drug users attending methadone maintenance clinics: A randomized controlled trial Drug and Alcohol Dependence 113 192 199 10.1016/j.drugalcdep.2010.07.025 20832196 Corbin J Strauss A 1990 Grounded theory method: Procedures, canons, and evaluative criteria Qualitative Sociology 13 3 21 Duncombe C Rosenblum S Hellmann N Holmes C Wilkinson L Biot M Garnett G 2015 Reframing HIV care: Putting people at the centre of antiretroviral delivery Tropical Medicine and International Health 20 4 430 447 10.1111/tmi.12460 25583302 Gruskin S Ferguson L Alfven T Rugg D Peersman G 2013 Identifying structural barriers to an effective HIV response: Using the national composite policy index data to evaluate the human rights, legal and policy environment Journal of the International AIDS Society 6 1 18000 10.7448/IAS.16.1.18000 Haire K Burton C Park R Reynolds J Stewart D Purushotham AD 2012 Integrated cancer system: A perspective on developing an integrated system for cancer services in London London Journal of Primary Care 5 1 29 34 Ichikawa M Natpratan C 2004 Quality of life among people living with HIV/AIDS in northern Thailand: MOS-HIV health survey Quality of Life Research 13 3 601 610 15130024 Jiamsakul A Kumarasamy N Ditangco R Li PC Phanuphak P Sirisanthana T TREAT Asia Studies to Evaluate Resistance Monitoring Study (TASER-M) 2014 Factors associated with suboptimal adherence to antiretroviral therapy in Asia Journal of the International AIDS Society 17 18911 10.7448/IAS.17.1.18911 24836775 Lambdin BH Mbwambo JK Josiah RM Bruce RD 2015 Service integration: opportunities to expand access to antiretroviral therapy for people who inject drugs in Tanzania Journal of the International AIDS Society 18 1 19936 10.7448/IAS.18.1.19936 26202669 Lee SJ Li L Jiraphongsa C Iamsirithaworn S Khumtong S Rotheram-Borus MJ 2010 Regional variations in HIV disclosure in Thailand: Implications for future interventions International Journal of STD & AIDS 21 3 161 165 10.1258/ijsa.2009.009008 20071440 Li L Wu Z Cao X Zhang L 2012 Provider-client interaction in methadone treatment clinics in China Journal of Drug Issues 42 2 147 155 10.1177/0022042612446593 Lin C Cao X Li L 2014 Integrating antiretroviral therapy in methadone maintenance therapy clinics: service provider perceptions The International Journal on Drug Policy 25 6 1066 1070 10.1016/j.drugpo.2014.04.021 24939555 Pearson CR Micek M Simoni JM Matediana E Martin DP Gloyd S 2006 Modified directly observed therapy to facilitate highly active antiretroviral therapy adherence in Beira, Mozambique. Development and implementation Journal of Acquired Immune Deficiency Syndromes 43 Suppl 1 S134 S141 10.1097/01.qai.0000248339.82567.17 17133197 Reddon H Milloy MJ Simo A Montaner J Wood E Kerr T 2014 Methadone maintenance therapy decreases the rate of antiretroviral therapy discontinuation among HIV-positive illicit drug users AIDS and Behavior 18 4 740 746 10.1007/s10461-013-0584-z 23918244 Saberi P Caswell NH Jamison R Estes M Tulsky JP 2012 Directly observed versus self-administered antiretroviral therapies: Preference of HIV-positive jailed inmates in San Francisco Journal of Urban Health 89 5 794 801 10.1007/s11524-012-9682-x 22547327 Sarang A Rhodes T Sheon N 2013 Systemic barriers accessing HIV treatment among people who inject drugs in Russia: A qualitative study Health Policy and Planning 28 7 681 691 10.1093/heapol/czs107 23197431 Suhr DD 2005 Principal component analysis vs. exploratory factor analysis Retrieved from http://www2.sas.com/proceedings/sugi30/203-30.pdf Tran BX Ohinmaa A Duong AT Nguyen LT Vu PX Mills S Jacobs P 2012 Cost-effectiveness of integrating methadone maintenance and antiretroviral treatment for HIV-positive drug users in Vietnam’s injection-driven HIV epidemics Drug and Alcohol Dependence 125 3 260 266 10.1016/j.drugalcdep.2012.02.021 22436971 Uhlmann S Milloy MJ Kerr T Zhang R Guillemi S Marsh D Wood E 2010 Methadone maintenance therapy promotes initiation of antiretroviral therapy among injection drug users Addiction 105 5 907 913 10.1111/j.1360-0443.2010.02905.x 20331553 Vlahov D Celentano DD 2006 Access to highly active antiretroviral therapy for injection drug users: Adherence, resistance, and death Cadernos de Saúdepública 22 705 718 Wohl DA Stephenson BL Golin CE Kiziah CN Rosen D Ngo B Kaplan AH 2003 Adherence to directly observed antiretroviral therapy among human immunodeficiency virus-infected prison inmates Clinical Infectious Diseases 36 12 1572 1576 10.1086/375076 12802758 Wolfe D Carrieri MP Shepard D 2010 Treatment and care for injecting drug users with HIV infection: A review of barriers and ways forward Lancet 376 9738 355 366 10.1016/S0140-6736(10)60832-X 20650513 Wu Z Sullivan SG Wang Y Rotheram-Borus MJ Detels R 2007 Evolution of China’s response to HIV/AIDS Lancet 369 9562 679 690 10.1016/S0140-6736(07)60315-8 17321313 Yin W Pang L Cao X McGoogan JM Liu M Zhang C Rou K 2015 Factors associated with depression and anxiety among patients attending community-based methadone maintenance treatment in China Addiction 110 Suppl 1 51 60 10.1111/add.12780 Zhang F Dou Z Ma Y Zhang Y Zhao Y Zhao D Chen RY 2011 Effect of earlier initiation of antiretroviral treatment and increased treatment coverage on HIV-related mortality in China: a national observational cohort study Lancet Infectious Diseases 11 7 516 524 10.1016/S1473-3099(11)70097-4 21600849 Zhao Y Shi CX McGoogan JM Rou K Zhang F Wu Z 2015 Predictors of accessing antiretroviral therapy among HIV-positive drug users in China’s national methadone maintenance treatment programme Addiction 110 Suppl 1 40 50 10.1111/add.12782

PMC005xxxxxx/PMC5127772.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 9216904 2419 Nat Genet Nat. Genet. Nature genetics 1061-4036 1546-1718 27749841 5127772 10.1038/ng.3683 NIHMS814330 Article Spatial intratumor heterogeneity of genetic, epigenetic alterations and temporal clonal evolution in esophageal squamous cell carcinoma Hao Jia-Jie 110 Lin De-Chen 231011 Dinh Huy Q. 410 Mayakonda Anand 510 Jiang Yan-Yi 510 Chang Chen 1 Jiang Ye 1 Lu Chen-Chen 1 Shi Zhi-Zhou 6 Xu Xin 1 Zhang Yu 1 Cai Yan 1 Wang Jin-Wu 7 Zhan Qi-Min 1 Wei Wen-Qiang 811 Berman Benjamin P. 411 Wang Ming-Rong 111 Koeffler H. Phillip 259 1 State Key Laboratory of Molecular Oncology, National Cancer Center/Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, 100021, China 2 Cedars-Sinai Medical Center, Division of Hematology/Oncology, UCLA School of Medicine, Los Angeles, USA 3 Guangdong Provincial Key Laboratory of Malignant Tumor Epigenetics and Gene Regulation, Medical Research Center, Sun Yat-Sen Memorial Hospital, Sun Yat-Sen University, Guangzhou, China 4 Center for Bioinformatics and Functional Genomics, Biomedical Sciences, Cedars-Sinai Medical Center, UCLA School of Medicine, Los Angeles, USA 5 Cancer Science Institute of Singapore, National University of Singapore, Singapore 6 Faculty of Medicine, Kunming University of Science and Technology, Kunming, Yunnan, China 7 Department of Pathology, Linzhou Cancer Hospital, Henan, China 8 Department of Cancer Epidemiology, National Cancer Center/Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, 100021, China 9 National University Cancer Institute, National University Hospital Singapore, Singapore 10 These authors contributed equally to this work. 11 Correspondence should be addressed to M-R.W. (wangmr2015@126.com), B.P.B (Benjamin.Berman@csmc.edu), D-C.L. (dchlin11@gmail.com) or W-Q W. (weiwq2006@126.com) 8 9 2016 17 10 2016 12 2016 17 4 2017 48 12 15001507 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. Esophageal squamous cell carcinoma (ESCC) is among the most common malignancies, but little is known about its spatial intratumor heterogeneity (ITH) and temporal clonal evolutionary processes. To address this, we performed multiregion whole-exome sequencing on 51 tumor regions from 13 ESCCs, and multiregion global methylation profiling on three of these 13 cases. We found an average of 35.8% heterogeneous somatic mutations with strong evidence of ITH. Half of driver mutations located on the branches targeted oncogenes, including PIK3CA, NFE2L2, MTOR, etc. By contrast, the majority of truncal and clonal driver mutations occurred in tumor suppressor genes, including TP53, KMT2D, ZNF750, etc. Interestingly, the phyloepigenetic trees robustly recapitulated the topologic structures of the phylogenetic ones, indicating the possible relationship between genetic and epigenetic alterations. Our integrated investigations of the spatial ITH and clonal evolution provide an important molecular foundation for enhanced understanding of the tumorigenesis and progression of ESCC. Esophageal carcinoma is among the most common human cancers, causing over 400,000 deaths worldwide annually1,2. The highest-risk areas are located in Eastern Asia, as well as Eastern and Southern Africa; and the most prevalent type is esophageal squamous cell carcinoma (ESCC)1,2. The five-year survival rates of ESCC patients undergoing surgery is below 30%, because a large proportion of the tumors are unresectable or have already metastasized before diagnosis3. Recently, several large-scale genomic studies including ours have characterized ESCC genomes with hundreds of somatic mutations and copy number alterations, and have identified significantly mutated genes, including TP53, PIK3CA, ZNF750, etc.4–9. The APOBEC signature is a predominant mutational spectrum, and contributes to the mutagenic processes of ESCCs6,8. However, the genomic alterations identified in all of these studies were obtained using only single samples representing individual cases, and little is known about the spatial intratumor heterogeneity (ITH) and the temporal clonal evolutionary processes of mutational spectrum in ESCC. Moreover, although alterations in DNA methylation have been observed in ESCC, the ITH of these epigenetic changes is still unknown, and whether it correlates with the genetic architecture remains unexplored. Precise understanding of both the genomic and epigenomic architecture of primary ESCC tumors is crucial for developing personalized patient treatment and molecular-based biomarkers10. Furthermore, an integrated investigation of the genomic and epigenomic evolutionary trajectory of ESCC may also reveal new insights into the relationship between the genome and epigenome. In the present study, we address these critical issues through integrative molecular approaches, including multiregional whole-exome sequencing (M-WES), global methylation profiling, as well as phylogenetic and phyloepigenetic tree construction. RESULTS Spatial ITH of ESCC M-WES was performed on the genomic DNA from 13 primary ESCC patients, and the clinical-pathological parameters of these patients were listed in Supplementary Table 1. In total, 51 tumor regions and 13 matched morphologically normal esophageal tissues (four tumor regions and one matched normal tissue per case, with the exception of ESCC04, which only had three tumor regions) were sequenced, with the mean coverage depth of 150×. A total of 1,610 non-silent somatic mutations (affecting 1,427 genes) and 568 silent mutations were identified, with a validation rate of 90% (Supplementary Tables 2–3). To explore ITH and genomic evolution of ESCC, phylogenetic trees were constructed based on somatic mutations (both silent and non-silent) identified from each tumor region. The trunk, shared branch and private branch of the tree represented mutations in all tumor regions, in some but not all regions, and only in one region, respectively. As shown in Fig. 1a and Supplementary Fig. 1, phylogenetic trees varied extensively among different cases, and all of the 13 ESCCs showed evidence of spatial ITH, with an average of 35.8% (780/2,178; range, 8.0%–60.9%) of somatic variants having spatial heterogeneity. Characterization of the relative timing of mutations affecting driver genes with possible biological relevance is essential for revealing the evolutionary processes of the cancer genome, as well as further improving precision medicine strategies. To address this, we identified potential driver mutations according to recent large-scale ESCC sequencing data4–8, COSMIC gene census11 and Pan-cancer analysis12; this was followed by tracing them within the phylogenetic trees (See Methods). Overall, driver mutations were significantly more enriched in the trunks than were passenger mutations (77.8% vs. 63.8%; P = 0.023; Fig. 1b). This indicates that drivers are mutated as relatively early events during the evolutionary process of the tumors, which is in accordance with previous findings in other tumor types13. We next separated putative driver mutations into those occurring either in oncogenes or tumor suppressor genes (TSGs). Importantly, half of the driver mutations (50.0%) that mapped to the branches were within oncogenes, including PIK3CA, KIT, NFE2L2, MTOR and FAM135B. In comparison, only 22.4% of driver mutations located on the trunks affected oncogenes, and the rest were in TSGs. For example, TP53 mutations were present in twelve of the thirteen cases, and were truncal in all of the mutated cases, in agreement with recent reports14,15. It is worthwhile to note that potentially actionable mutations such as those targeting PIK3CA and MTOR tended to be oncogenic branch events. These findings highlight the extra caution needed when considering inhibiting these mutants in ESCC, given previous studies showing that suppressing subclonal drivers led to growth acceleration of non-mutated subpopulations16. Clonal status of putative driver mutations We next investigated the clonal status of somatic mutations within individual regions. Cancer cell fraction (CCF) in each tumor region was calculated as described previously through integrative analysis of local copy number, variant allele frequency (VAF) and tumor cell purity16,17. Several driver mutations were subclonal and possibly occurred as late events in ESCC, including MTOR, KEAP1, PTPRB and FAM135B. In contrast, cancer genes on the trunks, such as TP53, NOTCH1, CREBBP, KMT2D and ZNF750, were predominantly mutated in a fully clonal manner (Fig. 2), further verifying our earlier phylogenetic tree analysis showing that these mutations were possibly early lesions during ESCC development. Of particularly noteworthy distinction, a number of driver variants detected as clonal within some individual tumor regions, were absent in others from the same individual, producing an “illusion” of clonal dominance. For example, a PIK3CA hotspot mutation (M1043I) was undetectable in tumor region T2 and T3 in case ESCC13 but was clonally dominant in the other two regions. Likewise, a hotspot mutation in KIT gene (E601K) was present in 100% tumor cells in regions T1 and T3 in case ESCC08, yet was absent in the rest of the tumor regions. Such clonal dominance was also observed in NFE2L2 in case ESCC12. Our results suggest that driver mutations can have mixed and complex intratumoral clonal status in ESCC, and that current single-sampling approach may misinterpret these critical genomic lesions because of the “illusion” of clonal dominance. We further investigated all the non-silent variants within genes and related pathways that have potential targeting approaches. As shown in Supplementary Fig. 2, mutations affecting members in PI3K/MTOR pathway, KIT, AURKA and CCND2 were always late events (branched/subclonal). By contrast, variants in ERBB4, FGFR2, BRCA2, ATM and TP53, were mutated as early events (truncal/clonal), suggesting their potentials as candidate actionable targets for ESCC. ITH of copy number alterations (CNA) We next analyzed the ITH at the copy number level (Supplementary Table 4). First, recurrent copy number alterations which involve important cancer genes in ESCC were identified based on our previous results6, and we confirmed that the present cohort harbored these recurrent CNAs with similar frequencies (Supplementary Fig. 3). Although CNAs were generally more similar within cases than between different cases, we found extensive CNA ITH, with 90% (9/10) of all recurrent CNAs being spatially heterogeneous. For example in ESCC08, chr7p11.2 amplification (encompassing EGFR) was observed in regions T1 and T4, but not in regions T2 and T3. Similarly, deletions of chr9p21.3 (harboring CDKN2A/B) were ubiquitous in some cases but also occurred as heterogeneous aberrations in other samples. The only driver CNA found as consistently ubiquitous was the copy number gain of 11q13, which encompassed a number of oncogenes including CCND1, ANO118–20 and CTTN21,22, highlighting the importance of this aberration as a founder genomic lesion in the development of ESCC. These results suggest that similar as somatic mutations, CNAs also show significant spatial ITH, concordant with the observations in several other types of cancers23–25. The within-patient mutational rate (mean = 168) was higher than the within-region rate (mean = 139, Supplementary Table 5), highlighting the improved resolution of our multi-biopsy approach for genomic interrogation. Particularly, in the case of branched cancer genes, current M-WES approach markedly increased the sensitivity of the detection rate (Table 1). For example, ATR and TSC1 mutations, which were detected in only 2% of tumor regions (in agreement with previous results), occurred in 7.7% of cases. In addition, the proportion of subclonal mutations detected in each tumor region was much lower than that in each case (Table 2 and Supplementary Fig. 4). These results again signify that analyzing sequencing data obtained from a single biopsy will likely underestimate the prevalence of the mutations, especially for those acquired late in the mutational process24. Temporal dissecting of mutational spectra and signatures To determine the temporal dynamics of the mutagenic processes in ESCC, the mutational spectrum of both the trunks and branches was analyzed using deconstructSigs26, which identifies the linear combination of pre-defined signatures that most accurately reconstructs the mutational profile of a single tumor sample. As shown in Fig. 3a, the overall mutational spectra were similar between trunk and branch mutations, with very strong enrichment of Signature 1 substitutions (associated with age), and more subtle but enriched representation of APOBEC-associated Signature 2 and Signature 13 substitutions (C>G and C>T in the TpCpW context). We next calculated the contributions of individual mutational signatures to each tumor (Fig. 3b), and identified several signatures within the tumors tested, including Signature 1 (Age), Signatures 2 and 13 (APOBEC), and Signatures 6 and 15 (DNA mismatch repair), in agreement with previous results in esophageal squamous and other squamous-type cancers6,8,27. Interestingly, we noticed that a number of tumors displayed a prominent decrease of the relative contribution of Signature 1 in the branch compared with trunk mutations, albeit without obtaining statistical significance due to the relatively small number of tumors analyzed. In some of these cases, we also observed an increase of signatures associated with DNA damage (including Signatures 3 and 15) in the branch mutations (such as ESCC10 and ESCC12, shown in Figs. 3c–d and Supplementary Fig. 5). To interpret these temporal differences of mutational spectra within the same tumor will require further investigations, but the data indicate that various mutational processes might play important roles in subclonal diversification during the progression of ESCC. ITH of DNA methylation in ESCC As with other cancers, epigenetic abnormalities have been associated with the development and pathogenesis of ESCC28–30. To decipher ESCC ITH at the epigenetic level and its potential relationship with subclonal gene mutations, the genome-wide methylation levels of fourteen M-WES-profiled tumor and normal tissues from three ESCC cases (ESCC01, ESCC03 and ESCC05) were profiled using the Illumina HumanMethylation450 (HM450) Bead array. We first identified CpG probes that showed significant differences between the tumor regions and normal tissues from the same patient (except for ESCC01, which did not have a matched normal tissue), then divided these differentially methylated probes into those with “shared” changes (i.e. consistent within all tumor regions from the same case), and “private” changes (those present in one or more of the regions, but not all). We used the probes with private changes to infer tumor evolution and constructed phyloepigenetic trees for each case based on the Euclidean pairwise distances of methylation profiles31,32 (See Methods). Topological similarities were tested between phyloepigenetic and phylogenetic trees in all three cases based on Robinson-Foulds (RF) distance for unrooted trees33 (Fig. 4a). Notably, in accordance with a recent report on glioma32, the RF distances (zeros for all 3 cases) suggested high concordance between genetic and epigenetic tree topologies in all three cases (see Methods). Since the distinction between the private and shared methylation changes was cutoff dependent, we further tested four different probe-selection cutoffs, and noted that the phyloepigenetic trees were robust to the cutoff and showed highly similar topological structures (Supplementary Fig. 6). Moreover, to alleviate the confounding effects from non-tumor DNA contamination, two different methods were performed to account for and mitigate the potential influence from immune cells (the major source of non-cancer cells in these samples, Supplementary Fig. 7), and again, similar results were observed between the trees using uncorrected methylation values, and trees using either correction method (Supplementary Fig. 8, see Methods). These findings suggest the possible relationship between genomic and epigenomic alterations during the clonal evolution of ESCC cells, and are indicative of the presence of multiple epigenetically distinct, subclonal cell populations, as recently observed in prostate cancer31, glioma32 and hepatocellular carcinoma [unpublished data, D-C. L., A. M., H. Q. D, Pinbo Huang, Lehang Lin, et al.]. We observed that a number of TSGs, including EPHA734,35, PCDH1036,37, DOK138,39, etc, were hypermethylated at their promoters within some but not all regions of the same case, indicating that their expression might be differentially suppressed in different tumor regions. Notably, some TSGs were both mutated and acquired promoter hypermethylation, such as ASXL1 and EPHA7. Interestingly, ASXL1 was subject to both truncal/clonal mutation and shared hypermethylation at its promoter, suggesting that ASXL1 was disrupted early during both the genomic and epigenomic evolutionary processes. To explore the potential biological significance of DNA methylation ITH in ESCC, we next sought to determine whether the differentially methylated DNA CpG loci in each case were enriched at particular functional genomic categories. We first divided CpG probes into those where tumor methylation was higher than the adjacent normal tissues (hyper-methylated) or lower (hypo-methylated). Shared probes were selected for their relatively consistent changes in different tumor regions (Supplementary Fig. 9), while the rest (private probes) exhibited prominent differences between the tumor regions (Fig. 4b) and reflected the extensive ITH seen in the phyloepigenetic trees. We next compared shared vs. private probes by assigning them to various relevant functional genomic categories including CpG Islands (CGIs), CGI Shores, promoters and enhancers, etc., and compared them to the background frequencies of these categories based on all probes on the array (Fig. 4c). As expected, shared CpG sites showed several methylation patterns commonly seen across cancer types40,41, including hypermethylated probes strongly enriched within CGI promoter regions, and depleted in both long-range partially methylated domains (PMDs) and enhancer regions (after removing CGIs). Shared hypomethylated probes showed an inverse distribution, i.e., markedly depleted in CGI promoters while enriched in PMDs as well as enhancer regions (Fig. 4c). Strikingly, private CpG sites for the most part resembled the distribution patterns of their shared counterparts (Fig. 4c). In light of the known contribution of tumor-specific methylation to cancer biology42,43, our results suggest that intratumoral methylation heterogeneity might play a role in the subclonal diversification of ESCC tumors. In support of this, GO analysis of the genes with privately-hypermethylated promoters showed that they were significantly enriched in cancer-related processes, including cell proliferation, differentiation, migration, adhesion and transcriptional regulation (Fig. 4d). In addition, we noticed that privately-hypermethylated probes were even more enriched in CGI Shores than shared-hypermethylated ones (Fig. 4c). Given the prior observations that i) cancer-specific differentially methylated regions occur more frequently within CGI shores than within CGIs44,45, and ii) CGI Shore methylation correlates with the expression of the associated genes44, our observations further suggest the potential involvement of heterogeneity of DNA methylation in the evolutionary biology of ESCC cells. DISCUSSION ESCC is one of the most common malignancies, with relatively low overall five-year survival rates. The main cause leading to unfavorable prognosis of ESCC patients is the lack of effective therapies. Currently, none of the targeted therapies have been established for clinical management of ESCC46. Hundreds of genomic alterations, including somatic mutations and copy number alterations, have recently been identified in ESCC4–9, but these data have not been translated into clinical applications. In addition, the genomic and epigenomic ITH and clonal evolution of ESCC tumors have not yet been characterized. In light of the evidence that ITH is the major cause of drug resistance and treatment failure47, deciphering the genomic diversity and clonal evolution of ESCC tumors will provide both a theoretical and translational basis for identifying new targets and designing personalized medicine strategies. In the present study, the genomic ITH of 13 ESCC cases, as well as the epigenetic ITH of three of these individuals, were investigated through a variety of molecular approaches, and concordant tumor evolutionary trajectories were found as inferred by both their DNA mutations and methylation. A very recent study of two ESCC cases reported that the ITH rate for somatic mutations was approximate 90%48, whereas the rates in our study were much lower, with an average of 35.8%. The discrepancy may well be due to the differences of sequencing depths between the two studies (50× V.S. 150×). Although the true extent of ITH is difficult to define, high sequencing coverage in our study offers an improved resolution to decipher the spatial heterogeneity and clonal evolution of ESCC. Although phylogeny analysis based on M-WES is not able to resolve completely the true temporal ordering of all the somatic variants, we calculated that an average of 93.5% (range from 87.8% to 97.7%) of somatic mutations were compatible with the present phylogenic trees (Supplementary Fig. 1). For example, in case ESCC13, 282 out of 294 variants (95.9%) were compatible with the evolution model based on the topological structure of the phylogenetic tree; and only 12 mutations, including PIK3CA, were incompatible with the phylogenetic tree (Supplementary Table 6). Therefore, the phylogeny method correctly resolves the temporal order of the vast majority of the somatic mutations. Moreover, the evolutionary models inferred from the M-WES-based phylogeny are strongly supported by our DNA-methylation phylogeny in all three cases (Fig. 4a). Hence, this reconstruction of the phylogenetic topologies, from a completely independent epigenomic event, strongly reinforces the validity of these evolutionary models. Resolving the clonal status of driver mutations will help to distinguish early from late events, and targeting clonally dominant driver mutations (early events) conceivably represents an optimized therapeutic strategy10,49. In this study, despite driver mutations having a tendency to be truncal/clonal compared with passenger mutations, approximate 40% of driver mutations were branched or subclonal. This observation suggests that these driver mutations were relatively late events during tumor evolution, and contributed to the emergence of distinct subclonal expansions after the founding clones were established. Notable examples included KIT, and members of the PI3K/MTOR pathway (PIK3CA and MTOR) and NFE2L2 pathway (NFE2L2 and KEAP1). These examples, most of which are oncogenes, were frequently mutated as late events in ESCC. Furthermore, evidence of “parallel evolution” was noted in some cases. For example, ESCC13 contained branch PIK3CA mutations derived in two spatially separated tumor regions, both harboring the M1043I variant, which is a hotspot mutation. Similar parallel evolution was also observed in NFE2L2 mutations in ESCC12. Interestingly, PIK3CA, KIT and NFE2L2 mutations were fully clonal in some tumor regions but were completely absent in others, showing an “illusion” of clonal dominance. In addition, the number of within-patient mutations was higher than the within-region mutations. These results strongly suggest that the prevalence of these driver events, and the rate of sub-clonality overall, are likely underestimated when using a single biopsy to represent an individual patient. Although ESCC DNA methylation alterations have been profiled using single-sampling approaches, their intratumoral diversity and the relationship to genetic lesions remain unknown. In the present study, we found a number of TSGs with private hypermethylation at the promoters, some of which have been associated with either tumorigenesis or progression of other cancer types, such as EPHA734,50, ABCB451, PCDH1052,53 and DOK138,39. This finding suggests that these genes might be differentially inactivated in different tumor regions from same individuals. We revealed profound epigenetic ITH in ESCC through global methylation analysis. Importantly, subclonal evolutions inferred from DNA methylation closely recapitulated phylogenetic trees, indicating the possible relationship between genetic and epigenetic alterations in ESCC. Therefore, integrative analysis of both phylogenetic and phyloepigenetic trees may generate an enhanced understanding of clonal architecture, and reveal the basis for subclonal epigenetic driver events. These features of epigenetic and genetic ITH revealed by our study may have important implications in ESCC biology. ONLINE METHODS Patients and specimens Tissue samples from 13 ESCC patients, including primary esophageal tumors and matched morphologically normal esophageal epithelial margins, were collected in Linzhou Esophagus Cancer Hospital, Henan province, China. All the samples used in this study were residual specimens collected after diagnosis sampling. All the patients received no treatment before surgery, and signed separate informed consent forms for the sampling and molecular analyses. We also considered the clinic-pathological parameters when selecting these ESCC patients, including gender, pathological tumor (pT) stage, regional lymph node metastasis, and tumor differentiation, to avoid bias towards particular pathological characteristics (Supplementary Table 1). Specifically, the male/female ratio in the current cohort was similar to that reported in the latest publication54. The number of patients with relatively early (pT1b/T2) and late (pT3) tumor stage were five and eight, respectively. The status of lymph node metastasis (negative, n = 4; positive, n = 9), as well as tumor differentiation (G1, n = 1; G2, n = 6; G2/3, n = 2; G3, n = 4) were also taken into account. This study has been approved by the Ethics Committee/IRB of Cancer Hospital/Institute, Peking Union Medical College and Chinese Academy of Medical Sciences (Approval No. NCC2013-066). The collection and publication of Chinese human genetic data used in the present study has been approved by the Ministry of Science and Technology. In 12 out of 13 cases, four spatially separated tumor specimens were obtained from each individual, with each section at least 0.5 cm away from the others. In the case of ESCC04, three tumor regions were sampled. We carefully reviewed the hematoxylin and eosin (H&E) slides of each tumor region before subjecting them to WES analysis, in order to make sure that the tumor cell content of the selected regions were comparable and were at least greater than 60% (Representative H&E photos in Supplementary Fig. 10). Multiregional whole-exome sequencing (M-WES) For each individual, genomic DNA (gDNA) of cells from different tumor regions and one matched normal epithelial tissue at the surgical margins were sequenced. Genomic DNA was extracted using Qiagen DNeasy Blood & Tissue Kit according to the manufacturer’s instructions. For cases ESCC01 and ESCC02, whole-exome capture of gDNA was performed by the Beijing Genome Institute (BGI), using the BGI Exome Enrichment Kit, and massively parallel sequencing of captured gDNA was performed and analyzed by BGI using the Complete Genomics platform. For the 11 other cases, the Agilent SureSelect Human All Exon v4 (51 Mb) kit was used for whole-exome capture of gDNA, and the captured DNA was sequenced by BGI using the Illumina HiSeq4000 sequencing platform, with 150 base pair paired-end sequencing. Alignment of sequencing reads and somatic variant detection 150 base pair paired-end fastq files were aligned to the human reference genome (build hg19) using bwa-mem aligner in default mode (URL). Alignments were then filtered for duplicate reads using Samblaster55, and bam files were indel realigned and base quality scores were recalibrated according to GATK best practices56. Somatic variants were detected using VarScan257. Tumor and matched normal pileup files were generated using the samtools “mpileup” command and fed into the VarScan “somatic” command58. Reference genome positions covered at least by 10 reads in normal and 14 reads in tumor samples were considered for variant calling. Variants with VAF less than 0.07 were discarded. Raw somatic variants were filtered using the VarScan “processSomatic” command with arguments–min-tumor-freq 0.07, --max-normal-freq 0.02 and –p-value set to 0.05. These high quality somatic variants were filtered for false positives using fpFilter perl script (URL). These filtered variants were annotated with annovar59 and filtered against dbSNP135 database for commonly occurring Single Nucleotide Polymorphisms (SNPs)59. Disease associated variants annotated in ClinVar database and COSMIC database were retained. Phylogenetic tree construction For mutations that have been detected from at least one tumor region, a method described by Stachler et al.60 was used to increase the sensitivity of detecting these mutations in other tumor regions from the same individual with low VAF. In brief, Bam-Readcount (URL) was used to obtain read counts for unique somatic variants across all tumor regions. A variant was considered as absent if either its VAF was less than 0.02 or the reads were fewer than 3. The VAFs across all tumor regions of each individual were then used to generate a binary table. Phylogenetic trees were constructed based on the binary tables using Discrete Character Parsimony, implemented in PHYLogeny Inference Package (URL), with the matched morphologically normal epithelial margins as outgroup roots. Based on the calculated branch/trunk lengths inferred from mutation counts, the final trees were drawn manually. Cancer cell fraction (CCF) analysis Copy number analysis from WES data was performed using ReCapSeg, which is implemented as part of the Genome Analysis Tool Kit (GATK v4). Briefly, read counts for each of the exome targets were extracted from all samples and were divided by the total number of reads to generate proportional coverage. A panel of normal controls14 was created by using proportional coverage from all of the normal samples. Each of the tumor samples was compared to PoN, followed by tangent normalization. These normalized coverage profiles were then segmented using circular binary segmentation61. Variants on the sex chromosomes (X and Y) were excluded from this analysis. Tumor cellularity was determined based on the VAF and segmented copy number data using ABSOLUTE62, in order to determine the cancer cell fraction (CCF) of each mutation, as was previously described by McGranahan et al.13. The clonal status was defined according to the confidence interval (CI) of the CCF. Mutations were classified as subclonal if the upper bound of their 95% CI was less than 1. Identification of putative driver mutations We first identified putative cancer driver genes based on recent large-scale ESCC sequencing data4–8, COSMIC cancer gene census (Aug 2015)11 and Pan-cancer analysis12. Next, non-silent variants in these genes were evaluated, and putative driver mutations were identified if they met one of the following requirements: i) Either the exact mutation, the same mutation site or at least three mutations located within 15 bp around the variant were found in COSMIC; ii) If the candidate gene was remarked to be recessive in COSMIC, and the variant was predicted to be deleterious, including stopgain, frameshift and splicing mutation and, had a SIFT score < 0.0563 or a Polyphen score > 0.99564,65. Mutational signature analysis Both silent and non-silent somatic mutations were classified as either truncal or branch as described earlier, and the mutational signatures of these variants were separately generated. We performed a multiple regression approach, deconstructSigs26, to extract the signatures based on Wellcome Trust Sanger Institute Mutational Signature Framework27, and to quantify statistically the contribution of each signature for each tumor. DNA methylation analysis and construction of phyloepigenetic tree DNA methylation profiles of 12 tumor regions and 2 matched normal esophageal epithelial tissues from three M-WES-examined ESCC cases (ESCC01, ESCC03 and ESCC05) were generated using the Illumina Infinium HumanMethylation450K platform (Illumina, San Diego, CA) by the University of Southern California Norris Comprehensive Cancer Center Genomics Core. We performed basic data processing of the HM450k data using many of the same processing steps that we have performed previously for TCGA data analysis, which is based on the Methylumi R package66 with several additional quality control steps. Probes with the detection p-value greater than 0.01 in any of the samples were removed, as were probes overlapping with dbSNP SNPs, and probes on the X or Y chromosomes. For intratumoral analysis, we defined a probe as “private” if the difference in beta values between any single pair of tumor regions was at least 0.3, and defined a probe as “shared” if the differences in beta values between all pairs of tumor regions were less than 0.1. Only private probes were used for construction of phyloepigenetic trees. For each tumor, pairwise Euclidean distances were calculated between all tumor regions using the complete set of private probes. The phyloepigenetic tree was constructed from these pairwise distances, using the Minimal Evolution method implemented by the fastme.bal function in the R package ape67. Different probe-selection cutoffs for calling private and shared probes produced similar results, with only minimal changes in case ESCC01 (at the cutoff of 0.5) and ESCC05 (at the cutoff of 0.2, Supplementary Fig. 6). The topological comparison of phylogenetic vs. phyloepigenetic and other tree pairs was performed using RF.dist function in the CRAN R package phangorn. The comparison in the case ESCC01 was done based on only tumor samples due to the lack of matched normal in DNA methylation data (for visualization purpose in Fig. 4a, we used normal samples from the other two cases as the root.) To mitigate confounding effects of non-cancer cells in phyloepigenetic tree reconstruction, we performed additional bioinformatic analyses, as described below. The major source of nonmalignant DNA contamination in the esophageal tumor is from immune cells68, which has been shown by TCGA to be the case for most solid tumors62,69. We confirmed this by review of all of our methylation-profiled samples through immunostaining of the leukocyte common antigen (LCA)/CD45 (representative immunohistochemical photos in Supplementary Fig. 7), which is a common marker of the immune cells and widely used in distinguishing the infiltration of immune cells70–73. To precisely determine the extent of immune cell contamination, we estimated the fraction of leukocyte cells in each sample using profiles of immune-specific methylation probes74, as described previously62,69. Using this method, we noted that case ESCC01 was highly pure (estimated immune-cell fraction 7.1%, ranging from 1% to 14% in various tumor regions) and case ESCC03 and ESCC05 contained an average of 20% and 32% immune cells, respectively (Supplementary Table 7). We re-calculated each phyloepigenetic tree using one of two methods to model the mixture of cancer and and immune cells within the samples: As demonstrated in several TCGA marker papers, performing an analysis using only the subset of Infinium probes that are unmethylated in purified leukocyte cells, and dichotomizing/binarizing the tumor beta value of these probes with a minimum beta value cutoff, could minimize the influence of contaminating leukocytes75–77. We used HM450k profiles from purified leukoctye populations74, and selected probes with a maximum beta value less than 0.3 across all leukocyte samples. We then binarized the tumor beta values as 1 if they were > 0.3 and 0 otherwise. We computed pairwise distances between binarized values using the Jaccard index (dist function in R package), and performed tree construction using these pairwise distances as described above. The resulting trees are labeled “Dichotomized” in Supplementary Fig. 8. In an independent approach, we modeled the tumor beta value as a linear combination of DNA from a mixture of cancer and leukocyte cells. The mixing ratio was estimated for each sample based on methylation of leukocyte-specific probes, as described above and in previously62,69. For each probe, we used the fixed mixing ratio, the average beta value of the probe in purified leukocytes74, and our measured beta value in the tumor, to estimate the methylation beta value of the cancer cells alone. This method was used to reconstruct phyloepigenetic trees for each case, and the resulting trees are labeled “Immune cell adjusted” in Supplementary Fig. 8. Trees for both methods (i) and (ii) were compared to original trees using the RF method described above, and RF values are listed in Supplementary Fig. 8. Determining the genomic context of shared vs. private methylation patterns Shared vs. Private probes were determined based on heterogeneity between different tumor regions, as described above. These were further divided “hypermethylated” and “hypomethylated”, based on comparisons between the tumor methylation and that of adjacent normal tissue. For hypermethylated probes within a specific case, we selected all probes with a methylation beta value < 0.3 in the adjacent normal sample (or a maximum beta value of two other normal samples < 0.3 for ESCC01) and a mean beta value across all tumor regions that was at least 0.3 above from mean of normal sample(s). Similarly for “hypomethylation”, we selected those probes with >= 0.6 in the normal, and at least below 0.3 for the tumor mean from mean of normals. For ESCC01, which had no matched normal, we averaged the beta values from the other two normals. Hyper- and hypomethylated probe sets are shown in Fig. 4b–d, and Supplementary Fig. 9. For the enrichment analysis in Fig. 4c, promoters were defined as 1.5 kbp up/down stream of RefSeq TSS, CpG islands were taken from the HMM-defined set78, shores and enhancers were defined using standard Illumina 450K annotation manifest. Partially Methylated Domains (PMD) were called using the Roadmap79 normal esophagus sample (ID: E079), using an HMM-based segmentation method80. Enrichment/Depletion p-values for the enrichment of private vs. shared probes within each genomic context were computed based on a hypergeometric test based, where null model frequencies were calculated based on all probes present on the array (shown as “Background” in Fig. 4c). Immunohistochemistry (IHC) Formalin-fixed and paraffin-embedded tissue slides were deparaffinized using xylene, rehydrated using xylene and ethanol, and then immersed in 3% hydrogen peroxide solution for 10 min, heated in citrate at 95°C for 25 min, and cooled at room temperature for 60 min. The slides were incubated overnight at 4°C with the leukocyte common antigen (LCA)/CD45 antibody (Cell Marque, 145M-96, USA; diluted at 1:100), and visualized using PV-9000 Polymer Detection System following the manufacturer’s instructions (Beijing Zhongshan Golden Bridge Biotechnology Co. Ltd., China). Counterstaining was carried out with hematoxylin. Supplementary Material 1 2 3 4 5 6 7 We thank Dr. Hui Shen (Van Andel Institute), Dr. Dan Weisenberger (University of Southern California) as well as Dr. Anand D Jeyasekharan (The Singapore Gastric Cancer Consortium) for their kind help on analysis and discussion. This work was funded by the Singapore Ministry of Health’s National Medical Research Council (NMRC) under its Singapore Translational Research (STaR) Investigator Award to H.P.K., NMRC Individual Research Grant (NMRC/1311/2011) and the NMRC Centre Grant awarded to National University Cancer Institute of Singapore, the National Research Foundation Singapore and the Singapore Ministry of Education under its Research Centres of Excellence initiatives to H.P.K. D.-C.L. was supported by American Society of Hematology Fellow Scholar Award, National Natural Science Foundation of China (81672786) and National Center for Advancing Translational Sciences UCLA CTSI Grant UL1TR000124. M.-R.W was supported by National Natural Science Foundation of China (81330052, 81520108023 and 81321091). Y.Z. was supported by Beijing Natural Science Foundation (7151008). This study was partially supported by a generous donation from the Melamed family, and NIH/NCI grant 1U01CA184826 as well as institutional support from the Samuel Oschin Comprehensive Cancer Institute to B.P.B and H.Q.D. Fig. 1 ITH of somatic mutations in 13 ESCCs generated by M-WES (a) Phylogenetic trees were constructed from all somatic mutations by the Wagner parsimony method using PHYLIP (See Method). Lengths of trunks and branches are proportional to the numbers of mutations acquired. Heat maps showed the presence (blue) or absence (gray) of a somatic mutation in each tumor region (T). Each gene was arranged in a row, and cancer genes with putative driver mutations were indicated. The total number of mutations (n), and the proportions of branched mutations in each case, were provided above each tree. (b) Bar plots showed the proportions of putative driver mutations versus other mutations on the trunks and branches. Statistical differences of truncal and branched proportions, between driver and other mutations across all cases, were analyzed using a χ2 test, and a significant P value was shown. Fig. 2 Clonal status of putative driver mutations in ESCC tumors A heatmap displayed the cancer cell fraction (CCF) of driver mutations in each region of the ESCC tumors. Genomic regions with no segmentation data available were shown as NA. Fig. 3 Temporal dissection of mutational signatures in ESCC tumors (a) The 96 trinucleotide mutational spectrum of truncal (Bottom panel) and branched (Top panel) mutations across all regions was inferred by deconstructSigs. (b) Dot plots displayed the contributions of individual mutational signatures to individual cases, with each dot representing one case. Signatures 1–30 were based on the Wellcome Trust Sanger Institute COSMIC Mutational Signature Framework. Inferred signatures included: Signature 1 (associated with age), Signatures 2 and 13 (associated with APOBEC), Signatures 6 and 15 (associated with DNA mismatch repair), Signature 3 (associated with DNA double-strand break-repair), Signature 7 (associated with UV exposure in squamous cancer). The bars represent the mean values. (c, d) Piecharts displayed the truncal and branch mutational signatures in cases ESCC10 and ESCC12, and only signatures with contributions over 10% were indicated. Fig. 4 Epigenetic ITH in ESCC (a) Phyloepigenetic trees of three ESCC cases. Lengths of trunks and branches were inferred using a phylogenetic approach, based on Euclidean distances between different tumor regions using private probes (see Methods). The total number of probes (n) was provided above each tree. For comparison, phylogenetic trees from Fig. 1 were reproduced below each phyloepigenetic tree. (b) Heatmaps showed the beta values of private probes for each case, separated into hyper- and hypo-methylation. (c) Overlap between each probe set from panel (b), and a variety of functional genomic contexts: non-CpG Island Promoters (nCGI-Prom), non-Promoter CpG Islands (CGI-nProm), CpG Island Promoters (CGI-Prom), CpG Island Shores (CGI-Shore), Partially Methylated Domains excluding CpG Islands (nCGI-PMD) and enhancers. Overlapping frequencies of private probes from panel (b) were shown in yellow, shared probes (Supplementary Fig. 9) in green, and gray showed the frequency for the entire set of probes on the array. The hypergeometric test (* = P < 10−5) was used to compare the frequency of each private and shared probe set category to that of array background (see Methods). (d) Enriched GO biological processes for the genes associated with privately hypermethylated promoters in ESCC01 and ESCC03 (case ESCC05 was excluded due to the lack of sufficient privately hypermethylated promoters). Table 1 Prevalence of non-silent mutations in ESCC (within-patient versus within-region) Cancer gene Prevalence (number of patients with mutations) in previous studies* Within-region prevalence (number of regions with mutations) n = 51 regions Within-patient prevalence (number of patients with mutations) n = 13 cases Within-patient/ within-region TP53 78.9% (430) 94.1% (48) 92.3% (12) 0.98 KMT2D 13.8% (63) 23.5% (12) 23.1% (3) 0.98 NOTCH1 12.8% (70) 21.6% (11) 23.1% (3) 1.07 FAT1 11.2% (51) 15.7% (8) 15.4% (2) 0.98 ZNF750 5.7% (26) 15.7% (8) 15.4% (2) 0.98 FAM135B 6.4% (29) 13.7% (7) 15.4% (2) 1.12 NFE2L2 5.7% (26) 7.8% (4) 15.4% (2) 1.97 PTPRB 2.9% (13) 7.8% (4) 15.4% (2) 1.97 ATM 1.8% (8) 7.8% (4) 7.7% (1) 0.98 BRCA2 3.1% (14) 7.8% (4) 7.7% (1) 0.98 CREBBP 4.2% (19) 7.8% (4) 7.7% (1) 0.98 KMT2A 1.1% (5) 7.8% (4) 7.7% (1) 0.98 NOTCH2 3.3% (18) 7.8% (4) 7.7% (1) 0.98 FAT2 6.4% (29) 5.9% (3) 7.7% (1) 1.31 KEAP1 1.8% (8) 5.9% (3) 7.7% (1) 1.31 MTOR 1.1% (5) 3.9% (2) 7.7% (1) 1.96 TP53BP1 0.9% (4) 3.9% (2) 7.7% (1) 1.96 KIT 0.7% (3) 3.9% (2) 7.7% (1) 1.96 PIK3CA 9.0% (41) 3.9% (2) 7.7% (1) 1.96 ATR 1.1% (5) 2.0% (1) 7.7% (1) 3.92 BRIP1 0.9% (4) 2.0% (1) 7.7% (1) 3.92 TSC1 1.1% (5) 2.0% (1) 7.7% (1) 3.92 * Summary of published data from Agrawal et al. (Ref. 4), Song et al. (Ref. 5), Lin et al. (Ref. 6), Gao et al. (Ref. 7), and Zhang et al. (Ref. 8). The total number of cases is 545 for TP53, NOTCH1 and NOTCH2 mutations, and is 456 for the rest gene mutations. The last column showed the fold change when the prevalence was analyzed using individual cases instead of individual tumor regions. Table 2 Prevalence of subclonal mutations in ESCC Case Within-region prevalence Within-patient prevalence T1 T2 T3 T4 ESCC01 10.1% 16.1% 26.7% 15.4% 40.0% ESCC02 14.7% 8.2% 10.4% 14.9% 20.5% ESCC03 13.6% 7.2% 8.4% 24.1% 33.2% ESCC04 10.7% 5.8% NA 1.2% 13.3% ESCC05 27.3% 21.4% 3.6% 33.3% 48.8% ESCC06 6.9% 28.3% 5.4% 6.1% 33.3% ESCC07 6.1% 21.1% 92.4% 61.1% 86.1% ESCC08 11.6% 12.7% 15.4% 16.2% 31.7% ESCC09 30.4% 41.3% 5.7% 20.0% 56.5% ESCC10 21.2% 2.0% 3.1% 6.1% 27.0% ESCC11 42.3% 35.5% 36.0% 41.7% 66.4% ESCC12 1.4% 38.6% 6.1% 46.3% 62.5% ESCC13 29.5% 3.0% 14.4% 29.5% 50.0% Note: The within-patient prevalence was derived through dividing the number of subclonal mutations by the number of total mutations in each patients. URLs BWA-MEM, http://arxiv.org/abs/1303.3997v2. fpFilter perl script, https://github.com/ckandoth/variant-filter. Bam-readcount, https://github.com/genome/bam-readcount. PHYLIP, http://evolution.genetics.washington.edu/phylip.html. Accession codes. Digital sequencing and HM450 Bead array files have been deposited into Sequence Read Archive (SRP072112) and Gene Expression Omnibus (GSE79366), respectively. AUTHOR CONTRIBUTIONS M.-R.W., D.-C.L., B.P.B. and H.P.K. conceived and designed the experiments. J.-J.H., D.-C.L., H.Q.D., W.-Q.W. B.P.B., M.-R.W., and H.P.K. wrote the manuscript. J.-J.H., D.-C.L., Y.J., C.C., C.-C.L., X.X., Y.C. performed the experiments. J.-J.H., H.Q.D., A.M., B.P.B., and Z.-Z.S. performed statistical analysis. J.-J.H., D.-C.L., H.Q.D., Y.-Y.J., B.P.B. and H.P.K. analyzed the data. X.X. contributed reagents. W.-Q.W. contributed materials. J.-W.W. and J.-J.H. read the H&E slides. D.-C.L., Y.Z., Q.-M.Z. and H.P.K. jointly supervised research. COMPETING FINANCIAL INTERESTS The authors declare no competing financial interests. REFERENCES 1 Torre LA Global cancer statistics, 2012 CA Cancer J Clin 2015 65 87 108 25651787 2 Ferlay J Cancer incidence and mortality worldwide: sources, methods and major patterns in GLOBOCAN 2012 Int J Cancer 2015 136 E359 E386 25220842 3 Enzinger PC Mayer RJ Esophageal cancer N Engl J Med 2003 349 2241 2252 14657432 4 Agrawal N Comparative genomic analysis of esophageal adenocarcinoma and squamous cell carcinoma Cancer Discov 2012 2 899 905 22877736 5 Song Y Identification of genomic alterations in oesophageal squamous cell cancer Nature 2014 509 91 95 24670651 6 Lin DC Genomic and molecular characterization of esophageal squamous cell carcinoma Nat Genet 2014 46 467 473 24686850 7 Gao YB Genetic landscape of esophageal squamous cell carcinoma Nat Genet 2014 46 1097 1102 25151357 8 Zhang L Genomic analyses reveal mutational signatures and frequently altered genes in esophageal squamous cell carcinoma Am J Hum Genet 2015 96 597 611 25839328 9 Cheng C Whole-Genome Sequencing Reveals Diverse Models of Structural Variations in Esophageal Squamous Cell Carcinoma Am J Hum Genet 2016 98 256 274 26833333 10 McGranahan N Swanton C Biological and therapeutic impact of intratumor heterogeneity in cancer evolution Cancer Cell 2015 27 15 26 25584892 11 Futreal PA A census of human cancer genes Nat Rev Cancer 2004 4 177 183 14993899 12 Lawrence MS Discovery and saturation analysis of cancer genes across 21 tumour types Nature 2014 505 495 501 24390350 13 McGranahan N Clonal status of actionable driver events and the timing of mutational processes in cancer evolution Sci Transl Med 2015 7 283ra54 14 Durinck S Temporal dissection of tumorigenesis in primary cancers Cancer Discov 2011 1 137 143 21984974 15 Shah SP The clonal and mutational evolution spectrum of primary triple-negative breast cancers Nature 2012 486 395 399 22495314 16 Lohr JG Widespread genetic heterogeneity in multiple myeloma: implications for targeted therapy Cancer Cell 2014 25 91 101 24434212 17 Landau DA Evolution and impact of subclonal mutations in chronic lymphocytic leukemia Cell 2013 152 714 726 23415222 18 Shi ZZ Consistent and differential genetic aberrations between esophageal dysplasia and squamous cell carcinoma detected by array comparative genomic hybridization Clin Cancer Res 2013 19 5867 5878 24009147 19 Shang L ANO1 protein as a potential biomarker for esophageal cancer prognosis and precancerous lesion development prediction Oncotarget 2016 7 24374 24382 27016410 20 Britschgi A Calcium-activated chloride channel ANO1 promotes breast cancer progression by activating EGFR and CAMK signaling Proc Natl Acad Sci U S A 2013 110 E1026 E1034 23431153 21 Luo ML Amplification and overexpression of CTTN (EMS1) contribute to the metastasis of esophageal squamous cell carcinoma by promoting cell migration and anoikis resistance Cancer Res 2006 66 11690 11699 17178864 22 Lu P Genome-wide gene expression profile analyses identify CTTN as a potential prognostic marker in esophageal cancer PLoS One 2014 9 e88918 24551190 23 de Bruin EC Spatial and temporal diversity in genomic instability processes defines lung cancer evolution Science 2014 346 251 256 25301630 24 Gerlinger M Genomic architecture and evolution of clear cell renal cell carcinomas defined by multiregion sequencing Nat Genet 2014 46 225 233 24487277 25 Murugaesu N Tracking the genomic evolution of esophageal adenocarcinoma through neoadjuvant chemotherapy Cancer Discov 2015 5 821 831 26003801 26 Rosenthal R McGranahan N Herrero J Taylor BS Swanton C deconstructSigs: delineating mutational processes in single tumors distinguishes DNA repair deficiencies and patterns of carcinoma evolution Genome Biol 2016 17 31 26899170 27 Alexandrov LB Signatures of mutational processes in human cancer Nature 2013 500 415 421 23945592 28 Toh Y Egashira A Yamamoto M Epigenetic alterations and their clinical implications in esophageal squamous cell carcinoma Gen Thorac Cardiovasc Surg 2013 61 262 269 23529258 29 Agarwal R Epigenomic program of Barrett's-associated neoplastic progression reveals possible involvement of insulin signaling pathways Endocr Relat Cancer 2012 19 L5 L9 22194443 30 Alvarez H Widespread hypomethylation occurs early and synergizes with gene amplification during esophageal carcinogenesis PLoS Genet 2011 7 e1001356 21483804 31 Brocks D Intratumor DNA methylation heterogeneity reflects clonal evolution in aggressive prostate cancer Cell Rep 2014 8 798 806 25066126 32 Mazor T DNA Methylation and Somatic Mutations Converge on the Cell Cycle and Define Similar Evolutionary Histories in Brain Tumors Cancer Cell 2015 28 307 317 26373278 33 Robinson DF Foulds LR Comparison of phylogenetic trees Math Biosci 1981 53 131 147 34 Oricchio E The Eph-receptor A7 is a soluble tumor suppressor for follicular lymphoma Cell 2011 147 554 564 22036564 35 Lopez-Nieva P EPHA7, a new target gene for 6q deletion in T-cell lymphoblastic lymphomas Carcinogenesis 2012 33 452 458 22114070 36 Yu J Methylation of protocadherin 10, a novel tumor suppressor, is associated with poor prognosis in patients with gastric cancer Gastroenterology 2009 136 640.e1 651.e1 19084528 37 Zhao Y A novel wnt regulatory axis in endometrioid endometrial cancer Cancer Res 2014 74 5103 5117 25085246 38 Saulnier A Inactivation of the putative suppressor gene DOK1 by promoter hypermethylation in primary human cancers Int J Cancer 2012 130 2484 2494 21796618 39 Mercier PL Characterization of DOK1, a candidate tumor suppressor gene, in epithelial ovarian cancer Mol Oncol 2011 5 438 453 21856257 40 Bergman Y Cedar H DNA methylation dynamics in health and disease Nat Struct Mol Biol 2013 20 274 281 23463312 41 Quante T Bird A Do short, frequent DNA sequence motifs mould the epigenome? Nat Rev Mol Cell Biol 2016 17 257 262 26837845 42 Baylin SB Jones PA A decade of exploring the cancer epigenome - biological and translational implications Nat Rev Cancer 2011 11 726 734 21941284 43 Lay FD The role of DNA methylation in directing the functional organization of the cancer epigenome Genome Res 2015 25 467 477 25747664 44 Irizarry RA The human colon cancer methylome shows similar hypo- and hypermethylation at conserved tissue-specific CpG island shores Nat Genet 2009 41 178 186 19151715 45 Doi A Differential methylation of tissue- and cancer-specific CpG island shores distinguishes human induced pluripotent stem cells, embryonic stem cells and fibroblasts Nat Genet 2009 41 1350 1353 19881528 46 Gharwan H Groninger H Kinase inhibitors and monoclonal antibodies in oncology: clinical implications Nat Rev Clin Oncol 2016 13 209 227 26718105 47 Gerlinger M Cancer: evolution within a lifetime Annu Rev Genet 2014 48 215 236 25292359 48 Cao W Multiple region whole-exome sequencing reveals dramatically evolving intratumor genomic heterogeneity in esophageal squamous cell carcinoma Oncogenesis 2015 4 e175 26619400 49 Yap TA Gerlinger M Futreal PA Pusztai L Swanton C Intratumor heterogeneity: seeing the wood for the trees Sci Transl Med 2012 4 127ps10 50 Wang J Downregulation of EphA7 by hypermethylation in colorectal cancer Oncogene 2005 24 5637 5647 16007213 51 Kiehl S ABCB4 is frequently epigenetically silenced in human cancers and inhibits tumor growth Sci Rep 2014 4 6899 25367630 52 Jao TM Protocadherin 10 suppresses tumorigenesis and metastasis in colorectal cancer and its genetic loss predicts adverse prognosis Int J Cancer 2014 135 2593 2603 24740680 53 Narayan G PCDH10 promoter hypermethylation is frequent in most histologic subtypes of mature lymphoid malignancies and occurs early in lymphomagenesis Genes Chromosomes Cancer 2013 52 1030 1041 23929756 METHODS-ONLY REFERENCES 54 Chen W Cancer statistics in China, 2015 CA Cancer J Clin 2016 66 115 132 26808342 55 Faust GG Hall IM SAMBLASTER: fast duplicate marking and structural variant read extraction Bioinformatics 2014 30 2503 2505 24812344 56 Van der Auwera GA From FastQ data to high confidence variant calls: the Genome Analysis Toolkit best practices pipeline Curr Protoc Bioinformatics 2013 43 11.10.1 11.10.33 25431634 57 Koboldt DC VarScan 2: somatic mutation and copy number alteration discovery in cancer by exome sequencing Genome Res 2012 22 568 576 22300766 58 Li H The Sequence Alignment/Map format and SAMtools Bioinformatics 2009 25 2078 2079 19505943 59 Wang K Li M Hakonarson H ANNOVAR: functional annotation of genetic variants from high-throughput sequencing data Nucleic Acids Res 2010 38 e164 20601685 60 Stachler MD Paired exome analysis of Barrett's esophagus and adenocarcinoma Nat Genet 2015 47 1047 1055 26192918 61 Olshen AB Venkatraman ES Lucito R Wigler M Circular binary segmentation for the analysis of array-based DNA copy number data Biostatistics 2004 5 557 572 15475419 62 Carter SL Absolute quantification of somatic DNA alterations in human cancer Nat Biotechnol 2012 30 413 421 22544022 63 Kumar P Henikoff S Ng PC Predicting the effects of coding non-synonymous variants on protein function using the SIFT algorithm Nat Protoc 2009 4 1073 1081 19561590 64 Adzhubei IA A method and server for predicting damaging missense mutations Nat Methods 2010 7 248 249 20354512 65 Adzhubei I Jordan DM Sunyaev SR Predicting functional effect of human missense mutations using PolyPhen-2 Curr Protoc Hum Genet 2013 Chapter 7 Unit 7 20 66 Triche TJ Jr Weisenberger DJ Van Den Berg D Laird PW Siegmund KD Low-level processing of Illumina Infinium DNA Methylation BeadArrays Nucleic Acids Res 2013 41 e90 23476028 67 Desper R Gascuel O Fast and accurate phylogeny reconstruction algorithms based on the minimum-evolution principle J Comput Biol 2002 9 687 705 12487758 68 Takahashi T Estimation of the fraction of cancer cells in a tumor DNA sample using DNA methylation PLoS One 2013 8 e82302 24312652 69 Zack TI Pan-cancer patterns of somatic copy number alteration Nat Genet 2013 45 1134 1140 24071852 70 Pages F In situ cytotoxic and memory T cells predict outcome in patients with early-stage colorectal cancer J Clin Oncol 2009 27 5944 5951 19858404 71 de Miranda NF Infiltration of Lynch colorectal cancers by activated immune cells associates with early staging of the primary tumor and absence of lymph node metastases Clin Cancer Res 2012 18 1237 1245 22261803 72 Punt S Whole-transcriptome analysis of flow-sorted cervical cancer samples reveals that B cell expressed TCL1A is correlated with improved survival Oncotarget 2015 6 38681 38694 26299617 73 Gorter A Prins F van Diepen M Punt S van der Burg SH The tumor area occupied by Tbet+ cells in deeply invading cervical cancer predicts clinical outcome J Transl Med 2015 13 295 26357849 74 Reinius LE Differential DNA methylation in purified human blood cells: implications for cell lineage and studies on disease susceptibility PLoS One 2012 7 e41361 22848472 75 Cancer Genome Atlas Research N Comprehensive molecular characterization of gastric adenocarcinoma Nature 2014 513 202 209 25079317 76 Ceccarelli M Molecular Profiling Reveals Biologically Discrete Subsets and Pathways of Progression in Diffuse Glioma Cell 2016 164 550 563 26824661 77 Cancer Genome Atlas Research N The Molecular Taxonomy of Primary Prostate Cancer Cell 2015 163 1011 1025 26544944 78 Wu H Caffo B Jaffee HA Irizarry RA Feinberg AP Redefining CpG islands using hidden Markov models Biostatistics 2010 11 499 514 20212320 79 Roadmap Epigenomics C. et al Integrative analysis of 111 reference human epigenomes Nature 2015 518 317 330 25693563 80 Song Q A reference methylome database and analysis pipeline to facilitate integrative and comparative epigenomics PLoS One 2013 8 e81148 24324667

PMC005xxxxxx/PMC5127919.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 0404511 7473 Science Science Science (New York, N.Y.) 0036-8075 1095-9203 27634526 5127919 10.1126/science.aag3042 NIHMS831116 Article Virulence factors enhance Citrobacter rodentium expansion through aerobic respiration Lopez Christopher A. 1 Miller Brittany M. 1 Rivera-Chávez Fabian 1 Velazquez Eric 1 Byndloss Mariana X. 1 Chávez-Arroyo Alfredo 1 Lokken Kristen L. 1 Tsolis Renée M. 1 Winter Sebastian E. 2 Bäumler Andreas J. 1* 1 Department of Medical Microbiology and Immunology, School of Medicine, University of California, Davis, One Shields Avenue, Davis, CA, USA 2 Department of Microbiology, University of Texas Southwestern Medical Center, 5323 Harry Hines Blvd, Dallas, TX, USA * To whom correspondence should be addressed. ajbaumler@ucdavis.edu 21 11 2016 15 9 2016 16 9 2016 30 11 2016 353 6305 12491253 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. Citrobacter rodentium uses a type III secretion system (T3SS) to induce colonic crypt hyperplasia in mice, thereby gaining an edge during its competition with the gut microbiota through an unknown mechanism. Here we show that by triggering colonic crypt hyperplasia, the C. rodentium T3SS induced an excessive expansion of undifferentiated Ki67-positive epithelial cells, which increased oxygenation of the mucosal surface and drove an aerobic C. rodentium expansion in the colon. Treatment of mice with the γ-secretase inhibitor dibenzazepine to diminish Notch-driven colonic crypt hyperplasia curtailed the fitness advantage conferred by aerobic respiration during C. rodentium infection. We conclude that C. rodentium uses its T3SS to induce histopathological lesions that generate an intestinal microenvironment in which growth of the pathogen is fueled by aerobic respiration. Attaching and effacing (AE) pathogens are defined by virulence characteristics encoded by a pathogenicity island known as the locus of enterocyte effacement (LEE) (1, 2), which contains genes encoding for a T3SS (3) and an adhesin termed intimin (4). The T3SS injects the intimin receptor Tir into the host cell cytosol (5), resulting in intimate attachment of bacteria to the effaced epithelial surface (6). The LEE encoded T3SS of the AE pathogen C. rodentium injects additional effector proteins that are required for causing transmissible colonic crypt hyperplasia in mice (7, 8) (Fig. S1). Following the development of colonic crypt hyperplasia, C. rodentium blooms in the lumen of the murine large bowel (9). The LEE encoded T3SS is required for this rapid luminal expansion possibly by allowing C. rodentium to compete with the microbiota for carbon sources, because the T3SS provides no benefit in germ-free mice (10). These data suggest that the T3SS places C. rodentium in a microenvironment that somehow provides a growth advantage during its competition with the resident microbiota, but it remains obscure which resources might become available in this niche to fuel pathogen expansion. Electron acceptors, such as nitrate, are produced as a by-product of the inflammatory host response and boost luminal growth of pathogenic Salmonella enterica or commensal Escherichia coli by anaerobic respiration in mouse models of colitis (11–13). Since C. rodentium infection triggers colonic crypt hyperplasia, we wanted to determine whether the inflammatory host response would enable the pathogen to grow by anaerobic respiration. The respiratory reductases for nitrate, dimethyl sulfoxide (DMSO) and trimethylamine N-oxide (TMAO) as well as the formate dehydrogenases FdnGHI and FdoGHI contain a molybdopterin cofactor. Thus, to explore a possible role of anaerobic respiration during C. rodentium growth in the mouse gut, we constructed a C. rodentium mutant lacking a gene required for molybdopterin cofactor biosynthesis (moaA mutant) (Fig. S2A) (14). Mice (C57BL/6) were inoculated with an equal mixture of wild-type C. rodentium and an isogenic moaA mutant to compare the fitness of both strains. Mice developed intestinal inflammation as indicated by increased transcript levels of pro-inflammatory markers in the colonic mucosa (Fig. S3A and S3B). The C. rodentium wild type was recovered in significantly (P < 0.05) higher numbers than the moaA mutant (Fig. 1A). Similar results were observed with genetically susceptible C3H/HeJ mice that experience more severe intestinal inflammation during C. rodentium infection (Fig. S3C and S3D). In contrast, when germ-free mice were inoculated with an equal mixture of the C. rodentium wild type and a moaA mutant, both strains were recovered in similar numbers (Fig. 1B and S3E), suggesting that either anaerobic respiration or the utilization of formate provided an edge during competition of the pathogen with the resident microbiota. C. rodentium infection resulted in a markedly increased colonic expression of Nos2 (Fig. S3B and S3D), the gene encoding inducible nitric oxide synthase (iNOS), an enzyme necessary for generating nitric oxide, which is converted to nitrate in the intestinal lumen (11). To determine whether nitrate respiration enhances growth, we constructed a mutant deficient for the three nitrate reductases encoded by C. rodentium (narG napA narZ mutant) (Fig. S2B and S2F). Remarkably, in mice inoculated with an equal mixture of the C. rodentium wild type and a narG napA narZ mutant, both strains were recovered in equal numbers from colon contents and feces (Fig. 1A, S3C and S3E), suggesting that nitrate respiration did not provide a fitness advantage. We next examined the possibility that the phenotype of the moaA mutant was due to an inability to utilize formate as an electron donor rather than nitrate as an electron acceptor. Since formate dehydrogenases FdnGHI and FdoGHI can couple electron transfer from formate through the quinone pool to nitrate (15) or oxygen (16), respectively, we envisioned that analysis of this pathway would provide indirect information about the electron acceptor used by C. rodentium in the gut. To test whether the nitrate-dependent and/or the oxygen-dependent formate dehydrogenase contributed to growth of C. rodentium during colitis, mice were inoculated with an equal mixture of the C. rodentium wild type and either an fdnG mutant or fdoG mutant (Fig. S2C and S2D). Consistent with our observation that nitrate respiration did not contribute to fitness (Fig. 1A), inactivation of the nitrate-dependent fdnG did not reduce growth of C. rodentium (Fig. 1B). However, inactivation of the oxygen-dependent fdoG reduced fitness of C. rodentium in conventional (C57BL/6) mice (P < 0.05), while little benefit was observed in germ-free (Swiss Webster) mice (Fig. 1B and S3E). The finding that fdoG was required for pathogen expansion was intriguing, because it suggested that C. rodentium might utilize oxygen as a terminal electron acceptor during luminal growth. We thus wanted to test the hypothesis that aerobic respiration fuels C. rodentium growth during infection. To test this idea, we deleted the C. rodentium cydAB genes (Fig. S2E), which encode a high-affinity cytochrome bd oxidase required for aerobic respiration in microaerobic environments. In vitro, the C. rodentium wild type was recovered in higher numbers than the cydAB mutant when both strains were co-cultured under conditions mimicking hypoxia (1% O2) or normal tissue oxygenation (4% or 8% O2), but not under anaerobic conditions (0% O2) or at atmospheric oxygen levels (21% O2) (Fig. 1C). Next, we compared the fitness of the C. rodentium wild type and a cydAB mutant by infecting C57BL/6 mice with a 1:1 mixture of both strains. Remarkably, the C. rodentium wild type was recovered in 20,000-fold higher numbers (P < 0.01) than the cydAB mutant from colon contents seven days after infection (Fig. 1D), suggesting that aerobic respiration contributed to growth of the pathogen. Expression of ler, encoding the master regulator of T3SS synthesis, was reduced in the cydAB mutant, but this difference did not reach statistical significance (Fig. S4A). When infection with a 1:1 mixture of the C. rodentium wild type and a cydAB mutant was repeated in germ-free (Swiss Webster) mice, the fitness advantage provided by cydAB was greatly reduced (Fig. 1D). Collectively, these data suggested that aerobic respiration drove an uncontrolled luminal expansion of C. rodentium in an environment occupied by the gut microbiota. To test whether tissue/mucus-associated bacteria have a respiratory metabolism compared to luminal bacteria, we investigated transcription of the sucA gene, which encodes a subunit of 2-ketoglutarate dehydrogenase, an enzyme required for the aerobic tricarboxylic acid (TCA) cycle. Under anaerobic conditions, SucA is no longer required as the normally cyclic TCA pathway switches to a non-cyclic series of reactions. Higher expression of sucA under aerobic than anaerobic conditions can serve as an indicator for a respiratory central metabolism (17). C. rodentium sucA was transcribed at significantly (P < 0.05) higher levels when bacteria were cultured in vitro under microaerobic conditions compared to anaerobic conditions (Fig. 1E and S4B). To investigate sucA expression in vivo, RNA was isolated from mucus scrapings or colon contents of mice infected with C. rodentium. C. rodentium residing in close proximity to tissue (i.e. mucus scrapings) transcribed sucA at significantly (P < 0.05) higher levels than bacteria located in colon contents (Fig. 1F and S4C). These results provided further support for the idea that bacteria in close proximity to the mucosal surface had an oxidative metabolism in vivo. Intestinal inflammation caused by Salmonella enterica drives a depletion of Clostridia, which in turn increases epithelial oxygenation (18). We thus wanted to determine whether C. rodentium-induced colonic crypt hyperplasia would increase oxygen availability in the gut by driving a depletion of Clostridia. The numbers of C. rodentium recovered from colon contents of mice infected with the C. rodentium wild type were 1,000-fold higher (P < 0.05) than those recovered from colon contents of mice infected with a cydAB mutant (Fig. 1G), which correlated with reduced colonic inflammation (P < 0.05) (Fig. 1H and S5). Consistent with a previous report (9), infection of mice with the C. rodentium wild type resulted in a shift in the microbial community structure (Fig. 1I and S6). The absolute abundances of members of the Enterobacteriaceae (Fig. S7A) and Erythrobacteraceae (Fig. S7B) were significantly (P < 0.05) increased in mice infected with the C. rodentium wild type, a shift in the community structure previously associated with increased oxygen availability in the gut (19). However, infection with the C. rodentium wild type resulted in an increased (P < 0.05) abundance of Clostridia (Fig. 1I), which was most pronounced for members of the Lachnospiraceae belonging to the genus Dorea (Fig. S7C) and the genus Coprococcus (Fig. S7D). These changes in the microbial population profile were reduced or absent in mice infected with a cydAB mutant (Fig. 1I, S6 and S7). In conclusion, our data were not consistent with the hypothesis that colonic crypt hyperplasia increased oxygen availability by depleting Clostridia. C. rodentium attaches to murine colonic epithelial cells (colonocytes) using a Type IV pilus encoded by the cfc operon (20). To test whether this adhesin contributed to establishing a niche in which the pathogen could grow by CydAB-mediated respiration, we constructed a cfcH mutant (Fig. S1F). Genetic ablation of Type IV pilus biosynthesis did not change the CydAB-dependent growth advantage seven days after infection of mice (Fig. 2A and Fig. S8A), possibly because a cfcH mutant still occupied a niche near the epithelium (Fig. S8B). C. rodentium intimately attaches to colonocytes in a T3SS-dependent fashion using intimin. To test whether the T3SS and intimin contributed to CydAB-dependent growth, we constructed mutations in escN, encoding a component of the T3SS, and eae, the gene encoding intimin (Fig. S1G and S1H). Interestingly, the fitness advantage conferred by CydAB-dependent respiration seven days after infection was significantly reduced in C. rodentium strains lacking eae or escN (Fig. 2A and S8A), suggesting that in the absence of intimate attachment or a functional T3SS, oxygen respiration no longer conferred a marked fitness advantage. However, at three days after infection, when intestinal inflammation was just beginning to develop (Fig. S8C and S8D), inactivation of escN did not significantly reduce the fitness advantage conferred by the cydAB genes (Fig. S8E). Since the escN mutation did not reduce the fitness advantage conferred by the cydAB genes in the absence of overt crypt hyperplasia (i.e. at three days after infection), we considered the possibility that enhanced access to oxygen might not depend on intimate attachment, but require T3SS-mediated host responses. Colonocytes mature as they migrate from the bottom of the crypt to the luminal surface. The mature surface colonocyte population functions in water absorption, which is driven by electrolyte transport (21). To energize absorption of electrolytes, mature colonocytes respire butyrate in their mitochondria (22), a process consuming oxygen and rendering the mucosal surface hypoxic (< 7.6 mmHg pO2 or < 1% O2) (23, 24) (Fig. S1). To investigate whether T3SS-induced colonic crypt hyperplasia, which developed by seven days after infection (Fig. S9A and S9B), would alter epithelial oxygenation in vivo, we visualized the “physiologic hypoxia” of surface colonocytes using the exogenous hypoxic marker pimonidazole (25) (Fig. 2B). Remarkably, a lack of hypoxia staining at the mucosal surface suggested that infection with the C. rodentium wild type significantly (P < 0.05) reduced epithelial hypoxia by seven days after infection (Fig. 2B, 2C and S10), indicative of a marked increase in epithelial oxygenation. This increased epithelial oxygenation was T3SS-dependent, because it was not observed in mice infected with the escN mutant (Fig. 2B and 2C), which did not develop inflammation in the colon (Fig. S8). T3SS-induced colonic crypt hyperplasia is characterized by an excessive intestinal epithelial repair response (reviewed in (26)). To further investigate the mechanism by which the T3SS increases epithelial oxygenation, we deleted the map, cesF and espH genes to reduce damage to colonocytes (Fig. S2H). Map and EspH are T3SS effector proteins that damage mitochondria (which might impair respiration of butyrate) or activate caspase-3 to induce cytotoxicity, respectively (27, 28), while the cesF gene encodes a chaperone for the mitochondria-associated T3SS effector EspF (29). Deletion of the map, cesF and espH genes significantly (P < 0.05) reduced the fitness advantage conferred by CydAB-dependent respiration seven days after infection of mice (Fig. 2A), but it was not clear whether this was simply due to an overall lower level of colonization (Fig. S8A). When mice were infected with either the C. rodentium wild type or an espH cesF map mutant, both strains were detected in association with the mucosal surface, although the espH cesF map mutant appeared impaired for colonization (Fig. S11A) and the C. rodentium wild type was recovered in higher numbers than a espH cesF map mutant from colon contents seven days after infection (Fig. 3A). Only infection with the C. rodentium wild type significantly reduced epithelial hypoxia, while the colonic surface remained hypoxic in mice infected with a espH cesF map mutant (Fig. S11B and S11C). While mature colonocytes at the luminal surface respire butyrate, undifferentiated colonocytes in the crypts exhibit the Warburg metabolism of dividing cells, which is characterized by fermenting glucose to lactate (30). C. rodentium infection induces epithelial regeneration and repair mechanisms, which drives a marked expansion of undifferentiated colonocytes, resulting in crypt elongation and the presence of undifferentiated colonocytes at the luminal surface (31, 32)(Fig. S1). Since the Warburg metabolism does not consume oxygen, we wanted to determine whether reduced hypoxia of surface colonocytes observed during C. rodentium infection (Fig. 2B, 2C, S11B and S11C) was due to an accumulation of undifferentiated colonocytes at the luminal surface. Interestingly, infection with a espH cesF map mutant triggered significantly (P < 0.05) less mitotic divisions in the colonic epithelium than infection with wild-type C. rodentium (Fig. S8C), suggesting that deletion of the espH, cesF and map genes might reduce crypt hyperplasia. To test this idea, we stained colonic sections from mock-infected mice or mice infected with the C. rodentium wild type or an espH cesF map mutant with Ki67, a cellular marker for proliferation. Infection with the C. rodentium wild type induced a marked expansion of Ki67-positive colonocytes, resulting in crypt elongation, a thickening of the colonic mucosa and an accumulation of Ki67-positive colonocytes along the length of the crypt and at the luminal surface (Fig. 3B and 3C). Interestingly, this host response was blunted in mice infected with a C. rodentium espH cesF map mutant (Fig. 3B and 3C), which correlated with increased epithelial hypoxia (Fig. S11B and S11C). We next wanted to determine whether colonic crypt hyperplasia was a driver of CydAB-dependent pathogen expansion. To this end, mice were mock infected or infected with a 1:1 mixture of the C. rodentium wild type and cydAB mutant and then treated with dibenzazepine (DBZ) or with vehicle control. DBZ is a Notch and Wnt (wingless-related integration site) signaling pathway inhibitor that prevents colonic crypt hyperplasia during C. rodentium infection (31). Ki67 staining suggested that mice infected with the C. rodentium strain mixture developed colonic crypt hyperplasia, which was significantly (P < 0.05) blunted by DBZ treatment (Fig. 3D and 3E), although DBZ treatment did not reduce the severity of other histopathological changes (Fig. S12A and S12B). Remarkably, inhibition of colonic crypt hyperplasia by DBZ treatment abrogated the fitness advantage conferred by the cydAB genes (Fig. 3F), suggesting that crypt hyperplasia was a main driver of aerobic pathogen expansion. Our finding that the LEE-encoded T3SS provides C. rodentium access to oxygen in vivo helps explain why the mechanism by which this trademark virulence factor confers a benefit to AE pathogens has not been apparent from in vitro studies, because such experiments are commonly performed in a 95% air (79% N2/21% O2) atmosphere, supplemented by 5% of carbon dioxide, thus providing 19.95% O2 (150 mmHg). This oxygenation is considerably higher than normal tissue oxygenation (23–70 mmHg or 3%–10% oxygen) (35) or oxygenation of colonocytes in vivo (< 7.6 mmHg or < 1% O2) (23, 24). While oxygen emanating from the mucosal surface is a very limited resource in the lumen of the large intestine (36), epithelial oxygenation was markedly elevated during T3SS-induced colonic crypt hyperplasia. As a result, T3SS-induced colonic crypt hyperplasia drove growth of C. rodentium through CydAB-mediated aerobic respiration, presumably because a respiratory metabolism enables the pathogen to utilize carbon sources more effectively than competing microbes that rely on fermentation for growth. In conclusion, our results revealed how histopathological changes triggered by a trademark virulence factor of AE pathogens create a unique nutrient-niche to fuel an uncontrolled luminal expansion of C. rodentium. Materials and Methods Bacterial strains and culture conditions For the Escherichia coli and Citrobacter rodentium strains used in this study, see Table S1. Strains were routinely grown in Luria-Burtani broth (BD Biosciences #244620) or on LB plates unless otherwise indicated. Antibiotics were used at the following concentrations: carbenicillin (Carb), 0.1 mg/ml; chloramphenicol (Cm), 0.015mg/ml; kanamycin (Kan), 0.1 mg/ml; tetracycline (Tet), 0.02 mg/ml. In vitro competitive growth assay Bacteria were grown in 10 ml M9 minimal media (2mM MgS04, 0.2mM CaCl2, M9 salts [Na2HPO4, 6.8 g/l; KH2PO4, 3 g/l; NaCl, 0.5 g/l; NH4Cl, 1 g/l]) with 0.4% glycerol, 1% casamino acids (BD Biosciences). Bacteria for the assay were grown overnight at 37°C in an anaerobe chamber (Shellab) with an environment of 5 % CO2, 5 % H2, and the remainder N2. Pre-reduced media was inoculated with a 1:1 ratio of two C. rodentium strains at a concentration of 1×105 total CFU in the anaerobe chamber. The inoculated media was then either left in the anaerobe chamber (0% oxygen) or moved to a hypoxia chamber in an airtight container. The hypoxia chamber was set at either 1 %, 4 %, 8 %, or 21 % oxygen, with the remaining atmosphere being composed of 5 % CO2 and 95 % N2. Cultures were left to incubate for 24 hours and then plated on selective MacConkey agar to differentiate between the two strains. Data represents eight replicates per group. Nitrate reductase assay Bacterial cultures were grown overnight in LB broth, then diluted 1:1000 into fresh LB broth containing 40 mM NaNO3. Cultures were grown statically at 37°C for 3 hours. The assay was then performed as described previously (37). Briefly, 50 μl of a washed cell suspension was permeabilized in 0.1% sodium dodecyl sulfate and chloroform. Next methyl viologen (0.5 mg/ml) and a solution of bicarbonate (8 mg/ml), sodium hydrosulfite (8 mg/ml), and nitrate (0.5 M) were added to cells and left to incubate at room temperature for 5 minutes. Assay was terminated with vigorous vortexing, then sulfanilic HCl (1% sulfanilic acid and 20% HCl) and Marshall’s reagent (0.13 % N-1-naphthylethylenediamine 2HCl) were added to the solution. Absorbance was next read at 540 nm and 420 nm. To calculate nitrate reductase activity, we used the following equation: {[A540 −(0.72 * A420)]/(t * v * A600)} * 100 where t is the time (minutes) of the reaction, v is the volume (ml) of the cell suspension in the assay, and A600 is the optical density of the original cell suspension. Construction of C rodentium mutants To construct a napA mutant, upstream and downstream regions of approximately 1kb in length flanking C. rodentium napA were amplified via PCR, run on an agarose gel, and purified using the QiaexII kit (Qiagen). These flanking regions were digested with XbaI (NEB), ligated, and PCR amplified to form one fragment of approximately 2kb that was then inserted into pCR2.1. The plasmid was propagated in Top10 E. coli. This plasmid was next digested with SphI (NEB) and the fragment was inserted into pRDH10 to form pCAL7 and cloned into E. coli DH5α λpir. To insert a tetracylin resistance cassette, pCAL7 was digested with XbaI while pSPN23 was digested with NheI (NEB). The tetracycline resistance cassette from pSPN23 was ligated with pCAL7 to form pCAL17. Plasmid pCAL17 was then transformed into E. coli strain S17-1 λpir and conjugation performed with C. rodentium that was transformed with pWSK129 to have a selectable marker for the recipient. After selection to pick C. rodentium colonies containing a single crossover mutation, sucrose selection was performed to select for colonies that lost the vector backbone along with the napA gene, but retained the tetracycline resistance cassette. After confirming the mutation via PCR, this strain was named CAL60. To generate a nonpolar napA mutant, CAL60 was conjugated with E. coli S17-1 λpir containing pCAL7. After selection for single crossover C. rodentium colonies, sucrose selection was performed to select for colonies that lost the tetracycline cassette (CAL77). To construct the narZ mutation, upstream and downstream regions of approximately 1kb in length flanking C. rodentium narZ were amplified via PCR, run on an agarose gel, and purified using the QiaexII kit. These flanking regions were digested with XbaI, ligated, and PCR amplified to form one fragment of approximately 2kb that was then inserted into pCR2.1 and propagated in E. coli Top10. This plasmid was next digested with SphI and the fragment was inserted into pRDH10 to form pCAL8, which was propagated in DH5α λpir E. coli. To insert a tetracycline resistance cassette, pCAL8 was digested with XbaI while pSPN23 was digested with NheI. The tetracycline resistance cassette was ligated with pCAL8 to form pCAL15. Plasmid pCAL15 was transformed into E. coli S17-1 λpir and conjugation performed with CAL77. After selection to pick C. rodentium colonies containing a single crossover mutation, sucrose selection was performed to select for colonies that lost the vector backbone and narZ gene, but retained the tetracycline resistance cassette. After confirming the mutation via PCR, the napA narZ mutant was named CAL92. To construct the narG mutation, a region within C. rodentium narG was amplified via PCR, inserted into pCR2.1, and propagated in Top10 E. coli. This plasmid and the plasmid pGP704 were then digested with XbaI and SphI. The narG fragment was ligated into pGP704 to produce the plasmid pCAL6 and propagated in E. coli DH5α λpir. Plasmid pCAL6 was then transformed into E. coli S17-1 λpir conjugation performed with CAL92. C. rodentium colonies with a single crossover were selected and PCR confirmed the mutation in narG. The napA narZ narG mutant was named CAL93. To construct a moaA mutant, a region within C. rodentium moaA was amplified via PCR, inserted into pCR2.1, and propagated in E. coli Top10. This plasmid and the plasmid pEP185.2 were then digested with XhoI (NEB) and SacI (NEB). The moaA fragment was ligated into pEP185.2 to generate the plasmid pCAL34, which was subsequently propagated in E. coli DH5α λpir. To aid in selection during conjugation, wild type C. rodentium was transformed with the temperature-sensitive plasmid pSW172. Plasmid pCAL34 was then introduced into E. coli S17-1 λpir and conjugation performed with C. rodentium (SW172) at 30°C. Since C. rodentium exhibits low-level resistance to carbenicillin naturally, a higher concentration of this antibiotic (0.2 mg/ml) was used to select for C. rodentium colonies that retained the CarbR plasmid pSW172 and also contained a single crossover event at the moaA locus. The plasmid pSW172 was cured after growth at 40°C and PCR was used to confirm the moaA mutation. The moaA mutant was named CAL142. To construct a non-polar cydAB mutant, upstream and downstream regions of approximately 1kb in length flanking the C. rodentium cydAB operon were amplified via PCR, run on an agarose gel, and purified using the QiaexII kit. These flanking regions were digested with XbaI, ligated, and PCR amplified to form one fragment of approximately 2kb that was then inserted into pCR2.1 and propagated in E. coli Top10. While cloning in C. rodentium, we observed that sucrose selection does not efficiently prevent the growth of Citrobacter colonies that retained a suicide vector encoding sacRB. To improve screening for double crossover events, we amplified the phoN gene of Salmonella enterica serovar Typhimurium via PCR and inserted it into pCR2.1 followed by propagation in E. coli Top10. This plasmid was subsequently digested with EcoRI (NEB) and phoN was inserted into the EcoRI-digested pRDH10 plasmid to generate pCAL52. C. rodentium colonies with this plasmid will turn blue when grown on plates containing X-phos (5-bromo-4-chloro-3-indolyl phosphate). The plasmid pCR2.1 containing the cydAB flanking regions was next digested with SalI (NEB) and the fragment was inserted into pCAL52 to generate pCAL59. To insert a kanamycin resistance cassette, pCAL59 and pUC4 KSAC were first digested with XbaI. The kanamycin resistance cassette from pUC4 KSAC was then ligated with pCAL59 to form pCAL60 and propagated in E. coli DH5α λpir. Plasmid pCAL60 was then transformed into E. coli strain S17-1 λpir and conjugation performed with C. rodentium (SW172). C. rodentium colonies containing a single crossover mutation were then grown overnight in LB broth at 30°C and plated on LB agar plates containing X-phos (40mg/l). White colonies were screened for resistance to kanamycin, but sensitivity to chloramphenicol, confirming loss of the pCAL52 vector backbone but retaining the kanamycin resistance cassette. After confirming the cydAB deletion via PCR, the plasmid pSW172 was cured after growth at 40°C and the strain named CAL247. To generate a non-polar cydAB mutant, CAL247 was conjugated with E. coli S17-1 λpir containing pCAL59. After selection for single crossover C. rodentium colonies, we performed blue-white screening on LB agar plates containing X-phos as described above to select for colonies that lost the kanamycin cassette and generated CAL267. To construct a non-polar fdnG mutant, upstream and downstream regions of approximately 1kb in length flanking the C. rodentium fdnG gene were amplified via PCR, run on an agarose gel, and purified using the QiaexII kit. These flanking regions were digested with XbaI, ligated, and PCR amplified to form one fragment of approximately 2kb in size that was then inserted into pCR2.1 and propagated in E. coli Top10. This plasmid was next digested with SphI and inserted into pCAL52 to generate pCAL55 and propagated in E. coli DH5α λpir. Plasmid pCAL55 was transformed into E. coli S17-1 λpir, conjugation performed with C. rodentium (SW172) and single crossover events were selected for using the appropriate antibiotics. Blue-white screening with X-phos was performed as described above. A double crossover event was confirmed via PCR. The plasmid pSW172 was cured after growth at 40°C and the non-polar fdnG mutant was named CAL210. To aid in selection during mouse experiments, CAL210 was transformed with pWSK129. To construct the fdoG mutant, upstream and downstream regions of approximately 1kb in length flanking the C. rodentium fdoG gene were amplified via PCR, run on an agarose gel, and purified using the QiaexII kit. These flanking regions were digested with XbaI, ligated, and PCR amplified to form one fragment of approximately 2kb that was then inserted into pCR2.1 and propagated in E. coli Top10. This plasmid, containing the fdoG flanking regions, was next digested with SphI and the fragment was inserted into pCAL52 to generate pCAL72. To insert a kanamycin resistance cassette, pCAL72 and pUC4 KSAC were first digested with XbaI. The kanamycin resistance cassette from pUC4 KSAC was then ligated with pCAL72 to form pCAL74 and propagated in E. coli DH5α λpir. Plasmid pCAL74 was then transformed into E. coli strain S17-1 λpir and conjugation performed with C. rodentium (SW172). Single crossover events were selected for by growing on the appropriate antibiotics. Blue-white screening with Xphos was performed as described above to identify a colony that lost the vector backbone and the fdoG gene, but retained the kanamycin resistance cassette. Plasmid pSW172 was cured after growth at 40°C. After PCR confirmation of the fdoG mutation, this strain was named CAL261. To construct a escN mutant, the Gibson Assembly method (New England Biolabs) was used to generate plasmid pCAL81 with a pEP185.2 vector backbone and an insert of an internal amplified region of escN. Plasmid pCAL81 was first propagated in E. coli DH5α λpir, then transformed into E. coli S17-1 λpir and conjugation performed with C. rodentium (SW172) at 30°C. After selection for single crossover events, the plasmid pSW172 was cured after growth at 40°C and PCR was used to confirm the escN mutation. This escN mutant was named CAL286. To construct a cydAB escN double mutant, the E. coli S17-1 λpir strain containing pCAL81 was conjugated with CAL247. After selection for single crossover events, PCR was used to confirm the escN mutation. The cydAB escN mutant was named CAL287. For generation of a cfcH mutant, a region within the C. rodentium cfcH was amplified via PCR. Fragments were run on an agarose gel and purified using the Zymoclean Gel DNA Recovery kit (Zymo Research). Creation of pBMM1 was accomplished by insertion of the purified fragment into pEP185.2 via the Gibson Assembly (New England Biolabs), and propagation in E. coli DH5α λpir. pBMM1 was then transformed into E. coli S17-1 λpir, followed by conjugation with C. rodentium (SW172). C. rodentium colonies with a single crossover were selected, and PCR was used to confirm the mutation in cfcH. The cfcH mutant was named BMM12. For construction of a cydAB cfcH double mutant, the E. coli S17-1 λpir strain containing pBMM1 was conjugated with a cydAB mutant. C. rodentium colonies carrying a single crossover were selected, and PCR was used to confirm the cfcH mutation. This cydAB cfcH double mutant was then named BMM11. To construct an eae mutant, primers were designed using NEBuilder Assembly Tool (New England Biolabs) to amplify an eae intragenic region with overlapping regions with the plasmid pEP185.2. The intragenic region was amplified using Q5 high-fidelity master mix (New England Biolabs) and ligated via Gibson Assembly with pEP185.2 after the vector was digested with SacI. The assembled plasmid, named pCAL82, was transformed into E. coli DH5α λpir then into E. coli S17-1 λpir. C. rodentium (SW172) or the ΔcydAB::KanR mutant (CAL247) were next conjugated with E. coli S17-1 λpir containing pCAL82. A C. rodentium colony resistant to Cm was verified by PCR to contain a single crossover at the eae locus. The strain was then cured of the temperature sensitive plasmid after growth at 40°C and designated CAL290. A C. rodentium ΔcydAB mutant resistant to Kan and Cm was verified by PCR to contain a single crossover at the eae locus in addition to the deletion of cydAB and this strain was named CAL291. To construct a non-polar espH cesF map mutant, regions of approximately 400bp in length flanking the espH, cesF, and map genes were amplified via PCR using Q5 high fidelity master mix, run on an agarose gel, and purified using Zymoclean Gel DNA Recovery kit (Zymo Research). Fragments were ligated into pRDH10 via Gibson Assembly after the vector was digested with SphI. The resulting plasmid pBMM3 was transformed into E. coli E. coli S17-1 λpir and conjugated with C. rodentium wild type or a cydAB mutant (CAL247), which had been transformed with the temperature-sensitive plasmid pSW172. Exconjugants were selected for loss of the plasmid using sucrose selection and screened via PCR to confirm deletion of the target genes. The strains were then cured of the temperature sensitive plasmid after growth overnight at 42°C and were named BMM30 and BMM40, respectively. Animal experiments The Institutional Animal Care and Use Committee at the University of California, Davis, approved all animal experiments. For experiments involving mice colonized with conventional specific pathogen free microbiota, either C57BL/6J or C3H/HeJ strains (Jackson Laboratories) aged 6–8 weeks were inoculated with 1×109 colony forming units (CFU)/mouse of a single C. rodentium strain in LB broth intragastrically. In competitive infections, a 1:1 ratio of two C. rodentium strains were given intragastrically at a combined final concentration of 1×109 CFU/mouse. Fecal pellets were collected on days 3, 7, or 10 and mice were sacrificed between 3 and 14 days after infection. Colon contents were stored in phosphate-buffered saline (PBS) on ice. Colon tissue for histopathology was fixed in 10% buffered formalin phosphate while colon sections for murine RNA analysis were flash frozen and stored at −80°C. C. rodentium from feces and colon contents were enumerated by plating serial ten-fold dilutions of samples on LB agar or MacConkey agar (BD Biosciences) containing the appropriate antibiotics based on the resistance markers listed for each strain in Table S1. Plates were incubated overnight at 37°C under atmospheric oxygen conditions. Germ-free Swiss Webster mice were bred and housed at the Teaching and Research Animal Care Services facilities at UC Davis. These mice were infected and euthanized as described above. To inhibit hyperplasia, mice were first mock infected or infected with a 1:1 mixture of the indicated C. rodentium strains as described above. Dibenzazepine (DBZ) was initially dissolved in DMSO and then suspended in a solution of hydroxypropyl methylcellulose (0.5 %)(Sigma) and Tween80 (0.1 %). Mice were given 0.1ml of the DBZ solution or DMSO vehicle control starting at day 2 after infection (approximately 10 μmol/kg) and continuing daily up to day 6 post-infection, or a total of 5 doses. On day 7 after infection, mice were given pimonidazole (PMDZ) one hour before euthanasia was performed. Microbiota analysis Seven days after treatment of mice, DNA from the colon content was extracted using the PowerSoil DNA Isolation kit (Mo-Bio, Carlsbad, CA) according to the manufacturer’s protocol. Bacterial DNA was amplified via a PCR enrichment of the 16S rDNA (V4 region) using primers 515F and 806R modified by addition of barcodes for multiplexing. Libraries were sequenced with an Illumina MiSeq system. Sample sequences were demultiplexed and trimmed, followed by filtering for quality. QIIME open-source software (http://qiime.org) (38) was used for initial identification of operational taxonomic units (OTU), clustering, and phylogenetic analysis. Samples containing less than 1000 quality reads were removed from dataset. Subsequent data transformation was performed using MEGAN 4 software (http://ab.inf.uni-tuebingen.de/software/megan4/) (39) and then further analyzed using Explicet version 2.10.5 (http://www.explicet.org/) (40). Data analyzed using METAGENassist (http://www.metagenassist.ca/) was first normalized by logarithmically transforming the reads to yield a more normal distribution (41). Histopathology Fixed colon sections were stained with hematoxylin and eosin. A veterinary pathologist scored histopathological changes of blinded samples using the scoring scheme described in Figure S4A. For mitotic divisions, the data represent the number of mitotic figures per high power field (40X), with the average of five fields per slide displayed. RNA isolation For murine RNA isolation, colon tissue sections were homogenized in a Mini-Beadbeater (BioSpec Products, Bartlesville, OK) and RNA was isolated by the TRI-Reagent method (Molecular Research Center, Inc.) following the manufacturer’s protocol. Contaminating DNA was removed using the DNA-free kit (Applied Biosystems) and RNA was stored at −80°C. Upon C. rodentium infection of mice, bacterial RNA was isolated from two compartments: the intestinal colon contents and mucus layer. During necropsy, a section of the middle colon containing a fecal pellet was removed and cut longitudinally. The fecal pellet was carefully removed to avoid pellet disintegration and stored in RNAlater (Ambion) overnight at 4°C. Forceps were next used to gently scrape off the mucus layer from the underlying tissue. The mucus layer was also stored in RNAlater overnight at 4°C. Sample isolation was performed at atmospheric oxygen tensions. Excess solution was subsequently removed from the pellet and Trizol (Invitrogen) was added. Next chloroform was added and the solution was transferred to 2 ml phase-lock tubes (5 Prime) and centrifuged at 14000g. The upper phase was removed and RNA was precipitated with 100% ethanol at −20°C overnight. RNA was collected after centrifugation and washed with 70% ethanol. Contaminating DNA was removed as described above. For bacterial RNA isolation from cultures, 1 ml of bacterial culture was processed using the Aurum Total RNA mini kit (BioRad) following the manufacturer’s protocol. Contaminating DNA was removed as described above. Quantitative real-time PCR Isolated RNA was reverse transcribed using random hexamers and Moloney murine leukemia virus reverse transcriptase (Applied Biosystems). Quantitative real-time PCR was performed using SYBR green (Applied Biosystems) PCR mix and the appropriate primer sets (see Table S2) at a final concentration of 0.25 mM. Absolute values were calculated using a series of standards and a standard curve. To generate standards, the above mentioned primer sets were used in standard PCR reaction to amplify the target genes, which were subsequently inserted into pCR2.1 using the TOPO cloning kit (Life Technologies). The resulting plasmids were sequenced for accuracy and the concentration quantified to create a set of standards ranging from 108 to 101 copies/μl diluted in a 0.02 mg/ml yeast RNA (Sigma) solution. To quantify transcription of the regulator of the LEE pathogenicity island, ler, RNA was isolated from the colon contents (see above) and qRT-PCR was applied with a published primer pair (42). To quantify transcription of sucA expression in vivo, RNA was isolated from both the mucus layer and colon contents (see above) and qRT-PCR was applied using the primers in Table S2. Flow cytometry Flow cytometry analysis was performed for detection of PMDZ staining on whole colon epithelial cells isolated from mock, C. rodentium-infected and C. rodentium escN mutant-infected mice. Single cell suspensions of colonocytes were obtained as described previously (43). Cells were suspended in 2 ml of dPBS and stained with Aqua Live/Dead cell discriminator (Invitrogen) according to the manufacturer’s protocol. After Live/Dead staining, cells were washed with dPBS and re-suspended in 100 μl of PBS containing 3.6% of M.O.M.™ mouse Ig blocking reagent (Vector Laboratories) and incubated in the dark at 4°C for 20 min. Cells were washed with 2 ml of dPBS containing 1% bovine serum albumin and 2 mM EDTA (fluorescence-activated cell sorter [FACS] buffer) and then re-suspended in 50 μl of FACS. Cells were stained with a FC receptor blocking antibody, anti-CD16/32 (Biolegend), for 5 minutes at 4°C and then stained for 30 min at 4°C with a cocktail of anti-CD45 PerCp-Cy5.5 (Biolegend) and anti-EpCam FITC (Biolegend). Cells were washed in FACS buffer, fixed and permeabilized with BD Cytofix/Cytoperm™ (BD Biosciences) for 20 minutes at 4°C and then washed with BD Perm/Wash buffer™ (BD Biosciences). Cells were re-suspended in 100 μl of BD Perm/Wash buffer™ containing 3.6% of M.O.M.™ mouse Ig blocking reagent and incubated in the dark at 4°C for 20 min. Cells were washed with BD Perm/Wash buffer™ and then stained with an anti-PMDZ PE antibody (Hypoxyprobe), diluted in BD Perm/Wash buffer™ for 30 min at 4°C. Cells were washed in BD Perm/Wash buffer™ and re-suspended in FACS. For quantification of cell populations, 50 μl of SPHERO AccuCount Fluorescent Particles 10.1 μm (Spherotech) were added to each sample prior to analysis. Calculation of absolute counts was performed according to manufacturer’s protocol. Flow cytometry analysis was performed using a BD (Becton Dickinson) LSRII, and 4.0 × 105 events were collected per mouse. Data were analyzed using FlowJo software (Treestar, inc. Ashland, OR) and gates were based on fluorescence-minus-one (FMO) controls. Fluorescent Imaging For mouse colon imaging, 10–15 mm sections of mid-colon were frozen in Optimal Cutting Temperature (OCT) compound (Fisher HealthCare) and cut in transverse sections to a thickness of 7μm. Sections were fixed with 4% PFA and permeabilized with 0.2% Triton-X 100. C. rodentium cells were stained with a rabbit anti-Citrobacter primary antibody (Abcam, ab37056) and a goat anti-rabbit Alexafluor 647conjugate (Abcam, ab150079). Proliferating cells were visualized with a primary rabbit antibody targeting the nuclear protein Ki67 (Abcam, ab15580) and a goat anti-rabbit Alexafluor 647 conjugate. Nuclei were stained with DAPI in SlowFade mounting media (Invitrogen). Actin was stained using phalloidin-tetramethylrhodamine B isothiocyanate (Sigma Aldrich). For imaging of intestinal hypoxia, at the day of euthanasia mice were treated intraperitoneally with 20 mg per mouse of pimonidazole HCl (Hypoxyprobe). One hour after treatment, mice were euthanized and colon sections were frozen and cryosectioned as stated above. These PMDZ adducts were detected with a mouse monoclonal antibody MAb1 (Hypoxyprobe) followed by a secondary goat anti-mouse IgG antibody (Alexafluor 546, ThermoFisher). After fixation and permeabilization sections were incubated first with Universal Block (KPL) then with an IgG blocking reagent from a Mouse on Mouse (M.O.M) kit (Vector Laboratories). Sections were finally incubated with the anti-PMDZ primary antibody and the secondary anti-mouse Alexafluor 546 antibody. PMDZ and Ki67 staining was quantified using ImageJ 1.49v software (44). For PMDZ quantification, staining intensity of the outer colonocytes was normalized to the staining intensity of DAPI. This ratio was then expressed as fold change over the mock control group. For Ki67 quantification, staining intensity of all the cells within a colonic crypt was normalized to the staining intensity of DAPI. This ratio was then expressed as fold change over the mock control group. For both PMDZ and Ki67 quantification, three to four non-overlapping images were used per mouse. In vitro bacterial RNA expression To measure gene expression in vitro, wild type C. rodentium was grown in M9 media supplemented with 1% casamino acids and 0.4% mannose for 16 hours at 37°C under either microaerobic or anaerobic conditions. A microaerobic environment was created by growing bacteria without shaking under environmental oxygen tension while anaerobiosis was achieved by growing bacteria in pre-reduced media in an anaerobe chamber. Experiments were performed in triplicate. RNA was then isolated as described above. Statistical analysis Data for our experiments displayed as bar graphs represent the geometric mean and the standard error of the mean. For most experiments, data points were first log transformed and differences between experimental groups were determined on the transformed data using a Student’s T-test (for comparing two groups) or ANOVA followed by Fisher’s LSD post hoc test (for comparison of more than two groups). If appropriate, outliers were removed using Grubb’s test. Data representing the percent of PMDZ positive colonocytes were first arc-sine transformed before applying the ANOVA test. A one-way ANOVA using METAGENassist online software was used to determine statistical significance on the changes in the proportion of microbial taxa from the16S sequencing reads. A P-value of less than 0.05 was considered significant. Supplementary Material We would like to acknowledge the Host-Microbe Systems Biology Core (HMSB Core) at the UC Davis School of Medicine for expert technical assistance with microbiota sequence analysis. The data reported in the manuscript are tabulated in the main paper and in the supplementary materials. Work in A.J.B.’s laboratory was supported by Public Health Service Grants AI044170, AI096528, AI107393, and AI112949. Work in R.M.T.’s lab was supported by Public Health Service Grant AI098078. Work in S.E.W.’s lab was supported by Public Health Service Grant AI103248. C.A.L was supported by Public Health Service Grant AI112241. E.M.V. was supported by Public Health Service Grant OD010931. Figure 1 Oxygen respiration supports C. rodentium expansion in the mouse colon (A) C57BL/6 (C57) mice were infected with C. rodentium wild type (wt, DBS100) and either a moaA mutant (CAL142) or a napA narG narZ mutant (CAL93). (B) Conventional C57 or germ-free Swiss Webster (SW) mice were infected with wt C. rodentium and either a moaA mutant, a fdnG mutant (CAL210 [pWSK129]), or a fdoG mutant (CAL261). N is indicated in Fig. S3E. (C) Competitive in vitro growth (N =8) of C. rodentium wild type (wt) and a cydAB mutant (CAL247) for 16 hours in minimal medium in the presence of the indicated oxygen levels (% O2). (D) Conventional or germ-free mice were infected with an equal mixture of the C. rodentium wild type (wt) and a cydAB mutant. (A and D) N = 4. (E) C. rodentium was grown in minimal medium supplemented with mannose as a carbon source under either microaerobic or anaerobic conditions. (F) Bacterial RNA was isolated from either mucus scrapings or colon contents of C. rodentium-infected mice. (E–F) The transcript levels of sucA were quantified by real-time PCR, normalized to 16S rRNA levels and shown as fold-changes. N is shown in Fig. S4B and S4C. (G–I) Mice (N indicated in I) were either mock-treated, infected with the C. rodentium wild type (wt) or with a cydAB mutant (CAL247). (G) Enumeration of C. rodentium by plating on selective media. (A–G) Bars represent geometric mean ± standard error. (H) Boxes in Whisker plots represent the second and third quartiles of combined histopathology scores, while lines indicate the first and fourth quartiles. (I) Microbial representation at the class-level based on 16S rRNA gene sequencing of colon contents 7 days after infection. Color-coding for classes is shown on the right. *, P < 0.05; **, P < 0.01; ns, not statistically significantly different. Figure 2 A functional T3SS increases epithelial oxygenation and drives a CydAB-dependent C. rodentium expansion (A) Mice (N is indicated in Fig. S8A) were infected with the indicated C. rodentium strain mixtures. Bars represent geometric means of the competitive index (CI) ± standard error. (B) Representative images of colonic sections stained to detect pimonidazole hypoxia stain (red fluorescence) and counterstained with DAPI nuclear stain (blue fluorescence). Arrow heads point to the mucosal surface. E, epithelial cell; L, lumen. (C) Pimonidazole staining (PMDZ+) was quantified by flow cytometry in colonic epithelial (CD45− CD326+) cells. White bars indicate the level of background staining observed in mice that were not injected with PMDZ. Bars represent geometric mean ± standard error. **, P < 0.01; *, P < 0.05; ns, not statistically significantly different; wt, DBS100; cydAB, CAL247; cfcH, BMM12; cfcH cydAB, BMM11; eae, CAL290; eae cydAB, CAL291; escN, CAL286; escN cydAB, CAL287; espH cesF map, BMM30; espH cesF map cydAB, BMM40. Figure 3 Colonic crypt hyperplasia drives a CydAB-dependent C. rodentium expansion (A–C) C57BL/6 mice were mock infected or infected with the C. rodentium wild type (wt, DBS100) or an espH cesF map mutant (BMM30). (D–F) C57BL/6 mice were mock infected or infected with a 1:1 mixture of the C. rodentium wild type (wt) and a cydAB mutant (CAL247). Mice were either treated with dibenzazepine (DBZ) or with vehicle control. (A–F) Organs were collected seven days after infection. (A) CFU recovered from colon contents. (B and D) Representative images of colonic sections were stained to detect Ki67 (yellow fluorescence) and counterstained with DAPI nuclear stain (blue fluorescence). (C and E) Ki67 staining was quantified by image analysis. (F) Enumeration of wt and cydAB mutant recovered from colon contents was used to calculate the competitive index (CI). (A, C, E and F) Bars represent geometric mean ± standard error. **, P < 0.01; *, P < 0.05; ns, not statistically significantly different. References and notes 1 Staley TE Jones EW Corley LD Attachment and penetration of Escherichia coli into intestinal epithelium of the ileum in newborn pigs The American journal of pathology 56 371 392 1969 4898367 2 Moon HW Whipp SC Argenzio RA Levine MM Giannella RA Attaching and effacing activities of rabbit and human enteropathogenic Escherichia coli in pig and rabbit intestines Infection and immunity 41 1340 1351 1983 6350186 3 Jarvis KG Enteropathogenic Escherichia coli contains a putative type III secretion system necessary for the export of proteins involved in attaching and effacing lesion formation Proceedings of the National Academy of Sciences of the United States of America 92 7996 8000 1995 7644527 4 Jerse AE Yu J Tall BD Kaper JB A genetic locus of enteropathogenic Escherichia coli necessary for the production of attaching and effacing lesions on tissue culture cells Proceedings of the National Academy of Sciences of the United States of America 87 7839 7843 1990 2172966 5 Kenny B Enteropathogenic E. coli (EPEC) transfers its receptor for intimate adherence into mammalian cells Cell 91 511 520 1997 9390560 6 Deng W Vallance BA Li Y Puente JL Finlay BB Citrobacter rodentium translocated intimin receptor (Tir) is an essential virulence factor needed for actin condensation, intestinal colonization and colonic hyperplasia in mice Mol Microbiol 48 95 115 2003 12657048 7 Nataro JP Kaper JB Diarrheagenic Escherichia coli Clin Microbiol Rev 11 142 201 1998 9457432 8 Luperchio SA Schauer DB Molecular pathogenesis of Citrobacter rodentium and transmissible murine colonic hyperplasia Microbes Infect 3 333 340 2001 11334751 9 Lupp C Host-mediated inflammation disrupts the intestinal microbiota and promotes the overgrowth of Enterobacteriaceae Cell Host Microbe 2 204 2007 18030708 10 Kamada N Regulated virulence controls the ability of a pathogen to compete with the gut microbiota Science 336 1325 1329 2012 22582016 11 Lopez CA Phage-mediated acquisition of a type III secreted effector protein boosts growth of salmonella by nitrate respiration MBio 3 2012 12 Winter SE Host-derived nitrate boosts growth of E. coli in the inflamed gut Science 339 708 711 2013 23393266 13 Winter SE Gut inflammation provides a respiratory electron acceptor for Salmonella Nature 467 426 429 2010 20864996 14 Rivers SL McNairn E Blasco F Giordano G Boxer DH Molecular genetic analysis of the moa operon of Escherichia coli K-12 required for molybdenum cofactor biosynthesis Molecular microbiology 8 1071 1081 1993 8361352 15 Berg BL Stewart V Structural genes for nitrate-inducible formate dehydrogenase in Escherichia coli K-12 Genetics 125 691 702 1990 2168848 16 Sawers G Heider J Zehelein E Bock A Expression and operon structure of the sel genes of Escherichia coli and identification of a third selenium-containing formate dehydrogenase isoenzyme Journal of bacteriology 173 4983 4993 1991 1650339 17 Amarasingham CR Davis BD Regulation of alpha-ketoglutarate dehydrogenase formation in Escherichia coli J Biol Chem 240 3664 3668 1965 5319784 18 Rivera-Chavez F Depletion of Butyrate-Producing Clostridia from the Gut Microbiota Drives an Aerobic Luminal Expansion of Salmonella Cell Host Microbe 19 443 454 2016 27078066 19 Albenberg L Correlation between intraluminal oxygen gradient and radial partitioning of intestinal microbiota Gastroenterology 147 1055 1063 e1058 2014 25046162 20 Mundy R Identification of a novel type IV pilus gene cluster required for gastrointestinal colonization of Citrobacter rodentium Molecular microbiology 48 795 809 2003 12694622 21 Lomax RB McNicholas CM Lombes M Sandle GI Aldosterone-induced apical Na+ and K+ conductances are located predominantly in surface cells in rat distal colon Am J Physiol 266 G71 82 1994 8304460 22 Velazquez OC Lederer HM Rombeau JL Butyrate and the colonocyte. Production, absorption, metabolism, and therapeutic implications Adv Exp Med Biol 427 123 134 1997 9361838 23 Furuta GT Hypoxia-inducible factor 1-dependent induction of intestinal trefoil factor protects barrier function during hypoxia J Exp Med 193 1027 1034 2001 11342587 24 Kelly CJ Crosstalk between Microbiota-Derived Short-Chain Fatty Acids and Intestinal Epithelial HIF Augments Tissue Barrier Function Cell Host Microbe 17 662 671 2015 25865369 25 Terada N Ohno N Saitoh S Ohno S Immunohistochemical detection of hypoxia in mouse liver tissues treated with pimonidazole using “in vivo cryotechnique” Histochem Cell Biol 128 253 261 2007 17680263 26 Collins JW Citrobacter rodentium: infection, inflammation and the microbiota Nat Rev Microbiol 12 612 623 2014 25088150 27 Wong AR Clements A Raymond B Crepin VF Frankel G The interplay between the Escherichia coli Rho guanine nucleotide exchange factor effectors and the mammalian RhoGEF inhibitor EspH MBio 3 2012 28 Ma C Citrobacter rodentium infection causes both mitochondrial dysfunction and intestinal epithelial barrier disruption in vivo: role of mitochondrial associated protein (Map) Cell Microbiol 8 1669 1686 2006 16759225 29 Nougayrede JP Donnenberg MS Enteropathogenic Escherichia coli EspF is targeted to mitochondria and is required to initiate the mitochondrial death pathway Cell Microbiol 6 1097 1111 2004 15469437 30 Fan YY A bioassay to measure energy metabolism in mouse colonic crypts, organoids, and sorted stem cells American journal of physiology Gastrointestinal and liver physiology 309 G1 9 2015 25977509 31 Ahmed I Critical roles of Notch and Wnt/beta-catenin pathways in the regulation of hyperplasia and/or colitis in response to bacterial infection Infect Immun 80 3107 3121 2012 22710872 32 Chandrakesan P Novel changes in NF-{kappa}B activity during progression and regression phases of hyperplasia: role of MEK, ERK, and p38 J Biol Chem 285 33485 33498 2010 20710027 33 Caron E Subversion of actin dynamics by EPEC and EHEC Curr Opin Microbiol 9 40 45 2006 16406772 34 Frankel G Enteropathogenic and enterohaemorrhagic Escherichia coli: more subversive elements Molecular microbiology 30 911 921 1998 9988469 35 Carreau A El Hafny-Rahbi B Matejuk A Grillon C Kieda C Why is the partial oxygen pressure of human tissues a crucial parameter? Small molecules and hypoxia J Cell Mol Med 15 1239 1253 2011 21251211 36 Espey MG Role of oxygen gradients in shaping redox relationships between the human intestine and its microbiota Free radical biology & medicine 55 130 140 2013 23127782 37 Stewart V Parales J Jr Identification and expression of genes narL and narX of the nar (nitrate reductase) locus in Escherichia coli K-12 Journal of bacteriology 170 1589 1597 1988 2832370 38 Caporaso JG QIIME allows analysis of high-throughput community sequencing data Nat Methods 7 335 336 2010 20383131 39 Huson DH Mitra S Ruscheweyh HJ Weber N Schuster SC Integrative analysis of environmental sequences using MEGAN4 Genome Res 21 1552 1560 2011 21690186 40 Robertson CE Explicet: graphical user interface software for metadata-driven management, analysis and visualization of microbiome data Bioinformatics 29 3100 3101 2013 24021386 41 Arndt D METAGENassist: a comprehensive web server for comparative metagenomics Nucleic Acids Res 40 W88 95 2012 22645318 42 Kamada N Humoral Immunity in the Gut Selectively Targets Phenotypically Virulent Attaching-and-Effacing Bacteria for Intraluminal Elimination Cell Host Microbe 17 617 627 2015 25936799 43 Godinez I Interleukin-23 orchestrates mucosal responses to Salmonella enterica serotype Typhimurium in the intestine Infect Immun 77 387 398 2009 18955477 44 Schneider CA Rasband WS Eliceiri KW NIH Image to ImageJ: 25 years of image analysis Nat Methods 9 671 675 2012 22930834 45 Barthold SW Coleman GL Bhatt PN Osbaldiston GW Jonas AM The etiology of transmissible murine colonic hyperplasia Lab Anim Sci 26 889 894 1976 1018473 46 Pal D Venkova-Canova T Srivastava P Chattoraj DK Multipartite regulation of rctB, the replication initiator gene of Vibrio cholerae chromosome II Journal of bacteriology 187 7167 7175 2005 16237000 47 Simon R Priefer U Puhler A A Broad Host Range Mobilization System for Invivo Genetic-Engineering - Transposon Mutagenesis in Gram-Negative Bacteria Bio-Technology 1 784 791 1983 48 Wang RF Kushner SR Construction of versatile low-copy-number vectors for cloning, sequencing and gene expression in Escherichia coli Gene 100 195 199 1991 2055470 49 Miller VL Mekalanos JJ A novel suicide vector and its use in construction of insertion mutations: osmoregulation of outer membrane proteins and virulence determinants in Vibrio cholerae requires toxR Journal of bacteriology 170 2575 2583 1988 2836362 50 Kingsley RA Ferrioxamine-mediated Iron(III) utilization by Salmonella enterica Appl Environ Microbiol 65 1610 1618 1999 10103258 51 Kinder SA Badger JL Bryant GO Pepe JC Miller VL Cloning of the YenI restriction endonuclease and methyltransferase from Yersinia enterocolitica serotype O8 and construction of a transformable R-M+ mutant Gene 136 271 275 1993 8294016 52 Raffatellu M Lipocalin-2 resistance confers an advantage to Salmonella enterica serotype Typhimurium for growth and survival in the inflamed intestine Cell Host Microbe 5 476 486 2009 19454351

PMC005xxxxxx/PMC5129607.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 101561091 39170 Epidemiol Psychiatr Sci Epidemiol Psychiatr Sci Epidemiology and psychiatric sciences 2045-7960 2045-7979 25720357 5129607 10.1017/S2045796015000189 NIHMS830479 Article Anxious and non-anxious major depressive disorder in the World Health Organization World Mental Health Surveys Kessler R.C. Ph.D. 1 Sampson N.A. B.A. 1 Berglund P. M.B.A. 2 Gruber M.J. M.S. 1 Al-Hamzawi A. MBCh.B, MD, FICMS 3 Andrade L. M.D., Ph.D. 4 Bunting B. Ph.D. 5 Demyttenaere K. M.D., Ph.D. 6 Florescu S. M.D., Ph.D. 7 de Girolamo G. M.D. 8 Gureje O. M.D., Ph.D., DS.C. 9 He Y. M.D. 10 Hu C. M.D., Ph.D. 11 Huang Y. MD, M.P.H, Ph.D. 12 Karam E. M.D. 13 Kovess-Masfety V. M.Sc., M.D., Ph.D. 14 Lee S M.B., B.S., FRCPsych 15 Levinson D. Ph.D. 16 Mora M.E. Medina Ph.D. 17 Moskalewicz J. Ph.D. 18 Nakamura Y. M.D., MPH, FFPH 19 Navarro-Mateu F. M.D., Ph.D. 20 Oakley Browne Mark A. F.R.A.N.Z.C.P., Ph.D. 21 Piazza M. ScD, M.P.H. 22 Posada-Villa J. M.D. 23 Slade T. Ph.D. 24 ten Have M. Ph.D. 25 Torres Y. M.PH, Dra.HC. 26 Vilagut G. M.Sc 27 Xavier M. M.D., Ph.D 28 Zarkov Z. M.D. 29 Shahly V. Ph.D. 1 Wilcox M.A. Ed.D., Sc.D. 30 1 Department of Health Care Policy, Harvard Medical School, Boston, MA, USA 2 Institute for Social Research, University of Michigan, Ann Arbor, MI, USA 3 College of Medicine, Al-Qadisiya University, Diwania governorate, Iraq 4 Institute of Psychiatry, University of São Paulo Medical School, São Paulo, Brazil 5 School of Psychology, Ulster University 6 Department of Psychiatry, University Hospital Gasthuisberg, Katholieke Universiteit Leuven, Leuven, Belgium 7 National School of Public Health, Management and Professional Development, Bucharest, Romania 8 Unit of Epidemiological and Evaluation Psychiatry, IRCCS St John of God Clinical Research Centre, Brescia, Italy 9 Centre for Research and Training in Mental Health, Neurosciences, Drug and Alcohol Abuse, Department of Psychiatry, University of Ibadan, Ibadan, Nigeria 10 Shanghai Mental Health Center, Shanghai Jiao Tong University, School of Medicine, Shanghai, PRC 11 Shenzhen Institute of Mental Health & Shenzhen Kanging Hospital, Guangdon Province, PRC 12 Institute of Mental Health, Peking University, Beijing, China 13 Department of Psychiatry and Clinical Psychology, Faculty of Medicine, Balamand University, Beirut, Lebanon; Department of Psychiatry and Clinical Psychology, St George Hospital University Medical Center, Beirut, Lebanon; Institute for Development Research Advocacy and Applied Care (IDRAAC), Beirut, Lebanon 14 Ecole des Hautes Etudes en Santé Publique (EHESP), EA 4057 Paris Descartes University,Paris, France 15 Department of Psychiatry, The Chinese University of Hong Kong 16 Ministry of Health Israel, Mental Health Services 17 Ramond e la Fuente Muñiz National Institute of Psychiatry, Mexico 18 Institute of psychiatry and Neurology, Warsaw, Poland 19 Department of Public Health, Jichi Medical University, Shimotsuke, Tochigi, Japan 20 Unidad de Docencia, Investigación y Formación en Salud Mental (UDIF-SM). Servicio Murciano de Salud. IMIB-Arrixaca. CIBERESP-Nodo Murcia, Spain 21 Department of Psychiatry, School of Medicine, University of Tasmania, Australia 22 National Institute of Health, Peru 23 Universidad Colegio Mayor de Cundinamarca, Bogota, Colombia 24 National Drug and Alcohol Research Centre, University of New South Wales, Sydney, Australia 25 Netherlands Institute of Mental Health and Addiction, Utrecht, The Netherlands 26 Center for Excellence on Research in Mental Health, CES University, Medellin, Colombia 27 Health Services Research Unit, IMIM (Institut Hospital del Mar d’Investigacions Mèdiques), Carrer del Doctor Aiguader, 88, Edifici PRBB, 08003, Barcelona, Spain 28 Department of Mental Health – CEDOC and Faculdade Ciencias Medicas, Universidade Nova de Lisboa, Lisbon, Portugal 29 Department Mental Health, NCPHA, Bulgaria 30 Janssen Pharmaceutical Research & Development, Titusville, NJ, USA Correspondence: Ronald C. Kessler, Department of Health Care Policy, Harvard Medical School, 180 Longwood Avenue, Boston, MA 02115. Tel. (617) 432-3587, Fax (617) 432-3588, kessler@hcp.med.harvard.edu 17 11 2016 27 2 2015 6 2015 30 11 2016 24 3 210226 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. AIMS To examine cross-national patterns and correlates of lifetime and 12-month comorbid DSM-IV anxiety disorders among people with lifetime and 12-month DSM-IV major depressive disorder (MDD). METHODS Nationally or regionally representative epidemiological interviews were administered to 74,045 adults in 27 surveys across 24 countries in the WHO World Mental Health (WMH) Surveys. DSM-IV MDD, a wide range of comorbid DSM-IV anxiety disorders, and a number of correlates were assessed with the WHO Composite International Diagnostic Interview (CIDI). RESULTS 45.7% of respondents with lifetime MDD (32.0–46.5% inter-quartile range [IQR] across surveys) had one of more lifetime anxiety disorders. A slightly higher proportion of respondents with 12-month MDD had lifetime anxiety disorders (51.7%, 37.8–54.0% IQR) and only slightly lower proportions of respondents with 12-month MDD had 12-month anxiety disorders (41.6%, 29.9–47.2% IQR). Two-thirds (68%) of respondents with lifetime comorbid anxiety disorders and MDD reported an earlier age-of-onset of their first anxiety disorder than their MDD, while 13.5% reported an earlier age-of-onset of MDD and the remaining 18.5% reported the same age-of-onset of both disorders. Women and previously married people had consistently elevated rates of lifetime and 12-month MDD as well as comorbid anxiety disorders. Consistently higher proportions of respondents with 12-month anxious than non-anxious MDD reported severe role impairment (64.4% vs. 46.0%; χ21=187.0, p<.001) and suicide ideation (19.5% vs. 8.9%; χ21=71.6, p<.001). Significantly more respondents with 12-month anxious than non-anxious MDD received treatment for their depression in the 12 months before interview, but this difference was more pronounced in high income countries (68.8% vs. 45.4%; χ21=108.8, p<.001) than low/middle income countries (30.3% vs. 20.6%; χ21=11.7, p<.001). CONCLUSIONS Patterns and correlates of comorbid DSM-IV anxiety disorders among people with DSM-IV MDD are similar across WMH countries. The narrow IQR of the proportion of respondents with temporally prior AOO of anxiety disorders than comorbid MDD (69.6–74.7%) is especially noteworthy. However, the fact that these proportions are not higher among respondents with 12-month than lifetime comorbidity means that temporal priority between lifetime anxiety disorders and MDD is not related to MDD persistence among people with anxious MDD. This, in turn, raises complex questions about the relative importance of temporally primary anxiety disorders as risk markers versus causal risk factors for subsequent MDD onset and persistence, including the possibility that anxiety disorders might primarily be risk markers for MDD onset and causal risk factors for MDD persistence. anxious depression comorbidity epidemiology INTRODUCTION The Global Burden of Disease (GBD) 2010 Study ranked major depressive disorder (MDD) as the 2nd leading cause of years lived with disability in the world (exceeded only by low back pain) and the 1st–4th leading cause (out of nearly 300 considered) in each region of the world (Vos et al. 2012). These high estimates are due to MDD having both high prevalence (estimated by the GBD 2010 investigators to be the 19th most common disease in the world) and high severity (indicated by higher ranking of MDD as a cause of disability than prevalent disease). Yet MDD severity is highly variable (Birnbaum et al. 2010; Li et al. 2014). Indeed, severity is the most consistent discriminating characteristic in empirical studies of MDD symptom subtypes (van Loo et al. 2012). One of the strongest predictors of MDD severity is comorbid anxiety disorder (Mineka and Vrshek-Schallhorn, 2008; Wu and Fang, 2014). Epidemiological studies show consistently that MDD is highly comorbid with numerous anxiety disorders (Andrade et al. 2003; Kessler et al. 2011b; Lamers et al. 2011) and more severe and persistent when accompanied by comorbid anxiety disorders (Ormel et al. 1994; Roy-Byrne et al. 2000; McLaughlin et al. 2006; Fichter et al. 2010). People with anxious MDD are also significantly more likely to seek treatment (Kessler et al. 2001; Jacobi et al. 2004) but significantly less likely to respond to treatment (Jakubovski and Bloch, 2014; Saveanu et al. 2014) than those with non-anxious MDD. Comorbid anxiety disorders have been found consistently to have earlier age-of-onset (AOO) than MDD both in cross-sectional surveys that assess AOO retrospectively (Kessler, 1995; Kessler et al. 2011a) and prospective studies that examine unfolding of comorbidity over time (Murphy et al. 1986; Bittner et al. 2004; Copeland et al. 2009; Klein et al. 2013). Two noteworthy limitations of existing research on comorbid anxiety in MDD are that a narrow definition of comorbid anxiety is often used that either focuses on current (but not lifetime) comorbidity or examines only one anxiety disorder (typically generalized anxiety disorder or panic disorder) and that these studies are typically, although not always (Lin et al. 2014), carried out in high income Western countries. We address both limitations here by presenting cross-national epidemiological data on comorbidities of DSM-IV anxiety disorders and MDD using a composite measure that includes a wide range of anxiety disorders in a coordinated series of 27 community epidemiological surveys carried out in 24 countries throughout the world. We estimate the proportions of survey respondents with lifetime and 12-month DSM-IV MDD who also met criteria for one or more lifetime and 12-month DSM-IV anxiety disorders. We examine cross-national consistencies in AOO priorities between comorbid anxiety disorders and MDD, whether anxious MDD is more severe and persistent than non-anxious MDD, and whether people with anxious MDD are more likely than those with non-anxious MDD to obtain professional treatment for MDD. We also examine cross-national consistency in basic socio-demographic correlates of anxious and non-anxious MDD. METHODS Sample Data come from the WHO World Mental Health (WMH) surveys, a series of community epidemiological surveys administered in 10 countries classified by the World Bank (World Bank, 2009) as low or middle income (Brazil, Bulgaria, Colombia, Iraq, Lebanon, Mexico, Nigeria, Peru, Peoples Republic of China [PRC], and Romania) and 14 high income (Australia, Belgium, France, Germany, Israel, Italy, Japan, Netherlands, New Zealand, Northern Ireland, Poland, Portugal, Spain, and the United States). The majority of surveys (5 in low-middle income countries, 13 in high income countries) were based on nationally representative household samples. Two were representative of all urban areas in their countries (Colombia, Mexico). Two were representative of selected regions in their countries (Japan, Nigeria). And a final 5 were representative of selected Metropolitan Areas in their countries (Sao Paulo in Brazil; Medellin in Colombia; Murcia in Spain; Beijing-Shanghai and Shenzhen in PRC). Standardized interviewer training and quality control procedures were used in each survey (Pennell et al. 2008). Informed consent was obtained before administering interviews. The institutional review boards of the organizations coordinating the surveys approved and monitored compliance with procedures for informed consent and protecting human subjects. Interviews were administered face-to-face by trained lay interviewers in respondents’ homes. A total of 138,602 adults (age 18+) completed interviews. The weighted (by sample size) average response rate was 68.7%. To reduce respondent burden, the interview was divided into two parts. Part I, which assessed core mental disorders, was administered to all respondents. Part II, which assessed additional disorders and correlates, was administered to all Part I respondents who met criteria for any Part I disorder plus a probability subsample of other Part I respondents. Part II interviews (n= 74,045), the focus of the current report, were weighted by the inverse of their probabilities of selection into Part II and additionally weighted to adjust samples to match population distributions on the cross-classification of key socio-demographic and geographic variables. Further details about WMH sampling and weighting are available elsewhere (Heeringa et al. 2008). Measures Mental disorders Mental disorders were assessed with the WHO Composite International Diagnostic Interview (CIDI) Version 3.0, (Kessler and Üstün, 2004), a fully-structured lay-administered interview generating lifetime and 12-month prevalence estimates of 20 mood (major depressive, dysthymic, bipolar I–II, sub-threshold bipolar), anxiety (generalized anxiety, panic, agoraphobia, specific phobia, social phobia, post-traumatic stress, separation anxiety), behavior (attention-deficit/hyperactivity, oppositional-defiant, conduct, intermittent explosive), and substance (alcohol and drug abuse, alcohol and drug dependence with abuse) disorders. The WMH interview translation, back-translation, and harmonization protocol required culturally competent bilingual clinicians to review, modify, and approve key phrases describing symptoms (Harkness et al. 2008). However, no attempt was made to go beyond DSM-IV criteria to assess depression-equivalents that might be unique to specific countries. The latter expansion might have led to a change in results, although previous research has shown that the latent structure of major depression is quite consistent across countries (Simon et al. 2002; Bernert et al. 2009; Schrier et al. 2010). Blinded clinical reappraisal interviews with the Structured Clinical Interview for DSM-IV (SCID; First et al. 2002) were carried out in four WMH countries. Good concordance was found with diagnoses based on the CIDI (Haro et al. 2006). AOO was assessed using a special probing sequence shown experimentally to yield more plausible distributions than conventional AOO questions (Knäuper et al. 1999). MDD was defined as meeting lifetime DSM-IV/CIDI criteria for major depressive episode (MDE) and not meeting lifetime DSM-IV/CIDI criteria for broadly-defined bipolar disorder (bipolar I-II or sub-threshold). As detailed elsewhere (Merikangas et al. 2011), our definition of sub-threshold bipolar disorder includes both hypomania without history of MDE and sub-threshold hypomania with history of MDE. Anxious MDD is defined as MDD in conjunction with any of the anxiety disorders assessed in the surveys. Comorbid anxiety is considered temporally primary if at least one lifetime anxiety disorder had an AOO earlier than that of MDD. MDD is considered temporally primary if MDD AOO is earlier than that of all lifetime comorbid anxiety disorders. A third category consists of respondents who reported that MDD AOO was the same as anxiety disorder AOO. Impairment in role functioning Severe role impairment in the 12 months before interview was assessed with a modified version of the Sheehan Disability Scales (SDS; Leon et al. 1997) that asked respondents with 12-month MDD to think of the one month in the year when their depression was most severe and rate how much their depression interfered with their functioning in each of four role domains (home management, ability to work, social life, and close relationships) during that month using a 0–10 response scale with labels of None (0), Mild (1–3), Moderate (4–6), Severe (7–9), and Very Severe (10) interference. Severe role impairment was defined as having any SDS score of 7–10. The SDS has excellent internal consistency reliability (Leon et al. 1997) and good concordance with objective measures of role functioning (Ormel et al. 2008). Suicide ideation was assessed with a single question that asked respondents whether there was ever a time in the 12 months before interview when they “seriously thought about committing suicide.” Socio-demographics We examined associations of MDD with respondent age (18–34, 35–49, 50–64, 65+), gender, current marital status (married, never married, previously married [combining separated, divorced, and widowed]), current income (low, low-average, high-average, and high based on country-specific quartiles of gross household income per family member), and education (none, some primary, completed primary, some secondary, completed secondary, some college or other post-secondary, completed college). Treatment Respondents with lifetime MDD were asked if they ever obtained professional treatment for their depression and, if so, if they did so in the past 12 months. Those with 12-month treatment were asked if they saw a mental health specialty treatment provider (psychiatrist, psychologist, other mental health professional in any setting, social worker or counselor in a mental health specialty treatment setting, used a mental health hotline) general medical treatment provider (primary care doctor, other general medical doctor, any other health care profession seen in a general medical setting), or nonmedical treatment provider (religious or spiritual advisor, social worker or counselor, any other type of healer) for a mental health problem. A more detailed description of WMH 12-month treatment measures is presented elsewhere (Wang et al. 2007). Statistical Analyses Cross-tabulations were used to estimate lifetime and 12-month DSM-IV/CIDI MDD prevalence, the proportions of lifetime and 12-month cases with comorbid DSM-IV anxiety disorders, the proportions of lifetime comorbid cases with anxiety disorder or MDD temporally primary AOO, 12-month prevalence of severe role impairment and suicide ideation related to comorbid anxiety disorders among respondents with 12-month MDD, and 12-month MDD treatment as a function of comorbid anxiety disorders among respondents with 12-month MDD. Person-level logistic regression was used to examine multivariate associations of socio-demographic variables with lifetime and 12-month MDD in the total sample, lifetime anxious MDD among respondents with lifetime MDD, and 12-month anxious MDD among respondents with 12-month MDD. Time-varying socio-demographics (i.e., marital status, income, and education) were defined as of the time of interview (rather than at time of disorder onset). Standard errors were estimated using the Taylor series linearization method (Wolter, 1985) implemented in the SUDAAN software system (Research Triangle Institute, 2002) to adjust weighting and clustering. Multivariate significance of predictor sets was evaluated using Wald χ2 tests based on design-corrected coefficient variance-covariance matrices. Statistical significance was evaluated using two-sided .05-level tests. RESULTS Prevalence Lifetime prevalence of DSM-IV/CIDI MDD averaged 11.2% across surveys, 8.1% in low/middle income countries, and 13.0% in high income countries. (Table 1) The inter-quartile range (IQR; 25th–75th percentiles) of lifetime prevalence estimates was 6.8–15.3%. 45.7% of respondents with lifetime MDD (42.4% in low/middle income countries, 46.9% in high income countries, 32.0–46.5% IQR) also had one of more lifetime DSM-IV/CIDI anxiety disorders. A comparison of the ratios of (75th percentile – 25th percentile)/Mean shows that lifetime prevalence varied much more across surveys than did the proportion of lifetime cases with lifetime comorbid anxiety disorders (.75 vs. .32). Twelve-month MDD prevalence averaged 4.7% across surveys (4.0% in low/middle income countries, 5.1% in high income countries, 3.0–5.6 IQR), while 51.7% of respondents with 12-month MDD (48.0% in low/middle income countries, 52.2% in high income countries, 37.8–54.0% IQR) also had one or more lifetime DSM-IV/CIDI anxiety disorders. Only slightly lower proportions of respondents with 12-month MDD had 12-month comorbid anxiety disorders (41.6% in the total sample, 38.8% in low/middle income countries, 42.9% in high income countries, 29.9–47.2% IQR). As with lifetime prevalence, a comparison of the ratios of (75th percentile – 25th percentile)/Mean shows that 12-month prevalence varied more across surveys than did the proportion of 12-month cases with 12-month comorbid anxiety disorders (.55 vs. .42). Age-of-onset priorities Two-thirds (68.0%) of respondents with lifetime anxious MDD reported first onset of anxiety disorders at earlier ages than MDD (71.5% in low/middle income countries, 66.9% in high income countries, 69.6–74.7% IQR), (Table 2) while 13.5% reported earlier AOO of MDD than anxiety disorder (11.3% in low/middle income countries, 14.2% in high income countries, 10.2–15.6% IQR) and the remaining 18.5% (17.2% in low/middle income countries, 18.9% in high income countries, 10.6–23.7% IQR) reported the same AOO of anxiety disorders and MDD. The dominant temporal priority of anxiety disorders before MDD occurred in all surveys other than in Israel, where the proportions with temporally primary anxiety (30.0%) and MDD (33.3%) were virtually the same (χ21=0.2, p=.62). Comparable proportions of respondents reporting temporally primary anxiety disorders were found for 12-month comorbid cases (68.5% in the total sample, 72.9% in low/middle income countries, 66.7% in high income countries, 62.0–72.8% IQR), again with the exception of Israel in addition to Murcia in Spain. It is noteworthy that rates of comorbid anxiety disorder among respondents with MDD (reported Table 1) were comparatively low in both Israel and Murcia (21.9–32.0% compared to IQR 32.0–46.5%). Socio-demographic correlates Significantly higher rates of lifetime MDD were found among respondents in middle age (ages 34–64) compared to ages 65+ (OR=1.8–1.9), women compared to men (OR=1.8), the previously-married compared to currently-married (OR=2.0), and those with less than high incomes compared to those with high incomes (OR=1.1). (Table 3) Slightly lower lifetime prevalence of MDD was found among respondents with less than some college education compared to those with at least some college education (OR=0.7–0.9). Country-specific analyses (available online) showed that the most consistent of these associations were being female (significant in 23 surveys, OR IQR 1.6–2.2) and previously-married (significant in 24 surveys, OR IQR 1.8–2.4). The elevated ORs associated with being middle-aged were significant in 13 surveys (OR IQR 1.7–2.9). The associations of income and education with lifetime MDD were inconsistent across surveys. Significant associations of these same socio-demographic variables were found with comorbid anxiety disorders among respondents with lifetime MDD, but the ORs for age, sex, and income were higher than in predicting lifetime MDD: OR=1.8–2.7 for ages 18–44 compared to 65+; OR=2.1 for women compared to men; and OR=1.3 for low compared to high income. ORs associated with marital status and education were virtually identical to those predicting lifetime MDD. Country-specific analyses (results available online) showed that the most consistently elevated ORs predicting comorbid anxiety among people with MDD were being female (significant in 22 surveys, OR IQR 1.7–2.7) and previously-married (significant in 15 surveys, OR IQR 1.5–2.1). The elevated ORs associated with being middle-aged were significant in 15 surveys (OR IQR 2.1–4.2), while the associations of income and education with lifetime anxious vs. non-anxious MDD were inconsistent across surveys. Socio-demographics were also significantly associated with 12-month MDD. The positive ORs of young age, not being married, low to high-average income, and less than college education were consistently somewhat larger than those of the same predictors with lifetime MDD. The OR associated with female gender, in comparison, was identical in predicting lifetime and 12-month MDD (OR=1.8), indicating that women did not differ significantly from men in 12-month prevalence among lifetime cases. Broadly similar patterns were also found in predicting 12-month anxious MDD vs. non-anxious MDD in that the ORs were all significant as a set (albeit with some ORs for specific education categories not significant even though education was significant overall [χ26=32.3, p<.001]) and either equal or higher in magnitude than those associated with MDD in the total sample. This means that these socio-demographics were all more strongly associated with persistence of anxious MDD than persistence of non-anxious MDD. The results of more detailed within-country analyses were unstable due to small numbers of cases (results available online). Severity of anxious versus non-anxious MDD The proportion of respondents with 12-month MDD who reported severe role impairment was significantly higher in the presence (64.4%) than absence (46.0%) of 12-month anxiety disorders (χ21=187.0, p<.001). (Table 4) Very similar overall patterns were found in low/middle income countries (61.5% vs. 41.9%; χ21=97.5, p<.001) and high income countries (68.8% vs. 49.6%; χ21=128.0, p<.001) although the pattern was less consistent in low/middle income countries (7 of 12 surveys, 6 statistically significant) than high income countries (all 15 surveys, 10 significant). The proportion of respondents with 12-month MDD who reported suicide ideation was also significantly higher in the presence 19.5%) than absence (8.9%) of 12-month anxiety disorders (χ21=71.6, p<.001). Very similar patterns were found in low/middle income countries (15.2% vs. 9.4%; χ21=7.9, p=.005) and high income countries (21.3% vs. 8.7%; χ21=72.8, p<.001) overall, although consistency of the pattern was again somewhat lower in low/middle income countries (9 of 12 surveys, 3 statistically significant) than high income countries (14 of 15 surveys, 13 significant). Treatment Twelve-month treatment of MDD was significantly more common in the presence than absence of 12-month anxiety disorders (56.2% vs. 37.3%; χ21=21.8, p<.001). (Table 5) Very similar relative treatment rates were found in low/middle income (30.3% vs. 20.6%; χ21=11.7, p<.001) and high income (68.8% vs. 45.4%; χ21=108.8, p<.001) countries (i.e., respondents with anxious MDD were roughly 50% more likely to receive treatment than those with non-anxious MDD). However, the absolute difference in treatment rates was much higher in high income (a 23.4% higher treatment rate of anxious than non-anxious MDD [68.8%–45.4%]) than low/middle income (a 9.7% higher treatment rate of anxious than non-anxious MDD [30.3%–20.6%]) countries due to the overall treatment rate being much higher in high income than low/middle income countries. The pattern of higher treatment of anxious than non-anxious MDD was also more consistent in high income (all 15 surveys, 10 significant) than low/middle income (9 of 12 surveys, 3 significant) countries. Similarly significant patterns were found in separate treatment sectors other than the nonmedical sector in low/middle income countries (χ21=4.7–7.5; p=.030–.006) and in all sectors in high income countries (χ21=36.6–77.1; p<.001). DISCUSSION The above results are limited by between-survey differences in response rates and sample frames (most notably, under-representation of rural areas in developing countries), diagnoses being based on fully-structured lay interviews rather than semi-structured clinician-administered interviews (although available evidence documents good concordance between the two types of diagnoses in WMH; Haro et al. 2006), the fact that we examined only a summary measure of any DSM-IV anxiety disorder rather than disaggregated disorder-specific measures, and the fact that the WMH surveys were cross-sectional. The latter limitation meant that both lifetime prevalence and AOO were assessed retrospectively. Previous methodological studies (Moffitt et al. 2010; Hamdi and Iacono, 2014; Takayanagi et al. 2014) suggest that use of retrospective recall probably led to under-estimation of lifetime prevalence and over-estimation of persistence. Long-term prospective studies are needed to resolve this problem. Within the context of these limitations, we found a relatively narrow IQR across surveys (32.0–46.5%) in estimated rates of lifetime anxiety disorders among people with lifetime MDD, somewhat higher rates but an equally narrow IQR (37.8–54.0%) of lifetime comorbid anxiety disorders among respondents with 12-month MDD, and only slightly lower rates with a similarly narrow IQR (29.9–47.2%) of 12-month comorbid anxiety disorders among respondents with 12-month MDD. The fact that lifetime comorbid anxiety disorders were more prevalent among respondents with 12-month than lifetime MDD suggests that lifetime comorbid anxiety disorders predict MDD persistence, while the fact that 12-month comorbid anxiety disorders were only slightly less prevalent than lifetime comorbid anxiety disorders among respondents with 12-month MDD suggests that anxiety disorder persistence is positively associated with MDD persistence. These patterns are broadly consistent with previous epidemiological studies (Andrade et al. 2003; Kessler et al. 2011b; Lamers et al. 2011). We are unaware, though, of previous studies that examined either the differences we did in the magnitudes of lifetime comorbidity, 12-month comorbidity or comorbidity between lifetime anxiety disorders and 12-month MDD. Our results also go beyond previous studies in documenting considerable cross-national consistency in comorbidity between anxiety disorders and MDD. The socio-demographic associations documented here are broadly consistent with previous studies in finding higher rates of both anxiety disorders and MDD among women (Parker and Brotchie, 2010; Altemus et al. 2014) and the previously-married (Scott et al. 2010; Leach et al. 2013) along with less consistent inverse associations with age (de Graaf et al. 2013; McDowell et al. 2014). However, we are unaware of prior systematic efforts to examine nested associations in the way we did here. It is noteworthy, though, that we did not examine disaggregated associations (e.g., the extent to which socio-demographics predict onset of secondary MDD among people with a history of temporally primary anxiety disorders). The strength and consistency of the associations we documented across nested outcomes suggest that more detailed studies of these specifications might be useful. Our finding of higher role impairment and suicidality in anxious than non-anxious MDD is broadly consistent with previous findings (Roy-Byrne et al. 2000; McLaughlin et al. 2006; Ormel et al. 2008), although we showed that this pattern generalizes to many more countries than in previous research. We found stronger and more consistent associations of comorbid anxiety disorders with elevated MDD treatment rates in high income than low/middle income countries. Previous studies of this pattern, which were limited to high income Western countries (Kessler et al. 2001; Jacobi et al. 2004), found similar associations to those in the high income WMH surveys. The over-representation of anxious MDD in treatment populations is important because comorbid anxiety disorders predict both low MDD treatment persistence (Shippee et al. 2014) and low MDD treatment response (Stiles-Shields et al. 2014). Our finding that the vast majority of WMH respondents with anxious MDD reported earlier AOO of anxiety disorders than MDD is consistent with previous research in both cross-sectional/retrospective (Kessler, 1995; Kessler et al. 2011a) and prospective (Murphy et al. 1986; Bittner et al. 2004; Copeland et al. 2009; Klein et al. 2013) samples. The narrow IQR of the proportion of respondents who reported earlier AOO of anxiety disorders than MDD (69.6–74.7%) is especially noteworthy. It is also striking, though, that these proportions are not higher among respondents with 12-month than lifetime comorbidity, as we might expect the latter rates to be higher if temporally primary comorbid anxiety was more important than temporally secondary comorbid anxiety in predicting MDD persistence. We are unaware of any previous research on this distinction. The finding that comorbid anxiety disorder is related to MDD persistence equally whether or not the anxiety is temporally primary might reflect influences of common underlying causes accounting for the lifetime comorbidities of MDD with anxiety disorders, although another possibility consistent with this pattern is that the causal processes accounting for the effects of anxiety disorders on MDD onset differ from the causal processes accounting for the effects of anxiety disorders on MDD persistence. We have no way to adjudicate between these competing possibilities with the WMH data. If temporally primary anxiety disorders are causal risk factors for MDD, interventions to treat pure anxiety disorders would be expected to reduce subsequent onset of MDD. However, no well-controlled long-term treatment studies have evaluated this possibility despite calls to do so (Flannery-Schroeder, 2006; Garber and Weersing, 2010). Consistent with this possibility, though, two observational studies based on community epidemiological surveys found that individuals who received treatment for temporally primary panic disorder (Goodwin and Olfson, 2001) and generalized anxiety disorder (Goodwin and Gorman, 2002) were significantly less likely than others with these disorders to go on to develop temporally secondary MDD. While selection bias into treatment is a possible explanation for these patterns, the most plausible type of selection bias (i.e., selection into treatment based on severity) would be expected to lead to the opposite association with subsequent MDD, arguing indirectly for the possibility that treatment of anxiety disorders might lead to a reduction in risk of subsequent MDD. Other results consistent with this possibility include those from controlled studies of focused psychotherapy for anxiety disorders that showed these treatments reduced concurrent symptoms of MDD (reviewed in (Hofmann and Smits, 2008; Cuijpers et al. 2014), although this result is not entirely consistent (McLean et al. 1998; Woody et al. 1999). In addition, one small controlled study of CBT for social phobia among adolescent girls found that treatment reduced relapse of MDD among patients with a history of comorbid MDD over the subsequent year (Hayward et al. 2000). Despite the suggestive evidence in the above studies, more definitive long-term controlled efficacy trials are needed to evaluate the impact of interventions to treat temporally primary anxiety disorders on the subsequent onset and persistence of MDD. An intriguing observation related to the need for this kind of definitive long-term controlled treatment study is that several epidemiological studies have found distinct risk factors for anxiety disorders that are not risk factors for MDD (Moffitt et al. 2007; Beesdo et al. 2010; Mathew et al. 2011; Asselmann et al. 2015). For example, an extensive literature shows that stressful life events associated with danger predict anxiety but not depression (Finlay-Jones and Brown, 1981; Kendler et al. 2003; Asselmann et al. 2015). This specificity should not exist if anxiety disorders caused MDD, as the latter causal process would lead to attenuated associations of the risk factors with MDD. To find that this is not the case suggests that something more complex is at work linking anxiety disorders with MDD and that common causes are involved in the comorbidity of anxiety disorders with MDD. The existence of common causes would impose an upper bound on how much secondary MDD could be prevented by successful treatment of temporally primary anxiety disorders. Common causes might also help account for the fact that concurrent comorbidity is associated with poor treatment response for both anxiety disorders (Rapee et al. 2013; Kelly et al. 2014) and MDD (Jakubovski and Bloch, 2014; Saveanu et al. 2014). Less is known, though, about the associations of lifetime comorbidity with treatment response among patients who do not have concurrent comorbid symptoms. An examination of this specification would be useful in helping distinguish differential treatment response associated with the amelioration of life stressors surrounding particular anxious-MDD episodes and risk factors associated with more fundamental causes of lifetime anxious-MDD. The latter suggestion highlights the fact that little research has attempting to distinguish the determinants of first onsets from the determinants of the subsequent course of either anxiety disorders or MDD. Virtually all the research cited above on the association between treating anxiety disorders and subsequent chance in depression as well as on the associations of comorbidity with differential treatment response implicitly focused on the course of depression, as only a small minority of MDD cases in clinical studies are first-onset cases. As noted above, it is quite possible that different processes are at work in bringing about lifetime comorbidity and episode comorbidity. Research is needed to investigate such differences explicitly. It is noteworthy in this regard that epidemiological research assuming the existence of common causes has shown that the coefficients describing the cross-lagged prospective associations of temporally primary lifetime anxiety with subsequent first lifetime onset of MDD and vice versa can be parsimoniously described by assuming a latent intervening predisposition to all internalizing disorders (Kessler et al. 2011a; Kessler et al. 2011b). While some hypotheses have been advanced for asymmetries in these associations (Cummings et al. 2014), available evidence suggests that these asymmetries are weak. If this model is accurate, then temporally primary lifetime anxiety disorders might be risk markers rather than causal risk factors for the subsequent first onset of lifetime MDD even though anxiety disorders might have causal effects on the subsequent persistence of MDD. If this is the case, then successful intervention to treat early-onset primary anxiety disorders might not prevent the subsequent first onset of MDD even though it would reduce MDD persistence. A more consistent distinction between lifetime and concurrent comorbidity needs to be made in future observational and clinical studies of anxious MDD to shed light on these possibilities. As part of this increased focus, any attempt to carry out long-term controlled treatment studies to evaluate the effects of treating temporally primary anxiety on subsequent MDD should be designed to have a sufficiently large sample size and a sufficient duration of follow-up to examine effects on both MDD onset and MDD persistence and to include an assessment of plausible biomarkers. Our understanding of the causal determinants of the high comorbidity of MDD with anxiety disorders will remain at its current relatively primitive level until studies of this sort are carried out. Supplementary Material Appendix Tables The authors appreciate the helpful contributions to WMH of Herbert Matschinger, PhD. FINANCIAL SUPPORT: The WMH surveys were supported by the United States National Institute of Mental Health (R01MH070884), the John D. and Catherine T. MacArthur Foundation, the Pfizer Foundation, the US Public Health Service (R13-MH066849, R01-MH069864, and R01 DA016558), the Fogarty International Center (FIRCA R03-TW006481), the Pan American Health Organization, the Eli Lilly & Company Foundation, Ortho-McNeil Pharmaceutical, Inc., GlaxoSmithKline, Bristol-Myers Squibb, and Shire. The São Paulo Megacity Mental Health Survey is supported by the State of São Paulo Research Foundation (FAPESP) Thematic Project Grant 03/00204-3. The Bulgarian Epidemiological Study of common mental disorders EPIBUL is supported by the Ministry of Health and the National Center for Public Health Protection. The Chinese World Mental Health Survey Initiative is supported by the Pfizer Foundation. The Shenzhen Mental Health Survey is supported by the Shenzhen Bureau of Health and the Shenzhen Bureau of Science, Technology, and Information. The Colombian National Study of Mental Health (NSMH) is supported by the Ministry of Social Protection. The ESEMeD project is funded by the European Commission (Contracts QLG5-1999-01042; SANCO 2004123), the Piedmont Region (Italy), Fondo de Investigación Sanitaria, Instituto de Salud Carlos III, Spain (FIS 00/0028), Ministerio de Ciencia y Tecnología, Spain (SAF 2000-158-CE), Departament de Salut, Generalitat de Catalunya, Spain, Instituto de Salud Carlos III (CIBER CB06/02/0046, RETICS RD06/0011 REM-TAP), and other local agencies and by an unrestricted educational grant from GlaxoSmithKline. The WMHI was funded by WHO (India) and helped by Dr R Chandrasekaran, JIPMER. Implementation of the Iraq Mental Health Survey (IMHS) and data entry were carried out by the staff of the Iraqi MOH and MOP with direct support from the Iraqi IMHS team with funding from both the Japanese and European Funds through United Nations Development Group Iraq Trust Fund (UNDG ITF). The Israel National Health Survey is funded by the Ministry of Health with support from the Israel National Institute for Health Policy and Health Services Research and the National Insurance Institute of Israel. The World Mental Health Japan (WMHJ) Survey is supported by the Grant for Research on Psychiatric and Neurological Diseases and Mental Health (H13-SHOGAI-023, H14-TOKUBETSU-026, H16-KOKORO-013) from the Japan Ministry of Health, Labour and Welfare. The Lebanese National Mental Health Survey (LEBANON) is supported by the Lebanese Ministry of Public Health, the WHO (Lebanon), Fogarty International, Act for Lebanon, anonymous private donations to IDRAAC, Lebanon, and unrestricted grants from Janssen Cilag, Eli Lilly, GlaxoSmithKline, Roche, and Novartis. The Mexican National Comorbidity Survey (MNCS) is supported by The National Institute of Psychiatry Ramon de la Fuente (INPRFMDIES 4280) and by the National Council on Science and Technology (CONACyT-G30544- H), with supplemental support from the PanAmerican Health Organization (PAHO). Te Rau Hinengaro: The New Zealand Mental Health Survey (NZMHS) is supported by the New Zealand Ministry of Health, Alcohol Advisory Council, and the Health Research Council. The Nigerian Survey of Mental Health and Wellbeing (NSMHW) is supported by the WHO (Geneva), the WHO (Nigeria), and the Federal Ministry of Health, Abuja, Nigeria. The Romania WMH study projects “Policies in Mental Health Area” and “National Study regarding Mental Health and Services Use” were carried out by National School of Public Health & Health Services Management (former National Institute for Research & Development in Health), with technical support of Metro Media Transilvania, the National Institute of Statistics-National Centre for Training in Statistics, SC. Cheyenne Services SRL, Statistics Netherlands and were funded by Ministry of Public Health (former Ministry of Health) with supplemental support of Eli Lilly Romania SRL. The EZOP – Poland (Epidemiology of Mental Disorders and Access to Care) survey was supported by the grant from the EAA/Norwegian Financial Mechanisms as well as by the Polish Ministry of Health). The South Africa Stress and Health Study (SASH) is supported by the US National Institute of Mental Health (R01-MH059575) and National Institute of Drug Abuse with supplemental funding from the South African Department of Health and the University of Michigan. The US National Comorbidity Survey Replication (NCS-R) is supported by the National Institute of Mental Health (NIMH; U01-MH60220) with supplemental support from the National Institute of Drug Abuse (NIDA), the Substance Abuse and Mental Health Services Administration (SAMHSA), the Robert Wood Johnson Foundation (RWJF; Grant 044708), and the John W. Alden Trust. Table 1 Lifetime and 12-month prevalence of DSM-IV/CIDI major depressive disorder (MDD) along with the proportions of respondents with lifetime and 12-month MDD who have comorbid DSM-IV/CIDI anxiety disordersa in the WHO World Mental Health Surveys Lifetime MDD 12-Month MDD Lifetime MDD Lifetime anxiety/Lifetime MDD 12-month MDD Lifetime anxiety/12-month MDD 12-month anxiety/12-month MDD % (SE) % (SE) % (SE) % (SE) % (SE) (n) I. Low/Middle Income  Brazil – São Paulo 18.0 (0.8) 50.0 (2.2) 10.1 (0.6) 51.2 (2.9) 39.9 (2.8) (2,942)  Bulgaria   6.7 (0.5) 32.5 (3.3)   3.0 (0.3) 37.8 (5.0) 36.7 (4.9) (2,233)  Colombia 11.8 (0.7) 50.4 (3.0)   5.3 (0.5) 63.8 (4.1) 50.7 (4.4) (2,381)  Columbia – Medellin   9.9 (0.7) 51.4 (3.6)   3.8 (0.4) 54.0 (5.2) 47.3 (5.1) (1,673)  Iraq   7.2 (0.6) 46.1 (4.2)   3.9 (0.4) 50.7 (5.8) 42.1 (5.7) (4,332)  Lebanon 10.3 (0.8) 44.7 (3.6)   4.9 (0.5) 43.6 (5.3) 35.1 (5.0) (1,031)  Mexico   7.6 (0.5) 46.5 (2.9)   3.7 (0.3) 59.4 (3.9) 46.0 (4.1) (2,362)  Nigeria   3.2 (0.3) 19.1 (3.5)   1.1 (0.2) 20.3 (5.9) 18.7 (5.8) (2,143)  Peru   6.4 (0.4) 35.9 (3.3)   2.7 (0.3) 49.5 (5.4) 37.2 (5.1) (1,801)  PRCb – Beijing/Shanghai   3.8 (0.4) 25.3 (6.0)   2.0 (0.3) 36.8 (9.5) 33.6 (9.7) (1,628)  PRCb – Shenzhen   6.8 (0.6) 18.7 (3.4)   3.6 (0.4) 22.0 (3.8) 16.2 (3.4) (2,475)  Romania   2.9 (0.4) 27.0 (5.7)   1.5 (0.3) 31.5 (8.8) 25.9 (8.4) (2,357)  Total   8.1 (0.2) 42.4 (1.1)   4.0 (0.1) 48.0 (1.6) 38.8 (1.5) (27,358) II. High Income  Australia 12.8 (0.5) 51.4 (2.1)   4.8 (0.3) 59.0 (3.5) 49.0 (3.4) (8,841)  Belgium 14.1 (1.1) 29.8 (3.3)   5.2 (0.7) 37.7 (6.3) 29.9 (5.6) (1,043)  France 20.4 (1.2) 41.6 (2.6)   5.6 (0.6) 51.0 (5.3) 41.8 (5.3) (1,436)  Germany 10.3 (0.7) 45.0 (3.1)   3.1 (0.4) 57.9 (5.6) 48.7 (5.7) (1,323)  Israel   9.8 (0.4) 21.9 (2.0)   5.9 (0.4) 23.7 (2.7) 18.1 (2.4) (4,859)  Italy   9.7 (0.5) 39.3 (2.6)   2.9 (0.3) 46.2 (5.1) 41.3 (5.0) (1,779)  Japan   6.8 (0.5) 27.6 (3.5)   2.4 (0.3) 42.2 (6.7) 29.1 (5.9) (1,682)  Netherlands 18.0 (1.3) 40.6 (3.3)   4.9 (0.7) 49.3 (6.6) 32.5 (5.9) (1,094)  New Zealand 15.8 (0.5) 52.4 (1.4)   5.7 (0.3) 60.3 (2.4) 49.5 (2.4) (7,312)  Northern Ireland 17.7 (1.0) 56.8 (2.8)   8.8 (0.7) 61.4 (4.4) 47.2 (4.3) (1,986)  Poland   3.8 (0.3) 32.2 (3.6)   1.6 (0.2) 36.7 (5.1) 27.8 (4.6) (4,000)  Portugal 17.4 (0.8) 45.3 (2.1)   7.0 (0.5) 49.8 (3.3) 42.5 (3.3) (2,060)  Spain 10.4 (0.6) 34.9 (2.6)   3.8 (0.3) 48.4 (4.5) 40.0 (4.3) (2,121)  Spain – Murcia 15.3 (1.0) 32.0 (3.0)   6.9 (0.7) 36.7 (4.8) 23.2 (3.8) (1,459)  US 16.6 (0.5) 62.6 (1.4)   6.7 (0.3) 71.9 (1.9) 58.5 (2.2) (5,692)  Total 13.0 (0.2) 46.9 (0.7)   5.1 (0.1) 53.3 (1.1) 42.9 (1.1) (46,687) III. Total 11.2 (0.1) 45.7 (0.6)   4.7 (0.1) 51.7 (0.9) 41.6 (0.9) (74,045) a Anxiety disorders include generalized anxiety disorder, panic disorder with or without agoraphobia, agoraphobia without a history of panic disorder, specific phobia, social phobia, separation anxiety, and post-traumatic stress disorder. b People’s Republic of China. Table 2 Temporal priority in age-of-onset (AOO) distributions of lifetime DSM-IV/CIDI major depressive disorder (MDD) and anxiety disordersa among respondents with lifetime and 12-month comorbid MDD and anxiety disorders in the WHO World Mental Health Surveys Among lifetime comorbid casesb Among 12-Month comorbid casesb Anxiety first MDD first Anxiety first MDD first % (SE) % (SE) (n) % (SE) % (SE) (n) I. Low/Middle Income  Brazil – São Paulo 76.6* (2.6) 12.8 (2.0) (439) 75.6* (3.4) 15.7 (3.0) (205)  Bulgaria 62.8* (5.6) 11.0 (3.1) (145) 62.0* (7.6) 13.1 (4.2) (54)  Colombia 78.3* (3.0)   8.2 (1.8) (285) 84.4* (3.3)   9.3 (2.7) (110)  Columbia – Medellin 79.2* (3.4) 10.1 (2.5) (186) 87.4* (4.1)   6.9 (3.0) (73)  Iraq 44.4* (6.3) 12.2 (3.5) (167) 47.6* (8.2) 11.0 (5.1) (81)  Lebanon 68.1* (4.8) 17.2 (4.1) (128) 68.1* (4.8) 18.8 (6.4) (49)  Mexico 78.7* (3.6) 13.4 (3.1) (212) 77.8* (4.1) 15.8 (3.7) (101)  Nigeria 88.9* (4.7)   5.2 (3.3) (36) 86.1* (9.1)   7.8 (7.9) (12)  Peru 66.5* (5.3) 15.6 (4.0) (85) 69.3* (6.9) 17.3 (5.6) (37)  PRCc – Beijing/Shanghai 74.4* (9.5)   5.1 (3.6) (40) 70.0* (13.1)   5.1 (4.5) (21)  PRCc – Shenzhen 87.5* (5.20   4.0 (2.2) (70) 84.7* (7.1)   2.9 (2.2) (37)  Romania 89.9* (5.70   3.9 (2.5) (25) 97.2* (3.0)   2.8 (3.0) (11)  Total 71.5* (1.5) 11.3 (1.0) (1,782) 72.9* (2.0) 12.4 (1.4) (791) II. High Income  Australia 62.8* (2.6) 13.5 (1.8) (623) 62.0* (4.0) 16.2 (2.8) (229)  Belgium 64.4* (6.8) 14.2 (5.4) (106) 65.9* (10.1) 21.5 (10.0) (39)  France 70.2* (3.8) 15.6 (2.9) (261) 64.1* (7.3) 18.0 (5.4) (68)  Germany 69.6* (4.9) 17.1 (4.8) (164) 72.8* (8.0) 19.9 (8.0) (55)  Israel 30.0 (4.8) 33.3 (4.7) (104) 31.0 (6.1) 32.5 (5.9) (48)  Italy 57.9* (4.3) 12.8 (3.0) (172) 56.9* (7.4) 13.8 (5.8) (51)  Japan 55.8* (7.4)   5.4 (3.2) (64) 59.1* (10.4)   7.2 (5.3) (25)  Netherlands 58.1* (4.8) 18.7 (4.4) (186) 52.8 (8.5) 31.4 (8.5) (40)  New Zealand 70.2* (1.8) 10.2 (1.3) (1,125) 71.1* (2.8) 12.8 (2.3) (370)  Northern Ireland 74.7* (3.1) 17.1 (2.7) (282) 77.2* (4.0) 13.8 (3.2) (110)  Poland 73.3* (5.7) 17.5 (4.9) (89) 71.1* (8.2) 21.9 (7.9) (36)  Portugal 69.5* (2.8) 17.3 (2.3) (338) 72.8* (4.1) 17.9 (3.7) (126)  Spain 56.8* (4.4) 11.2 (2.5) (217) 50.7* (6.1) 14.5 (4.2) (87)  Spain – Murcia 49.7* (5.5) 22.3 (4.8) (131) 44.7 (7.8) 31.5 (7.8) (45)  US 75.6* (1.6) 12.5 (1.3) (957) 77.5* (2.2) 13.2 (1.8) (375)  Total 66.9* (0.9) 14.2 (0.7) (4,818) 66.7* (1.3) 16.5 (1.1) (1,704) III. Total 68.0* (0.8) 13.5 (0.6) (6,600) 68.5 (1.1) 15.3 (0.9) (2,495) * Significant difference between the proportion of cases with anxiety temporally primary vs. MDD temporally primary at the .05 level, two-sided test. a Anxiety disorders include generalized anxiety disorder, panic disorder with or without agoraphobia, agoraphobia without a history of panic disorder, specific phobia, social phobia, separation anxiety, and post-traumatic stress disorder. b Percentages with MDD first and anxiety first sum to less than 100% because some respondents reported that their MDD and anxiety disorders started at the same age. In cases where respondents had multiple anxiety disorders, earliest AOO was used. c People’s Republic of China Table 3 Socio-demographic correlates of lifetime and 12-month DSM-IV/CIDI major depressive disorder (MDD) and of comorbid anxiety disordersa given MDD in the WHO World Mental Health Surveysb Lifetime 12-Month MDD Anxiety disorder/MDD MDD Anxiety disorder/MDD OR (95% CI) OR (95% CI) OR (95% CI) OR (95% CI) Age  18–34 1.4* (1.25–1.52) 1.8* (1.56–2.09) 2.1* (1.78–2.42) 2.6* (2.03–3.24)  35–49 1.9* (1.73–2.06) 2.6* (2.29–2.96) 2.5* (2.16–2.84) 3.3* (2.67–4.07)  50–64 1.7* (1.55–1.85) 2.2* (1.93–2.49) 1.9* (1.68–2.21) 2.6* (2.14–3.24)  65+ 1.0 – 1.0 – 1.0 – 1.0 –   χ23 256.9* 247.2* 172.8* 126.0* Gender  Male 1.0 – 1.0 – 1.0 – 1.0 –  Female 1.8* (1.73–1.91) 2.1* (1.97–2.28) 1.8* (1.66–1.96) 2.1* (1.83–2.36)   χ21 536.0* 419.0* 199.5* 126.8* Marital status  Married 1.0 – 1.0 – 1.0 – 1.0 –  Never Married 1.2* (1.07–1.23) 1.2* (1.04–1.27) 1.4* (1.22–1.49) 1.4* (1.23–1.69)  Sep/Wid/Divorced 2.0* (1.90–2.19) 2.0* (1.81–2.19) 2.2* (1.97–2.42) 2.3* (2.01–2.72)   χ22 387.5* 203.1* 237.5* 128.5* Incomec  Low 1.1 (0.99–1.15) 1.3* (1.16–1.43) 1.4* (1.23–1.54) 1.6* (1.36–1.89)  Low-Mid 1.1* (1.05–1.22) 1.3* (1.20–1.47) 1.3* (1.19–1.49) 1.6* (1.31–1.83)  Mid-High 1.1* (1.05–1.20) 1.3* (1.15–1.39) 1.2* (1.05–1.30) 1.3* (1.08–1.49)  High 1.0 – 1.0 – 1.0 – 1.0 –   χ23 15.3* 36.6* 37.3* 36.5* Education level  None 0.7* (0.59–0.88) 0.7* (0.51–0.93) 1.0 (0.75–1.37) 0.9 (0.58–1.40)  Some primary 0.9* (0.79–0.98) 0.9 (0.79–1.08) 1.2* (1.03–1.42) 1.2 (0.98–1.53)  Completed primary 0.7* (0.64–0.80) 0.8* (0.66–0.90) 0.9 (0.74–1.04) 1.0 (0.78–1.29)  Some secondary 0.8* (0.72–0.85) 0.9* (0.76–0.97) 0.9* (0.76–0.99) 0.9 (0.77–1.16)  Completed secondary 0.8* (0.70–0.82) 0.8* (0.68–0.84) 0.8* (0.72–0.91) 0.8* (0.62–0.90)  Some College 1.0 (0.87–1.04) 1.0 (0.91–1.15) 1.0 (0.91–1.19) 1.1 (0.87–1.29)  Completed college 1.0 – 1.0 – 1.0 – 1.0 –   χ26 85.3* 51.0* 46.2* 32.3*    (n) 74,045 14,430 74,045 5,898 * Significant at the .05 level, two-sided test a Anxiety disorders include generalized anxiety disorder, panic disorder with or without agoraphobia, agoraphobia without a history of panic disorder, specific phobia, social phobia, separation anxiety, and post-traumatic stress disorder. b Based on person-level logistic regression models pooled across surveys. Time-varying socio-demographic variables were coded as of time of interview rather than as of AOO. c Income was coded within country. Low = less than 50% of the median value of the ratio of before-tax income to number of family members; Low-average = 50–100% of the median value of the ratio of before-tax income to number of family members; High-average = more than 100% to 300% of the median value of the ratio of before-tax income to number of family members; High = more than 300% of the median value of the ratio of before-tax income to number of family members. Table 4 Two indicators of severity (proportion of cases reporting severe role impairment due to depression and proportion of cases reporting suicide ideation) among respondents with 12-month DSM-IV/CIDI major depressive disorder (MDD) depending on presence or absence of comorbid anxiety disordersa in the WHO World Mental Health Surveys Prevalence of severe role impairmentb when comorbid anxiety disorders are … Prevalence of suicide ideationc when comorbid anxiety disorders are … Number of respondents with 12-month MDD where comorbid anxiety disorders are … Present Absent Present Absent Present Absent % (SE) % (SE) % (SE) % (SE) (n) (n) I. Low/Middle Income  Brazil – São Paulo 57.0* (3.5) 41.2 (2.9) 18.7* (3.4)   7.1 (1.6) (205) (284)  Bulgaria 51.7* (6.9) 41.5 (5.2)   8.7* (4.5) 16.1 (6.7) (54) (91)  Colombia 49.3* (4.8) 42.1 (4.8) 20.8* (5.9) 13.5 (3.6) (110) (131)  Colombia – Medellin 51.0 (5.9) 51.1 (5.7) 21.3 (5.8) 18.5 (5.3) (73) (78)  Iraq 71.8* (5.0) 36.3 (4.8) 11.8* (5.9)   4.7 (1.8) (81) (101)  Lebanon 63.3 (7.0) 64.5 (5.5)   5.3* (3.8)   9.2 (3.9) (49) (77)  Mexico 45.2* (5.0) 38.6 (4.3) 16.4 (4.3) 15.9 (3.9) (101) (130)  Nigeria 8.0 (8.2) 16.2 (4.8) 20.8 (18.2) 17.7 (6.5) (12) (60)  Peru 35.3 (8.0) 39.7 (6.3)   6.9* (4.7) 18.3 (5.0) (37) (62)  PRCd – Beijing/Shanghai 22.1 (9.3) 32.2 (5.8)   8.2 (5.8)   4.2 (2.8) (21) (66)  PRCd – Shenzhen 41.1* (8.2) 19.4 (2.9)   5.0 (2.9)   3.2 (1.6) (37) (185)  Romania 46.9 (15.8) 40.8 (9.3)   5.0 (5.4)   4.6 (4.6) (11) (29)  Total 61.5* (1.5) 41.9 (1.3) 15.2* (1.8)   9.4 (1.0) (791) (1,294) II. High Income  Australia 77.3* (2.8) 52.4 (3.4) 24.9* (3.9)   7.1 (1.7) (229) (217)  Belgium 66.0 (12.6) 52.4 (9.6) 12.7 (5.2) 10.9 (5.0) (39) (67)  France 67.2 (8.4) 49.3 (7.9) 19.3* (6.3)   7.0 (2.8) (68) (89)  Germany 64.5 (9.6) 54.2 (8.7) 19.1* (6.2)   2.1 (1.5) (55) (54)  Israel 70.8* (6.6) 51.2 (3.3) 15.3* (5.5)   9.7 (2.0) (48) (232)  Italy 59.8 (11.8) 53.6 (8.9) 10.8* (4.4)   5.0 (2.4) (51) (67)  Japan 68.9* (9.4) 28.5 (6.1) 32.7* (11.0)   6.9 (3.6) (25) (56)  Netherlands 68.1 (14.7) 55.7 (9.4) 19.0 (7.2) 14.1 (9.0) (40) (81)  New Zealand 73.2* (2.3) 56.2 (2.6) 28.9* (3.3) 14.7 (2.3) (370) (365)  Northern Ireland 62.6* (4.6) 51.1 (4.8) 14.7* (3.7)   1.2 (0.9) (110) (109)  Poland 56.5* (8.4) 41.7 (5.3) 18.5* (7.0)   8.4 (3.3) (36) (87)  Portugal 59.3* (4.4) 42.1 (3.9) 13.5* (3.1)   8.7 (2.4) (126) (164)  Spain 61.8 (8.4) 52.6 (7.6)   6.8 (3.0)   7.3 (2.7) (87) (142)  Spain – Murcia 50.8* (7.5) 36.4 (4.6) 15.2* (6.0)   6.8 (2.9) (45) (109)  US 67.0* (2.4) 46.0 (3.0) 21.5* (2.4)   7.8 (1.7) (375) (271)  Total 68.8* (1.2) 49.6 (1.2) 21.3* (1.3)   8.7 (0.7) (1,704) (2,110) III. Total 64.4* (1.0) 46.0 (0.9) 19.5* (1.1)   8.9 (0.6) (2,495) (3,404) * Significant difference depending on whether comorbid anxiety disorders are present versus absent at the .05 level, two-sided test. a Anxiety disorders include generalized anxiety disorder, panic disorder with or without agoraphobia, agoraphobia without a history of panic disorder, specific phobia, social phobia, separation anxiety, and post-traumatic stress disorder b Ratings of severe or very severe on one or more Sheehan Disability Scale dimensions in the one month in the 12 before interview when the respondent’s MDD was most severe c At any time in the 12 months before interview d People’s Republic of China Table 5 Treatment of 12-month DSM-IV/CIDI major depressive disorder (MDD) in the presence versus absence of comorbid anxiety disordersa in the WHO World Mental Health Surveysb Any Specialty Psychiatricc General Medicalc Non-medicalc Present Absent Present Absent Present Absent Present Absent % (SE) % (SE) % (SE) % (SE) % (SE) % (SE) % (SE) % (SE) I. Low/Middle Income  Brazil – São Paulo 43.9* (4.5) 30.4 (3.5) 27.2 (4.0) 17.7 (2.8) 19.4 (3.7) 12.6 (2.6)   9.6 (2.9)   9.1 (2.3)  Bulgaria 39.7 (8.1) 27.8 (6.1) 13.9 (5.7)   8.7 (3.9) 34.6 (8.0) 22.7 (5.6)   0.5 (0.5)   0.9 (0.9)  Colombia 24.2 (5.9) 16.4 (3.7) 17.3 (5.6)   8.7 (2.7)   6.2 (2.2)   6.9 (2.5)   5.6 (3.9)   1.7 (1.1)  Colombia – Medellin 28.6 (6.1) 26.9 (6.1) 15.5 (4.9) 24.2 (5.9) 11.4 (4.1)   4.9 (2.8)   5.9 (2.9)   1.0 (1.0)  Iraq 16.5 (7.6)   8.4 (4.3)   7.0 (5.7)   5.8 (4.2)   2.6 (1.5)   0.5 (0.4)   7.0 (5.6)   2.1 (1.0)  Lebanon 18.2 (6.1) 26.0 (5.9)   7.4 (4.2) 13.0 (5.1) 11.8 (5.3) 13.1 (4.0)   2.0 (1.5)   1.8 (1.5)  Mexico 18.9 (4.6) 30.7 (4.7) 11.0 (4.0) 12.8 (3.4)   5.7* (1.9) 13.8 (3.5)   6.1 (3.6)   5.8 (2.2)  Nigeria 58.5* (18.1) 10.1 (5.7)   3.0 (3.2)   2.8 (2.1) 47.6* (18.1)   8.3 (5.5)   8.0 (8.2)   1.7 (1.8)  Peru 25.9 (7.4) 33.2 (6.6)   6.7 (4.2) 19.6 (5.3) 12.5 (5.5) 12.0 (4.9) 12.8 (5.7)   8.7 (3.8)  PRCd – Beijing/Shanghai 56.2* (18.5)   7.2 (4.4)   3.9 (4.1)   6.2 (4.4) 54.8* (18.9)   0.6 (0.6) 37.5 (24.1)   0.5 (0.5)  PRCd – Shenzhen 14.1 (7.0)   5.9 (2.3)   9.9 (6.4)   0.2 (0.2)   9.5 (6.4)   0.7 (0.4)   2.5 (2.6)   5.3 (2.2)  Romania 48.5 (19.7) 12.0 (6.4) 30.4 (19.8)   7.5 (5.5) 25.1 (13.2)   8.3 (5.9) – –   1.2 (1.3)  Total 30.3* (2.4) 20.6 (1.5) 16.1* (1.9) 11.2 (1.2) 14.0* (1.7)   8.6 (1.0)   7.5* (1.7)   4.6 (0.7) II. High Income  Australia 80.2* (2.6) 47.0 (3.4) 54.4* (3.3) 22.6 (2.8) 61.9* (3.2) 34.1 (3.2) 20.0* (2.6)   6.9 (1.7)  Belgium 61.4 (10.1) 49.3 (8.6) 34.1 (9.1) 34.4 (8.5) 49.4 (10.1) 40.5 (8.6)   9.1 (6.2)   2.6 (2.0)  France 62.2* (8.5) 36.4 (6.4) 28.7 (8.6) 14.7 (4.8) 49.7* (8.1) 29.3 (5.7)   0.5 (0.6)   2.8 (1.6)  Germany 51.6 (7.5) 43.8 (8.5) 39.2* (7.3) 17.6 (5.6) 21.2 (6.5) 33.8 (8.2) 10.4 (4.2)   5.9 (3.3)  Israel 55.8* (7.5) 35.5 (3.3) 36.4* (7.3) 19.7 (2.7) 35.0* (7.3) 14.8 (2.4) 15.7 (5.6)   8.4 (1.8)  Italy 64.3* (7.2) 20.2 (4.9) 21.8 (7.2)   9.9 (3.8) 61.4* (7.4) 16.8 (4.7)   3.1 (2.2)   2.2 (1.6)  Japan 60.0* (11.1) 32.4 (8.3) 46.6 (11.9) 20.4 (6.6) 18.1 (9.6)   9.7 (7.0) 34.0* (11.7)   7.9 (4.3)  Netherlands 77.4* (7.7) 36.8 (7.7) 52.7* (10.5) 17.9 (4.5) 63.8* (10.0) 27.3 (7.1) 10.3 (5.3)   3.7 (1.8)  New Zealand 64.7* (3.3) 53.4 (3.5) 30.9* (3.1) 21.7 (2.8) 46.8 (3.3) 42.1 (3.4) 19.9* (2.8)   9.9 (1.9)  Northern Ireland 61.3* (6.2) 41.0 (6.1) 25.2 (4.7) 11.4 (4.1) 59.6* (6.2) 38.2 (6.0)   9.4 (3.2)   6.8 (3.6)  Poland 49.8 (9.7) 36.2 (6.0) 33.2 (8.8) 25.3 (5.1) 28.7 (9.4) 16.6 (4.7) 12.0 (5.6)   7.7 (3.5)  Portugal 60.8 (4.9) 55.4 (4.4) 32.8 (4.6) 25.2 (3.9) 41.0 (4.9) 36.6 (4.3)   6.9 (2.4)   3.5 (1.7)  Spain 70.2* (6.1) 51.6 (5.6) 33.4 (6.5) 40.2 (5.3) 59.1** (6.6) 24.8 (4.3)   6.2 (4.0)   2.3 (1.0)  Spain – Murcia 73.9 (7.8) 61.8 (5.9) 56.9* (8.7) 33.6 (6.2) 21.0 (6.3) 33.0 (5.6)   4.9 (3.7)   1.1 (0.7)  US 63.5* (2.8) 48.6 (3.3) 37.7* (2.9) 26.0 (2.8) 38.0* (2.9) 26.4 (3.0) 19.1 (2.3) 14.9 (2.3)  Total 66.8* (1.5) 45.4 (1.4) 38.5* (1.6) 22.6 (1.2) 47.7* (1.6) 29.6 (1.3) 15.8* (1.2)   7.4 (0.7) III. Total 56.2* (3.1) 37.3 (2.6) 32.0* (3.0) 18.9 (2.1) 37.9* (3.1) 22.7 (2.3) 13.4* (2.2)   6.5 (1.3) * Significant difference depending on whether comorbid anxiety disorders are present versus absent at the .05 level, two-sided test. a Anxiety disorders include generalized anxiety disorder, panic disorder with or without agoraphobia, agoraphobia without a history of panic disorder, specific phobia, social phobia, and post-traumatic stress disorder b See Table 3 for denominator sample sizes c See the text for definitions of specialty, general medical, and non-medical treatments d People’s Republic of China CONFLICTS OF INTEREST: In the past three years, Dr. Kessler has been a consultant for Hoffman-La Roche, Inc., Johnson & Johnson Wellness and Prevention, and Sonofi-Aventis Groupe. Dr. Kessler has served on advisory boards for Mensante Corporation, Plus One Health Management, Lake Nona Institute, and U.S. Preventive Medicine. Dr. Kessler owns 25% share in DataStat, Inc. Dr. Wilcox is an employee of Janssen Pharmaceuticals. ETHICAL STANDARDS: The authors assert that all procedures contributing to this work comply with the ethical standards of the relevant national and institutional committees on human experimentation and with the Helsinki Declaration of 1975, as revised in 2008. Altemus M Sarvaiya N Neill Epperson C 2014 Sex differences in anxiety and depression clinical perspectives Frontiers in Neuroendocrinology 35 320 330 24887405 Andrade L Caraveo-Anduaga JJ Berglund P Bijl RV Dragomericka E Kohn R Keller MB Kessler RC Kawakami N Kilic C D O Ustun TB Vicente B Wittchen H 2003 The epidemiology of major depressive episodes: results from the International Consortium of Psychiatric Epidemiology (ICPE) Surveys International Journal of Methods in Psychiatric Research 12 3 21 12830306 Asselmann E Wittchen HU Lieb R Hofler M Beesdo-Baum K 2015 Danger and loss events and the incidence of anxiety and depressive disorders: a prospective-longitudinal community study of adolescents and young adults Psychological Medicine 45 153 163 25065411 Beesdo K Pine DS Lieb R Wittchen HU 2010 Incidence and risk patterns of anxiety and depressive disorders and categorization of generalized anxiety disorder Archives of General Psychiatry 67 47 57 20048222 Bernert S Matschinger H Alonso J Haro JM Brugha TS Angermeyer MC 2009 Is it always the same? Variability of depressive symptoms across six European countries Psychiatry Research 168 137 144 19481817 Birnbaum HG Kessler RC Kelley D Ben-Hamadi R Joish VN Greenberg PE 2010 Employer burden of mild, moderate, and severe major depressive disorder: mental health services utilization and costs, and work performance Depression and Anxiety 27 78 89 19569060 Bittner A Goodwin RD Wittchen HU Beesdo K Hofler M Lieb R 2004 What characteristics of primary anxiety disorders predict subsequent major depressive disorder? Journal of Clinical Psychiatry 65 618 626 quiz 730 15163247 Copeland WE Shanahan L Costello EJ Angold A 2009 Childhood and adolescent psychiatric disorders as predictors of young adult disorders Archives of General Psychiatry 66 764 772 19581568 Cuijpers P Sijbrandij M Koole S Huibers M Berking M Andersson G 2014 Psychological treatment of generalized anxiety disorder: a meta-analysis Clinical Psychology Review 34 130 140 24487344 Cummings CM Caporino NE Kendall PC 2014 Comorbidity of anxiety and depression in children and adolescents: 20 years after Psychological Bulletin 140 816 845 24219155 de Graaf R ten Have M Tuithof M van Dorsselaer S 2013 First-incidence of DSM-IV mood, anxiety and substance use disorders and its determinants: results from the Netherlands Mental Health Survey and Incidence Study-2 Journal of Affective Disorders 149 100 107 23399481 Fichter MM Quadflieg N Fischer UC Kohlboeck G 2010 Twenty-five-year course and outcome in anxiety and depression in the Upper Bavarian Longitudinal Community Study Acta Psychiatrica Scandinavica 122 75 85 19922523 Finlay-Jones R Brown GW 1981 Types of stressful life event and the onset of anxiety and depressive disorders Psychological Medicine 11 803 815 7323236 First M Spitzer R Gibbon M Williams J 2002 Structured Clinical Interview for DSM-IV Axis I Disorders, Research Version, Non-patient Edition (SCID-I/NP) Biometrics Research, New York State Psychiatric Institute Flannery-Schroeder EC 2006 Reducing anxiety to prevent depression American Journal of Preventive Medicine 31 S136 142 17175408 Garber J Weersing VR 2010 Comorbidity of Anxiety and Depression in Youth: Implications for Treatment and Prevention Clinical psychology: a publication of the Division of Clinical Psychology of the American Psychological Association 17 293 306 21499544 Goodwin R Olfson M 2001 Treatment of panic attack and risk of major depressive disorder in the community American Journal of Psychiatry 158 1146 1148 11431239 Goodwin RD Gorman JM 2002 Psychopharmacologic treatment of generalized anxiety disorder and the risk of major depression American Journal of Psychiatry 159 1935 1937 12411233 Hamdi NR Iacono WG 2014 Lifetime prevalence and co-morbidity of externalizing disorders and depression in prospective assessment Psychological Medicine 44 315 324 23590946 Harkness J Pennell BE Villar A Gebler N Aguilar-Gaxiola S Bilgen I 2008 Translation procedures and translation assessment in the World Mental Health Survey Initiative The WHO World Mental Health Surveys: Global perspectives on the epidemiology of mental disorders Kessler RC Üstün TB New York, NY Cambridge University Press 91 113 Haro JM Arbabzadeh-Bouchez S Brugha TS de Girolamo G Guyer ME Jin R Lepine JP Mazzi F Reneses B Vilagut G Sampson NA Kessler RC 2006 Concordance of the Composite International Diagnostic Interview Version 3.0 (CIDI 3.0) with standardized clinical assessments in the WHO World Mental Health surveys International Journal of Methods in Psychiatric Research 15 167 180 17266013 Hayward C Varady S Albano AM Thienemann M Henderson L Schatzberg AF 2000 Cognitive-behavioral group therapy for social phobia in female adolescents: results of a pilot study Journal of the American Academy of Child and Adolescent Psychiatry 39 721 726 10846306 Heeringa S Wells J Hubbard F Mneimneh Z Chiu W Sampson N 2008 Sample Designs and Sampling Procedures The WHO World Mental Health Surveys: Global Perspectives on the Epidemiology of Mental Disorders Kessler RC Üstün T Cambridge University Press Hofmann SG Smits JA 2008 Cognitive-behavioral therapy for adult anxiety disorders: a meta-analysis of randomized placebo-controlled trials Journal of Clinical Psychiatry 69 621 632 18363421 Jacobi F Wittchen HU Holting C Hofler M Pfister H Muller N Lieb R 2004 Prevalence, co-morbidity and correlates of mental disorders in the general population: results from the German Health Interview and Examination Survey (GHS) Psychological Medicine 34 597 611 15099415 Jakubovski E Bloch MH 2014 Prognostic subgroups for citalopram response in the STAR*D trial Journal of Clinical Psychiatry 75 738 747 24912106 Kelly JM Jakubovski E Bloch MH 2014 Prognostic subgroups for remission and response in the Coordinated Anxiety Learning and Management (CALM) trial Journal of Clinical Psychiatry Kendler KS Hettema JM Butera F Gardner CO Prescott CA 2003 Life event dimensions of loss, humiliation, entrapment, and danger in the prediction of onsets of major depression and generalized anxiety Archives of General Psychiatry 60 789 796 12912762 Kessler RC 1995 Epidemiology of psychiatric comorbidity Textbook in Psychiatric Epidemiology Tsuang MTMT Zahner GEP New York John Wiley & Sons, Inc 18 Kessler RC Keller MB Wittchen HU 2001 The epidemiology of generalized anxiety disorder Psychiatric Clinics of North America 24 19 39 11225507 Kessler RC Ormel J Petukhova M McLaughlin KA Green JG Russo LJ Stein DJ Zaslavsky AM Aguilar-Gaxiola S Alonso J Andrade L Benjet C de Girolamo G de Graaf R Demyttenaere K Fayyad J Haro JM Hu C Karam A Lee S Lepine JP Matchsinger H Mihaescu-Pintia C Posada-Villa J Sagar R Ustun TB 2011a Development of lifetime comorbidity in the World Health Organization world mental health surveys Archives of General Psychiatry 68 90 100 21199968 Kessler RC Petukhova M Zaslavsky AM 2011b The role of latent internalizing and externalizing predispositions in accounting for the development of comorbidity among common mental disorders Current Opinions in Psychiatry 24 307 312 Kessler RC Üstün TB 2004 The World Mental Health (WMH) Survey Initiative Version of the World Health Organization (WHO) Composite International Diagnostic Interview (CIDI) International Journal of Methods in Psychiatric Research 13 93 121 15297906 Klein DN Glenn CR Kosty DB Seeley JR Rohde P Lewinsohn PM 2013 Predictors of first lifetime onset of major depressive disorder in young adulthood Journal of Abnormal Psychology 122 1 6 22889243 Knäuper B Cannell CF Schwarz N Bruce ML Kessler RC 1999 Improving accuracy of major depression age-of-onset reports in the US National Comorbidity Survey International Journal of Methods in Psychiatric Research 8 39 48 Lamers F van Oppen P Comijs HC Smit JH Spinhoven P van Balkom AJ Nolen WA Zitman FG Beekman AT Penninx BW 2011 Comorbidity patterns of anxiety and depressive disorders in a large cohort study: the Netherlands Study of Depression and Anxiety (NESDA) Journal of Clinical Psychiatry 72 341 348 21294994 Leach LS Butterworth P Olesen SC Mackinnon A 2013 Relationship quality and levels of depression and anxiety in a large population-based survey Social Psychiatry and Psychiatric Epidemiology 48 417 425 22875222 Leon AC Olfson M Portera L Farber L Sheehan DV 1997 Assessing psychiatric impairment in primary care with the Sheehan Disability Scale International Journal of Psychiatry in Medicine 27 93 105 9565717 Li Y Aggen S Shi S Gao J Tao M Zhang K Wang X Gao C Yang L Liu Y Li K Shi J Wang G Liu L Zhang J Du B Jiang G Shen J Zhang Z Liang W Sun J Hu J Liu T Miao G Meng H Hu C Huang G Li G Ha B Deng H Mei Q Zhong H Gao S Sang H Zhang Y Fang X Yu F Yang D Chen Y Hong X Wu W Chen G Cai M Song Y Pan J Dong J Pan R Zhang W Shen Z Liu Z Gu D Liu X Zhang Q Flint J Kendler KS 2014 Subtypes of major depression: latent class analysis in depressed Han Chinese women Psychological Medicine 44 3275 3288 25065911 Lin CH Wang FC Lin SC Chen CC Huang CJ 2014 A comparison of inpatients with anxious depression to those with nonanxious depression Psychiatry Research 220 855 860 25262642 Mathew AR Pettit JW Lewinsohn PM Seeley JR Roberts RE 2011 Co-morbidity between major depressive disorder and anxiety disorders: shared etiology or direct causation? Psychological Medicine 41 2023 2034 21439108 McDowell RD Ryan A Bunting BP O’Neill SM Alonso J Bruffaerts R de Graaf R Florescu S Vilagut G de Almeida JM de Girolamo G Haro JM Hinkov H Kovess-Masfety V Matschinger H Tomov T 2014 Mood and anxiety disorders across the adult lifespan: a European perspective Psychological Medicine 44 707 722 23721650 McLaughlin TP Khandker RK Kruzikas DT Tummala R 2006 Overlap of anxiety and depression in a managed care population: Prevalence and association with resource utilization Journal of Clinical Psychiatry 67 1187 1193 16965195 McLean PD Woody S Taylor S Koch WJ 1998 Comorbid panic disorder and major depression: implications for cognitive-behavioral therapy Journal of Consulting and Clinical Psychology 66 240 247 9583327 Merikangas KR Jin R He JP Kessler RC Lee S Sampson NA Viana MC Andrade LH Hu C Karam EG Ladea M Medina-Mora ME Ono Y Posada-Villa J Sagar R Wells JE Zarkov Z 2011 Prevalence and correlates of bipolar spectrum disorder in the world mental health survey initiative Archives of General Psychiatry 68 241 251 21383262 Mineka S Vrshek-Schallhorn S 2008 Comorbidity of Unipolar Depressive and Anxiety Disorders Handbook of Depression Gotlib IH Hammen Constance L Guilford Press Moffitt TE Caspi A Harrington H Milne BJ Melchior M Goldberg D Poulton R 2007 Generalized anxiety disorder and depression: childhood risk factors in a birth cohort followed to age 32 Psychological Medicine 37 441 452 17201999 Moffitt TE Caspi A Taylor A Kokaua J Milne BJ Polanczyk G Poulton R 2010 How common are common mental disorders? Evidence that lifetime prevalence rates are doubled by prospective versus retrospective ascertainment Psychological Medicine 40 899 909 19719899 Murphy JM Olivier DC Sobol AM Monson RR Leighton AH 1986 Diagnosis and outcome: depression and anxiety in a general population Psychological Medicine 16 117 126 3515377 Ormel J Petukhova M Chatterji S Aguilar-Gaxiola S Alonso J Angermeyer MC Bromet EJ Burger H Demyttenaere K de Girolamo G Haro JM Hwang I Karam E Kawakami N Lepine JP Medina-Mora ME Posada-Villa J Sampson N Scott K Ustun TB Von Korff M Williams DR Zhang M Kessler RC 2008 Disability and treatment of specific mental and physical disorders across the world British Journal of Psychiatry 192 368 375 18450663 Ormel J VonKorff M Ustun TB Pini S Korten A Oldehinkel T 1994 Common mental disorders and disability across cultures. Results from the WHO Collaborative Study on Psychological Problems in General Health Care Journal of the American Medical Association 272 1741 1748 7966922 Parker G Brotchie H 2010 Gender differences in depression International Review of Psychiatry 22 429 436 21047157 Pennell B Mneimneh Z Bowers A Chardoul S Wells J Viana M Dinkelmann K Gebler N Florescu S He Y Huang Y Tomov T Vilagut G 2008 Implementation of the World Mental Health Surveys The WHO World Mental Health Surveys: Global Perspectives on the Epidemiology of Mental Disorders Kessler R Ustun T Cambridge, UK Cambridge University Press 33 57 Rapee RM Lyneham HJ Hudson JL Kangas M Wuthrich VM Schniering CA 2013 Effect of comorbidity on treatment of anxious children and adolescents: results from a large, combined sample Journal of the American Academy of Child and Adolescent Psychiatry 52 47 56 23265633 Research Triangle Institute 2002 SUDAAN: Professional Software for Survey Data Analysis [computer program] Research Triangle Park, NC Research Triangle Institute Roy-Byrne PP Stang P Wittchen HU Ustun B Walters EE Kessler RC 2000 Lifetime panic-depression comorbidity in the National Comorbidity Survey. Association with symptoms, impairment, course and help-seeking British Journal of Psychiatry 176 229 235 10755069 Saveanu R Etkin A Duchemin AM Goldstein-Piekarski A Gyurak A Debattista C Schatzberg AF Sood S Day CV Palmer DM Rekshan WR Gordon E Rush AJ Williams LM 2014 The International Study to Predict Optimized Treatment in Depression (iSPOT-D): Outcomes from the acute phase of antidepressant treatment Journal of Psychiatric Research Schrier AC de Wit MA Rijmen F Tuinebreijer WC Verhoeff AP Kupka RW Dekker J Beekman AT 2010 Similarity in depressive symptom profile in a population-based study of migrants in the Netherlands Social Psychiatry and Psychiatric Epidemiology 45 941 951 19763363 Scott KM Wells JE Angermeyer M Brugha TS Bromet E Demyttenaere K de Girolamo G Gureje O Haro JM Jin R Karam AN Kovess V Lara C Levinson D Ormel J Posada-Villa J Sampson N Takeshima T Zhang M Kessler RC 2010 Gender and the relationship between marital status and first onset of mood, anxiety and substance use disorders Psychological Medicine 40 1495 1505 19939327 Shippee ND Rosen BH Angstman KB Fuentes ME DeJesus RS Bruce SM Williams MD 2014 Baseline screening tools as indicators for symptom outcomes and health services utilization in a collaborative care model for depression in primary care: a practice-based observational study General Hospital Psychiatry 36 563 569 25179215 Simon GE Goldberg DP Von Korff M Ustun TB 2002 Understanding cross-national differences in depression prevalence Psychological Medicine 32 585 594 12102373 Stiles-Shields C Kwasny MJ Cai X Mohr DC 2014 Comorbid anxiety as a differential treatment predictor for telephone versus face-to-face administered cognitive behavioral therapy for depression Depression and Anxiety 31 934 940 25407583 Takayanagi Y Spira AP Roth KB Gallo JJ Eaton WW Mojtabai R 2014 Accuracy of reports of lifetime mental and physical disorders: results from the Baltimore Epidemiological Catchment Area study JAMA Psychiatry 71 273 280 24402003 van Loo HM de Jonge P Romeijn JW Kessler RC Schoevers RA 2012 Data-driven subtypes of major depressive disorder: a systematic review BMC Medicine 10 156 23210727 Vos T Flaxman AD Naghavi M Lozano R Michaud C Ezzati M Shibuya K Salomon JA Abdalla S Aboyans V Abraham J Ackerman I Aggarwal R Ahn SY Ali MK Alvarado M Anderson HR Anderson LM Andrews KG Atkinson C Baddour LM Bahalim AN Barker-Collo S Barrero LH Bartels DH Basanez MG Baxter A Bell ML Benjamin EJ Bennett D Bernabe E Bhalla K Bhandari B Bikbov B Bin Abdulhak A Birbeck G Black JA Blencowe H Blore JD Blyth F Bolliger I Bonaventure A Boufous S Bourne R Boussinesq M Braithwaite T Brayne C Bridgett L Brooker S Brooks P Brugha TS Bryan-Hancock C Bucello C Buchbinder R Buckle G Budke CM Burch M Burney P Burstein R Calabria B Campbell B Canter CE Carabin H Carapetis J Carmona L Cella C Charlson F Chen H Cheng AT Chou D Chugh SS Coffeng LE Colan SD Colquhoun S Colson KE Condon J Connor MD Cooper LT Corriere M Cortinovis M de Vaccaro KC Couser W Cowie BC Criqui MH Cross M Dabhadkar KC Dahiya M Dahodwala N Damsere-Derry J Danaei G Davis A De Leo D Degenhardt L Dellavalle R Delossantos A Denenberg J Derrett S Des Jarlais DC Dharmaratne SD Dherani M Diaz-Torne C Dolk H Dorsey ER Driscoll T Duber H Ebel B Edmond K Elbaz A Ali SE Erskine H Erwin PJ Espindola P Ewoigbokhan SE Farzadfar F Feigin V Felson DT Ferrari A Ferri CP Fevre EM Finucane MM Flaxman S Flood L Foreman K Forouzanfar MH Fowkes FG Franklin R Fransen M Freeman MK Gabbe BJ Gabriel SE Gakidou E Ganatra HA Garcia B Gaspari F Gillum RF Gmel G Gosselin R Grainger R Groeger J Guillemin F Gunnell D Gupta R Haagsma J Hagan H Halasa YA Hall W Haring D Haro JM Harrison JE Havmoeller R Hay RJ Higashi H Hill C Hoen B Hoffman H Hotez PJ Hoy D Huang JJ Ibeanusi SE Jacobsen KH James SL Jarvis D Jasrasaria R Jayaraman S Johns N Jonas JB Karthikeyan G Kassebaum N Kawakami N Keren A Khoo JP King CH Knowlton LM Kobusingye O Koranteng A Krishnamurthi R Lalloo R Laslett LL Lathlean T Leasher JL Lee YY Leigh J Lim SS Limb E Lin JK Lipnick M Lipshultz SE Liu W Loane M Ohno SL Lyons R Ma J Mabweijano J MacIntyre MF Malekzadeh R Mallinger L Manivannan S Marcenes W March L Margolis DJ Marks GB Marks R Matsumori A Matzopoulos R Mayosi BM McAnulty JH McDermott MM McGill N McGrath J Medina-Mora ME Meltzer M Mensah GA Merriman TR Meyer AC Miglioli V Miller M Miller TR Mitchell PB Mocumbi AO Moffitt TE Mokdad AA Monasta L Montico M Moradi-Lakeh M Moran A Morawska L Mori R Murdoch ME Mwaniki MK Naidoo K Nair MN Naldi L Narayan KM Nelson PK Nelson RG Nevitt MC Newton CR Nolte S Norman P Norman R O’Donnell M O’Hanlon S Olives C Omer SB Ortblad K Osborne R Ozgediz D Page A Pahari B Pandian JD Rivero AP Patten SB Pearce N Padilla RP Perez-Ruiz F Perico N Pesudovs K Phillips D Phillips MR Pierce K Pion S Polanczyk GV Polinder S Pope CA 3rd Popova S Porrini E Pourmalek F Prince M Pullan EL Ramaiah KD Ranganathan D Razavi H Regan M Rehm JT Rein DB Remuzzi G Richardson K Rivara FP Roberts T Robinson C De Leon FR Ronfani L Room R Rosenfeld LC Rushton L Sacco RL Saha S Sampson U Sanchez-Riera L Sanman E Schwebel DC Scott JG Segui-Gomez M Shahraz S Shepard DS Shin H Shivakoti R Singh D Singh GM Singh JA Singleton J Sleet DA Sliwa K Smith E Smith JL Stapelberg NJ Steer A Steiner T Stolk WA Stovner LJ Sudfeld C Syed S Tamburlini G Tavakkoli M Taylor HR Taylor JA Taylor WJ Thomas B Thomson WM Thurston GD Tleyjeh IM Tonelli M Towbin JA Truelsen T Tsilimbaris MK Ubeda C Undurraga EA van der Werf MJ van Os J Vavilala MS Venketasubramanian N Wang M Wang W Watt K Weatherall DJ Weinstock MA Weintraub R Weisskopf MG Weissman MM White RA Whiteford H Wiersma ST Wilkinson JD Williams HC Williams SR Witt E Wolfe F Woolf AD Wulf S Yeh PH Zaidi AK Zheng ZJ Zonies D Lopez AD Murray CJ AlMazroa MA Memish ZA 2012 Years lived with disability (YLDs) for 1160 sequelae of 289 diseases and injuries 1990–2010: a systematic analysis for the Global Burden of Disease Study 2010 Lancet 380 2163 2196 23245607 Wang PS Aguilar-Gaxiola S Alonso J Angermeyer MC Borges G Bromet EJ Bruffaerts R de Girolamo G de Graaf R Gureje O Haro JM Karam EG Kessler RC Kovess V Lane MC Lee S Levinson D Ono Y Petukhova M Posada-Villa J Seedat S Wells JE 2007 Use of mental health services for anxiety, mood, and substance disorders in 17 countries in the WHO world mental health surveys Lancet 370 841 850 17826169 Wolter KM 1985 Introduction to Variance Estimation New York, NY Springer-Verlag Woody S McLean PD Taylor S Koch WJ 1999 Treatment of major depression in the context of panic disorder Journal of Affective Disorders 53 163 174 10360411 World Bank 2009 Data & Statistics, Country Groups by Income from http://go.worldbank.org/D7SN0B8YU0 Wu Z Fang Y 2014 Comorbidity of depressive and anxiety disorders: challenges in diagnosis and assessment Shanghai Arch Psychiatry 26 227 231 25317009

PMC005xxxxxx/PMC5129629.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 8111104 1254 Breast Cancer Res Treat Breast Cancer Res. Treat. Breast cancer research and treatment 0167-6806 1573-7217 27283835 5129629 10.1007/s10549-016-3847-3 NIHMS794571 Article Breast cancers from black women exhibit higher numbers of immunosuppressive macrophages with proliferative activity and of crown-like structures associated with lower survival compared to non-black Latinas and Caucasians Koru-Sengul Tulay 15 Santander Ana M. 2 Miao Feng 5 Sanchez Lidia G. 2 Jorda Merce 35 Glück Stefan 4512 Ince Tan A. 35 Nadji Mehrad 35 Chen Zhibin 25 Penichet Manuel L 678910 Cleary Margot P. 11 Torroella-Kouri Marta 215 1 Department of Public Health Sciences, University of Miami Miller School of Medicine, Miami, FL, USA 2 Department of Microbiology and Immunology, University of Miami Miller School of Medicine, 1600 NW 10th Avenue Rosenstiel Medical School Building Suite 3123A, P.O. Box 016960 (R-138), Miami, FL 33101, USA 3 Department of Pathology, University of Miami Miller School of Medicine, Miami, FL, USA 4 Department of Medicine, University of Miami Miller School of Medicine, Miami, FL, USA 5 Sylvester Comprehensive Cancer Center, University of Miami Miller School of Medicine, Miami, FL, USA 6 Division of Surgical Oncology, Department of Surgery, UCLA, Los Angeles, CA, USA 7 Department of Microbiology, Immunology, and Molecular Genetics, David Geffen School of Medicine at UCLA, UCLA, Los Angeles, CA, USA 8 Jonsson Comprehensive Cancer Center, UCLA, Los Angeles, CA, USA 9 UCLA AIDS Institute, UCLA, Los Angeles, CA, USA 10 The Molecular Biology Institute, UCLA, Los Angeles, CA, USA 11 Hormel Institute, University of Minnesota, Austin, MN 55912, USA 12 Present Address: Celgene Corporation, Summit, NJ, USA Tulay Koru-Sengul, Ana M. Santander share first authorship. Marta Torroella-Kouri mtorroella@med.miami.edu 9 7 2016 9 6 2016 7 2016 01 7 2017 158 1 113126 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. Racial disparities in breast cancer incidence and outcome are a major health care challenge. Patients in the black race group more likely present with an early onset and more aggressive disease. The occurrence of high numbers of macrophages is associated with tumor progression and poor prognosis in solid malignancies. Macrophages are observed in adipose tissues surrounding dead adipocytes in “crown-like structures” (CLS). Here we investigated whether the numbers of CD163+ tumor-associated macrophages (TAMs) and/or CD163+ CLS are associated with patient survival and whether there are significant differences across blacks, non-black Latinas, and Caucasians. Our findings confirm that race is statistically significantly associated with the numbers of TAMs and CLS in breast cancer, and demonstrate that the highest numbers of CD163+ TAM/CLS are found in black breast cancer patients. Our results reveal that the density of CD206 (M2) macrophages is a significant predictor of progression-free survival univariately and is also significant after adjusting for race and for HER2, respectively. We examined whether the high numbers of TAMs detected in tumors from black women were associated with macrophage proliferation, using the Ki-67 nuclear proliferation marker. Our results reveal that TAMs actively divide when in contact with tumor cells. There is a higher ratio of proliferating macrophages in tumors from black patients. These findings suggest that interventions based on targeting TAMs may not only benefit breast cancer patients in general but also serve as an approach to remedy racial disparity resulting in better prognosis patients from minority racial groups. Breast cancer Race/ethnicities Macrophages Crown-like structures Inflammation Introduction Breast cancer is the second leading cause of cancer mortality in women in the United States (US) [1]. Disparities in cancer incidence and outcome due to ethnicity and race-related differences are a major health care challenge [2]. Even after considering socio-economic factors, education and access to health care, African-American (AA) women still exhibit breast cancer health disparities in incidence and outcome [3, 4]. These patients present with a particular pattern of breast cancer pathogenesis, i.e., early onset, higher incidence in younger women, and more aggressive disease with less favorable prognosis. A number of biological differences may account for the nature of this disease in AA as compared to Caucasian (CA) women [5–11]. Some of these differences may be inheritable; inherited predisposition to breast cancer has been shown among AA women [12]. However, whether and how immune inflammatory components in the tumor microenvironment (TME) may correlate with the aggressiveness of breast cancer in AA women remains unclear. The occurrence of high numbers of macrophages in the TME has been associated with tumor progression and poor tumor prognosis in breast and other solid malignancies [13, 14]. Pro-inflammatory (M1) TAMs can be cytotoxic to tumor cells [15, 16]. However, M1 TAMs can also contribute to tumor initiation through the mutagenic reactive oxygen and nitrogen species they generate [17, 18]. TAMs further impact tumor progression by promoting invasion, angiogenesis, and metastasis when in their immunosuppressive (M2) mode [19–21]. Actually, mixtures of pro-inflammatory M1 and immunosuppressive M2 macrophages co-exist in the TME of advanced mouse and human solid malignancies [22–26]. In contrast, high numbers of M1 macrophages colonize obese adipose tissues, providing obese adipose tissues a pro-inflammatory condition, while M2 macrophages are found in lean adipose tissues [27]. Macrophages are also observed in visceral adipose tissues of obese humans and mice surrounding dying adipocytes, forming particular structures known as “crown-like structures” (CLS) [28]. In the breast TME, macrophages are observed as isolated or grouped TAMs and also in CLS within the breast adipose tissue. Increased numbers of breast CLS have been associated with enhanced aromatase expression/activity and with inflammation in mammary tumors of obese mice and in obese women with breast cancer [29–31]. Aromatase converts androgens to estrogens [32] and promotes ER+ cell growth [33–35]; importantly, ER+ breast cancers are the most prevalent breast malignancies [1]. Using samples from an archival breast cancer tumor bank with cases obtained from 1978 to 1997 at the University of Miami/Jackson Memorial Hospital in Miami, Florida, we previously investigated in a pilot study (30 cases) whether the presence of CD68+ TAMs was an independent prognostic factor in small T1 ER+ breast cancers across three different racial groups [blacks (BL), non-black Latinas (NBLA), and CA women [36]. In Miami, where these cases were recruited and treated, there are many black Islanders, Caribbean, Haitians, South Americans (Brazilians and others), who do not identify themselves as AA; additionally, there are North Africans and South Africans of dark or light black skin color and these also are not AA. This led us to classify the patients from this archival bank on the basis of their race, not necessarily of their ethnicity. We demonstrated that TAM numbers were significantly higher in tumors from BL and NBLA than in CA patients [36]. These findings led us to design the present study, in which we used a larger number of tumors of all types and stages from the same archival bank and a different pan macrophage marker, CD163. CD163 is more sensitive than CD68 in detecting macrophages [37–39]. We also examined whether TAMs and CLS are associated with survival in these racial groups, and whether there are any differences in TAMs proliferation. Furthermore, we aimed to correlate the density and M1/M2 phenotypes of TAMs and also of CLS with clinical pathologic characteristics of the tumors across the three different racial groups. Materials and methods Case selection One hundred fifty (150) consecutive cases (50 BL, 50 NBLA, and 50 CA) of women treated for breast cancer between 1978 and 1997 at Jackson Memorial Hospital (JMH) and University of Miami's Sylvester Comprehensive Cancer Center (UM/SCCC) in Miami, Florida, were selected. Tumors were formalin-fixed paraffin-embedded (FFPE) specimens from the Cancer Center's Tumor Bank Core Facility. Tumors of any size and any ER and PR status were included. Patient inclusion criteria of samples were any stages, any hormonal receptor, and HER2 receptor. These cases were followed up for at least 5 years. Patient exclusion criteria included: the presence of previous cancers, exposure to previous chemotherapy, radio-therapy or hormonal therapy, or presenting bilateral or multifocal breast cancers. Patients were not treated with systemic anti-cancer therapy at the moment of the tumor sample collection in the surgery, and all (if any) systemic therapy was delivered post-surgery. Information concerning patient demographics, clinical characteristics, pathologic reports, and administered treatments was gathered from both UM/Jackson Memorial Hospital's Tumor Registry and medical records and UM/SCCC. The characteristics of patients and their tumors (de-identified) by race are described in Supplementary Table 1. The study was approved by the University of Miami's Institutional Review Board (IRB). Immunohistochemistry (IHC) IHC was carried out and assessed as briefly described in the legends of Figs. 1, 2 and 3 and as previously published [36]. Primary antibodies used were CD163 (pan macrophage marker), CD206 (M2 marker), CD40 (M1 marker), and Ki-67 (proliferation marker). Qualitative and quantitative analysis of TAMs and CLS densities was assessed blindly and independently by two investigators (MJ and AMS), including a breast pathologist (MJ). Photomicrographs and measurement of macrophages were conducted blind to patient identity. Macrophage proliferation was independently assessed by a breast pathologist (TAI) and two other study investigators (LS and AMS). Statistical analysis The clinical data were summarized by mean, standard deviation (SD) for continuous variables, and by frequencies and percentages for categorical variables for overall sample as well as by race (BL, NBLA, CA). Differences in means were tested by either Student's t test or one-way ANOVA with Bonferroni correction for multiple comparisons. Differences in proportions were tested by Chi-square or Fisher's exact test for categorical variables. Overall survival (OS) time is calculated as the elapsed time between the dates of diagnosis and death from any cause or last follow-up for alive patients. Progression-free survival (PFS) time is calculated as the elapsed time between the dates of diagnosis and earliest progression (local recurrence or distant metastasis or death) or last follow-up for patients without progression. Median survival and survival rates for both OS and PFS at 12, 24, 36, 60, 96, and 120 months were calculated by Kaplan–Meier method for all patients as well as by race. Log-rank test was used to test the differences in survival between the groups. Unadjusted and adjusted hazard ratio (HR), its 90 % confidence interval (90 %CI) along with p value were calculated from fitting several univariate and bivariate Cox proportional hazard regression models to identify significant predictors of OS and PFS, respectively. Race, ER, PR, HER2, CD163, CD206, and CD40 included one at a time in the univariate models. The predictors for bivariate models included paired variables between each of conventional breast cancer markers (ER, PR, and HER2) with each of macrophages surface markers (CD163, CD206, and CD40), and between race and each marker (ER, PR, HER2, CD163, CD206, and CD40). Patients with unknown ER or PR were excluded from all regression models. Due to the exploratory nature of the study, we set the type-I error rate as 10 % for all analysis, where p values <0.10 were considered as statistically significant. Statistical analyses were performed by SAS v9.4 (SAS Institute Inc., Cary, NC, USA). Results Higher densities of tumor-associated macrophages (TAMs) in breast tumors from black patients compared to non-black Latino and Caucasian patients Our IHC staining procedures successfully identified CD163+ TAMs in breast tumor sections from the archived samples (Fig. 1a). We calculated the densities of TAMs (cells/mm2) in the different tumors. We found significantly different occurrences of TAMs in breast cancers across these races with BL showing the highest numbers and CA the lowest (Fig. 1b). As detailed in Supplementary Table-2A, densities of TAMs are significantly different among the three ethnicities, with tumors from BL patients presenting with the highest densities of CD163+ TAMs (mean = 142.21 cells/mm2), followed by tumors from NBLA (110.16 cells/mm2), with tumors from CA showing the lowest TAM densities (62.72 cells/mm2). With Bonferroni multiple comparison analysis, both BL and NBLA tumors exhibited significantly higher TAMs densities than CA (p < 0.0001, Supplementary Table-2A). Higher densities of immunosuppressive M2 macrophages predominate in breast tumors from black patients compared to non-black Latino and Caucasian patients To assess the pro-inflammatory (M1) vs. immunosuppressive (M2) profiles of TAMs, CD40+ (M1) and CD206+ (M2) macrophage densities were estimated (Supplementary Table-2B). As shown in Fig. 2a, tumors from BL patients contained the highest estimated densities of CD163+ total TAM (p = 0.0009). For all three races, the majority of TAMs were immunosuppressive M2, with BL showing the highest densities of M2, followed by NBLA (Fig. 2b; Supplementary Table-2B, p = 0.0238). In contrast, pro-inflammatory CD40 (M1) macrophages were detected at the lowest densities in tumors from NBLA patients (2 %), followed by BL (12.5 %), with CA showing the highest (56.3 %) (Fig. 2c; Supplementary Table-2B, p < 0.0001). TAMs actively proliferate when in contact with breast tumor cells TAMs have been recently shown to proliferate [40], thus we investigated the proliferative status of TAMs across different races. We used double staining IHC (CD163+/Ki-67+) to detect dividing TAMs. Ki-67 is a nuclear protein associated with proliferation [41]; during interphase, the Ki-67 antigen can be exclusively detected within the cell nucleus, whereas in mitosis, most of the protein is relocated to the surface of the chromosomes. Therefore, we characterized proliferating macrophages as (a) CD163+ macrophages expressing Ki-67 in the nucleus in the absence of mitotic figures, suggestive of non-mitotic phases of the cell cycle and (b) CD163+ macrophages with mitotic condensed chromatin, indicative of active proliferation. As shown in Fig. 3a, CD163+ TAMs not adjacent to tumor cells are not Ki-67+. However, TAMs in contact with tumor cells in the TME exhibited Ki-67 staining (Fig. 3b). Higher proliferative capacity for TAMs in breast tumors from black patients compared to non-black Latino and Caucasian patients To analyze whether any differences in the proliferative capacity of TAMs could be detected across races, we randomly selected 23 cases representatives of the three racial groups studied and blindly assessed TAMs’ proliferation. The majority of the samples with high TAM proliferation activity (CD163+/Ki-67+ cells with or without mitotic chromatin) were tumors from BL patients, with NBLA and CA showing lower proliferative activity (Fig. 4a, b). Supplementary Table-3A shows the calculated densities of TAMs (total, CD163+ and proliferating, CD163+/Ki-67+), as well as the average numbers of CD163+/Ki-67+ proliferating TAMs in the three races within these 23 cases when all CD163+/Ki-67+ TAMs were considered. Supplementary Table-3B is similar to Table-3A except that it shows only actively dividing TAMs, exclusively expressing Ki-67 in mitotic figures. The calculated densities of total number of TAMs (CD163+) and of total dividing TAMs (CD163+/Ki-67+) in these 23 patients are shown in Fig. 4c and d, respectively; Fig. 4e depicts the proliferating ratios of TAMs across the three races, which are significantly different. When all proliferating TAMs are considered (Supplementary Table-3A), the ratios resulting from dividing the density of proliferating TAMs by the density of total number of TAMs (proliferation ratio) are approximately 1:18 (or 0.05) for NBLA, 1:12 (0.091) for CA and 1:10 (0.095) for BL. When the comparison is done considering only actively dividing TAMs (Supplementary Table-3B), the ratios are 1:56 (0.018) for NBLA, 1:59 (0.017) for CA and 1:23 (0.044) for BL, indicating a significantly higher proliferative capacity for TAMs in tumors from BL than in the other two races (p = 0.0353). Importantly, when actively proliferating TAMs were only considered (Supplementary Table-3B), the average numbers of proliferating TAMs as well as the density of proliferating TAMs were all statistically significantly different across the races. Even within the small group of n = 23 patients used in the TAMs proliferation study, the calculated densities of CD163+ TAMs show distributions similar to those in the large sample of 145 patients (Supplementary Table-2A, p < 0.001), with statistically significant differences across the three racial groups, revealing that CA patients exhibit on average significantly less (mean = 169.0) TAM density than BL (mean = 236.9) (Bonferroni multiple comparison analysis, p = 0.0701) (Supplementary Table-3A, B). Density of CD163+ TAMs (cells/mm2), average number of cells of CD163/Ki67+, and density of CD163/Ki67+ were summarized by ER, PR, and HER2 status (Supplementary Tables 4–6, respectively). There were no significant differences observed except the mean density of CD163+ (cells/mm2) of ER (p = 0.0269) and PR (p = 0.0722) positive was less than ER and PR negative patients. Crown-like structures (CLS), like TAMs, are differently expressed in the breast cancer microenvironment across diverse races Since the numbers and inflammatory profiles of TAMs present differently in breast cancers across three different racial groups of women, we examined whether macrophages within CLS in the adipose tissues of these tumors would follow a similar expression pattern. To this aim, we inspected the breast adipose tissues in tumors for CD163+ CLS using the same sections analyzed for CD163+ TAMs and calculated the densities of these structures in the different sections. Moreover, the M1 or M2 phenotypes of the macrophages within these CLS were investigated. Supplementary Table-7 reveals that, as expected, the frequency of CLS is much lower than the frequency of TAMs, because CLS are structures only present in the fat tissue of the breast TME, whereas TAMs occur throughout the entire TME. Supplementary Table-7 also shows a signifi-cant difference in the densities of CLS (p = 0.0167) among BL, NBLA, and CA patients, with BL exhibiting significantly higher densities than CA, and NBLA being in between, as happened with TAMs. However, the densities of neither CD40 nor CD206 CLS were significantly different among the three racial groups. Figure 5a shows representative IHC showing staining of CLS (CD163+ macrophages surrounding dying adipocytes) in breast cancers from patients of the three different races. These results are summarized in Fig. 5b, c, and d. Overall survival (OS) and progression-free survival (PFS) analysis OS and PFS curves were similar across three racial groups (Supplementary Figure 1). However, survival rates for BL patients were lower than NBLA and CA patients for both OS and PFS, and survival rates for NBLA were lower than for CA (Supplementary Table-8). Surprisingly, densities of CLS with CD40+ macrophages were significant predictors of OS (Table 1). The higher the density of CLS for CD40, the worse the OS was (HR = 12.15; p = 0.058) in univariate analysis, and in bivariate analysis after adjusting the effect of race (HR = 9.14; p = 0.100), PR status (HR = 17.43; p = 0.036), or HER2 status (HR = 13.59; p = 0.047). On the other hand, a higher estimated density of M2 TAMs (CD206+) was a predictor for lower PFS (HR = 1.65; p = 0.019) in the univariate model (Table 2). In addition, estimated densities of CD206+ M2 TAMs were significant predictors of PFS after adjusted with race in the bivariate models (HR = 1.55; p = 0.056). Furthermore, a higher density CD206 was associated with worse PFS (HR = 1.59; p = 0.031) after adjusting the effect for HER2 in the bivariate model. Discussion Here, we provide evidence that the breast TME of BL women contains significantly higher numbers of TAMs compared to the other two races studied. We show that these are mainly tumor-promoting, M2, immunosuppressive macrophages with higher proliferative capacity, as compared with tumors from NBLA and CA women. Our results offer another potential explanation for the aggressiveness of breast cancers among BL women. To the best of our knowledge, our work is one of the first to report the existence of statistically significant differences in the numbers, densities, activation profiles, and proliferative capacity of TAMs and of CLS among breast cancer patients of different races. We found that tumors were mainly populated by immunosuppressive M2 TAMs (CD206+), as previously reported in other solid malignancies and specifically in breast cancer [42, 43]; M2 TAMs have been associated with tumor invasion, angiogenesis, and metastasis [44–46]. Interestingly, the highest M2 densities were found in tumors from BL women, followed by tumors from NBLA, with CA exhibiting the lowest M2 densities. In contrast, tumors were in general less populated by M1 pro-inflammatory TAMs (CD40+), and interestingly, the highest M1 densities were observed in tumors from CA, followed by tumors from BL, with tumors from NBLA showing the lowest densities of M1 TAMs. We further showed that tumors from BL patients exhibit significantly higher densities of CLS than those from NBLA, with CA having the lowest densities, paralleling the situation with TAM. Previously, Morris et al. [29] found increased numbers of CLS in breast cancers from obese women. Due to an absence of height information, we were not able to calculate body mass index (BMI). It has been reported that obese individuals have CLS with M1 macrophages in their visceral adipose tissues [27]. In our study, we assessed the inflammatory profiles of macrophages within CLS in the fat portions of the tumor sections, and even though we saw no statistically significant differences across races for both M2 (CD206+) or M1 (CD40+) CLS, we did observe a trend for a prevalence of an M2 immunosuppressive CLS phenotype in all three groups, and a trend that showed a density of CD206+ M2 CLS that decreased from highest in BL to the lowest in CA (BL > NBLA > CA). Therefore, our results suggest that the patients in our study were unlikely to have been obese because a predominance of M2 CLS and not of M1 CLS was observed in all tumors. The high numbers of CLS—and thus, the amount of breast fat inflammation—observed in the breast tumors of BL patients was more likely associated with their biological race rather than potential obesity. Interestingly though, the only tumors showing few M1 CLS were from BL patients. Importantly, the high numbers of CLS in BL parallels the high numbers of immunosuppressive TAMs in the same tumors. Nevertheless, future studies are needed wherein not only races but also particular ethnicities within races and patient's BMI information will be gathered, to corroborate our findings in a larger Florida-based population, as well as the patient population in all states and worldwide. Macrophages have been considered terminally differentiated cells without proliferative capacity, thought to originate exclusively by replenishment and differentiation from blood monocytes. However, evidence on the diverse ontogeny and proliferative capacity of resident versus inflammatory macrophages has recently emerged [40, 47]. These new data demonstrate the embryonal origin of the majority of tissue resident macrophages with self-renewal capabilities, in contrast to macrophages recruited to the tissues after a pathogenic or damaging insult, which in the majority of the cases seem to differentiate from blood monocytes and lack proliferative capacity [48]. Proliferation of macrophages has now been identified in various settings, such as in nematode-infected tissues, obese adipose tissues, glomerulonephritis, atherosclerosis, AIDS-related dementia, and in a variety of murine tumors, when the resident versus inflammatory origin (ontogeny) of these macrophages is in many cases uncertain [49, 50]. Proliferating macrophages have also been recently recognized in human lymphomas and in breast tumors [51–54]. We identified Ki-67+ nuclei in various stages of the cell cycle in CD163+ TAMs. Interestingly, macrophage proliferation was only observed when they were in direct contact with tumor cells and not when they were in contact with stromal cells of the TME, suggesting that cytokines or other factors produced by tumor cells could be acting in paracrine fashion stimulating macrophage proliferation. Of relevance, when assessing ratios of proliferating vs. total number of TAMs across the three races, our data also strongly demonstrate a significantly higher active (mitotic) proliferation in TAMs from BL tumors compared with TAMs from the other two races. A recent study on young BL breast cancer patients reports that the prevalence of BRCA mutations among a Florida-based sample of young black women with breast cancer exceeds that previously reported for non-Hispanic white women [55]. However, it is becoming increasingly clear that some of the differences in cancer risk, incidence and survival among individuals of different racial and ethnic backgrounds can be attributed to biological factors other than the inheritance of predisposing tumor suppressor genes [3]. We propose that significant differences in the cellular and molecular components of the TMEs do exist across various races and may contribute to these differences. Future studies are needed to determine if there is a difference in inflammatory gene expression profiles of the TME across these three races with or without obesity. Difference in epigenetic regulation may also exist that contribute to TAM abundance and function. This might also impact the future use of novel immunomodulating agents in patients with different TME and should be considered as potential companion diagnostics in future clinical trials. In summary, our work demonstrates that breast cancers from BL patients contain significantly higher densities of TAMs and of CLS compared with NBLA and CA. Importantly, numbers of M2 TAMs and M1 CLS are associated with worse survival in all racial groups. Our results also confirm that TAMs do have the capacity to divide in breast cancers, and that they do it when in direct contact with tumor cells. Our work also demonstrates the existence of significant differences in the average numbers and densities of actively proliferating TAMs across races, with TAMs from BL patients proliferating in higher numbers than those in tumors from NBLA and CA. Possibly, significant differences in the cellular and molecular regulation of the TME exist across various races. Further studies are needed to elucidate the particular cytokines, chemokines, and other molecules that predominate in TME of breast cancer patients in the minority race groups. Those molecular differences may account for the particular TAMs’ functional patterns that may contribute to aggressiveness of the tumors. Overall, our findings suggest the potential application of third-generation checkpoint inhibitors [56] to activate macrophage's M1 killing activities in breast cancer immunotherapies, and its possible use as a biological intervention to address breast cancer disparity in minority race groups. Supplementary Material 10549_2016_3847_MOESM1_ESM Acknowledgments We would like to extend our thanks to the University of Miami Sylvester Comprehensive Cancer Center's Tissue Bank Core Facility and the Tumor Registry, and particularly to Drs. Consuelo Alvarez and Clara Milikowski without whom this investigation would not have been carried out. This study was funded by the Braman Family Breast Cancer Institute Development Grant from Sylvester Comprehensive Cancer Center at University of Miami Miller School of Medicine and also in part by the NCI/NIH R21CA176055 both to MTK. Research for this article was supported in part by funding to T.A.I from Breast Cancer Research Foundation and Play for P.I.N.K., NIEHS R01-ES024991, Women's Cancer Association of UM and Sylvester Comprehensive Cancer Center; to L.G.S from a supplement to R21CA176055, to ZC from R21CA178675; to MLP from R01CA181115 and to M.P.C. from R01CA157012. Fig. 1 TAMs occur at significantly higher densities in breast cancers from blacks. a IHC for CD163 macrophage marker (AbD Serotec, Raleigh, NC, USA) was done in 4 μm sections from FFPE tumor blocks. Sections were deparaffinized in xylene and hydrated in series of graded alcohols (100, 95 and 75 %). Heat-induced antigen retrieval was carried out in water bath (90 °C) in the presence of antigen unmasking solution (Citric Acid Based from Vector Laboratories, Burlingame, CA, USA). Antigen retrieval was followed by one incubation step with Peroxidase blocking reagent (dilution 1:100 of Perdrogen 30 %, Sigma-Aldrich, St. Louis, MO, USA) for 30 min at room temperature, followed by another step of blocking serum (normal horse serum 1.5 %, Vector Laboratories, in PBS) for 20 min at RT. CD163 (1:250) was incubated for 1 h at RT. As a detection system, we used Vectastain Elite ABC kit (Biotin/Avidin System) from Vector Laboratories; slides were counterstained with hematoxylin for 30 s. Pictures (×40) show the different densities of CD163 expression across the three racial groups. b Quantitative (calculated) macrophage density was determined in one tumor section slide per tumor. Density was calculated by dividing the number of CD163+ cells observed using a ×40 lens of a Olympus BX41 microscope by the area of the visual field, calculated as πr2, where the radius of the visual field was determined using a calibration graduated slide. TAMs’ density was calculated for 20 different and randomly selected visual fields in each tissue section, and a final average was determined and reported as the TAMs’ density per each case as cells/mm2. The graph shows TAMs’ density distribution across the three racial groups studied Fig. 2 Estimated densities of TAMs and their M1/M2 activation phenotypes in breast cancers exhibit different distributions across races. IHC for CD206 (1:50, M2 macrophage marker) and CD40 (1:200, M1 macrophage marker), both from R&D, Minneapolis, MN, was done as in Fig. 1, except Antigen Unmasking Solution High pH from Vector Laboratories was used. After staining, each histological sample was assessed qualitatively for CD163, CD206, and CD40. These estimated densities were assessed by visually scanning the whole tissue slide with a ×10 lens and arbitrarily assigning 1+ (low), 2+ (medium) or 3+ (high) as follows: 1+ = 1–150 cells/mm2; 2+ = 151–300 cells/mm2, and 3+ = >300 cells/mm2. a Estimated density of CD163+ TAMs expressed in % of total patients with low, medium, or high densities per each race. b Estimated density of CD206+ (M2) expressed in % of total patients with low, medium, or high densities per each race. c Estimated density of CD40+ (M1) expressed in % of total patients with low, medium, or high densities per each race Fig. 3 TAMs in breast cancers proliferate when in contact with tumor cells. For the analysis of proliferating macrophages (double staining with Ki-67 and CD163), tumor sections were first incubated with Ki-67 (Dako, Carpinteria, CA, USA) for 180 min (1:25 dilution) at RT and developed in brown. Then, the slides were incubated with CD163 (1:250) for 30 min (RT) and developed in red. The slides were then counterstained with hematoxylin for 30 s. As a detection system, Vectastain Elite ABC and ABC-AP Mouse kits (Biotin/AvidinSystem) from Vector Laboratories were used. In this case, breast biopsies were photographed at ×20 using an Olympus BX41 microscope with an Olympus DP15 digital camera. Images were stored in JPEG format, and density was calculated by dividing the number of proliferating TAMs in one image per case by the area observed in each picture of the analyzed tissue, calculated using the reference of 50 μm divisions. Proliferating macrophages were considered the cells simultaneously expressing CD163 (red) and Ki-67 (brown) staining. a TAMs far from the tumor cells marked negative for proliferative activity (CD163+/Ki67−, see inset with amplification); b TAMs in close contact with tumor cells exhibit proliferative activity (CD163+/Ki-67+, see inset with amplification) Fig. 4 Breast cancers from BL patients show the highest numbers of proliferating TAMs. a Double staining (CD163+/Ki-67+) shows proliferative activity in the three racial groups. b Double staining (CD163+/Ki-67+) show proliferative activity with mitotic figures in the three racial groups. c Graph shows the density distribution of total CD163+ TAMs in 23 patients of the three racial groups studied. d Graph shows the density distribution of proliferating CD163+/Ki67+ TAMs in the 23 patients of the three racial groups studied. e TAMs from Black patients exhibit increased proliferating ratios as compared with the other two races Fig. 5 Crown-like structure (CLS) occur at significantly higher numbers in breast cancers from Blacks. To calculate the density of CLS, total numbers of CLS present in one slide per tumor sample were counted, and this number was divided by the tumor tissue area (approximately calculated as a rectangle). This was done for CD163, CD206, and CD40 markers in CLS. a IHC results showing CD163 + CLS in the three racial groups (see inset and amplification). b Graph shows density of CD163+ CLS in the three racial groups. c Graph shows density of CD206+ CLS in the three racial groups. d Graph shows density of CD40+ CLS in the three racial groups Table 1 Cox regression models for overall survival Category Univariate Bivariate With Race With ER With PR With HER2 HR (90 % CI) p HR (90 % CI) p HR (90 % CI) p HR (90 % CI) p HR (90 % CI) p Qual. Density (CD 163) 2+/3+ vs 1+ 1.15 (0.78, 1.70) 0.549 1.04 (0.69, 1.55) 0.885 0.69 (0.43, 1.10) 0.193 0.73 (0.46, 1.14) 0.248 1.07 (0.72, 1.60) 0.781 Qual. Density(CD206) 2+/3+ vs 1+ 1.44 (0.99, 2.09) 0.111 1.27 (0.86, 1.89) 0.314 1.09 (0.72, 1.65) 0.736 1.14 (0.77, 1.71) 0.584 1.38 (0.95, 2.02) 0.155 Qual. density (CD40) 2+/1+ vs 0 0.97 (0.61, 1.54) 0.922 1.23 (0.73, 2.08) 0.518 0.89 (0.56, 1.41) 0.680 0.90 (0.57, 1.43) 0.712 1.01 (0.63, 1.60) 0.980 Density of TAMs (CD163) Density 1.00 (1.00, 1.00) 0.195 1.00 (1.00, 1.00) 0.549 1.00 (1.00, 1.00) 0.599 1.00 (1.00, 1.00) 0.740 1.00 (1.00, 1.00) 0.436 Density of CLS (CD163) Density 2.97 (0.69, 12.81) 0.222 2.14 (0.46, 9.96) 0.415 1.68 (0.35, 8.01) 0.584 1.99 (0.42, 9.31) 0.465 2.42 (0.54, 10.89) 0.334 Density of CLS (CD206) Density 0.94 (0.05, 17.62) 0.970 0.65 (0.03, 12.58) 0.809 0.32 (0.01, 7.03) 0.544 0.36 (0.02, 7.59) 0.581 0.74 (0.04, 15.55) 0.872 Density of CLS (CD40) Density 12.15 (1.39, 106.49) 0.058 9.14 (1.00, 83.60) 0.100 7.41 (0.80, 68.54) 0.139 17.43 (1.85, 163.75) 0.036 13.59 (1.56, 118.16) 0.047 ER Pos vs Neg 0.52 (0.36, 0.75) 0.004 0.54 (0.37, 0.79) 0.007 - - - - - - PR Pos vs Neg 0.52 (0.36, 0.75) 0.003 0.53 (0.37, 0.78) 0.006 - - - - - - HER2 Pos vs Neg 1.71 (1.00, 2.94) 0.102 1.57 (0.91, 2.72) 0.177 - - - - - - Ethnicity BL vs CA 1.72 (1.07, 2.78) 0.061 - - - - - - - - NBLA vs CA 1.39 (0.85, 2.26) 0.267 - - - - - - - - To be consisted across models with the same patients, 11 patients with unknown ER or PR were excluded resulted in 134 patients. Bivariate model included two predictors only (predictors in row + predictor in column) and displayed the HR (90 % CI) and p value for row predictor only HR (90 % CI) hazard ratio and 90% confidence interval Type-I error rate=10% The bold values indicate statically significant probabilities Table 2 Cox regression models for progression-free survival Category Univariate Bivariate With Race With ER With PR With HER2 HR (90 % CI) p HR (90 % CI) p HR (90 % CI) p HR (90 % CI) p HR (90 % CI) p Qual. density (CD163) 2+/3+ vs 1+ 1.20 (0.83, 1.73) 0.420 1.11 (0.76, 1.62) 0.663 0.75 (0.48, 1.16) 0.273 0.81 (0.53, 1.25) 0.421 1.11 (0.76, 1.62) 0.664 Qual. density (CD206) 2+/3+ vs 1+ 1.65 (1.16, 2.35) 0.019 1.55 (1.06, 2.25) 0.056 1.30 (0.88, 1.92) 0.274 1.37 (0.94, 2.01) 0.170 1.59 (1.12, 2.27) 0.031 Qual. density (CD40) 2+/1+ vs 0 0.89 (0.57, 1.39) 0.668 1.03 (0.62, 1.71) 0.914 0.82 (0.52, 1.28) 0.455 0.84 (0.54, 1.31) 0.517 0.93 (0.59, 1.45) 0.786 Density of TAMs (CD163) Density 1.00 (1.00, 1.00) 0.161 1.00 (1.00, 1.00) 0.408 1.00 (1.00, 1.00) 0.658 1.00 (1.00, 1.00) 0.919 1.00 (1.00, 1.00) 0.459 Density of CLS (CD163) Density 2.88 (0.87, 9.51) 0.145 2.30 (0.66, 8.03) 0.275 1.71 (0.48, 6.12) 0.486 2.12 (0.61, 7.39) 0.322 2.21 (0.65, 7.59) 0.288 Density of CLS (CD206) Density 1.56 (0.13, 18.84) 0.768 1.16 (0.09, 14.28) 0.924 0.56 (0.04, 7.64) 0.717 0.68 (0.05, 8.84) 0.803 1.17 (0.09, 15.35) 0.922 Density of CLS (CD40) Density 5.20 (0.64, 42.25) 0.196 4.12 (0.49, 34.92) 0.276 3.09 (0.36, 26.71) 0.390 6.94 (0.80, 60.40) 0.141 5.87 (0.73, 47.23) 0.162 ER Pos vs Neg 0.52 (0.37, 0.74) 0.002 0.54 (0.38, 0.77) 0.004 - - - - - - PR Pos vs Neg 0.55 (0.39, 0.78) 0.005 0.56 (0.39, 0.80) 0.007 - - - - - - HER2 Pos vs Neg 1.90 (1.15, 3.16) 0.036 1.81 (1.08, 3.02) 0.057 - - - - - - Ethnicity BL vs CA 1.50 (0.97, 2.34) 0.126 - - - - - - - - NBLA vs CA 1.25 (0.80, 1.96) 0.411 - - - - - - - - To be consisted across models with the same patients, 11 Patients with unknown ER or PR were excluded resulted in 134 patients. Bivariate model included two predictors only (predictors in row + predictor in column) and displayed the HR (90 % CI) and p value for row predictor only HR (90 % CI) hazard ratio and 90 % confidence interval The bold values indicate statically significant probabilities Electronic supplementary material The online version of this article (doi:10.1007/s10549-016-3847-3) contains supplementary material, which is available to authorized users. Compliance with ethical standards Conflict of interest The authors declare that they have no conflict of interest. References 1 Cancer Facts and Figures 2014 2014 American Cancer Society Atlanta 2 Porter PL Lund MJ Lin MG Yuan X Liff JM Flagg EW Coates RJ Eley JW Racial differences in the expression of cell cycle-regulatory proteins in breast carcinoma. Cancer 2004 100 12 2533 2542 doi:10.1002/cncr.20279 15197793 3 Henderson BE Lee NH Seewaldt V Shen H The influence of race and ethnicity on the biology of cancer. Nat Rev Cancer 2012 12 9 648 653 doi:10.1038/nrc3341 22854838 4 Kolonel LN Altshuler D Henderson BE The multiethnic cohort study: exploring genes, lifestyle and cancer risk. Nat Rev Cancer 2004 4 7 519 527 doi:10.1038/nrc1389 15229477 5 Amend K Hicks D Ambrosone CB Breast cancer in African-American women: differences in tumor biology from European-American women. Cancer Res 2006 66 17 8327 8330 doi:10.1158/0008-5472.CAN-06-1927 16951137 6 Furberg H Millikan R Dressler L Newman B Geradts J Tumor characteristics in African American and white women. Breast Cancer Res Treat 2001 68 1 33 43 doi:10.1023/A:1017994726207 11678307 7 Martin DN Boersma BJ Yi M Reimers M Howe TM Yfantis HG Tsai YC Williams EH Lee DH Stephens RM Weissman AM Ambs S Differences in the Tumor Microenvironment between African-American and European-American Breast Cancer Patients. PLoS One 2009 4 2 e4531 doi:10.1371/journal.pone.0004531 19225562 8 Jatoi I Becher H Leake CR Widening disparity in survival between white and African-American patients with breast carcinoma treated in the U. S. Department of Defense Healthcare system. Cancer 2003 98 5 894 899 doi:10.1002/cncr.11604 12942554 9 Newman LA Griffith KA Jatoi I Simon MS Crowe JP Colditz GA Meta-analysis of survival in African American and white American patients with breast cancer: ethnicity compared with socioeconomic status. J Clin Oncol Off J Am Soc Clin Oncol 2006 24 9 1342 1349 doi:10.1200/JCO.2005.03.3472 10 Elledge RM Clark GM Chamness GC Osborne CK Tumor biologic factors and breast cancer prognosis among white, Hispanic, and black women in the United States. J Natl Cancer Inst 1994 86 9 705 712 doi:10.1158/0008-5472.CAN-06-1927 7908990 11 Chlebowski RT Chen Z Anderson GL Rohan T Aragaki A Lane D Dolan NC Paskett ED McTiernan A Hubbell FA Adams-Campbell LL Prentice R Ethnicity and breast cancer: factors influencing differences in incidence and outcome. J Natl Cancer Inst 2005 97 6 439 448 doi:10.1093/jnci/dji064 15770008 12 Churpek JE Walsh T Zheng Y Moton Z Thornton AM Lee MK Casadei S Watts A Neistadt B Churpek MM Huo D Zvosec C Liu F Niu Q Marquez R Zhang J Fackenthal J King MC Olopade OI Inherited predisposition to breast cancer among African American women. Breast Cancer Res Treat 2015 149 1 31 39 doi:10.1007/s10549-014-3195-0 25428789 13 Mahmoud SM Lee AH Paish EC Macmillan RD Ellis IO Green AR Tumour-infiltrating macrophages and clinical outcome in breast cancer. J Clin Pathol 2012 65 2 159 163 doi:10.1136/jclinpath-2011-200355 22049225 14 Bingle L Brown NJ Lewis CE The role of tumour-associated macrophages in tumour progression: implications for new anticancer therapies. J Pathol 2002 196 3 254 265 doi:10.1002/path.1027 11857487 15 Beatty GL Chiorean EG Fishman MP Saboury B Teitelbaum UR Sun W Huhn RD Song W Li D Sharp LL Torigian DA O'Dwyer PJ Vonderheide RH CD40 agonists alter tumor stroma and show efficacy against pancreatic carcinoma in mice and humans. Science 2011 331 6024 1612 1616 doi:10.1126/science.1198443 21436454 16 Kondo T Tsunematsu T Yamada A Arakaki R Saito M Otsuka K Kujiraoka S Ushio A Kurosawa M Kudo Y Ishimaru N Acceleration of tumor growth due to dysfunction in M1 macrophages and enhanced angiogenesis in an animal model of autoimmune disease. Lab Investig J Tech Methods Pathol 2016 96 4 468 480 doi:10.1038/labinvest.2015.166 17 Balkwill F Charles KA Mantovani A Smoldering and polarized inflammation in the initiation and promotion of malignant disease. Cancer Cell 2005 7 3 211 217 doi:10.1016/j.ccr.2005.02.013 15766659 18 Grivennikov SI Greten FR Karin M Immunity, inflammation, and cancer. Cell 2010 140 6 883 899 doi:10.1016/j.cell.2010.01.025 20303878 19 Mantovani A Sozzani S Locati M Allavena P Sica A Macrophage polarization: tumor-associated macrophages as a paradigm for polarized M2 mononuclear phagocytes. Trends Immunol 2002 23 11 549 555 doi:10.1016/S1471-4906(02)02302-5 12401408 20 Pollard JW Tumour-educated macrophages promote tumour progression and metastasis. Nat Rev Cancer 2004 4 1 71 78 doi:10.1038/nrc1256 14708027 21 Noy R Pollard JW Tumor-associated macrophages: from mechanisms to therapy. Immunity 2014 41 1 49 61 doi:10.1016/j.immuni.2014.06.010 25035953 22 Torroella-Kouri M Ma X Perry G Ivanova M Cejas PJ Owen JL Iragavarapu-Charyulu V Lopez DM Diminished expression of transcription factors nuclear factor kappaB and CCAAT/enhancer binding protein underlies a novel tumor evasion mechanism affecting macrophages of mammary tumor-bearing mice. Cancer Res 2005 65 22 10578 10584 doi:10.1158/0008-5472.CAN-05-0365 16288051 23 Torroella-Kouri M Silvera R Rodriguez D Caso R Shatry A Opiela S Ilkovitch D Schwendener RA Iragavarapu-Charyulu V Cardentey Y Strbo N Lopez DM Identification of a subpopulation of macrophages in mammary tumor-bearing mice that are neither M1 nor M2 and are less differentiated. Cancer Res 2009 69 11 4800 4809 doi:10.1158/0008-5472.CAN-08-3427 19458073 24 Rodriguez D Silvera R Carrio R Nadji M Caso R Rodriguez G Iragavarapu-Charyulu V Torroella-Kouri M Tumor microenvironment profoundly modifies functional status of macrophages: peritoneal and tumor-associated macrophages are two very different subpopulations. Cell Immunol 2013 283 1–2 51 60 doi:10.1016/j.cellimm.2013.06.008 23850963 25 Torroella-Kouri M Rodriguez D Caso R Alterations in macrophages and monocytes from tumor-bearing mice: evidence of local and systemic immune impairment. Immunol Res 2013 57 1–3 86 98 doi:10.1007/s12026-013-8438-3 24203436 26 Yuan A Hsiao YJ Chen HY Chen HW Ho CC Chen YY Liu YC Hong TH Yu SL Chen JJ Yang PC Opposite effects of M1 and M2 macrophage subtypes on lung cancer progression. Sci Rep 2015 5 14273 doi:10.1038/srep14273 26399191 27 Lumeng CN Bodzin JL Saltiel AR Obesity induces a phenotypic switch in adipose tissue macrophage polarization. J Clin Investig 2007 117 1 175 184 doi:10.1172/JCI29881 17200717 28 Cinti S Mitchell G Barbatelli G Murano I Ceresi E Faloia E Wang S Fortier M Greenberg AS Obin MS Adipocyte death defines macrophage localization and function in adipose tissue of obese mice and humans. J Lipid Res 2005 46 11 2347 2355 doi:10.1194/jlr.M500294-JLR200 16150820 29 Morris PG Hudis CA Giri D Morrow M Falcone DJ Zhou XK Du B Brogi E Crawford CB Kopelovich L Subbaramaiah K Dannenberg AJ Inflammation and increased aromatase expression occur in the breast tissue of obese women with breast cancer. Cancer Prev Res (Phila) 2011 4 7 1021 1029 doi:10.1158/1940-6207.CAPR-11-0110 21622727 30 Subbaramaiah K Howe LR Bhardwaj P Du B Gravaghi C Yantiss RK Zhou XK Blaho VA Hla T Yang P Kopelovich L Hudis CA Dannenberg AJ Obesity is associated with inflammation and elevated aromatase expression in the mouse mammary gland. Cancer Prev Res (Phila) 2011 4 3 329 346 doi:10.1158/1940-6207.CAPR-10-0381 21372033 31 Santander AM Lopez-Ocejo O Casas O Agostini T Sanchez L Lamas-Basulto E Carrio R Cleary MP Gonzalez-Perez RR Torroella-Kouri M Paracrine interactions between adipocytes and tumor cells recruit and modify macrophages to the mammary tumor microenvironment: the role of obesity and inflammation in breast adipose tissue. Cancers 2015 7 1 143 178 doi:10.3390/cancers7010143 25599228 32 Nelson LR Bulun SE Estrogen production and action. J Am Acad Dermatol 2001 45 3 Suppl S116 S124 doi:10.1067/mjd.2001.117432 11511861 33 Fanelli MA Vargas-Roig LM Gago FE Tello O Lucero De Angelis R Ciocca DR Estrogen receptors, progesterone receptors, and cell proliferation in human breast cancer. Breast Cancer Res Treat 1996 37 3 217 228 doi:10.1007/BF01806503 8825133 34 Tan H Zhong Y Pan Z Autocrine regulation of cell proliferation by estrogen receptor-alpha in estrogen receptor-alpha-positive breast cancer cell lines. BMC Cancer 2009 9 31 doi:10.1186/1471-2407-9-31 19171042 35 Liao XH Lu DL Wang N Liu LY Wang Y Li YQ Yan TB Sun XG Hu P Zhang TC Estrogen receptor alpha mediates proliferation of breast cancer MCF-7 cells via a p21/PCNA/E2F1-dependent pathway. FEBS J 2014 281 3 927 942 doi:10.1111/febs.12658 24283290 36 Carrio R Koru-Sengul T Miao F Gluck S Lopez O Selman Y Alvarez C Milikowski C Gomez C Jorda M Nadji M Torroella-Kouri M Macrophages as independent prognostic factors in small T1 breast cancers. Oncol Rep 2013 29 1 141 148 doi:10.3892/or.2012.2088 23076599 37 Pettersen JS Fuentes-Duculan J Suarez-Farinas M Pierson KC Pitts-Kiefer A Fan L Belkin DA Wang CQ Bhuvanendran S Johnson-Huang LM Bluth MJ Krueger JG Lowes MA Carucci JA Tumor-associated macrophages in the cutaneous SCC microenvironment are heterogeneously activated. J Invest Dermatol 2011 131 6 1322 1330 doi:10.103/jid.2011.9 21307877 38 Shabo I Stal O Olsson H Dore S Svanvik J Breast cancer expression of CD163, a macrophage scavenger receptor, is related to early distant recurrence and reduced patient survival. Int J Cancer J Int du Cancer 2008 123 4 780 786 doi:10.1002/ijc.23527 39 Lau SK Chu PG Weiss LM CD163: a specific marker of macrophages in paraffin-embedded tissue samples. Am J Clin Pathol 2004 122 5 794 801 doi:10.1309/QHD6-YFN8-1KQX-UUH6 15491976 40 Lahmar Q Keirsse J Laoui D Movahedi K Van Overmeire E Van Ginderachter JA Tissue-resident versus monocyte-derived macrophages in the tumor microenvironment. Biochim Biophys Acta 1865 2016 1 23 34 doi:10.1016/j.bbcan.2015.06.009 41 Scholzen T Gerdes J The Ki-67 protein: from the known and the unknown. J Cell Physiol 2000 182 3 311 322 doi:10.1002/(SICI)1097-4652(200003)182:3\311:AID-JCP1[3.0.CO;2-9 10653597 42 Zhang Y Cheng S Zhang M Zhen L Pang D Zhang Q Li Z High-infiltration of tumor-associated macrophages predicts unfavorable clinical outcome for node-negative breast cancer. PLoS One 2013 8 9 e76147 doi:10.1371/journal.pone.0076147 24098773 43 Medrek C Ponten F Jirstrom K Leandersson K The presence of tumor associated macrophages in tumor stroma as a prognostic marker for breast cancer patients. BMC Cancer 2012 12 306 doi:10.1186/1471-2407-12-306 22824040 44 Qian BZ Zhang H Li J He T Yeo EJ Soong DY Carragher NO Munro A Chang A Bresnick AR Lang RA Pollard JW FLT1 signaling in metastasis-associated macrophages activates an inflammatory signature that promotes breast cancer metastasis. J Exp Med 2015 212 9 1433 1448 doi:10.1084/jem.20141555 26261265 45 Pollard JW Macrophages define the invasive microenvironment in breast cancer. J Leukoc Biol 2008 84 3 623 630 doi:10.1189/jlb.1107762 18467655 46 Mantovani A Locati M Tumor-associated macrophages as a paradigm of macrophage plasticity, diversity, and polarization: lessons and open questions. Arterioscler Thromb Vasc Biol 2013 33 7 1478 1483 doi:10.1161/ATVBAHA.113.300168 23766387 47 Hashimoto D Chow A Noizat C Teo P Beasley MB Leboeuf M Becker CD See P Price J Lucas D Greter M Mortha A Boyer SW Forsberg EC Tanaka M van Rooijen N Garcia-Sastre A Stanley ER Ginhoux F Frenette PS Merad M Tissue-resident macrophages self-maintain locally throughout adult life with minimal contribution from circulating monocytes. Immunity 2013 38 4 792 804 doi:10.1016/j.immuni.2013.04.004 23601688 48 Gordon S Pluddemann A Martinez Estrada F Macrophage heterogeneity in tissues: phenotypic diversity and functions. Immunol Rev 2014 262 1 36 55 doi:10.1111/imr.12223 25319326 49 Zheng C Yang Q Cao J Xie N Liu K Shou P Qian F Wang Y Shi Y Local proliferation initiates macrophage accumulation in adipose tissue during obesity. Cell Death Dis 2016 7 e2167 doi:10.1038/cddis.2016.54 27031964 50 Lindau A Hardtner C Hergeth SP Blanz KD Dufner B Hoppe N Anto-Michel N Kornemann J Zou J Gerhardt LM Heidt T Willecke F Geis S Stachon P Wolf D Libby P Swirski FK Robbins CS McPheat W Hawley S Braddock M Gilsbach R Hein L von Zur Muhlen C Bode C Zirlik A Hilgendorf I Atheroprotection through SYK inhibition fails in established disease when local macrophage proliferation dominates lesion progression. Basic Res Cardiol 2016 111 2 20 doi:10.1007/s00395-016-0535-8 26891724 51 Zhang W He KF Yang JG Ren JG Sun YF Zhao JH Zhao YF Infiltration of M2-polarized macrophages in infected lymphatic malformations: possible role in disease progression. Br J Dermatol 2016 doi:10.1111/bjd.14471 52 Campbell MJ Tonlaar NY Garwood ER Huo D Moore DH Khramtsov AI Au A Baehner F Chen Y Malaka DO Lin A Adeyanju OO Li S Gong C McGrath M Olopade OI Esserman LJ Proliferating macrophages associated with high grade, hormone receptor negative breast cancer and poor clinical outcome. Breast Cancer Res Treat 2011 128 3 703 711 doi:10.1007/s10549-010-1154-y 20842526 53 Campbell MJ Wolf D Mukhtar RA Tandon V Yau C Au A Baehner F van't Veer L Berry D Esserman LJ The prognostic implications of macrophages expressing proliferating cell nuclear antigen in breast cancer depend on immune context. PLoS One 2013 8 10 e79114 doi:10.1371/journal.pone.0079114 24205370 54 Mukhtar RA Moore AP Tandon VJ Nseyo O Twomey P Adisa CA Eleweke N Au A Baehner FL Moore DH McGrath MS Olopade OI Gray JW Campbell MJ Esserman LJ Elevated levels of proliferating and recently migrated tumor-associated macrophages confer increased aggressiveness and worse outcomes in breast cancer. Ann Surg Oncol 2012 19 12 3979 3986 doi:10.1245/s10434-012-2415-2 22622474 55 Pal T Bonner D Kim J Monteiro AN Kessler L Royer R Narod SA Vadaparampil ST Early onset breast cancer in a registry-based sample of African-american women: BRCA mutation prevalence, and other personal and system-level clinical characteristics. Breast J 2013 19 2 189 192 doi:10.1111/tbj.12083 23320992 56 Kong F Gao F Li H Liu H Zhang Y Zheng R Chen J Li X Liu G Jia Y CD47: a potential immunotherapy target for eliminating cancer cells. Clin Transl Oncol Off Publ Fed Spanish Oncol Soc Natl Cancer Inst Mexico 2016 doi:10.1007/s12094-016-1489-x

PMC005xxxxxx/PMC5129655.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 0376574 32844 J Anim Ecol J Anim Ecol The Journal of animal ecology 0021-8790 1365-2656 26937627 5129655 10.1111/1365-2656.12512 NIHMS829326 Article The contribution of developmental experience vs. condition to life history, trait variation, and individual differences DiRienzo Nicholas ab Montiglio Pierre-Olivier ac a University of California at Davis, One Shields avenue, California, U.S.A. 95616 b University of Arizona, PO Box 210088, Tucson, AZ, USA 85721 c McGill University, 1205 Dr Penfield Avenue, Montreal, Quebec, Canada, H3A 1B1 Corresponding author: Nicholas DiRienzo, (630)269-4128, ndirienzo@gmail.com. University of Arizona, PO Box 210088, Tucson, AZ, USA 85721 12 11 2016 24 3 2016 7 2016 01 7 2017 85 4 915926 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. SUMMARY Developmental experience, for example food abundance during juvenile stages, is known to affect life history and behaviour. However, the life history and behavioural consequences of developmental experience have rarely been studied in concert. As a result it is still unclear whether developmental experience affects behaviour through changes in life history, or independently of it. The effect of developmental experience on life history and behaviour may also be masked or affected by individual condition during adulthood. Thus, it is critical to tease apart the effects of developmental experience and current individual condition on life history and behaviour. In this study we manipulated food abundance during development in the western black widow spider, Latrodectus hesperus, by rearing spiders on either a restricted or ad lib diet. We separated developmental from condition dependent effects by assaying adult foraging behaviour (tendency to attack prey and to stay on out of the refuge following an attack) and web structure multiple times under different levels of satiation following different developmental treatments. Spiders reared under food restriction matured slower and at a smaller size than spiders reared in ad lib conditions. Spiders reared on a restricted diet were more aggressive towards prey and built webs structured for prey capture while spiders reared on an ad lib diet were less aggressive and build safer webs. Developmental treatment affected which traits were plastic as adults: restricted spiders built safer webs when their adult condition increased, while ad-lib spiders reduced their aggression when their adult condition increased. The amount of individual variation in behaviour and web structure varied with developmental treatment. Spiders reared on a restricted diet exhibited consistent variation in all aspects of foraging behaviour and web structure, while spiders reared on an ad lib diet exhibited consistent individual variation in aggression and web weight only. Developmental experience affected the average life history, behaviour, and web structure of spiders, but also shaped the amount of phenotypic variation observed among individuals. Surprisingly, developmental experience also determined the particular way in which individuals plastically adjusted their behaviour and web structure to changes in adult condition. black widow development extended phenotype foraging Latrodectus life history personality behavioural plasticity animal architecture INTRODUCTION Animals within populations often express alternative life history strategies in response to trade-offs (Stearns 1989; Stearns 1992) and varying environmental conditions such as food abundance. Such life history strategies often range from ‘fast’ putting more emphasis on early reproduction to ‘slow’ putting more emphasis on survival (Gaillard et al. 1989; Bielby et al. 2007). Individuals expressing a fast strategy (e.g. a rapid growth, early maturation, and high fecundity) should also express behaviours associated with resource acquisition and higher mortality risk (e.g. exploration, aggression) compared to slow strategies (Réale et al. 2010), thus resulting in the emergence of consistent individual differences in behaviour (i.e. animal personalities Stamps 2007; Biro & Stamps 2008). Importantly, more insights are needed on the role of developmental plasticity expressed in response to variation in resource abundance and condition-dependence in generating joint life history and behavioural variation among individuals. Both individual life history and behaviour (i.e. its behavioural type) are known to be sensitive to conditions during development. Factors such as temperature, population density, social status, and resource availability during development all impact growth rate, maturation time, and body size (Metcalfe et al. 1989; Pepin 1991; Bonenfant et al. 2002). Yet, few studies have investigated how developmental plasticity affect the integration between life history and behaviour, or have elucidated the mechanism generating this integration (but see Stamps & Groothuis 2010; Guenther & Trillmich 2013; Urszán et al. 2015). Developmental experience can also affect the correlations between different behaviours, as well as the presence/absence of consistent individual differences (Butler et al. 2012; Bengston, Pruitt & Riechert 2014; DiRienzo et al. 2015). Variation in life history and behaviour among individuals could be observed because animals express developmental plasticity of both life history and behaviour in response to environmental conditions as juveniles (West-Eberhard 2003). Variation in food abundance, for example, affects both maturation rate and behaviour in mustard leaf beetles (Phaedon cochlearine), with those being reared on low-quality food showing slower development and lower mass at maturation while also exhibiting bold behavioural types as adults (Tremmel & Müller 2013). Importantly, developmental plasticity in life history and behaviour may be shaped by feedback loops between the an individual’s behaviour expressed and individual condition (Sih et al. 2015). Under high food availability, aggressive, bold, active individuals acquire high amounts of prey, the energy of which facilitates further increases in those behaviours, while those who are less aggressive, bold, and active are unable to acquire prey, thus further decreasing those traits (McElreath et al. 2007; Luttbeg & Sih 2010). Such positive feedback loops might increase the amount of variation in behaviour and life history expressed between individuals (Sih et al. 2015). Alternatively, individuals who acquire an abundance of food should decrease aggression, activity, or boldness (e.g. boldness, aggression), while those in poor condition should increase risky behaviour in order to acquire additional resources (Clark 1994; Luttbeg & Sih 2010). Such negative feedback loops might reduce the amount of life history and behavioural variation expressed between individuals. Variation in food abundance may also affect behaviour and life history during adulthood through condition-dependence. Thus, it is also necessary to manipulate adult condition to determine the relative contributions of current vs. developmental conditions to both the life history-behaviour relationship as well as the level of between-individual variation present in the population. Satiation level, for example, is known to affect aggression in a number of organisms (Clark 1994; Heithaus et al. 2007; Philip & Shillington 2010). In tarantulas, individuals who have received large quantities of food demonstrate reduced foraging aggression relative to those who have received smaller quantities (Philip & Shillington 2010). If an animal experiences high food abundance throughout development and their adult life, then any observed relationship between life history and behaviour may simply be a result of adult body condition. Thus, large condition-dependent shifts in behaviour as adults would suggest that any behaviour expressed is simply a carry-over from the juvenile stages (e.g. from excess energy stores), and any between-individual differences may be a byproduct of variation in condition that arose during ontogeny. Alternatively, if developmental experience drives a functional integration between life history, and behaviour, then changes in condition should have little effect on adult behaviour and between-individual differences. In this study we use the western black widow spider, Latrodectus hesperus, to assess the role of developmental experience in shaping life history and foraging decision. In black widows, foraging decisions encompass aspects of both an individual’s behaviour as well as the architecture of their web. Spiders with high levels of foraging aggression rapidly respond to prey cues present in the web, thus increasing their ability to subdue prey that are larger or can quickly escape the web (Stowe 1986; Pruitt, Riechert & Jones 2008). Web structure contributes to foraging as well. Latrodectus spp. build webs that contain both gumfooted lines, which are long, sticky lines that aid in capturing prey (Argintean et al. 2006; Blackledge & Zevenbergen 2007), and structural lines which are firmly rooted and provide a protective function (Blackledge & Zevenbergen 2007). Previous research has shown that L. hesperus build more gumfooted lines when starved as adults, and that these webs are more effective at capturing prey (Zevenbergen, Schneider & Blackledge 2008). Furthermore, spiders also show consistent individual differences in both their foraging aggression and web structure, with some being more aggressive and building webs with a large number of gumfooted lines, while others are less aggressive and build denser webs with more structural lines (DiRienzo & Montiglio, unpublished data). These individual differences are maintained even when condition is equalized, yet little is known about the contribution of developmental experience to differences in web structure. The primary goals of this study were as follows: 1) Determine the effects of experience during development on individual variation in life history, foraging behaviour, web structure, and 2) Separate developmental from condition-dependent effects on these traits to determine contribution of individual condition to adult behavioural/web phenotype. We reared spiders under different food availabilities (restricted vs. ad lib), then repeatedly measured adult behaviour and web structure under alternate nutritional states (starved vs. satiated). We hypothesize that being reared under ad lib food will result in spiders maturing earlier at a larger size, while they being less aggressive and bold, and building webs structured for safety. Alternatively, being reared under restricted food will result in spiders to maturing slowly and at a smaller size, while being more aggressive and building webs structured for capturing prey. Furthermore, we predict that these differences in developmental experience will drive consistent individual differences in behaviour and web structure, and that changes in condition as adults will drive minimal changes in behaviour and web structure. MATERIALS AND METHODS Subjects and developmental treatments We used field caught immature female Latrodectus hesperus black widow spiders (N=69), captured from two locations in Davis, California. Given the spiders were obtained from the field, the actual instar of the spiders was unknown. Once in the laboratory, all spiders were given an individual container (6 cm high × 8.5 cm diameter) and randomly assigned to one of two developmental conditions (restricted or ad lib). Spiders assigned to the ad lib treatment were provided with three large (approximately equal to the spider body size) Acheta domesticus crickets per week, while restricted spiders were provided with one small cricket per week (less than the spider body size). All spiders were held at a 12:12 h light:dark cycle at 24°C +/-1°C. Spiders were checked once a week for their final adult molt, the date of which was recorded. Spiders were assigned to bins consisting of all individuals that matured within a 2-week interval. Condition alterations and experimental timeline In order to assess condition effects, we quantified behaviour and web structure three times as adults. Once mature, spiders were assigned to one of two feeding sequences (starved-satiated-starved or satiated-starved-satiated) indicating the amount of food a spider received before being allowed to build a web and have its behaviour assayed (Fig. S1). Starved spiders received one small cricket, while satiated spiders received three large crickets. All spiders, regardless of feeding sequence or type received their food on the same day, and were subsequently allowed to feed for five days. After five days the spiders were weighed and transferred to a standardized container and allowed to build a web for seven days (see ‘spider web assessment’ for methods). We then assayed the spiders’ behaviour on three consecutive days, three times within each day (see ‘spider foraging aggression assay’ for methods). After the three days of behavioural trials were complete the spiders were removed from the web structure, returned to their home containers and provided with the number of crickets indicated by the next step of the feeding sequence. For example, if a spider was starved in the previous trial, it was then satiated (or vice-versa). Thus, all spiders built a total of three webs and had their behaviour assessed three times on nine days for a total of 27 behavioural assays per spider. Spider web assessment To assess web structure, we provided spiders with a standardized structure on which to build. Five days after being given the food allotted by their feeding sequence, the spiders were each transferred to a skeletonized cardboard box (L 27.5 × W 21 × H 14 cm). The box had three walls and all but 3cm of the top removed, thus leaving a rectangular cardboard frame with the back, bottom, and the portion of the top walls remaining (Fig. S2). The bottom and back walls were covered in black paper in order to ease the counting of individual web components. This structure provided a shelter along with a frame on which to build their web. This box was placed inside a plastic container (L 42.5 X W 27.75 × H 16.25 cm), and the spider was given seven days to construct a web. These methods are similar to those used by other researchers in order to assess web structure (Blackledge & Zevenbergen 2007; Zevenbergen, Schneider & Blackledge 2008, DiRienzo & Montiglio, unpublished data). After seven days of web construction we removed the box and counted the total number of structural and gumfooted lines connected to the floor of the box. In order to reduce any handling effects, the webs were assessed on the first day of the behavioural trials, but after the behavioural trials themselves were conducted. After the three days of behavioural trials were complete, the spiders were transferred back to their home containers. Once the spider was removed we gathered the web by winding it onto a plastic rod. The webs were placed in a desiccation chamber for 48 hours and then weighed using a Mettler Toledo MT-26 microbalance. In this species, web weight is highly correlated (R2 = 0.8) with web density (measured as the amount of reflectance from an illuminated web), and thus provide a measure as to the level of overall web investment (denser vs sparse webs) (Blackledge & Zevenbergen 2007). Spider foraging aggression assay After being allowed to build a web for seven days we assessed spider aggression towards a prey cue. In order to remove any affect of prey behaviour (DiRienzo, Pruitt & Hedrick 2013), we used a standardized vibrating mechanism (Classical Silicone Vibrator, Liler, Shenzhen, China) (Keiser & Pruitt 2014b; Keiser & Pruitt 2014a). Attached to the vibrating mechanism was a 10cm long plastic cable tie, which allowed us to apply the vibratory cue in specific locations of the web, while also reducing the intensity of the vibrations and risk of damaging the web. The vibrator provided one second long pulses at 100 cycles/second separated by approximately half second long periods of reduced frequency. This vibratory pattern and frequency is in the range of vibrations produced by houseflies caught in a spiderweb (Walcott 1963), and frequency that has been shown to elicit a response in other species of spiders (Parry 1965). Spiders respond to the prey cue as they would a prey trapped in their web. They rapidly approach the cue, then rotate 180 degrees and quickly apply sticky silk in an attempt to subdue the apparent prey item. The prey cue was applied three times, once near the shelter (within 2cm), once at the end farthest from the shelter, and once in between. If the spider did not build in the predetermined location, the observation was left blank. The cue was held to the web for 15 seconds at each location, with 10 seconds separating each application. We recorded if spiders did or did not attack the prey cue in the 15 second interval as a binary response, and if they attacked if they subsequently retreated back to their shelter as a binary response. The cue application was applied in a random order (e.g. near, far, between; far, between, near; etc.). This assay was repeated again 24 and 48 hours later. Spider morphology After all web building and behavioural trials were complete we assessed spider body size by measuring the total length of the front femur and patella. We lightly anesthetized each spider using CO2. and photographed the front limb using a Canon EOS M digital camera. Images included a ruler as reference, enabling us to measure femur-patella length using the program imageJ (Schneider, Rasband & Eliceiri 2012). Statistical methods We assessed the role of experience during ontogeny on life history variables using linear models. Average adult body mass, femur-patella length, and time to maturation were response variables and developmental treatment was included as a main effect. We used the statistical software package R version 3.1.1 (R Core Team 2015) and the package BBMLE (Bolker 2010). The relative contributions of life history, measured as the average adult weight, and current condition, measured as the deviation from the average weight, on behaviour and web structure were assessed using generalized linear mixed models. In the models we included one aspect of web structure as the response variable (number of gumfooted lines, number of structural lines, web weight). We included individual ID, bin number, and date of trial as random effects. In order to separate between-individual differences from within-individual differences, we included each individual’s average adult weight calculated over the three sets of web/behavioural assays and change in condition, expressed as a deviation from the average weight (within each trial) as main effects (van de Pol & Wright 2009). Thus, average weights represents between-individual differences in weight arising from developmental treatments, while weight deviation represents within-individual differences that arise from alterations in condition. We also included an interaction between developmental treatment and weight deviation in order to determine if there were treatment-specific responses to change in condition. Web order (first, second, or third) and population location were also included as a main effect in order to control for effects of repeated web construction and initial collection location. The number of gumfooted and structural lines were modeled with a Poisson error distribution, while web weight was modeled with a Gaussian distribution. We accounted for overdispersion in the Poisson models by assigning each observation an observation-level random effect (OLRE) (Harrison 2014). Such an effect effectively acts to account for the surplus of variation (i.e. for the overdispersion), hence ensuring that the residual variance estimated adequately, and that the statistical significance of the terms is not over-estimated. We analyzed foraging aggression with a binomial error distribution. We used the same model structure that was used to analyze web structure, but with the addition of predetermined distance from the refugee the prey cue was presented (e.g. near = 1, middle = 2, or far = 3, Distance), the order of each spiders’ trial within the day (ID order), the number of times the prey cue was presented within a day (e.g. 1, 2 or 3, Assay within day), and day of the behavioural assay within the week (e.g. 1, 2 or 3) (Day within week) as main effects. These additional main effects allowed us to control for any effects of order or repeatedly measuring the spiders in short time intervals. All explanatory variables were standardized to a mean of zero and standard deviation of one prior to fitting the models. Population had no significant effect in all models and ID order, day within week, and web number had no significant effect in all models relating to attack and retreat behaviour. Those parameter estimates are presented in the supplementary material (Table S1-4). Developmental treatment can affect average adult mass, potentially generating colinearity between these explanatory variables. Thus, we compared the full models previously described to models including only developmental treatment or average adult mass (Table S5-9). With the exception of retreat behaviour, both AIC comparisons and likelihood ratio tests provide significant support for the full models (Table S10). We calculated the adjusted repeatability of behaviour and web structure following the methods of Nakagawa & Schielzeth, 2010 (Nakagawa & Schielzeth 2010). Across-treatment repeatabilities were calculated using the variance components of the fully parameterized models described in the previous paragraph. Within-treatment repeatablities were calculated using the variance components from models with the same structure as the across-treatment models, minus the main effect of developmental treatment and fitted to only the appropriate subset of data (e.g. restricted only or ad lib only). 95% confidence intervals were obtained through parametric bootstrapping procedures (number of simulations = 1000) (Bates et al. 2013). We tested for significance of consistent individual differences in web structure and aggressive behaviour through likelihood ratio tests (one degree of freedom, (Pinheiro & Bates 2000) comparing the model with the individual random effect to a model without it while keeping the fixed effect structure constant. We calculated, the proportion of variance described by the fixed effects within the model (marginal R2), and the proportion of variance described by both the fixed and random effects (conditional R2) values for all models following Nakagawa and Schielzeth, (2013) (Nakagawa & Schielzeth 2013). All mixed models were fit with the lme4 package (Bates et al. 2013). The relationship between foraging behaviour and web structure, and how it was affected by treatment and condition, was assessed using generalized linear mixed models and a model comparison approach (Burnham & Anderson 1998) using Akaike information criteria (AIC) (Akaike 1987). A set of four models were created to predict the probability of attacking the stimuli or retreating after attacking. Each model used a binomial error distribution and consisted of the same structure as the models predicting the probability of attacking: Individual ID, bin number, and date as random effects, and average weight, change in condition, web number, distance of presentation, trial order, and assay number as main effects. There was again an interaction between weight deviation and developmental treatment in each model. The models varied in which web component they included as a main effect (gumfooted lines, structural lines, web weight, null). We also tested for an interaction between the web component and treatment to determine treatment-specific effects on web structure. After fitting they were compared using Akaike information criteria (AIC) (Akaike 1987), which computes the difference in AIC between the models (ΔAIC) as well as the Akaike weights (ωi). If the ΔAIC between two models is greater than two, the model with the lower AIC is considered to fit statistically better (Richards 2005). The Akaike weights estimate the probability of a model being the best fit for the data relative to the other models in the set. The probability of retreating was found not to be related to web structure, and thus the results are not presented (ΔAIC of top model from null = 0.8). RESULTS Effect of developmental treatment on life history Spiders reared on a restricted diet matured at a lighter weight (ß = - 323.26 ± 24.51 See Table 1), had a smaller body size as indicated by the femur-patella length (ß = -0.320 ± 0.127), and took longer to mature (ß = 23.593 ± 4.741) than spiders reared on an ad lib diet. Effect of developmental treatment on web structure and behaviour Spiders with a heavier average weight built fewer gumfooted lines (ß = -0.576 ± 0.290). Spiders reared on a restricted diet built more gumfooted lines than spiders on the ad lib diet (ß = 1.663 ± 0.552, Figure 1, Table 2). Average weight also affected web mass and the number of structural lines, with heavier spiders building more structural lines (ß = 0.364 ± 0.159, Figure 2, Table 2) and produced heavier webs (ß = 1.173 ± 0.209, Figure 3, Table 2). Spiders that were heavier on average were less likely to attack the prey cue (ß = -1.811 ± 0.759). However, the spider’s average weight did not predict the probability of retreating after attacking (Figure 4, Table 3). There were significant between-individual differences in all aspects of web structure, and behaviour (Table 4). However, the extent of individual differences in web structure and behaviour varied within each developmental treatment (Table 4). Food restricted spiders showed consistent differences in all aspects of their web structure (gumfooted lines: R = 0.76, p < 0.001; structural lines: R = 0.44, p < 0.001; web weight: R = 0.59, p < 0.001) and in their behaviour (attack: R = 0.66, p < 0.001; retreat: R = 0 .32, p < 0.001). In contrast, spiders reared on ad lib food expressed consistent individual differences only in web weight and in their probability to attack (web weight: R = 0.46, p < 0.001; attack: R = 0.68, p < 0.001). Effect of changes in adult condition on web structure and behaviour Within individual changes in adult condition interacted with developmental treatment to affect both web structure and behaviour. Spiders reared on a restricted diet responded to increases in their adult condition by building heavier webs (ß = 0.463 ± 0.136, Figure 3, Table 2), and producing more structural lines (ß = 0.390 ± 0.102, Figures 2, Table 2). Yet, due to opposing effects of adult condition (ß = -0.394 ± 0.124) and an interaction between the restricted developmental treatment and adult condition (ß = 0.472 ± 0.216), spiders reared on a restricted diet showed a minimal change in their behaviour in response to changes in adult condition (Table 3). On the other hand, spiders reared on an ad lib diet were more likely to attack if they experienced a decrease in adult condition (ß = -0.394 ± 0.124, Figure 4, Table 3). Finally, there was a trend (p = 0.055) indicating that spiders reared on a food restricted diet had a greater tendency to retreat if they experienced an increase in adult condition (β = -0.540 ± 0.281 Table 3). Spiders from the ad lib development treatment gained 20.58mg, a 3.73% increase from their average adult body mass in response to satiation, while spiders from the restricted treatment gained 17.44mg, a 7.02% increase from their average adult body mass. In response to starvation as adults spiders from the ad lib treatment lost 20.68mg, a 3.58% decrease from their average adult body mass, while spiders from the restricted treatment lost 14.34mg, a 6.22% decrease from their average adult body mass. A post-hoc analysis using percent deviation in place of average deviation in body mass yielded virtually identical outputs. Collectively, this indicates that the deviations in mass presented here do indeed reflect changes in energetic state. Relationship between web structure and behaviour There were strong correlations between the measured aspects of web structure. Webs with more gumfooted lines had fewer structural lines (rho = -0.592, n = 173, p <0.001) and a lower web weight (rho = -0.439, n = 173, p < 0.001). Yet, webs with a greater number of structural lines were also heavier (rho = 0.735, n = 173, p < 0.001). When considering how the relationship between web structure and behaviour is affected by development, models including web mass (AIC = 1095.1, delta AIC = 0.0, K = 16, weight = 0.544) and the number of structural lines (AIC = 1096.1, delta AIC = 1.0, K = 16, weight = 0.323) fit better than both the null model (AIC = 1098.5, delta AIC = 3.5, K = 14, weight = 0.095) and the model with the number of gumfooted lines as a covariate (AIC = 1100.4, delta AIC = 5.4, K = 16, weight = 0.037). The top two models fit equally well and carried the vast majority of the AIC weight. Both models were nearly identical in terms of significant parameters and sign of parameter estimates (Table 5). Overall, there was an interaction between developmental treatment and the two web components. Among spiders reared on a restricted diet, more aggressive individuals had lighter webs (ß = -1.141 ± 0.246, Table 5) and produced fewer structural lines (ß = -0.711 ± 0.359, Table 5). Yet, there was no relationship between aggression and web structure for spiders reared on an ad lib diet. DISCUSSION Here we investigated how food availability during development affected adult life history, behaviour, web structure, and their relationships. We also assessed the effect of adult condition on behaviour and web characteristics. We found that spiders reared on ad lib food were heavier, larger, and matured faster than those reared on a restricted diet. These life history differences were associated with variation in both behaviour and web structure. Heavier spiders built webs that were denser and contained more structural lines, but fewer gumfooted lines than lighter spiders. Instead, lighter spiders were more likely to attack. Further, changes in adult condition affected spiders reared on ad lib and restricted food differently. Spiders reared on a restricted diet responded to increases in adult condition by building denser webs with more structural lines. Yet, spiders reared on an ad lib diet responded to increases in adult condition by becoming less likely to attack the prey cue. Furthermore, web structure was related to behaviour such that the probability of attacking the prey cue was influenced by an interaction between developmental treatment and adult condition. Spiders reared on a restricted diet that also built denser webs with more structural lines were less likely to attack the prey cue. Behaviour and web structure showed large consistent individual differences (repeatabilities ranging from 0.29-0.67). Within treatments, restricted spiders showed consistent individual differences in all traits, while ad lib spiders showed consistent individual differences in only web mass and the probability of attacking. Individual life history strategies appear to be linked to behaviour (Biro & Stamps 2008; Réale et al. 2010). Yet, the contribution of developmental experience to this relationship is still unclear. Our results indicate that spiders reared on an ad lib diet had higher average adult body mass, which in turn influenced all measured aspects of web structure, as well as the probability of attacking prey cues. Heavier spiders appeared to take a more risk-averse strategy, building more protective webs through increased density and more structural lines (Blackledge, Coddington & Gillespie 2003; Blackledge & Zevenbergen 2007), while also being less likely to attack the prey cue. The infrequent prey encounter rates of the restricted rearing treatment may indicate lower overall prey densities, thus prompting the building of webs that increase the probability of capturing the few prey in the environment. These lighter spiders appeared to take a risk-prone strategy by instead building more gumfooted lines and being highly likely to attack the prey cue. Thus, developmental conditions can generate large differences in both behavioural and other phenotypic traits (e.g. web structure) through changes in life history (e.g. average body weight at adulthood). It is important to note that although average mass influenced the number of gumfooted lines, the number of gumfooted lines was also influenced by developmental treatment. Specifically, those spiders reared on a restricted diet built more gumfooted lines, even after accounting for average adult mass. The developmental effects of food abundance on individual behaviour (or BT) are generally thought to be adaptive in that the experience during development may predict the future environment (West-Eberhard 2003). In our study, changes in adult condition did not appear to counteract the effect of developmental experience. Even when satiated, spiders reared on a restricted diet were still more likely to attack prey cues and still built webs more suited for foraging relative to safer webs than those reared on an ad lib diet. Across both treatments all aspects of behaviour and web structure were highly repeatable over spiders’ adult lives, even though their conditions were repeatedly manipulated. Zevenbergen et al. (2008) demonstrated that starving L. hesperus females for seven days resulted in them building webs with more gumfooted lines than spiders that were fed continuously (Zevenbergen, Schneider & Blackledge 2008). Given this, the seven day interval after feeding plus the five days allowed for web construction likely ensured that our starvation attempts were successful. The relative contribution of condition dependence to consistent behavioural (and in our case web) differences, despite its potential importance, is relatively unstudied. This is unfortunate as condition-dependent effects may alter the role of individual differences in a host of ecological processes they are known to influence (Sih et al. 2012). Extensive theoretical work has proposed different ways in which individual differences may be affected through condition-dependent feedback loops (Luttbeg & Sih 2010; Sih et al. 2015). Yet, empirical studies investigating these proposed links are rare (but see Dosmann, Brooks & Mateo 2015), and experimental manipulations are virtually nonexistent. Dosmann et al (2015) found that condition was related to activity and exploration in free-ranging Belding’s ground squirrels (Urocitellus beldingi) only at the within-individual level, indicating that condition does affect behaviour, but does not affect the amount of individual behavioural variation (i.e. personality) (Dosmann, Brooks & Mateo 2015). Our results suggest that spiders that experienced an increase in adult condition took less risk, either by altering behaviour or web structure. Models indicate that such negative feedback mechanisms alone should drive a convergence of condition and behaviour, eroding personality over time (McElreath et al. 2007; Luttbeg & Sih 2010; Sih et al. 2015). Indeed, our results show that spiders reared in ad lib conditions lost 4.2% of their body weight over the course of the experiment, while those reared in restricted conditions increased in weight by 8.4%. Despite this convergence in condition, spiders still maintained large between-individual behavioural and web structure differences. Collectively, this suggests that although there are indeed short term condition-dependent effects on behaviour and web structure, the majority of the behavioural variation is driven by developmental plasticity of personality and web structure in response to food abundance. More interestingly, we also found different amounts of behavioural variation in restricted and ad-lib spiders. Specifically, juvenile food restriction resulted in large consistent individual differences in all the measured traits as adults, while ad lib food provisioning during development resulted in consistent individual differences in only web mass and the probability of attacking. One possible explanation for this result is that the limited energetic resources associated with food restriction may drive tradeoffs between the traits we measured in this study and ones which were not accounted for (e.g. physiology, immune function, egg production). Individuals making different investments in behaviour, web structure, and physiology could result in the large between individual differences we observed. Yet, under ad lib food conditions such energetic restrictions may not occur. Thus, all spiders could invest equally in all traits, measured or unmeasured, resulting in a lack of between-individual variation. Indeed, according to our models there was no individual variance in the number of structural lines produced within the ad lib development treatment, and little between individual variation in the number of gumfooted lines and in the probability of retreating. An alternative explanation for the different patterns of individual differences is that developmental treatment may drive a different functional integration between traits. Another potential alternative is the existence of ceiling or floor effects. For example, it is possible that ad lib spiders all build the maximum number of structural lines that can fit into the provided area, thus preventing the expression of between-individual differences. Additional experiments that vary the area provided for web building will be conducted to determine if this is a true ceiling effect or a byproduct of experimental constraints. In addition to influencing between individual differences, we also found that the developmental treatments differed in their responses to changes in adult condition. Specifically, spiders from the restricted development treatment who experienced positive increases in body mass from an influx of food (average weight gain = 7.0% of average body mass) did not change their behavioural response to the prey cue, but instead increased web safety by building a denser web with more structural lines. Yet, spiders from the ad lib treatment responded to decreases in adult body condition (mean weight loss = -3.6% of average body mass) not by altering web structure, but instead by being more likely to attack the prey cue. Thus, although developmental conditions generates consistent individual differences in behaviour and web structure, it also affects which traits responds to the same environmental stimuli as adults (i.e. increases or decreases in food availability). We are unaware of any other study documenting such an effect of developmental conditions. One major evolutionary implication is that developmental conditions may subsequently determine which trait, in this case behaviour vs. web structure, responds to selection pressure from environmental variability as adults. Spiders reared in conditions with minimal food will have less plasticity in behaviour relative to web structure as adults, while those who experience an abundance of food may have more plasticity in behaviour. These differences may limit an adult’s ability to respond to large swings in condition, with restricted spiders maintaining risky behaviour long after environmental conditions changed to have abundant prey, while ad lib spiders may fail to build webs with a large number of gumfooted lines if prey abundance decreases. Such developmental differences in food abundance may also carry over and limit plasticity in other contexts, such as mating or predator encounters, potentially creating maladaptive responses to the demographic profile or predator density. Spider webs are a primary example of animal architecture, or architectural constructions created by an organism that are involved in fitness-related process. Architectural constructions have been shown to be plastic in a variety of species (Smith, Ostwald & Seeley 2015). For example, spiders modify their webs in response to changes in conditions (Blackledge & Zevenbergen 2007, this study). In addition to demonstrating large consistent individual differences in web structure, our results also suggest that the construction and behaviour interact in a potentially complementary manner as parts of a foraging syndrome. Generally, web structure appears to support an individual’s foraging behaviour, with aggressive spiders building webs that aid in foraging, and non-aggressive spiders building protective webs. Overall, these results suggest a greater need for considering how individual differences in architectural constructions affect evolutionary tradeoffs, as the makeup of the structure itself may act to either mitigate or amplify the fitness benefits associated with behavioural differences. Furthermore, this study stresses the need to understand how developmental experience may affect the integration between behavioural traits and construction architecture. We thank the American Society of Naturalist for providing a Student Research Award to ND to fund this project. We also thank Ann V. Hedrick for allowing us to use her laboratory space, and Louie H. Yang for allowing us to use a microbalance. ND was supported by a University of California Davis Dissertation Year Fellowship. POM was supported by a Fonds Recherche Nature et Technologie postdoctoral fellowship. Figure 1 Relationship between average spider weight (standardized) and a) the number of gumfooted lines (N = 173 webs from 69 individuals). Red triangles represent individuals reared on a restricted diet, while black crosses represent individuals reared on an ad lib diet. The fit lines represent the relationship between the number of gumfooted lines and spider weight as predicted by the model (dotted red line = restricted, black dashed line = ad lib). Figure 2 Relationship between average spider weight (standardized) and the number of structural lines (N = 173 webs from 69 individuals). Red triangles represent individuals reared on a restricted diet, while black crosses represent individuals reared on an ad lib diet. The fit lines represent the relationship between the number of structural lines and spider weight as predicted by the model (dotted red line = restricted, black dashed line = ad lib). Given the restricted treatment interacted with spider weight deviation, the thickness of the red fit lines represent the highest, mean, and lowest weight deviation. Figure 3 Relationship between average spider weight (standardized) and the web weight (N = 173 webs from 69 individuals). Red triangles represent individuals reared on a restricted diet, while black crosses represent individuals reared on an ad lib diet. The fit lines represent the relationship between the web weight and spider weight as predicted by the model (dotted red line = restricted, black dashed line = ad lib). The thickness of the fit line represents the change in prediction associated with the greatest, mean, and lowest weight deviation experienced in each treatment. Figure 4 Relationship between average spider weight (standardized) foraging aggressiveness (1504 behavioural assays on 69 individuals over the course of the experiment). Red triangles represent individuals reared on a restricted diet, while black crosses represent individuals reared on an ad lib diet. The fit lines represent the relationship between the predicted probability of attacking and web weight as predicted by the model (dotted red line = restricted, black dashed line = ad lib). The thickness of the fit line represents the change in prediction associated with the greatest, mean, and lowest weight deviation experienced in each treatment. A jitter function was added to the plot in order to separate overlapping values. Table 1 Linear models predicting the effect of developmental treatment on body mass (mg) and leg length (mm) at maturation and the number of days until maturation on 63 individuals. Development treatment represents the effect of being reared on a restricted diet. Weight deviance = 565296.8 on 61 residual degrees of freedom, leg length deviance = 13.4 on 58 degrees of freedom, and day to maturation deviance = 21153.9 on 61 degrees of freedom. Weight Leg length Days to maturation Effect ß SE t p ß SE t p ß SE t p Intercept 561.14 18.53 30.29 <0.001 9.127 0.098 93.070 <0.001 52.074 3.584 14.530 <0.001 Developmental treatment -323.26 24.51 -13.19 <0.001 -0.320 0.127 -2.529 <0.001 23.593 4.741 4.967 <0.001 p < 0.001, F1,61= 174.00, R2 = 0.764 p = 0.014, F1,58= 6.40, R2 = 0.084 p < 0.001, F1,61= 18.62, R2 = 0.277 Table 2 Generalized linear mixed model outputs predicting the number of gumfooted and structural lines in each web and the web weight of each web built by the spiders (N = 173 webs from 69 individuals). Development treatment represents the effect of being reared on a restricted diet. Gumfooted lines deviance = 1198.6 on 162 residual degrees of freedom. Structural lines deviance 1243.0 on 162 residual degrees of freedom. Web weight deviance = 1243.0 on 162 residual degrees of freedom Gumfooted lines Structural lines Web Weight Random effects Estimate LRT p Estimate LRT p Estimate LRT p ID 0.828 24.67 <0.001 0.133 12.322 <0.001 0.533 39.276 <0.001 Date 0.141 9.975 0.002 0.002 0.059 0.808 0.000 0.000 1 Bin 0.000 0.011 0.915 0.004 0.060 0.806 0.000 0.000 1 OLRE 0.424 137.293 <0.001 0.200 138.036 <0.001 - - - Residual - - - - - 0.519 - - Fixed Effects ß SE z p ß SE z p ß SE z p Intercept 0.944 0.471 2.006 0.045 2.942 0.207 14.236 <0.001 3.120 0.308 10.133 <0.001 Developmental treatment 1.663 0.552 3.015 0.003 -0.362 0.289 -1.212 0.226 0.173 0.425 0.408 0.683 Average weight -0.576 0.290 -1.988 0.047 0.364 0.159 2.828 0.023 1.173 0.209 5.621 <0.001 Weight deviation -0.171 0.090 -1.890 0.059 0.056 0.046 1.218 0.223 0.173 0.067 2.591 0.010 Web number 0.095 0.146 0.651 0.515 -0.127 0.057 -2.243 0.029 -0.240 0.073 -3.300 0.001 Development * Weight deviation 0.119 0.151 0.786 0.432 0.390 0.102 3.810 <0.001 0.463 0.136 3.211 0.001 Marginal R2 0.504 0.437 0.552 Conditional R2 0.830 0.626 0.779 Table 3 Generalized linear mixed model predicting spider response to a simulated prey cue (binomial response) and subsequent retreat behaviour (binomial response) (1504 behavioural assays on 69 individuals over the course of the experiment). Development treatment represents the effect of being reared on a restricted diet. Attack deviance = 1070.5 on 1489 residual degrees of freedom. Retreat deviance 750.0 on 608 residual degrees of freedom Attack Retreat Random Effects Variance LRT p Variance LRT p ID 6.652 471.844 <0.001 1.316 65.343 <0.001 Date 0.106 1.934 0.164 0.145 2.201 0.138 Bin 0.196 0.199 0.655 0.106 0.238 0.629 Fixed Effects ß SE z p ß SE z p Intercept 1.826 1.059 1.724 0.085 0.951 0.783 1.214 0.224 Developmental treatment 1.114 1.429 0.780 0.436 0.945 0.961 0.984 0.325 Average weight -1.811 0.759 -2.420 0.016 0.953 0.604 1.577 0.115 Weight deviation -0.394 0.124 -3.174 0.002 0.298 0.202 1.476 0.140 Distance -1.522 0.126 -12.059 <0.001 0.298 0.128 2.334 0.020 Assay w/i day 0.036 0.105 0.340 0.734 -0.588 0.120 -4.900 <0.001 Development * Weight deviation 0.472 0.216 2.185 0.029 -0.540 0.281 -1.917 0.055 Marginal R2 0.403 0.110 Conditional R2 0.808 0.397 Table 4 Adjusted repeatabilities (R adj.), 95% confidence intervals (95%CI), and likelihood ratios (LRT) across and within developmental treatments for all measured behavioural and web traits. Gumfooted R adj 95% CI LRT P All 0.52 0.40-0.70 24.67 < 0.001 Ad lib 0.36 0.00-0.46 4.090 0.043 Restricted 0.76 0.63-0.78 33.334 < 0.001 Structural R adj 95% CI LRT P All 0.35 0.16-0.49 12.322 < 0.001 Ad lib 0.00 0.00-0.19 0.000 1 Restricted 0.44 0.16-0.50 11.303 0.001 Web Weight R adj 95% CI LRT P All 0.51 0.44-0.55 39.376 < 0.001 Ad lib 0.46 0.27-0.57 15.908 < 0.001 Restricted 0.59 0.47-0.64 26.276 < 0.001 Attack R adj 95% CI LRT P All 0.67 0.56-0.82 471.841 < 0.001 Ad lib 0.68 0.42-0.90 157.306 < 0.001 Restricted 0.66 0.48-0.78 284.579 < 0.001 Retreat R adj 95% CI LRT P All 0.29 0.12-0.40 65.344 < 0.001 Ad lib 0.17 0.00-0.40 4.0143 0.045 Restricted 0.32 0.14-0.45 62.642 <0.001 Table 5 The best fit generalized linear mixed models predicting the probability to attack as a function of web structure (1504 behavioural assays on 69 individuals over the course of the experiment). Development treatment represents the effect of being reared on a restricted diet. Web weight deviance = 1063.1 on 1488 residual degrees of freedom. Structural deviance = 1064.1 on 1488 residual degrees of freedom. Web weight Structural lines Random Effects Variance LRT p Variance LRT p ID 6.277 423.283 <0.001 6.288 434.440 <0.001 Date 0.124 2.739 0.098 0.085 1.393 0.247 Bin 0.217 0.185 0.668 0.172 0.169 0.681 Fixed Effects ß SE z p ß SE z p Intercept 1.599 1.050 1.523 0.128 1.782 1.036 1.720 0.086 Developmental treatment 0.950 1.403 0.677 0.498 0.888 1.393 0.637 0.524 Average weight -1.945 0.765 -2.544 0.011 -1.655 0.729 -2.270 0.023 Weight deviation -0.470 0.133 -3.537 <0.001 -0.403 0.125 -3.222 0.002 Distance -1.534 0.127 -12.066 <0.001 -1.523 0.126 -12.049 <0.001 Assay w/i day 0.041 0.105 0.387 0.699 0.039 0.105 0.379 0.705 Development * Weight deviation 0.768 0.259 2.971 0.003 0.747 0.245 3.049 0.002 Web component 0.564 0.292 1.933 0.053 -0.068 0.193 -0.350 0.727 Development * web component -1.141 0.426 -2.680 0.007 -0.711 0.359 -1.979 0.048 Marginal R2 0.419 0.416 Conditional R2 0.807 0.805 DATA ACCESSIBILITY The data set and the scripts used to replicate the analyses and figures presented in this paper are submitted as supporting material with the manuscript. Should the manuscript be accepted, the data set and scripts will be uploaded to Dryad and shared publicly. References cited Akaike H 1987 Factor analysis and AIC Psychometrika 52 317 332 Argintean S Chen J Kim M Moore A 2006 Resilient silk captures prey in black widow cobwebs Applied Physics A 82 235 241 Bates D Maechler M Bolker B Walker S 2013 lme4: Linear mixed-effects models using Eigen and S4 R package version 1 Bengston SE Pruitt JN Riechert SE 2014 Differences in environmental enrichment generate contrasting behavioural syndromes in a basal spider lineage Animal Behaviour 93 105 110 Bielby J Mace G Bininda-Emonds O Cardillo M Gittleman J Jones K Orme C Purvis A 2007 The fast-slow continuum in mammalian life history: an empirical reevaluation The American Naturalist 169 748 757 Biro PA Stamps JA 2008 Are animal personality traits linked to life-history productivity? Trends in Ecology & Evolution 23 361 368 18501468 Blackledge TA Coddington JA Gillespie RG 2003 Are three-dimensional spider webs defensive adaptations? Ecology Letters 6 13 18 Blackledge TA Zevenbergen JM 2007 Condition-dependent spider web architecture in the western black widow, Latrodectus hesperus Animal Behaviour 73 855 864 Bolker B 2010 bbmle: Tools for general maximum likelihood estimation R package version 0.9 Bonenfant C Gaillard JM Klein F Loison A 2002 Sex-and age-dependent effects of population density on life history traits of red deer Cervus elaphus in a temperate forest Ecography 25 446 458 Burnham K Anderson D 1998 Model selection and inference Springer New York, NY Butler MW Toomey MB McGraw KJ Rowe M 2012 Ontogenetic immune challenges shape adult personality in mallard ducks Proceedings of the Royal Society B: Biological Sciences 279 326 333 21653587 Clark CW 1994 Antipredator behavior and the asset-protection principle Behavioral Ecology 5 159 170 DiRienzo N Niemelä PT Skog A Vainikka A Kortet R 2015 Juvenile pathogen exposure affects the presence of personality in adult field crickets Frontiers in Ecology and Evolution 3 36 DiRienzo N Pruitt JN Hedrick AV 2013 The combined behavioural tendencies of predator and prey mediate the outcome of their interaction Animal Behaviour 86 317 322 Dosmann AJ Brooks KC Mateo JM 2015 Within-Individual Correlations Reveal Link Between a Behavioral Syndrome, Condition, and Cortisol in Free-Ranging Belding’s Ground Squirrels Ethology 121 125 134 25598565 Gaillard J-M Pontier D Allaine D Lebreton J Trouvilliez J Clobert J 1989 An analysis of demographic tactics in birds and mammals Oikos 59 76 Guenther A Trillmich F 2013 Photoperiod influences the behavioral and physiological phenotype during ontogeny Behavioral Ecology 24 402 411 Harrison XA 2014 Using observation-level random effects to model overdispersion in count data in ecology and evolution PeerJ 2 e616 25320683 Heithaus MR Frid A Wirsing AJ Dill LM Fourqurean JW Burkholder D Thomson J Bejder L 2007 State-dependent risk-taking by green sea turtles mediates top-down effects of tiger shark intimidation in a marine ecosystem Journal of Animal Ecology 76 837 844 17714261 Keiser CN Pruitt JN 2014a Personality composition is more important than group size in determining collective foraging behaviour in the wild Proceedings of the Royal Society of London B: Biological Sciences 281 20141424 Keiser CN Pruitt JN 2014b Spider aggressiveness determines the bidirectional consequences of host–inquiline interactions Behavioral Ecology 25 142 151 Luttbeg B Sih A 2010 Risk, resources and state-dependent adaptive behavioural syndromes Philosophical Transactions of the Royal Society B: Biological Sciences 365 3977 3990 McElreath R Luttbeg B Fogarty SP Brodin T Sih A 2007 Evolution of animal personalities Nature 450 E5 E5 17994035 Metcalfe N Huntingford FA Graham W Thorpe J 1989 Early social status and the development of life-history strategies in Atlantic salmon Proceedings of the Royal Society of London B: Biological Sciences 236 7 19 2565574 Nakagawa S Schielzeth H 2010 Repeatability for Gaussian and non-Gaussian data: a practical guide for biologists Biological Reviews 85 935 956 20569253 Nakagawa S Schielzeth H 2013 A general and simple method for obtaining R2 from generalized linear mixed-effects models Methods in Ecology and Evolution 4 133 142 Parry D 1965 The signal generated by an insect in a spider’s web Journal of Experimental Biology 43 185 192 5884066 Pepin P 1991 Effect of temperature and size on development, mortality, and survival rates of the pelagic early life history stages of marine fish Canadian Journal of Fisheries and Aquatic Sciences 48 503 518 Philip B Shillington C 2010 The effect of prey availability on metabolism and activity in the tarantula Phormictopus cancerides Canadian journal of zoology 88 90 98 Pinheiro JC Bates DM 2000 Linear mixed-effects models: basic concepts and examples Mixed-Effects Models in S and S-Plus 3 56 Pruitt JN Riechert SE Jones TC 2008 Behavioural syndromes and their fitness consequences in a socially polymorphic spider, Anelosimus studiosus Animal Behaviour 76 871 879 R Core Team 2015 R: A language and Environment for Statistical Computing Available from CRAN sites Réale D Garant D Humphries MM Bergeron P Careau V Montiglio P-O 2010 Personality and the emergence of the pace-of-life syndrome concept at the population level Philosophical Transactions of the Royal Society B: Biological Sciences 365 4051 4063 Richards SA 2005 Testing ecological theory using the information-theoretic approach: examples and cautionary results Ecology 86 2805 2814 Schneider CA Rasband WS Eliceiri KW 2012 NIH Image to ImageJ: 25 years of image analysis Nature methods 9 671 675 22930834 Sih A Cote J Evans M Fogarty S Pruitt J 2012 Ecological implications of behavioural syndromes Ecology Letters 15 278 289 22239107 Sih A Mathot KJ Moirón M Montiglio P-O Wolf M Dingemanse NJ 2015 Animal personality and state–behaviour feedbacks: a review and guide for empiricists Trends in Ecology & Evolution 30 50 60 25498413 Smith ML Ostwald MM Seeley TD 2015 Adaptive tuning of an extended phenotype: honeybees seasonally shift their honey storage to optimize male production Animal Behaviour 103 29 33 Stamps JA 2007 Growth-mortality tradeoffs and ‘personality traits’ in animals Ecology Letters 10 355 363 17498134 Stamps JA Groothuis TG 2010 Developmental perspectives on personality: implications for ecological and evolutionary studies of individual differences Philosophical Transactions of the Royal Society B: Biological Sciences 365 4029 4041 Stearns SC 1989 Trade-offs in life-history evolution Functional Ecology 3 259 268 Stearns SC 1992 The evolution of life histories Oxford University Press Oxford Stowe MK 1986 Prey specialization in the Araneidae 1 Stanford University Press Redwood City, California Tremmel M Müller C 2013 Insect personality depends on environmental conditions Behavioral Ecology 24 386 392 Urszán TJ Garamszegi LZ Nagy G Hettyey A Török J Herczeg G 2015 No personality without experience? A test on Rana dalmatina tadpoles Ecology and Evolution 5 5847 5856 26811759 van de Pol M Wright J 2009 A simple method for distinguishing within-versus between-subject effects using mixed models Animal Behaviour 77 753 758 Walcott C 1963 The effect of the web on vibration sensitivity in the spider, Achaearanea tepidariorum (Koch) Journal of Experimental Biology 40 595 611 West-Eberhard MJ 2003 Developmental plasticity and evolution Oxford University Press Zevenbergen JM Schneider NK Blackledge TA 2008 Fine dining or fortress? Functional shifts in spider web architecture by the western black widow Latrodectus hesperus Animal Behaviour 76 823 829

PMC005xxxxxx/PMC5129664.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 0404243 5057 J Nutr J. Nutr. The Journal of nutrition 0022-3166 1541-6100 25740908 5129664 10.3945/jn.114.206078 HHSPA826271 Article The Global Nutrition Report 2014: Actions and Accountability to Accelerate the World’s Progress on Nutrition1–4 Haddad Lawrence 5* Achadi Endang 6 Bendech Mohamed Ag 7 Ahuja Arti 8 Bhatia Komal 9 Bhutta Zulfiqar 1011 Blössner Monika 12 Borghi Elaine 12 Colecraft Esi 13 de Onis Mercedes 12 Eriksen Kamilla 14 Fanzo Jessica 15 Flores-Ayala Rafael 16 Fracassi Patrizia 17 Kimani-Murage Elizabeth 18 Koukoubou Eunice Nago 19 Krasevec Julia 20 Newby Holly 20 Nugent Rachel 21 Oenema Stineke 22 Martin-Prével Yves 23 Randel Judith 24 Requejo Jennifer 25 Shyam Tara 9 Udomkesmalee Emorn 26 Reddy K Srinath 27 5 International Food Policy Research Institute, Washington, DC 6 University of Indonesia, Jakarta, Indonesia 7 Food and Agriculture Organization of the UN, Rome, Italy 8 Women and Child Development, Odisha, India 9 Institute of Development Studies, Brighton, United Kingdom 10 Center for Global Child Health, Hospital for Sick Children, Toronto, Canada 11 Center of Excellence in Women and Child Health, Aga Khan University, Karachi, Pakistan 12 WHO, Geneva, Switzerland 13 University of Ghana, Accra, Ghana 14 University of Cambridge, Cambridge, United Kingdom 15 Columbia University, New York, NY 16 CDC, Atlanta, GA 17 Scaling Up Nutrition Secretariat, Geneva, Switzerland 18 African Population and Health Research Center (APHRC), Nairobi, Kenya 19 University of Abomey Calavi, Cotonou, Benin 20 UNICEF, New York, NY 21 University of Washington, Seattle, WA 22 Interchurch Organization for Development Cooperation (ICCO) Alliance, Utrecht, The Netherlands 23 Institute of Research for Development, Marseille, France 24 Development Initiatives, Bristol, United Kingdom 25 Partnership for Maternal, Newborn and Child Health, WHO, Geneva, Switzerland 26 Institute of Nutrition, Mahidol University, Nakhon Pathom, Thailand 27 Public Health Foundation of India, New Delhi, India * To whom correspondence should be addressed. l.haddad@cgiar.org 22 11 2016 4 3 2015 4 2015 30 11 2016 145 4 663671 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. In 2013, the Nutrition for Growth Summit called for a Global Nutrition Report (GNR) to strengthen accountability in nutrition so that progress in reducing malnutrition could be accelerated. This article summarizes the results of the first GNR. By focusing on undernutrition and overweight, the GNR puts malnutrition in a new light. Nearly every country in the world is affected by malnutrition, and multiple malnutrition burdens are the “new normal.” Unfortunately, the world is off track to meet the 2025 World Health Assembly (WHA) targets for nutrition. Many countries are, however, making good progress on WHA indicators, providing inspiration and guidance for others. Beyond the WHA goals, nutrition needs to be more strongly represented in the Sustainable Development Goal (SDG) framework. At present, it is only explicitly mentioned in 1 of 169 SDG targets despite the many contributions improved nutritional status will make to their attainment. To achieve improvements in nutrition status, it is vital to scale up nutrition programs. We identify bottlenecks in the scale-up of nutrition-specific and nutrition-sensitive approaches and highlight actions to accelerate coverage and reach. Holding stakeholders to account for delivery on nutrition actions requires a well-functioning accountability infrastructure, which is lacking in nutrition. New accountability mechanisms need piloting and evaluation, financial resource flows to nutrition need to be made explicit, nutrition spending targets should be established, and some key data gaps need to be filled. For example, many UN member states cannot report on their WHA progress and those that can often rely on data >5 y old. The world can accelerate malnutrition reduction substantially, but this will require stronger accountability mechanisms to hold all stakeholders to account. malnutrition SDGs accountability progress indicators Introduction Malnutrition, encompassing both undernutrition and overweight, is a problem facing virtually every country in the world. The consequences of malnutrition have fundamental implications throughout the life cycle: reduced chances of survival, increased risk of acute and chronic disease, impaired learning in school, and lower economic productivity. The consequences are transmitted across generations via maternal-child nutrition linkages (1–3). Currently, political commitment to malnutrition reduction is high. For example, 54 countries have chosen to become members of the Scaling Up Nutrition (SUN)28 movement; the second International Conference on Nutrition in 2014 had a strong focus on all forms of malnutrition; resource allocation for the prevention and treatment of malnutrition is on the rise as indicated by the commitments made by many organizations at the Nutrition for Growth (N4G) summit in 2013; the Lancet nutrition series of 2008 and 2013 brought together what we know about preventing and reducing malnutrition; and, unlike the Millennium Development Goal framework, the 2013 High Level Panel on the Post 2015 Development Agenda (4) explicitly recommends nutrition as an explicit feature of one of its proposed goals. The imperative now is to sustain and intensify this commitment and turn it into action that accelerates nutrition improvements. To this end, in 2013 the N4G stakeholders called for a Global Nutrition Report (GNR) to track the world’s progress in improving nutrition, to strengthen its accountability to meet commitments, and to identify actions to accelerate progress (5). The GNR is produced by an independent expert group empowered by the N4G stakeholder group [Governance details of the GNR are found on the GNR website (6)]. This article summarizes results and conclusions from the inaugural GNR. The GNR is global in nature, bringing together disparate data on 76 indicators of nutrition status, programs, policies, resources, legislation, and institutional arrangements, establishing a baseline for them to be tracked annually for the 193 UN member states. It draws on the data to describe progress in improving nutrition status, the coexisting burdens of malnutrition, the state of coverage of nutrition interventions and practices, trends in the sectors that support nutrition, nutrition spending trends, the state of data on nutrition policies, laws, and institutional arrangements. The GNR also collects data on progress toward the recent N4G commitments made in 2013 by a wide range of stakeholders. The GNR is action oriented. Drawing on original analyses and several highlighted country case studies of change, it identifies actions to accelerate progress in nutrition, suggests ways of filling key data gaps, and makes concrete recommendations on how nutrition accountability can be strengthened to better deliver action (Table 1). The report’s framework is summarized in Figure 1. Data Sources and Methods The GNR presents 76 indicators from different data sources selected on the basis of their methodologic quality and global representativeness. Most data for child anthropometry, intervention coverage, and child feeding practices come from nationally representative household surveys, primarily the Demographic and Health Surveys and the Multiple Indicator Cluster Survey. The report uses databases provided by UNICEF, WHO, the World Bank, FAO, the UN Population Division, and the UN Educational, Scientific and Cultural Organization (UNESCO). The UNICEF/WHO/World Bank’s 2013 Joint Child Malnutrition Estimates form the backbone of analysis of child nutritional status and of tracking progress toward the World Health Assembly (WHA) targets (7). The report uses some modeled estimates in the absence of adequate survey data. The WHO Global Health Observatory Data Repository is used for estimates of the prevalence of adult overweight and obesity and risk factors for noncommunicable diseases. Data on sanitation and safe water coverage are from WHO/UNICEF’s Joint Monitoring Program, and data on anemia in women of reproductive age are from a study by Stevens et al. (8). Data on food supply are from FAO and data on severe acute malnutrition geographical coverage are from UNICEF/Coverage Monitoring Network/Action Against Hunger (ACF International). The GNR’s data for financial resources, policy and legislation, and institutional arrangements indicators are from UNICEF, WHO, the International Food Policy Research Institute, the Food Fortification Initiative, the International Labor Organization, the Institute of Development Studies, and the SUN Movement. Data analysis was carried out by the GNR’s Secretariat based at the Institute of Development Studies, with additional technical input and assistance from partners at WHO and UNICEF. The report follows UN country and regional classification and naming conventions. The summary measure for the prevalence of malnutrition and intervention coverage rates at the regional and subregional levels is the population-weighted mean, with the UN Development Program’s 2012 population estimates used as analytical weights. Rates are calculated when available data covered ≥50% of the regional or subregional population. Trends in intervention coverage (nutrition-specific and nutrition-sensitive), underlying determinants (including food supply, water and sanitation, female education, and health worker density), and government expenditure in sectors related to nutrition (health, agriculture, education, and social protection) are calculated for milestone years: 1990, 2000, 2010, and for the most recent year when data are available. Progress on WHA indicators at country levels, and summarized by regions, is assessed by presenting baseline (latest estimate in 2005–2012 for stunting, wasting, and overweight and 2011 for anemia) levels of malnutrition and current annual average rates of reduction or increase (AARRs or AARIs, respectively) compared with required rates when global targets apply. These on- and off-course rules are simply to allow global comparisons using a common denominator (9). Countries will need to set their own targets and assess whether they are on and off course on the basis of them. Assessments of progress toward N4G commitments was made as follows: 1) identify the specific commitment in the N4G Compact document; 2) remind the signatory of this commitment asking them to report progress via a template, tailored to each group; 3) clarify issues with those who responded; 4) enter the final responses into a set of detailed online N4G Commitment Tracking tables; and 5) make an assessment of progress. Assessment consisted of 2 of the writing team reviewing the detailed online tables that show progress for each signatory, making independent assessments and then (twice) jointly reviewing each of the 2 independent assessments. All data sources are reported in full in the Global Nutrition Report 2014 (10). Findings Multiple malnutrition burdens are the “new normal” To date, the worlds of undernutrition and of overweight (and obesity and diet-related noncommunicable disease) have operated largely independently of each other. However, once undernutrition and overweight data are brought together, the global picture of malnutrition changes dramatically. For example, of the 122 countries in the world with comparable data, all except 2 countries experience 1 of these 3 common forms of malnutrition: under-5 stunting, anemia in women of reproductive age, or adult overweight. Most countries experience multiple forms of malnutrition, and 24 countries show all 3 forms of malnutrition (Figure 2). These different forms of malnutrition need to be considered holistically. They are connected at a political level because they compete for resources and attention (11) and at the programmatic level because the potential for unintended consequences is substantial (12). However, complexity is not an excuse for inaction. Rather, it is a call for more careful prioritizing of actions, with an enhanced understanding of potential trade-offs. The need for the development of tools and strategies for sequencing of nutrition-relevant actions in complex contexts is urgent. Multiple burdens and trends toward decentralization of nutrition programming [e.g., (13)] highlight the need for disaggregated data on nutrition outcomes. The realization that all countries face malnutrition opens the door for more cross-national collaboration and learning on what works to achieve good nutrition. Although global progress in nutrition status outcomes is too slow, many countries are making good progress Unfortunately, the world is not on course to meet any of the 6 nutrition WHA global targets for 2025. Table 2 lists the 6 WHA nutrition global targets and the extent of global progress toward them. For stunting and exclusive breastfeeding there is some progress, but for anemia, low birth weight, and wasting the global figures are static, and for under-5 overweight rates they are increasing. Countries make up the global numbers, and their progress can be assessed in 3 ways. First and foremost, individual nations will set their own targets and these are under development. Second, country rates of progress can be compared to the required global rates to meet the global targets, as shown in the last column of Table 2. Nearly one-fifth of countries are above the rate of reduction in stunting required to meet the global target. For under-5 overweight, half the countries show declining rates. For exclusive breastfeeding, over half of the countries are increasing their rates faster than the global required rate. And well over half of the countries that have data on wasting show declines. Amid the global picture, many countries are making good progress. The third way of assessing country progress is to apply the global targets on a country-by-country basis. In other words, how many countries would be on course to make their proportionate contribution to the WHA global targets? By using data from the most recent UNICEF/WHO/World Bank joint global database (7) and estimates from WHO on the required rates of change in country-level indicators to meet the global target applied at the country level (9), we apply rules proposed by WHO (9) for determining whether a country is on or off course to meet the global WHA targets.29 Of the 99 countries that have data on all 4 WHA indicators for which rules exist (stunting, wasting, overweight, and anemia), only one—Colombia—is on course to meet all 4 targets by 2025. Thirty-one countries are not on course to meet any of the 4 targets. More encouragingly, more than two-thirds of all countries that have data on all 4 indicators will meet at least one goal. There is no regional pattern to whether or not countries are on or off course. On an indicator-by-indicator basis, anemia is the indicator for which most countries are finding it difficult to make progress (Figure 3).30 The 5 countries on course for anemia reduction are Burundi, Colombia, Kenya, Vanuatu, and Vietnam. Wasting is the indicator for which the largest number of countries are making progress. This result is based not on AARRs but rather on whether wasting is <5% (on course) or ≥5% (off course).31 The 59 countries on course for wasting reduction include Brazil, China, and the United States. These 59 countries represent 39% of all children under age 5 in the 123 countries with available data. For stunting reduction, 22 countries are on course, including China, Turkey, and Vietnam. These 22 countries represent 23% of all children under age 5 in the 109 countries with available data. Finally, 31 countries are on course for under-5 overweight reduction of the 107 with available data. The 31 countries, which include India, Nigeria, and the United States, represent 45% of all children under age 5 in these 107 countries. Nutrition needs a stronger place within the Sustainable Development Goals The WHA nutrition goals are a key accountability tool for nutrition. For development more generally, the Sustainable Development Goals (SDGs) will be the world’s post-2015 accountability mechanism. The UN’s Open Working Group has made an initial proposal for 17 SDGs and 169 accompanying targets (17). Improvements in nutrition status will make substantial contributions to the attainment of many of SDGs (Table 3) and, in turn, will benefit from improvements in them. Despite these contributions, nutrition is presently under-represented in the SDG framework. For example, 45% of global deaths of children under the age of 5 can be attributed jointly to fetal growth restriction, suboptimum breastfeeding, stunting, wasting, and deficiencies of vitamin A and zinc (2). In addition, the economic benefit-cost ratio of preventing undernutrition in Africa is 16:1 and the economic costs of obesity are estimated at 5% of gross national product in China. Despite the substantial burdens imposed on development by malnutrition, only 2 of the 6 WHA global nutrition goals are mentioned in the SDGs (stunting and wasting in children under 5 y of age) and nutrition is explicitly referred to in only 1 of the 169 targets. The GNR concludes that the remaining WHA nutrition indicators (under-5 overweight, anemia in women of reproductive age, exclusive breastfeeding, and low birth weight) need to be incorporated under the other remaining SDGs, subgoals, and targets. For example, low birth weight, exclusive breastfeeding, and anemia in women of reproductive age could be indicators for the stated subgoal (3.2) of “by 2030 end preventable deaths of newborns and under 5 children,” (17) and under-5 overweight rates could be an indicator for the stated subgoal (3.4) of “by 2030 reduce by one-third pre-mature morality from non-communicable diseases (NCDs) through prevention and treatment” (17). In addition, other indicators that are important for a range of SDGs but also for nutrition improvement need to be identified and lobbied for. For example, coverage rates for nutrition-specific interventions could be indicators for the stated subgoal of “achieve universal health coverage” and there should be discussion and negotiation on the subgoals for hunger, water, sanitation, social protection, poverty and inequality, education, and women’s participation to determine if they can be framed in ways that are more helpful to nutrition. As yet, no SDG nutrition targets have been set for 2030. The WHA targets are set for 2025. The GNR recommends that the 2030 targets should not be “business as usual” 5-y extrapolations of the WHA 2025 targets. The WHA targets are determined on the basis of historical trends as of 2012, based largely on data up to 2010. We should consider being more ambitious for the 2030 targets because of new data, new analysis, and new commitments. The new data are surveys for India (preliminary) and for the Indian state of Maharashtra. Both surveys show rates of decline in stunting that are far in excess of the rates assumed in 2012. Because India accounts for >40% of global stunting, this matters a great deal. By using a new analysis of the relation between 6 underlying drivers of nutrition (e.g., improved water and sanitation coverage, gender equity, and measures of the quantity and quality of food supply) and stunting rates (24), the GNR suggests that with challenging but realistic increases in the levels of these drivers the WHA targets can be met and even exceeded. Finally, the new pledges from the N4G in 2013, the expanding membership of SUN, and the commitments that will be embodied in the second International Conference on Nutrition all need to be taken into account when formulating the 2030 targets. Nutrition interventions need to be scaled up much more quickly Because a person’s nutritional status depends on a range of immediate, underlying, and basic determinants and their interactions, nutrition investments may take various forms to address these determinants (2). Nutrition-specific programs address the immediate determinants (e.g., inadequate diet and disease burden) of nutrition status and are found in a range of policy areas, such as health, humanitarian relief, and food processing (3). Nutrition-sensitive programs and approaches incorporate explicit nutritional goals or actions and address the underlying determinants of nutrition status (e.g., food security, health access, healthy household environment, and care practices) and are found in a wide range of policy areas such as the following: agriculture; education; water, sanitation, and hygiene; social protection; women’s empowerment; and health (25). Enabling environment investments address the basic determinants of nutrition status such as governance, income, and equity. These take the form of laws, regulations, policies, investments in economic growth, and improvements in governance capacity (26). Efforts to improve nutritional status can come from investments that address all 3 levels of determinants: immediate, underlying, and basic. The aim should be to find the most potent blend of determinants, at scale, given the need, capacities, and political opportunities in each context. Nutrition-specific intervention coverage is too low The scaled-up coverage of nutrition-specific interventions is crucial for undernutrition reduction (3). The GNR summarizes the state of coverage data for the 10 nutrition-specific interventions in Bhutta et al. (3), plus zinc treatment for diarrhea (27) and universal salt iodization (also a proven nutrition-specific intervention). Data are only readily available for more than a handful of countries for vitamin A supplementation for 6- to 59-mo-olds, universal salt iodization, and zinc treatment for diarrhea. This is because programs have not been scaled up (e.g., preventive zinc supplementation, multiple micronutrient supplementation, calcium supplementation in pregnancy) or because coverage data are missing (e.g., the treatment of moderate or severe acute malnutrition). Although expanded program coverage is vital, it is only valuable if it leads to expanded impact. It is thus important to focus on maintaining and improving program effectiveness. Implementation research is important here (28). Trends in sectors that are important for nutrition are positive but too slow The FAO’s estimate of hunger (what it terms “undernourishment”) is based on food supply data that are converted into the percentage of the population below a minimum energy cutoff. Undernourishment rates are declining, although the absolute numbers of hungry individuals in sub-Saharan Africa are increasing (29). At the same time, the GNR shows that the percentage of the population that is above an upper energy threshold is increasing steadily, leaving the share of the population in the healthy range between the 2 cutoffs static at 60%. Access to improved water and sanitation services is steadily improving, but there are large coverage gaps in Eastern, Western, and Middle Africa for water and in South and South East Asia for sanitation. Girls’ secondary education enrollment is increasing steadily and now exceeds 50% in Africa. Health worker population density remains very low in Africa and half the rate of Asia. Government expenditures on these broad categories—agriculture, education, health, and social protection—vary between and within regions. Social protection spending is increasing rapidly in many African and Asian countries, providing an opportunity to incorporate nutrition into those programs. The evidence base is weak on how to make interventions that address underlying determinants more nutrition sensitive. Although the GNR offers some specific ideas for agriculture, social protection, education, health, water, sanitation, and hygiene, there are several actions that nutrition stakeholders need to take, whatever the sector. These include making the case to other sectors that they can further their own goals by using a nutrition lens; incorporating nutrition goals, indicators, and targets in sector strategies and log-frames; work with partners to use a nutrition lens to develop specific nutrition-enhancing practices and actions within their sector; deploy interventions in high malnutrition areas; engage women in design and implementation; focus on key stages in the life cycle; and incorporate nutrition-specific interventions within broader platforms (25, 30–33). Investments in nutrition are probably increasing, but the picture is unclear Most countries are unable, at present, to identify and track their financial commitments to nutrition (34). This is due to a combination of factors: lack of agreement on what to include and exclude, an absence of data that are sufficiently disaggregated to apportion spending to nutrition, and a lack of capacity to undertake a classification exercise. Several tools exist to accomplish this, and investments will need to be made to build the organizational capacity to use them. Thirteen external funders provided data to the GNR on their nutrition investments. Between 2010 and 2012, commitments and disbursements to nutrition-specific interventions increased by 39% and 30%, respectively. Nutrition-sensitive commitments declined by 14%, but nutrition-sensitive disbursements for the 10 donors that reported such data increased by 19%. Although these changes are encouraging, the percentage of external funding to nutrition in total in 2012 was only marginally above 1% of all overseas development assistance. A nutrition spending target for governments and external funders would be a way of focusing more attention on financial resources to nutrition. It would need to be complemented by improved tracking of spending to ensure that the quality of spending is also improved. The scope and quality of policies, laws, and institutions is important for scale-up Policies, laws, and institutions are important for scaling up nutrition actions and impacts. The GNR reports on efforts to assess commitment to nutrition of governments [Hunger and Nutrition Commitment Index (35)], businesses [Access to Nutrition Index (36)], and the extent to which the food-health environment supports healthy choices [FOOD-INFO (37)]. The SUN process score approach is noteworthy for being a participatory measurement process that stimulates reflection and action by nutrition stakeholders on whether they are aligning their institutions to nutrition priorities (38). Accountability in nutrition needs strengthening The features of nutrition outcomes and actions—short- and long-term effects, invisibility of some consequences, and the need for alliances—make the process of identifying commitments, and then monitoring them for accountability, more complex than for many other development issues. Reporting on N4G commitments: challenges and conclusions As well as reporting on publicly available data, the GNR also sought primary, self-reported data on whether the 96 N4G signatories met their 2013 commitments. Reporting on the commitments was challenging for all groups of signatories. Nevertheless, 90% of them responded to requests for updates against their N4G commitments. Very few signatories were assessed by the GNR as “off course,” although there were many “not clear” assessments due to the vagueness of the initial commitment and of the response. In addition, financial tracking systems were not in place to distinguish between nutrition-specific and -sensitive spending and the infrequent implementation of national nutrition surveys made tracking nutrition status commitments difficult. In brief, the signatories were willing to be held to account, but the mechanisms to do so were weak. Notwithstanding these challenges, there were no obvious causes for concern from any group in terms of progress toward N4G targets at this stage in the 2013–2020 reporting period. Ways to strengthen accountability in nutrition The GNR outlines many ways in which nutrition accountability can be strengthened. The role of civil society actors is particularly important, although they need support if they are to be more effective. The reach and ability of civil society organizations to mobilize are invaluable and they are mostly credible in holding governments and others to account. There is potential for citizens to hold duty bearers to account via community scorecards, social audits, and other citizen-level accountability mechanisms (39, 40). National evaluation platforms (41) that bring together existing data and capacities are another promising set of interventions to be piloted and evaluated. A data revolution is needed Credible and timely data are the bedrock of accountability (42). Unfortunately, there are many data gaps in nutrition outcomes, outputs, and inputs. For example, 40% of the 193 UN member countries cannot track >2 of the 4 WHA indicators. For the countries that can report on WHA indicator progress, data are often outdated. For example, nearly 40% of the baseline surveys in Table 2 are from the period 2005–2009. It is challenging to make public policy on the basis of outdated data. To identify data gaps beyond the WHA indicators, we asked, “Where are data gaps constraining attention to important issues and where are they holding back important actions?” We identified 4 indicators—anemia, overweight, wasting, and low birth weight—in which weak data (lack of detailed food consumption data, more accurate assessments of adult obesity, and outdated adjustments to birth weight) are likely to be holding back action. We also identified gaps—in program coverage data for nutrition-specific programs, financial tracking data, data on a system’s capacity to scale, program cost data, and disaggregated nutrition status data—that are likely holding back the scale-up of nutrition-specific and nutrition-sensitive interventions. A lack of broadly based but high-quality country case studies is also stifling learning about the effectiveness of different strategies, approaches, and combinations of interventions in reducing malnutrition at scale. Decisions about which data gaps are most important to fill need to be undertaken at the national level, based on nutrition policies, plans, and strategies. Priority actions for the next 12 mo Several actions are essential if current levels of commitment to nutrition are to be sustained, intensified, and more effectively converted into impact. First, influential nutrition stakeholders need to work hard to ensure that nutrition is embedded more firmly within the SDG framework. This entails 1) making the case for the inclusion of all 6 WHA indicators within the suite of SDGs, 2) identifying other indicators that are relevant for nutrition (e.g., improved water and sanitation coverage) and making sure they are included and defined in ways that are most useful to nutrition, and 3) establishing inspiring yet attainable 2030 targets for the WHA indicators. Second, several countries need to demonstrate that it is feasible to track resource allocations to nutrition actions. The methods do not have to be exactly comparable across countries if this is likely to slow down the production of credible estimates. Without such estimates, it is impossible to hold governments to account for nutrition-specific or nutrition-sensitive investments. A target for spending on nutrition also needs to be developed, either in relation to health or total government expenditure. This will promote transparency and help shift norms about optimal nutrition spending levels. Third, more analysis needs to be done on coverage levels of nutrition-specific interventions. What are they for different countries and regions and why are some countries doing better than others in scaling up? External funding figures suggest that investments in nutrition-sensitive interventions are only twice those of nutrition-specific interventions. If this is a reliable guide to country spending on nutrition sensitive actions—and given the large sectoral budgets that the nutrition-sensitive interventions are carved out of—the current levels of spending seem low. More guidance and examples are needed on what design features make an intervention at the underlying determinant level become “nutrition sensitive” (43). This will give confidence to those who are being asked to scale up spending in this area. Fourth, innovations in improving stakeholder accountability in nutrition need to be encouraged, piloted, and evaluated. Can community feedback play a key role in signaling a lack of fidelity in program implementation and will this improve fidelity? Will the tracking of commitments at the national or business levels change the behavior of governments or businesses? Can mobile phones be used to highlight failures in implementation of nutrition programs and will this lead to improved implementation? These questions need to be answered with rigorous impact assessments. Fifth, the data gaps that are holding back priority actions need to be addressed. We identify 5 ways to fill key gaps: 1) use existing data better, 2) strengthen existing data collection quality, 3) improve data comparability across countries, 4) improve the frequency of data collection, and 5) collect new data where there is not enough for good accountability. All of these approaches require complementary investments in capacity. Capacities at the individual, organization, and system level are needed if nutrition actions are to be scaled and deliver the expected impact, yet they are often overlooked by potential funders who often do not know where to begin investing. Better data on capacity needs and gaps will help mobilize and guide investments in this vital area (44). Finally, we need more country-level analyses of what is facilitating and holding back nutrition improvements. It is at the national and subnational level where nutrition theory becomes practice. How do interventions fit together? How do stakeholders work together? Where are the weak links in the nutrition chain? How do political, capacity, and technical considerations interact? Documenting the different pathways to improved nutrition status provides inspiration and learning for all countries. Research funders and scientific journals need to encourage more high-quality country case studies. Improving nutrition is a quintessential 21st century challenge Malnutrition affects all countries and crosses generational boundaries. Confronting it requires collaborations across sectors and disciplines. It requires all groups in society—governments, civic organizations, and businesses—to come together to address it. These features are typical of other 21st century development challenges such as mitigating and adapting to climate change, managing resource scarcity, and building resilient societies. But the world does not have to wait until the end of the 21st century to see the demise of malnutrition. We know what works. Commitment levels have never been higher. Stronger accountability mechanisms can help convert this energy into impact. They can provide greater transparency in commitments and greater clarity on progress and can offer more accurate signposts to action. Critically, stronger accountability mechanisms empower all stakeholders to put pressure on those who are most responsible for accelerating the world’s nutrition status to actually do so. We acknowledge the contributions of Catherine Gee of the International Food Policy Research Institute for her support in preparing this manuscript. LH, KB, KE, and JF analyzed the data, performed the statistical analysis, and generated the figures; MB, EB, and JK provided data and advice; LH wrote the manuscript and had primary responsibility for the final content; and EA, MAB, AA, ZB, EC, MdO, RF-A, PF, EK-M, ENK, HN, RN, SO, YM-P, J Randel, J Requejo, TS, EU, and KSR added revisions to the text. All authors read and approved the final manuscript. FIGURE 1 The Global Nutrition Report’s conceptual framework. FIGURE 2 Number of countries experiencing multiple burdens of malnutrition. FIGURE 3 Number of countries that are on course to meet each World Health Assembly global target. TABLE 1 Key messages of the Global Nutrition Report1 1 Multiple malnutrition burdens are the “new normal.” Countries are increasingly dealing with complex combinations of malnutrition problems. This strengthens the case for strategic whole-of-society approaches to addressing malnutrition. 2 Nearly every country in the world is affected by malnutrition. The potential for lesson learning is enormous but is not being exploited. 3 The world is off track to meet the 2025 WHA targets for nutrition. Nevertheless, many countries are making good progress on key WHA indicators. More needs to be understood about how this progress is being achieved. 4 Nutrition needs to be more strongly represented in the SDG framework. At present, it is only explicitly mentioned in 1 of 169 targets. 5 Nutrition-specific interventions have poor coverage. Countries and external funders need to invest in the capacity to scale up these interventions and their impact. 6 Nutrition-sensitive approaches are slow to take off. More guidance needs to be given to nutrition allies in sectors such as agriculture, social protection, water, sanitation, and hygiene about how and why their investments can become more nutrition sensitive. 7 Accountability in nutrition is in need of strengthening. New accountability mechanisms need piloting and evaluation. Financial resources to nutrition need to be made explicit. Nutrition spending targets need to be established to influence resource allocation norms. 8 Data gaps in nutrition need to be filled by governments. Forty percent of UN member states cannot report on the WHA progress. Of the countries that can, 40% of the data they are using are >5 y old. 9 Malnutrition rates can be reduced more quickly than they are now. Strengthening accountability is one way to build greater political commitment to do so. 1 SDG, Sustainable Development Goal; WHA, World Health Assembly. TABLE 2 Progress toward the global WHA nutrition targets1 WHA target Baseline year(s) Baseline status Target for 2025 Required global average annual rate of change Globally on course? Comments2 Number of countries above and below required global rate of change Stunting: 40% reduction in the number of children under 5 who are stunted3 2012 162 million ~100 million (~15% prevalence) 3.9% AARR No Current pace projects 130 million by 2025 (20% reduction) AARR is above or equal to required rate: 21 countries; AARR is below required rate: 89 countries Anemia: 50% reduction in anemia in women of reproductive age 2011 29% 15% 5.2% AARR No Very little movement (was 32% in 2000) — Low birth weight: 30% reduction in low birth weight 2008–2012 15% 10% 2.74% AARR No Little progress globally — Under-5 overweight: no increase in childhood overweight 2012 7% 7% No Upward trajectory is unchecked AARR is constant or decreasing: 50 countries; AARR is increasing: 51 countries Exclusive breastfeeding: increase the rate of exclusive breastfeeding in the first 6 mo up to at least 50% 2008–2012 38% 50% 2.11% AARI No 37% in 2000, 41% in 2012 AARI is above or equal to required rate: 59 countries4; AARI is below required rate: 48 countries Wasting: reduce and maintain childhood wasting to <5% 2012 8% <5% No No progress (was 8% globally in 2013)4 Wasting rate is constant or decreasing: 73 countries; rate is increasing: 51 countries [see Table 3.2 in the GNR (10)] 1 Data from reference 9. AARI, average annual rate of increase; AARR, average annual rate of reduction; GNR, Global Nutrition Report; WHA, World Health Assembly. 2 For more on the methods behind the WHA stunting target, see reference 14. 3 These are AARIs estimated by the writing team on the basis of the last 2 available estimates for exclusive breastfeeding in UNICEF (15). Formal AARIs from UNICEF/WHO are not available at this time. 4 These figures are from reference 7. TABLE 3 Contribution of nutrition to the SDGs1 Proposed SDG Contribution of nutrition to SDG End poverty in all its forms everywhere Preventing stunting in children <36 mo old makes it less likely that they will live in households below the poverty line (18) End hunger, achieve food security and improved nutrition, and promote sustainable agriculture Explicitly refers to 2 WHA indicators (stunting and wasting) Optimal breastfeeding and complementary feeding represent the best dietary start in life (3) A focus on pre-pregnancy and the first part of 1000 d reduces risk of low birth weight and improves women’s nutrition status (2) Ensure healthy lives and promote well-being for all at all ages Micronutrient malnutrition and women stunted in childhood, link to maternal mortality and low birth weight; 45% of all under-5 deaths linked to undernutrition (2) Link of stunting to later NCD onset (19) Reducing overweight and obesity for fewer NCDs (20) Reducing infectious diseases that are linked to nutrition-related morbidity and mortality Ensure inclusive and equitable quality education and promote lifelong learning opportunities Links between nutrition status in first 1000 d, early child development and school grade completion, and achievement, particularly for adolescent girls (1) Achieve gender equality and empower all women and girls Improving the nutrition status of girls, adolescents, and women increases their ability to perform well at school and in the workforce (18) Ensure availability and sustainable management of water and sanitation for all Improvements in nutrition outcomes help reinforce the need for action on water, sanitation, and hygiene as critical determinants of nutrition (21) Promote sustained, inclusive, and sustainable economic growth; full and productive employment; and decent work for all Undernutrition cuts GNP by at least 8–11% (22) Earned income is improved by stunting prevention (18) Reduce inequality within and among countries Stunting rates by wealth quintile demonstrate how current inequality perpetuates future inequality (2) Ensure sustainable consumption and production Research on sustainable food systems and sustainable diets can offer structure and indicators to this policy debate (23) 1 GNP, gross national product; NCD, noncommunicable disease; SDG, Sustainable Development Goal; WHA, World Health Assembly. 1 Supported by 1000 Days, the Bill & Melinda Gates Foundation, the Children’s Investment Fund Foundation, the European Commission, the Government of Canada, Irish Aid, the UK Department for International Development, and the CGIAR Agriculture for Nutrition and Health Programme. This is a free access article, distributed under terms (http://www.nutrition.org/publications/guidelines-and-policies/license/) that permit unrestricted noncommercial use, distribution, and reproduction in any medium, provided the original work is properly cited. 2 Author disclosures: L Haddad, E Achadi, M Ag Bendech, A Ahuja, K Bhatia, Z Bhutta, M Blössner, E Borghi, E Colecraft, M de Onis, K Eriksen, J Fanzo, R Flores-Ayala, P Fracassi, E Kimani-Murage, E Nago Koukoubou, J Krasevec, H Newby, R Nugent, S Oenema, Y Martin-Prével, J Randel, J Requejo, T Shyam, E Udomkesmalee, and KS Reddy, no conflicts of interest. 3 The findings and conclusions in this report are those of the authors and do not necessarily represent the views or the official position of the CDC. 4 The full report, technical notes, and nutrition country profiles can be found at: www.globalnutritionreport.org. Statements of related interests of authors are available on the Global Nutrition Report website at: http://globalnutritionreport.org/governance/ieg/. 28 GNR, Global Nutrition Report; N4G, Nutrition for Growth; SDG, Sustainable Development Goal; SUN, Scaling Up Nutrition; WHA, World Health Assembly. 29 The Global Nutrition Report Independent Expert Group alone is responsible for the classification of countries in this report, which does not necessarily represent the views or assessments of WHO. WHO will report on progress made toward the achievement of the WHA global nutrition targets at its 68th WHA session in May 2015. 30 Mason et al. (16) argue that addressing anemia urgently requires scaling up effective intervention programs such as supplementation with iron and folic acid or multiple micronutrients, fortification of staple foods or condiments, and disease control measures such as malaria control and deworming. They suggest that the lack of attention to this issue stems from the lack of awareness of both its pervasiveness and the slow rate of progress in reducing it. 31 Wasting trends between surveys that are several years apart are not considered meaningful by WHO, and so AARR is not used as a rule for determining whether countries are on or off course. 1 Adair LS Fall CHD Osmond C Stein AD Martorell R Ramirez-Zea M Sachdev HS Dahly DL Bas I Norris SA Associations of linear growth and relative weight gain during early life with adult health and human capital in countries of low and middle income: findings from five birth cohort studies Lancet 2013 382 525 34 23541370 2 Black RE Victora CG Walker SP Bhutta ZA Christian P de Onis M Ezzati M Grantham-McGregor S Katz J Martorell R Maternal and child undernutrition and overweight in low-income and middle-income countries Lancet 2013 382 427 51 23746772 3 Bhutta ZA Das JK Rizvi A Gaffey MF Walker N Horton S Webb P Lartey A Black RE Evidence-based interventions for improvement of maternal and child nutrition: what can be done and at what cost? Lancet 2013 382 452 77 23746776 4 High Level Panel on the Post-2015 Development Agenda Goal 5: ensure food security and good nutrition Report of the High-Level Panel of Eminent Persons on the Post-2015 Development Agenda [cited 2014 Sep 23]. Available from: http://report.post2015hlp.org/digital-report-goal-5-ensure-food-security-and-good-nutrition.html 5 UK Government Nutrition for Growth commitments: executive summary 2013 [cited 2014 Nov 11]. Available from: https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/207274/nutrition-for-growth-commitments.pdf 6 International Food Policy Research Institute Global Nutrition Report: governance 2014 [cited 2014 Dec 12]. Available from: http://global-nutritionreport.org/governance/ 7 United Nations Children’s Fund; World Health Organization; The World Bank UNICEF-WHO-World Bank 2013 Joint Child Malnutrition Estimates 2014 [cited 2014 Nov 11]. Available from: http://www.who.int/nutgrowthdb/estimates2013 8 Stevens GA Finucane MM De-Regil LM Paciorek CJ Flaxman SR Branca F Peña-Rosas JP Bhutta ZA Ezzati M Global, regional, and national trends in haemoglobin concentration and prevalence of total and severe anemia in children and pregnant and non-pregnant women for 1995–2011: a systematic analysis of population-representative data Lancet Glob Health 2013 1 e16 25 25103581 9 WHO Global targets 2025: what is measured gets done 2014 [cited 2014 Aug 26]. Available from: http://www.who.int/nutrition/globaltargets_indicators 10 International Food Policy Research Institute Global Nutrition Report 2014: actions and accountability to accelerate the world’s progress on nutrition Washington (DC) International Food Policy Research Institute 2014 11 Nisbett N Gillespie S Haddad L Harris J Why worry about the politics of childhood undernutrition? World Dev 2014 64 420 33 12 Leroy JL Gadsden P González de Cossío T Gertler P Cash and in-kind transfers lead to excess weight gain in a population of women with a high prevalence of overweight in rural Mexico J Nutr 2013 143 378 83 23343672 13 Lapping K Frongillo EA Nguyen PH Coates J Webb P Menon P Organizational factors, planning capacity, and integration challenges constrain provincial planning processes for nutrition in decentralizing Vietnam Food Nutr Bull 2014 35 382 91 25902597 14 de Onis M Dewey KG Borghi E Onyango AW Blossner M Daelmans B Piwoz E Branca F The World Health Organization’s global target for reducing childhood stunting by 2025: rationale and proposed actions Matern Child Nutr 2013 9 6 26 15 UNICEF Global Databases Nutrition: infant and young child feeding 2014 [cited 2014 Jun 1]. Available from: http://data.unicef.org/nutrition/iycf 16 Mason JB Shrimpton R Saldanha LS Ramakrishnan U Victora CG Girard AW McFarland DA Martorell R The first 500 days of life: policies to support maternal nutrition Glob Health Action 2014 7 2362 17 UN Open Working Group Open Working Group proposal for Sustainable Development Goals 2014 [cited 2014 Nov 11]. Available from: http://sustainabledevelopment.un.org/focussdgs.html 18 Hoddinott J Alderman H Behrman JR Haddad L Horton S The economic rationale for investing in stunting reduction Matern Child Nutr 2013 9 69 82 24074319 19 Uauy R Kain J Corvalan C How can the Developmental Origins of Health and Disease (DOHaD) hypothesis contribute to improving health in developing countries? Am J Clin Nutr 2011 94 6, Suppl 1759S 64S 21543534 20 Popkin BM Adair LS Ng SW Global nutrition transition and the pandemic of obesity in developing countries Nutr Rev 2012 70 3 21 22221213 21 Spears D Ghosh A Cumming O Open defecation and childhood stunting in India: an ecological analysis of new data from 112 districts PLoS ONE 2013 8 e73784 24066070 22 Horton S Steckel RH Malnutrition: global economic losses attributable to malnutrition 1900–2000 and projections to 2050 Lomborg B How much have global problems cost the earth? A scorecard from 1900 to 2050 New York Cambridge University Press 2013 247 72 23 Macdiarmid JI Kyle J Horgan GW Loe J Fyfe C Johnstone A McNeill G Sustainable diets for the future: can we contribute to reducing greenhouse gas emissions by eating a healthy diet? Am J Clin Nutr 2012 96 632 9 22854399 24 Smith L Haddad L Reducing child undernutrition: past drivers and priorities for the post-MDG Era Brighton (United Kingdom) Institute of Development Studies 2014 IDS Working Paper No. 441 25 Ruel MT Alderman H Nutrition-sensitive interventions and programmes: how can they help to accelerate progress in improving maternal and child nutrition? Lancet 2013 382 536 51 23746780 26 Gillespie S Haddad L Mannar V Menon P Nisbett N The politics of reducing malnutrition: building commitment and accelerating progress Lancet 2013 382 552 69 23746781 27 Bhutta ZA Das JK Walker N Rizvi A Campbell H Rudan I Black RE Interventions to address deaths from childhood pneumonia and diarrhoea equitably: what works and at what cost? Lancet 2013 381 1417 29 23582723 28 Menon P Covic NM Harrigan PB Horton SE Kazi NM Lamstein S Neufield L Oakley E Pelletier D Strengthening implementation and utilization of nutrition interventions through research: a framework and research agenda Ann N Y Acad Sci 2014 1332 39 59 24934307 29 FAO The state of food insecurity in the world Rome (Italy) FAO 2014 30 World Bank Improving nutrition through multisectoral approaches Washington (DC) World Bank 2013 31 UN Standing Committee on Nutrition The nutrition sensitivity of agriculture and food policies: a synthesis of eight country case studies Geneva (Switzerland) UN Standing Committee on Nutrition 2013 32 World Vision International Importance of nutrition-sensitive agriculture 2014 [cited 2014 Jun 1]. Available from: www.wvi.org/nutrition/nutrition-sensitive-agriculture 33 Fanzo J Curran S Remans R Mara V Briseño JS Cisewski D Denning G Fracassi P Simulating potential of nutrition-sensitive interventions New York Columbia University, Earth Institute, Center on Globalization and Sustainable Development 2014 [cited 2014 Sep 1]. Available from: http://scalingupnutrition.org/wp-content/uploads/2014/07/SUN-Report-CU_Jan-21-2014.pdf 34 Picanyol C Tracking investments on nutrition Working Paper—final Geneva (Switzerland) SUN Movement 2014 [cited 2014 Sep 1]. Available from: http://scalingupnutrition.org/wp-content/uploads/2013/02/140120-Tracking-Investments-on-Nutrition.pdf 35 te Lintelo DJH Haddad LJ Lakshman R Gatellier K The Hunger and Nutrition Commitment Index (HANCI 2013): measuring the political commitment to reduce hunger and undernutrition in developing countries Brighton (United Kingdom) Institute of Development Studies 2014 IDS Evidence Report No. 78 36 Access to Nutrition Foundation Access to Nutrition Index: Global Index 2013 [cited 2014 Dec 12]. Available from: http://www.accesstonutrition.org/global-index-2013 37 Swinburn B Sacks G Vandevijvere S Kumanyika S Lobstein T Neal B Barquera S Friel S Hawkes C Kelly B INFORMAS (International Network for Food and Obesity/Non-Communicable Diseases Research, Monitoring and Action Support): overview and key principles Obes Rev 2013 14 1 12 38 Scaling Up Nutrition Monitoring progress—SUN Monitoring Pilot Workshop 2014 [cited 2014 Nov 11]. Available from: http://scalin-gupnutrition.org/news/monitoring-progress-sun-monitoring-evaluation-workshop-in-rwanda#.VC7RTTleflK 39 Björkman M Svenssonn J Power to the people: evidence from a randomized field experiment on community-based monitoring in Uganda Q J Econ 2009 124 735 69 40 Swain B Sen PD Bridging the malnutrition gap with social audits and community participation IDS Bull 2009 40 95 102 41 Bryce J Vitora CG National Evaluation Platform Approach [policy brief] Baltimore (MD) Institute for International Programmes 2011 Available from: http://www.jhsph.edu/departments/international-health/centers-and-institutes/institute-for-international-programs/projects/nep-docs/NEP-Policy-Brief-2011.pdf 42 UNICEF; WHO Countdown to 2015 Fulfilling the health agenda for women and children: the 2014 report Geneva (Switzerland) UNICEF/World Health Organization 2014 43 Haddad L Isenman P Which aid spending categories have the greatest untapped potential to support the reduction of undernutrition? Some ideas on moving forward Food Nutr Bull 2014 35 266 76 25076774 44 Sodjinou R Bosu WK Fanou N Déart L Kupka R Tchibindat F Baker S A systematic assessment of the current capacity to act in nutrition in West Africa: cross-country similarities and differences Glob Health Action 2014 7 24763 25034256

PMC005xxxxxx/PMC5129747.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 7503056 4435 J Am Chem Soc J. Am. Chem. Soc. Journal of the American Chemical Society 0002-7863 1520-5126 26938470 5129747 10.1021/jacs.6b00633 NIHMS828791 Article Molecular Characterization of the Cercosporin Biosynthetic Pathway in the Fungal Plant Pathogen Cercospora nicotianae Newman Adam G. Townsend Craig A. * Department of Chemistry, Johns Hopkins University, Baltimore, MD, 21218, USA Corresponding Author: ctownsend@jhu.edu 9 11 2016 16 3 2016 30 3 2016 30 11 2016 138 12 42194228 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. Perylenequinones are a class of photoactivated polyketide mycotoxins produced by fungal plant pathogens that notably produce reactive oxygen species with visible light. The best-studied perylenequinone is cercosporin—a product of the Cercospora species. While the cercosporin biosynthetic gene cluster has been described in the tobacco pathogen Cercospora nicotianae, little is known of the metabolite’s biosynthesis. Furthermore, in vitro investigations of the polyketide synthase central to cercosporin biosynthesis identified the naphthopyrone nor-toralactone as its direct product—an observation in conflict with published biosynthetic proposals. Here, we present an alternative biosynthetic pathway to cercosporin based on metabolites characterized from a series of biosynthetic gene knockouts. We show that nor-toralactone is the key polyketide intermediate and the substrate for the unusual didomain protein CTB3. We demonstrate the unique oxidative cleavage activity of the CTB3 monooxygenase domain in vitro. These data advance our understanding of perylenequinone biosynthesis and expand the biochemical repertoire of flavin-dependent monooxygenases. INTRODUCTION The plant pathogenic Cercospora species are a widespread and destructive genus of ascomycetous fungi characterized by their production of the phytotoxin cercosporin (1, Figure 1b).1 Cercosporin belongs to the perylenequinone natural product family. The perylenequinone metabolites share a characteristic core architecture (2) that is essential to their toxicity.2,3 Multiple perylenequinone natural products have been identified from fungal and aphidian sources.4 Cercosporin was first isolated in 1957 from Cercospora kikuchii T. Matsu & Tomoyasu—a fungal pathogen of soybeans.5,6 Perylenequinone metabolites contain the same highly oxidized conjugated pentacyclic core regardless of varied substituents—primarily at positions C7 and C7′—and are typically C2-symmetric. Notably, most perylenequinone metabolites are helically chiral, demonstrating a preference for one particular atropisomer despite their conjugated, sp2-hybridized carbon core.7–9 Steric clashes between substituents in the mature dimeric metabolites account for the observed helical chirality. The biosynthetic origin of this property is unknown. Cercosporin—like all perylenequinone metabolites—functions as a photosensitizing agent.10 Upon absorption of visible light, cercosporin facilitates remarkably efficient transfer of this energy to O2 (quantum yield 0.8111), leading to the production of the potent reactive oxygen species (ROS) singlet oxygen (1O2) and superoxide radical (O2•−). These ROS cause indiscriminate damage to a variety of cellular targets including cell membranes, nucleic acids, proteins, and lipids.12–14 Peroxidation of the cell membrane is particularly pernicious and is the primary mode of toxicity, causing ion leakage from the host organism.12,13 A direct cellular target of cercosporin has not been found but it is believed that its toxicity is entirely attributed to indiscriminant damage by ROS. As a result, cercosporin exerts broad toxicity to bacteria, fungi, and mice.15–17 Production of cercosporin by Cercospora nicotianae—a fungal pathogen of tobacco—is crucial to the fungus’s pathogenicity and is concomitant with lesion formation on tobacco leaves.18–20 Multiple putative resistance mechanisms have been proposed in C. nicotianae, but the primary mode of resistance is through reversible reduction of the perylenequinone moiety to dihydrocercosporin (3)—a species with little photosensitizing capacity. Dihydrocercopsorin is oxidized to cercosporin spontaneously upon export from the fungus, restoring its photoactivated toxicity.21–23 Cercosporin was classified early on as a polyketide natural product and bore the characteristic alternating polyketide labeling pattern from acetyl and malonyl building blocks.24 Eventually a gene from C. nicotianae encoding a fungal polyketide synthase—dubbed CTB1 (cercosporin toxin biosynthesis 1)—was identified through restriction enzyme-mediated integration mutagenesis.18 CTB1 is absolutely necessary for cercosporin production and bears all the hallmarks of an iterative, non-reducing polyketide synthase (NR-PKS).25 Using CTB1 as a benchmark, the complete cercosporin biosynthetic gene cluster from C. nicotianae was determined (Figure 1a, Table 1).26 The cluster comprises eight genes, six of which are believed to be responsible for cercosporin assembly (CTB1, 2, 3, 5, 6, and 7).18,20,26,27 The zinc finger transcription factor CTB8 co-regulates expression of the cluster,26 while the major facilitator superfamily (MFS) transporter CTB4 exports the final metabolite.19 In addition to the identified gene cluster, the Zn(II)Cys6 transcription factor CRG1 is implicated in both cercosporin production and resistance.28–30 The Zn(II)Cys6 family of transcription factors is unique to fungi that regulate diverse cellular processes.31 Regulation by CRG1 is complex and poorly understood; however, its expression is implicated in regulation of chemical detoxification, multidrug membrane transport, and antioxidant biosynthesis pathways.30 Cercosporin production is completely dependent upon exposure to light, and C. kikuchii grown in the dark will not accumulate cercosporin.32 The regulatory mechanism governing light dependence is unknown. On the basis of individual homologies of the C. nicotianae gene cluster and retrobiosynthetic analysis, Chen and coworkers proposed a biosynthetic pathway for cercosporin (Figure 1b).26 This proposal hinges upon the metabolite’s C2-symmetry and the authors argued that dimerization of two identical aromatic intermediates 4 would lead to the perylenequinone core. They postulated that CTB1 produces carboxylic acid 5 from which the aromatic intermediate 4 is derived. While this biogenesis appears reasonable and accounts for the putative activity of all the biosynthetic gene cluster members, it is nevertheless likely incorrect. Using an enzyme-deconstruction approach, we previously characterized the in vitro activity of CTB1.25 Unexpectedly, the naphthopyrone nor-toralactone (6) is the unambiguous in vitro product of CTB1 (Figure 1b). The identification of this intermediate is problematic, as there is no clear member of the gene cluster that could presumably open the pyrone moiety—an event that must occur to access the perylenequinone core architecture. We present an alternative biosynthetic pathway bolstered by metabolites accumulated in pathway-interrupted C. nicotianae knockouts. Furthermore, we characterize the in vitro activity of CTB3—an unusual didomain protein containing O-methyltransferase and flavin-dependent monooxygenase domains—and demonstrate its role in toralactone formation—an O-methylated congener of nor-toralactone—and subsequent pyrone opening. RESULTS Chemical Identification of Intermediates C. nicotianae has been previously cultivated in liquid potato dextrose broth (PDB) and on solid PDA. While functional cercosporin biosynthetic gene knockout strains of C. nicotianae grew in PDB, reproducible metabolite profiles could not be obtained. Therefore, metabolites were isolated from cultures grown on PDA under constant light, which provided reproducible metabolite profiles. Cultures of wild-type C. nicotianae resulted in a clean metabolic profile with the red-pigmented cercosporin (1) as the principal product (Figure 2a). Cercosporin was absent in the HPLC profiles of all functional cercosporin biosynthetic gene knockout strains. As expected, secondary metabolite accumulation was missing in the ΔCTB1 mutant—the knockout mutant of the central NR-PKS of cercosporin biosynthesis (Figure 2b). The structures of the major accumulated metabolites from C. nicotianae CTB gene cluster knockout strains were elucidated by characteristic UV spectra, exact masses, mass fragmentation patterns, and where possible NMR spectra. We identified the previously characterized naphthopyrones nor-toralactone (6) and toralactone (7) both as metabolites of the ΔCTB3c mutant (Figure 2d). Also accumulated in the ΔCTB3c mutant was the oxidation product of nor-toralactone, naphthoquinone 8. Two previously unobserved naphthoquinones were isolated from the ΔCTB6 and ΔCTB5 mutants, which we called cercoquinone A (9) and cercoquinone B (10), respectively (Figure 2f,e). No major compounds were observed in the extracted metabolite profiles for the ΔCTB2 and ΔCTB7 mutants (Figure 2c,g). The profiles for each of these mutants were similar to that of the ΔCTB1 mutant. Proposed structural assignments for new compounds cercoquinones A and B are supported by their 1H NMR spectra, HRMS, mass fragmentation patterns and UV spectra (see Supporting Information). Phenotypic Analysis of Gene Knockout Strains The mycelia of wild-type C. nicotianae were blood red in color, with pigmentation occurring at about 4 days after inoculation (Figure 2a). The pigmented metabolites were primarily concentrated in the mycelia with a small amount exported into the agar surrounding individual colonies. The mycelia of each knockout strain displayed a different pigmentation from wild-type C. nicotianae, with pigmentation similarly occurring at about 4 days after inoculation. The ΔCTB1 mutant did not display any pigmentation (Figure 2b). The ΔCTB3c mutant adopted a dark yellow-brown coloration, with slight export of pigmented metabolites into the agar (Figure 2d). The mycelia of both ΔCTB6 and ΔCTB5 mutants turned a dark orange-red with significant export of colored compounds into the agar (Figure 2f,e). Although they did not accumulate any observed extractable secondary metabolite observed by HPLC, the mycelia of the ΔCTB2 and ΔCTB7 mutants both adopted a yellow-brown color with some export of pigmented compounds into the agar (Figure 2c,g). In addition to the pigmentation phenotypes, the average colony diameter for each individual strain was distinctive after 7 days of growth (Figure 3b). The average colony diameters for the wild-type, ΔCTB3c, and ΔCTB6 strains were the smallest at 15.5 ± 0.8, 15.7 ± 1.1, and 15.9 ± 1.0 mm (average ± standard deviation, n = 37). The average colony diameter for the ΔCTB5 mutant was of intermediate size at 17.1 ± 1.1 mm, while the average colony diameters for the ΔCTB1, ΔCTB2, and ΔCTB7 mutants were the largest at 20.8 ± 1.3, 19.2 ± 1.4, and 18.7 ± 1.8 mm, respectively. The average colony diameter showed an inverse correlation with small molecule accumulation; the strains with identifiable cercosporin metabolites (wild-type, ΔCTB3c, ΔCTB6, and ΔCTB5) displayed smaller colony diameters while the strains without identifiable extractable intermediates (ΔCTB1, ΔCTB2, and ΔCTB7) displayed larger colony diameters. Cercosporin Complementation Assay To test whether accumulated metabolites represented on-pathway intermediates, we conducted a complementation assay in which pairs of CTB gene cluster knockout strains were grown adjacent to one another. Cercosporin biosynthetic complementation was indicated by red pigmentation at the colony-colony interface. The only pair that could successfully complement cercosporin biosynthesis was the ΔCTB1 and ΔCTB3c mutant pair (Figure 3a). Clear complementation was observed between 4 and 5 days after inoculation between colonies spotted approximately 0.5 cm apart. Pigmentation was isolated to the ΔCTB1 mutant with no clear accumulation of cercosporin in the ΔCTB3c mutant. We attempted to detect cercosporin extracted from agar plugs of the colony-colony interface using a previously described spectrophotometric assay33; however, the amount of accumulated cercosporin was below the detection limit. No other mutant pair showed clear cercosporin complementation (Figure S3, Supporting Information). Genetic Analysis of the Cercosporin Gene Cluster The antiSMASH algorithm was used to identify gene clusters homologous to the CTB gene cluster.34 The top five sequences all came from fungal plant pathogens (Figure 3c). Of these sequences, only one was associated with a known natural product, the hypocrellin A biosynthetic cluster from Shiraia sp. slf14 responsible for the formation of the perylenequinone hypocrellin A (13).35 Five homologous CTB genes were shared among all identified clusters: homologs of CTB1, 3, 2, 5, and 4. The CTB1 homologs contained the canonical NR-PKS domain architecture as determined by a BLASTp search. All of the homologs of CTB3 displayed the same unique didomain architecture with a predicted N-terminal O-methyltransferase and a C-terminal flavin-dependent monooxygenase. In Vitro Analysis of CTB3 Activity Initially deducing that it followed CTB1 on the cercosporin pathway, we investigated the in vitro activity of CTB3 toward nor-toralactone—the in vitro product of CTB1. Because of its unique didomain architecture, we dissected CTB3 into its constituent domains—the N-terminal O-methyltransferase (CTB3-MT) and the C-terminal flavin-dependent monooxygenase (CTB3-MO)—for ease of expression and to investigate the activity of each domain individually (Figure S1, Supporting Information). Both CTB3 and CTB3-MO were purified as holo enzymes with the FAD cofactor bound to the proteins. The presence of FAD was confirmed by HPLC and spectrophotometric analysis (Figure S2, Supporting Information). Cosubstrates SAM and NADH were included in reactions of CTB3 as needed. SAM served as the methyl donor for CTB3-MT and NADH served as the reductant for CTB3-MO—transferring a hydride equivalent to FAD at the initiation of a catalytic cycle. In vitro reactions were conducted with both nor-toralactone (6) and toralactone (7) as substrates. Initial reactions containing 10 μM CTB3 clearly processed these substrates; however, the product profiles were ambiguous. Furthermore, the CTB3-MO domain processed both potential substrates with nor-toralactone being converted to a host of products. While these species were not fully characterized, MS and UV data were indicative of perylenequinone-like products with the clear loss of a single carbon. In order to simplify analysis of these enzymatic reactions, significant effort went into optimizing the experimental conditions. We found that lower protein concentrations (1 μM) and simplified sample preparation (filtration only before HPLC analysis) were necessary for reproducible and robust analysis. Reactions of CTB3-MT showed turnover of nor-toralactone to toralactone (Figure 4a). Similarly, reactions containing a low concentration of protein (0.1 μM) showed partial methylation of nor-toralactone to toralactone. Together, these results indicated that toralactone served as an intermediate of the CTB3 catalytic cycle and was the true substrate for the CTB3-MO domain despite its apparent activity towards nor-toralactone. Postulating that the programming of the catalytic cycle between the MT and MO domains was in part kinetically controlled, we conducted reactions that were allowed to proceed for 60 min before quenching—significantly longer than the 7–10 min reactions initially employed. Reactions of the CTB3-MO domain with toralactone serving as the substrate resulted in turnover to o-naphthoquinone 12—cercoquinone C (Figure 4b). Cercoquinone C was also observed in reactions of deconstructed CTB3 with nor-toralactone as the substrate—albeit with a coterie of other uncharacterized products (Figure 4c). These additional products were absent in reactions with toralactone. Additionally, toralactone is an observed intermediate of this reaction along with the off-pathway oxidation product, naphthoquinone 8. The formation of cercoquinone C was unexpected. Presumably, this product is formed through its hydroquinone (14) even though that species was not directly observed. Assuming that oxidation was spontaneous, we postulated that conducting in vitro reactions under reductive conditions could have prevented the presumptive oxidation. Reactions under reductive conditions (1 mM DTT) resulted in the formation of a new p-naphthoquinone 11—cercoquinone D (Figure 4d–f). Cercoquinone D was only observed when DTT was present. It was observed in either reactions of combined CTB3-MT and CTB3-MO or intact CTB3 with nor-toralactone and reactions of the CTB3-MO domain with toralactone. Time-course experiments showed that the appearance of either cercoquinone C under nonreductive conditions and cercoquinone D under reductive conditions was coincident with the consumption of toralactone by CTB3-MO and no other intermediates were detectable. Exhaustive attempts to capture the putative intermediate 14 through methylation by CTB2, an O-methyltransferase, were unsuccessful. Reactions contained either CTB2 and CTB3 or CTB2 and CTB3-MO or CTB2 and CTB3-MT, with nor-toralactone or toralactone, respectively, serving as a substrate. As expected, CTB2 was unsuccessful at methylating nor-toralactone. Furthermore, it did not alter cercoquinone C and D formation by CTB3, as previously observed (Figure S4, Supporting Information). Structural assignments of cercoquinones C and D were supported by HRMS, mass fragmentation, and UV spectral data. Insufficient amounts of material made structural analysis by NMR for these products unfeasible. Structural analogs of both cercoquinones C and D containing the same quinone cores display similar UV spectra.36 The fragmentation data were highly characteristic (Schemes S2 and S3, Supporting Information). As with cercoquinones A and B, fragmentation ions were observed that comport with the naphthoquinone core. It should be noted that 1,2-naphthoquinones and 1,4-naphthoquinones each fragment in distinct ways.37 Additionally, substituents at the C3 position—the 2-oxopropyl moieties in cercoquinones C and D—can influence the fragmentation of the quinone ring.37–39 Mass fragmentation results for cercoquinones C and D are presented in detail in the Supporting Information. DISCUSSION The results presented herein resolve the uncertainties about cercosporin biosynthesis and point to a common biosynthetic pathway for the perylenequinone natural products.25,26 We propose a revised biosynthetic scheme for cercosporin based upon the accumulated metabolites of functional biosynthetic gene knockout strains (Figure 5). First, CTB1 acts according to its established in vitro chemistry producing nor-toralactone (6).25 The bifunctional enzyme CTB3 methylates nor-toralactone to toralactone (7) before conducting an unusual oxidative aromatic ring opening producing metabolite 14. The O-methyltransferase CTB2 is proposed to methylate the nascent OH-6 of intermediate 14—blocking further oxidation at this site and yielding compound 15—before the reductase CTB6 reduces the 2-oxopropyl ketone at position C7, giving naphthalene 16. CTB5 is thought to be responsible for homodimerization of intermediate 16 with CTB7 installing the dioxepine moiety, finally producing cercosporin (1). We suggest that several of the accumulated metabolites correspond to oxidation products of true on-pathway intermediates. The metabolites cercoquinone A (9) and cercoquinone B (10) from the ΔCTB6 and ΔCTB5 mutants, respectively, fall into this category. Both quinone products derive from the oxidation of electron-rich naphthalene derivatives. The oxidation of electron-rich aromatic metabolites is common and spontaneous—often leading to stabilized products.40 1H NMR and mass fragmentation data support these structural assignments, especially with respect to the sites of oxidation and methylation. Indeed, the observed fragmentation ions agreed with established fragmentation mechanisms for naphthoquinone products.37–39 The inability of either ΔCTB6 or ΔCTB5 mutants to complement cercosporin activity as secretor-converter pairs bolsters the argument that cercoquinones A and B are rapidly formed off-pathway byproducts by spontaneous oxidation. Interestingly, cercoquinone A displays the loss of a methyl group installed earlier through the activity of the O-methyltransferase domain of CTB3. The loss of the methyl at this position is, however, unambiguous given the metabolite’s characteristic mass fragmentation pattern; in particular, the retention of the methoxyl in both the ring contraction carbocation and the ring-opened oxonium ion confirms its structural assignment. We envision two possible scenarios that could account for the elimination of this moiety (Scheme 1). In the first possibility, it is simply eliminated as methanol following oxidation and addition of water to the p-quinone activated by internal hydrogen bonding. In the second possibility, the methyl position is removed enzymatically. The methoxy at this position is eventually incorporated into the dioxepine functionality of the final product we believe through the activity of CTB7. While the mechanism of this transformation is unknown, by analogy to studies of methylenedioxy formations,41,42 it would likely involve oxidative loss. Given the structural similarity of cercoquinone A to the monomeric unit of cercosporin, we suggest that CTB7 could accept this metabolite as an alternative substrate and catalyze the elimination of the methyl in question. The absence of any extractable cercosporin pathway metabolites in the ΔCTB2 and ΔCTB7 mutants is notable given that they are absolutely necessary to sustain cercosporin biosynthesis. Naphthoquinone secondary metabolites in fungi have been shown to inhibit fungal growth.43 Both ΔCTB2 and ΔCTB7 mutant strains have more robust growth than either wild-type or mutant strains that accumulate significant amounts of naphthoquinone byproducts. Despite their limited metabolic profiles, both ΔCTB2 and ΔCTB7 mutant strains appear pigmented while the ΔCTB1 mutant strain is colorless. The putative cercosporin pathway intermediates produced in these mutant strains would likely be electron-rich, hydroxylated naphthalene analogs similar to the components of melanin.44 Melanin—a crucial biological pigment—is produced in fungi by the polymerization of polyketide-derived 1,8-dihydroxynaphthalene monomers. It is a critical metabolite in plant pathogens as it provides mechanical strength to cellular structures responsible for host invasion.45 Furthermore, melanin is highly insoluble and would not contribute to the extracted metabolite profiles analyzed in the current study. Together, the data suggest that the metabolites of the ΔCTB2 and ΔCTB7 mutants are likely captured by a melanin-like self-polymerization pathway resulting in incorporation into insoluble, pigmented polymers. As has been shown previously, the in vitro product of CTB1, an NR-PKS, is naphthopyrone nor-toralactone (6).25 The evolution of nor-toralactone by the NR-PKS presents an immediate problem. In order to synthesize cercosporin from nor-toralactone, the pyrone ring must be opened; however, there is no apparent enzyme in the CTB cluster that could carry out this transformation through conventional lactonase-like hydrolysis—the common pathway for this type of reaction.46 The accumulation of nor-toralactone in the ΔCTB3c mutant simultaneously confirms the importance of this metabolite while implicating a potential candidate in CTB3 for pyrone ring opening. Furthermore, the successful complementation of cercosporin biosynthesis in the ΔCTB1/ΔCTB3c mutant pair corroborates the view that nor-toralactone is an on-pathway intermediate. CTB3 is an unusual didomain enzyme with an O-methyltransferase domain and a flavin-dependent monooxygenase domain. The in vitro enzymatic reactions presented herein confirm that both domains work in tandem to transform the polyketide product nor-toralactone to the on-pathway naphthalene analog 14. Although naphthalene 14 is not directly observed, it is the implied shared precursor to both cercoquinone C (12) and cercoquinone D (11) whose structures are secured by characteristic mass fragments for both ortho- and para-quinones that confirm the positions of oxidation (Schemes S2 and S3, Supporting Information). The direct methylation of nor-toralactone to toralactone by CTB3-MT, as well as the appearance of toralactone in reactions of CTB3 with nor-toralactone are strong indicators that toralactone is a true intermediate of the CTB3 catalytic cycle. Furthermore, its appearance in the metabolite profile of the ΔCTB3c mutant suggests its biosynthetic importance. The ΔCTB3c mutant strain was produced by inserting the knockout cassette into the 5′-terminus of the CTB3 gene corresponding to the CTB3-MO domain. The sequence of the CTB3-MT domain is largely unaltered suggesting the possibility of a functional O-methyltransferase. We suspect that the ΔCTB3c mutant strain could retain some endogenous CTB3-MT activity resulting in the partial methylation of nor-toralactone in the disruption mutant strain. Alternatively, another O-methyltransferase could conceivably convert nor-toralactone to toralactone in vivo; however, this hypothetical O-methyltransferase could not be CTB2—the only other methyltransferase in the CTB cluster as it is incapable of this transformation in vitro. The CTB3-MO domain shares strong primary sequence homology to the canonical p-hydroxybenzoate (17) hydrolase family of flavin-dependent monooxygenases.47 This is a well-studied family of enzymes with a shared catalytic mechanism and preference for electron-rich aromatic substrates. They are responsible for the hydroxylation of simple phenolic systems with a preference for the ortho and para positions. The reaction cycle proceeds in two halves: reductive and oxidative (Scheme 2a).48 The reductive half cycle is initiated with the formation of a ternary complex of enzyme containing oxidized FAD (FADox, 18), NAD(P)H, and substrate. Upon formation of the ternary complex, rapid reduction of FAD by NAD(P)H occurs (FADred, 19). Reduction is contingent on substrate binding or in some cases analogs of the native substrate. NAD(P)+ dissociation occurs at a rate similar to that of reduction. The resulting reduced enzyme and substrate complex has significant kinetic stability. In the oxidative half cycle, oxygen reacts with FADred, forming the characteristic flavin-C4a-hydroperoxide reactive intermediate (21). The deprotonated flavin-C4a-peroxy species (20) has never been observed in these enzymes. Hydroxylation of the substrate occurs through electrophilic aromatic substitution resulting in an enzyme-bound product and flavin-C4a-hydroxide (22) complex. Water is eliminated from the flavin species resulting in FADox and the product is finally released. We argue that a directly analogous mechanism is at play in the CTB3-MO domain with the enzyme installing hydroxyl at the bridgehead position of toralactone (7, Scheme 2b). The hydroxylated intermediate 24 would be far more susceptible to lactone hydrolysis due to the loss of conjugation and aromaticity and the favorable enolate leaving group. Hydrolysis—enzyme catalyzed or spontaneous—would result in acid 25. Loss of carbon dioxide from acid 25 to generate the o-hydroquinone 14 would likely be spontaneous and rapid with the re-establishment of aromaticity providing a large thermodynamic driving force. Alternatively, one could also consider a Baeyer-Villiger-like oxidation, with a flavin-C4a-peroxy nucleophile attacking the pyrone carbonyl. Ring expansion followed by hydrolysis and loss of carbon dioxide would generate the same product 14. We consider this mechanism to be unlikely. The p-hydroxybenzoate monooxygenase enzyme family invariably proceeds through an electrophilic flavin-C4a-hydroperoxy species while flavoenzyme catalyzed Baeyer-Villiger oxidations rely on the nucleophilic flavin-C4a-peroxy species.49 There is precedence for p-hydroxybenzoate monooxygenase family members participating in oxidative aromatic ring cleavage. The enzymes 2-methyl-3-hydroxypyridine-5-carboxylic acid (MHPC, 26a) oxygenase (MHPCO) and 5-pyridoxic acid (5PA, 26b) oxygenase (5PAO) are flavoenzymes of the p-hydroxybenzoate monooxygenase family catalyzing aromatic hydroxylation and subsequent aromatic ring cleavage reactions (Scheme 2c).50 Like conventional flavin monooxygenases, these enzymes use the flavin-C4a-hydroperoxy reactive intermediate (21) in their initial electrophilic aromatic oxidation; however, they undergo a subsequent ring-opening reaction. The mechanism of ring opening has not been definitively established, but it is known that ring opening is enzyme catalyzed in MHPCO and involves the addition of water to the carbonyl intermediate (27a).51 The evolution of cercoquinones C and D and their putative shared intermediate 14 are in keeping with the proposed mechanism (Scheme 2b). The transfer of a hydride equivalent from dihydroquinone 14 to FADox, while not anticipated, is not unreasonable. In the absence of a reductive environment, one could envision this oxidation is the dominant pathway towards cercoquinone C as we observe. Furthermore, the observed consumption of nor-toralactone by CTB3-MO implies an ortho-quinone intermediate. Ortho-quinone analogs of cercoquinone C have been used previously to access the perylenequinone core through simple coupling under acidic conditions.52–54 Assuming nor-toralactone serves as an alternative substrate for CTB3-MO, the observed but uncharacterized dimeric products could likely arise through analogous couplings. Under reductive conditions (DTT), cercoquinone C upon release would be rapidly reduced back to species 14. Electron-rich naphthalene species such as compound 14 are highly prone to spontaneous oxidation. One could envision the oxidation of compound 14 to cercoquinone D as being rapid and complete, even under reductive conditions. The apparent instability of intermediate 14 could explain the observed preference for oxidation of ortho-quinone cercosporin C. In the quinone oxidation state, this species is protected from subsequent oxidation, ensuring biosynthetic fidelity. Structural analysis of known fungal perylenequinone natural products reveals several common features among this family of metabolites (Figure 6).54 First, the 7, 7′ moieties are derived from 2-oxypropyl side chains. Second, the 2, 2′ substituents are always methoxyl, or in the case of cercosporin a derivative thereof. Third, the 6, 6′ substituents are always methoxyl or derivatives thereof. Fourth, the perylenequinone core architecture remains unaltered. These core structural features suggest a common biosynthetic pathway for the perylenequinone metabolites. Indeed, the activities of CTB1, CTB3, CTB2, and CTB5 delineated here would account for each of these features with CTB1 producing the common intermediate nor-toralactone, CTB3 methylating this intermediate at the OH-2 and opening the pyrone ring thereby installing a C6-OH, CTB2 methylating the nascent C6-OH, and CTB5 catalyzing dimerization yielding the perylenequinone carbon core. Interestingly, homologs of each of these enzymes are found in known gene clusters most similar to the CTB cluster. Of these clusters, only one is linked to a known product, hypocrellin A (13).35 Hypocrellin A differs from cercosporin in three key features: the 7, 7′ substituents are retained in the 2-oxypropyl oxidation state and are linked through an intramolecular aldol reaction, the 2, 2′ substituents are methoxyl, and the compound adopts the opposite atropisomeric configuration. Comparing the CTB and hypocrellin A biosynthetic clusters, CTB6 and CTB7 appear unique to the CTB cluster. Given that these enzymes presumptively reduce the 2-oxypropyl substituents and install the dioxepine moiety, respectively, it makes sense that they would be missing in the hypocrellin A cluster, where these structural features are not present. Altogether, these data imply that perylenequinone natural products proceed through a common biosynthetic pathway in which nor-toralactone is initially produced and further processed by a CTB3 didomain homolog. Given its unique coupled activities, CTB3 and its homologs make appealing targets for antifungal agents. CONCLUSION Here we resolve the long-standing ambiguity of cercosporin biosynthesis and propose an alternative biosynthetic pathway. We demonstrate that the naphthoquinone nor-toralactone is an essential intermediate of the cercosporin pathway, further corroborating the previously observed in vitro activity of the NR-PKS CTB1.25 We also characterize the activity of the unusual didomain protein CTB3 showing that its flavin-dependent monooxygenase domain is responsible for an uncommon oxidative aromatic cleavage. Together, these findings further not only our understanding of perylenequinone biosynthesis but also expand the known biochemical repertoire of flavin-dependent monooxygenases—an important class of enzymes in both primary and secondary metabolism. While addressing in detail the initial steps of cercosporin biosynthesis, data presented only partially resolve the later portion of the biosynthetic pathway. The crucial perylenequinone dimerization step(s) and how the absolute configurations of the atropisomers are set remain intriguing, unanswered mechanistic questions. Given the importance of cercosporin and its congeners for plant pathogenicity, disruption of these steps or the rare confluence of CTB3 activities present opportunities for the development of selective antifungal agents against Cercospora species. Supplementary Material SI We thank Dr. Kuang-Ren Chung (National Chung Hsing University, Taichung, Taiwan) for providing C. nicotianae (ATCC© 18366™) wild type and CTB gene cluster knockout strains. We thank Philip A. Storm and Callie R. Huitt-Roehl (Department of Chemistry, Johns Hopkins University, Baltimore, MD, USA) for helpful conversations. The authors declare no financial conflicts of interest. This work was supported by NIH Grant ES001670. Figure 1 The currently proposed cercosporin biosynthetic pathway. (a) The cercosporin toxin biosynthetic (CTB) gene cluster has been identified in C. nicotianae. (b) The proposed cercosporin biosynthesis hinges upon the formation of carboxylic acid 5 by the NR-PKS CTB1. The direct product of CTB1 is nor-toralactone (6), precluding the proposed biosynthetic scheme. Figure 2 Metabolic profiles of CTB gene cluster mutant strains. Chromatograms at 250 nm of extracted metabolite profiles for (a) wild-type, (b) ΔCTB1, (c) ΔCTB2, (d) ΔCTB3c, (e) ΔCTB5, (f) ΔCTB6, and (g) ΔCTB7 strains displayed along with images of the mycelia for each strain. The metabolites were prepared at a concentration of 10 cm2 colony surface area per mL in methanol. Identified cercosporin intermediate metabolites are displayed. Figure 3 Phenotypic and genetic analysis of the CTB cluster. (a) Cercosporin biosynthesis was complemented at the colony-colony interface of the ΔCTB1/ΔCTB3c mutant pair (top/bottom, respectively). The numbers indicate the distance that colonies were inoculated apart from one another in cm. (b) Average colony diameters of CTB disruption mutants are displayed (n = 37). Error bars represent the 95% confidence interval. Stars indicate statistically significant difference from wild-type (p < 0.01). (c) Comparison of gene clusters similar to the CTB gene cluster of C. nicotianae (top). Figure 4 Product profiles of in vitro reactions of CTB3. The 280 nm chromatograms of the following reactions are displayed: (a) CTB3-MT with nor-toralactone, (b) CTB3-MO with toralactone, (c) CTB3-MT and CTB3-MO with nor-toralactone, (d) CTB3-MT and CTB3-MO with nor-toralactone under reductive conditions, (e) CTB3-MO with toralactone under reductive conditions, and (f) CTB3 with nor-toralactone under reductive conditions. Peaks for nor-toralactone (6) and toralactone (7) are indicated along with peaks for products cercoquinone C (12) and cercoquinone D (11). A peak for DTT and other cosubstrates are observed, as applicable. Figure 5 Proposed cercosporin biosynthetic pathway. On the strength of observed pathway intermediates, in vitro chemistry, phenotypic, genetic, and pairwise complementation a revised biosynthetic scheme for cercosporin is presented. Figure 6 Structures of known fungal perylenequinone natural products. Common architectural features are observed in the family. The occurrence of methyoxy at 2, 2′ and 6, 6′ positions are invariable. As is the 2-oxypropyl derivatives at 7, 7′. The formation of the perylenequinone core also appears to be conserved. Scheme 1 Proposed formation of cercoquinone A. Two mechanisms were considered. Addition of water followed by elimination of methanol (pictured) or enzymatic demethylation. Cercosporin is shown with a monomeric unit shown in red. CTB7 is postulated to form the dioxepine ring, a transformation that would require the elimination of a methyl group at the position of demethylation in cercoquinone A. Scheme 2 Proposed mechanism for CTB3 flavin-dependent monooxygenase domain. (a) General mechanism of p-hydroxybenzoate (17) hydroxylase family members. (b) Proposed mechanism for CTB3 catalyzed oxidative aromatic ring cleavage of toralactone (7). (c) Mechanism of MHPCO and 5PAO oxidative aromatic ring cleavage. Table 1 The cercosporin biosynthetic gene cluster from C. nicotianae. Gene Accession Number Length (bp) Intron Amino Acids Domains and Motifs InterPro Annotation Predicted Function CTB1 AY649543 7036 8 2196 PKS β-ketoacyl synthase IPR020841 nor-toralactone PKS acyl transferase IPR020801 synthase Polyketide product template domain IPR030918 Acyl carrier protein-like IPR009081 PKS thioesterase domain IPR020802 CTB2 DQ991505 1439 1 461 O-methyltransferase, family 2 IPR001077 O-methyltransferase CTB3 DQ355149 2734 2 871 O-methyltransferase, family 2 IPR001077 O-methyltransferase Monooxygenase, FAD- binding IPR002938 FAD-dependent monooxygenase CTB4 DQ991506 1696 3 512 Major facilitator superfamily domain IPR020846 MFS transporter CTB5 DQ991507 1380 0 459 FAD linked oxidase, N- terminal IPR006094 FAD-dependent oxidoreductase CTB6 DQ991508 1074 0 357 NAD(P)-binding domain IPR016040 NAD(P)H-dependent ketone reductase CTB7 DQ991509 1401 1 450 Monooxygenase, FAD- binding IPR002938 FAD-dependent monooxygenase CTB8 DQ991510 1245 1 397 Zn(2)-C6 fungal-type DNA binding domain IPR001138 Transcription factor Aflatoxin regulatory protein IPR013700 Supporting Information Experimental details, CTB3 and CTB2 sequence details, full results of complementation assay and CTB3 + CTB2 in vitro reactions, and compound spectra. This material is available free of charge on the Internet via http://pubs.acs.org. 1 Daub ME Phytopathology 1981 71 213 213 2 Morgan BJ Dey S Johnson SW Kozlowski MC J Am Chem Soc 2009 131 9413 9425 19489582 3 Morgan BJ Mulrooney CA Kozlowski MC J Org Chem 2010 75 44 56 19894744 4 Daub ME Herrero S Chung KR FEMS Microbiol Lett 2005 252 197 206 16165316 5 Kuyama S Tamura T J Am Chem Soc 1957 79 5725 5726 6 Kuyama S Tamura T J Am Chem Soc 1957 79 5726 5729 7 Lousberg RJJC Weiss U Salemink CA Arnone A Merlini L Nasini G J Chem Soc D 1971 1463 1464 8 Yamazaki S Ogawa T Agric Biol Chem 1972 36 1707 1718 9 Nasini G Merlini L Andreetti GD Bocelli G Sgarabotto P Tetrahedron 1982 38 2787 2796 10 Daub ME Hangarter RP Plant Physiol 1983 73 855 857 16663313 11 Dobrowolski DC Foote CS Angew Chem Int Ed Engl 1983 22 720 721 12 Daub ME Plant Physiol 1982 69 1361 1364 16662404 13 Daub ME Briggs SP Plant Physiol 1983 71 763 766 16662903 14 Daub ME Ehrenshaft M Annu Rev Phytopathol 2000 38 461 490 11701851 15 Yamazaki S Okubo A Akiyama Y Fuwa K Agric Biol Chem 1975 39 287 288 16 Fajola AO Physiol Plant Pathol 1978 13 157 164 17 Daub ME Phytopathology 1982 72 370 374 18 Choquer M Dekkers KL Chen HQ Cao LH Ueng PP Daub ME Chung KR Mol Plant Microbe Interact 2005 18 468 476 15915645 19 Choquer M Lee MH Bau HJ Chung KR FEBS Lett 2007 581 489 494 17250832 20 Dekkers KL You BJ Gowda VS Liao HL Lee MH Bau HH Ueng PP Chung KR Fungal Genet Biol 2007 44 444 454 17074519 21 Daub ME Leisman GB Clark RA Bowden EF Proc Natl Acad Sci USA 1992 89 9588 9592 1409670 22 Leisman GB Daub ME Photochem Photobiol 1992 55 373 379 23 Sollod CC Jenns AE Daub ME Appl Environ Microbiol 1992 58 444 449 1610167 24 Okubo A Yamazaki S Fuwa K Agric Biol Chem 1975 39 1173 1175 25 Newman AG Vagstad AL Belecki K Scheerer JR Townsend CA Chem Commun 2012 48 11772 11774 26 Chen HQ Lee MH Daub ME Chung KR Mol Microbiol 2007 64 755 770 17462021 27 Chen HQ Lee MH Chung KR Microbiology 2007 153 2781 2790 17660442 28 Chung KR Jenns AE Ehrenshaft M Daub ME Mol Gen Genet 1999 262 382 389 10517336 29 Chung KR Daub ME Kuchler K Schuller C Biochem Biophys Res Commun 2003 302 302 310 12604346 30 Herrero S Amnuaykanjanasin A Daub ME FEMS Microbiol Lett 2007 275 326 337 17850326 31 Akache B Wu K Turcotte B Nucleic Acids Res 2001 29 2181 2190 11353088 32 Ehrenshaft M Upchurch RG Appl Environ Microbiol 1991 57 2671 2676 16348553 33 Jenns AE Daub ME Phytopathology 1995 85 906 912 34 Weber T Blin K Duddela S Krug D Kim HU Bruccoleri R Lee SY Fischbach MA Muller R Wohlleben W Breitling R Takano E Medema MH Nucleic Acids Res 2015 43 W237 243 25948579 35 Yang H Wang Y Zhang Z Yan R Zhu D Genome Announc 2014 2 e00011 14 24503982 36 Ebnother A Meijer TM Schmid H Helv Chim Acta 1952 35 910 928 37 Vessecchi R Emery FS Lopes NP Galembeck SE Rapid Commun Mass Spectrom 2013 27 816 824 23495028 38 Vessecchi R Emery FS Galembeck SE Lopes NP Rapid Commun Mass Spectrom 2010 24 2101 2108 20552691 39 Vessecchi R Emery FS Galembeck SE Lopes NP J Mass Spectrom 2012 47 1648 1659 23280754 40 Wheeler MH Klich MA Pestic Biochem Physiol 1995 52 125 136 41 Bauer W Zenk MH Tetrahedron Lett 1989 30 5257 5260 42 Rueffer M Zenk MH Phytochemistry 1994 36 1219 1223 43 Medentsev AG Akimenko VK Phytochemistry 1998 47 935 959 9564730 44 Eisenman HC Casadevall A Appl Microbiol Biotechnol 2012 93 931 940 22173481 45 Chen Z Nunes MA Silva MC Rodrigues CJ Jr Mycologia 2004 96 1199 1208 21148942 46 Thomas PW Stone EM Costello AL Tierney DL Fast W Biochemistry 2005 44 7559 7569 15895999 47 Huijbers MME Montersino S Westphal AH Tischler D van Berkel WJH Arch Biochem Biophys 2014 544 2 17 24361254 48 Entsch B van Berkel WJ The FASEB Journal 1995 9 476 483 7737455 49 Mirza IA Yachnin BJ Wang S Grosse S Bergeron H Imura A Iwaki H Hasegawa Y Lau PC Berghuis AM J Am Chem Soc 2009 131 8848 8854 19385644 50 Chaiyen P Arch Biochem Biophys 2010 493 62 70 19728986 51 Chaiyen P Brissette P Ballou DP Massey V Biochemistry 1997 36 8060 8070 9201954 52 Hauser FM Sengupta D Corlett SA J Org Chem 1994 59 1967 1969 53 Merlic CA Aldrich CC Albaneze-Walker J Saghatelian A Mammen J J Org Chem 2001 66 1297 1309 11312960 54 Mulrooney CA Morgan BJ Li X Kozlowski MC J Org Chem 2010 75 16 29 19894746

PMC005xxxxxx/PMC5129837.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 0413066 2830 Cell Cell Cell 0092-8674 1097-4172 27716507 5129837 10.1016/j.cell.2016.08.076 NIHMS819484 Article Host-protozoan interactions protect from mucosal infections through activation of the inflammasome Chudnovskiy Aleksey 12310 Mortha Arthur 123910 Kana Veronika 123 Kennard Andrea 6 Ramirez Juan David 56 Rahman Adeeb 34 Remark Romain 123 Mogno Ilaria 4 Ng Ruby 3 Gnjatic Sasha 23 Amir El-ad David 123 Solovyov Alexander 2 Greenbaum Benjamin 2 Clemente Jose 34 Faith Jeremiah 34 Belkaid Yasmine 78 Grigg Michael E. 6 Merad Miriam 12311* 1 Department of Oncological Science, 1475 Madison Avenue New York, NY 10028 2 The Tisch Cancer Institute, 1475 Madison Avenue New York, NY 10028 3 The Immunological Institute, 1475 Madison Avenue New York, NY 10028 4 Icahn Institute or Genomics and Multiscale Biology and Department of Genetics & Genomic Sciences Icahn School of Medicine at Mount Sinai, 1475 Madison Avenue New York, NY 10028 5 Grupo de Investigaciones Microbiologicas – UR (GIMUR), Universidad del Rosario, Bogotá, Colombia 6 Molecular Parasitology Section, National Institutes of Health, 9000 Rockville Pike, Bethesda, MD 20892 7 Mucosal Immunology Section, National Institutes of Health, 9000 Rockville Pike, Bethesda, MD 20892 8 NIAID Microbiome Program, Laboratory of Parasitic Diseases, National Institute of Allergy and Infectious Diseases, National Institutes of Health, 9000 Rockville Pike, Bethesda, MD 20892 * Correspondence : miriam.merad@mssm.edu (M.M.), Address : 1470 madison avenue, New york NY 10029 9 present address: Department of Immunology, University of Toronto, Toronto, 1 King’s College Cir, ON M5S 1A8, Canada. 10 Co-first authors 11 Lead Contact 10 11 2016 6 10 2016 06 10 2017 167 2 444456.e14 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. While conventional pathogenic protists have been extensively studied, there is an underappreciated constitutive protist microbiota that is an integral part of the vertebrate microbiome. The impact of these species on the host and their potential contributions to mucosal immune homeostasis remain poorly studied. Here, we show that the protozoan Tritrichomonas musculis activates the host epithelial inflammasome to induce IL-18 release. Epithelial-derived IL-18 promotes dendritic cell-driven Th1 and Th17 immunity and confers dramatic protection from mucosal bacterial infections. Along with its role as a “protistic” antibiotic, colonization with T. musculis exacerbates the development of T cell driven colitis and sporadic colorectal tumors. Our findings demonstrate a novel mutualistic host-protozoan interaction that increases mucosal host defenses at the cost of an increased risk of inflammatory disease. Graphical Abstract Introduction The mammalian gut is host to a wide consortium of microbes from diverse phyla including viruses, prokaryotic bacteria and eukaryotic microbes. The latter, broadly referred to as the eukaryome (Lukes et al., 2015), is comprised of a myriad of fungi, helminths and protists. Several protists are known pathogens of the mouse and human intestine, these include the microsporidia (Encephalitizoan cuniculi) (Stentiford et al., 2016), Entamoeba histolytica (Moonah et al., 2013), Toxoplasma gondii (Molloy et al., 2013), Giardia spp. and Cryptosporidium spp. (Kotloff et al., 2013), and the host immune response induced upon colonization with these unicellular protozoan parasites is well studied in both patients and experimental settings. In contrast, it is increasingly evident that a constitutive protistic microbiota, which exists as an integral part of the vertebrate microbiome, inhabits mammalian intestinal tracts. The prevalence and classification of these protists, including stramenopiles (Blastocystis spp.), diplomonads (Enteromonas spp.), amoebozoa (Entamoeba dispar, Entamoeba coli) and parabasalids (Pentatrichomonas hominis, Dientamoeba fragilis), as commensal, pathobionts, or pathogens remains enigmatic and debated (Lukes et al., 2015). The impact of these species on the host in general and, in particular, on the immune system has been practically neglected. In this study, we describe the critical contribution of the rodent parabasalid Tritrichomonas musculis (T.mu), a previously unrecognized commensal of the rodent microbial flora, in shaping the intestinal immune landscape. We show that intestinal colonization with T.mu leads to inflammasome activation in the epithelial compartment and the release of the inflammatory cytokine IL-18, which in turn contributes to host protection against mucosal bacterial infections but exacerbates disease sequelae in animal models of colitis and tumorogenesis. These results uncover a previously unappreciated mutualistic relationship between a protist and its host, and identify the critical contribution of protozoa to mucosal defenses. Results Identification of a gut protozoan commensal in mice Routine phenotypic analysis of gut tissue revealed a significant expansion of the CD45+ hematopoietic cell compartment in the C57BL/6 (B6) mouse colony maintained “in-house” at the Mount Sinai animal facility (Fig. 1A–B) together with increased levels of IgA in the sera and colonic explants compared to commercially available mice purchased from Jackson Laboratories (Jax) (Fig. 1C). Remarkably, analysis of colonic luminal content revealed a large number of IgAintDAPI+ cells in in-house colonies, which were absent in commercial mice (Fig. 1D). Microscopic analysis of fecal material from in-house mice revealed the presence of unicellular flagellated microorganisms that resembled a parabasalid protozoan parasite (Fig. 1E) which were closely adherent to the intestinal epithelial surface (Fig. 1F). Molecular PCR-DNA sequencing at the 18S (Supplementary Fig. 1C) and ITS (Fig. 1H and Supplementary Fig. 1D–E) rDNA locus identified a new protozoan parasite referred to hereafter as Tritrichomonas musculis (T.mu), which is similar to, but distinct from, Tritrichomonas muris. Phylogenetic analysis of T.mu sequences obtained for GAPDH, a-tubulin, EF1a and MDH from metagenomic sequences obtained from FACS-purified T.mu isolated from infected B6 mice established that T.mu is indeed unique, with close ancestry to Tritrichomonas foetus (Supplementary Figure 1F–I). T.mu was also identified within 4 separate animal facilities within the intramural NIH animal facilities (Bethesda, MD) in addition to Mount Sinai animal facility indicating that the parasite was both widespread and common within East Coast research facilities. D. fragilis is the closest human ortholog Humans are likewise host to several enteric parabasalids, such as Pentatrichomonas hominis and Dientamoeba fragilis, but whether these protists are commensals, pathobionts, or pathogens is still debated. To determine if a T.mu orthologous sequence type is common in people, we screened 188 fecal samples collected from healthy adults with no gastro-intestinal clinical symptoms obtained from 9 health districts as part of a Colombian NIH Health Survey. A heterogeneous array of Dientamoeba fragilis sequences was detected in all health districts sampled (31/188; 16.5%) with the highest incidence (6/19; 31.6%) in the Fomeque region (Supplementary Table 1). To assay whether enteric parabasalids were found globally, we screened available Giardia positive fecal material obtained from 96 human individuals collected from South America, Africa, Europe and Asia. We identified a widespread distribution of the human T. mu ortholog (11/96; 11.5%) emphasizing the potentially unappreciated prevalence of T.mu orthologs in humans (Supplementary Fig. 1E). Consistent with prior studies (Treuting et al., 2012), we found little if any T.mu in the small intestine of T.mu-bearing animals prompting us to focus mainly on immune changes in the colonic mucosa (data not shown). To directly address whether the presence of T.mu was responsible for the expansion of gut tissue-resident immune cells and increased IgA levels, we colonized commercial B6 mice with purified T.mu. We used fluorescent activated cell-sorting (FACS) to purify the protozoan (Supplementary Fig. 1A and data not shown) which was quadruple sorted and we confirmed the purity and reproducibility of our sorting method using mass spectrometry analysis of the FACS-sorted T.mu from B6 mice (Supplementary Fig. 1B). Commercial B6 mice were colonized using a single gavage of purified T.mu and analyzed at different times after colonization. The colonization was life-long, and persisted in the offspring of the colonized animals for several generations (data not shown). Importantly, T.mu colonization frequency was identical in mice inoculated via oral gavage to in-house T.mu infected animals (Fig. 1G). Strikingly, and despite the extensive expansion of gut tissue-resident hematopoietic cells we failed to detect histological signs of mucosal injury in T.mu bearing mice, although a mild epithelial and goblet cell hyperplasia was noted in infected animals compared to non-infected controls (Fig. 1I,J). Altogether these results establish that T.mu is a constituent member of the microbial flora of adult mice capable of colonizing mice with no acute harm to the host and regardless of the preexisting microbial communities present. T.mu colonization rapidly shapes the colonic tissue immune landscape To directly address whether T.mu was responsible for the expansion of gut tissue-resident hematopoietic cells, we characterized the immune cell composition of T.mu-colonized mice using Flow cytometry and Cytometry by Time of Flight (CyTOF) analysis of colonic lamina propria (LP) isolated at different times after colonization (Fig. 2A–C). We observed a substantial increase in Ly6Chi monocytes, CD103−CD11b+ DCs, macrophages and neutrophils as early as one week after colonization, compared to PBS-gavaged mice (Fig. 2C–D). In contrast, migratory CD103+DCs and CD103+CD11b+ DCs were reduced in the colonic mucosa, in line with their superior ability to migrate to the mesenteric draining lymph nodes (MLNs) in response to inflammatory signals (Bogunovic et al., 2009; Bogunovic et al., 2012). Changes in myeloid cell composition were associated with increased production of tumor necrosis factor alpha (TNFα) by macrophages and infiltrating Ly6Chi monocytes (Fig. 2D,E). In addition to changes of the innate myeloid compartment, T.mu colonization also affected innate lymphoid cells (ILC) (Spits et al., 2013). While ILC did not expand in T.mu-colonized mice, we observed a substantial increase of IL-5 and IL-13 producing ILC2 in the colonic mucosa, similar to a recent report in the small intestine of mice colonized with the protozoa Tritrichomonas spp (Howitt et al., 2016). However, we also found that activation of ILC2 in T.mu-colonized mice was associated with a significant increase of GM-CSF-producing ILC3 (Supplementary Fig. 2 A – C). We also detected a drastic expansion of interferon γ (IFNγ)-producing CD4+ T helper cells (Th1) cells as early as one week after T.mu colonization, whereas we failed to detect any expansion of IL-5, IL-13 or IL-4 producing Th2 cells (Fig. 2F–I). CD4+IFNγ+Th1 cells peaked and plateaued around 2 months post colonization (Fig. 2J). We also observed an expansion of interleukin-17 (IL-17)-producing CD4+ Th cells (Th17), which increased around 2–3 weeks after colonization, but to a lower level than Th1 cells (Fig. 2F, 2J). In addition, we found that the number of activated CD44hiCD4+ T cells increased in the MLN (data not shown) and colonic LP (Supplementary Fig. 2D), while the number of proliferating CD4+Ki67+ T cells and CXCR3+CD4+ T cells expanded in the colonic LP two weeks after T.mu colonization (Supplementary Fig. 2D–E). These results suggest that colonization with T.mu promotes the priming of tissue-specific T effector responses in the MLN, followed by the accumulation and/or expansion of effector T cells in the colonic LP. To assess whether the drastic colonic immune modulation observed upon T.mu colonization was not solely due to a collateral effect related to the transfer of endosymbiotic or contaminating bacterial taxa we inoculated T.mu free mice with purified T.mu cultured in the presence of 5 antibiotics (Abx), as previously described (Saeki et al., 1983). Importantly, we observed a similar induction of colonic Th1 and Th17 immune responses in mice inoculated with purified and cultured Abx treated T.mu or untreated FACS-sorted T.mu (Supplementary Fig. 2F and Fig. 2F), suggesting that T.mu is the main driver of the colonic immune-modulatory effects observed in T.mu colonized animals. Consistent with the dominant Th1 cell induction, we observed a strong IFNγ-driven gene expression signature in the colonic mucosal epithelia of T.mu colonized mice (Supplementary Fig. 3), prompting us to probe the pathways that promoted T. mu-induction of Th1 and Th17 responses in mice. Division of labor between distinct colonic tissue DC subsets helps shape mucosal immunity upon T.mu colonization Tissue-resident DCs are most potent at driving mucosal-specific T cell effector responses upon migration, and antigen presentation in the T cell zone of tissue-draining LNs enriched in naïve T cells (Iwasaki, 2007). CCR7 is a chemokine receptor that controls DC migration from the mucosal tissues to the MLN, and Ccr7−/− mice lack migratory DCs in the MLN (Jang et al., 2006). Colonization with T.mu did not lead to Th1 and Th17 cell expansion in Ccr7-deficient mice (Fig. 3A), suggesting that DC migration may be required to promote T effector cell responses upon T.mu colonization in vivo. However, since CCR7 is also required for T cell entry into the LN (Forster et al., 1999), failure to induce effector cell responses in T.mu colonized Ccr7-deficient mice may also be due to T cell intrinsic defects. LP DCs are heterogeneous and consist of functionally specialized subsets that arise from distinct differentiation lineages. CD103+CD11b+F4/80−Flt3+ DCs which require the transcription factor Interferon Regulatory Factor 4 (IRF4) for their differentiation, are more potent at driving Th2 and Th17 differentiation in the mucosal tissues (Gao et al., 2013; Murphy, 2013; Persson et al., 2013; Schlitzer et al., 2013; Tussiwand et al., 2015). CD103+CD11b−F4/80−Flt3+ DCs arise independently of IRF4 and require the transcription factors IRF8, Basic Leucine Zipper Transcription Factor ATF-like 3 (BATF3) and Inhibitor of DNA binding 2 (ID2) for their differentiation, and are more potent at driving CD8+ T cell immunity (Bogunovic et al., 2009; Edelson et al., 2010; Helft et al., 2010; Varol et al., 2009). To directly examine the contribution of each DC subset to the induction of T effector cells upon T.mu colonization, we used DC-specific knock-out lines that lack distinct DC subsets. We found that Batf3−/− mice that lack CD103+CD11b− DCs but not CD103+CD11b+ DCs in the colonic LP, failed to expand CD4+IFNγ+ Th cells upon T.mu colonization, whereas expansion of CD4+IL-17+ Th17 cells was intact and slightly increased in the colonic mucosa of T.mu-colonized mice, compared to mice free of T.mu (Fig. 3B,D). In addition to Batf3-deficient mice, Irf8-deficient mice should also be unable to mount T.mu-specific Th1 responses, since IRF8 is required for the differentiation of the CD103+ DC lineage (Helft et al., 2010). Because IRF8 has pleiotropic effects in hematopoietic cell differentiation, we crossed Cd11cCre mice with Irf8flox/flox mice (Irf8ΔDC mice) to specifically delete Irf8 in CD11c+ DCs. Importantly, similar to Batf3−/− mice, Irf8ΔDC mice lacked CD103+CD11b− DC and failed to expand CD4+IFNγ+ Th1 cells upon T.mu colonization (Fig. 3E–F), whereas CD4+IL-17+ Th17 cells expanded similarly in Irf8ΔDC and Irf8flox/flox mice colonized with T.mu. To probe the contribution of CD103+CD11b+F4/80−Flt3+ DC to immune changes induced upon T.mu-colonization, we deleted Irf4 in CD11c+ DCs (Cd11cCre x Irf4flox/flox (Irf4ΔDC)) leading to the depletion of CD103+CD11b+ DCs in the colonic LP (Fig. 3G). Two weeks post colonization with T.mu, expansion of CD4+IFNγ+ Th1 cells was similar in Irf4ΔDC and Irf4flox/flox mice, whereas CD4+IL-17+ Th17 cells expanded less prominently in the colonic mucosa of Irf4ΔDC mice (Fig. 3H). Altogether, these results establish the distinct contribution of CD103+CD11b− DC and CD103+CD11b+ DC to the induction of Th1 and Th17 immunity, respectively, upon T.mu colonization. Given the significant accumulation of Ly6Chi monocytes observed in T.mu-colonized mice, we asked whether monocytes and macrophages help shape T cell responses observed after T.mu colonization. CCR2 is a chemokine receptor that is highly expressed on monocytes and is required for efficient monocyte egress from the bone marrow leading to reduced monocyte accumulation in tissues of Ccr2-deficient mice (Serbina and Pamer, 2006). Interestingly, while Ccr2-deletion led to a drastic reduction of colonic tissue-infiltrating monocytes, we did not observe any changes in the number of IFNγ+ and IL-17+ T effector cells at two weeks following T.mu colonization (Supplementary Fig. 4A), in addition T cell responses were independent of TLR5 signaling and Myd88 signaling in macrophages and intestinal epithelial cells (Supplementary Fig. 6). Importantly, T.mu engraftment efficiency was identical in mice lacking DC or monocytes populations suggesting that differences in T.mu colonization resistance were not responsible for the immune phenotype observed in these mice (Supplementary Fig. 4B). Altogether, these results reveal a critical role for tissue-resident DC to the induction of a diverse effector response to T.mu-colonization. To examine whether T.mu was directly responsible for the induction of Th1 and Th17 responses in the colonic mucosa, we colonized germ-free (GF) mice with T.mu and probed the expansion of IFNγ and IL-17 T cells in the colonic mucosa two weeks after colonization. We found that similar to our observations in conventional mice, T.mu colonization led to an expansion of Th1 and Th17 cells in GF mice. Interestingly, T.mu colonization of GF mice led to a slightly stronger Th17 cell response compared to conventional mice, (Supplementary Fig. 5A–B). Increased Th17 response observed in these mice may be due either to increased T.mu colonization (Supplementary Fig. 5C) or modulation of Th17 responses by the commensal bacterial flora. T.mu shapes colonic adaptive immunity via ASC and IL-18 activation The initiation of T helper responses requires the orchestrated actions of cytokines to induce and amplify their effector polarization. Interleukin-12 (IL-12) and IL-18 have been associated with induction of Th1 immunity (Boraschi and Dinarello, 2006; Sims and Smith, 2010; Trinchieri, 2003). We found that IL-18 was released at high levels in the colonic tissue upon T.mu colonization (Fig. 4A) and that IL-18 receptor alpha (IL-18Rα) was expressed at high levels in colonic-infiltrating T cells (Supplementary Fig. 2G) prompting us to examine whether increased production of IL-18 contributed to the induction of Th1 response in T.mu colonized mice. We found that absence of IL-18 completely abrogated the expansion not only of CD4+IFNγ+ Th1 but also CD4+IL-17+ Th17 cells upon colonization with T.mu, compared to control mice (Fig. 4B). Abrogation of the Th1 and Th17 response observed in T.mu infected il18−/− mice was not due to reduced T.mu colonization as T.mu colonization efficiency was identical in il18−/− mice that have been either gavaged with purified T.mu or cohoused with T.mu infected mice (Supplementary Fig. 4B) Release of mature and bioactive IL-18 requires cleavage of the precursor form of IL-18 by active Caspase-1 (Gu et al., 1997). Activation of Caspase-1 is driven by the assembly of cytosolic multiprotein complexes called ‘inflammasomes’ and requires, in most cases, the adaptor apoptosis-associated speck-like protein containing a CARD (ASC) (Schroder and Tschopp, 2010). IL-18 is produced by many different cell types including intestinal epithelial cells, stromal and hematopoietic cells (Boraschi and Dinarello, 2006; Dinarello et al., 2013; Sims and Smith, 2010). To examine the contribution of the epithelial or hematopoietic compartment to IL-18 production upon T.mu infection, we generated (Il18−/− -> wild type) bone marrow chimeric mice that specifically lack IL-18 in the mucosal hematopoietic compartment. We found that (Il18−/−-> wild type) bone marrow chimeric mice responded similarly to (wild type -> wild type) bone marrow chimeric controls to T.mu colonization, suggesting that epithelial production of IL-18 controlled T.mu driven Th1 and Th17 response in the colon (Fig. 4C). To further probe the key role of IL-18 in host response to T.mu infection, we colonized Asc−/− mice that are unable to produce active Caspase 1 and bioactive IL-18. Consistent with the results observed in Il18−/− mice, Asc−/− mice failed to mount Th1 and Th17 responses upon T.mu colonization, further establishing the key role of the inflammasome in T.mu-associated adaptive immune responses (Fig. 4D). ASC-driven Caspase 1 activation can also cleave the precursor form of IL-1β (Schroder and Tschopp, 2010; Sims and Smith, 2010). However, we found that induction of CD4+IFNγ+ Th1 and CD4+IL-17+ Th17 response upon T.mu colonization was less affected in mice lacking Il1r1 (Fig. 4E). T.mu colonization protects from mucosal bacterial infection through activation of the inflammasome The expansion of a strong Th1 immune response mostly in the large intestine and cecum of colonized mice led us to explore a potential mutualistic role for T.mu in shaping host mucosal defenses. Thus, we probed the contribution of T.mu to modulating the clinical outcome of enterocolitis driven by Salmonella typhimurium infection, one of the leading causes of enterocolitis in developing countries (Barthel et al., 2003). Strikingly, we found that while Salmonella infection causes massive cecal inflammation in mice free of T.mu, the cecum of T.mu-colonized mice remained almost unaffected by the bacterial pathogen (Fig. 5A), despite the fact that Salmonella colonized as efficiently T.mu-inoculated as T.mu-free mice (Fig. 5B). All pathological parameters including colonic infiltration of polymorphonuclear granulocytes, submucosal edema, reduction of goblet cells number and epithelial cell integrity were dramatically improved in mice that were colonized with T.mu prior to the infection with Salmonella (Fig. 5A). Accordingly, we found that Salmonella colony-forming units (CFU) were strongly reduced in the MLN and spleen of mice inoculated with T.mu compared to T.mu free animals (Fig. 5B). Since our results above established that epithelial release of IL-18 was critical to shape T.mu–induced immunity, we asked whether IL-18 contributed to T.mu-mediated protection against mucosal Salmonella infection. Importantly, we found that injection of neutralizing anti-IL-18 antibody strongly reduced the protective effects of the T.mu-colonization in Salmonella-infected mice compared to control mice reflected by increased inflammation (Fig. 5C) and increased Salmonella infection (Fig. 5D) observed in the group treated with IL-18 blocking antibody in comparison to isotype control treated mice. These data establish that the presence of T.mu in the microbial flora significantly protects against mucosal infection through the induction of inflammasome-driven IL-18 release by the colonic epithelium. T.mu colonization increases susceptibility to pathogenic inflammation The sustained levels of inflammatory molecules found in the colonic tissues prompted us to test whether T.mu could also contribute to pathogenic inflammation in infected mice. Thus, conventional or GF, immunodeficient recombination-activating gene 2 (Rag2−/−) mice were colonized with T.mu two weeks prior to injecting highly purified CD4+CD45RBhi effector T cells. Consistent with prior reports, injection of effector T cells in Rag2−/− animals led to consistent weight loss, associated with intestinal inflammation, whereas GF mice were protected from disease (Feng et al., 2010; Read and Powrie, 2001). Importantly, conventional and GF animals colonized with T.mu prior to T cell transfer had a more aggravated disease, with increased weight loss starting approximately two weeks after T cell transfer, which correlated with more severe pathological score compared to control mice (Fig. 6A). Sustained tissue inflammation also prompted us to assess the impact of T.mu on epithelial cell damage. We found that as early as 30 days post-colonization, T.mu led to a significant increase in apoptotic epithelial cells at the tip of the villi and an increased number of proliferating epithelial cells at the villi crypts (Fig. 6B–C), confirming our finding of mild epithelial hyperplasia in T.mu infected mice (Fig. 1I–J). T.mu-induced epithelial cell activation was associated with increased expression levels of epithelial Nitric Oxide Synthase 2 (Nos2), TNFα (Tnfa) and IL-22 target genes (Fig. 6D and Supplementary Fig. 3), which together with IL-17 have been implicated in exacerbating colorectal cancer (CRC) development (Grivennikov et al., 2012; Kirchberger et al., 2013; Popivanova et al., 2008; Shaked et al., 2012 ). The loss of the tumor suppressor gene Adenomatous polyposis coli (Apc) increases the development of intestinal tumors in humans. Apcmin/+ mice that carry a truncation mutation in the Apc gene are commonly used as a model of sporadic CRC (Heyer et al., 1999). Colonization of Apcmin/+ mice with T.mu revealed a significant increase in the burden, but not the size, of the colonic tumors around 3–4 months following colonization (Fig. 6F). These results demonstrate that prolonged T.mu colonization can exacerbate the development of tumors in the large intestine of its vertebrate host (Fig. 6E–F). Discussion Here we report the identification of a previously unknown mutualistic interaction between the intestinal protozoan commensal T.mu and its vertebrate host. We found that T.mu engrafts into an established ecosystem, leading to long-lasting colonization without causing any pathological injuries to its host. We also show that T.mu colonization completely remodeled the composition and functional state of the mucosal tissue-resident immune system and increased host protection against mucosal bacterial infections. Colonization with T.mu led to a significant expansion of Th1 cells and Th17 effector cells in the colonic mucosa which was dependent on distinct DC subsets and their ability to migrate to the draining LN but also required the production of IL-18 by the epithelial cells. These results together with the high expression of IL-18Rα on colonic-infiltrating effector T cells (Supplementary Fig. 2G) suggest that T.mu-specific T cell immunity is likely initiated in the draining LN by migratory colonic DCs and is likely further propagated at the tissue site by epithelial IL-18. It is well established that trichomonad flagellates of insects (i.e., Oxymonads, Trichonymphids and Cristomonads) form endosymbiotic partnerships with bacteria (Hongoh et al., 2008) (Noda et al., 2009), but it is less clear whether the same is true for flagellated trichomonads of the enteric cavities of animals. Whereas many animal-sourced flagellates can only be maintained as xenic cultures, others such as the oral and vaginal flagellates of humans, Trichomonas tenax and Trichomonas vaginalis respectively, can be readily adapted to grow axenically (Diamond, 1962). We found that T.mu cultured in vitro in Abx-rich media or isolated from Abx-treated mice harbor different bacterial taxa (data not shown) but yet produce an equivalent immunomodulation of the colonic mucosa as that from FACS-sorted T.mu transferred from untreated mice, which argues against specific bacterial taxa as the sole agent(s) responsible for the modulation of colonic immunity and supports an instructive role for T.mu in remodeling the mucosal tissue-resident immune system. Regardless, whether the immunomodulatory phenotype observed in infected animals is mediated exclusively by T.mu, an associated microbial symbiont, or from the composite of a protozoan-microbial endosymbiotic interaction, it is our belief that the protist and any associated microbial symbiont should be treated as a single microbial unit that possesses distinct immunomodulatory properties. Importantly, differences in the microbial composition from different mouse strains did not influence T.mu colonization efficiency. Future studies, however, are needed to assess the degree to which shifts in microbial composition or function that naturally occur in response to T.mu colonization contribute to or modulate the colonic immune system. Importantly, the presence of T.mu in several East Coast research institutions and its highly immune-modulatory effects in the host further emphasizes the need for animal cohousing in experimental immunology research. The cost of cohousing experiments would thus need to be taken into account by funding agencies to facilitate the implementation of these additional controls throughout our research community. Interestingly, T.mu colonization induced the activation of ILC3 and ILC2. ILC2 activation observed in our study was consistent with recent data showing an expansion of IL-13 and IL-5 producing ILC2 in animals infected with Tritrichomonas spp, which was critical to restrain protozoan loads in mucosal sites (Howitt et al., 2016). We were unable to verify the degree of homology between the Tritrichomonas species described here and the one described in this recently published study, and thus it remains unclear at this stage how related these two protozoan species are. However, despite the consistent activation of ILC2 in our study, Th2 effector cells were not expanded or activated in T.mu-colonized mice, whereas Th1 and Th17 responses dominated the colonic immune landscape of our mice. Strikingly, we found that T.mu colonization conferred significant protection from Salmonella infection-driven enteritis in a manner that was fully dependent on the IL-18 inflammasome. Altogether these results uncover a formidable symbiotic host-protozoan interaction that promotes the release of anti-parasitic cytokines to control protozoan loads (Howitt et al., 2016) while inducing a T effector response to strengthen mucosal defenses against collateral infections such as Salmonella. However while serving a mutualistic relationship, T.mu colonization also increased murine susceptibility to inflammatory disease and cancer, consistent with prior literature showing that the IL-18/IL-18 receptor axis increases susceptibility to IBD and cancer in patients (Zitvogel et al., 2012). Protists are common members of the constitutive microbiota in humans. However, while most attention has been focused on their pathogenic role in mucosal disease, little is known about their symbiotic relationships with their hosts. Our study identified T.mu’s closest human ortholog as D. fragilis, a unicellular parasite capable of inducing gastro-intestinal symptoms in infected human individuals (Barratt et al., 2011). Controversies are ongoing whether D. fragilis, and for that matter, other protist species such as Enteromonas spp., Entamoeba dispar, Entamoeba coli and the related parabasalid Pentatrichomonas hominis are commensals, pathobionts, or pathogens of the human intestinal tract (Lukes et al., 2015). Certainly D. fragilis has been identified as an etiological agent of irritable bowel syndrome (IBS) (Stark et al., 2007), but it has also been recognized for its potential protective effect against disease flares in patients with inflammatory bowel disease (IBD) (Petersen et al., 2013). Our findings in the murine system demonstrate a host-protozoan interaction that increases mucosal host defenses at the cost of increased risk of inflammatory disease and this may gleen new perspective on the etiology of inflammation and immune homeostasis in individuals colonized with D. fragilis The strong mutualistic relationship between the protist T.mu and its rodent hosts highlight the need to expand our understanding of host-protozoan relationships in humans and explore the potential key contributions of commensal protozoa in shaping animal and human mucosal defenses. Interestingly, DNA sequencing of T.mu and D. fragilis in mouse and human reveal a strong genetic diversification of the protist between infected carriers. Phylogenetic classification may thus uncover strong geographical heterogeneity on the species level and could furthermore inform about functional and pathogenic differences between subspecies. STAR METHODS CONTACT FOR REAGENT AND RESOURCE SHARING Further information and requests for reagents may be directed to, and will be fulfilled by the corresponding author Miriam Merad (miriam.merad@mssm.edu) EXPERIMENTAL MODEL AND SUBJECT DETAILS Animals C57BL/6, B6.129P2(SJL)-Myd88tm1.1Defr/J (Myd88flox/flox), B6.129P2-Lyz2tm1(cre)Ifo/J (LysMCre), B6.Cg-Tg(Vil-cre) (VillinCre), B6.129S4-Ccr2tm1Ifc/J (Ccr2−/−), B6.129P2(C)-Ccr7tm1Rfor/J (Ccr7−/−), C57BL/6J-ApcMin/J (Apc+/Min), B6.129S1-Irf4tm1Rdf/J (Irf4flox/flox), B6(Cg)-Irf8tm1.1Hm/J (Irf8flox/flox), B6.129S2-Ifnar1tm1Agt/Mmjax (Ifnar−/−), 129S-Batf3tm1Kmm/J (Batf3−/−), NOD.Cg-Tg(Itgax-cre)1-1Reiz/PesaJ (CD11cCre), B6.129S1-Tlr5tm1Flv/J (Tlr5−/−), B6.129S7-Il1r1-tm1lmx/J (Il1r1−/−) and B6.129P2-Il18r1-tm1Aki/J (Il18−/−) mice were purchased from Jackson Laboratory. B6.129S6-Rag2tm1FwaN12 (Rag2−/−) mice were purchased from Taconic. Asc−/− mice were obtained from Millenium Inc. GF B6 and Rag2−/− mice were maintained in isolators at the Icahn School of Medicine at Mount Sinai Animal Facility. Unless otherwise stated, mice were used at 6–12 weeks of age. Experiments were carried out using age and gender matched groups. All animal procedures performed in this study were approved by the Institutional Animal Care and Use Committee (IACUC) of Mount Sinai School of Medicine. Animals housed within the vivarium were maintained at specific pathogen free (SPF) health status in individually ventilated cage units that are changed on a weekly basis; cage bottoms are covered with autoclaved bedding and nesting material for enrichment. Mice were fed irradiated food and drink reverse osmosis water from autoclaved bottles or automatic watering system. All animals were manipulated under HEPA filtered air exchange stations or biosafety cabinets and users were required to use chlorine-based disinfectant (MB-10) on gloves and work surfaces within manipulations with animals. All animal users, non-animal users, and husbandry staff were required to wear personnel protection equipment (PPE) to enter the facility i.e. disposable gowns, gloves, head caps, and shoe covers. Human Specimen Assessing prevalence of T.mu orthologs in humans Two human cohort populations were available for diagnostic screening. The first was comprised of fecal samples preserved in 100% ethanol collected from 188 otherwise healthy adult individuals from 9 regions in equatorial Colombia as part of a Colombian NIH health survey. The age of the individuals tested ranged between 16–41years-old. All samples were first screened by microscopy for the presence of Giardia spp., Entamoeba spp., Blastocystis hominis and Cryptosporidium spp. The second comprised 96 globally distributed fecal samples preserved in 100% ethanol from patients diagnosed Giardia positive by microscopy 16 samples each were collected from China, Colombia, Denmark, Ecuador, Egypt, and South Africa. All fecal samples were collected from nine locations across Colombia from asymptomatic adults. The adults provided written informed consent and ethical clearance for this study followed the ethics of Helsinki declaration. The study protocol was approved by the ethics committee from Universidad INCCA de Colombia under the Number 237894 METHOD DETAILS Isolation of lamina propria and mesenteric lymph nodes leukocytes Lamina propria lymphocytes were isolated as described (Vonarbourg et al., 2010). Briefly, the large intestines were incubated in EDTA supplemented Hank’s balanced salt solution (HBSS) without Ca2+ and Mg2+(Gibco) for 15–20 min at 37°C with mild agitation. The epithelial cell layer was removed by vortexing. Remaining sheets of LP were digested in collagenase (Sigma) and DNaseI (Sigma). The cells were resuspended in 5ml of 40% Percoll (GE Healthcare) and overlaid onto 5ml of 80% Percoll in a 15ml tube. Lymphocytes were collected at the interphase of the Percoll gradient, washed once and resuspended in media. Mesenteric lymph nodes were removed and cleaned from remaining mesenteric fat and digested with collagenase (Sigma) in Hank’s balanced salt solution (HBSS) with Ca2+ and Mg2+(Gibco) and 10% FBS for 15–20 min at 37°C. Cells we re washed using FACS buffer and filtered through 70µm cell strainer. Antibodies CD3e (clone 145-2C11, eBioscience), CD4 (clone GK1.5, eBioscience), CD11b (clone M1/70, eBioscience), CD11c (clone N418, eBioscience), CD45 (clone 30F11, eBioscience ), CD45.2 (clone 104, eBioscience) CD45RB (clone C363.16A, eBioscience), CXCR3 (clone CXCR3-173, Biolegend), CD44 (clone IM7, eBioscience), CD62L (clone MEL-14, Biolegend), Ly6C (clone AL-21, BD), B220 (clone RA3-6B2, eBioscience), ICOS (clone C398.4A , Biolegend), KLRG1 (clone 2F1/KLRG1, Biolegend), RORγt (clone B2D, eBioscience) IgA (clone S107, SouthernBiotech), CD64 (clone X54-5/7.1, Biolegend), CD103 (clone 2E7, eBioscience), F4/80 (clone BM8, Biolegend), I-A/I-E (clone M5/114.15.2, eBioscience), IFNγ (clone XMG1.2, eBioscience), IL-17A (clone eBio17B7, eBioscience), IL-22 (clone IL22JOP, eBioscience) Ki67 (clone SolA15, eBioscience) GM-CSF (clone MP1-22E9, eBioscience), IL-5 (clone TRFK5, eBioscience), IL-4 (clone 11B11, eBioscience), IL-13 (clone eBio1316H, eBioscience). Secondary antibodies, isotype controls and fluorophore-conjugated streptavidins were purchased from Biolegend and eBioscience. Time of flight Mass cytometry (CyTOF) analysis C57Bl/6 mice were gavaged with PBS or 2×106 highly purified T. mu prior to analysis. 14 days after inoculation, colonic LP cells were isolated from 2 PBS-treated and 2 T. musculis inoculated mice. Cells from individual samples were bar-coded with anti-CD45 antibodies, pooled and stained with the following antibodies: Ly6G (141 Pr), CD11c (142 Nd), TCRb (143 Nd), CD8 (146 Nd), CD11b (148 Nd), CD19 (149 Sm), CD24 (150 Nd), CD25 (151 Eu), Siglec-F (152 Sm), NKp46 (153 Eu), CD64 (156 Gd), Foxp3 (158 Gd), CD62L (160 Gd), Tbet (161 Dy), Ly6c (162 Dy), RORgt (163 Dy), CD103 (164 Dy), CD117 (166 Er), Gata3 (167 Er), ST2 (168 Er), NK1.1 (170 Er), CD44 (171 Yb), CD4 (172 Yb), MHCII (174 Yb), CD127 (175 Lu), B220 (176 Yb), Cisplatin (195Pt) All mass cytometry reagents were purchased from Fluidigm Inc. (former DVS) unless otherwise noted. LN and colons were dissociated into single-cell suspensions, washed with PBS containing 0.1% BSA and blocked with a commercial Fc-blocking reagent (BD Bioscience) to minimize non-specific antibody binding. The cells were then stained with a panel of metal-labeled antibodies against 26 cell surface markers for 30 minutes on ice, and then washed. All antibodies were either purchased pre-conjugated to metal tags, or conjugated in-house using MaxPar X8 conjugation kits according to the manufacturer’s instructions. After antibody staining, the cells were incubated with cisplatin for 5 minutes at RT as a viability dye for dead cell exclusion. The cells were then fixed and permeabilized with a commercial fix/perm buffer (BD Biosciences) and stained with antibodies against intracellular epitopes and stored in PBS containing 1.6% formaldehyde and a 1:4000 dilution of Ir nucleic acid intercalator to label all nucleated cells. Immediately prior to acquisition, the cells were washed in PBS, then in diH20 and resuspended in diH20 containing a 1/10 dilution of EQ 4 Element Calibration beads. After routine instrument tuning and optimization, the samples were acquired on a CyTOF2 Mass Cytometer in sequential 10 minute acquisitions at an acquisition rate of <500 events /s. The resulting FCS files were concatenated and normalized using a bead-based normalization algorithm in the CyTOF acquisition software and uploaded to Cytobank for analysis. FCS files were manually pre-gated on Ir193 DNA+ CD45+ events, excluding dead cells, doublets and DNA-negative debris, and the gated populations were then analyzed using viSNE (Amir el et al., 2013) based on all myeloid phenotypic markers. Putative cell populations on the resulting viSNE map were manually gated based on the expression of canonical markers, while allowing for visualization of additional heterogeneity within and outside of the labeled population bubbles. Flow Cytometry analysis Isolated cells were surface stained in FACS buffer (PBS w/o Ca2+ Mg2+ supplemented with 2% heat inactivated FBS and 5mM EDTA) for 20–30 min on ice. Staining for CXCR3 was performed for 30min at room temperature. Multiparameter analysis was performed on a LSR II Fortessa (BD) and analyzed with FlowJo software (Tree Star). Dead cells and doublets were excluded from all analysis. For the detection of cytokines, cells were cultured in media in the presence of Brefeldin A (10µg/ml) for 4h at 37°C. Ex vivo stimulations were carried out in the presence of Brefeldin A, phorbol 12-myristate 13-acetate (PMA) (Sigma) and Ionomycin (Sigma). Cells were fixed post stimulation using Cytofix/Cytoperm buffer (BD). Intracellular staining with anti-IFNγ, anti-IL-17A and anti-IL-22 was performed in FACS buffer containing 0.5% Saponin (Sigma). For the detection of transcription factors, cells were fixed and stained using the Foxp3-staining kit (eBioscience) according to the manufacturer’s instructions. Detection of bacterial bound IgA The cecum of mice was isolated, cut open and luminal content was washed out into 25ml of fresh cold PBS. Content was vortexed and filtered 2–3 times through 70µm cell strainer. Flow through was spun for 10min at 4000rpm at 4°C. Colle cted content was resuspended in 5–10ml of PBS and 500µl were stained with anti-IgA antibodies and DAPI. Cecal content from Rag2−/− mice served as a negative control for IgA binding. Scanning Electron Microscopy T.mu containing colonic and cecal tissue was fixed in 3% Glutaraldehyde/0.2M Sodium Cacodylate Buffer at pH7.3 for more than one hour. Fixative was washed away using 0.2M Sodium Cacodylate Buffer pH 7.3 for 10 minutes. Tissues were fixed in 1% Osmium Tetraoxide/0.2M Cacodylate Buffer pH 7.3 for one hour. After washing, tissues were dehydrated using increasing concentrations of Ethyl Alcohol (1× 10min 50%, 1× 10min 70%, 1× 10min 95%, 3× 10min 100%). After dehydration, tissues were critical point dried, mounted and sputter coated prior to imaging. Scanning electron pictures were taken on a Hitachi Field Emission S4300 scanning electron microscope. Giemsa staining T.mu were triple sorted into PBS and spun onto glass slides. Cells were semi-dried and fixed in cold methanol. Cells were stained with Giemsa (Sigma) according to the manufacturer’s instructions. Purification of Tritrichomonas musculis (T.mu) from cecal content The cecal content of T. mu containing mice was harvested into 30ml of sterile PBS and filtered several times through a 70µm cell strainer. The suspension was spun at 1000rpm for 5min at 4°C. The supernatant was discarded and the pellet was washed twice with 30ml PBS. The pellet was resuspended in sterile PBS after the final wash and cells were triple sorted into PBS on a FACS ARIA II using the 70µm nozzle at 70psi at 4°C. Purity was >99%. One to two million collected T. mu were gavaged into mice immediately after the sort. Culture of Tritrichomonas musculis (T.mu) Culture of T.musculis was performed as described in Saeki et al.(Saeki et al., 1983). Briefly, cecums of T.mu infected mice were flushed with PBS, passed through the 70 um cell strainers and spun down at 1000 rpm for 5 min. In order to enrich for T.mu containing fraction the 40/80 % percoll gradient centrifugation step was performed at 1000g for 15 minutes with breaks off and the T.mu were collected at the interphase. The interphase containing T.mu was spun down again at 1750 rpm for 10 min and resuspended in T.mu culture medium. T.mu culture medium was prepared as described by Saeki et al with the following modifications. Briefly cecums of mice were harvested and homogenized in PBS with 25 volumes of PBS per corresponding cecum weights. In order to get homogeneous suspension the cecum extract was stirred at 4 °C for 6 hrs and then spun down at 3500rpm for 10 min and the supernatant was filtered. The filtered cecal extract was used to resuspend the TrichoselTM broth (Becton Dickinson) and titrated to pH 7 with NaOH. The medium was then autoclaved prior to be supplemented with 10% heat-inactivated horse serum and with Abx against Gram positive and negative bacteria including vancomycin, gentamicin, streptomycin, penicillin and amphotericin B. 2 × 106 T.mu were inoculated per ml of growth media and cultured under anaerobic conditions for 3 days. After 3 days of cultures, the T.mu containing media was centrifuged at 1750 rpm for 10 min and 2 × 106 T.mu was inoculated per mouse as described above. Quantification of T.mu colonization efficiency Cecum of mice colonized with T.mu was excised weighted and cut longitudinally. Cecal content was harvested and resuspended in 10 ml of sterile PBS. Trophozoites were counted using hemocytometer and the numbers of T.mu trophozoites per gram of cecum were measured. Pathological assessment of intestinal tissue Cecums and colons were fixed, embedded in paraffin and sections were cut and stained with hematoxylin and eosin (H&E) and Periodic acid-Schiff (PAS) according to standard procedures. Pathological evaluation was performed by a pathologist assessing epithelial thickness, and epithelial integrity, submucosal edema, Goblet cell numbers, presence of polymorphonuclear granulocytes and a pathological score was assigned according to Barthel et al 2003. Quantitative real-time PCR (Q-PCR) Conventional reverse transcription, using the Sprint PowerScript reverse transcriptase (Clontech) was performed in accordance with the manufacturers’ instructions. Q-PCR was performed with SYBR GREEN on an ABI PRISM 7900HT Sequence Detection System (Applied Biosystems). The PCR protocol consisted of one cycle at 95°C (10 min) followed by 40 cycles of 95°C (15 s) and 58°C (1 min). Expressi on of hypoxanthine-guanine phosphoribosyltransferase (Hprt) was used as a standard. The average threshold cycle number (CRtR) for each tested mRNA was used to quantify the relative expression of each gene: 2^-[Ct(gene)-Ct(Hprt)]. Primers are listed below: Hprt (Fwd)_CCTGGTTCATCATCGCTAATC Hprt (Rev)_TCCTCCTCAGACCGCTTTT Tnfa (Fwd) TCTTCTCATTCCTGCTTGTGG Tnfa (Rev) GGTCTGGGCCATAGAACTGA Nos2 (Fwd) GGGCTGTCACGGAGATCA Nos2 (Rev) CCATGATGGTCACATTCTGC Reg3b (Fwd) CAGACCTGGTTTGATGCAGA Reg3b (Rev) GAAGCCTCAGCGCTATTGAG Reg3g (Fwd) ATGGCTCCTATTGCTATGCC Reg3g (Rev) GATGTCCTGAGGGCCTCTT ELISA 300–500µl of mouse blood was collected and serum was isolated using Serum/Gel Z/1.1 serum separation tubes (SARSTEDT, Germany) according to the manufacturers protocol. Serum was stored at −80°C. Immunoglobulins were abs orbed on high-bound plates (NUNC) and anti-IgA ELISA was performed using the anti-mouse IgA-HRP antibody (Southern Biotech). 200–300µg of ascending colonic tissue were cultured in 500ul of complete RPMI at 37°C and 5% CO 2. Supernatant was collected 24h later, transferred into sterile tubes and stored at −80°C. Detection of IL-18 from overnight tissue explants was performed using the IL-18 ELISA kit (eBioscience) according to the manufacturers protocol. MALDI-TOF T.mu cells were quadrupel sorted into sterile PBS. The cells were pelleted for 3 minutes at 15,000g. PBS was removed and the pellet resuspended in 100µL 70% ethanol. Fixed Trichomonas were pelleted again for 3 minutes at 15,000 g. The 70% ethanol was removed and the pellets were resuspended in 10µL 70% formic acid for lysis followed by the addition of 10µL acetonitrile. The formic acid and acetonitrile mixture was spun for 3 minutes at 15,000g and 1µL of the supernatant was spotted into the MALDI-TOF target plate. After the lysate was dry, 1 µl of α-cyano-4-hydroxycinnamic acid matrix was added on top of each spot. The target was then analyzed with a MALDI TOF BioTyper (Bruker) according to manufacturer’s instructions. Repetitive recordings were collected and these spectra used to create a reference spectrum, which was used for future identification and quality control of the Trichomonas at other time points. RNAseq analysis of epithelial cells RNA sequencing was performed on RNA isolated from whole colonic tissues or epithelial cells. Groups of age and gender-matched mice were colonized with purified T.mu and RNA was isolated using Trizol. RNA was purified using phenol/chloroform extraction. Quantification was performed using the Qubit RNA assay kit HS according to the manufacturers protocol. RNA was sequenced at the Broad Technology Services using the High-throughput eukaryote 3’ DGE library. RNA count analysis and visualization was done in R (https://www.R-project.org/). RNA counts were converted to log2-counts per million using the voom package (Law et al., 2014). Differential expression was quantified using the lmFit function in the limma package (Ritchie et al., 2015). Gene sets were downloaded from the MSigDB mouse ortholog website (Subramanian et al., 2005) (http://bioinf.wehi.edu.au/software/MSigDB/) and enrichment was computed using a hypergeometric test. Data was visualized using ggplot2. The reads were quality filtered as follows. First, the bases at the ends of the reads with quality less than 10 were trimmed. If the quality of the leftover part at the 20-th percentile was less than 5 or the length of the leftover part was less than 20 the read was discarded. If one of the paired end reads failed the quality control, the mate was discarded as well. Reads were aligned using STAR (Dobin et al., 2013) to the mouse genome. Counts for reads mapping to each gene were evaluated using the Subread package (Liao et al., 2013). Phylogenetic and metagenomic sequencing of T.mu DNA was isolated from FACS-purified protozoa and subjected to ITS PCR-DNA sequence analysis using pan-parabasalid primers anchored in the 18S [AATACGTCCCCTGCCCTTTGT] and 28S [TCCTCCGCTTAATGAGATGC] rDNA gene array that amplify across the polymorphic ITS region. PCR amplification consisted of 35 cycles of 40 sec at 95° C, 40 sec at 58° C, and 40 sec at 72° C. Phylogenetic analysis was inferred using RAxML methodology after alignment to related parabasalid sequences downloaded from GenBank. The sequence deposited in GenBank has accession number (T.musculis_18S.sqn Tritrichomonas KX000921 and T.musculis_ITS.sqn Tritrichomonas KX000922).Branching support was assessed using ML bootstrap analysis (1000 replications). Major groups are bracketed and labeled. The sequence placed the rodent parabasalid, which we hereafter refer to as T. musculis (T.mu), as distinct and sister to a clade comprising the Tritrichomonadidae that shared 95% identity with T. muris. T.mu was FACS-sorted and pulled from five in-house mice naturally colonized with T.mu. DNA was extracted using the Qiagen stool extraction kit according to manufacturer specifications. A paired-end TrueSeq nano DNA sequencing library (Illumina, Cat. No. 15041110) with an average insert size of 550 bp was prepared using 200 ng of input T.mu genomic DNA and run on an Illumina Mi-Seq platform. 35.61 million reads passed filtering T.mu sequences for the following genes alpha-Tubulin, EF1-alpha, GAPDH, and MDH were pulled from the metagenomic data for phylogenetic discrimination of T.mu from other enteric parabasailds as follows: reference sequences from the related parabasalids T. foetus and T.vaginalis for the 4 marker genes were downloaded from NCBI. BWA (v0.7.5a) was used to map reads against these 4 sequences. GATK (v3.5) in conjunction with Picard (v1.131) were used to select only high quality reads (average read depth for single copy loci was 20X) to produce a consensus sequence. No heterozygous SNP positions were identified in the regions mapped. Sequences were deposited in GenBank, with the following Accession IDs: KX492911, 12, 13, 14. Salmonella Infection Groups of age and gender matched C57BL/6 mice were either left untreated or were orally gavaged with 2×106 purified T.mu. 14 days after treatment mice were orally gavaged with Streptomycin and infected with Salmonella typhimurium as previously described (Barthel et al., 2003), Infection and Immunity. Colony-forming units of S typhimurium in fecal pellets, the cecum, the spleen and the mesenteric lymph nodes were measured on MacConkey agar plates containing Streptomycin 48h after infection. Cecal weight was recorded and caeca were fixed, embedded in paraffin and sections were cut and stained with hematoxylin and eosin (H&E) according to standard procedures. Pathological evaluation was performed by a pathologist in a blinded manner and scored as previously described (Barthel et al., 2003). For neutralization of IL-18, groups of Trichomonas infected mice were injected 3 times with 200µg of neutralizing anti-IL-18 antibody (YIGIF74-1G7, Bioxcell) prior to infection with Salmonella. T cell transfer colitis Groups of age and gender matched SPF or GF Rag2−/− mice were either gavaged with PBS or 1–2×106 highly purified T.mu. Conventionalized GF Rag2−/− mice were colonized with SPF Rag2−/− flora through fecal transplantation 4 weeks prior to gavage with PBS or T. musculis. Two weeks after gavage, 0.75–1×106 highly purified CD4+ CD45RBhi T cells were injected i.p. into mice. To purify CD45RBhi T cells spleens and lymph nodes of 6–8 week old mice were isolated and organs were mashed through a 70µm cell strainer. Leukocytes were obtained through red cell lysis and stained with biotinylated anti-CD4 antibodies followed by incubation with magnetic anti-biotin beads. The enriched lymphocyte fraction was stained with DAPI, anti-TCRb, anti-CD4, anti-CD45RB and anti-CD25 antibodies. TCRb+ CD4+ CD45RBhi CD25−cells were sorted on a BD FACS ARIA II into complete RPMI media containing 10% (heat inactivated) FBS. 0.75–1×106 T cells were injected i.p. into Rag2−/− mice. Weight loss was monitored weekly and pathological evaluation of colons was performed as endpoint analysis on H&E stained sections of paraffin-embedded colons. Pathological scores were assessed by experimental pathologist in a blind fashion and calculated based on the previously established scoring method (Asseman et al., 1999). Colorectal cancer model Apc+/min Groups of 4-week-old Apc+/min mice were left untreated or were orally gavaged with 2×106 purified T.mu. 3–4 months later, mice were analyzed and tumor burden and tumor size were measured. Immunofluorescence analysis TUNEL staining was performed using the TUNEL staining kit (Roche) according to the manufacturer’s protocol. Ki67 staining was performed using anti-Ki67 antibody (clone: VPRMO4 Vector Lab) on paraffin-embedded sections. Antigen retrieval was performed in Demasking Solution at 800W for 15min (Vector Lab H-3300). QUANTIFICATION AND STATISTICAL ANALYSIS In each experiment, multiple mice were analyzed as biological replicates as indicated in all the figures. Graphs show mean ± SEM. Student’s t test or ONE-way ANOVA Bonferroni’s Multiple Comparison Test was used to determine significance. * p < 0.05, ** p < 0.01, *** p < 0.001. DATA AND SOFTWARE AVAILABILITY Data Resources The accession number for the sequencing data reported in this paper is GSE8611. The T.mu sequences have been deposited to the GenBank. The accession number for the T.mu sequences are as follows; T.musculis_18S.sqn Tritrichomonas KX000921, T.musculis_ITS.sqn Tritrichomonas KX000922, T.musculis a-tubulin, KX492911; T.musculis EF1a, KX492912; T.musculis GAPDH, KX492913; T.musculis, MDH, KX492914. Supplementary Material 1 2 3 4 5 6 7 08 We thank the Merad laboratory for helpful discussions and input. We would like to thank Fiona Desland for critical review of the manuscript. We would like to thank Jordi Ochando and the Flow Cytometry facility for their excellent technical support and assistance with cell sorting. We would like to thank Steve Porcella (Genomics Unit, RTB, Rocky Mountain Labs, NIAID), Mariam Quinones (Bioinformatics and Computational Biology Branch, NIAID) and Jacquice Davis (NIAID Microbiome Program) for their assistance with metagenomic and 16S sequencing and analysis. This study used the Nephele platform from the NIAID Office of Cyber Infrastructure and Computational Biology (OCICB) in Bethesda, MD. M.M. is funded by NIH grants R01 CA154947A, R01 CA173861 and U01 AI095611. This work was supported in part by the National Institute of Allergy and Infectious Diseases, National Institutes of Health (Y.B. and M.E.G.). M.E.G. Is a Scholar of the Canadian Institute for Advanced Research (CIFAR) Integrated Microbial Biodiversity Program Figure 1 Identification of a new protistic commensal in mice (A) Colonic LP cells were isolated from B6 mice obtained from commercial sources or bred at the Mount Sinai animal facility (in-house). Cells were stained with anti-CD45 antibodies and analyzed by flow cytometry. Colored contour plots show staining for CD45 on gated live LP cells. Numbers adjacent to gates represent percentages (+/− SEM). (B) Bar graph shows quantification of CD45.2+ cells. (C) Bar graph shows IgA concentrations in sera (mg/ml, left) or colonic tissue explants (mg/g, right). (D) Surface staining of IgA bound to intestinal microbes. Numbers adjacent to gates represent percentages. (E) Cytospins of sorted IgA+ cells in (D) were stained with Giemsa. Micrograph shows a representative staining of colonic lavage (40x magnification). Bar indicates 10µm. (F) Scanning electron microscopy (SEM) of colonic tissues from in-house mice. (G) Protozoa per gram of cecum were quantified in five in-house B6 animals naturally colonized with protozoa (B6 Nat) or five animals gavaged with 2 × 106 FACS sorted protozoa (B6 Gavage). Bar graph represents number of protozoa per gram of cecum. (H) DNA was isolated from FACS-purified protozoa and subjected to ITS PCR-DNA sequence analysis. Phylogenetic analysis was performed as described in Material and Methods. The sequence placed the rodent parabasalid, which we hereafter refer to as T. musculis (T.mu), as distinct and sister to a clade comprising the Tritrichomonadidae that shared 95% identity with T. muris. (I) Intestinal tissues were isolated from B6 mice 2 weeks after inoculation with purified T. mu, and paraffin-embedded sections were stained with H&E and PAS. (J) Pathological scoring of T. mu-inoculated cecums and colons was performed, assesing combined pathological score for colon tissue and colonic goblet cell hyperplasia (left) and cecum tissue and cecum goblet cell hyperplasia (right). Red dashed bar denotes borderline pathological score found in normal mice (score 0–2) (Barthel et al., 2003). Data shown represent 3 mice per group, whereof 4–5 high-power fields per tissue and animal were assessed. Statistical analysis was performed using unpaired Student’s t-test. Statistical significance is indicated by *P < 0.05, **P < 0.01, and ***P < 0.001Scale bar 100 mm (A–C) Data shown is representative of at least two independent experiments with at least three animals per experimental group. All data are shown as mean ± SEM. Student’s t-test was performed. Statistical significance is indicated by *P < 0.05, **P < 0.01, ***P < 0.001; ns, not significant. See also Supplementary Fig. 1 and Supplementary Table 1. Figure 2 Colonization with T.mu remodels the colonic mucosal immune tone (A) Groups of B6 mice were gavaged with 2 × 106 highly purified T.mu. Eight weeks later, colonic LP cells were stained with anti-CD45.2 antibodies and analyzed by flow cytometry. Colored contour plots show CD45+ cells. Numbers adjacent to gates represent percentages. (B) Bar graph shows absolute number of CD45.2+ cells in control (Ctrl) mice or mice inoculated with T.mu. (C) CyTOF analysis of colonic myeloid immune populations after colonization with purified T.mu. ViSNE dot plots show changes in immune cell composition after T.mu colonization. (D) Intracellular cytokine staining and flow cytometric analysis of TNFα-producing cells in the colonic tissue of T.mu colonized mice. Colored contour plots show staining for TNFα and MHCII within all CD45+ cells. Gates were drawn on TNFα+ MHCII+ cells and the expression of Ly6C and CD64 was analyzed. Numbers adjacent to gates represent percentages (E) Bar graphs show absolute numbers of all TNFα-producing cells and their distribution into the macrophage and monocyte population. (F–G) Mice were gavaged with purified T.mu and 14 days later, colonic LP CD4+ T cells were analyzed for the production of IL-17 and IFNγ (F) and for the production of IL-5 and IL-13 (G) by intracellular cytokine staining. Numbers adjacent to gates represent percentages. (H) Bar graphs show quantification of cytokine-producing cells. (I) Bar graphs show absolute numbers of IFNγ+, IL-17+ and IL-17+IFNγ+ (DP) Th cells. (J) IFNγ, IL-17 and IL-17+IFNγ+, production by CD4+ T cells isolated 7 days, 14 days, 28 days and 56 days after colonization with T.mu. Bar graphs show absolute numbers of Th cell-subsets. Data shown is representative of at least two independent experiments with at least 3 mice per group and experiment. All data are shown as mean ± SEM. Student’s t-test was performed. Statistical significance is indicated by *P < 0.05, **P < 0.01, ***P < 0.001; ns, not significant. See also Supplementary Figures 2 and 3. Figure 3 Distinct DC subsets control T.mu associated immune responses (A–G) Mice were gavaged with purified T.mu 14 days prior to analysis. Colonic LP cells were isolated from (A) Ccr7−/−, (B) Batf3−/−, (E) Irf8ΔDC, (H) Irf4ΔDC. Cells were incubated in the presence of brefeldin A, Ionomycin and PMA for 4h and stained with anti-CD45, CD3, CD4, IL-17 and IFNγ. IL-17 and IFNγ-producing CD4+ T cells were analyzed using flow cytometry. Bar graphs show quantification of absolute numbers of cytokine-producing colonic Th cells. Colonic LP cells isolated from indicated knock out mice were stained with anti-CD45, MHCII, CD11c, CD11b, CD103, CD64 and Ly6C. DC populations were analyzed by gating on CD45+ MHCII+ CD11c+ CD64− Ly6c− cells. (C, D, F and G) contour plots show the abundance of CD103+ DC, CD103+CD11b+ DC and CD103−CD11b+ DC within the colonic LP of WT, Batf3−/−, Irf8ΔDC, Irf4ΔDC mice. Data shown is representative of at least 3 independent experiments with at least 3 mice per experimental group for each genotype. All data are shown as mean ± SEM. Student’s t-test was performed. Statistical significance is indicated by *P < 0.05, **P < 0.01, ***P < 0.001; ns, not significant. See also Supplementary Figure 4. Figure 4 Inflammasome activation promotes T.mu associated immune responses (A) Tissue culture supernatants of tissue explants or RNA, isolated from total colonic tissue were analyzed. Supernatants were analyzed for the presence of IL-18 protein using ELISA. Quantitative RT-PCR was performed to measure IL-18 RNA on whole colonic tissue. Bar graphs show pg/ml of IL-18 or relative expression of Il18 normalized to Hprt. (B–E) Mice were gavaged with purified T.mu 14 days prior to analysis. Colonic LP cells were isolated from (B) Il18−/−, (C) Lethally irradiated bone marrow chimeric mice reconstituted with Il18+/+ or Il18−/−bone marrow cells, at least 8 weeks prior to inoculation with T.mu (D) Asc−/−, (E) Il1r1−/−, mice. Cells were incubated in the presence of brefeldin A, Ionomycin and PMA for 4h and then stained with anti-CD45, CD3, CD4, IL-17 and IFNγ. IFNγ and IL-17 production by CD4+ T cells was analyzed using flow cytometry. Bar graphs show quantification of colonic Th cells producing IFNγ only, IL-17 only or IL-17 and IFNγ (DP). Data shown is representative of at least 3 independent experiments with at least 3–4 mice per experimental group for each genotype. All data are shown as mean ± SEM. Student’s t-test was performed. Statistical significance is indicated by *P < 0.05, **P < 0.01, ***P < 0.001; ns, not significant. See also Supplementary Figures 4 and 6. Figure 5 T.mu protects against mucosal infection through inflammasome activation (A) Groups of mice were either gavaged with purified T.mu or left untreated 14 days prior to Salmonella typhimurium infection. Cecum weight was measured and pathological scoring was performed on H&E stained sections. Data shown is representative of 3 individual experiments with 5 mice per experiment and experimental condition. (B) S. typhimurium colony-forming units (CFU) were measured in the fecal pellet, the cecum, spleen and MLN 48h after infection. (C) Groups of mice were gavaged with purified T.mu 14 days prior to S. typhimurium infection and injected for three consecutive days with 200 ug of either isotype control IgG or anti-IL18 neutralizing antibody right before S. typhimurium infection. Cecum weight was measured and pathological scoring was performed on H&E stained sections. (D) CFU of S. typhimurium were measured in the MLN and spleen 48h after infection. Data shown is representative of at least 2–3 independent experiments with at least 5 mice per group and experiment. All data are shown as mean ± SEM. Student’s t-test was performed. Statistical significance is indicated by *P < 0.05, **P < 0.01, ***P < 0.001; ns, not significant. Scale bar 100 mm. Figure 6 T.mu exacerbates pathogenic inflammation (A) Groups of GF Rag2−/− mice were left untreated (black), conventionalized (pink) or inoculated with quadruple sorted T.mu (blue). Groups of conventional Rag2−/− mice were left untreated (grey) or inoculated with purified T.mu (purple). Conventionalization of mice was carried out 4 to 6 weeks prior to adoptive transfer of sorted splenic CD45RBhi CD4+ T cells. Gavage with T.mu was performed 2 weeks prior to T cell transfer. Weight loss was monitored weekly and end-point assessment of the disease score was performed using histology assessing degree of inflammation, epithelial erosion, ulceration and hyperplasia, mucin depletion and goblet cell numbers. Pie charts and histology show the clinical assessment of colitis. Data shown is representative of at least 3–4 individual experiments with at least 3–4 mice per experimental group. Scale bar 100 mm. (B and C) Epithelial cell death and epithelial proliferation in colonic tissues of naïve or T.mu inoculated mice. Representative staining of DNA (DAPI blue) and TUNEL (green) in naïve (top row) or T.mu colonized (bottom row) groups. (C) Ki67 staining in colonic tissue of naïve or T.mu inoculated mice. Representative staining of DNA (DAPI blue) and Ki67 (red) in naïve (top row) or T.mu colonized (bottom row). Scatter dot plots show quantification of TUNEL+ cells and Ki67+ cells. (D) Quantitative RT-PCR on T.mu-driven intestinal target genes (Nos2, Tnfa, Reg3g and Reg3b), normalized to Hprt. (E and F) Apc+/min mice were left untreated or inoculated with purified T.mu. (E) Tumor burden and size was measured 8–12 weeks later. Micrographs of formalin-fixed colonic tissue counterstained with methylene blue (0.5% PBS) show representative tumor burden in Apc+/min mice untreated or inoculated with T.mu. (F) Bar graphs show tumor counts and size. Data shown is representative of at least 3 individual experiments with 3–5 mice per experimental condition. All data are shown as mean ± SEM. Student’s t-test or One-way ANOVA Bonferroni’s multiple comparison T test was performed for (A) Statistical significance is indicated by *P < 0.05, **P < 0.01, and ***P < 0.001. See also Supplementary Figure 5. Methods and Resources KEY RESOURCES TABLE REAGENT or RESOURCE SOURCE IDENTIFIER Antibodies Anti-mouse Ly6G-141Pr Biolegend RRID:AB_1089179 Anti-mouse CD11c-142Nd Fluidigm/DVS Sciences Cat#:3142003C Anti-mouse TCRb-143Nd Fluidigm/DVS Sciences Cat#:3143010B Anti-mouse CD16/32-144Nd Fluidigm/DVS Sciences Cat#:3144009B Anti-mouse CD8a-146Nd Fluidigm/DVS Sciences Cat#:3146003B Anti-mouse CD45-147Sm Fluidigm/DVS Sciences Cat#:3147003B Anti-mouse CD11b-148Nd Fluidigm/DVS Sciences Cat#:3148003B Anti-mouse CD19-149Sm Fluidigm/DVS Sciences Cat#:3149002B Anti-mouse IgM-151Eu Fluidigm/DVS Sciences Cat#:3151006B Anti-mouse SiglecF-152Sm eBioscience RRID:AB_2572866 Anti-mouse NKP46-153Eu Fluidigm/DVS Sciences Cat#:3153006B Anti-mouse CD64-156Gd Biolegend RRID:AB_314486 Anti-mouse FoxP3-158Gd Fluidigm/DVS Sciences Cat#:3158003A Anti-mouse F4/80-159Tb Fluidigm/DVS Sciences Cat#:3159009B Anti-mouse CD62L-160Gd Fluidigm/DVS Sciences Cat#:3160008B Anti-human/mouse Tbet-161Dy Biolegend Cat#:644802 Anti-mouse Ly6C-162Dy Biolegend Cat#:128002 Anti-mouse RORgt-163Dy eBioscience RRID:AB_925759 Anti-mouse CD103-164Dy Biolegend RRID:AB_535944 Anti-mouse Gata3-167Er Fluidigm/DVS Sciences Cat#:3167007A Anti-mouse ST2-168Er Biolegend RRID:AB_2561843 Anti-mouse NK1.1-170Er Fluidigm/DVS Sciences Cat#:3170002B Anti-human/mouse CD44-171Yb Fluidigm/DVS Sciences Cat#:3171003B Anti-mouse CD4-172Yb Fluidigm/DVS Sciences Cat#:3145002B Anti-mouse I-A/I-E-174Yb Fluidigm/DVS Sciences Cat#:3174003B Anti-mouse CD127-175Lu Fluidigm/DVS Sciences Cat#:3175006B Anti-mouse CD45R/B220-176Yb Fluidigm/DVS Sciences Cat#:3160012B CD3e clone 145-2C11 eBioscience Cat#:45-0031-82 CD4 clone GK1.5 eBioscience Cat#:25-0041-82 CD11b clone M1/70 eBioscience Cat#:45-0112-82 CD11c clone N418 eBioscience Cat#:25-0114-82 CD45 clone 30F11 CD45.2 clone 104 eBioscience eBioscience Cat#:47-0451-82 Cat#:47-0454-82 CD45RB clone C363.16A eBioscience Cat#:17-0455-81 CXCR3 clone CXCR3-173 Biolegend Cat#:126514 CD44 clone IM7 eBioscience Cat#:25-0441-82 CD62L clone MEL-14 Biolegend Cat#:104406 Ly6C clone AL-21 BD Cat#:553104 B220 clone RA3-6B2 eBioscience Cat#:45-0452-82 ICOS clone C398.4A Biolegend Cat#:313514 KLRG1 clone 2F1/KLRG1 Biolegend Cat#:138410 RORγt clone B2D eBioscience Cat#:14-6981-82 IgA clone S107 SouthernBiotech Cat#:1040-02 CD64 clone X54-5/7.1 Biolegend Cat#:139306 CD103 clone 2E7 eBioscience Cat#:12-1031-83 F4/80 clone BM8 I-A/I-E clone M5/114.15.2 Biolegend eBioscience Cat#:123120 Cat#:48-5321-82 IFNγ clone XMG1.2 eBioscience Cat#:12-7311-82 IL-17A eBio17B7 eBioscience Cat#:17-7177-81 IL-22 clone IL22JOP eBioscience Cat#:17-7222-80 Ki67 clone SolA15 eBioscience Cat#:12-5698-82 GM-CSF clone MP1-22E9 eBioscience Cat#:12-7331-82 IL-5 clone TRFK5 eBioscience Cat#:BMS179 IL-4 clone 11B11 eBioscience Cat#:12-7041-81 IL-13 clone eBio1316H eBioscience Cat#:25-7133-82 CD3e clone 145-2C11 eBioscience Cat#:45-0031-82 CD4 clone GK1.5 eBioscience Cat#:25-0041-82 CD11b clone M1/70 eBioscience Cat#:45-0112-82 CD11c clone N418 eBioscience Cat#:25-0114-82 CD45 clone 30F11 eBioscience Cat#:47-0451-82 CD45.2 clone 104 CD45RB clone C363.16A eBioscience eBioscience Cat#:47-0454-82 Cat#:17-0455-81 CXCR3 clone CXCR3-173 Biolegend Cat#:126514 CD44 clone IM7 eBioscience Cat#:25-0441-82 CD62L clone MEL-14 Biolegend Cat#:104406 Ly6C clone AL-21 BD Cat#:553104 B220 clone RA3-6B2 eBioscience Cat#:45-0452-82 ICOS clone C398.4A Biolegend Cat#:313514 Chemicals, Peptides, and Recombinant Proteins Brefeldin A SIGMA Cat#:B7651 Phorbol 12-myristate 13-acetate SIGMA Cat#:79346 Ionomycin SIGMA Cat#:I0634 TRizol-reagent Ambion Cat#:15596026 Formic Acid SIGMA Cat#:94318-250ML-F Acetonitrile SIGMA Cat#:34967-250ML HCCA Bruker Cat#:255344 Critical Commercial Assays Foxp3 / Transcription Factor Fixation/Permeabilization Concentrate and Diluent eBioscience Cat#:00-5521-00 Foxp3 / Transcription Factor Staining Buffer Set eBioscience Cat#:00-5523-00 Fixation/Permeabilization Concentrate eBioscience Cat#:00-5123 Permeabilization Buffer eBioscience Cat#:00-8333 QIAquick 96 PCR Purification Kit Qiagen Cat#:28181 Deposited Data Raw and analyzed data This paper GEO: GSE63473 B-RAF RBD (apo) structure This paper PDB: 5J17 human reference genome NCBI build 37, GRCh37 Genome Reference Consortium http://www.ncbi.nlm.nih.gov/projects/genome/assembly/grc/human/ Experimental Models: Cell Lines n.a. Experimental Models: Organisms/Strains Salmonella enterica serovar typhimurium ATCC Stock#:14028 B6.129P2(SJL)-Myd88tm1.1Defr/J (Myd88flox/flox) Jackson Laboratory Stock#:006148 B6.129P2-Lyz2tm1(cre)Ifo/J (LysMCre) Jackson Laboratory Stock#:004781 B6.Cg-Tg(Vil-cre) (VillinCre) Jackson Laboratory Stock#:021504 B6.129S4-Ccr2tm1Ifc/J (Ccr2−/−) Jackson Laboratory Stock#:004999 B6.129P2(C)-Ccr7tm1Rfor/J (Ccr7−/−) Jackson Laboratory Stock#:006621 C57BL/6J–ApcMin/J (Apc+/Min) Jackson Laboratory Stock#:002020 B6.129S1-Irf4tm1Rdf/J (Irf4flox/flox) Jackson Laboratory Stock#:009380 B6(Cg)-Irf8tm1.1Hm/J (Irf8flox/flox) Jackson Laboratory Stock#:014175 B6.129S2-Ifnar1tm1Agt/Mmjax (Ifnar−/−) Jackson Laboratory Stock#:32045 129S–Batf3tm1Kmm/J (Batf3−/−) Jackson Laboratory Stock#:013755 NOD.Cg-Tg(Itgax-cre)1-1Reiz/PesaJ (CD11cCre) Jackson Laboratory Stock#:008068 B6.129S1-Tlr5tm1Flv/J (Tlr5−/−) Jackson Laboratory Stock#:008377 B6.129S7-Il1r1-tm1lmx/J (Il1r1−/−) Jackson Laboratory Stock#:003245 B6.129P2-Il18r1-tm1Aki/J (Il18−/−) Jackson Laboratory Stock#:004130 B6.129S6-Rag2tm1FwaN12 (Rag2−/−) Taconic Stock#:RAGN12 Asc−/− Millenium Inc. n.a. Recombinant DNA n.a. Sequence-Based Reagents Hprt (Fwd)_CCTGGTTCATCATCGCTA ATC Sigma n.a. Hprt (Rev)_TCCTCCTCAGACCGCTTTT Sigma n.a. Tnfa (Fwd) TCTTCTCATTCCTGCTTGTGG Sigma n.a. Tnfa (Rev) GGTCTGGGCCATAGAACTGA Sigma n.a. Nos2 (Fwd) GGGCTGTCACGGAGATCA Sigma n.a. Nos2 (Rev) CCATGATGGTCACATTCTGC Sigma n.a. Reg3b (Fwd) CAGACCTGGTTTGATGCAGA Sigma n.a. Reg3b (Rev) GAAGCCTCAGCGCTATTGAG Sigma n.a. Reg3g (Fwd) ATGGCTCCTATTGCTATGCC Sigma n.a. Reg3g (Rev) GATGTCCTGAGGGCCTCTT Sigma n.a. Software and Algorithms R R Development Core Team (2008). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. ISBN 3-900051-07- 0 http://www.R-project.org limma Ritchie, M.E., Phipson, B., Wu, D., Hu, Y., Law, C.W., Shi, W., and Smyth, G.K. (2015). limma powers differential expression analyses for RNA- sequencing and microarray studies. Nucleic Acids Research 43(7), e47. https://bioconductor.org/packages/release/bioc/html/limma.html ggplot2 H. Wickham. ggplot2: Elegant Graphics for Data Analysis. Springer-Verlag New York, 2009. http://ggplot2.org/ This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. Authors Contribution A.C., A.M., V.K. designed and analyzed the experiments and wrote the manuscript. A.K., J.D.R., A.R., R.R., I.M., R.N. performed experiments and analyzed data. E.D.A. and A.S. analyzed the RNAseq data. S.G., B.G., J.C. and J.F. provided intellectual expertise. Y.B. designed and funded experiments. M.E.G. designed and funded experiments and contributed to writing the manuscript. M.M designed and funded the study and wrote the manuscript . REFERENCES Amir el AD Davis KL Tadmor MD Simonds EF Levine JH Bendall SC Shenfeld DK Krishnaswamy S Nolan GP Pe’er D viSNE enables visualization of high dimensional single-cell data and reveals phenotypic heterogeneity of leukemia Nat Biotechnol 2013 31 545 552 23685480 Asseman C Mauze S Leach MW Coffman RL Powrie F An essential role for interleukin 10 in the function of regulatory T cells that inhibit intestinal inflammation J Exp Med 1999 190 995 1004 10510089 Barratt JL Harkness J Marriott D Ellis JT Stark D The ambiguous life of Dientamoeba fragilis: the need to investigate current hypotheses on transmission Parasitology 2011 138 557 572 21349214 Barthel M Hapfelmeier S Quintanilla-Martinez L Kremer M Rohde M Hogardt M Pfeffer K Russmann H Hardt WD Pretreatment of mice with streptomycin provides a Salmonella enterica serovar Typhimurium colitis model that allows analysis of both pathogen and host Infect Immun 2003 71 2839 2858 12704158 Bogunovic M Ginhoux F Helft J Shang L Hashimoto D Greter M Liu K Jakubzick C Ingersoll MA Leboeuf M Origin of the lamina propria dendritic cell network Immunity 2009 31 513 525 19733489 Bogunovic M Mortha A Muller PA Merad M Mononuclear phagocyte diversity in the intestine Immunol Res 2012 Boraschi D Dinarello CA IL-18 in autoimmunity: review Eur Cytokine Netw 2006 17 224 252 17353157 Diamond LS Axenic Cultivation of Trichomonas tenax, the Oral Flagellate of Man Journal of Eukaryotic Microbiology 1962 9 442 444 Dinarello CA Novick D Kim S Kaplanski G Interleukin-18 and IL-18 binding protein Front Immunol 2013 4 289 24115947 Dobin A Davis CA Schlesinger F Drenkow J Zaleski C Jha S Batut P Chaisson M Gingeras TR STAR: ultrafast universal RNA-seq aligner Bioinformatics 2013 29 15 21 23104886 Edelson BT Kc W Juang R Kohyama M Benoit LA Klekotka PA Moon C Albring JC Ise W Michael DG Peripheral CD103+ dendritic cells form a unified subset developmentally related to CD8alpha+ conventional dendritic cells J Exp Med 2010 207 823 836 20351058 Feng T Wang L Schoeb TR Elson CO Cong Y Microbiota innate stimulation is a prerequisite for T cell spontaneous proliferation and induction of experimental colitis J Exp Med 2010 207 1321 1332 20498021 Forster R Schubel A Breitfeld D Kremmer E Renner-Muller I Wolf E Lipp M CCR7 coordinates the primary immune response by establishing functional microenvironments in secondary lymphoid organs Cell 1999 99 23 33 10520991 Gao Y Nish SA Jiang R Hou L Licona-Limon P Weinstein JS Zhao H Medzhitov R Control of T helper 2 responses by transcription factor IRF4-dependent dendritic cells Immunity 2013 39 722 732 24076050 Grivennikov SI Wang K Mucida D Stewart CA Schnabl B Jauch D Taniguchi K Yu GY Osterreicher CH Hung KE Adenoma-linked barrier defects and microbial products drive IL-23/IL-17-mediated tumour growth Nature 2012 491 254 258 23034650 Gu Y Kuida K Tsutsui H Ku G Hsiao K Fleming MA Hayashi N Higashino K Okamura H Nakanishi K Activation of interferon-gamma inducing factor mediated by interleukin-1beta converting enzyme Science 1997 275 206 209 8999548 Helft J Ginhoux F Bogunovic M Merad M Origin and functional heterogeneity of non-lymphoid tissue dendritic cells in mice Immunol Rev 2010 234 55 75 20193012 Heyer J Yang K Lipkin M Edelmann W Kucherlapati R Mouse models for colorectal cancer Oncogene 1999 18 5325 5333 10498885 Hongoh Y Sharma VK Prakash T Noda S Taylor TD Kudo T Sakaki Y Toyoda A Hattori M Ohkuma M Complete genome of the uncultured Termite Group 1 bacteria in a single host protist cell Proc Natl Acad Sci U S A 2008 105 5555 5560 18391199 Howitt MR Lavoie S Michaud M Blum AM Tran SV Weinstock JV Gallini CA Redding K Margolskee RF Osborne LC Tuft cells, taste-chemosensory cells, orchestrate parasite type 2 immunity in the gut Science 2016 351 1329 1333 26847546 Iwasaki A Mucosal dendritic cells Annu Rev Immunol 2007 25 381 418 17378762 Jang MH Sougawa N Tanaka T Hirata T Hiroi T Tohya K Guo Z Umemoto E Ebisuno Y Yang BG CCR7 is critically important for migration of dendritic cells in intestinal lamina propria to mesenteric lymph nodes J Immunol 2006 176 803 810 16393963 Kirchberger S Royston DJ Boulard O Thornton E Franchini F Szabady RL Harrison O Powrie F Innate lymphoid cells sustain colon cancer through production of interleukin-22 in a mouse model J Exp Med 2013 210 917 931 23589566 Kotloff KL Nataro JP Blackwelder WC Nasrin D Farag TH Panchalingam S Wu Y Sow SO Sur D Breiman RF Burden and aetiology of diarrhoeal disease in infants and young children in developing countries (the Global Enteric Multicenter Study, GEMS): a prospective, case-control study Lancet 2013 382 209 222 23680352 Law CW Chen Y Shi W Smyth GK voom: Precision weights unlock linear model analysis tools for RNA-seq read counts Genome Biol 2014 15 R29 24485249 Liao Y Smyth GK Shi W The Subread aligner: fast, accurate and scalable read mapping by seed-and-vote Nucleic Acids Res 2013 41 e108 23558742 Lukes J Stensvold CR Jirku-Pomajbikova K Wegener Parfrey L Are Human Intestinal Eukaryotes Beneficial or Commensals? PLoS Pathog 2015 11 e1005039 26270819 Molloy MJ Grainger JR Bouladoux N Hand TW Koo LY Naik S Quinones M Dzutsev AK Gao JL Trinchieri G Intraluminal containment of commensal outgrowth in the gut during infection-induced dysbiosis Cell Host Microbe 2013 14 318 328 24034617 Moonah SN Jiang NM Petri WA Jr Host immune response to intestinal amebiasis PLoS Pathog 2013 9 e1003489 23990778 Murphy KM Transcriptional control of dendritic cell development Adv Immunol 2013 120 239 267 24070387 Noda S Hongoh Y Sato T Ohkuma M Complex coevolutionary history of symbiotic Bacteroidales bacteria of various protists in the gut of termites BMC Evol Biol 2009 9 158 19586555 Persson EK Uronen-Hansson H Semmrich M Rivollier A Hagerbrand K Marsal J Gudjonsson S Hakansson U Reizis B Kotarsky K Agace WW IRF4 transcription-factor-dependent CD103(+)CD11b(+) dendritic cells drive mucosal T helper 17 cell differentiation Immunity 2013 38 958 969 23664832 Petersen AM Stensvold CR Mirsepasi H Engberg J Friis-Moller A Porsbo LJ Hammerum AM Nordgaard-Lassen I Nielsen HV Krogfelt KA Active ulcerative colitis associated with low prevalence of Blastocystis and Dientamoeba fragilis infection Scand J Gastroenterol 2013 48 638 639 23528075 Popivanova BK Kitamura K Wu Y Kondo T Kagaya T Kaneko S Oshima M Fujii C Mukaida N Blocking TNF-alpha in mice reduces colorectal carcinogenesis associated with chronic colitis J Clin Invest 2008 118 560 570 18219394 Read S Powrie F Induction of inflammatory bowel disease in immunodeficient mice by depletion of regulatory T cells Curr Protoc Immunol 2001 Chapter 15 Unit 15 13 Ritchie ME Phipson B Wu D Hu Y Law CW Shi W Smyth GK limma powers differential expression analyses for RNA-sequencing and microarray studies Nucleic Acids Res 2015 43 e47 25605792 Saeki H Togo M Imai S Ishii T A new method for the serial cultivation of Tritrichomonas muris Nihon Juigaku Zasshi 1983 45 151 156 6632453 Schlitzer A McGovern N Teo P Zelante T Atarashi K Low D Ho AW See P Shin A Wasan PS IRF4 transcription factor-dependent CD11b+ dendritic cells in human and mouse control mucosal IL-17 cytokine responses Immunity 2013 38 970 983 23706669 Schroder K Tschopp J The inflammasomes Cell 2010 140 821 832 20303873 Serbina NV Pamer EG Monocyte emigration from bone marrow during bacterial infection requires signals mediated by chemokine receptor CCR2 Nat Immunol 2006 7 311 317 16462739 Shaked H Hofseth LJ Chumanevich A Chumanevich AA Wang J Wang Y Taniguchi K Guma M Shenouda S Clevers H Chronic epithelial NF-kappaB activation accelerates APC loss and intestinal tumor initiation through iNOS up-regulation Proc Natl Acad Sci U S A 2012 109 14007 14012 22893683 Sims JE Smith DE The IL-1 family: regulators of immunity Nat Rev Immunol 2010 10 89 102 20081871 Spits H Artis D Colonna M Diefenbach A Di Santo JP Eberl G Koyasu S Locksley RM McKenzie AN Mebius RE Innate lymphoid cells - a proposal for uniform nomenclature Nat Rev Immunol 2013 13 145 149 23348417 Stark D van Hal S Marriott D Ellis J Harkness J Irritable bowel syndrome: a review on the role of intestinal protozoa and the importance of their detection and diagnosis Int J Parasitol 2007 37 11 20 17070814 Stentiford GD Becnel JJ Weiss LM Keeling PJ Didier ES Williams BA Bjornson S Kent ML Freeman MA Brown MJ Microsporidia-Emergent Pathogens in the Global Food Chain Trends in Parasitology 2016 32 336 348 April 2, 2016 26796229 Subramanian A Tamayo P Mootha VK Mukherjee S Ebert BL Gillette MA Paulovich A Pomeroy SL Golub TR Lander ES Mesirov JP Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles Proc Natl Acad Sci U S A 2005 102 15545 15550 16199517 Treuting PM Clifford CB Sellers RS Brayton CF Of mice and microflora: considerations for genetically engineered mice Vet Pathol 2012 49 44 63 22173977 Trinchieri G Interleukin-12 and the regulation of innate resistance and adaptive immunity Nat Rev Immunol 2003 3 133 146 12563297 Tussiwand R Everts B Grajales-Reyes GE Kretzer NM Iwata A Bagaitkar J Wu X Wong R Anderson DA Murphy TL Klf4 expression in conventional dendritic cells is required for T helper 2 cell responses Immunity 2015 42 916 928 25992862 Varol C Vallon-Eberhard A Elinav E Aychek T Shapira Y Luche H Fehling HJ Hardt WD Shakhar G Jung S Intestinal lamina propria dendritic cell subsets have different origin and functions Immunity 2009 31 502 512 19733097 Vonarbourg C Mortha A Bui VL Hernandez PP Kiss EA Hoyler T Flach M Bengsch B Thimme R Holscher C Regulated expression of nuclear receptor RORgammat confers distinct functional fates to NK cell receptor-expressing RORgammat(+) innate lymphocytes Immunity 2010 33 736 751 21093318 Zitvogel L Kepp O Galluzzi L Kroemer G Inflammasomes in carcinogenesis and anticancer immune responses Nat Immunol 2012 13 343 351 22430787

PMC005xxxxxx/PMC5129872.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 8300331 2806 Cardiol Clin Cardiol Clin Cardiology clinics 0733-8651 1558-2264 27886791 5129872 10.1016/j.ccl.2016.09.001 NIHMS825291 Article Environmental Exposures and Cardiovascular Disease: A Challenge for Health and Development in Low- and Middle-Income Countries Burroughs Peña Melissa S. MD, MS 1 Rollins Allman MD 2 1 Division of Cardiology, Department of Medicine, University of California, San Francisco, CA 2 Department of Medicine, University of California, San Francisco, CA Corresponding Author: Melissa S. Burroughs Peña, MD, MS, 505 Parnassus Avenue, 11th floor, Room 1180D, San Francisco, CA 94143, Phone: 415-353-4934, Melissa.Burroughspena@ucsf.edu 3 11 2016 2 2017 01 2 2018 35 1 7186 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. Summary Environmental exposures in low- and middle-income countries lie at the intersection of increased economic development and the rising public health burden of cardiovascular disease. Increasing evidence suggests an association of exposure to ambient air pollution, household air pollution from biomass fuel, lead, arsenic, and cadmium with multiple cardiovascular disease outcomes including hypertension, coronary heart disease, stroke, and cardiovascular mortality. While populations in low- and middle-income countries are disproportionately exposed to environmental pollution, the bulk of evidence that links these exposures to cardiovascular disease is derived from populations in high-income countries. More research is needed to further characterize the extent of environmental exposures and develop targeted interventions towards reducing cardiovascular disease in at-risk populations in low- and middle-income countries. environmental health air pollution household air pollution heavy metals lead arsenic cadmium cardiovascular disease global health Introduction In the wake of large-scale economic development in low- and middle-income countries (LMIC), environmental pollution has been a challenge that has spurred tension within countries and across regions1. The use of fossil fuel combustion to increase access to electricity and transportation for millions of people has simultaneously modernized a multitude of rural and urban communities while locally polluting the air and globally increasing air temperatures2–4. Extractive industries, such as mining, have fueled the economies of many middle-income countries, lifting large swaths of the population out of poverty while contaminating water with heavy metals5. The conflict over environmental pollution is so intense in some regions that large-scale demonstrations and even violence have erupted, thus threatening national and regional security5. While many have argued that poverty reduction and economic growth justify the subsequent damage to the environment, the health consequences of environmental pollution, particularly for the populations residing in LMIC must also be taken into account6,7. Exposure to environmental pollution is associated with multiple adverse health outcomes in children and adults. While environmental pollution often evokes concerns for neurological development, cancer and pulmonary disease, cardiovascular disease must be considered as well8. Cardiovascular disease is the top cause of mortality worldwide, and has been identified as a target for large-scale, multisectoral intervention at the population level9,10. Taking into account the necessary integration of public and private sector activities to reduce the population burden of cardiovascular disease, the substantial impact of environmental exposures on the burden of cardiovascular disease at the population-level must be acknowledged and addressed11–13. Understanding the impact of environmental exposures on cardiovascular disease has the potential to yield greater insight into the full human cost of economic development14. This review will discuss the extent of the exposure, mechanisms of disease pathogenesis and the impact on cardiovascular disease for the following 5 environmental exposures: air pollution, household air pollution, lead, arsenic, and cadmium (Figure 1). While the selected environmental exposures described in this review do not represent an exhaustive list of every exposure with an observed association with cardiovascular disease, these pollutants represent the most widely studied exposures. While the focus of this review is to discuss the impact of these exposures on cardiovascular disease in LMIC, data from studies of high-income countries will be incorporated as needed to better illustrate the full impact of these exposures on cardiovascular disease risk factors and outcomes (Table 1). Ambient Air Pollution Fossil fuels power economic development in LIMC, fueling the expansion of industry, homes, and transportation. However, fossil fuel combustion releases a heterogeneous mixture of gases and particles, all of which are components of ambient air pollution. Particulate matter is defined as particles suspended in the air of varying chemical composition and can be separated by particle size: coarse particulate matter less than 10 micrometers in diameter (PM10), fine particulate matter less than 2.5 micrometers in diameter (PM2.5), and ultrafine particulate matter less than 0.1 micrometers in diameter (PM<0.1). The gaseous products of fossil fuel combustion include: carbon monoxide (CO), nitrogen dioxide (NO2), sulfur dioxide (SO2), nitrogen oxides (NOx), and ozone (O3). PM are heterogeneous in chemical composition and can contain different metallic and nonmetallic compounds from different sources and may exert differential health effects. Ambient air pollution is the most robustly studied environmental exposure that has been linked to cardiovascular disease. Extent of the Exposure Exposure to ambient air pollution in urban and periurban communities in LMIC is often much higher than what is observed in the large metropolitan areas of high-income countries. According to the 2016 Urban Air Quality Database 98% of urban centers with more than 100,000 inhabitants in LMIC are annually exposed to PM2.5 levels greater than 10 µg/m3 and PM10 greater than 20 µg/m3, these levels are guidelines set forth by WHO15, 198. Lack of robust regulation of the sources of air pollution likely contribute to disproportionate air pollution exposure in LMIC208. While this database did not identify the sources of PM, common sources include diesel exhaust, industrial smokestack emissions, and biomass combustion. For example, all of the Latin American major metropolitan areas with 2013 air quality data exceed the World Health Organization standards for PM2.5 and PM10, with Bogotá, Colombia and Lima, Peru leading the cities with highest annual mean PM2.5 concentration at 35.1 µg/m3 and 31.5 µg/m3, respectively2,3. However, air pollution in China’s capital Beijing largely exceeds cities in Latin America with an annual mean PM2.5 concentration over 80 µg/m3 in 201516. Additionally, while much of the air pollution exposure in sub-Saharan Africa results from household air pollution from biomass fuel combustion and data on ambient air pollution exposure in the region are few, it is estimated the 32% of all West Africans are exposed to PM2.5 levels that exceed the World Health Organization limit16. Considerable heterogeneity of air pollution exposure can exist within large metropolitan areas as well, often disproportionately affecting low-income communities17. Looking to the future, the effect of temperature on PM2.5 concentration raises concern that the impact of air pollution exposure on health might continue to increase in the wake of climate change, differentially affecting LMIC with warm climates4,18. Mechanisms of Disease Ambient air pollution affects cardiovascular health largely due to systemic inflammation from the incorporation of fine particulate matter into the pulmonary interstitium19–21. Additionally, ultrafine particulate matter and the gaseous components of air pollution have the potential to directly enter the bloodstream22,23. In the presence of air pollutants, multiple biochemical effects have been observed including: increased oxidative stress through increased production of reactive oxygen species; increased inflammatory biomarkers including IL-6 and CRP; increased pro-thrombotic factors including D-Dimer, platelet activation, increased fibrinogen, thrombin generation and impaired fibrinolysis; increased expression of adhesive molecules on monocytes and leukocytes; and impaired endothelial function, including NO-mediated vasodilation20,24–33. The acute physiological response to exposure to ambient air pollution includes increased plasma viscosity, reduced heart rate variability, impaired vasoreactivity, vasoconstriction, increased blood pressure, and increased insulin resistance26,30,31,34–46. Impact on Cardiovascular Disease Chronic exposure to ambient air pollution has been associated with risk factors for cardiovascular disease in multiple cohorts. The association between chronic air pollution exposure and elevated blood pressure has been extensively studied, including data from multiethnic cohorts in several countries37,47–49. Additionally, some evidence has emerged supporting an association between air pollution exposure with elevated fasting glucose and Type 2 Diabetes Mellitus47,50,51. Yet, the data have not been entirely consistent and additional studies on the factors that increase vulnerability to the blood pressure effects of air pollution exposure are needed, including a greater understanding of the specific air pollutants that account for the observed cardiometabolic effects52,53. The majority of the studies of ambient air pollution and cardiovascular risk factors were conducted in high-income countries, with very few studies conducted in LMIC41. Exposure to ambient air pollution is associated with multiple measures of subclinical cardiovascular disease. Ambient air pollution has been associated with measures of subclinical atherosclerosis including carotid-intimal thickness and aortic atherosclerotic plaques54–58. There is evidence that air pollution exposure is also associated with the progression of coronary calcium59. Additionally, air pollution exposure has also been associated with adverse cardiac remodeling, including right and left ventricular hypertrophy55,60,61. While most of these studies were conducted in high-income countries, several small studies in LMIC have recently emerged including a study of occupational air pollution exposure and cardiac structure and function in Iran62. Beyond subclinical cardiovascular disease, large studies have demonstrated a strong association between ambient air pollution exposure and adverse cardiovascular outcomes. Acute ambient air pollution exposure has been associated with angina, stroke, acute myocardial infarction, heart failure hospitalization, arrhythmias, cardiac arrest, heart failure hospitalization and cardiovascular mortality48,63–81. Data that are specific to LMIC are largely conducted in upper middle-income countries including China and in Latin America48,82–89. Of note, almost no studies of air pollution and cardiovascular disease in sub-Saharan Africa have been published. The discrepancy between the relatively high exposure to ambient air pollution in LMIC and the lack of data specific to LIMC suggests that the public health impact is potentially underestimated. Household Air Pollution from Biomass Fuel Use While economic development in LMIC has improved access to electricity, natural gas and liquefied petroleum gas, many communities depend on biomass fuels for daily needs90. Biomass fuels include wood, charcoal, dung and crop residue, which are burned in indoor and outdoor stoves for cooking and heating. Similar to fossil fuel combustion, biomass fuels produce gases and particulate matter that are suspended in air, including carbon monoxide and fine particulate matter (PM2.5). Exposure to the components of biomass fuel combustion has been studied in several contexts in relation to cardiovascular disease risk factors and outcomes. The Extent of the Exposure Household air pollution from biomass fuel use affects 3 billion people worldwide, including 6.5 million Americans90, 199. While biomass fuel use can be found on every continent, it is more prevalent in resource-poor settings, disproportionately affecting low-income individuals in HIC and LMIC199. In many cultures women are more likely to perform household cooking, and thus are more highly exposed to smoke from biomass fuel use along with small children in the home. The geographic distribution of biomass fuel use can vary by region due to social, cultural, economic, and climate differences. For example, in the Andean region of South America, daily biomass fuel use is primarily confined to rural communities91. In contrast, a large study in peri-urban Malawi found that 70.9% of the 6,445 households surveyed use wood and/or charcoal for cooking92. Furthermore, older age and low education were associated with the use of wood for cooking. Understanding and addressing the social and cultural factors that contribute to biomass fuel use is critical and has implications for the implementation of improved cook-stove interventions. Mechanisms of Disease The biochemical and physiologic response to the air pollutants released from biomass fuel combustion has not been as extensively studied as air pollution from fossil fuel combustion. While both forms of combustion release fine particulate matter, the chemical composition of the particulate matter vary according to fuel source, and some studies suggest that the chemical composition and diameter of particulate matter has differential impact on cardiovascular disease outcomes69,93–95. Coarse particulate matter is often found in ocean spray, dust, and construction byproducts. Acute exposure to wood smoke has been showed to cause arterial stiffness and decreased heart rate variability96. Additionally, observational studies conducted in women in villages in eastern India observed increased pro-inflammatory cytokines, higher serum c-reactive protein, and higher reactive oxygen species generation in the women exposed to biomass fuel smoke97. Another study of women in rural India observed an increase in systolic blood pressure during cooking times during which there was also an increase in exposure to the air pollutant black carbon, a major component of soot98. Additional research on the acute biochemical and physiologic response to household air pollution from biomass fuel combustion is needed to better understand how this exposure differs from ambient air pollution. Impact on Cardiovascular Disease Exposure to biomass fuel smoke has been associated with cardiovascular risk factors in multiple observational studies. The most common cardiovascular risk factor associated with biomass fuel use is elevated blood pressure. Multiple cohort studies in China, Peru, Guatemala and Nicaragua have identified an association between exposure to biomass fuel smoke and elevated blood pressure99–104. Replacement of traditional cookstoves with cleaner burning cookstoves was associated with lower blood pressure101. In addition to observing differences in blood pressure, 2 large studies in China and Peru also observed an increased prevalence of hypertension in daily biomass fuel users100,103. Exposure to biomass fuel smoke has also been associated with subclinical cardiovascular disease in several small studies. In Guatemala, biomass fuel use was associated with changes in the ST segment of the electrocardiogram in women prior to participating in an improved cookstove trial105. These changes improved after the cookstove intervention, suggesting an improvement in myocardial ischemia. Additionally, a cross sectional study of 266 individuals in Puno, Peru found that chronic exposure to biomass fuel smoke was associated with increased carotid intima-media thickness and a higher prevalence of carotid atherosclerotic plaques106. However, contrary to what was previously hypothesized, in a sample from the same Peruvian cohort there was no association between biomass fuel use with elevated NT pro-BNP or right ventricular systolic pressure by echocardiography107. A small echocardiography study in a single hospital in Turkey observed that biomass fuel users had increased right ventricular systolic pressure and decreased left and right ventricular myocardial indices, indicating decreased biventricular systolic function108. However, the relationship between biomass fuel smoke exposure and cardiac structure and function is currently undergoing further examination in population-based cohorts. There have been conflicting results in studies of the association of household air pollution from biomass fuel use with outcomes, such as coronary heart disease and cardiovascular mortality. While the Global Burden of Disease Study estimated the global impact of household air pollution due to biomass fuel use based on the observed relationship between ambient air pollution exposure and cardiovascular events, very few studies have examined cardiovascular outcomes in biomass fuel users. Emerging data suggest an association between biomass fuel use and coronary heart disease109. In a study of participants living in the Brazilian Amazon, elderly individuals with increased exposure to biomass fuel smoke had increased cardiovascular mortality when compared to age-matched controls110. However, large cohorts in Iran and Bangladesh have failed to demonstrate an association between chronic biomass fuel use and cardiovascular mortality111,112. Additional studies that prospectively study cardiovascular outcomes in biomass fuel users compared to nonusers are needed to better quantify the impact of household air pollution on cardiovascular disease. Lead The acute and chronic neurological effects of lead exposure have been widely described in both high-income countries and LMIC113. However, less public attention has been paid to the cardiovascular impact of chronic lead exposure and the contribution of heavy metal exposure on the burden of cardiovascular disease in LMIC. Globally, it is estimated the lead exposure ranks #26 as a risk factor for disability-adjusted life-years lost, yet in sub-regions of Latin America and Southern Africa this ranking rises to #2090. Lead exposure in LMIC deserves close examination as a modifiable risk factor for cardiovascular disease and a potential target for intervention at the population level. Extent of the Exposure Globally, an estimated 26 million people are at risk for lead toxicity, resulting in a loss of 9 million disability-adjusted life-years114. While lead exposure exists in high-income countries and LMIC alike from lead pipes and paint, in general, the prevalence of lead exposure has not decreased in LMIC to the degree that has been observed in many high-income countries115,116. Tobacco use is a common mode of lead exposure in HIC and LMIC however, there are multiple sources of lead exposure that are specific to the industries and cultures of LMIC117. While leaded petroleum was banned from high-income countries many decades ago, its use in LMIC continues in Yemen, Algeria, and Iraq, polluting the air and soil114,118, 200. Additionally, occupational exposures in battery manufacturing and recycling factories have been well described, particularly in Kenya and several South Asian countries119,120, 203. Mining operations in Peru, Tanzania, Nigeria and Zambia have been associated with lead exposure not only for the workers at the mine, but also for the local communities located near the mines121–123. Toxic waste from other industrial sources, is also known to contaminate water and soil with lead113. Fishing and hunting with lead tools fashioned from industrial sources are associated with chronic lead exposure in Peruvian Amazon River Basin communities204. Moreover, the artisanal use of lead in pottery has also be a source of lead exposure in Latin America and Africa114,115,124,125, and leaded paints are still being sold and used in some LMIC, as noted in a recent study in Cambodia126. Independent of the source of the lead contamination, children are often the most vulnerable population exposed to lead, with often unmeasured detriment to the present and future neurological and cardiovascular health113,122,125,127–129. Mechanisms of Disease By promoting the generation of reactive oxygen species, lead increases oxidative stress in cardiovascular tissues and endothelial cells130. The increase in oxidative stress in the setting of lead exposure is also associated with decreased nitric oxide (NO) availability. Decreased NO availability in turn has been shown to cause sodium retention, vasoconstriction, and increased adrenergic tone130. Additionally, NFκB activation due to increased oxidative stress in the setting of lead exposure leads to the oxidation of LDL, increases the expression of adhesive molecules on monocytes and increases foam cell formation130. These processes in addition to platelet activation and vascular remodeling are the basis by which lead associated cardiovascular disease occurs130. Impact of Cardiovascular Disease Hypertension is the cardiovascular risk factor most greatly associated with lead exposure. Multiple studies in the United States in addition to several studies in LMIC have demonstrated a convincing association between even low levels of lead exposure and increased blood pressure, gestational hypertension incidence, and hypertension prevalence119,131–133. Moreover, some evidence suggests the lead exposure is also associated in with decreased heart rate variability131. However, emerging evidence suggests that lead exposure is also associated with other cardiometabolic derangements including increased fasting glucose, decreased HDL, increased total cholesterol, and increased prevalence of the metabolic syndrome124,134,135. Several of these studies of cardiometabolic impairment in the setting of lead exposure were conducted in LMIC settings, including multiple settings in West Africa and the Americas, thus highlighting the potential role of environmental exposures on non-communicable disease risk in LMIC. Lead exposure is associated with subclinical cardiovascular disease and cardiovascular outcomes. Increased carotid intimal medial thickness has been observed in association with increased serum lead levels in a Turkish population with concomitant renal disease136. Lead exposure has also been associated with reduced heart rate variability and abnormalities of cardiac structure and function, including increased left ventricular hypertrophy and decreased ejection fraction131. Clinical atherosclerotic disease has been observed in association with lead exposure including stroke, peripheral arterial disease, and coronary heart disease131,137. Increased exposure to lead has also been associated with increased cardiovascular mortality in several studies of the United States population131,138,139. Despite the considerable exposure to lead in LMIC, there are limited published data on lead exposure and cardiovascular disease outcomes in LMIC populations. Arsenic Arsenic is a naturally occurring metalloid and a contaminant of drinking water, soil, and food. In contrast to acute arsenic poisoning, chronic arsenic exposure can be more difficult to identify, but ultimately is association with multiple adverse health outcomes, including CVD114. Extent of the exposure Chronic arsenic exposure has been described in countries of all income levels and most commonly occurs from drinking wells contaminated with arsenic naturally present in the soil114. While arsenic contamination of wells within the US have been well documented, particularly in Native American reservations, arsenic contamination in LMIC countries including Bangladesh, India, Taiwan, and Turkey has also been well-documented140–144. However, as safe drinking water becomes more readily available within high-income countries, arsenic contamination from well water disproportionately affects low-income communities, such as Bangladesh where an estimated 20 million inhabitants consume arsenic contaminated water142,144, 205. Arsenic contamination of food such as rice represents a particularly important exposure risk factor for inhabitants of LMIC and constituents of global trade partners205–207. Mechanisms of disease Arsenic typically enters the body through the gastrointestinal tract and is metabolized in the liver where it undergoes methylation, yielding toxic intermediates145. Arsenic exposure is associated with increased inflammatory markers, including IL-6 and IL-8, and matrix metalloproteinase- 2 and -9146,147. In animal models, arsenic exposure leads to myocardial fibrosis, which is proposed to be the mechanism by which QT prolongation in electrocardiogram occurs in response to arsenic toxicity148. Additionally, endothelial dysfunction associated with arsenic exposure has also been observed. In Bangladesh, gene by environment interaction in relation to increases in blood pressure from arsenic exposure has been well described, demonstrating variable cardiotoxicity due to variable methylation of arsenic149–151. In a separate Bangladesh study, folate supplementation promoted urinary excretion of arsenic and may attenuate arsenic toxicity201. Impact on Cardiovascular Disease Chronic arsenic exposure has been associated with cardiovascular risk factors. While elevated blood pressure and hypertension in response to arsenic exposure has been observed in multiple LIMC settings including India, Bangladesh, Mexico and China, this observation has not been consistent152–158. Additionally, Type 2 Diabetes Mellitus, elevated triglycerides, and elevated total cholesterol have also been observed in association with arsenic exposure142,156,159. Chronic arsenic exposure is associated with subclinical cardiovascular disease, including increased carotid intimal-medial thickness, which has been observed in several studies in LMIC including Mexico and Bangladesh156,160,161. Moreover, left ventricular ejection fraction is reduced in children chronically exposed to arsenic in Mexico157. In terms of clinical cardiovascular disease, arsenic exposure is associated with peripheral arterial disease, cardiomyopathy, coronary heart disease, acute myocardial infarction, stroke, stroke mortality, and cardiovascular mortality162–171. Not only does arsenic exposure increase the risk of acquired heart disease, but also is associated with increased risk of congenital heart disease172. As more evidence is generated regarding the full spectrum of cardiovascular disease associated with chronic arsenic exposure, the potential cost of arsenic contamination in LMIC is being appreciated. Cadmium Cadmium does receive the same degree of attention from the lay public as lead or arsenic, however the public health burden in relation exposure to cadmium remains significant114. An estimated 5 million people are exposed chronically to cadmium, which has implications for cardiovascular disease risk at the population level in many LMIC114. Extent of the Exposure Similar to lead, cadmium exposure commonly occurs from tobacco smoking, an exposure that has been well described in high-income countries and LMIC132,173. Additionally, cadmium from mining, smelting, refining and industrial waste can also pollute air, water, and soil leading to the contamination of foods including leafy vegetables, fish, and shellfish114,173. Cadmium is also used in the production of plastics, fertilizers, and batteries114,174. Communities in LMIC, particularly low-income communities, may be chronically exposed to cadmium, an exposure that is only recently made apparent as heavy metal monitoring is implemented in communities, as illustrated by studies from Ghana and Uganda175,176. In fact, cadmium exposure is likely to increase in the coming decades in part due to electronic waste disposal, as seen in Nigeria,202. Beyond contamination of the environment, serum levels of cadmium from individuals living in LIMC can be several orders of magnitude greater than what observed in high-income countries177. Mechanisms of Disease Cadmium increases oxidative stress through the increased production and decreased metabolism of reactive oxygen species178. Moreover, cadmium has been shown to impair endothelial function179. Cadmium also has been associated with increased serum levels of galetin-3, a biomarker for myocardial fibrosis, in a population in Turkey180. Through these multiple mechanisms cadmium exposure is thought to cause cardiovascular disease. Impact on Cardiovascular Disease Similar to other environmental exposures, cadmium exposure is associated with elevated blood pressure and hypertension181–184. While much of the evidence was generated in high-income countries, a number of studies have been conducted in LIMC, including Thailand, China and Pakistan185–187. Of note, there are several studies that did not find an association between measured cadmium exposure and hypertension, suggesting that additional data on the genetic and environmental risk factors for cadmium-related hypertension is needed188,189. Cadmium has also been associated with cardiometabolic derangement including Type 2 Diabetes, as noted in a study from China187. Additionally, increased carotid intimal-medial thickness and carotid plaques also have been associated with cadmium exposure136. The evidence regarding the association between cadmium exposure and cardiovascular disease outcomes overwhelmingly comes from high-income countries. Cadmium exposure is associated with diseases of atherosclerosis including peripheral arterial disease, stroke, ischemic heart disease, and acute coronary syndromes182,190–193. Cadmium exposure has also been associated with incident heart failure, although it unclear what percentage of heart failure cases are ischemic versus nonischemic in etiology182,192,194. The largest studies of cadmium and cardiovascular disease are from US NHANES data and the Strong Heart Study of US Native Americans. In these cohorts, cadmium exposure was associated with cardiovascular mortality, thus highlighting the likely unmeasured mortality burden that cadmium exposure potentially has in LMIC195–197. Summary Environmental exposures in LIMC lie at the intersection of increased economic development and the rising public health burden of cardiovascular disease. Increasing evidence suggests an association of exposure to ambient air pollution, household air pollution from biomass fuel, lead, arsenic, and cadmium with multiple cardiovascular disease outcomes including hypertension, coronary heart disease, stroke, and cardiovascular mortality. While populations in LMIC are disproportionately exposed to environmental pollution, the bulk of evidence that links these exposures to cardiovascular disease is derived from populations in high-income countries. Low-income regions of high-income countries are at high risk of exposure. In order to better understand the extent to which environmental exposures contribute to the rising epidemic of cardiovascular disease in LMIC and develop interventions to reduce cardiovascular disease risk at the population level, additional research is needed. Figure 1 Summary of the Association Between Environmental Exposures, Pathophysiologic Mechanisms and Cardiovascular Disease. Panel A illustrates the multiple mechanisms by which selected environmental exposures cause cardiovascular injury. Panel B shows the multiple cardiovascular outcomes that are associated with environmental exposures. Table 1 The Origin of the Peer-Reviewed Literature on the Association Between Selected Environmental Exposures and Cardiovascular Disease by World Bank Country Income-Level Hypertension Subclinical Atherosclerosis Cardiac Structure and Function Coronary Heart Disease Heart Failure Cardiovascular Hospitalization Arrhythmia Stroke Cardiovascular Mortality Ambient Air Pollution Biomass Fuel Air Pollution Lead Arsenic Cadmium Data derived from low- and middle-income countries Data derived from high-income countries Data derived from both low- and middle-income countries and high-income countries Key Points Environmental exposures, including air pollution and heavy metal and metalloid contamination, are more prevalent in low- and middle-income countries. Exposure to air pollution in the form of ambient air pollution and household air pollution from biomass fuel use is associated with hypertension, acute myocardial infarction, heart failure, arrhythmia, sudden cardiac death and cardiovascular mortality. Lead, arsenic and cadmium exposure is associated with hypertension, coronary heart disease and cardiovascular mortality. There is increasing epidemiological evidence of an association of environmental exposures with cardiovascular risk factors and cardiovascular disease, yet most of the research has been conducted in high-income countries. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. The authors report no conflicts of interest. 1 Landrigan PJ Fuller R Horton R Environmental pollution, health, and development: a Lancet-Global Alliance on Health and Pollution-Icahn School of Medicine at Mount Sinai Commission Lancet 2015 386 10002 1429 1431 26466029 2 Bell ML Cifuentes LA Davis DL Cushing E Telles AG Gouveia N Environmental health indicators and a case study of air pollution in Latin American cities Environ. Res 2011 111 1 57 66 21075365 3 Green J Sanchez S Air Quality in Latin America: An Overview 2013 Washington, DC Clear Air Institute 4 Patz JA Frumkin H Holloway T Vimont DJ Haines A Climate Change: Challenges and Opportunities for Global Health JAMA : the journal of the American Medical Association 2014 5 Mining in Latin America: From conflict to cooperation The Economist 2016 6 Briggs D Environmental pollution and the global burden of disease Br Med Bull 2003 68 1 24 14757707 7 Landrigan PJ Fuller R Global health and environmental pollution Int J Public Health 2015 60 7 761 762 26135237 8 Bhatnagar A Environmental cardiology: studying mechanistic links between pollution and heart disease Circulation research 2006 99 7 692 705 17008598 9 Beaglehole R Bonita R Alleyne G UN High-Level Meeting on Non-Communicable Diseases: addressing four questions Lancet 2011 378 9789 449 455 21665266 10 Lozano R Naghavi M Foreman K Global and regional mortality from 235 causes of death for 20 age groups in 1990 and 2010: a systematic analysis for the Global Burden of Disease Study 2010 Lancet 2012 380 9859 2095 2128 23245604 11 Ordunez P Cardiovascular health in the Americas: facts, priorities and the UN high-level meeting on non-communicable diseases MEDICC review 2011 13 4 6 10 12 Ebrahim S Pearce N Smeeth L Casas JP Jaffar S Piot P Tackling non-communicable diseases in low- and middle-income countries: is the evidence from high-income countries all we need? PLoS medicine 2013 10 1 e1001377 23382655 13 Pearce N Ebrahim S McKee M Global prevention and control of NCDs: Limitations of the standard approach J Public Health Policy 2015 36 4 408 425 26377446 14 Burroughs Pena MS Bloomfield GS Cardiovascular disease research and the development agenda in low- and middle-income countries Glob Heart 2015 10 1 71 73 25754569 15 N O Chriscaden K Air Pollution levels rising in many of the world's poorest cities 2016 Accessed July, 6, 2016 2016; http://www.who.int/mediacentre/news/releases/2016/air-pollution-rising/en 16 Energy and Air Pollution 2016 Paris, France International Energy Agency/OECD 17 Bravo MA Bell ML Spatial heterogeneity of PM10 and O3 in Sao Paulo, Brazil, and implications for human health studies Journal of the Air & Waste Management Association (1995) 2011 61 1 69 77 21305890 18 Kioumourtzoglou MA Schwartz J James P Dominici F Zanobetti A PM2.5 and Mortality in 207 US Cities: Modification by Temperature and City Characteristics Epidemiology (Cambridge, Mass.) 2016 27 2 221 227 19 Seaton A MacNee W Donaldson K Godden D Particulate air pollution and acute health effects Lancet 1995 345 8943 176 178 7741860 20 Mills NL Donaldson K Hadoke PW Adverse cardiovascular effects of air pollution Nature clinical practice. Cardiovascular medicine 2009 6 1 36 44 21 Brook RD Rajagopalan S Pope CA 3rd Particulate matter air pollution and cardiovascular disease: An update to the scientific statement from the American Heart Association Circulation 2010 121 21 2331 2378 20458016 22 Nemmar A Hoet PH Vanquickenborne B Passage of inhaled particles into the blood circulation in humans Circulation 2002 105 4 411 414 11815420 23 Mills NL Amin N Robinson SD Do inhaled carbon nanoparticles translocate directly into the circulation in humans? American journal of respiratory and critical care medicine 2006 173 4 426 431 16339922 24 Pekkanen J Brunner EJ Anderson HR Tiittanen P Atkinson RW Daily concentrations of air pollution and plasma fibrinogen in London Occupational and environmental medicine 2000 57 12 818 822 11077010 25 Peters A Frohlich M Doring A Particulate air pollution is associated with an acute phase response in men; results from the MONICA-Augsburg Study European heart journal 2001 22 14 1198 1204 11440492 26 Mills NL Tornqvist H Robinson SD Diesel exhaust inhalation causes vascular dysfunction and impaired endogenous fibrinolysis Circulation 2005 112 25 3930 3936 16365212 27 Frampton MW Stewart JC Oberdorster G Inhalation of ultrafine particles alters blood leukocyte expression of adhesion molecules in humans Environmental health perspectives 2006 114 1 51 58 28 Tornqvist H Mills NL Gonzalez M Persistent endothelial dysfunction in humans after diesel exhaust inhalation American journal of respiratory and critical care medicine 2007 176 4 395 400 17446340 29 Lucking AJ Lundback M Mills NL Diesel exhaust inhalation increases thrombus formation in man European heart journal 2008 29 24 3043 3051 18952612 30 Strak M Hoek G Steenhof M Components of ambient air pollution affect thrombin generation in healthy humans: the RAPTES project Occupational and environmental medicine 2013 70 5 332 340 23378445 31 Hajat A Allison M Diez-Roux AV Long-term exposure to air pollution and markers of inflammation, coagulation, and endothelial activation: a repeat-measures analysis in the Multi-Ethnic Study of Atherosclerosis (MESA) Epidemiology (Cambridge, Mass.) 2015 26 3 310 320 32 Adar SD D'Souza J Mendelsohn-Victor K Markers of inflammation and coagulation after long-term exposure to coarse particulate matter: a cross-sectional analysis from the multi-ethnic study of atherosclerosis Environmental health perspectives 2015 123 6 541 548 25616153 33 Li W Wilker EH Dorans KS Short-Term Exposure to Air Pollution and Biomarkers of Oxidative Stress: The Framingham Heart Study J Am Heart Assoc 2016 5 5 34 Peters A Doring A Wichmann HE Koenig W Increased plasma viscosity during an air pollution episode: a link to mortality? Lancet 1997 349 9065 1582 1587 9174559 35 Brook RD Brook JR Urch B Vincent R Rajagopalan S Silverman F Inhalation of fine particulate air pollution and ozone causes acute arterial vasoconstriction in healthy adults Circulation 2002 105 13 1534 1536 11927516 36 Urch B Silverman F Corey P Acute blood pressure responses in healthy adults during controlled air pollution exposures Environmental health perspectives 2005 113 8 1052 1055 16079078 37 Auchincloss AH Diez Roux AV Dvonch JT Associations between recent exposure to ambient fine particulate matter and blood pressure in the Multi-ethnic Study of Atherosclerosis (MESA) Environmental health perspectives 2008 116 4 486 491 18414631 38 Brook RD Rajagopalan S Particulate matter, air pollution, and blood pressure Journal of the American Society of Hypertension : JASH 2009 3 5 332 350 20409976 39 Brook RD Urch B Dvonch JT Insights into the mechanisms and mediators of the effects of air pollution exposure on blood pressure and vascular function in healthy humans Hypertension 2009 54 3 659 667 19620518 40 Park SK Auchincloss AH O'Neill MS Particulate air pollution, metabolic syndrome, and heart rate variability: the multi-ethnic study of atherosclerosis (MESA) Environmental health perspectives 2010 118 10 1406 1411 20529761 41 Baccarelli A Barretta F Dou C Effects of particulate air pollution on blood pressure in a highly exposed population in Beijing, China: a repeated-measure study Environmental health : a global access science source 2011 10 108 22188661 42 Shields KN Cavallari JM Hunt MJ Traffic-related air pollution exposures and changes in heart rate variability in Mexico City: a panel study Environmental health : a global access science source 2013 12 7 23327098 43 Brook RD Sun Z Brook JR Extreme Air Pollution Conditions Adversely Affect Blood Pressure and Insulin Resistance: The Air Pollution and Cardiometabolic Disease Study Hypertension 2015 44 Green R Broadwin R Malig B Long- and Short-term Exposure to Air Pollution and Inflammatory/Hemostatic Markers in Midlife Women Epidemiology (Cambridge, Mass.) 2016 27 2 211 220 45 Haberzettl P O'Toole TE Bhatnagar A Conklin DJ Exposure to Fine Particulate Air Pollution Causes Vascular Insulin Resistance by Inducing Pulmonary Oxidative Stress Environmental health perspectives 2016 46 Ljungman PL Wilker EH Rice MB The Impact of Multipollutant Clusters on the Association Between Fine Particulate Air Pollution and Microvascular Function Epidemiology (Cambridge, Mass.) 2016 27 2 194 201 47 Coogan PF White LF Jerrett M Air pollution and incidence of hypertension and diabetes mellitus in black women living in Los Angeles Circulation 2012 125 6 767 772 22219348 48 Langrish JP Li X Wang S Reducing personal exposure to particulate air pollution improves cardiovascular health in patients with coronary heart disease Environmental health perspectives 2012 120 3 367 372 22389220 49 Kirwa K Eliot MN Wang Y Residential Proximity to Major Roadways and Prevalent Hypertension Among Postmenopausal Women: Results From the Women's Health Initiative San Diego Cohort J Am Heart Assoc 2014 3 5 50 Peng C Bind MC Colicino E Particulate Air Pollution and Fasting Blood Glucose in Non-Diabetic Individuals: Associations and Epigenetic Mediation in the Normative Aging Study, 2000–2011 Environmental health perspectives 2016 51 Yitshak Sade M Kloog I Liberty IF Schwartz J Novack V The Association Between Air Pollution Exposure and Glucose and Lipids Levels The Journal of clinical endocrinology and metabolism 2016 101 6 2460 2467 27218271 52 Park SK Adar SD O'Neill MS Long-term exposure to air pollution and type 2 diabetes mellitus in a multiethnic cohort American journal of epidemiology 2015 181 5 327 336 25693777 53 Coogan PF White LF Yu J PM2.5 and Diabetes and Hypertension Incidence in the Black Women's Health Study Epidemiology (Cambridge, Mass.) 2016 27 2 202 210 54 Breton CV Wang X Mack WJ Carotid artery intima-media thickness in college students: race/ethnicity matters Atherosclerosis 2011 217 2 441 446 21679950 55 Gill EA Curl CL Adar SD Air pollution and cardiovascular disease in the Multi-Ethnic Study of Atherosclerosis Progress in cardiovascular diseases 2011 53 5 353 360 21414470 56 Breton CV Wang X Mack WJ Childhood air pollutant exposure and carotid artery intima-media thickness in young adults Circulation 2012 126 13 1614 1620 22896588 57 Rivera M Basagana X Aguilera I Association between long-term exposure to traffic-related air pollution and subclinical atherosclerosis: the REGICOR study Environmental health perspectives 2013 121 2 223 230 23384708 58 Wilker EH Mittleman MA Coull BA Long-term Exposure to Black Carbon and Carotid Intima-Media Thickness: The Normative Aging Study Environmental health perspectives 2013 121 9 1061 1067 23820848 59 Kaufman JD Adar SD Barr RG Association between air pollution and coronary artery calcification within six metropolitan areas in the USA (the Multi-Ethnic Study of Atherosclerosis and Air Pollution): a longitudinal cohort study Lancet 2016 60 Leary PJ Kaufman JD Barr RG Traffic-related air pollution and the right ventricle. The multi-ethnic study of atherosclerosis American journal of respiratory and critical care medicine 2014 189 9 1093 1100 24593877 61 Liu Y Goodson JM Zhang B Chin MT Air pollution and adverse cardiac remodeling: clinical effects and basic mechanisms Front Physiol 2015 6 162 26042051 62 Golshahi J Sadeghi M Saqira M Zavar R Sadeghifar M Roohafza H Exposure to occupational air pollution and cardiac function in workers of the Esfahan Steel Industry, Iran Environmental science and pollution research international 2016 23 12 11759 11765 26946505 63 Dockery DW Pope CA 3rd Xu X An association between air pollution and mortality in six U.S. cities N Engl J. Med 1993 329 24 1753 1759 8179653 64 Hoffmann B Moebus S Stang A Residence close to high traffic and prevalence of coronary heart disease European heart journal 2006 27 22 2696 2702 17003049 65 Miller KA Siscovick DS Sheppard L Long-term exposure to air pollution and incidence of cardiovascular events in women N Engl J. Med 2007 356 5 447 458 17267905 66 Peng RD Chang HH Bell ML Coarse particulate matter air pollution and hospital admissions for cardiovascular and respiratory diseases among Medicare patients JAMA : the journal of the American Medical Association 2008 299 18 2172 2179 18477784 67 Puett RC Schwartz J Hart JE Chronic particulate exposure, mortality, and coronary heart disease in the nurses' health study American journal of epidemiology 2008 168 10 1161 1168 18835862 68 Bell ML Peng RD Dominici F Samet JM Emergency hospital admissions for cardiovascular diseases and ambient levels of carbon monoxide: results for 126 United States urban counties, 1999–2005 Circulation 2009 120 11 949 955 19720933 69 Peng RD Bell ML Geyh AS Emergency admissions for cardiovascular and respiratory diseases and the chemical composition of fine particle air pollution Environmental health perspectives 2009 117 6 957 963 19590690 70 Mustafic H Jabre P Caussin C Main air pollutants and myocardial infarction: a systematic review and meta-analysis JAMA : the journal of the American Medical Association 2012 307 7 713 721 22337682 71 Cesaroni G Badaloni C Gariazzo C Long-term exposure to urban air pollution and mortality in a cohort of more than a million adults in Rome Environmental health perspectives 2013 121 3 324 331 23308401 72 Chen H Goldberg MS Burnett RT Jerrett M Wheeler AJ Villeneuve PJ Long-term exposure to traffic-related air pollution and cardiovascular mortality Epidemiology (Cambridge, Mass.) 2013 24 1 35 43 73 Ensor KB Raun LH Persse D A case-crossover analysis of out-of-hospital cardiac arrest and air pollution Circulation 2013 127 11 1192 1199 23406673 74 Hart JE Rimm EB Rexrode KM Laden F Changes in traffic exposure and the risk of incident myocardial infarction and all-cause mortality Epidemiology (Cambridge, Mass.) 2013 24 5 734 742 75 Hoek G Krishnan RM Beelen R Long-term air pollution exposure and cardio- respiratory mortality: a review Environmental health : a global access science source 2013 12 1 43 23714370 76 Madrigano J Kloog I Goldberg R Coull BA Mittleman MA Schwartz J Long-term exposure to PM2.5 and incidence of acute myocardial infarction Environmental health perspectives 2013 121 2 192 196 23204289 77 Shah AS Langrish JP Nair H Global association of air pollution and heart failure: a systematic review and meta-analysis Lancet 2013 78 Wichmann J Folke F Torp-Pedersen C Out-of-hospital cardiac arrests and outdoor air pollution exposure in Copenhagen, Denmark PloS. one 2013 8 1 e53684 23341975 79 Cesaroni G Forastiere F Stafoggia M Long term exposure to ambient air pollution and incidence of acute coronary events: prospective cohort study and meta-analysis in 11 European cohorts from the ESCAPE Project BMJ 2014 348 f7412 24452269 80 Milojevic A Wilkinson P Armstrong B Bhaskaran K Smeeth L Hajat S Short-term effects of air pollution on a range of cardiovascular events in England and Wales: case-crossover analysis of the MINAP database, hospital admissions and mortality Heart 2014 100 14 1093 1098 24952943 81 Stockfelt L Andersson EM Molnar P Long term effects of residential NOx exposure on total and cause-specific mortality and incidence of myocardial infarction in a Swedish cohort Environ. Res 2015 142 197 206 26163761 82 Martins LC Pereira LA Lin CA The effects of air pollution on cardiovascular diseases: lag structures Rev Saude Publica 2006 40 4 677 683 17063245 83 O'Neill MS Bell ML Ranjit N Air pollution and mortality in Latin America: the role of education Epidemiology (Cambridge, Mass.) 2008 19 6 810 819 84 Liu L Breitner S Schneider A Size-fractioned particulate air pollution and cardiovascular emergency room visits in Beijing, China Environ. Res 2013 121 52 63 23375554 85 Freitas CU Leon AP Juger W Gouveia N Air pollution and its impacts on health in Vitoria, Espirito Santo, Brazil Rev Saude Publica 2016 50 4 26982960 86 Huang F Chen R Shen Y Kan H Kuang X The Impact of the 2013 Eastern China Smog on Outpatient Visits for Coronary Heart Disease in Shanghai, China International journal of environmental research and public health 2016 13 7 87 Phung D Hien TT Linh HN Air pollution and risk of respiratory and cardiovascular hospitalizations in the most populous city in Vietnam The Science of the total environment 2016 557–558 322 330 88 Ye X Peng L Kan H Acute Effects of Particulate Air Pollution on the Incidence of Coronary Heart Disease in Shanghai, China PloS. one 2016 11 3 e0151119 26942767 89 Zuniga J Tarajia M Herrera V Urriola W Gomez B Motta J Assessment of the Possible Association of Air Pollutants PM10, O3, NO2 With an Increase in Cardiovascular, Respiratory, and Diabetes Mortality in Panama City: A 2003 to 2013 Data Analysis Medicine 2016 95 2 e2464 26765444 90 Lim SS Vos T Flaxman AD A comparative risk assessment of burden of disease and injury attributable to 67 risk factors and risk factor clusters in 21 regions, 1990–2010: a systematic analysis for the Global Burden of Disease Study 2010 Lancet 2012 380 9859 2224 2260 23245609 91 Pollard SL Williams DL Breysse PN A cross-sectional study of determinants of indoor environmental exposures in households with and without chronic exposure to biomass fuel smoke Environmental health : a global access science source 2014 13 1 21 24655424 92 Piddock KC Gordon SB Ngwira A A cross-sectional study of household biomass fuel use among a periurban population in Malawi Ann Am Thorac. Soc 2014 11 6 915 924 24960156 93 Meng Q Richmond-Bryant J Lu SE Cardiovascular outcomes and the physical and chemical properties of metal ions found in particulate matter air pollution: a QICAR study Environmental health perspectives 2013 121 5 558 564 23462649 94 Sun M Kaufman JD Kim SY Particulate matter components and subclinical atherosclerosis: common approaches to estimating exposure in a Multi-Ethnic Study of Atherosclerosis cross-sectional study Environmental health : a global access science source 2013 12 39 23641873 95 Kim SY Sheppard L Kaufman JD Individual-level concentrations of fine particulate matter chemical components and subclinical atherosclerosis: a cross-sectional analysis based on 2 advanced exposure prediction models in the multi-ethnic study of atherosclerosis American journal of epidemiology 2014 180 7 718 728 25164422 96 Unosson J Blomberg A Sandstrom T Exposure to wood smoke increases arterial stiffness and decreases heart rate variability in humans Particle and fibre toxicology 2013 10 20 23742058 97 Dutta A Ray MR Banerjee A Systemic inflammatory changes and increased oxidative stress in rural Indian women cooking with biomass fuels Toxicology and applied pharmacology 2012 261 3 255 262 22521606 98 Norris C Goldberg MS Marshall JD A panel study of the acute effects of personal exposure to household air pollution on ambulatory blood pressure in rural Indian women Environ. Res 2016 147 331 342 26928412 99 Baumgartner J Schauer JJ Ezzati M Indoor air pollution and blood pressure in adult women living in rural China Environmental health perspectives 2011 119 10 1390 1395 21724522 100 Lee MS Hang JQ Zhang FY Dai HL Su L Christiani DC In-home solid fuel use and cardiovascular disease: a cross-sectional analysis of the Shanghai Putuo study Environmental health : a global access science source 2012 11 18 22455369 101 Clark ML Bachand AM Heiderscheidt JM Impact of a cleaner-burning cookstove intervention on blood pressure in Nicaraguan women Indoor. air 2013 23 2 105 114 22913364 102 Baumgartner J Zhang Y Schauer JJ Huang W Wang Y Ezzati M Highway proximity and black carbon from cookstoves as a risk factor for higher blood pressure in rural China Proceedings of the National Academy of Sciences of the United States of America 2014 111 36 13229 13234 25157159 103 Burroughs Pena M Romero KM Velazquez EJ Relationship between daily exposure to biomass fuel smoke and blood pressure in high-altitude Peru Hypertension 2015 65 5 1134 1140 25753976 104 Dutta A Roychoudhury S Chowdhury S Ray MR Changes in sputum cytology, airway inflammation and oxidative stress due to chronic inhalation of biomass smoke during cooking in premenopausal rural Indian women International journal of hygiene and environmental health 2013 216 3 301 308 22771078 105 McCracken J Smith KR Stone P Diaz A Arana B Schwartz J Intervention to lower household wood smoke exposure in Guatemala reduces ST-segment depression on electrocardiograms Environmental health perspectives 2011 119 11 1562 1568 21669557 106 Painschab MS Davila-Roman VG Gilman RH Chronic exposure to biomass fuel is associated with increased carotid artery intima-media thickness and a higher prevalence of atherosclerotic plaque Heart 2013 99 14 984 991 23619984 107 Caravedo MA Painschab MS Davila-Roman VG Lack of association between chronic exposure to biomass fuel smoke and markers of right ventricular pressure overload at high altitude American heart journal 2014 168 5 731 738 25440802 108 Kargin R Kargin F Mutlu H Long-term exposure to biomass fuel and its relation to systolic and diastolic biventricular performance in addition to obstructive and restrictive lung diseases Echocardiography 2011 28 1 52 61 20738366 109 Fatmi Z Coggon D Coronary heart disease and household air pollution from use of solid fuel: a systematic review British medical bulletin 2016 118 1 91 109 27151956 110 Nunes KV Ignotti E Hacon Sde S Circulatory disease mortality rates in the elderly and exposure to PM(2.5) generated by biomass burning in the Brazilian Amazon in 2005 Cadernos de saude publica 2013 29 3 589 598 23532293 111 Alam D Chowdhury M Siddiquee A Adult cardiopulmonary mortality and indoor air pollution: A 10-year retrospective cohort study in a low-income rural setting Global Heart 2012 7 3 215 221 25691484 112 Mitter S Vedanthan R Islami F Household Fuel Use and Cardiovascular Disease Mortality: Golestan Cohort Study Circulation 2016 133 2360 2369 27297340 113 Chatham-Stephens K Caravanos J Ericson B Landrigan P Fuller R The pediatric burden of disease from lead exposure at toxic waste sites in low and middle income countries Environ. Res 2014 132 379 383 24853976 114 World's Worst Pollution Problems 2015 New York Pure Earth and Green Cross Switzerland 115 Tong S von Schirnding YE Prapamontol T Environmental lead exposure: a public health problem of global dimensions Bulletin of the World Health Organization 2000 78 9 1068 1077 11019456 116 Muntner P Menke A DeSalvo KB Rabito FA Batuman V Continued decline in blood lead levels among adults in the United States: the National Health and Nutrition Examination Surveys Arch Intern. Med 2005 165 18 2155 2161 16217007 117 Richter PA Bishop EE Wang J Kaufmann R Trends in tobacco smoke exposure and blood lead levels among youths and adults in the United States: the National Health and Nutrition Examination Survey, 1999–2008 Prev Chronic. Dis 2013 10 E213 24355106 118 Jones D Diop A Block M Smith-Jones A Smith-Jones A Assessment and remediation of lead contamination in Senegal J Health Pollution 2011 2 37 47 119 Ahmad SA Khan MH Khandker S Blood lead levels and health problems of lead acid battery workers in Bangladesh Scientific World Journal 2014 2014 974104 24707223 120 Basit S Karim N Munshi AB Occupational lead toxicity in battery workers Pak J. Med Sci 2015 31 4 775 780 26430401 121 van Geen A Bravo C Gil V Sherpa S Jack D Lead exposure from soil in Peruvian mining towns: a national assessment supported by two contrasting examples Bulletin of the World Health Organization 2012 90 12 878 886 23284193 122 Yabe J Nakayama SM Ikenaka Y Lead poisoning in children from townships in the vicinity of a lead-zinc mine in Kabwe, Zambia Chemosphere 2015 119 941 947 25303652 123 Getso K Hadejia I Sabitu K Prevalence and determinants of childhood lead poisoning in Zamfar State, Nigeria J Health Pollution 2014 6 1 9 124 Ademuyiwa O Ugbaja RN Idumebor F Adebawo O Plasma lipid profiles and risk of cardiovascular disease in occupational lead exposure in Abeokuta, Nigeria Lipids Health. Dis 2005 4 19 16191200 125 Caravanos J Dowling R Tellez-Rojo MM Blood lead levels in Mexico and pediatric burden of disease implications Ann Glob Health 2014 80 4 269 277 25459328 126 Lim S Murphy T Wilson K Irvine K Leaded paint in cambodia- Pilot-scale assessment J Health Pollution 2015 9 18 24 127 Caravanos J Fuller R Robinson S Centers for Disease C, Prevention Notes from the field: severe environmental contamination and elevated blood lead levels among children - Zambia, 2014 MMWR Morb Mortal Wkly. Rep 2014 63 44 1013 25375074 128 Laborde A Tomasina F Bianchi F Children's health in Latin America: the influence of environmental exposures Environ Health Perspect 2015 123 3 201 209 25499717 129 Suk WA Ahanchian H Asante KA Environmental Pollution: An Under-recognized Threat to Children's Health, Especially in Low- and Middle-Income Countries Environ Health Perspect 2016 124 3 A41 A45 26930243 130 Vaziri ND Mechanisms of lead-induced hypertension and cardiovascular disease Am J Physiol Heart Circ Physiol 2008 295 2 H454 H465 18567711 131 Navas-Acien A Guallar E Silbergeld EK Rothenberg SJ Lead exposure and cardiovascular disease--a systematic review Environ Health Perspect 2007 115 3 472 482 17431501 132 Afridi HI Kazi TG Kazi NG Evaluation of cadmium, lead, nickel and zinc status in biological samples of smokers and nonsmokers hypertensive patients J Hum Hypertens 2010 24 1 34 43 20010608 133 Shiue I Higher urinary heavy metal, arsenic, and phthalate concentrations in people with high blood pressure: US NHANES, 2009–2010 Blood pressure 2014 23 6 363 369 24945898 134 Rhee SY Hwang YC Woo JT Blood lead is significantly associated with metabolic syndrome in Korean adults: an analysis based on the Korea National Health and Nutrition Examination Survey (KNHANES), 2008 Cardiovascular diabetology 2013 12 9 23302150 135 Ettinger AS Bovet P Plange-Rhule J Distribution of metals exposure and associations with cardiometabolic risk factors in the "Modeling the Epidemiologic Transition Study" Environ Health 2014 13 90 25374160 136 Ari E Kaya Y Demir H Asicioglu E Keskin S The correlation of serum trace elements and heavy metals with carotid artery atherosclerosis in maintenance hemodialysis patients Biological trace element research 2011 144 1–3 351 359 21647752 137 Arslan C Altan H Akgun OO Trace elements and toxic heavy metals play a role in Buerger disease and atherosclerotic peripheral arterial occlusive disease International angiology : a journal of the International Union of Angiology 2010 29 6 489 495 21173730 138 Schober SE Mirel LB Graubard BI Brody DJ Flegal KM Blood lead levels and death from all causes, cardiovascular disease, and cancer: results from the NHANES III mortality study Environ Health Perspect 2006 114 10 1538 1541 17035139 139 Aoki Y Brody DJ Flegal KM Fakhouri TH Parker JD Axelrad DA Blood Lead and Other Metal Biomarkers as Risk Factors for Cardiovascular Disease Mortality Medicine 2016 95 1 e2223 26735529 140 Flanagan SV Johnston RB Zheng Y Arsenic in tube well water in Bangladesh: health and economic impacts and implications for arsenic mitigation Bulletin of the World Health Organization 2012 90 11 839 846 23226896 141 Chen Y Wu F Liu M A prospective study of arsenic exposure, arsenic methylation capacity, and risk of cardiovascular disease in Bangladesh Environ Health Perspect 2013 121 7 832 838 23665672 142 Chen CJ Health hazards and mitigation of chronic poisoning from arsenic in drinking water: Taiwan experiences Reviews on environmental health 2014 29 1–2 13 19 24552958 143 Gunduz O Bakar C Simsek C Statistical analysis of causes of death (2005–2010) in villages of Simav Plain, Turkey, with high arsenic levels in drinking water supplies Archives of environmental & occupational health 2015 70 1 35 46 24455995 144 Joca L Sacks JD Moore D Lee JS Sams R 2nd Cowden J Systematic review of differential inorganic arsenic exposure in minority, low-income, and indigenous populations in the United States Environment international 2016 145 Sidhu MS Desai KP Lynch HN Rhomberg LR Beck BD Venditti FJ Mechanisms of action for arsenic in cardiovascular toxicity and implications for risk assessment Toxicology 2015 331 78 99 25771173 146 Das N Paul S Chatterjee D Arsenic exposure through drinking water increases the risk of liver and cardiovascular diseases in the population of West Bengal, India BMC public health 2012 12 639 22883023 147 Islam MS Mohanto NC Karim MR Elevated concentrations of serum matrix metalloproteinase-2 and-9 and their associations with circulating markers of cardiovascular diseases in chronic arsenic-exposed individuals Environ Health 2015 14 92 26637202 148 Chu W Li C Qu X Arsenic-induced interstitial myocardial fibrosis reveals a new insight into drug-induced long QT syndrome Cardiovascular research 2012 96 1 90 98 22853924 149 Wu F Jasmine F Kibriya MG Interaction between arsenic exposure from drinking water and genetic susceptibility in carotid intima-media thickness in Bangladesh Toxicology and applied pharmacology 2014 276 3 195 203 24593923 150 Farzan SF Karagas MR Jiang J Gene-arsenic interaction in longitudinal changes of blood pressure: Findings from the Health Effects of Arsenic Longitudinal Study (HEALS) in Bangladesh Toxicology and applied pharmacology 2015 288 1 95 105 26220686 151 Wu F Jasmine F Kibriya MG Interaction between arsenic exposure from drinking water and genetic polymorphisms on cardiovascular disease in Bangladesh: a prospective case-cohort study Environ Health Perspect 2015 123 5 451 457 25575156 152 Abhyankar LN Jones MR Guallar E Navas-Acien A Arsenic exposure and hypertension: a systematic review Environ Health Perspect 2012 120 4 494 500 22138666 153 Guha Mazumder D Purkayastha I Ghose A Hypertension in chronic arsenic exposure: A case control study in West Bengal Journal of environmental science and health. Part. A, Toxic/hazardous substances & environmental engineering 2012 47 11 1514 1520 154 Islam MR Khan I Attia J Association between hypertension and chronic arsenic exposure in drinking water: a cross-sectional study in Bangladesh Int J Environ Res Public Health 2012 9 12 4522 4536 23222207 155 Li X Li B Xi S Zheng Q Lv X Sun G Prolonged environmental exposure of arsenic through drinking water on the risk of hypertension and type 2 diabetes Environmental science and pollution research international 2013 20 11 8151 8161 23649600 156 Stea F Bianchi F Cori L Sicari R Cardiovascular effects of arsenic: clinical and epidemiological findings Environmental science and pollution research international 2014 21 1 244 251 24019140 157 Osorio-Yanez C Ayllon-Vergara JC Arreola-Mendoza L Blood pressure, left ventricular geometry, and systolic function in children exposed to inorganic arsenic Environ Health Perspect 2015 123 6 629 635 25738397 158 Ameer SS Engstrom K Harari F Concha G Vahter M Broberg K The effects of arsenic exposure on blood pressure and early risk markers of cardiovascular disease: Evidence for population differences Environ. Res 2015 140 32 36 25825128 159 Mendez MA Gonzalez-Horta C Sanchez-Ramirez B Chronic Exposure to Arsenic and Markers of Cardiometabolic Risk: A Cross- Sectional Study in Chihuahua, Mexico Environ Health Perspect 2016 124 1 104 111 26068977 160 Osorio-Yanez C Ayllon-Vergara JC Aguilar-Madrid G Carotid intima-media thickness and plasma asymmetric dimethylarginine in Mexican children exposed to inorganic arsenic Environ Health Perspect 2013 121 9 1090 1096 23757599 161 Wu F Molinaro P Chen Y Arsenic Exposure and Subclinical Endpoints of Cardiovascular Diseases Current environmental health reports 2014 1 2 148 162 25013752 162 Cheng TJ Ke DS Guo HR The association between arsenic exposure from drinking water and cerebrovascular disease mortality in Taiwan Water research 2010 44 19 5770 5776 20561663 163 Jovanovic DD Paunovic K Manojlovic DD Jakovljevic B Rasic-Milutinovic Z Dojcinovic BP Arsenic in drinking water and acute coronary syndrome in Zrenjanin municipality, Serbia Environ. Res 2012 117 75 82 22626473 164 Moon K Guallar E Navas-Acien A Arsenic exposure and cardiovascular disease: an updated systematic review Current atherosclerosis reports 2012 14 6 542 555 22968315 165 Chen Y Karagas MR Arsenic and cardiovascular disease: new evidence from the United States Annals of internal medicine 2013 159 10 713 714 24061555 166 Moon KA Guallar E Umans JG Association between exposure to low to moderate arsenic levels and incident cardiovascular disease. A prospective cohort study Annals of internal medicine 2013 159 10 649 659 24061511 167 Rahman M Sohel N Yunus M A prospective cohort study of stroke mortality and arsenic in drinking water in Bangladeshi adults BMC public health 2014 14 174 24548416 168 D'Ippoliti D Santelli E De Sario M Scortichini M Davoli M Michelozzi P Arsenic in Drinking Water and Mortality for Cancer and Chronic Diseases in Central Italy, 1990–2010 PloS. one 2015 10 9 e0138182 26383851 169 Farzan SF Chen Y Rees JR Zens MS Karagas MR Risk of death from cardiovascular disease associated with low-level arsenic exposure among long-term smokers in a US population-based study Toxicology and applied pharmacology 2015 287 2 93 97 26048586 170 James KA Byers T Hokanson JE Meliker JR Zerbe GO Marshall JA Association between lifetime exposure to inorganic arsenic in drinking water and coronary heart disease in Colorado residents Environ Health Perspect 2015 123 2 128 134 25350952 171 Wade TJ Xia Y Mumford J Cardiovascular disease and arsenic exposure in Inner Mongolia, China: a case control study Environ Health 2015 14 35 25889926 172 Rudnai T Sandor J Kadar M Arsenic in drinking water and congenital heart anomalies in Hungary International journal of hygiene and environmental health 2014 217 8 813 818 24916166 173 Riederer AM Belova A George BJ Anastas PT Urinary cadmium in the 1999–2008 U.S. National Health and Nutrition Examination Survey (NHANES) Environmental science & technology 2013 47 2 1137 1147 23253114 174 Tellez-Plaza M Jones MR Dominguez-Lucas A Guallar E Navas-Acien A Cadmium exposure and clinical cardiovascular disease: a systematic review Current atherosclerosis reports 2013 15 10 356 23955722 175 Vowotor M Hood C Sackey S An assessment of heavy metal pollution in sediments of a tropical lagoon: A case study of the Benya Lagoon, Komenda Edina Eguafo Abrem Municipality KEEA)- Ghana J Health Pollution 2014 6 26 39 176 Namuhani N Kimumwe C Soil contamination with heavy metals around Jinia Steel Rolling Mills in Jinja Municipality, Uganda J Health Pollution 2015 9 61 67 177 Orisakwe O Blum J Zelikoff J Metal pollution in Nigeria: A biomonitoring update J Health Pollution 2014 6 40 52 178 Valko M Morris H Cronin MT Metals, toxicity and oxidative stress Curr Med Chem 2005 12 10 1161 1208 15892631 179 Kaya Y Ari E Demir H Gecit I Beytur A Kaspar C Serum cadmium levels are independently associated with endothelial function in hemodialysis patients International urology and nephrology 2012 44 5 1487 1492 21904850 180 Yazihan N Kocak MK Akcil E Involvement of galectin-3 in cadmium-induced cardiac toxicity Anadolu kardiyoloji dergisi : AKD = the Anatolian journal of cardiology 2011 11 6 479 484 21788204 181 Al-Saleh I Shinwari N Mashhour A Cadmium and mercury levels in Saudi women and its possible relationship with hypertension Biological trace element research 2006 112 1 13 29 16943613 182 Peters JL Perlstein TS Perry MJ McNeely E Weuve J Cadmium exposure in association with history of stroke and heart failure Environ. Res 2010 110 2 199 206 20060521 183 Lee BK Kim Y Association of blood cadmium with hypertension in the Korean general population: analysis of the 2008–2010 Korean National Health and Nutrition Examination Survey data American journal of industrial medicine 2012 55 11 1060 1067 22692952 184 Caciari T Sancini A Fioravanti M Cadmium and hypertension in exposed workers: A meta-analysis International journal of occupational medicine and environmental health 2013 26 3 440 456 23857371 185 Swaddiwudhipong W Nguntra P Kaewnate Y Human health effects from cadmium exposure: Comparison between persons living in cadmium-contaminated and non-contaminated areas in northwestern Thailand The Southeast Asian journal of tropical medicine and public health 2015 46 1 133 142 26513915 186 Afridi HI Kazi TG Talpur FN Distribution of arsenic, cadmium, lead, and nickel levels in biological samples of Pakistani hypertensive patients and control subjects Clinical laboratory 2014 60 8 1309 1318 25185416 187 Liu B Feng W Wang J Association of urinary metals levels with type 2 diabetes risk in coke oven workers Environ Pollut 2015 210 1 8 26689646 188 Gallagher CM Meliker JR Blood and urine cadmium, blood pressure, and hypertension: a systematic review and meta-analysis Environ Health Perspect 2010 118 12 1676 1684 20716508 189 Kurihara I Kobayashi E Suwazono Y Association between exposure to cadmium and blood pressure in Japanese peoples Arch Environ Health 2004 59 12 711 716 16789481 190 Lee MS Park SK Hu H Lee S Cadmium exposure and cardiovascular disease in the 2005 Korea National Health and Nutrition Examination Survey Environ. Res 2011 111 1 171 176 21055738 191 Tellez-Plaza M Navas-Acien A Crainiceanu CM Sharrett AR Guallar E Cadmium and peripheral arterial disease: gender differences in the 1999–2004 US National Health and Nutrition Examination Survey American journal of epidemiology 2010 172 6 671 681 20693268 192 Tellez-Plaza M Guallar E Howard BV Cadmium exposure and incident cardiovascular disease Epidemiology (Cambridge, Mass.) 2013 24 3 421 429 193 Barregard L Sallsten G Fagerberg B Blood Cadmium Levels and Incident Cardiovascular Events during Follow-up in a Population-Based Cohort of Swedish Adults: The Malmo Diet and Cancer Study Environ Health Perspect 2015 194 Borne Y Barregard L Persson M Hedblad B Fagerberg B Engstrom G Cadmium exposure and incidence of heart failure and atrial fibrillation: a population-based prospective cohort study BMJ open 2015 5 6 e007366 195 Menke A Muntner P Silbergeld EK Platz EA Guallar E Cadmium levels in urine and mortality among U.S. adults Environ Health Perspect 2009 117 2 190 196 19270787 196 Tellez-Plaza M Navas-Acien A Menke A Crainiceanu CM Pastor-Barriuso R Guallar E Cadmium exposure and all-cause and cardiovascular mortality in the U.S. general population Environ Health Perspect 2012 120 7 1017 1022 22472185 197 Larsson SC Wolk A Urinary cadmium and mortality from all causes, cancer and cardiovascular disease in the general population: systematic review and meta-analysis of cohort studies International journal of epidemiology 2015 198 Ambient (Outdoor) Air Quality and Health WHO website Accessed August 7, 2016 http://www.who.int/mediacentre/factsheets/fs313/en. Updated March 2014 199 Rogalsky D Mendola P Metts T Martin W Estimating the number of low-income Americans exposed to household air pollution from burning solid fuels Environ Health Perspect 2014 122 8 1 5 24184886 200 Leaded Petrol Phase-out: Global Status as at June 2016, United Nations Environment Programme website Accessed August 8, 2016 http://www.unep.org/Transport/new/PCFV/pdf/Maps_Matrices/world/lead/MapWorldLead_June2016.pdf. Updated June 2016 201 Gamble MV Liu X Ahsan H Folate and arsenic metabolism: a double-blind, placebo-controlled folic acid–supplementation trial in Bangladesh The American journal of clinical nutrition 2006 84 5 1093 1101 17093162 202 Schmidt CW Unfair Trade e-Waste in Africa Environmental Health Perspectives 2006 114 4 A232 A235 16581530 203 Were FH Moturi MC Gottesfeld P Lead exposure and blood pressure among workers in diverse industrial plants in Kenya J Occup Environ. Hyg 2014 11 11 706 715 24690073 204 Anticona C Bergdahl IA San Sebastian M Lead exposure among children from native communities of the Peruvian Amazon basin Revista Panamericana de Salud Pública 2012 31 4 296 302 22652969 205 Smith AH Lingas EO Rahman M Contamination of drinking-water by arsenic in Bangladesh: a public health emergency Bulletin of the World Health Organization 2000 78 9 1093 1103 11019458 206 Gilbert-Diamond D Cottingham KL Gruber JF Rice consumption contributes to arsenic exposure in US women Proceedings of the National Academy of Sciences of the United States of America 2011 108 51 20656 20660 22143778 207 Davis MA Mackenzie TA Cottingham KL Gilbert-Diamond D Punshon T Karagas MR Rice Consumption and Urinary Arsenic Concentrations in U.S. Children Environmental Health Perspectives 2012 120 10 1418 1424 23008276 208 Actions on Air Quality: Regional Reports, United Nations Environment Programme website Accessed August 10, 2016 http://www.unep.org/transport/airquality/regionalreports.asp. Updated June 2014

PMC005xxxxxx/PMC5130158.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 101573691 39703 Cell Rep Cell Rep Cell reports 2211-1247 27829147 5130158 10.1016/j.celrep.2016.10.030 NIHMS827641 Article Adaptive resistance to an inhibitor of chromosomal instability in human cancer cells Orr Bernardo 12 Talje Lama 3 Liu Zhexian 3 Kwok Benjamin H. 34 Compton Duane A. 12 1 Dept. of Biochemistry and Cell Biology, Geisel School of Medicine at Dartmouth, Hanover, NH 03755, United States 2 Norris Cotton Cancer Center, Lebanon, NH 03766, United States 3 Institute for Research in Immunology and Cancer, Université de Montréal, Québec H3T 1J4, Canada 4 Département de Médecine, Université de Montréal, Québec H3T 1J4, Canada General correspondence: Duane A. Compton, Ph.D., Geisel School of Medicine at Dartmouth, Department of Biochemistry and Cell Biology, Hanover, NH, 03755, USA, duane.a.compton@dartmouth.edu Correspondence related to chemical compounds: Benjamin H. Kwok, PhD., Institute for Research in Immunology and Cancer, Université de Montréal, Québec H3T 1J4, Canada, benjamin.kwok@gmail.com 5 11 2016 8 11 2016 30 11 2016 17 7 17551763 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. Summary Karyotype diversity is a hallmark of solid tumors that contributes to intratumor heterogeneity. This diversity is generated by persistent chromosome mis-segregation associated with chromosomal instability (CIN). CIN correlates with tumor relapse and is thought to promote drug resistance by creating a vast genomic landscape through which karyotypically unique clones survive lethal drug selection. We explore this proposition using a small molecule (UMK57) that suppresses chromosome mis-segregation in CIN cancer cells by potentiating the activity of the kinesin-13 protein MCAK. Sub-lethal doses of UMK57 destabilize kinetochore-microtubule (k-MT) attachments during mitosis to increase chromosome segregation fidelity. Surprisingly, chromosome mis-segregation rebounds in UMK57-treated cancer cells within a few days. This rapid relapse is driven by alterations in the Aurora B signaling pathway that hyper-stabilize k-MT attachments, and is reversible following UMK57 removal. Thus, cancer cells display adaptive resistance to therapies targeting CIN through rapid and reversible changes to mitotic signaling networks. Graphical Abstract Introduction Aneuploidy is a hallmark of solid tumors (Luo et al., 2009) and commonly arises in tumors through the frequent mis-segregation of whole chromosomes as a consequence of CIN (Geigl et al., 2008; Lengauer et al., 1997). Persistent chromosome mis-segregation is a major driver of intra-tumor heterogeneity (Heppner, 1984), a genomic change that is proposed to allow cells to acquire new phenotypes (Duesberg et al., 2000; Gerlinger and Swanton, 2010). Accordingly, CIN positively correlates with poor patient prognosis (Bakhoum et al., 2011), multidrug resistance (Lee et al., 2011) and tumor relapse (Sotillo et al., 2010). The prevailing model posits that CIN generates a genomic landscape from which clones and sub-clones with specific karyotypes emerge from the population through survival of targeted therapy and/or other selective pressures (Greaves and Maley, 2012). Directly testing this model requires the development of tools that specifically suppress CIN in human cancer cells. The root cause of CIN is the persistence of errors in k-MT attachments in mitosis (Thompson and Compton, 2008). Errors in k-MT attachment arise spontaneously during mitosis and are efficiently corrected in diploid cells to preserve genome integrity. The correction process relies on the frequent detachment of microtubules from kinetochores to allow for microtubules with the proper orientation to make attachments. It was previously demonstrated that many CIN cancer cells have hyper-stable k-MT attachments and fail to efficiently correct k-MT attachment errors (Bakhoum et al., 2009a). Importantly, strategically destabilizing k-MT attachments by over-expressing the microtubule destabilizing kinesin-13 proteins Kif2b and MCAK suppresses CIN in cancer cells and establishes a causative relationship between the stability of k-MT attachments and the rate of chromosome mis-segregation (Bakhoum et al., 2009a; 2014; 2009b; Kleyman et al., 2014). These data provide proof of concept for a strategy to suppress CIN in human cancer cells. Unfortunately, this strategy is severely limited by the requirement for protein overexpression in tumor cells. To overcome this technical limitation and to examine how cancer cells respond to the suppression of CIN, we examine the effects of a cell permeable small molecule that specifically activates the kinesin-13 protein MCAK. Results and Discussion UMK57 potentiates MCAK activity Current strategies for the suppression CIN in cancer cells rely on the manipulation of proteins involved in the regulation of k-MT attachments during mitosis (Bakhoum et al., 2009b; Ertych et al., 2014), which prove to be limiting outside of cell culture. To overcome these limitations, a high throughput screen was performed to identify small molecules that modulate the activities of kinesin-13 proteins (Talje et al., 2014). This screen identified a kinesin-13 inhibitor that was previously reported (Talje et al., 2014). This screen also identified a family of compounds that potentiate the microtubule depolymerizing activity of kinesin-13 proteins in vitro, and the complete characterization of how these compounds affect the biochemistry of kinesin-13 proteins in vitro will be provided elsewhere. Here, we focus on the effects of one of these compounds (UMK57) on chromosome segregation during mitosis in vivo. We focused on UMK57 because it shows no inhibitory effect on the ATPase activity of any kinesins tested (Figure S1A), but specifically enhances MCAK-dependent microtubule depolymerization as measured using in vitro ultracentrifugation microtubule sedimentation (Figure S1B) and microscopy (Figure S1C) assays. Additionally, UMK57 inhibits cell proliferation in a dose-dependent manner (Figure S1D). In contrast, a chemically related analog differing only in one chemical group (UMK95) has no effect on MCAK-mediated microtubule depolymerization (Figure S1B) or cell proliferation (Figure S1D), demonstrating the potency and specificity of UMK57 (Figure S1E & S1F). Titration experiments in U2OS cells demonstrate that 100nM UMK57 is the optimal dose to achieve the maximal effect on the fidelity of chromosome segregation, without significantly affecting mitotic progression (Figure 1A) and therefore all treatments were done at this concentration unless stated otherwise. Treatment of cells with UMK95, a chemically related but inactive compound (Figure S1F), shows no detectable effect on chromosome segregation (Figure 1B) underscoring the specificity of UMK57 (Arrowsmith et al., 2015) (Figure S1E). It is important to note that treatment of cells with 100nM UMK57 does not alter total MCAK levels (Figure 1C), MCAK localization or spindle organization at different stages of mitosis (Figures 1D). Also, very few UMK57-treated cells transiently expressing GFP or GFP-Kif2b show discernible defects in chromosome alignment or spindle bipolarity (Figures 1E and 1F). In contrast, a substantial fraction of UMK57-treated cells transiently expressing GFP-MCAK display defects in chromosome alignment and spindle organization indicating that MCAK over-expression renders cells hypersensitive to UMK57 (Figures 1E and 1F). In addition, UMK57 reduces the rate of chromosome mis-segregation in control and Kif2b-depleted cells (Figures 1G and 1H). However, there is no change in the rate of chromosome mis-segregation in MCAK-depleted cells (Figures 1G and 1H). This indicates that the effects of UMK57 in mitosis are dependent on the presence of MCAK. UMK57 suppresses CIN in human cancer cells To determine the effect of UMK57 on chromosome segregation in different human cell lines, we quantified lagging chromosome rates in anaphase in CIN cancer lines (U2OS, HeLa and SW-620) and in non-transformed diploid cell lines (hTERT-immortalized RPE-1 and BJ). Treatment with UMK57 significantly reduces lagging chromosome rates in all cancer cell lines tested but has no effect on chromosome segregation in the non-transformed diploid cell lines (Figures 2A and 2B). Given that many lagging chromosomes in anaphase can segregate to the correct daughter cell (Thompson and Compton, 2011), we tested whether UMK57 treatment also reduces the rate of chromosome non-disjunction. We measured the fate of sister chromatids of a single chromosome in HCT116 cells recovering from a monastrol-induced mitotic arrest using previously described methods (Lampson and Kapoor, 2004; Thompson and Compton, 2011; 2008). Cells treated with UMK57 display a significant reduction in the rate of chromosome non-disjunction as compared to control cells (Figures 2C and 2D). Next, we quantified the stability of k-MT attachments in UMK57 treated U2OS cells by measuring the rate of dissipation of fluorescence after photoactivation of GFP-tubulin (Zhai et al., 1995). The stability of k-MT attachments during metaphase in mitotic cells treated with UMK57 is reduced by more than 35% when compared to control cells, consistent with the potentiation of the microtubule depolymerase activity of MCAK (Figures 2E, S2A–C). There is no significant change in the stability of k-MT attachments in prometaphase in agreement with previous reports indicating that MCAK preferentially destabilizes metaphase k-MTs in this cell line (Bakhoum et al., 2009b). Moreover, there are no significant differences in the fractions of microtubules in the stable (k-MTs) versus unstable (non-k-MTs) populations (Figure S2D) or in the turnover rate of the unstable population in UMK57-treated cells (Figure S2E). Despite the destabilization of k-MTs during metaphase, UMK57 treatment at this dose does not significantly alter mitotic progression (Figure 1A) but induces a subtle, yet significant reduction in the inter-kinetochore distance of aligned kinetochore pairs (Figure 2F). However, UMK57 treatment does not affect chromosome bi-orientation or the spindle assembly checkpoint as judged by the localization patterns of Aurora B kinase (Figure S2F), Astrin (Figure S2G) and the levels of mitotic checkpoint protein BubR1 at kinetochores (Figure 2G and 2H). Furthermore, RPE-1 and U2OS cell lines display differential sensitivity to UMK57 but not Vinblastine, indicating that the effects of UMK57 are distinct from a compound that directly targets tubulin (Figures S2H–K). Collectively, these results demonstrate that UMK57 acts through an MCAK-dependent process to specifically destabilize k-MT attachments during metaphase and reduces chromosome mis-segregation with little or no appreciable effects on other mitotic processes. Thus, UMK57 is a small molecule that specifically promotes the correction of k-MT attachment errors to suppress chromosome mis-segregation in CIN cancer cells with no detectable effects on non-transformed diploid cells. Resistance to UMK57 treatment arises rapidly The gold standard for assaying CIN is through the measurement of karyotypic diversity in colonies derived from single cells where chromosomally stable cells generate karyotypically homogeneous populations and CIN cells generate karyotypically heterogeneous populations (Bakhoum et al., 2009b; Lengauer et al., 1997; Thompson and Compton, 2008). To determine if UMK57 suppresses CIN in this context we used fluorescence in situ hybridization (FISH) to determine the modal chromosome number for two different chromosomes and scored the percentage of cells that deviate from that mode in two independent colonies grown from single cells for each condition. Non-transformed RPE-1 cells maintain stable diploid karyotypes and the fraction of cells that deviate from the modal chromosome copy number within each colony is very low and insensitive to UMK57 treatment (Figures 3A and 3B). U2OS cells are CIN and the fraction of cells that deviate from the modal chromosome copy number within each colony is high. Surprisingly, treatment of U2OS cells with UMK57 continuously during colony growth (6–8 weeks) did not reduce the fraction of cells that deviate from the modal chromosome copy number within each colony (Figures 3A and 3B). Thus, CIN cells treated with UMK57 for extended time periods maintain karyotypic heterogeneity despite the fact that it suppresses chromosome mis-segregation in this cell line when measured immediately (Figures 2A–D). This cellular response was not observed previously when CIN was suppressed by overexpression of MCAK or Kif2b (Bakhoum et al., 2009b) indicating a distinction between protein overexpression and chemical enhancement of protein activity as methods to suppress CIN in cancer cells. Serial transfer of culture media over multiple days followed by immediate quantification of chromosome segregation defects in U2OS cells verifies that UMK57 retains biological activity in culture media for up to 120hr (Figure S3A). Nevertheless, lagging chromosome rates in cells treated with UMK57 return to control levels within 72hr (Figure 3C) and this phenomenon was also observed if media containing UMK57 was renewed every 24hr (Figure S3B). This time-dependent decrease in responsiveness to UMK57 was also observed in multiple CIN cancer cell lines (Figures 3D–F). In addition, Verapamil, a potent inhibitor of the multidrug efflux pump encoded by the MDR1 gene, was effective at enhancing the response of these cells to doxorubicin (Figures S3C and S3D) but it did not significantly alter the decrease in responsiveness to UMK57 (Figures 3D–F). This suggests that resistance to UMK57 occurs independently of multidrug efflux pump activity. Aurora B kinase influences the rate of resistance to UMK57 We then focused on the events that change in mitotic cells during extended treatment with UMK57. Kinetochore-microtubule attachment stability in metaphase cells is reduced equivalently in cells treated with UMK57 for 1hr or 72–120hr, thus verifying the persistent UMK57-dependent MCAK activation during extended treatment (Figures 2E and 4A). Accordingly, the quantity of phosphorylated MCAK at centromeres does not change between cells treated with UMK57 for 1hr or >72hr (Figure 4B), suggesting MCAK levels remain unchanged within this time frame. Moreover, extended treatment with UMK57 did not change the fractions of microtubules in the stable (k-MTs) versus unstable (non-k-MTs) populations or in the turnover rate of the unstable population in UMK57-treated cells (Figures S3E–H). In contrast, k-MT attachment stability in prometaphase cells significantly increases upon exposure to UMK57 for 72–120hr or 6–8 weeks (Figure 4A), the latter being a time frame that mimics the clonal cell growth assay (Figures 3A and 3B). Strikingly, these changes in k-MT attachment stability in response to UMK57 treatment are reversible upon UMK57 removal (Figure 4A). In addition to the stabilization of k-MT attachments in prometaphase, there is a substantial reduction in active Aurora B kinase at centromeres in cells treated with UMK57 for >72hr (Figures 4C and 4D). However, this did not lead to corresponding changes in substrate phosphorylation as phopsho-KNL1 levels were specifically increased at unaligned kinetochores (Figure S3I) and there were no changes in phospho-Hec1 levels (Figure S3J). This suggests that the Aurora B signaling network (e.g. the spatially regulated activities of both the kinase and the counteracting phosphatases) is altered in cells exposed to UMK57 for >72hr. Thus, the capacity of UMK57 to initially improve chromosome segregation fidelity in CIN cancer cells by destabilizing k-MT attachments in metaphase, is superseded after just a few days of UMK57 treatment by changes in Aurora B kinase signaling and hyperstabilization of k-MT attachments in prometaphase. Consistently, single cell clones grown continuously in the presence of UMK57 display reduced sensitivity to the Aurora B inhibitor ZM447439, but show no changes in sensitivity to other compounds such as Doxorubicin, Taxol, Vinblastine or BI-2536 (Figure S4A–G). Next, we dampened Aurora B kinase signaling using the inhibitor ZM447439 at a dose (250nM) sufficient to partially inhibit Aurora B kinase activity as determined by the effect on phosphorylation of Histone H3 at Ser10 in early prometaphase (Figure S4H). Partial inhibition of Aurora B kinase signaling does not increase lagging chromosome rates or affect the ability of UMK57 to reduce lagging chromosome rates in U2OS cells when measured within 1hr (Figure S4I). However, partial inhibition of Aurora B kinase activity prevents cells from regaining high levels of lagging chromosomes in anaphase after treatment with UMK57 for 72hr (Figure 4E and S4J). Partial Aurora B kinase inhibition also delayed the emergence of resistance to UMK57 in SW-620 cells (Figure S4K), but did not do so in HeLa cells (Figure S4L). This demonstrates that changes in the Aurora B kinase signaling pathway is a common means to develop resistance to UMK57, and that the full dynamic range of Aurora B kinase signaling is required for the emergence of this resistance within a 72hr time frame. The effect of partial Aurora B kinase inhibition in preventing resistance to UMK57 was only transient because cells ultimately regained high rates of lagging chromosomes in anaphase by 96hr in the presence of UMK57 and the Aurora B kinase inhibitor (Figures 4E and S4K). Thus, cells achieve resistance to UMK57 through multiple pathways suggesting that there is strong selective pressure to maintain high levels of chromosome mis-segregation in CIN cancer cells. This work demonstrates that CIN cancer cells can systematically alter mitotic signaling pathways to generate resistance to therapies designed to target the specific mitotic defects that cause persistent chromosome mis-segregation. Emergence of resistance through rapid and reversible alterations of cellular signaling pathways is defined as adaptive resistance (Figure 4F). This is distinct from acquired resistance (Duesberg et al., 2000; Gerlinger and Swanton, 2010; Gottesman, 2002) that arises through the selection of rare, genetically unique cells from heterogeneous populations (Figure 4F). Other distinguishing features of cells that become drug resistant through an adaptive mechanism is the retention of genome complexity in the population (Figure 4F) as we observed in our FISH experiments (Figure 3A and 3B) and the emergence of resistance at sub-lethal doses (Figures 3D–F and S2A). However, it is likely that both adaptive and acquired resistance contribute to drug resistance in a tumor cell population. Adaptive resistance has been previously described in bacterial (Fernández et al., 2011) and fungal (Walker et al., 2010) systems and likely surfaces in tumor cells based on their need to tolerate the myriad stresses associated with their microenvironment and unrestrained growth (proteotoxic, mitotic, oxidative, DNA damage and metabolic) (Luo et al., 2009). The capacity of cells to undergo adaptive resistance is rooted in cellular signaling and/or transcriptional networks that involve feedback-dependent homeostatic control (Chandarlapaty, 2012; Pisco et al., 2013) such as the mechanisms proposed to regulate k-MT attachment stability during mitosis (Godek et al., 2014). Importantly, these data suggest that the oncogenic pathways that drive cellular transformation impose changes onto the signaling networks that regulate chromosome segregation fidelity during mitosis (Orr and Compton, 2013) to promote and preserve CIN. These results provide insight into the observations that CIN cancer cells frequently display multi-drug resistance (Lee et al., 2011) and that the efficacy of targeted therapies based on molecular profiling in tumor cells (Barretina et al., 2012; Garnett et al., 2012) has been limited by rapid alterations of cellular signaling and/or transcriptional networks responsible for generating adaptive resistance (Litvin et al., 2015; Muranen et al., 2012; Sun et al., 2014). Experimental Procedures Drug Treatments UMK57 was used for all assays at 100nM (~5-fold lower than the calculated IC50 in U2OS cells) since this is the lowest concentration of UMK57 that results in maximal suppression of lagging chromosomes and was used at 100–2000nM for cell proliferation assays only. The following drugs were also used: Monastrol (100μM; Tocris Bioscience), MG132 (5μM; Sigma Aldrich), Paclitaxel/Taxol (1–100nM; Biotang, Inc), Doxorubicin (1–100nM; Sigma Aldrich), Vinblastine (0.1–2nM, Sigma Aldrich), BI-2536 (2–500nM; synthesized in house), ZM447439 (100–3000nM; Tocris Bioscience), Verapamil (6μM; Sigma Aldrich). All controls for drug treatment were performed using 0.1% DMSO. Photoactivation experiments k–MT attachment stability was measured in U2OS cells expressing α-tubulin tagged with photoactivatable GFP (plasmid provided by Alexey Khodjakov). Differential interference contrast (DIC) microscopy was used to identify mitotic cells and fluorescent images were generated and acquired using Quorum WaveFX-X1 spinning disk confocal system (Quorum Technologies) equipped with Mosaic digital mirror for photoactivation (Andor Technology) and Hamamatsu ImageEM camera. Cells were defined as being in prometaphase or metaphase on the basis of chromosome alignment using DIC optics. Microtubules were locally activated in one half-spindle. Fluorescence images captured every 15s for 4min with a 100X oil-immersion 1.4 numerical aperture objective. To quantify fluorescence dissipation after photoactivation, mean pixel intensities were quantified within a rectangular area surrounding the region of highest fluorescence intensity and background subtracted using an equally sized area from the non-activated half-spindle using Quorum WaveFX software. Fluorescence intensities were normalized to the first time point after photoactivation for each cell following background subtraction and correction for photobleaching. Correction for photobleaching was calculated by normalizing to values of fluorescence loss obtained from photoactivated 100nM Taxol-stabilized spindles where the photoactivated region did not dissipate. To measure microtubule turnover, the average intensity at each time point was fit to a double exponential curve A1*exp(k1*t) + A2*exp(k2*t) using MatLab (Mathworks), in which t is time, A1 represents the less stable (non-k-MT) population and A2 the more stable (k-MT) population with decay rates of k1 and k2, respectively. When the equation for the double exponential curve is solved, the rate constants as well as the percentage of microtubules for the fast (non-k-MTs) and the slow (k-MTs) process are obtained. The turnover half-life for each process was calculated as -ln2/k for each population of microtubules. All experiments were performed in the presence of MG132 (5μM) to prevent mitotic exit. Cell transfections Plasmid transfections were done with FuGENE 6 (Roche Diagnostics), and cells analyzed 24hr later by immunofluorescence. Plasmids used were GFP (pEGFP plasmid with modified multiple cloning site to incorporate AscI and PacI sites, gift from Aaron Straight), GFP-Kif2b and GFP-MCAK (gifts from Linda Wordeman). Short interfering RNA (siRNA) transfections were performed using Oligofectamine (Invitrogen), and cells analyzed 48hr later. RNA duplexes for Mock siRNA (Silencer® Negative Control No. 4), Kif2b (Silencer® Select Pre-designed; 5′-GGACCUGGAUAUCAUCACCtt-3′) or MCAK (Silencer® Select Validated siRNA; 5′-CAAAGUAUCUGGAGAACCAtt-3′) were purchased from Ambion and used at a concentration of 200nM. Chromosome mis-segregation assay For analyzing chromosome segregation in interphase cells, cells were arrested in mitosis with monastrol (100 μM) for 15h and then media replaced with media containing DMSO or UMK57 (100nM). Mitotic cells were isolated by shake-off and plated at low density on coverslips for 8h (to allow cells to complete cytokinesis and reenter interphase). For fixing, cells were pre-extracted with pre-warmed PHEM with 1% Triton X-100 for 5min then fixed with 3.5% paraformaldehyde. Cells stained with DAPI and mounted using ProLong® Gold antifade reagent (Molecular Probes). Chromosomal instability (CIN) assay U2OS or RPE-1 cells plated at a low density (50–100 cells/plate) in 100mm plates and single-cell clones were subsequently isolated and expanded gradually until each clone occupied 2 × 100mm plates as a monolayer (6–8 weeks for U2OS cells; 4–6 weeks for RPE-1 cells). Cells were then rinsed with PBS, treated with KCl (75mM) for 30min and then fixed in and washed twice with methanol-acetic acid (3:1). FISH was performed using centromeric α-satellite probes for chromosomes 2 and 3 (Cytocell) according to manufacturer’s instructions. Supplementary Material supplement We thank all members of our laboratory for stimulating discussions and critical feedback in the interpretation and analysis of results. We also thank Jennifer DeLuca, Iain Cheeseman, Ryoma Ohi and Dave Cortez for kindly providing important materials and reagents. This work was supported by a grant from the National Institutes of Health (GM51542 to D.A.C.). B.H.K acknowledges funding from the Canadian Institute of Health Research (MOP-97928), the Cancer Research Society (CRS), the Institute for Research in Immunology and Cancer (IRIC) and IRICoR. The authors declare no conflict of interests. Figure 1 UMK57 targets MCAK in vivo but does not alter its levels or localization in cells (A) Percentage of mitotic cells in prometaphase, metaphase or anaphase in cells treated with DMSO or increasing concentrations of UMK57 for <1h (n≥200 mitotic cells). For each concentration of UMK57, the percentage of anaphases with lagging chromosomes is shown below (n≥150 anaphases). (B) Lagging chromosome rates of U2OS cells treated with DMSO or UMK95, a chemical probe which is a structurally similar to UMK57, but is inactive in vitro (n≥150 anaphases); n.s. = P > 0.05 using Fisher’s Exact two-tailed test. (C) Quantitative western blot of mitotic U2OS cell extracts treated with MG132 for 4hr to enrich the mitotic population and prevent mitotic exit, and then treated with DMSO, ZM447439 or UMK57 for 1hr and blotted for total MCAK and α-Tubulin. Numbers below indicate the relative quantities (%) for each protein. (D) U2OS cells treated with DMSO or UMK57 for <1hr and then fixed and stained to reveal DNA, ACA (blue), MCAK (green) or α-Tubulin. Insets represent 5X magnifications of selected regions (further contrasted for better visualization). Scale bar = 5μm. (E) U2OS cells transiently over-expressing GFP, GFP-Kif2b or GFP-MCAK were treated with UMK57 (100nM) and then fixed and stained to reveal DNA, α-Tubulin and ACA. Scale bar = 5μm. (F) Percentage of prometaphase cells with defects in spindle organization in cells over-expressing GFP, GFP-Kif2b or GFP-MCAK and treated with DMSO or UMK57 (n>150 prometaphase cells). (G) Total cell lysates of Mock-, Kif2b- and MCAK-depleted U2OS cells were immunoblotted for MCAK and α-Tubulin (loading control). (H) Quantification of lagging chromosome rates in Mock-, Kif2b- or MCAK-depleted cells incubated with DMSO or UMK57 for <1h (n>100 anaphases). **P ≤ 0.01 and n.s. = P > 0.05 using Fisher’s Exact two-tailed test. Figure 2 UMK57 reduces lagging chromosome rates in human cancer cell lines by destabilizing k-MT attachments in metaphase (A) Representative immunofluorescence image of U2OS cells in anaphase treated with DMSO or UMK57 for <1hr and then fixed and stained to reveal DNA (blue) and ACA (red). Inset represents 5X magnification of selected region (further contrasted for better visualization). Scale bar = 5μm. (B) Percentage of lagging chromosomes in anaphase in multiple cell lines after treatment with DMSO or UMK57 for <1hr (n>300 anaphases). Error bars represent SD from three independent experiments. *P ≤ 0.05, **P ≤ 0.01 and n.s.= P > 0.05 using Fisher’s Exact two-tailed test. (C) Representative immunofluorescence image of segregation phenotypes in HCT116 cells expressing LacI-GFP with LacO arrays integrated into a single chromosome, that were fixed and imaged after completion of cytokinesis. Scale bar = 5μm. (D) Quantification of total mis-segregation events in HCT116 cells expressing LacI-GFP with LacO arrays integrated into a single chromosome subjected to a monastrol washout into media containing DMSO or UMK57 and allowed to complete mitosis and re-enter interphase. Total mis-segregation was quantified as depicted in panel (C) (n>600 pairs of daughter nuclei). **P ≤ 0.01 using Fisher’s Exact two-tailed test. (E) Calculated k-MT half-life in U2OS cells treated with 5μM MG132 (Control) or 5μM MG132 + UMK57 (100nM); n=12–19 cells. Graph shows mean ± SEM; **P ≤ 0.01 and n.s.= P > 0.05 using two-tailed t-test. (F) Quantification of inter-kinetochore distance as measured by distance between Hec1 staining in sister kinetochores in cells treated with DMSO or UMK57. Red bars indicate the mean and red error bars the SEM; n ≥ 150 kinetochores; *P ≤ 0.05 and n.s.= P > 0.05 using two-tailed t-test. (G) Cells treated with DMSO or UMK57 for <1h and then fixed and stained to reveal DNA (blue), BubR1 (green) and Hec1. Insets represent 5X magnifications of selected regions (further contrasted for better visualization). Scale bar = 5μm. (H) Quantification of BubR1 levels at kinetochores (normalized to Hec1) in U2OS cells treated with UMK57 (100nM) for <1hr as quantified by immunofluorescence. n≥150 kinetochores from 10–20 cells; n.s.= P > 0.05 using two-tailed t-test. Figure 3 Long-term exposure to UMK57 results in decreased responsiveness to UMK57-mediated suppression of CIN (A) Example of FISH data from U2OS single cell clones fixed and stained with DNA (blue) to visualize nuclei and with probes specific for centromeric α-satellite DNA of chromosomes 2 (green) and 3 (red). White arrows indicate cells whose chromosome copy number deviates from the modal chromosome number (Chr. 2 = mode of 4; Chr. 3 = mode of 5). Scale bar = 5μm. (B) Percentage of nuclei that deviate from the modal chromosome number in RPE-1 or U2OS single cell clones grown in the presence of DMSO or UMK57. Chromosome copy number quantified for chromosomes 2 and 3 (n>300 nuclei). (C) Media with DMSO or UMK57 was prepared and added to U2OS cells at time 0. Cells were fixed and stained every 24hr and lagging chromosome rates quantified every 24hr. *P ≤ 0.05, **P ≤ 0.01 and n.s.= P > 0.05 using Fisher’s Exact two-tailed test. Media with DMSO, UMK57 or UMK57+Verapamil (MDR1 inhibitor) was prepared and added to cells at time 0. Cells were fixed and stained at <1h and 72h and lagging chromosome rates quantified for (D) U2OS cells (n>300) (E) HeLa cells (n>300) and (F) SW-620 cells (n>200). Total lagging chromosome rates were pooled from two independent experiments; *P ≤ 0.05, **P ≤ 0.01 and n.s.= P > 0.05 using Fisher’s exact two-tailed test. Figure 4 Rate of adaptive resistance in UMK57-treated cells is dependent on Aurora B activity (A) Calculated k-MT stability for U2OS cells treated with UMK57 for 72–120h, 6–8 weeks or for 6–8 weeks followed by UMK57 washout (72–120h post-washout). n=10–42 cells. Graph shows mean ± SEM. Statistical analysis performed between Control and UMK57-treated cells for each mitotic phase. *P ≤ 0.05 and **P ≤ 0.01 using two-tailed t-test. (B) Quantification of the relative levels of phospho-MCAK Ser95 at kinetochores/centromeres in U2OS cells treated with UMK57 for <1hr or >72h as quantified by immunofluorescence. (C) Representative image of U2OS cells in prometaphase and metaphase treated with DMSO or UMK57 for 72h, fixed and stained to reveal DNA (blue), phospho-Aurora B Thr232 (green) and ACA. Scale bar = 5μm. (D) Quantification of phospho-Aurora B Thr232 at kinetochores/centromeres in U2OS cells treated with UMK57 for <1hr or >72h as quantified by immunofluorescence. n ≥ 300 kinetochores from 10–20 cells for quantification of phospho-MCAK S95 and phospho-Aurora B Thr232. **P ≤ 0.01 and n.s.= P > 0.05 using two-tailed t-test. (E) Media with DMSO or ZM447439 + UMK57 was prepared and added to cells at time 0. Cells were fixed and stained at <1h, 72h, 96h and 120h and lagging chromosome rates quantified for each time point (n ≥ 150 anaphases). *P ≤ 0.05 and n.s.= P > 0.05 using Fisher’s exact two-tailed test. (F) Conceptual model for adaptive and acquired drug resistance in CIN cancer cells. Non-transformed diploid cells are non-adaptable and genetically homogeneous. Upon oncogenic transformation, the resulting tumor cells become both adaptable to stressful conditions and karyotypically heterogeneous through chromosomal instability. Adaptive resistance arises through the rapid and reversible changes in intracellular signaling networks within the genetically heterogeneous population. Acquired drug resistance occurs through the selection of rare, genetically unique cells from the population of heterogeneous cells. These mechanisms of resistance are not mutually exclusive. Author Contributions Conceptualization, B.O. and D.A.C.; Methodology, B.O. and D.A.C.; Formal Analysis, B.O.; Resources, L.T. and B.H.K.; Writing – Original Draft, B.O. and D.A.C.; Writing – Review & Editing, B.H.K., B.O. and D.A.C.; Visualization, B.O. and D.A.C.; Supervision, D.A.C.; Funding Acquisition, B.H.K. and D.A.C. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. Arrowsmith CH Audia JE Austin C Baell J Bennett J Blagg J Bountra C Brennan PE Brown PJ Bunnage ME Buser-Doepner C Campbell RM Carter AJ Cohen P Copeland RA Cravatt B Dahlin JL Dhanak D Edwards AM Frederiksen M Frye SV Gray N Grimshaw CE Hepworth D Howe T Huber KVM Jin J Knapp S Kotz JD Kruger RG Lowe D Mader MM Marsden B Mueller-Fahrnow A Müller S O’Hagan RC Overington JP Owen DR Rosenberg SH Ross R Roth B Schapira M Schreiber SL Shoichet B Sundström M Superti-Furga G Taunton J Toledo-Sherman L Walpole C Walters MA Willson TM Workman P Young RN Zuercher WJ 2015 The promise and peril of chemical probes Nature Chemical Biology 11 536 541 10.1038/nchembio.1867 26196764 Bakhoum SF Danilova OV Kaur P Levy NB Compton DA 2011 Chromosomal Instability Substantiates Poor Prognosis in Patients with Diffuse Large B-cell Lymphoma Clinical Cancer Research 17 7704 7711 10.1158/1078-0432.CCR-11-2049 22184286 Bakhoum SF Genovese G Compton DA 2009a Deviant Kinetochore Microtubule Dynamics Underlie Chromosomal Instability Current Biology 19 1937 1942 10.1016/j.cub.2009.09.055 19879145 Bakhoum SF Silkworth WT Nardi IK Nicholson JM Compton DA Cimini D 2014 The mitotic origin of chromosomal instability Current Biology 24 R148 R149 10.1016/j.cub.2014.01.019 24556433 Bakhoum SF Thompson SL Manning AL Compton DA 2009b Genome stability is ensured by temporal control of kinetochore–microtubule dynamics Nat Cell Biol 11 27 35 10.1038/ncb1809 19060894 Barretina J Caponigro G Stransky N Venkatesan K Margolin AA Kim S Wilson CJ Lehár J Kryukov GV Sonkin D Reddy A Liu M Murray L Berger MF Monahan JE Morais P Meltzer J Korejwa A Jané-Valbuena J Mapa FA Thibault J Bric-Furlong E Raman P Shipway A Engels IH Cheng J Yu GK Yu J Aspesi P de Silva M Jagtap K Jones MD Wang L Hatton C Palescandolo E Gupta S Mahan S Sougnez C Onofrio RC Liefeld T MacConaill L Winckler W Reich M Li N Mesirov JP Gabriel SB Getz G Ardlie K Chan V Myer VE Weber BL Porter J Warmuth M Finan P Harris JL Meyerson M Golub TR Morrissey MP Sellers WR Schlegel R Garraway LA 2012 The Cancer Cell Line Encyclopedia enables predictive modelling of anticancer drug sensitivity Nature 483 603 307 10.1038/nature11003 22460905 Chandarlapaty S 2012 Negative Feedback and Adaptive Resistance to the Targeted Therapy of Cancer Cancer Discovery 2 311 319 10.1158/2159-8290.CD-12-0018 22576208 Duesberg P Stindl R Hehlmann R 2000 Explaining the high mutation rates of cancer cells to drug and multidrug resistance by chromosome reassortments that are catalyzed by aneuploidy Proc Natl Acad Sci USA 97 14295 14300 10.1073/pnas.97.26.14295 11121035 Ertych N Stolz A Stenzinger A Weichert W Kaulfuß S Burfeind P Aigner A Wordeman L Bastians H 2014 Increased microtubule assembly rates influence chromosomal instability in colorectal cancer cells Nat Cell Biol 16 779 791 10.1038/ncb2994 24976383 Fernández L Breidenstein EBM Hancock REW 2011 Drug Resistance Updates Drug Resistance Updates 14 1 21 10.1016/j.drup.2011.01.001 21288762 Garnett MJ Edelman EJ Heidorn SJ Greenman CD Dastur A Lau KW Greninger P Thompson IR Luo X Soares J Liu Q Iorio F Surdez D Chen L Milano RJ Bignell GR Tam AT Davies H Stevenson JA Barthorpe S Lutz SR Kogera F Lawrence K McLaren-Douglas A Mitropoulos X Mironenko T Thi H Richardson L Zhou W Jewitt F Zhang T O’Brien P Boisvert JL Price S Hur W Yang W Deng X Butler A Choi HG Chang JW Baselga J Stamenkovic I Engelman JA Sharma SV Delattre O Saez-Rodriguez J Gray NS Settleman J Futreal PA Haber DA Stratton MR Ramaswamy S McDermott U Benes CH 2012 Systematic identification of genomic markers of drug sensitivity in cancer cells Nature 483 570 575 10.1038/nature11005 22460902 Geigl JB Obenauf AC Schwarzbraun T Speicher MR 2008 Defining “chromosomal instability Trends in Genetics 24 64 69 10.1016/j.tig.2007.11.006 18192061 Gerlinger M Swanton C 2010 How Darwinian models inform therapeutic failure initiated by clonal heterogeneity in cancer medicine British Journal of Cancer 103 1139 1143 10.1038/sj.bjc.6605912 20877357 Godek KM Kabeche L Compton DA 2014 Regulation of kinetochore–microtubule attachments through homeostatic control during mitosis Nat Rev Mol Cell Biol 1 8 10.1038/nrm3916 Gottesman MM 2002 Mechanisms of cancer drug resistance Annu Rev Med 53 615 627 10.1146/annurev.med.53.082901.103929 11818492 Greaves M Maley CC 2012 Clonal evolution in cancer Nature 481 306 313 10.1038/nature10762 22258609 Heppner GH 1984 Tumor heterogeneity Cancer Research 44 2259 2265 6372991 Kleyman M Kabeche L Compton DA 2014 STAG2 promotes error correction in mitosis by regulating kinetochore-microtubule attachments Journal of Cell Science 127 4225 4233 10.1242/jcs.151613 25074805 Lampson MA Kapoor TM 2004 The human mitotic checkpoint protein BubR1 regulates chromosome–spindle attachments Nat Cell Biol 7 93 98 10.1038/ncb1208 15592459 Lee AJX Endesfelder D Rowan AJ Walther A Birkbak NJ Futreal PA Downward J Szallasi Z Tomlinson IPM Howell M Kschischo M Swanton C 2011 Chromosomal Instability Confers Intrinsic Multidrug Resistance Cancer Research 71 1858 1870 10.1158/0008-5472.CAN-10-3604 21363922 Lengauer C Kinzler KW Vogelstein B 1997 Genetic instability in colorectal cancers Nature 386 623 627 10.1038/386623a0 9121588 Litvin O Schwartz S Wan Z Schild T Rocco M Oh NL Chen BJ Goddard N Pratilas C Pe’er D 2015 Interferon a/b Enhances the Cytotoxic Response of MEK Inhibition in Melanoma Molecular Cell 57 784 796 10.1016/j.molcel.2014.12.030 25684207 Luo J Solimini NL Elledge SJ 2009 Principles of Cancer Therapy: Oncogene and Non-oncogene Addiction Cell 136 823 837 10.1016/j.cell.2009.02.024 19269363 Muranen T Selfors LM Worster DT Iwanicki MP Song L Morales FC Gao S Mills GB Brugge JS 2012 Inhibition of PI3K/mTOR Leads to Adaptive Resistancein Matrix-Attached Cancer Cells Cancer Cell 21 227 239 10.1016/j.ccr.2011.12.024 22340595 Orr B Compton DA 2013 A double-edged sword: how oncogenes and tumor suppressor genes can contribute to chromosomal instability Front Oncol 3 164 10.3389/fonc.2013.00164 23825799 Pisco AO Brock A Zhou J Moor A Mojtahedi M Jackson D Huang S 2013 Non-Darwinian dynamics in therapy-induced cancer drug resistance Nature Communications 4 1 11 10.1038/ncomms3467 Sotillo R Schvartzman JM Socci ND Benezra R 2010 Mad2-induced chromosome instability leads to lung tumour relapse after oncogene withdrawal Nature 464 436 440 10.1038/nature08803 20173739 Sun C Wang L Huang S Heynen GJJE Prahallad A Robert C Haanen J Blank C Wesseling J Willems SM Zecchin D Hobor S Bajpe PK Lieftink C Mateus C Vagner S Grernrum W Hofland I Schlicker A Wessels LFA Beijersbergen RL Bardelli A Di Nicolantonio F Eggermont AMM Bernards R 2014 Reversible and adaptive resistance to BRAF(V600E) inhibition in melanoma Nature 508 118 122 10.1038/nature13121 24670642 Talje L Ben El Kadhi K Atchia K Tremblay-Boudreault T Carreno S Kwok BH 2014 DHTP is an allosteric inhibitor of the kinesin-13 family of microtubule depolymerases FEBS Letters 588 2315 2320 10.1016/j.febslet.2014.05.024 24859087 Thompson SL Compton DA 2011 Chromosome missegregation in human cells arises through specific types of kinetochore-microtubule attachment errors Proceedings of the National Academy of Sciences 108 17974 17978 10.1073/pnas.1109720108 Thompson SL Compton DA 2008 Examining the link between chromosomal instability and aneuploidy in human cells The Journal of Cell Biology 180 665 672 10.1083/jcb.200712029 18283116 Walker LA Gow NAR Munro CA 2010 Fungal Genetics and Biology Fungal Genetics and Biology 47 117 126 10.1016/j.fgb.2009.09.003 19770064 Zhai Y Kronebusch PJ Borisy GG 1995 Kinetochore microtubule dynamics and the metaphase-anaphase transition The Journal of Cell Biology 131 721 734 7593192

PMC005xxxxxx/PMC5130160.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 101689588 45568 Remediation (N Y) Remediation (N Y) Remediation (New York, N.Y.) 1051-5658 1520-6831 27917031 5130160 10.1002/rem.21458 NIHMS806519 Article A New Perspective on Sustainable Soil Remediation—Case Study Suggests Novel Fungal Genera Could Facilitate in situ Biodegradation of Hazardous Contaminants Czaplicki L.M. Ph.D. Candidate and Dean’s Graduate Fellow in the Department of Civil and Environmental Engineering at Duke University in Durham, North Carolina. Her doctoral thesis focuses on fungal bioremediation of high molecular weight polycyclic aromatic hydrocarbon contaminated soils. She received her M.S. from Duke University and her B.S. in Environmental Engineering from The Ohio State University Cooper E. Ph.D. research scientist and she manages the Duke Superfund Analytical Chemistry Core in Durham, North Carolina. Dr. Cooper is interested in analyzing environmentally important organic compounds in a variety of matrices including sediments, water, biological samples, and polyurethane foam. She received her Ph.D. in Environmental Sciences from Duke University. She earned her B.S in Plant Science and her M.S. in Plant and Soil Sciences from the University of Delaware Ferguson P.L. Ph.D. an associate professor of Environmental Chemistry and Engineering in the Department of Civil and Environmental Engineering and the Nicholas School of the Environment at Duke University in Durham, North Carolina. His research focuses on developing new methods for trace analysis of organic and nanoparticulate contaminants in the aquatic environment. Dr. Ferguson received his Ph.D. from the State University of New York at Stony Brook in Coastal Oceanography. He received his B.S. in Marine Science and Chemistry from the University of South Carolina Stapleton H.M. Ph.D. an associate professor in the Nicholas School of the Environment. Her research increases the understanding of the fate and transformation of organic contaminants in aquatic systems and indoor environments. Dr. Stapleton received her Ph.D. and M.S. from the University of Maryland, and her B.S. from Long Island University Southampton College Vilgalys R. Ph.D. professor in the Department of Biology and adjunct professor in the Department of Molecular Genetics and Microbiology at Duke University in Durham, North Carolina. His research focuses on fungal evolution, genetics and systematics. Dr. Vilgalys received his Ph.D. in Botany from Virginia Polytechnic Institute and State University. He received his M.S. in Botany from Virginia Tech and his B.A. in Biology from the State University of New York College at Genesco Gunsch C.K. Ph.D. an associate professor in the Department of Civil and Environmental Engineering at Duke University in Durham, North Carolina. Her research focuses on characterizing and engineering environmental microbiomes. Dr. Gunsch received her Ph.D. in Civil Engineering from the University of Texas at Austin. She received her M.S. in Environmental Engineering and Science from Clemson University and her B.S. in Civil Engineering from Purdue University lmc58@duke.edu 6 8 2016 2 3 2016 Spring 2016 01 4 2017 26 2 5972 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. Deciding upon a cost effective and sustainable method to address soil pollution is a challenge for many remedial project managers. High pressure to quickly achieve cleanup goals pushes for energy-intensive remedies that rapidly address the contaminants of concern with established technologies, often leaving little room for research and development especially for slower treatment technologies, such as bioremediation, for the more heavily polluted sites. In the present case study, new genomic approaches have been leveraged to assess fungal biostimulation potential in soils polluted with particularly persistent hydrophobic contaminants. This new approach provides insights into the genetic functions available at a given site in a way never before possible. In particular, this article presents a case study where next generation sequencing (NGS) has been used to categorize fungi in soils from the Atlantic Wood Industries Superfund site in Portsmouth, Virginia. Data suggest that original attempts to harness fungi for bioremediation may have focused on fungal genera poorly suited to survive under heavily polluted site conditions, and that more targeted approaches relying on native indigenous fungi which are better equipped to survive under site specific conditions may be more appropriate. INTRODUCTION Sites polluted with hydrophobic contaminants, such as heavy polycyclic aromatic hydrocarbons (PAHs), are challenging to remedy sustainably because these contaminants sorb strongly onto soils. The location of the present case study, Atlantic Wood Industries Superfund site (AWI) has PAHs as its main contaminants of concern. In particular, high molecular weight PAHs are of concern because they are highly recalcitrant in soil and contain known carcinogenic constituents such as benzo[a]pyrene, indeno(1,2,3-c,d)pyrene, and benz[a]anthracene. Similar to other sites, attempts to minimize risk at AWI have focused on excavation and in situ stabilization because remediation options that are both sustainable and degrade the contaminants of interest are limited. Thus, it is clear that more sustainable remediation options are needed as excavation and in situ stabilization are expensive, draw heavily on resources, and have a large carbon footprint. Bioremediation is a sustainable alternative to physico-chemical treatment which has been relied upon since the late 1980s (Sloan, 1987). Inherently sustainable because it harnesses natural biological processes to degrade pollutants, bioremediation has grown more favorable in recent years since awareness has spread about the need to minimize remedy carbon footprint (Forum, 2009). Fungal degradation processes have been of interest for over 30 years since they were discovered to cause decay in wooden utility poles despite their creosote treatment (Morrell and Zabel, 1985). The ability of fungi to grow and thrive in environments with mixtures of the toxic, hydrophobic PAHs in creosote motivated further interest into the metabolism of these fungi as a possible remedy for contaminated soil treatment. A series of studies on white rot fungi showed that fungal metabolism relied on extracellular enzymes secreted outside the cell. These extracellular enzymes have evolved to be highly nonspecific in order to break down the different bonds holding lignin and cellulose together in wood, but these same enzymes have been shown to also degrade other contaminants with similar chemical structures including creosote, pesticides, munitions, and chlorinated solvents (Cameron et al., 2000). Further studies showed that fungal processes could improve soil structure and that fungi may be better suited to challenging site conditions including, for example, lower nutrient and pH conditions (Miles and Chang, 1997). Furthermore, site conditions where nutrients are unevenly distributed or scarce may be better suited for fungi than bacteria because fungi have strategies to face those scenarios such as forming mycelia that conduct oxygen in their interior, requiring less nitrogen, and sporulating in times of extreme scarcity (Dowson et al., 1988; Tuisel et al., 1990; Mancera-López et al., 2008; Leitao et al., 2011). In addition, more inclusive studies on microbes showed degradation synergies between fungi and bacterial degraders where fungi initiated degradation and then bacteria carried it forward (Lade et al., 2012). Fungi with hydrophilic filaments have been seen to conduct bacterial degraders throughout the network, aiding in dispersal and resulting in better biodegradation than in the absence of their filaments (Kohlmeier et al., 2005; Warmink et al., 2011). Thus, involving fungal processes in a bioremediation strategy have the potential to lead to sustainable treatment of otherwise recalcitrant pollutants in soils. All of these studies motivated the development of fungal bioaugmentation strategies to introduce these well-known fungi to soils polluted with recalcitrant compounds wherein the fungi would be added along with materials to help them colonize the soil. However, when strategies using this field-scale fungal bioaugmentation approach proved disappointing, remedies involving fungi dropped in popularity. Bioremediation failures were difficult to directly study in the time that preceded the genomics era because one could not assess the soils for full fungal or bacterial colonization. However, it was thought that the native microbial communities might have had some role to play in these failures. To answer questions about the role of native microbiota, fungi were examined in sterile and nonsterile soils. Fungi were seen to exhibit different behavior in sterile soils as compared to nonsterile soils and it was concluded that the fungi that failed did not have adequate ability to compete with the native microbes for a space in the soil’s ecology (Andersson et al., 2000). It became clear that a new strategy involving the site’s native fungi in a more active role was needed (Harms et al., 2011). In this article, the authors present a discussion on fungal biostimulation, propose a framework wherein native fungi are centrally involved in remediation, and provide a case study wherein the genomic framework method was applied. Background A strategy focusing on a site’s indigenous fungi must first categorize the site’s fungi and assess them for bioremediation potential before fungal biostimulation can proceed. In a worst case scenario, conditions may exist where some sites are too heavily polluted to allow fungi to survive. Recent advances employing next generation sequencing (NGS) have revolutionized the study of microbiomes and made it possible to scan the entire soil’s DNA to provide information about: 1) the identity of microbes present at a given site; 2) the genetic capacities the microbes carry; and, 3) the specific microbial functions which are actively being used. These three pieces of information about a site’s native microbes can provide a more thorough explanation of the site microbial ecology. In the past, NGS has been leveraged in water pollution scenarios to explain why microbial mats remove pollutants from wastewater (Akyon et al., 2015). NGS has also been used to explain why specific bioremediation strategies failed, by tracking microbial function over a specific timescale. In previous studies carried out prior to NGS advances, signs pointed to the introduced fungi not being competitive enough in their new environment. Here, the framework proposed leverages NGS to examine biostimulation strategies focusing on already established indigenous fungi as opposed to bioaugmentation (i.e., addition of exogenous fungi to the soils). The likelihood is high that fungi found at the site beyond the classic white rot fungi will have adapted an ability to degrade the site pollutants. In the time since early interest in white rot fungi, many other fungi have been identified which are capable of degrading a range of pollutants, raising questions about the degree of shared degradation ability in fungi. The kingdom Fungi is broken down into phyla that are further divided into classes, orders, families, genera, and species. The fungi that garnered much attention in the 1980s (i.e., white rot fungi) belong to the phylum Basidiomycota. However, Harms et al later summarized findings from studies on fungi inhabiting different areas of the environment and discussed how the degradation capacity spread well beyond Basidiomycota into Ascomycota, Glomeromycota, Chytridiomycota, Mucoromycotina, and other phyla. In particular, many common fungi have been identified in soils which belong to the Ascomycota phylum. Using culture-based techniques, fungal species capable of degrading soil contaminants have been identified and further characterized. For example, the ascomycete Lasiodiplodia theobromae was cultured from a soil contaminated by the Beijing Coking Plant in China that could degrade benzo[a]pyrene, pyrene, and phenanthrene using its laccases and lignin peroxidases (Wang et al., 2014). Trichoderma asperellum, another member of the Ascomycota phylum, was cultured from a heavy crude oil-contaminated soil and found to degrade benzo[a]pyrene, pyrene, and phenanthrene using laccases and peroxidases (Zafra et al., 2015). Ye and coauthors cultured yet another fungus within the Ascomycota phylum, Aspergillus fumigatus, that could degrade anthracene present in contaminated soil near a gas station (Ye et al., 2011). Although degradation mechanisms have not been studied extensively throughout the kingdom Fungi, it is clear that various fungi have evolved different enzyme systems to suit specific niche environments. Even within the wood rotting fungi, a variety of enzymes are used by each genus in degrading pollutants (Barr and Aust, 1994; Rivera-Hoyos et al., 2013). Mechanisms of action resulting in contaminant degradation range from enzyme attack and resulting oxidation outside the cell to uptake of the contaminant into the cell where it becomes accessible to intracellular enzymes for degradation (Barr and Aust, 1994). Current literature suggests that various strategies will be used to degrade contaminants in different fungi, though it is generally accepted that degradation ability is common throughout the kingdom Fungi. Purpose To some, it may seem that categorizing indigenous fungi does not contribute much to the ultimate remedy because the different strains may use different enzymes. However, it should be noted that the value in categorizing indigenous fungi lies in its power to inform meaningful fungal biostimulation. Despite the different enzyme systems that white rot fungi use to degrade pollutants, all their enzymes can be stimulated by wood substrate addition because it is known that they thrive in the presence of wood. Thus, site amendments can be recommended to stimulate site-specific fungal growth and overall contaminant degradation. For example, fungi that have evolved chitin-degrading enzymes may be stimulated more by chitin additions than by wood additions. The same could be said about fungi that have evolved closely with plants, which may benefit more from xylan-rich substrates. Furthermore, species-level differences in metabolism may be less important if site fungi can co-metabolically degrade the contaminant in response to stimulation by the same substrate. As we continue to develop a better understanding of the enzyme systems associated with specific groups of fungi and the substrates which stimulate those enzymes, fungal biostimulation strategies can be devised. However, the first step towards this goal is identifying and characterizing the indigenous fungal population at a given site using NGS. MATERIALS AND METHODS In this article, we present the result of efforts to categorize fungi using NGS applied to severely contaminated soil in order to inform biostimulation strategies as shown in Exhibit 1. In contrast to the fungal bioaugmentation approach of years past, this framework centers around fungal biostimulation. The framework in Exhibit 1 answers the question of whether there are fungi inhabiting the soils by collecting samples from different levels of pollution then leveraging NGS to identify and categorize fungi already established in the soils. Results from the NGS effort feed into the second step wherein the list of categorized fungi inhabiting each contaminated soil is cross-referenced with fungal literature where fungi with demonstrated bioremediation ability have been characterized. Because the NGS effort results in a list of closest-related fungal categories down to the genus, a further literature review and laboratory studies can then be used to determine if degradation is feasible either as a metabolic or cometabolic process. This stage requires verification of degradation using a laboratory-scale bioreactor. Finally, the results can be used to inform a biostimulation strategy that encourages growth in situ based on the types of indigenous fungi present. NGS Efforts Samples were collected from the Atlantic Woods Industries Superfund Site in Portsmouth, Virginia and subjected to DNA extraction for NGS analysis. To this end, first, Mo BIO’s PowerSoil Powerlyzer DNA extraction kit was used to extract DNA from the samples (MO BIO Laboratories, Inc., Carlsbad, California). The suggested protocol was followed with two modifications. The recommended mass to start with was increased to 0.3-0.4 grams (g). Also, in order to get a more thorough separation between the PAH and the DNA, between 400-450 microliters (μL) of phenol-chloroform was added before the bead beating step. Then bead beating lysis was performed on a minibeadbeater (Biospec Products, Bartlesville, Oklahoma) for 20 seconds. Next, polymerase chain reaction (PCR) was used to amplify genes informative for fungal characterization. The ribosomal large sub-unit (LSU) was amplified using primers LR3 and LR0R linked with adapter sequences for Illumina MiSeq (Ilumina, Inc., San Diego, California) (Lundberg et al., 2013). DNA was subjected to 25-30 cycles of LSU amplification with these primers depending on the sample and how difficult amplification proved. The PCR protocol involved initial denaturation at 95 °C for 10 minutes followed by 25-30 cycles of 1 minute at 95 °C, 1 minute at 55.3 °C, and 90 seconds at 72 °C, with a final extension of 10 minutes at 72 °C. Ribsomal amplicon sequencing was performed on the Illumina MiSeq platform, 250 bp paired-end sequencing. The resulting sequences were processed using CutAdapt, USEARCH, and QIIME (Bittinger et al., 2010; Martin, 2012; Edgar and Flyvbjerg, 2015). Once the sequences were processed for length and quality, fungi were categorized into their respective phyla, classes, orders, families, and genera using the Ribosomal Database Project’s Classifier LSU training set 11 (Wang et al., 2007). Once fungi were identified, they were cross-referenced with relevant literature to highlight fungal genera that contain species shown to be involved in degradation. These were termed “candidate” genera. Chemical Analysis A total of 34 PAHs were analyzed in the same samples used for the NGS analysis using ultrasonication extraction in hexane:acetone (1:1) followed by silica gel cleanup then measured via gas chromatography combined with electron impact mass spectrometry based on a method published in Clark et al. (Clark et al., 2013). CASE STUDY-ATLANTIC WOOD INDUSTRIES SUPERFUND SITE Background To validate the framework presented in Exhibit 1, the described method was applied to several samples collected from AWI. The main contaminants at AWI are PAHs that had originated from historic operations and improper disposal of creosote when AWI operated a creosote wood treatment plant on the site. AWI is composed of approximately 48 acres of industrialized waterfront which extends into the sediments of the Elizabeth River. The sediments in this area have PAH concentrations which are orders of magnitude greater than background levels (Clark et al., 2013). In fact, this waterfront property has some of the highest PAH concentrations in the world (Di Giulio and Clark, 2015). Soils with high contamination were excavated temporarily on site before they were contained fully as part of the selected remedy. For the present study, approximately 50 g was collected from the surface of four excavated soil and wood chip piles generated during the excavation. These piles were produced during an excavation of a sewer junction which was conveying site contaminants to the Elizabeth River and, thus, these samples were expected to have high PAH concentrations. Creosote was visible in the bottom of the pit post-excavation and the piles appeared to have a range of creosote concentrations based on visual inspection. In total, samples were collected from three soils of different visible creosote pollutant levels, and one woodchip pile, consisting of woodchips roughly 100 mm x 50 mm x 20 mm in size, which had less visible creosote but still caused glove-staining upon collection, suggesting significant creosote contamination. In addition, clean soil was sampled from an uncontaminated area near the site as a control. Samples were kept at 4 °C until DNA extraction. This proof of concept case study measured fungal communities in the clean soil as well as the woodchip and three differentially polluted excavated soil piles. Results Exhibit 2 shows the average PAH concentration measured at each site. The average consists of 34 PAHs that range in molecular weight from the two-ringed naphthalene to the six-ringed dibenzo(a,l)pyrene. The full list of PAHs includes: 1-methylnapthalene, 2,6-dimethylnapthalene, naphthalene, acenapthene, acenapthylene, carbazole, dibenzofuran, dibenzothiophene, fluorene, 1-methylphenanthrene, 2-methylphenanthrene, anthracene, phenanthrene, retene, 1,2-benzofluorene, 3,4-benzofluorene, fluoranthene, 1,2-benzanthracene, benzo(c)phenanthrene, chrysene, pyrene, 3-methylcholanthrene, benzo(a)fluoranthene, benzo(b,k,f)fluoranthrene, benzo(a)pyrene, benzo(b)chrysene, benzo(e)pyrene, dibenzo(a,j)anthracene, dibenzo(a,h)anthracene, perylene, picene, indeno(1,2,3,c,d)pyrene, benzo(g,h,i)perylene, and dibenzo(a,l)pyrene. The lowest contaminated soil (Sample 1), had a total PAH concentration of 180 ± 0.023 micrograms per gram (μg/g; Exhibit 2). The total PAH concentration in Sample 2 was 974 ± 0.074 μg/g. Sample 3, the woodchip deposit, had a total PAH concentration of 5,854 ± 1.473 μg/g. The most contaminated soil was Sample 4 with a total PAH concentration of 18,407 ± 0.825 μg/g. Using NGS, we detected 99 different taxa from the media sampled. The phylum-level classification is shown in Exhibit 3. It should be noted that because these data focus on DNA, they only provide a snapshot of the fungi present and not their activity. As may be expected, there is relatively high diversity in the control soil and Sample 1 (the least contaminated soil). The control soil is dominated by members of Ascomycota (shown in dark grey), with members of Basidiomycota (in diagonal stripes) second most prevalent. Ascomycota are also seen here to dominate the fungal community in Samples 2 and 3. Sample 1 is split equally between Ascomycota and unclassified fungi (in light grey) with a smaller amount of Chytridiomycota (in white) and Basidiomycota. The fungi from Sample 4 fall mostly within Basidiomycota with secondary prevalence of Ascomycota. The phylum-level grouping is informative because it suggests that fungal biostimulation strategies which target members of Basidiomycota may not match up with the most prevalent members of the fungal community, at least based on the fungi detected in AWI’s contaminated media sampled. In fact, this analysis suggests that Ascomycota may be a better phylum to target with a biostimulation strategy in the majority of these samples. Exhibit 4 shows how the fungal genera grouped into different groups of “candidates” warranting further attention varied across the contaminated media. This exhibit describes the results of grouping fungal genera together based on a literature review of species demonstrating degradation capacity. The literature review behind Exhibit 4 is of importance because it identifies conditions that produced the results shown, conditions that can inform biostimulation strategies. It is important to note that the fungi not classified as candidates have not yet appeared in bioremediation or functional studies and, thus, have not yet been characterized in terms of bioremediation capacity. Thus, the proportion of these does not imply that they cannot degrade contaminants, but rather that they have not been looked at through a bioremediation lens yet. These data show that the candidate genera fraction out of the total population range from a minimum of 20 percent in Sample 1 to a maximum of 100 percent in Sample 2. In the more polluted samples, Samples 3 and 4, approximately 60 and 30 percent of the total fungal genera were found to be candidates, respectively. These results were somewhat surprising, as it may be unexpected to find any microorganisms at the extremely high PAH concentrations found in Samples 3 and 4 which had contamination levels on the order of milligrams per gram (mg/g). Some fungal genera detected in AWI samples were assumed to have degradation capacity because they contain species known as agents of PAH degradation. These groups are represented in black in Exhibit 4 (Ravelet et al., 2000; Potin et al., 2004; Cerniglia and Sutherland, 2010; Wang et al., 2014; Zafra et al., 2015). This fraction only represented 5 percent of the control soil which is expected because of the low PAH concentration. The fraction increased to approximately 15 percent of Sample 1, and then reached nearly 70 percent of Sample 2’s fungal community. Again, these results are encouraging as they suggest that the presence of the contaminants enriches for potential fungal genera capable of breaking down the contaminants. Sample 3 also had many fungal genera containing agents of PAH degradation, at just over 40 percent of the total fungal community. In Sample 4, this fraction returned to around 5 percent. Based on these data, it either appears that the known PAH degrading fungal genera are incapable of surviving beyond a certain threshold, or that there is the possibility that previously uncharacterized fungi outcompete at the higher concentration. Other fungal species previously shown to degrade contaminants other than PAHs were also detected at AWI (Daechul and Hyun, 2008; Cosgrove et al., 2010; de Oliveira et al., 2013; de Souza Pereira Silva et al., 2015). This suggests that these genera may also be able to degrade PAHs since they were enriched under high contamination concentrations. The fraction of non-PAH degrading fungi is shown in dark grey in Exhibit 4 and comprises 35 percent of the control soil and 5 percent of the lesser polluted Sample 1. Sample 2 contained a larger fraction (~25 percent of the total) while Sample 3 had less than 5 percent. The majority of the candidate genera within Sample 4 were in this category, however, at 20 percent of the total. This analysis suggests that the fungi shown to be agents of non-PAH degradation should be examined for PAH degradation, since they are so prevalent in the presence of high PAH concentration. Because fungal enzymes have been shown to have substrate promiscuity, enzyme production was also used as a proxy for characterizing likely degradation capacity. The analysis behind Exhibit 4 revealed fungi capable of producing such enzymes. Fungal genera producing polyphenol oxidases and laccases, which are known to degrade PAHs, are shown in light grey in the exhibit, while enzymes of lesser degradation certainty are shown in white with a black outline (Cameron et al., 2000). The fraction of fungal genera that produce extracellular enzymes known to degrade PAHs is low across all the sampled communities, comprising less than 5 percent of all samples and the control, except for Sample 4, which had about 10 percent (Verdin et al., 2004; Yang et al., 2005; Qasemian et al., 2012; Barbi et al., 2014; Zafra et al., 2015). Fungal genera that have been shown to contain species producing enzymes yet unstudied in a bioremediation context are only present in the control soil and Sample 3, at 10 percent each (Fenice et al., 1997; Bojsen et al., 1999; Kang et al., 2004; Yu et al., 2004; Zhao et al., 2013). This analysis suggests that biodegradation studies may need to target other enzymes with known substrate promiscuity besides lignin peroxidase, manganese peroxidase, and laccase. CONCLUSION This case study suggests that fungi are present and that biostimulation may be possible in highly contaminated soils. Here, NGS was shown to be a tool capable of tremendous insight into the contaminated soil ecology. Specifically, NGS was demonstrated to be useful as a first pass identification tool to characterize fungi present in polluted soils. NGS data revealed that the polluted media was suitable for more than Basidiomycota, and that the soils can be comprised entirely of fungal genera with observed degradation capacity. It was surprising that a high prevalence of the well-studied Basidiomycota was not reflected in all of the contaminated media sampled. Rather, the fungal communities were more composed of a blend of Ascomycota and Basidiomycota. This case study suggests that existing strategies developed to stimulate members of Basidiomycota may be insufficient for remediation sites such as AWI. There is a clear need to also develop ascomycete targeting strategies to take advantage of established fungi that have shown to be capable of degradation. Ultimately, the case study provides evidence to suggest usual fungal bioremediation strategies may not target the soil fungi that have established themselves in polluted soils. Moving forward, it would serve the remediation community to use NGS more often. As the proposed framework becomes applied to more polluted sites, more fungi can be identified that can survive under immense environmental stress. This will improve remedy sustainability by harnessing fungi more often in sites with recalcitrant soil pollutants. Future work should also incorporate a functional level analysis (RNA) to infer what functions are possible under the site’s specific conditions. ACKNOWLEDGMENTS We wish to thank the High Throughput Sequencing Facility at the University of North Carolina-Chapel Hill for the sequencing involved in this work. Additionally, we would like to thank Marc Gutterman, Randy Sturgeon, and Joe Alfano, site personnel at Atlantic Wood Industries Superfund site for help with logistics and site access. Also, we wish to thank Lauren Redfern, Dan Brown, Jordan Kozal, Nishad Jayasandara, Bryan Clark, and Josh Osterberg for helping collect samples in the field. Katherine Davis should also be recognized for contributions to chemical analyses involved in this work. Funding supported this research through the National Institute of Environmental Health Sciences’ Superfund Research Program grant P42-ES010356. Exhibit 1 Proposed fungal bioremediation framework Exhibit 2 Total PAH concentration measured in Atlantic Wood Industries contaminated media; numbers shown represent the average concentration +/− one standard deviation (n=6) Exhibit 3 Fungi inhabiting control and polluted media categorized at the phylum-level Exhibit 4 Fungi categorized at the genus level and grouped by type of potential demonstrated REFERENCES Akyon B Stachler E Wei N Bibby K Microbial mats as a biological treatment approach for saline wastewaters: The case of produced water from hydraulic fracturing Environmental Science & Technology 2015 49 10 6172 6180 25867284 Andersson BE Welinder L Olsson PA Olsson S Henrysson T Growth of inoculated white-rot fungi and their interactions with the bacterial community in soil contaminated with polycyclic aromatic hydrocarbons, as measured by phospholipid fatty acids Bioresource Technology 2000 73 1 29 36 Barbi F Bragalini C Vallon L Prudent E Dubost A Fraissinet-Tachet L Marmeisse R Luis P PCR primers to study the diversity of expressed fungal genes encoding lignocellulolytic enzymes in soils using high-throughput sequencing PLoS One 2014 9 12 Barr DP Aust SD Mechanisms white rot fungi use to degrade pollutants Environmental Science & Technology 1994 28 2 78A 87A Bittinger K Bushman FD Caporaso JG Costello EK Fierer N Goodrich JK Gordon JI Huttley GA Kelley ST Knight R Knights D Koenig JE Kuczynski J Ley RE Lozupone CA McDonald D Muegge BD Pena AG Pirrung M Reeder J Sevinsky JR Stombaugh J Turnbaugh PJ Walters WA Widmann J Yatsunenko T Zaneveld J QIIME allows analysis of high-throughput community sequencing data Nature Methods 2010 7 335 336 20383131 Bojsen K Yu SK Marcussen J A group of alpha-1,4-glucan lyase genes from the fungi Morchella costata, M-vulgaris and Peziza ostracoderma. Cloning, complete sequencing and heterologous expression Plant Molecular Biology 1999 40 3 445 454 10437828 Cameron MD Timofeevski S Aust SD Enzymology of Phanerochaete chrysosporium with respect to the degradation of recalcitrant compounds and xenobiotics Applied Microbiology & Biotechnology 2000 54 6 751 758 11152065 Cerniglia CE Sutherland JB Timmis K Degradation of polycyclic aromatic hydrocarbons by fungi Handbook of Hydrocarbon and Lipid Microbiology 2010 2079 2110 Springer Berlin, Heidelberg, Germany Clark BW Cooper EM Stapleton HM Di Giulio RT Compound- and mixture-specific differences in resistance to polycyclic aromatic hydrocarbons and PCB-126 among Fundulus heteroclitus subpopulations throughout the Elizabeth River Estuary (Virginia, USA) Environmental Science & Technology 2013 47 18 10556 10566 24003986 Cosgrove L McGeechan PL Handley PS Robson GD Effect of biostimulation and bioaugmentation on degradation of polyurethane buried in soil Applied and Environmental Microbiology 2010 76 3 810 819 19948849 Daechul C Hyun KS Comparison of cyanide degrading enzymes expressed from genes of fungal origin Journal of Environmental Science International 2008 17 11 1221 1226 de Oliveira JR Mizuno CM Seleghim MHR Javaroti DCD Rezende MOO Landgraf MD Sette LD Porto ALM Biotransformation of phenylacetonitrile to 2-hydroxyphenylacetic acid by marine fungi Marine Biotechnology 2013 15 1 97 103 22790719 de Souza Pereira Silva D de Lima Cavalcanti D de Melo EJV dos Santos PNF da Luz ELP de Gusmão NB de Queiroz Sousa M Bio-removal of diesel oil through a microbial consortium isolated from a polluted environment International Biodeterioration & Biodegradation 2015 97 85 89 Di Giulio RT Clark BW The Elizabeth River story: A case study in evolutionary toxicology. Journal of Toxicology and Environmental Health Part B, Critical reviews 2015 18 6 259 298 Dowson CG Rayner ADM Boddy L Foraging patterns of Phallus impudicus, Phanerochaete laevis and Steccherinum fimbriatum between discontinuous resource units in soil FEMS Microbiology Letters 1988 53 5 291 298 Edgar RC Flyvbjerg H Error filtering, pair assembly and error correction for next-generation sequencing reads Bioinformatics 2015 DOI: 10.1093/bioinformatics/btv401 Fenice M Selbmann L Zucconi L Onofri S Production of extracellular enzymes by Antarctic fungal strains Polar Biology 1997 17 3 275 280 Harms H Schlosser D Wick LY Untapped potential: Exploiting fungi in bioremediation of hazardous chemicals Nature Reviews Microbiology 2011 9 3 177 192 21297669 Kang Z Huang L Buchenauer H Ultrastructural and cytochemical studies on infection of wheat spikes by Microdochium nivale Zeitschrift Fur Pflanzenkrankheiten Und Pflanzenschutz-Journal of Plant Diseases and Protection 2004 111 4 351 361 Kohlmeier S Smits TH Ford RM Keel C Harms H Wick LY Taking the fungal highway: mobilization of pollutant-degrading bacteria by fungi Environmental Science & Technology 2005 39 12 4640 4646 16047804 Lade HS Waghmode TR Kadam AA Govindwar SP Enhanced biodegradation and detoxification of disperse azo dye Rubine GFL and textile industry effluent by defined fungal-bacterial consortium International Biodeterioration & Biodegradation 2012 72 94 107 Leitao AL Garcia-Estrada C Ullan RV Guedes SF Martin-Jimenez P Mendes B Martin JF Penicillium chrysogenum var. halophenolicum, A new halotolerant strain with potential in the remediation of aromatic compounds in high salt environments Microbiological Research 2011 167 2 79 89 21524896 Lundberg DS Yourstone S Mieczkowski P Jones CD Dangl JL Practical innovations for high-throughput amplicon sequencing Nature Methods 2013 10 999 1001 23995388 Mancera-López ME Esparza-García F Chávez-Gómez B Rodríguez-Vázquez R Saucedo-Castañeda G Barrera-Cortés J Bioremediation of an aged hydrocarbon-contaminated soil by a combined system of biostimulation–bioaugmentation with filamentous fungi International Biodeterioration & Biodegradation 2008 61 2 151 160 Martin M Cutadapt removes adapter sequences from high-throughput sequencing reads Bioinformatics in Action 2012 17 1 10 12 Miles PG Chang S-T Mushroom biology: Concise basics and current developments 1997 World Scientific Publishing Co. Pte. Ltd. Singapore Morrell JJ Zabel RA Wood strength and weight losses caused by soft rot fungi isolated from treated southern pine utility poles Wood and Fiber Science 1985 17 1 132 143 Potin O Veignie E Rafin C Biodegradation of polycyclic aromatic hydrocarbons (PAHs) by Cladosporium sphaerospermum isolated from an aged PAH contaminated soil FEMS Microbiology Ecology 2004 51 1 71 78 16329856 Qasemian L Billette C Guiral D Alazard E Moinard M Farnet A-M Halotolerant laccases from Chaetomium sp.,Xylogone sphaerospora, and Coprinopsis sp. isolated from a Mediterranean coastal area Fungal Biology 2012 116 10 1090 1098 23063188 Ravelet C Krivobok S Sage L Steiman R Biodegradation of pyrene by sediment fungi Chemosphere 2000 40 5 557 563 10665394 Rivera-Hoyos CM Morales-Álvarez ED Poutou-Piñales RA Pedroza-Rodríguez AM Rodríguez-Vázquez R Delgado-Boada JM Fungal laccases Fungal Biology Reviews 2013 27 3-4 67 82 Sloan R Bioremediation Demonstrated at Hazardous-Waste Site Oil & Gas Journal 1987 85 37 61 66 Sustainable Remediation Forum (SURF) Sustainable remediation white paper—Integrating sustainable principles, practices, and metrics into remediation projects Remediation Journal 2009 19 3 5 114 Tuisel H Sinclair R Bumpus JA Ashbaugh W Brock BJ Aust SD Lignin peroxidase H2 from Phanerochaete chrysosporium: purification, characterization and stability to temperature and pH Archive of Biochemistry and Biophysics 1990 279 1 158 166 Verdin A Sahraoui A Durand R Degradation of benzo[a]pyrene by mitosporic fungi and extracellular oxidative enzymes International Biodeterioration & Biodegradation 2004 53 2 65 70 L-H Wang C Liu H Li J Sun H Degradation of PAHs in soil by Lasiodiplodia theobromae and enhanced benzo[a]pyrene degradation by the addition of Tween-80 Environmental Science and Pollution Research 2014 21 18 10614 10625 24878554 Wang Q Garrity GM Tiedje JM Cole JR Naïve bayesian classifier for rapid assignment of rRNA sequences into the new bacterial taxonomy Applied and Environmental Microbiology 2007 73 16 5261 5267 17586664 Warmink JA Nazir R Corten B van Elsas JD Hitchhikers on the fungal highway: The helper effect for bacterial migration via fungal hyphae Soil Biology and Biochemistry 2011 43 4 760 765 Yang Q Yediler A Yang M Kettrup A Decolorization of an azo dye, Reactive Black 5 and MnP production by yeast isolate: Debaryomyces polymorphus Biochemical Engineering Journal 2005 24 3 249 253 Ye J-S Yin H Qiang J Peng H Qin H-M Zhang N He B-Y Biodegradation of anthracene by Aspergillus fumigatus Journal of Hazardous Materials 2011 185 1 174 181 20932640 Yu SK Refdahl C Lundt I Enzymatic description of the anhydrofructose pathway of glycogen degradation - I. Identification and purification of anhydrofructose dehydratase, ascopyrone tautomerase and alpha-1,4-glucan lyase in the fungus Anthracobia melaloma Biochimica Et Biophysica Acta-General Subjects 2004 1672 2 120 129 Zafra G Moreno-Montaño A Absalón Á Cortés-Espinosa D Degradation of polycyclic aromatic hydrocarbons in soil by a tolerant strain of Trichoderma asperellum Environmental Science and Pollution Research 2015 22 2 1034 1042 25106516 Zhao X-H Wang W Wei D-Z Identification of Petriella setifera LH and characterization of its crude carboxymethyl cellulase for application in denim biostoning Journal of Microbiology (Seoul, Korea) 2013 51 1 82 87

PMC005xxxxxx/PMC5130230.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 101139554 27040 J Chromatogr B Analyt Technol Biomed Life Sci J. Chromatogr. B Analyt. Technol. Biomed. Life Sci. Journal of chromatography. B, Analytical technologies in the biomedical and life sciences 1570-0232 1873-376X 27776327 5130230 10.1016/j.jchromb.2016.10.014 NIHMS824866 Article Component analysis and target cell-based neuroactivity screening of Panax ginseng by ultra-performance liquid chromatography coupled with quadrupole-time-of-flight mass spectrometry Yuan Jinbin ab* Chen Yang a Liang Jian c Wang Chong-Zhi b Liu Xiaofei a Yan Zhihong a Tang Yi a Li Jiankang a Yuan Chun-Su b a Key Laboratory of Modern Preparation of TCM, Ministry of Education, Jiangxi University of Traditional Chinese Medicine, Nanchang, 330004 China b Tang Center for Herbal Medicine Research, and Department of Anesthesia & Critical Care, The University of Chicago, Chicago, IL, 60637 USA c Research Center for the Resourcing of Traditional Chinese Medicine and Minority Medicine, Jiangxi University of Traditional Chinese Medicine, Nanchang, 33004 China * Corresponding author. kings2008@163.com; yuan001@uchicago.edu (J. Yuan) 28 10 2016 17 10 2016 1 12 2016 01 12 2017 1038 111 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. Ginseng is one of the most widely used natural medicines in the world. Recent studies have suggested Panax ginseng has a wide range of beneficial effects on aging, central nervous system disorders, and neurodegenerative diseases. However, knowledge about the specific bioactive components of ginseng is still limited. This work aimed to screen for the bioactive components in Panax ginseng that act against neurodegenerative diseases, using the target cell-based bioactivity screening method. Firstly, component analysis of Panax ginseng extracts was performed by UPLC-QTOF-MS, and a total of 54 compounds in white ginseng were characterized and identified according to the retention behaviors, accurate MW, MS characteristics, parent nucleus, aglycones, side chains, and literature data. Then target cell-based bioactivity screening method was developed to predict the candidate compounds in ginseng with SH-SY5Y cells. Four ginsenosides, Rg2, Rh1, Ro, and Rd, were observed to be active. The target cell-based bioactivity screening method coupled with UPLC-QTOF-MS technique has suitable sensitivity and it can be used as a screening tool for low content bioactive constituents in natural products. Graphical abstract Panax ginseng component analysis drug screening target cell extraction UPLC-QTOF-MS SH-SY5Y cell 1. Introduction Panax ginseng, the root of Panax ginseng C.A. Meyer (family Araliaceae), has been used as a traditional medicine for thousands of years in East Asian countries, and now is one of the most widely used natural medicines in the world. The chemical constituents in Panax ginseng include triterpene saponins, polysaccharides, peptides, polyacetylenic alcohols, phenolic compounds and fatty acids [1,2]. Ginsenosides, also called panaxosides or ginseng saponins, a type of triterpene glycosides, are the major active components in Panax ginseng. Recent studies have suggested Panax ginseng has a wide range of beneficial effects on aging, central nervous system disorders, and neurodegenerative diseases [2–5]. Panax ginseng extract [6], ginsenosides Rg1 [7–9], Rb1 [9,10], and Rd [11] have been found to protect SH-SY5Y cells against 1-methyl-4-phenylpyridinium-induced injury. In our preliminary experiments, ginseng extracts and the ginsenosides fractions were found to be capable of facilitating the proliferation of SH-SY5Y cells and protecting the cells against H2O2-induced injury. But what are the active constituents in the extracts and fractions? And are there any ginsenosides that can exert protective benefits other than Rg1, Rb1 and Rd? It is necessary to screen the active compounds in ginseng further. The classic screening procedures for bioactive components in natural products include isolation, purification, and then pharmacological evaluation. This conventional method is time-consuming, arduous, and has a low efficiency. Modern pharmacological investigation has shown that most drugs should enter the target cells or bind with some receptors, enzymes or channels on cell membranes to elicit activity [12,13]. A target cell-based screening method coupled with modern chromatographic technique has been developed using the principles of this theory [12,13]. In the screening procedures, the extracts are added into the target cells, the potential bioactive components may selectively bind to the cells, the unbound compounds can be washed away, and finally the active components released by digestion are analyzed by the hyphenated chromatography technique. Recently, the target cell-based screening method has been successfully applied to find the bioactive compounds in natural products [14–20]. Meanwhile, the analytical tools used in the screening method have undergone evolution from common liquid chromatography (LC) [16], LC-MS [14,15,20], to LC-MS/MS with low energy collision dissociation (CID) [18], especially UPLC-QTOF-MS (Ultra-high-performance liquid chromatography coupled with Quadrupole-time-of-flight mass spectrometry) [17]. The UPLC-QTOF-MS technique possesses high resolution, high sensitivity, high mass accuracy, and abundant fragment information, and has become a powerful analytical tool for complicated samples. This work aims to screen for the bioactive components in Panax ginseng that act against neurodegenerative disease with the target cell-based bioactivity screening method. SH-SY5Y human neuroblastoma cells have been widely used as in vitro models of neuronal function, differentiation, and neurodegenerative diseases. Thus, SH-SY5Y cells were selected as the target cells. Firstly, a component analysis of Panax ginseng extracts was performed by UPLC-QTOF-MS, and 54 compounds were identified. Secondly, target cell extraction was carried out, and the extracted compounds were analyzed by UPLC-QTOF-MS. Four potential neural active components in ginseng were identified in this work. 2. Experiment 2.1. Herbal materials and chemicals Panax ginseng was collected from Tonghua County, Jilin Province, China, and was authenticated by associate professor Yuye Zhu from Jiangxi University of Traditional Chinese Medicine (JXUTCM). A voucher specimen was preserved in the Key Lab of Modern Preparation of TCM, JXUTCM, Nanchang, China. HPLC grade acetonitrile (ACN) was from Fisher (USA). Purified water was obtained by a Milli-Q system (Millipore, Bedford, MA, USA). Other reagents and chemicals were of analytical grade. SH-SY5Y cells were obtained from the China Center for Type Culture Collection in Beijing University. Dulbecco’s Modified Eagle Medium (DMEM, LOT: 1471272), Ham’s F12 Nutrient Mixture (F12, LOT: 21127022), fetal bovine serum (FBS, LOT: 1414426), trypsin (LOT: 25200056) were purchased from Invitrogen Gibco (Carlsbad, CA, USA). 2.2. Sample preparation Ginseng samples were pulverized into powder. 60 grams of the powder were extracted twice by heat-reflux with 70% ethanol (600 mL, 2h; 480 mL, 1h). The combined extract was evaporated under vacuum and lyophilized. Two lyophilized samples were accurately weighed. One was re-diluted in 70% ethanol, and then passed through a 0.22 µm filter prior to analysis by UPLC-QTOF-MS. The other was dissolved in DMEM under a sterile environment. The final concentration of ginseng extract was 5 mg/mL, which was used for cell culture. 2.3. Cell culture SH-SY5Y cells were seeded into cell culture flasks and were incubated in DMEM/F12 (1:1, v/v) medium supplemented with 10% FBS, 100 U/mL penicillin, and 100 µg/mL streptomycin. They were maintained in a humidified 5% CO2 incubator (SANYO MCO-175, Osaka, Japan) at 37 °C. The medium was replaced at 24 hour intervals. The cells were subcultured (1:3) until 80% confluence. The cells in the logarithmic growth phase were used for the following experiments. A phase-contrast microscope (Nikon XDS-1B) and Tyrpan blue exclusion test were used to determine cell viability. 2.4 Cell-based bioactivity screening The screening scheme is shown in Fig. 1. The process includes incubation (drug-cell interaction), washing, digestion and LC-MS analysis. The cells in the logarithmic growth phase were incubated in a cell culture bottle until a density of 5.0 × 105 cells/mL, and then were cultured in DMEM medium free of serum in a humidified 5% CO2 incubator at 37 °C for 24 h. The culture medium was discarded, and the cells were resuspended in DMEM within the presence of ginseng extract (final concentration of 5 mg/mL), and incubated at 37 °C for 4 h. The suspension was centrifuged at 1,000 ×g for 5 min. The precipitate was then washed five times with PBS to remove unbounded components. The eluates were discarded except for the last one, which was collected as a control for LC-MS analysis. Finally, the cells were denatured with 2 mL of 40% acetic acid, and were extracted with 2 mL of methanol by ultrasonic extraction (80 Hz, 15 min). After centrifugation at 15,800 × g for 15 min, the obtained supernatant was evaporated to dryness under a stream of nitrogen gas at 40 °C. The residue was reconstituted in 100 µL of methanol by vortexing for 2 min and centrifuging at 15,800 × g for 10 min at 4 °C, and the suspension was used for UPLC-QTOF-MS analysis. The control samples free-of-drugs were prepared using the same procedures as the above. 2.5 UPLC-QTOF-MS analysis Ultra-performance liquid chromatography analysis was performed on a Nexera X2 series LC system (Nexera Technologies, Japan) equipped with a binary pump, micro-degasser, an auto-sampler, and a thermostatically controlled column compartment. The chromatographic separation was performed on a ZORBAX Eclipse plus C18 column (2.1×100 mm, 1.8-Micron) at 35 °C. The mobile phase was composed of water (A) and acetonitrile (B) both containing 0.1% formic acid, using the following gradient procedures: 15–30% (B) at 0–10 min, 30–30% (B) at 10–25 min, 30–50% (B) at 25–35 min, 50–70% (B) at 35–45 min, 70–90% (B) at 45–50 min, 90–5% (B) at 50–55 min. The sample volume injected was set at 5 µL. Mass detection was performed by on a Triple TOF 5600 plus Mass spectrometer (AB SCIEX, Framingham, USA) equipped with an ESI source. QTOF-MS analysis was performed in negative ion modes using full scan mode with a mass range of 100–2000 Da for both the TOF-MS and TOF-MS/MS scans. The following parameter settings were used: ion spray voltage, +/−4500 V; ion source heater, 500 °C; curtain gas (CUR, N2), 35 psi; nebulizing gas (GS1, Air), 60 psi; Tis gas (GS2, Air), 60 psi. The declustering potential (DP) was +/−100 V; collision energy (CE) was +/−45 V with a collision energy spread (CES) of 15 V. 2.6. Data analysis The accurate mass and composition for the precursor ions and fragment product ions were analyzed using the Peakview Software (AB SCIEX, version 1.2.0.3) integrated with the instrument. The exact mass calibration was performed automatically before each analysis employing the Automated Calibration Delivery System. Molecular formulas were generated by the molecular formula generator algorithm whose parameters were set as the following: C [0–80], H [0–150], O [0–60]. Other elements such as N, P, S, Br and Cl were not considered because of their rare presence in the ginseng. The empirical molecular formula was deduced from Peakview by comparing the theoretical mass of molecular ions and/or adduct ions with the determined values based on the limitation errors: mass accuracy <5 ppm, retention time <5.0%, and isotope abundance <10%. Components reported in the literature [1,2,21–31] in ginseng were summarized to establish a small-scale library for the rapid identification of non-target compounds. In a summary, retention time, accurate molecular weight, isotope abundance, fragment product ions, and literature data were reviewed to identify the compounds. 3. Results and discussions 3.1 UPLC-QTOF-MS conditions optimization In the studied ESI-MS the analytical conditions were optimized in order to enhance and achieve better resolution, higher sensitivity, and formation of abundant fragment ions. The addition of 0.1% formic acid in mobile phase appeared to significantly improve the detection sensitivity. In the negative ion mode, clearer ESI-MS were obtained with lower background noise, clearer mass spectrum, and higher detect sensitivity, so the ESI in the negative ion mode was chosen for the target-cell extract experiments. Other parameters, such as eluating gradient, ion source parameters, and mass analyzer settings, were also investigated. The optimum analytical conditions were set as described in the section “UPLC-QTOF-MS analysis”. 3.2. Identification of ginsenosides and component analysis of ginseng With respect to the structural characteristics of aglycone, ginseng saponins can be divided into several groups. The two major groups are the protopanaxadiol (PPD) group with sugar moieties attached to the -3 and/or C-20 and the protopanaxatriol (PPT) group with sugar moieties at C-6 and/or at C-20 [26,31–33]. Another family, the malonyl ginseng saponins, also called the acidic ginseng saponins, has a malonyl group attached at the 6-position of the glucosyl moiety. The minor group is oleanane (OLE) group with a nonsteroidal. The structure of other types of ginsenosides has some changes, but they still essentially belong to the original type of ginsenosides, only the side chain of the parent nucleus part is slightly different. The structures of ginseng saponins are shown in Fig. 2 and summarized in Table S1. The identification process of the detected compounds in ginseng was as follows. Firstly, the accurate molecular mass was obtained with the high-resolution QTOF-MS technique. When formic acid was added into the mobile phase, ginsenosides easily formed the deprotonated ion [M−H]− and adduct ion [M+HCOO]−, providing information about the molecular mass. Secondly, the formula was obtained by Peakview software according to the accurate molecular mass, element constituent, and isotope abundance. Thirdly, the differentiation and characterization of ginsenosides were completed according to the literature [20,26,32–34], and the types of ginsenosides were confirmed. The aglycone type can be distinguished according to the literatures in ESI+ mode [32–34]. The types could also be determined by characteristic fragmentations in ESI− mode. The relative characteristic aglycone ions at m/z 459, 477, 475, and/or 473 correspond to the PPD-type aglycone; m/z 475, 457, 493, and/or 473 to PPT-type; and m/z at 455.37 to OLE-type. In the low-energy CID-MS/MS (negetaive ion mode), the precursor deprotonated molecules formed characteristic sugar fragments product ions by the successive or simultaneous losses of the various sugar moieties: 162 Da (-Glc), and/or 146 Da (-Rha), and/or 132 Da (-Ara or-Xyl). Finally, the extracted compounds were identified according to their retention time behaviors, accurate MW, MS/MS fragmentation pathways, formation of aglycones product ions, and product ions produced by losses of side chains, and literature data. In this work, 54 compounds were identified in Panax ginseng extract (Fig. 3A). Ginsenoside Ra1 (PPD) was selected as an example to elucidate the analysis process. In the full scan ESI-MS of the UPLC peak #14 (at 21.708 min in Fig. 3A) gave high abundance ions at m/z 1255.6248 and m/z 1209.6272 (Fig. 3B); they were tentatively attributed as the [M+HCOO]− and [M−H]− according to the mass differences between the ions. So, the ion at m/z 1209.6272 is its quasi-molecular ion. Fig. 3B is the MS/MS spectrum of the peak 14. The fragment product ions at m/z 1077, 945, 783, 621, and 459 indicate the successive losses of 2 pentoses (−132 Da) and 3 hexoses (−162 Da) from the quasi-molecular ion at m/z 1209.6272. The product ion at m/z 459 is the characteristic ion of PPD type aglycone. Thus, the compound is a PPD ginsenoside that has a MW of 1210.6248 Da, and contains 2 pentoses (132 Da) and 3 hexoses (162 Da). Furthermore, relative high collision-induced dissociation (CID) energy was applied to determine sugar chain compositions in the range of m/z 70–400. As shown in Fig. 3C, m/z 161.04, 131.03, 119.03, and 113.02 indicate the presence of Glc; m/z 203.05, 323.09 indicate Glc–Glc composition; m/z 191.05 and 293.08 refer to Glc–Arap/Xly composition; and m/z 131.03, 149.04 indicate the presence of Arap/Xly. Therefore, the molecular mass, aglycone type, sugar numbers, and sugar composition were deduced. The results are consistent with the literature 28, so it is identified as Ra1. The typical MS/MS spectra of other types of ginsenosides are shown in Fig. 4. In Fig. 4A, the [M−H]− ion produced fragmentation at m/z 637, representing glycosidic cleavage by loss of one rhamnose residue (146 Da). The fragment ion at m/z 475, indicating the possibility of PPT-type aglycone, was produced by loss of glucose–rhamnose residue (162 Da +146 Da). The aglycone type was further verified with the characteristic ions at m/z 441.37, 423.36 and 405.35 in ESI+ mode. The results are consistent with the literature 23, so it is identified as as Rg2. In Fig. 4B, the [M−H]− ion at m/z 955 produced fragmentations at m/z 793, representing glycosidic cleavage by loss of one glucose residue (162 Da) in sugar chain at C-20 of Ro. The 6-position of the glucosyl moiety was attached to formic acid, which is easy to lose as CO2 and H2O (44 Da+18 Da). The ion at m/z 613 was generated by the loss of two glucose residue (162 Da) and one H2O residue (162 Da+162 Da+18 Da); The ions at m/z 731 and 569 were generated by the loss of CO2 and H2O; The ion at m/z 455, indicating OLE-type aglycone, was generated by loss of two glucose residues and glucose acid (162 Da+162 Da+176 Da). Malonyl ginsenosides are a special type of compounds, with a malonyl group attached at the 6-position of the glucosyl moiety. Most malonyl ginsenosides are derived from PPD-type ginsenosides. As shown in Fig. 4C, malonyl-ginsenosides Rd exhibited characteristic ions [M-H-malonyl]− by the loss of 86 Da and ions [M−H-CO2]− by the loss of 44 Da from [M-H]−. The other observed fragments are consistent with their corresponding ginsenosides. In addition, there is a special kind of ginsenosides, the 6-position of the glucosyl moiety was attached by butenoyl. These ginsenosides were derived from PPD-type ginsenosides or PPT-type ginsenosides. As shown in Fig. 4D, butenoyl-ginsenosides Rb2 exhibit characteristic ions [M-H-butenoyl]− by loss of 68 Da from [M-H]−. The other observed fragments are consistent with their corresponding ginsenosides. With the UPLC-QTOF-MS method, 54 compounds in white ginseng were identified and their MS characteristics are listed in Table 1. Many ginsenoside isomers are found in ginseng, and 8 isomers cannot be differentiated in this work (Table 1). Recent IonKey/MS Ion Mobility technique developed by Waters Corporation provide the feasibility of differentiating the isomers [35]. Compounds Rk1, Rg5, Rh2, Gypenoside XVII, Compound Mc-1, Compound O, Gypenoside IX, Majonoside R1, notoginsenoside N1 and notoginsenoside A were detected in white ginseng for the first time, and they were found in red ginseng [26] or the leaves of ginseng [27]. 3.3 Optimization of the conditions of cell extraction It is well known that the drug absorption of the cells plays a very important role in drug–cell interaction [15,20]. Influencing factors such as drug concentration, incubation time and digestion method were investigated carefully. The optimized conditions were: concentration of ginseng, 5 mg/mL; incubation time, 4 h; five washes with PBS buffer; digestion with methanol ultrasonic. 3.4. Find of bioactive candidates The typical total ion chromatograms (TIC) are shown in Fig. 5. Four compounds (peak No. 14, 15, 22, and 34) were found to be potential bioactive candidates by comparing the chromatograms between target cell-extract group (Fig. 5D) with the control samples (Fig. 5A, 5B and 5C). Furthermore, the detected four peaks were further confirmed by using a narrow mass window of 0.01 Da to restructure the extracted ion chromatograms (EICs) (Fig. 6A) with the main fragmentations of Rg2 (peak 14), Rh1 (peak 15), Ro (peak 22) and Rd (peak 34). The four candidates were characterized in Fig. 4A (Rg2), Fig. 4B (Ro), Fig. 6B (Rh1) and Fig. 6C (Rd), and their LC–MS data are summarized in Table 1. In general, when cells are incubated with drugs, the bioactive molecules may selectively bind with the cell or be transported into the cell. As can be seen from Fig. 4, some relatively high abundance compounds (peak no. 3 (Majonoside R1), 6 (Rg1), 7 (Re), 19 (Rb1) and 28 (Rb2), Fig. 5A) were not detected in the target cell-extract samples (Fig. 5D), which may indicate these compounds have no selective affinity for the target cells. Ginsenosides Rg1 [7–9], Rb1 [9,10], and Rd [11] have been found to protect SH-SY5Y cells, and Ro was shown to have a neuroprotective effect in animal experiments, but only Rd and Ro were verified in this work. Our experiments do not defy the pharmacological effects of the reported ginsenosides in the literature, and they may exert effects through other pathways. Except for Ro and Rd, the relatively low abundance Rg2 and Rh1 were found to be bioactive candidates, which indicates that the proposed method has suitable sensitivity and can be used as a screening tool for low content constituents in natural products. This case is interesting because most traditional pharmacological screening methods require relatively large amounts of purified compound. 4. Conclusion In this study, a total of 54 compounds in white ginseng were characterized and identified with the proposed UPLC-QTOF-MS. A target cell-based bioactivity screening method was developed and successfully applied to the predication of potential candidates in panax ginseng with SH-SY5Y cells as target cells, and four ginsenosedes, Rg2, Rh1, Ro, and Rd, were found to be the bioactive components. The target cell-based bioactivity screening method coupled with UPLC-QTOF-MS technique has suitable sensitivity and can be used as a screening tool for low content constituents in natural products. Supplementary Material This work was financially supported by the National Natural Science Foundation of China [grant numbers 81260605, 81560648] and the National Institutes of Health [grant numbers AT004418, AT005362]. Figure 1 Schematic drawing of target cell-based bioactivity screening process. Figure 2 Structures of skeleton (A) and monosaccharide & substituent group (B) of gensenosides. Figure 3 Typical TIC chromatogram of Panax ginseng extract by UPLC-QTOF-MS (A), MS/MS spectrum of the deprotonated ion and the proposed fragmentation pathway of Ra1 (B); MS/MS spectrum of Ra1 at m/z 70–400 and corresponding relationship between sugar moieties and its residue ions (C). Figure 4 Typical MS/MS spectra and the proposed fragmentation pathways. A: Rg2; B: Ro; C: Malonyl-Rd; D: Ra6. Figure 5 Typical TIC chromatograms from various samples. A, ginseng extract treated with DMEM; B, the fifth eluate; C, the extract of cells incubated without ginseng extract; D, the extract of cells incubated with ginseng extract. Figure 6 The extracted ion chromatogram (EIC) of the cells incubated with ginseng extract, A (A1, the experimental group; A2, the control group). The negative MS/MS spectra and the proposed fragmentation pathways. B, Rh1; C, Rd. Table 1 Compounds identified from ginseng extract No . Name RT (min ) Formul a MW (Da) Measu red (Da)a Err or (pp m) Main MS/MS fragment ions Ref . 1 Majonoside R1/isomer 3.21 1 C42H72 O15 816.48 72 815.48 15 2.5 653.4313 [M−H−Glc]−; 635.4169 [M−H−Glc−H2O]−; 491.3742 [M−H−2Glc]− 27 2 Notoginsenoside N1/isomer 3.64 3 C48H82 O19 962.54 5 961.54 05 3.4 799.4956 [M−H−Glc]−;781.4812 [M−H−Glc−H2O]− 27 3 Majonoside R1/isomer 6.31 1 C42H72 O15 816.48 72 815.48 00 0.7 653.4299 [M−H−Glc]−; 635.4299 [M−H−Glc−H2O]−; 491.3761 [M−H−2Glc]− 27 4 20-glc-Rf 7.43 2 C48H82 O19 962.54 50 961.53 85 1.4 799.4893 [M−H−Glc]−; 637.4316 [M−H−2Glc]−; 475.3767 [M−H−3Glc]− 23, 25 5 Notoginsenoside R1 7.85 1 C47H80 O18 932.53 45 931.52 75 0.8 799.4871 [M−H−Glc]−; 637.4315 [M−H−Araf−Glc]−; 475.3794 [M−H−Araf−2Glc]− 23, 28 6 Ginsenoside Rg1 8.58 5 C42H72 O14 800.49 22 799.48 68 2.4 637.4321 [M−H−Glc]−; 475.3798 [M−H−2Glc]− 23, 28 7 Ginsenoside Re 8.64 5 C48H82 O18 946.55 01 945.54 58 3.2 799.4918 [M−H−Rha]−; 783.4956 [M−H−Glc]−; 637.4348 [M−H−Rha−Glc]−; 475.3806 [M−H−Rha−Glc]− 23, 28 8 Notoginsenoside A/isomer 9.92 3 C54H92 O24 1124.5 979 1123.5 901 0 961.5431 [M−H−Glc]−; 799.4781 [M−H−2Glc]−; 781.4802 [M−H−2Glc−H2O]− 27 9 Ginsenoside Rf 13.2 36 C42H72 O14 800.49 22 799.48 19 −3.1 637.432 [M−H−Glc]−; 475.3768 [M−H−2Glc]− 23, 28 10 Bu-Rg1 13.3 57 C46H76 O15 868.51 84 867.51 34 3.2 799.4899 [M−H−Bu]−; 637.4337 [M−H−Bu−Glc]−; 475.3741 [M−H−Bu−2Glc]− 22 11 Ginsenoside Ra3 13.8 98 C59H100 O27 1240.6 452 1239.6 330 −3.5 1107.6207 [M−H−Xyl]−; 945.5357 [M−H−Xyl−Glc]−; 783.5173 [M−H−Xyl−2Glc]−; 621.4370 [M−H−Xyl−3Glc]−; 459.3882 [M−H−Xyl−4Glc]− 28 12 Notoginsenoside Fa 13.9 64 C59H100 O27 1240.6 452 1239.6 319 −4.4 1107.5786 [M−H−Xyl]−; 945.5450 [M−H−Xyl−Glc]−; 783.4880 [M−H−Xyl−2Glc]−; 621.4380 [M−H−Xyl−3Glc]−; 459.3845 [M−H−Xyl−4Glc]− 24 13 Notoginsenoside R2 14.5 32 C41H70 O13 770.48 16 769.47 20 − 2.3 637.4273 [M−H−Xyl]−; 475.3757 [M−H−Xyl−Glc]− 23, 28 14 Ginsenoside Rg2 16.3 16 C42H72 O13 784.49 73 783.48 70 −3.2 637.4311 [M−H−Rha]−; 475.3781 [M−H−Rha−Glc]− 23, 28 15 Ginsenoside Rh1 16.5 47 C36H62 O9 638.43 94 637.43 30 1.4 475.3765 [M−H−Glc]− 23, 28 16 Bu-Re 16.8 54 C52H86 O19 1014.5 763 1013.5 708 2.3 945.5451 [M−H−Bu]−; 799.4853 [M−H-Bu−Rha]−; 637.4331 [M−H-Bu−Rha−Glc]−; 475.3772 [M−H-Bu−Rha−2Glc]− 22 17 Ginsenoside Ra1 17.3 41 C58H98 O26 1210.6 346 1209.6 272 0.3 1077.6158 [M−H−Xyl]−; 945.5423 [M−H−Xyl−Arap]−; 783.4917 [M−H−Xyl−Arap−Glc]−; 765.4602 [M−H−Xyl−Arap−Glc−H2O] −; 621.4410 [M−H−Xyl−Arap−2Glc]−; 459.3818 [M−H−Xyl−Arap−3Glc]− 28 18 Bu-Rf 17.6 88 C46H76 O15 868.51 84 867.51 13 0.8 799.5486[M−H−Bu]−; 637.4282 [M−H−Bu−Glc]−; 475.3801 [M−H−Bu−2Glc]− 22 19 Ginsenoside Rb1 18.1 90 C54H92 O23 1108.6 029 1107.5 911 −3.6 945.5507 [M−H−Glc]−; 783.4896 [M−H−2Glc]−; 621.4381 [M−H−3Glc]−; 459.3838 [M−H−4Glc]− 23, 28 20 Malonyl-Rb1 20.3 06 C57H94 O26 1194.6 033 1193.5 993 2.8 1107.6072 [M−H−mal]−; 945.5593 [M−H−mal−Glc]−; 783.4674 [M−H−mal−2Glc]−; 765.4901 [M−H−mal-2Glc−H2O]−; 621.4435 [M−H−mal−3Glc]− 23, 28 21 Ginsenoside Rc 20.7 34 C53H90 O22 1078.5 924 1077.5 846 0 945.0000 [M−H−Araf]−; 783.0000 [M−H−Araf−Glc]−; 621.0000 [M−H−Araf−2Glc]−; 459.3812 [M−H−Araf−3Glc]− 23, 28 22 Ginsenoside Ro 20.9 29 C48H76 O19 956.49 81 955.47 36 −1.7 955.4828 [M−H]−; 793.4456 [M−H−Glc]−; 713.427[M−H−Glc−CO2−H2 O]−; 613.3784 [M−H-2Glc−H2O]−; 569.3766 [M−H−2Glc−CO2−H2O]−; 455.3549 [M−H-3Glc−Acid]− 23, 28 23 Ginsenoside Ra2 21.7 08 C58H98 O26 1210.6 346 1209.6 202 1.7 1077.6158 [M−H−Xyl]−; 945.5381 [M−H−Xyl−Araf]−; 783.4853 [M−H−Xyl−Araf−Glc]−; 765.4739 [M−H−Xyl−Araf−Glc−H2O] −; 621.4294 [M−H−Xyl−Araf−Glc−H2O] −; 459.3787 [M−H−Xyl−Araf−3Glc]− 28 24 Ginsenoside F1 23.5 59 C36H62 O9 638.43 94 637.43 39 3.6 475.3816 [M−H−Glc]− 28 25 Malonyl-Rc 23.8 07 C56H92 O25 1164.5 928 1163.5 887 2.7 1119.5898 [M−H−CO2]−; 1077.577 [M−H−mal]−; 945.5308 [M−H−mal-Araf]−; 783.4829 [M−H−mal-Araf−Glc]−; 621.4227 [M−H−mal-Araf−2Glc]−; 459.3813 [M−H−mal-Araf−3Glc]− 23, 28 26 Malonyl-Rb2 23.9 27 C56H92 O25 1164.5 928 1163.5 867 1.0 1119.5893 [M−H−CO2]−; 1077.5760 [M−H−mal-Arap]−; 945.5363 [M−H−mal-Arap]−; 783.4904 [M−H−mal-Arap−Glc]−; 621.4354 [M−H−mal-Arap−2Glc]− 23, 28 27 Malonyl-Ra1/ Malonyl-Ra2 24.5 01 C61H100 O29 1296.6 350 1295.6 299 1.7 1209.6272 [M−H−mal]−; 1077.6158 [M−H−mal−Xyl]−; 945.5423 [M−H−mal−Xyl−Ara]−; 783.4917[M−H−mal−Xyl−A ra−Glc]−; 765.4602 [M−H−mal]−Xyl−Ara−Glc− H2O]−; 621.4410[M−H−mal]−Xyl− Ara−2Glc]−; 459.3818[M−H−mal−Xyl−A ra−3Glc]− 28 28 Ginsenoside Rb2 25.0 78 C53H90 O22 1078.5 824 1077.5 761 0.9 945.5345 [M−H−Arap]−; 783.4844 [M−H−Arap−Glc]−; 621.4313 [M−H−Arap−2Glc]−; 459.3813 [M−H−Arap−3Glc]− 23, 28 29 Ginsenoside Rb3 26.7 16 C53H90 O22 1078.5 924 1077.5 846 0 945.0000 [M−H−Xyl]−; 783.0000 [M−H−Xyl−Glc]−; 621.4186 [M−H−Xyl−2Glc]−; 459.3711 [M−H−Xyl−3Glc]− 23, 28 30 Malonyl-Rb3 27.8 90 C56H92 O25 1164.5 928 1163.5 896 3.5 1119.5832 [M−H−CO2]−; 1077.5738 [M−H−mal]−; 945.5557 [M−H−mal−Xyl]−; 783.4785 [M−H−mal−Xyl−Glc]−; 621.4281 [M−H−mal−Xyl−2Glc]− 23, 28 31 Notoginsenoside Fc 28.2 78 C58H98 O26 1210.6 346 1209.6 222 −3.8 1077.6158 [M−H−Xyl]−; 945.5410 [M−H−Xyl−Arap]−; 783.4793 [M−H−Xyl−Arap−Glc]−; 621.4312 [M−H−Xyl−Arap−2Glc]−; 459.3830 [M−H−Xyl−Arap−3Glc]− 24 32 Quinquenoside R1 28.7 63 C56H94 O24 1150.6 135 1149.6 118 −2.8 1107.5836 [M−H−AC]−; 1089.5745 [M−H−AC−H2O]−; 945.5285 [M−H−AC−Glc]−; 783.4795 [M−H−AC−2Glc]−; 621.4282 [M−H−AC−3Glc]−; 459.3848 [M−H−AC−4Glc]− 24 33 Chikusetsusaponi n IVa 29.2 61 C42H66 O14 794.44 53 793.44 06 3.9 631.3770 [M−H−Glc]−; 613.7845 [M−H−Glc−H2O]−; 569.3789 [M−H−2Glc−CO2−H2O]−; 455.3485 [M−H−2Glc−Acid]− 21, 29 34 Ginsenoside Rd 29.5 32 C48H82 O18 946.55 01 945.54 57 3.0 783.4816 [M−H−Glc]−; 621.4299 [M−H−2Glc]−; 459.3781 [M−H−3Glc]− 23, 28 35 Ginsenoside Rs1 29.8 93 C55H92 O23 1120.6 029 1119.5 993 3.3 1077.5668 [M−H−Ac]−; 945.5472 [M−H−Ac−Arap]−; 783.4679 [M−H−Ac−Arap−Glc]−; 621.4302 [M−H−Ac−Arap−2Glc]−; 459.3709 [M−H−Ac−Arap−3Glc]− 28 36 Malonyl-Rd 30.2 5 C51H84 O21 1032.5 505 1031.5 408 −1.8 987.5372 [M−H−CO2]−; 945.5301 [M−H−mal]−; 783.4829 [M−H−mal−Glc]−; 621.4292 [M−H−mal−2Glc]−; 459.3823 [M−H−mal−3Glc]− 23, 28 37 Ginsenoside Rs2 30.6 49 C55H92 O23 1120.6 029 1119.6 003 4.6 1077.5742 [M−H−Ac]−; 945.5422 [M−H−Ac−Araf]−; 915.5121 [M−H−Ac−Glc]−; 783.4852 [M−H−Ac−Araf−Glc]−; 621.4312 [M−H−Ac−Araf−2Glc]−; 459.3818 [M−H−Ac−Araf−3Glc]− 28 38 Gypenoside XVII 31.0 41 C48H82 O18 946.55 01 945.54 24 −0.4 783.4831 [M−H−Glc]−; 621.4288 [M−H−2Glc]−; 459.3804 [M−H−3Glc]− 26 39 ginsenoside Ra6 31.1 36 C58H96 O24 1176.6 075 1175.6 025 2.0 1107.5847 [M−H−Bu]−; 945.5311 [M−H-Bu−Glc]−; 783.4826 [M−H-Bu−2Glc]−; 621.4338 [M−H-Bu−3Glc]−; 459.3887 [M−H-Bu−4Glc]− 22 40 ginsenoside Ra7 31.7 48 C57H94 O23 1146.6 186 1145.6 150 3.7 1077.5731 [M−H−Bu]−; 945.5274 [M−H−Bu−Araf]−; 783.4809 [M−H−Bu−Araf−Glc]−; 621.4315 [M−H−Bu−Araf−2Glc]−; 459.3795 [M−H−Bu−Araf−3Glc]− 22 41 Compound Mc-1 31.8 68 C47H80 O17 916.53 90 915.52 99 −2.0 783.4833 [M−H−Araf]−; 621.4284 [M−H−Araf−Glc]−; 459.3816 [M−H−Araf−2Glc]− 26 42 Compound O 32.0 33 C47H80 O17 916.53 90 915.53 00 −1.9 783.4795 [M−H−Arap]−; 621.4306 [M−H−Arap−Glc]−; 459.3778 [M−H−Arap−2Glc]− 26 43 Pseudoginsenosid e Rc1 32.1 24 C50H84 O19 988.56 07 987.55 61 3.2 945.5316 [M−H−AC]−; 783.4837 [M−H−AC−Glc]−; 621.4302 [M−H−AC−2Glc]−; 459.3802 [M−H−AC−3Glc]− 26 44 ginsenoside Ra8/Ra9 32.3 05 C57H94 O23 1146.6 18 1145.6 155 4.2 1077.5731 [M−H−Bu]−; 945.5274 [M−H−Bu−Araf]−; 783.4809 [M−H−Bu−Araf−Glc]−; 621.4315 [M−H−Bu−Araf−2Glc]−; 459.3795 [M−H−Bu−Araf−3Glc]− 22 45 Gypenoside IX 32.4 46 C47H80 O17 916.53 90 915.53 52 3.8 961.5394 [M+HCOO]−; 783.4812 [M−H−Xyl]−; 621.4284 [M−H−Xy−Glc]−; 459.3864 [M−H−Xy−2Glc]− 26 46 Bu-Gypenoside XVII 33.7 09 C52H86 O19 1014.5 757 1013.5 650 −3.4 945.5287 [M−H−Bu]−; 783.4796 [M−H-Bu−Glc]−; 621.4307 [M−H-Bu−2Glc]−; 459.3760 [M−H-Bu−3Glc]− 22 47 20(S)-Ginsenosid e Rg3/isomer 34.0 69 C42H72 O13 784.49 73 783.48 86 −1.8 621.4291 [M−H−Glc]−; 459.3784 [M−H−2Glc]− 23, 28 48 Zingibroside R1 34.6 13 C42H66 O14 794.44 53 793.44 10 4.4 631.3742 [M−H−Glc]−; 613.3672 [M−H−Glc−H2O]−; 569.3779 [M−H−2Glc−CO2−H2O]−; 455.3474 [M−H−2Glc−Acid]− 26, 29 49 Bu-Rd 34.7 09 C52H86 O19 1014.5 763 1013.5 672 −1.3 945.5314 [M−H−Bu]−; 783.4778 [M−H-Bu−Glc]−; 621.4255 [M−H-Bu−2Glc]−; 459.3775[M−H-Bu−3Glc]− 22 50 20(R)-Ginsenosid e Rg3/isomer 35.5 36 C42H72 O13 784.49 73 783.49 27 3.4 621.4245 [M−H−Glc]−; 459.3759 [M−H−2Glc]− 23, 28 51 24(R)-pseudogins enoside RT5 38.6 59 C33H58 O14 678.38 27 677.37 60 1.6 677.3823 [M−H]−; 24 52 Ginsenoside Rk1 39.3 51 C42H70 O12 766.48 67 765.47 69 −2.6 03.4125 [M−H−Glc]− 23, 28 53 Ginsenoside Rg5 39.7 82 C42H70 O12 766.48 67 765.47 59 −3.9 603.4125 [M−H−Glc]− 23, 28 54 Ginsenoside Rh2 40.1 17 C36H62 O8 622.44 45 621.43 86 3.1 459.3792 [M−H−Glc]− 23, 28 a [M−H]−; Glc=β-D-glucopyranosyl (162 Da); Rha=α-L-rhamnopyranosyl (146 Da); Ara=β-D-arabopyranosyl (132 Da); Xyl=β-D-Xylopyranosyl (132 Da); Glc-Acid=β-D-glucopyranosylacetyl acid (176 Da); Ac=acetyl (42 Da); Bu=trans-but-2-enoyl (68 Da); Mal=malonyl (86 Da). Highlights ✧ A total of 54 compounds in white ginseng were characterized and identified with UPLC-QTOF-MS. ✧ Four candidates (Rg2, Rh1, Ro, and Rd) in ginseng were found with the target cell-based bioactivity screening method. ✧ The target cell-based bioactivity screening method coupled with UPLC-QTOF-MS technique has suitable sensitivity and can be used as a screening tool for low content constituents in natural products. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. References 1 Ru WW Wang DL Xu YP He XX Sun YE Qian LY Zhou XS Qin YF Drug Discov. Ther 2015 9 23 32 25788049 2 Liu XF Hao JY Tang Y Li JK Yan ZH Chen HF Yuan JB Mod. Chin. Med 2016 18 76 81 3 Vaibhav R Juan SM Sylvain D Front. Cell. Neurosci 2015 8 1 13 4 Kim HJ Kim P Shin CY J. Ginseng Res 2013 37 8 29 23717153 5 Cho IH J. Ginseng Res 2012 36 342 353 23717136 6 Hu S Han R Mak S Han Y J. Ethnopharmacol 2011 135 34 42 21349320 7 Chen XC Zhou YC Chen Y Zhu YG Fang F Chen LM Acta Pharmacol. Sin 2005 26 56 62 15659115 8 Shi C Zheng DD Fang L Wu F Kwong WH Xu J Biochim. Biophys. Acta 2012 1820 453 460 22178929 9 Li NJ Zhou L Li W Liu Y Wang JH He P J. Chromatogr. B 2015 985 54 61 10 Ni N Liu Q Ren H Wu D Luo C Li P Wan JB Su H Molecules 2014 19 3012 3024 24662068 11 Liu Y Zhang RY Zhao J Dong Z Feng DY Wu R Shi M Zhao G Int. J. Mol. Sci 2015 16 14395 14408 26114390 12 Li SL Li P Sheng LH Li RY Qi LW Zhang LY J. Pharm. Biomed. Anal 2006 41 576 581 16488100 13 Zhang HY Hu CX Liu CP Li HF Wang JS Yuan KL Tang JW Xu GW J. Pharm. Biomed. Anal 2007 43 151 157 16876368 14 Su SL Yu L Hua YQ Duan JA Deng HS Tang YP Lu Y Ding AW Biomed. Chromatogr 2008 22 1385 1392 18655214 15 Qu FN Qi LW Wei YJ Wen XD Yi L Luo HW Li P Biol. Pharm. Bull 2008 31 501 506 18310918 16 Hong M Wang XZ Wang L Hua YQ Wen HM Duan JA J. Pharm. Biomed. Anal 2011 54 87 93 20828965 17 Li YH Wang PX Xiao W Zhao L Wang ZH Yu L Am. J. Chin. Med 2013 41 221 229 23336518 18 Qiu JY Chen X Zheng XX Jiang XL Yang DZ Yu YY Du Q Tang DQ Yin XY Biomed. Chromatogr 2015 29 226 232 24925151 19 Sun M Huang LM Zhu JL Bu WJ Sun J Fang ZH Arch. Phar. Res 2015 38 1044 1053 20 Otvos RA Nierop PV Niessen WMA Kini RM Somsen GW Smit AB Kool J Anal. Chem 2016 88 4825 4832 27046509 21 Li SL Lai SF Song JZ Qiao CF Liu X Zhou Y Cai H Cai BC Xu HX J. Pharm. Biomed. Anal 2010 53 946 957 20667431 22 Zhu GY Li YW Hau DK Jiang ZH Yu ZL Fong WF Chem. Biodivers 2011 8 1853 1863 22006713 23 Zhang HM Li SL Zhang H Wang Y Zhao ZL Chen SL Xu HX J. Pharm. Biomed. Anal 2012 62 258 273 22310552 24 Chu C Xu SJ Li XN Yan J Liu L J. Food Sci 2013 78 653 659 25 Yang XB Yang XW Liu JX Mod. Chin. Med 2013 15 349 358 26 Wang HY Hua HY Liu XY Liu JH Yu BY J. Pharm. Biomed. Anal 2014 98 296 306 24973593 27 Mao Q Bai M Xu JD Kong M Zhu LY Zhu H Wang Q Li SL J. Pharm. Biomed. Anal 2014 97 129 140 24867296 28 Wu W Sun L Zhang Z Guo Y Liu S J. Pharm. Biomed. Anal 2015 107 141 150 25590943 29 Yang H Lee DY Kang KB Kim SO Yoo YH Sung SH J. Pharm. Biomed. Anal 2015 109 91 104 25767906 30 Qiu S Yang WZ Shi XJ Yao CL Yang M Liu X Jiang BH Wu WY Guo DA Anal. Chim. Acta 2015 893 65 76 26398424 31 Bai HR Wang SJ Liu JJ Gao D Jiang YY Liu HX Cai ZW J. Chromagrogr. B 2016 1026 263 271 32 Qi LW Wang HY Zhang H Wang CZ Li P Yuan CS J. Chromatogr. A 2012 1230 93 99 22349142 33 Wan JY Liu P Wang HY Qi LW Wang CZ Li P Yuan CS J. Chromatogr. A 2013 1286 83 92 23499252 34 Wang LL Han LF Yu HS Sang MM Liu EW Zhang Y Fang SM Wang T Gao XM Molecules 2015 20 20518 20537 26593895 35 McCullagh M Rao R Chipperfield J Douce D Waters Corporation, Application Note 720005423EN http://www.waters.com/webassets/cms/library/docs/720005423en.pdf

PMC005xxxxxx/PMC5130349.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 9214478 2614 Mol Ecol Mol. Ecol. Molecular ecology 0962-1083 1365-294X 25319487 5130349 10.1111/mec.12967 NIHMS830803 Article Nasonia vitripennis venom causes targeted gene expression changes in its fly host MARTINSON ELLEN O. * WHEELER DAVID *1 WRIGHT JEREMY *2 MRINALINI * SIEBERT AISHA L. † WERREN JOHN H. * * Biology Department, University of Rochester, Rochester, NY 14627, USA † Translational Biomedical Science Department, University of Rochester School of Medicine and Dentistry, Rochester, NY 14627, USA Correspondence: Ellen O. Martinson, Fax: 585 275 2070; e.martinson@rochester.edu 1 Present address: Institute of Fundamental Sciences, Massey University, Palmerston North 4442, New Zealand 2 Present address: New York State Museum, Albany, NY 12230, USA 23 11 2016 10 11 2014 12 2014 30 11 2016 23 23 59185930 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. Parasitoid wasps are diverse and ecologically important insects that use venom to modify their host’s metabolism for the benefit of the parasitoid’s offspring. Thus, the effects of venom can be considered an ‘extended phenotype’ of the wasp. The model parasitoid wasp Nasonia vitripennis has approximately 100 venom proteins, 23 of which do not have sequence similarity to known proteins. Envenomation by N. vitripennis has previously been shown to induce developmental arrest, selective apoptosis and alterations in lipid metabolism in flesh fly hosts. However, the full effects of Nasonia venom are still largely unknown. In this study, we used high throughput RNA sequencing (RNA-Seq) to characterize global changes in Sarcophaga bullata (Diptera) gene expression in response to envenomation by N. vitripennis. Surprisingly, we show that Nasonia venom targets a small subset of S. bullata loci, with ~2% genes being differentially expressed in response to envenomation. Strong upregulation of enhancer of split complex genes provides a potential molecular mechanism that could explain the observed neural cell death and developmental arrest in envenomated hosts. Significant increases in antimicrobial peptides and their corresponding regulatory genes provide evidence that venom could be selectively activating certain immune responses of the hosts. Further, we found differential expression of genes in several metabolic pathways, including glycolysis and gluconeogenesis that may be responsible for the decrease in pyruvate levels found in envenomated hosts. The targeting of Nasonia venom effects to a specific and limited set of genes provides insight into the interaction between the ectoparasitoid wasp and its host. enhancer of split extended phenotype parasitoid wasps venom Introduction Parasitoid wasps are abundant and ecologically important organisms that inject venom into and lay their eggs in or on other insects (Godfray 1994; Quicke 1997). There are an estimated 150 000–600 000 species of parasitoid wasps that utilize a diverse array of host insect species and developmental stages (Quicke 1997; Heraty et al. 2011). Through their venom, these wasps are known to induce a variety of metabolic and immunologic manipulations of their host insects to provide resources for the development of the wasp’s offspring (Rivers & Denlinger 1994, 1995b; Parkinson et al. 2001, 2002; Rivers et al. 2002a,b; Federici & Bigot 2003; Danneels et al. 2010; De Graaf et al. 2010). As such, the venom of parasitoids can be considered an ‘extended phenotype’ (Dawkins 1999) that manipulates the host’s physiology to provide an environment more conducive for development of the wasp’s offspring. Parasitoid wasps are either endoparasitic or ectoparasitic. Endoparasitoid wasps lay their eggs within the tissue of their hosts, and the primary effects of their venoms on hosts are temporary paralysis and immune suppression to protect the wasp’s offspring from encapsulation and other negative immune responses (Parkinson et al. 2002; Vincent et al. 2010; Mortimer et al. 2013). Some endoparasitoid wasp species harbour polydnavirus symbionts that also alter the immune system and development of the parasitoid’s host (Burke & Strand 2012). In contrast, ectoparasitoid wasps lay their eggs on the surface of their hosts, and their venoms primarily alter host development and metabolism, to provide an improved resource for their feeding offspring (Rivers & Denlinger 1994; Nakamatsu & Tanaka 2003). The venom of parasitoid wasps holds enormous potential as sources for new bioactive compounds and drug discovery (Danneels et al. 2010; De Graaf et al. 2010); however, the full extent of the venom on host physiology and the mechanisms behind their effects are still unknown. Nasonia vitripennis (Hymenoptera: Pteromalidae) (hereafter Nasonia) is the genetic model for ectoparasitoid wasps, with a complete genome sequence and several powerful genetic tools available (Werren & Loehlin 2009a; Werren et al. 2010). Nasonia is a pupal ectoparasitoid and a generalist in terms of host preference, including the flesh fly Sarcophaga bullata (Diptera: Sarcophagidae) (Rivers & Denlinger 1995a). Because Nasonia lays its eggs between the host pupa and the puparium wall, eggs can be removed, thereby eliminating confounding effects of larval feeding and isolating the effects of envenomation. Nearly 100 venom proteins have been identified in Nasonia, 23 of which have no sequence similarity to known proteins (De Graaf et al. 2010; Werren et al. 2010). Nasonia whole venom induces developmental arrest of the host pupa (Rivers & Denlinger 1995b), depression in respiration (Rivers & Denlinger 1994), decrease in pyruvate metabolism (Rivers & Denlinger 1994), alteration in lipids (Rivers & Denlinger 1995b) and suppression of the host immune response (Rivers et al. 2002a,b). Little is known, however, about the specific mechanisms by which venoms induce these observed changes in the hosts or the full range of physiological and metabolic effects. Most previous studies on Nasonia venom functions have focused on specific phenotypic effects, such as oxygen consumption, pyruvate levels or lipid composition (Rivers & Denlinger 1994, 1995b; Rivers et al. 2002a), or have been limited to a predetermined set of genes in a microarray (Danneels et al. 2013) or metabolites (Mrinalini et al. 2014). However, new sequencing technologies allow for the comparison of gene expression changes across the entire genome of envenomated and normally developing hosts. These data can reveal both small and large venom-mediated changes in gene expression of major metabolic pathways, including previously unknown venom effects on host physiology and metabolism, as well as facilitate inference of potential mechanisms behind known venom effects. To determine the effects of N. vitripennis venom on gene expression in the host S. bullata, we used whole-body RNA sequencing (RNA-Seq) to sequence envenomated, normally developing and mechanically injured hosts at six time points ranging from 4 to 120 h post-envenomation. By comparing gene expression profiles within and between treatments as well as a global metabolic study of venom alteration in S. bullata (Mrinalini et al. 2014), we report (i) venom-induced changes in gene expression during S. bullata development, (ii) effects of simulated physical injury by Nasonia’s ovipositor and (iii) several possible mechanisms by which the venom initiates changes in host development, metabolism and immunity. Additionally, we report the S. bullata transcriptome assembly, thereby providing a valuable resource that will aid studies on important aspects of insect biology, such as diapause and immune stimulation. Methods and materials Sarcophaga bullata and Nasonia preparation A colony of S. bullata is maintained as previously described (Werren & Loehlin 2009b). For each of the three replicates, we used pupae that had pupated within a 12-h window of time to standardize the age of the hosts. Pupae were allowed to develop for a further 3 days at 30 °C and then kept at room temperature at 25 °C (<10 h) until needed. Newly enclosed and mated female Nasonia (strain AsymCx) (N = 80) were isolated in single test tubes containing two S. bullata pupa for 8 h to precondition stinging and host-feeding behaviours. After this period, S. bullata pupae were removed and the Nasonia were kept at 25 °C for 16 h without hosts to encourage stinging and oviposition during the experiment. Experimental setup and protocol The experiment is designed to identify the effects of Nasonia venom on gene expression in the host S. bullata by comparing envenomated, normally developing and mechanically injured hosts at 4, 8, 24, 36, 72 or 120 h after envenomation (Fig. 1). Three replicates for each treatment described below were performed on different dates (May, July and September 2012) to provide biological and technical replication. Sarcophaga bullata pupae (N = 240) were placed in drilled foam plugs leaving only the anterior portion of the puparium exposed for Nasonia stinging and oviposition, the anterior portion was chosen for ease of parasitoid egg removal. Prepared host pupae were randomly and equally divided into three experimental treatment groups: envenomated, mechanically injured and normally developing. For the envenomation treatment group, hosts were placed in tubes containing preconditioned Nasonia and kept at 25 °C for 4 h, allowing wasps sufficient time for stinging. Mechanically injured treatment group involved hosts punctured with an insect pin (size 00) between the second and third puparial segments. This was used to simulate insertion of Nasonia ovipositor into hosts to differentiate the effects of mechanical injury from the effects of envenomation. Normally developing and mechanically injured hosts were treated under identical conditions as envenomated hosts, with the exception that no Nasonia was added to the hosting tube. All hosts were removed from their plugs after 4 h, and 2–3 anterior segments of the pupal case were carefully removed to keep the host pupa intact within the puparium. Hosts were examined to confirm stinging or mechanical damage, both evidenced by a melanization reaction at the site. Nasonia eggs laid on envenomated hosts were removed using a paintbrush. Hosts that did not have a sting site and eggs were excluded from the study. Transparent gel caps (Medisa Inc.) were used to protect the anterior of host puparia from desiccation and injury. The processed hosts were kept at 25 °C until collection time. For each treatment, hosts were divided into six time points: collected at 4, 8, 24, 36, 72 or 120 h after the pupae were first placed into hosting tubes. For each time point, six S. bullata per treatment (N = 126) were collected for RNA extraction. Gel caps were removed, and each pupa was removed using fine-tipped forceps and placed into tubes containing 1 mL of TRIzol® (Life Technologies). Hosts were homogenized using plastic pestles, and the resulting homogenates were stored at −80 °C until all time points from a replicate set were processed. Total RNA extraction and sequencing Extraction of total RNA was carried out in TRIzol® following manufacturer’s instructions (Life Technologies). Basic quantification and quality assessments were performed using a NanoDrop ND-1000 Spectrophotometer (ThermoScientific). The five highest quality RNA extractions for each treatment and time point were then pooled ensuring equimolarity of each sample in a final 20 μL volume at a total RNA concentration of 1 μg/μL. Three replicates of the five pooled RNA samples were submitted to the Epigenomics Core Facility of Weill Cornell Medical College for cDNA library preparation and sequencing by 50 bp single-end read Truseq RNA sequencing on an Illumina HiSeq 2000 machine. Sarcophaga bullata reference transcriptome To ensure a high-quality reference transcriptome for mapping, we did additional sequencing of envenomated and normally developing hosts collected over the same six time periods with 454 sequencing. A normalized library of all samples pooled was sequenced using GS FLX titanium reagents on a full 454 sequencing plate at the Indiana University Sequencing Core, Bloomington, Indiana, USA. Reads obtained from the 454 plate were quality-trimmed to a minimum Phred score of 30 using SolexaQA v2.1 (Cox et al. 2010). Reads shorter than 50 base pairs (bp) were removed using prinseq-lite (version 0.20.3) ‘-min_len’ switch (Schmieder & Edwards 2011). Similarly, the Illumina reads were quality-trimmed to a minimum Phred score of 40 across the entire sequence and reads smaller than 20 bp or containing any Ns were removed. To reduce memory requirements and effects of PCR artefacts on the assembly, exact duplicate Illumina reads were removed using the derep switch in prinseq-lite. A transcriptome assembly of both 454 and Illumina reads was carried out with Oases v0.2.08 using the packages oases_pipeline python script with Kmer ranges 12–39 and coverage set to 6× (Schulz et al. 2012). The longest transcript from each oases locus was selected to represent that locus, to reduce sequence redundancy generated by transcript isoforms. To ensure that high-expressing, conserved genes were not erroneously removed in this step, previously removed contigs with good BLAST homology to Drosophila melanogaster transcripts (e-value <1e-50) and high expression (RPKM > 100) were added back to the data set. Next, a mapping scaffold was generated using an in-house script to join the transcripts using 150 N spacer sequences, as well as to generate a gene feature file that described the position of each transcript in the scaffold. To filter out loci that had very low expression, all of the good quality Illumina reads were mapped to the reference and counts were generated using HTSeq-count with the intersection-strict mode (Anders et al. 2014). Any transcripts with <1 count per million input reads were removed. The reference transcriptome was annotated using a BLASTX search of the nr database (e-value <1e-5). Additional local databases from the published genomes of several model organisms (Anopheles gambiae, Drosophila melanogaster, Xenopus laevis, Mus musculus, Homo sapiens, Rattus norvegicus and Danio rerio) were also searched using BLASTX to evaluate the consistency of the annotation across model organisms. Gene ontology (GO) terms were assigned by BLAST2GO v2.5.0 (Conesa et al. 2005). Data analyses Tophat2 v2.0.4 (default settings) was used to map Illumina reads to the reference transcriptome for fragments per kilobase per million (FPKM) values, and counts for each condition and replicate were generated using HTSeq-count with the intersection-strict mode (Kim et al. 2013). DESeq v1.12.1 was used to generate normalized read counts and differential expression calls for each time point and treatment (Anders & Huber 2010). In all comparisons, three biological replicates of each treatment group were used to calculate differential expression. Clustering was performed on temporal expression within treatment using the k-means analysis in JMP Pro®, v10.0.0 (SAS Institute Inc). A multidimensional scaling analysis and global linear analysis of all replicates were performed in EdgeR (Robinson et al. 2010; McCarthy et al. 2012). Three comparisons were made to determine groups of differentially expressed genes among treatments (Fig. 1): (i) to determine the effect of physical injury by the ovipositor on S. bullata, we compared normally developing and mechanically injured hosts during the first 24 h when almost all differential gene expression occured; (ii) to determine temporal patterns of gene expression within normally developing and envenomated hosts, we compared individual time points within treatments to the 4-h time point; and (iii) to determine the differences between normally developing and envenomated hosts, we directly compared envenomated to normally developing hosts at the same time point. To be considered significantly differentially expression, a contig needed an adjusted P-value <0.05 with DEseq and a false discovery rate (FDR) adjusted P-value <0.05 with EdgeR GLM. Only contigs that were significantly differentially expressed in both tests were included in further analyses. Significantly, overrepresented GO categories were determined using BINGO in Cytoscape with a adjusted P-value <0.01 (Maere et al. 2005). Results and discussion Sarcophaga bullata reference transcriptome We generated a de novo S. bullata pupal reference transcriptome assembly in Oases using a combination of high-quality 454 (1 357 655 reads) and Illumina (37 767 128 reads) sequencing data (see Methods). The raw assembly consisted of 113 689 transcripts with an average length of 1427 bp (Table 1). Filtering (see Methods and Table 1) resulted in a final set of 8234 contigs that we refer to as the ‘core group’ in the following analyses. The S. bullata reference transcriptome was assessed for completeness using BLASTX to the NCBI protein database. Among the 8234 core transcripts, 6653 (80%) had a hit, with dipteran species making up the majority of top hits. We assessed the level of transcript fragmentation in the core group by analysing BLASTX alignments between D. melanogaster proteins and S. bullata-translated transcripts. On average, 73% of each D. melanogaster protein aligned with a S. bullata sequence, suggesting that the assembly was able to capture most of the coding region. Additionally, the S. bullata assembly recovered 87% of the highly conserved pupal-expressed transcripts from D. melanogaster with reciprocal BLAST, indicating that the transcriptome is likely to be a good quality representation of the pupal life stage of S. bullata. Interestingly, 79 of the S. bullata transcripts lacked a clear BLASTX homologue in Drosophilidae and yet had other hits in NCBI. Most of these sequences (62%) have a top hit to uncharacterized and hypothetical proteins. The two major groups of annotated proteins had hits to hexamerins and transposable elements. The hexamerins are a conserved group of proteins that predominately act as storage proteins in insects. Previous studies in the brachyceran Diptera have shown significant turnover within this group (Burmester et al. 1998). Finally, we compared the S. bullata transcriptome to the recently completed transcriptome of the closely related Sarcophaga crassipalpis (Hahn et al. 2009). BLASTN comparisons showed good overlap among the S. bullata core group transcripts, with 89% having a strong hit (bit score ≥100) to one of 6451 S. crassipalpis sequences. The combination of 454 and Illumina data helped create a longer assembly for S. bullata, with the average query length in S. bullata being 1963 bp (N50 2480 bp) compared to the 454-only assembly of S. crassipalpis with an average of 331 bp (N50 321 bp). Summary of normally developing and envenomated host gene expression Reads from envenomated, normally developing and mechanically injured treatments were mapped to 8234 genes in the S. bullata reference transcriptome, with a range of FPKM values from 0 to 375 798 (average 273.0 RPKM). Among the 18 samples sequenced (six time points by three treatments), only 17.2% (1418/8234) of the genes differed significantly in their expression either within or between treatments, leaving 82.8% of the genes with a relatively constant expression level across all samples (Fig. S1, Supporting information). A multidimensional scaling analysis on the expression profile of all genes shows no universal clustering by treatment, collection time or replicate (Fig. 2); however, several clusters did emerge. All envenomated hosts from 24 to 120 h form a cluster, demonstrating that there is a unique gene expression profile in later envenomated hosts that is distinct from normally developing hosts. Additionally, normally developing hosts at 72 and 120 h formed a distinct cluster, indicating an abrupt shift in pupal development between 36 and 72 h. The lack of distinct groups in earlier time points (envenomated hosts from 4 to 8 h and normally developing hosts from 4 to 36 h) can be explained by the low level of significantly differentially expressed genes between envenomated and normally developing hosts collected before the 72-h time point (described below). Twenty Nasonia genes were identified by BLASTX in one 72-h and two 120-h envenomated replicates. Despite this contamination, we found no significant changes in S. bullata expression in these samples compared to replicates at the same time point without Nasonia genes (Fig. S2, Supporting information). This observation suggests that only a small number of Nasonia larvae escaped detection and their limited feeding had little effect on the host at the molecular level. Effects of mechanical injury To determine whether the effects of mechanical injury from the wasp’s ovipositor induce any part of the envenomated phenotype, we compared the gene expression of hosts that were mechanically injured (with an insect pin) to that of normally developing hosts (Fig. 1a). Mechanically injured hosts had a quick and limited response compared to envenomated hosts, with only 20 differentially expressed genes at 4 h, three at 8 h and one at 24 h out of 8234 S. bullata genes (Table S1, Supporting information). Overrepresented GO terms in this group include response to stimulus, immune response and response to wounding. The genes with the largest response were three attacin genes (A, B, and D), which are antibacterial genes in D. melanogaster, and the gene pale, which is a direct target of the wound-induced signal transduction pathway involved in cuticular wound repair (Mace et al. 2005). Whereas none of the mechanically injured differentially expressed genes overlapped with envenomated genes at a given time point, five of the 4-h genes were shared at later envenomated time points (i.e haemolectin, attacin and cytochrome c). However, envenomation had a larger effect upon these genes with an average upregulated fold change of 14.1 compared to 2.7 in the mechanically injured hosts. The small amount of overlap in differentially expressed genes, the difference in expression amplitude and the differences in timing of the response all indicate that the physical wounding of the host by Nasonia’s ovipositor plays only a trivial role in the venom response. Temporal changes in genes expression within normally developing hosts To determine gene expression changes throughout the development of normally developing and envenomated hosts, we compared the starting gene expression at 4 h to each subsequent time point in each treatment (Fig. 1b). From a total of 8234 S. bullata genes, normally developing hosts had 852 genes that were differentially expressed across time (Fig. 3b). Within the first 36 h, normally developing hosts had minimal gene expression change (average 1.0 differentially expressed gene per time point) (Fig. 3a). Strikingly, the number of differentially expressed genes increased to an average of 668 genes at 72 and 120 h, indicating a developmental shift somewhere between 36 and 72 h. Genes that make up this developmental shift are identified and discussed in the developmental arrest section below. Temporal changes in genes expression within envenomated hosts Surprisingly, only 2.9% of total genes in the S. bullata transcriptome changed temporally in envenomated hosts. Envenomated hosts had a total of 244 differentially expressed genes, 164 genes were only differentially expressed in envenomated hosts, and 80 genes were differentially expressed in both envenomated and normally developing hosts (Fig. 3b). In contrast with normally developing hosts, envenomated hosts maintained a relatively constant level of differentially expressed genes across all time points (average 99.8 genes at any time point). They lacked the sudden increase in expression shown in their normally developing counterparts, with an increase in only 83 genes between 36 and 72 h (Fig. 3a). After envenomation, the vast majority of gene expression levels remain constant relative to the expression profile for when the host was initially stung, as no significant changes in expression can be seen in 97.1% of host genes from when the host was stung to any subsequent time point. This suggests that venom directly affects a relatively small subset of genes that change temporally in envenomated hosts. Several significant differences found in metabolites from a global metabolomics study of envenomated hosts were not mirrored in this transcriptome study (Mrinalini et al. 2014), suggesting that venom may be manipulating proteins and metabolites directly (e.g., via protein–protein interactions), as opposed to transcriptionally regulating pathways, to modify metabolism. Alternatively, by sequencing the transcriptome of whole pupa, changes in tissue-specific gene expression may not be detectable. Comparing envenomated to normally developing host gene expression The direct comparison of envenomated hosts to the corresponding normally developing host time points was used to look at the net effect of Nasonia venom on S. bullata. Overall, we found 1217 genes significantly differentially expressed between normally developing and envenomated hosts. Among all differentially expressed genes between treatments, 610 of these were also differentially expressed in temporal comparisons among normally developing hosts but not in envenomated hosts (Fig. 3b). These could be identified as genes most likely involved in the developmental arrest of the host. Within this set, a large subset showed sudden and dramatic expression change between 36 and 72 h (Fig. S3b, Supporting information). A set of 147 differentially expressed genes between treatments were shared with the temporal comparisons in envenomated hosts and identified as genes affected by venom, and 58 genes were shared in the temporal comparisons of both envenomated and normally developing hosts (Fig. 3b). Additionally, there were 402 genes that were not found in the within-treatment comparisons (Fig. 3b). These may represent genes that have small expression changes in opposite directions in envenomated and normally developing hosts, so that they are not significant within a treatment but become so when compared between treatments. The general pattern shows that up to 36 h post-envenomation, there are a relatively small number of changes in gene expression and most are upregulated in envenomated hosts. After 36 h, there was a dramatic increase in the number of differentially expressed genes, most of which are associated with normal development genes that were not upregulated in envenomated hosts (Fig. S3b, Supporting information). Earlier time points (4–36 h) averaged 93.8 significantly differentially expressed genes per time point with 86.9% upregulated in envenomated hosts. However, later time points (72 and 120 h) had significantly more differentially expressed genes (average 867.5 genes per time point) with only 46.3% upregulated in envenomated hosts (Fig. S3b, Supporting information). Whereas the majority of the genes had some functional assignment, a subset (~30%) remained completely uncharacterized. This subset was relatively consistently represented across all treatments and time points and did not show any abnormally large expression changes; therefore, the current annotations have probably captured most of the large patterns involved in S. bullata envenomation. The majority of genes with a functional annotation that were differentially expressed genes between envenomated and normally developing hosts fell into three broad categories: immunity, metabolism and developmental arrest. Developmental arrest The most striking phenotypic alteration of the envenomated host is its developmental arrest, characterized by lack of eye pigment deposition and bristle formation, lack of melanization in the cuticle and decreased oxygen consumption (Rivers & Denlinger 1994). Whereas developmental arrest induced by venom may superficially resemble pupal diapause, venom-induced arrest cannot be rescued by exogenous ecdysteroids (Rivers & Denlinger 1994) and ultimately results in pupal death. Our study shows that this arrest can also be observed on a transcriptional level by the sudden upregulation of genes at 72 h in normally developing hosts that does not occur in envenomated flies (Fig. 2). When comparing envenomated to normally developing hosts, genes associated with developmental arrest increase from 9.9% of genes differentially expressed between 4 and 36 h to 54.8% between 72 and 120 h. This dramatic increase of differentially expressed normally developing genes between 36 and 72 h contain genes with annotations that we would expect in pupal development with overrepresented GO terms of chitin metabolism, myofibril assembly, and electron transport chain (Table S2, Supporting information). In 72-h envenomated hosts, cuticular proteins had some of the largest decreases, with ten transcripts ranging from a 100- to 2072-fold decrease compared to normally developing flies. Other chitin genes also experienced significantly decreased expression such as chitin deacetylase (max 2.3 fold decrease) and chitin-binding domain protein peritrophin-A (max 568.7-fold decrease). Genes involved in myofibril assembly (some with specificity to flight muscles) such as actin (max 47.8-fold decrease), flightin (max 7.2-fold decrease) and myosin light chain 2 (max 5.1-fold decrease) did not increase in expression in envenomated hosts. Finally, expression levels of genes involved in energy production via the electron transport chain and the TCA cycle, such as electron transfer flavoprotein (2.5-fold decrease), NADH dehydrogenase (2.7-fold decrease), succinate dehydrogenase (2.0-fold decrease) and aconitase (2.3-fold decrease), significantly decreased in envenomated hosts. The expression of these genes was arrested in envenomated hosts and maintained at constant expression levels similar to 4–36 h normally developing hosts. Possible mechanisms of developmental arrest The sudden upregulation of >600 genes at 72 h in normally developing hosts (Fig. 3a), which is an attribute of pupal development, was not mirrored in envenomated hosts. In envenomated hosts, the expression levels of these dramatically changing pupal genes remained similar to expression levels in normally developing hosts between 4 and 36 h. This suggests that the shift in gene expression associated with normal development is not triggered in envenomated hosts; therefore, envenomated hosts may simply maintain status quo gene expression levels. Rivers et al. (2011) found that envenomated hosts have significant and selective apoptosis of brain tissue relative to fat body tissue 24 h poststinging. The targeted destruction of brain tissue could interfere with hormones that trigger metamorphosis, specifically prothoracicotropic hormone (PTTH), which is produced in the insect brain and is essential for the secretion of ecdysteroids. Changes in ecdysteriod titres in insect pupae have been shown to initiate dramatic cascades in gene transcription (Andres et al. 1993). With selective damage to the neural tissue, Nasonia venom could induce developmental arrest and redirect the physiology of the fly for the benefit of its progeny (Rivers et al. 2011). A family of significantly upregulated genes, the enchancer of split complex (e(spl)-C), may be involved in directed apoptosis of neural tissue and subsequent developmental arrest (Fig. 4). The e(spl)-C members are partially redundant genes that encode transcription factors of the basic helix-loop-helix (bHLH) family and are known to suppress neurogenesis (Nakao & Campos-Ortega 1996). Proneural cells that retain high expression of e(spl)-C genes either undergo programmed cell death or differentiate into epidermal cells, whereas proneural cells with low e(spl)-C expression develop into neural cells. Persistent expression of e(spl)-C genes in D. melanogaster causes severe neural and imaginal sensory organ defects by apoptosis in proneural cells (Nakao & Campos-Ortega 1996). Whereas the exact mechanism of how e(spl)-C genes induce apoptosis in cells is not known, it is known that DNA binding is not essential (Nakao & Campos-Ortega 1996). All e(spl)-C genes present in our transcriptome were significantly upregulated in envenomated hosts in at least one time point sampled (4–120 h). Proteins m1, m2, m4, m5, m7, m8, malpha, mbeta and mdelta from the e(spl)-C were significantly upregulated ranging from a 2.6- to 6.5-fold increase, with their expression peaking at 24 or 36 h post-envenomation (Fig. 4). The e(spl)-C is part of the notch signalling pathway, which is a key pathway involved in the control of multiple cell differentiation processes during embryonic and adult life in metazoans (Guruharsha et al. 2012). To initiate the e(spl)-C, a notch protein is cleaved by a reprolysin-like metalloproteinase to release a Notch intracellular domain (NICD), which binds to Suppressor of Hairless (SU(H)) and subsequently induces the transcription of e(spl)-C, which typically inactivates the expression of multiple developmental genes. Interestingly, a metalloproteinase has also been found in Nasonia venom (Danneels et al. 2010; De Graaf et al. 2010). The top sequence hit to this protein outside of Nasonia, is to the venom gland-expressed metalloproteinase of the parasitic wasp Eulophus pennicornis (e-value = 3e−45). When recombinant E. pennicornis metalloproteinases were injected into fifth instar larvae of their host, Lacanobia oleracea, partial mortality was observed due to unsuccessful moulting, and surviving individuals showed significantly slowed development and growth (Price et al. 2009). Furthermore, injected larvae seemed to miss a developmental cue by continuing to gain weight, while normal larvae lost weight in preparation for pupation. Interference in the notch signalling pathway was therefore suggested as a possible mechanism for E. pennicornis (Price et al. 2009), and our results indicate that this may be shared with Nasonia. Defence and stress genes It has previously been reported that within an hour of envenomation, the number of circulating hemocytes (cells involved in insect immunity and found in haemolymph) is severely reduced (Rivers et al. 2002b; Zhang et al. 2005). Other immune responses, such as melanization and clotting of host haemolymph, are also suppressed in envenomated flies (Rivers et al. 2002a). We found no evidence of immune suppression in the host transcriptome, which may suggest that the previously observed immune responses were the result of protein–protein interactions and not transcriptionally regulated. In contrast, we found a large upregulation of the coagulation gene haemolectin (50 fold increase at 36 h) in envenomated hosts, which may be a compensatory response of S. bullata to the venom’s suppression of haemolymph clotting. Additionally, the plasmatocyte surface protein, Nimrod (max 39× fold increase) (Kurucz et al. 2007), and a mitogen-activated protein kinase (max 39× fold increase), which have been shown to regulate hemocyte production (Zettervall et al. 2004), were also upregulated. These responses do not seem to restore the circulating hemocyte number or clotting response, as neither recover in envenomated hosts (Rivers et al. 2002a). Unsurprisingly, several heat-shock proteins were upregulated in the envenomated host. Heat-shock proteins are highly expressed when organisms are exposed to elevated temperatures or other stresses, and are involved in the folding and unfolding of other proteins. Heat-shock proteins 27 (85×), 70 (77×), 23 (21×) and 83 (5.3×) all showed significantly increased expression in response to envenomation. It would be expected that envenomation would repress the host’s immune response to allow the wasp offspring to feed more readily; however, this may also expose the host to bacterial or fungal infections that would diminish its quality. With this is mind, it is interesting that several genes that encode antimicrobial peptides, which are specific to suppression of bacteria and fungi infection, were significantly upregulated following Nasonia envenomation. Cercropin (3.4×), diptericin (4.7×), attacin (5.1×) and sapecin (or defensin) (1.76×) showed the largest expression changes. Genes in both the toll and immune deficiency (IMD) pathways that regulate antimicrobials are also upregulated (Lemaitre et al. 1997). In the toll pathway, spätzle (8.6×) and cactus (1.8×) are upregulated; however, toll itself is slightly downregulated (−3.3-fold change, though not statistically significant). In the IMD pathway, relish (2.4×) is upregulated. Interestingly, when we examined the gene expression from the Nasonia larvae that were failed to be removed in some replicates, some of the highest expressing genes at 72 and 120 h in the Nasonia larvae were also genes encoding antimicrobial peptides. This included nahymenopteracecin, nasonin, abaecin, clavesin and apolipophorin. Whether Nasonia venom has a direct effect on the activation of the toll and IMD pathways is not known. It has been proposed that serine proteases present in the venom could process spätzle, thereby directly activating the toll pathway (Danneels et al. 2010) as a preventative measure against bacterial or fungal infection. However, it is also possible that the venom simply leaves the antimicrobial pathways unaltered and they are activated after microbes invade immune-suppressed hosts. These alternatives have not been resolved. Metabolism One of the more striking results from our study is that Nasonia venom seems to affect a relatively small number of temporally changing genes in envenomated hosts (244 of 8234). The vast majority of genes stay at a relatively consistent expression level that is similar with levels observed in 4–36-hr normally developing hosts. Among genes that show expression changes, most deal with stress and defence responses, whereas metabolic genes remain stable with two exceptions: tRNA synthetases and pyruvate metabolism. Glutaminyl-, lysyl-, valyl-, tyrosyl-, tryptophanyl- and aspartyl-tRNA synthetases were all significantly upregulated in envenomated hosts, and a further 10 of the 33 total tRNA synthetases were also upregulated, although not significantly. An increase in free amino acids in envenomated hosts (Mrinalini et al. 2014) may explain the upregulation in nearly half the tRNA synthetases present in the S. bullata transcriptome, which have the primary function of binding specific amino acids onto their appropriate tRNA. However, tRNA synthetases can be multifunctional proteins regulated by a diverse set of control mechanisms. For example, in humans, glutaminyl-tRNA synthetase can block apoptosis induced by heat shock (Ko et al. 2001). It is therefore possible that these tRNA synthetases are performing a yet unknown function in envenomated hosts. Pyruvate metabolism Changes in the regulation of multiple enzymes involved in pyruvate metabolism were found that are consistent with significant decreases in pyruvate levels in envenomated hosts found in previous studies (Rivers & Denlinger 1994; Mrinalini et al. 2014). Pyruvate is essential in several important metabolic pathways; it is synthesized via glycolysis and provides energy for both aerobic and anaerobic respiration (Fig. 5). It has been previously shown that oxygen consumption of the host is drastically reduced (>85%) soon after envenomation causing the host to function in a hypoxic state (Rivers & Denlinger 1994). The current study shows a significant upregulation of lactate dehydrogenase (max 16-fold increase) in envenomated hosts from 24 to 120 h, which converts pyruvate to lactic acid under anaerobic conditions. Figure 5 compares expression data to a prior metabolic analysis for pyruvate metabolism. Alterations were also found to other pathways that could be contributing to the depletion of pyruvate. Specifically, malic enzyme, which converts pyruvate into malate, was upregulated (max 10.3-fold increase) from 24 to 120 h and could indicate increased mobilization of pyruvate to form malate (Fig. 5). This mode of pyruvate depletion is further supported by the significant increase in malate in 72- and 120-h envenomated hosts (Mrinalini et al. 2014). This reaction could also explain the increase in malate in contrast to the decrease in most other TCA cycle metabolites and provide further support that venom could be interfering with oxidative metabolism (Mrinalini et al. 2014). Pyruvate levels could further diminish by its conversion to glucose via the gluconeogenesis pathway. Although genes shared between gluconeogenesis and glycolysis do not show any significant changes in expression, two genes unique to gluconeogenesis, pyruvate carboxylase (max 8.2-fold) and phosphoenolpyruvate carboxykinase (max 15-fold) are both significantly upregulated from 24 to 120 h in envenomated hosts (Fig. 5). These two enzymes act together to reverse the final step of glycolysis performed by pyruvate kinase, thereby initiating gluconeogenesis. Direct observation of significantly lower levels of pyruvate support this; however, glucose levels do not change between envenomated hosts and normally developing hosts (Mrinalini et al. 2014), which would be expected if gluconeogenesis were upregulated. Mrinalini et al. (2014) also found significant increases in sorbitol in envenomated hosts, suggesting that excess glucose could be converted into sorbitol via increased aldose reductase activity in the sorbitol pathway. A gene with the annotation of aldose reductase is upregulated (max 5.6 fold at 36 h) in envenomated hosts. Increases in sorbitol have also been recorded in diapause and temperature-stressed insects (Chino 1958; Salvucci 2000). The current study has several parallels with a microarray study conducted on another Nasonia host, S. crassipalpis (Danneels et al. 2013). Similar lists of overrepresented GO terms were identified during normal fly development. In both host species, Nasonia envenomation significantly increases lactate dehydrogenase and upregulated heat-shock proteins and other defensive genes. Sequences from the enhancer of split complex were not included in the microarray, so the developmental arrest hypothesis presented in this study could not be compared to the microarray study in S. crassipalpis. Our study identified additional pathways potentially depleting pyruvate (malic enzyme and gluconeogenesis) and increased antimicrobial peptide expression (e.g. cercropin and diptericin) that were not seen in S. crassipalpis. Differences between the studies could be accounted for by different sample times (3–24 h vs. 4–120 h), host species and preparation, and methods used to detect gene expression changes (microarray vs. de novo transcriptomes). Conclusions We ascertained genomewide alteration in host gene expression induced by Nasonia venom through the first 5 days of the envenomated phenotype. While envenomation causes large morphological effects, only a very small subset of genes (~2%) showed significant changes in their expression levels. From this small set of genes, we were able to generate a series of hypotheses for the mechanics driving different venom effects for further testing. Developmental arrest, which is the most striking phenotypic change, could clearly be observed at the transcriptional level. The venom-mediated mechanism that induces developmental arrest may involve the upregulation of the enhancer of split complex. Nasonia venom induces upregulation of antimicrobial peptides, which could be involved in protecting the host from microbial growth that could diminish the nutritional quality of the host for the wasp larvae. Furthermore, our results show that venom changes S. bullata metabolism by manipulating genes in multiple pathways leading to depletion in pyruvate levels. The profile of global transcriptional changes induced by venom of the parasitoid model Nasonia documented here is a first step towards putting venom function and evolution into an ecological context. Nasonia is a generalist parasitoid of flies from different families (e.g. Sarcophagidae, Calliphoridae, Muscidae), whereas its close congeners N. giraulti, N. longicornis and N. oneida specialize on Protocalliphora blowflies occurring in bird nests. Future research will explore how host use influences the venom repertoire of these parasitoids and to what extent physiological effects of venoms generalize across fly species, for example does venom of the generalist Nasonia use the same mechanisms in all host species or do different host species have unique responses to envenomation. Additionally, we will be able to test how universal these mechanisms are in closely related species, such as Nasonia giraulti, a specialist on a single host species, whose venom composition varies from N. vitripennis (J. H. Werren, unpublished). If the mechanisms for developmental arrest or metabolic changes are universal, this may be useful in the biological control of pest insects, in which parasitoid wasps are already used. Furthermore, the limited and specific targeting of Nasonia venom to a controlled set of genes supports the potential of Nasonia venom as a source of highly specific bioactive molecules, with potential pharmaceutical value. Future knockdowns of individual venom proteins can provide a finer scale and an improved mechanistic understanding of individual venom functions. Supplementary Material Host Response Supporting Data Table S1 Transcripts significantly differentially expressed between normally developing and mechanically injured hosts. Table S2 Overrepresented GO terms involved in the developmental arrest of the host. Fig. S1 Clusters of similar genes expression for all 8230 contigs present in the S. bullata transcriptome. Fig. S2 Clustering of all S. bullata gene expression for individual replicates for 72 and 120 h including Nasonia-contaminate transcripts. Fig. S3 Differentially expressed genes between envenomated and normally developing hosts at the same time point (directionality relative to envenomated hosts). Fig. S4 The four major patterns of gene expression in the comparison between normally developing and envenomated hosts. Supplementary Data We thank the National Institutes of Health (RO1GM098667) and the Provost Multidisciplinary Award (University of Rochester) for financial support, and M Williams, R Edwards and A Dolan for technical assistance and VG Martinson for helpful discussion. Fig. 1 Experimental design and comparisons conducted in present study. Each circle represents three replicates of five pooled hosts. (a) Comparisons to determine the effect of wounding by the ovipositor of Nasonia. (b) Comparisons within treatments to determine temporally changing gene expression. (c) Comparisons between treatments to determine the effects of Nasonia venom. Fig. 2 Multidimensional scaling analysis showing clustering of treatment and replicates. Squares are replicates from envenomated host transcriptions and circles from normal developing host transcriptomes. The number in each designates the hours after envenomation the sample was collected. Fig. 3 Comparing the number of differentially expressed genes in envenomated and normally developing hosts. (a) Number of significantly differentially expressed genes (adjusted P < 0.05) was determined by comparing 4-h hosts to each subsequent time point. (b) Overlap of differentially expressed genes among the temporally comparisons within envenomated and normally developing hosts as well as between the two treatments at each time point. Fig. 4 Differential expression in fragments per kilobase per million (FPKM) of the enchancer of split complex between envenomated (solid line) and normally developing (dashed line) hosts at multiple time points. Fig. 5 Changes in pyruvate metabolism in envenomated hosts. Metabolites measured in Mrinalini et al. (2014) are highlighted within circles, and genes measured in current study are labelled by their gene abbreviations listed below. Arrows indicate the direction of significantly changing genes or metabolites in envenomated hosts. ACO, aconitase; AR, aldose reductase; CS, citrate synthase; FUM, fumarase; IDH, isocitrate dehydrogenase;LDH, lactate dehydrogenase; MDH, malate dehydrogenase; ME, malic enzyme; OGDH, 2-oxoglutarate dehydrogenase complex; PC, pyruvate carboxylase; PDC, pyruvate dehydrogenase complex; PEPC, phosphoenolpyruvate carboxylase; PK, pyruvate kinase; SCS, succinyl CoA ligase; SDH, sorbitol dehydrogenase; SQR, succinate dehydrogenase. Table 1 Assembly and annotation information for the reference Sarcophaga bullata transcriptome at multiple filtering stages. All assembled transcripts Longest transcript per loci Longest transcript per loci filtered* Number of contigs 113 689 14 697 8234 Average length      1427    1194 1974 BLASTX hit NCBI —    7387 (50%) 6653 (81%) BLAST hit to Drosophila    63181 (56%)    6802 (46%) 6475 (79%) Hit to drosophila pupal expressed genes — — 5669 (69%) * Longest transcript per loci filtered contains transcripts with >1 count per million reads E.O.M., D.W., J.W. and J.H.W. wrote the manuscript; E.O.M., D.W. and J.H.W. performed the analysis; D.W., J.W., M., A.L.S. and J.H.W. contributed to experimental design; J.H.W. funded the research. Data accessibility Raw reads from all samples are submitted to SRA study accession SRP044688. The transcriptome assembly is submitted to BioProject Database (http://www.ncbi.nlm.nih.gov/bioproject/) accession PRJNA255811. Supporting information Additional supporting information may be found in the online version of this article. Anders S Huber W 2010 Differential expression analysis for sequence count data Genome Biology 11 R106 20979621 Anders S Pyl PT Huber W 2014 HTSeq–A Python framework to work with high-throughput sequencing data Bioinformatics 10.1093/bioinformatics/btu638 Andres AJ Fletcher JC Karim FD Thummel CS 1993 Molecular analysis of the initiation of insect metamorphosis: a comparative study of Drosophila ecdysteroid-regulated transcription Developmental Biology 160 388 404 8253272 Burke GR Strand MR 2012 Polydnaviruses of parasitic wasps: domestication of viruses to act as gene delivery vectors Insects 3 91 119 26467950 Burmester T Massey HC Jr Zakharkin SO Benes H 1998 The evolution of hexamerins and the phylogeny of insects Journal of Molecular Evolution 47 93 108 9664700 Chino H 1958 Carbohydrate metabolism in the diapause egg of the silkworm, Bombyx mori—II: conversion of glycogen into sorbitol and glycerol during diapause Journal of Insect Physiology 2 1 12 Conesa A Götz S García-Gómez JM 2005 Blast2GO: a universal tool for annotation, visualization and analysis in functional genomics research Bioinformatics 21 3674 3676 16081474 Cox MP Peterson DA Biggs PJ 2010 SolexaQA: at-a-glance quality assessment of Illumina second-generation sequencing data BMC Bioinformatics 11 485 491 20875133 Danneels EL Rivers DB De Graaf DC 2010 Venom proteins of the parasitoid wasp Nasonia vitripennis: recent discovery of an untapped pharmacopee Toxins 2 494 516 22069597 Danneels EL Formesyn EM Hahn DA 2013 Early changes in the pupal transcriptome of the flesh fly Sarcophagha crassipalpis to parasitization by the ectoparasitic wasp, Nasonia vitripennis Insect Biochemistry and Molecular Biology 43 1189 1200 24161520 Dawkins R 1999 The Extended Phenotype: The Long Reach of the Gene Oxford University Press Oxford De Graaf DC Aerts M Brunain M 2010 Insights into the venom composition of the ectoparasitoid wasp Nasonia vitripennis from bioinformatic and proteomic studies Insect Molecular Biology 19 11 26 20167014 Federici B Bigot Y 2003 Origin and evolution of polydnaviruses by symbiogenesis of insect DNA viruses in endoparasitic wasps Journal of Insect Physiology 49 419 432 12770621 Godfray HCJ 1994 Parasitoids: Behavioral and Evolutionary Ecology Princeton University Press Princeton, New Jersey Guruharsha K Kankel MW Artavanis-Tsakonas S 2012 The Notch signalling system: recent insights into the complexity of a conserved pathway Nature Reviews Genetics 13 654 666 Hahn DA Ragland GJ Shoemaker DD Denlinger DL 2009 Gene discovery using massively parallel pyrosequencing to develop ESTs for the flesh fly Sarcophaga crassipalpis BMC Genomics 10 234 19454017 Heraty J Ronquist F Carpenter JM 2011 Evolution of the hymenopteran megaradiation Molecular Phylogenetics and Evolution 60 73 88 21540117 Kim D Pertea G Trapnell C 2013 TopHat2: accurate alignment of transcriptomes in the presence of insertions, deletions and gene fusions Genome Biology 14 R36 23618408 Ko Y-G Kim E-K Kim T 2001 Glutamine-dependent antiapoptotic interaction of human glutaminyl-tRNA synthetase with apoptosis signal-regulating kinase 1 Journal of Biological Chemistry 276 6030 6036 11096076 Kurucz E Márkus R Zsámboki J 2007 Nimrod, a putative phagocytosis receptor with EGF repeats in Drosophila plasmatocytes Current Biology 17 649 654 17363253 Lemaitre B Reichhart J-M Hoffmann JA 1997 Drosophila host defense: differential induction of antimicrobial peptide genes after infection by various classes of microorganisms Proceedings of the National Academy of Sciences 94 14614 14619 Mace KA Pearson JC McGinnis W 2005 An epidermal barrier wound repair pathway in Drosophila is mediated by grainy head Science 308 381 385 15831751 Maere S Heymans K Kuiper M 2005 BiNGO: a Cytoscape plugin to assess overrepresentation of gene ontology categories in biological networks Bioinformatics 21 3448 3449 15972284 McCarthy DJ Chen Y Smyth GK 2012 Differential expression analysis of multifactor RNA-Seq experiments with respect to biological variation Nucleic Acids Research 40 4288 4297 22287627 Mortimer NT Goecks J Kacsoh BZ 2013 Parasitoid wasp venom SERCA regulates Drosophila calcium levels and inhibits cellular immunity Proceedings of the National Academy of Sciences 110 9427 9432 Mrinalini Siebert AL Wright J 2014 Parasitoid venom induces metabolic cascades in fly hosts Metabolomics 10.1007/s11306-014-0697-z Nakamatsu Y Tanaka T 2003 Venom of ectoparasitoid, Euplectrus sp. near plathypenae (Hymenoptera: Eulophidae) regulates the physiological state of Pseudaletia separata (Lepidoptera: Noctuidae) host as a food resource Journal of Insect Physiology 49 149 159 12770008 Nakao K Campos-Ortega JA 1996 Persistent expression of genes of the Enhancer of Split complex suppresses neural development in Drosophila Neuron 16 275 286 8789943 Parkinson N Smith I Weaver R Edwards JP 2001 A new form of arthropod phenoloxidase is abundant in venom of the parasitoid wasp, Pimpla hypochondriaca Insect Biochemistry and Molecular Biology 31 57 63 11102835 Parkinson N Richards E Conyers C Smith I Edwards J 2002 Analysis of venom constituents from the parasitoid wasp Pimpla hypochondriaca and cloning of a cDNA encoding a venom protein Insect Biochemistry and Molecular Biology 32 729 735 12044489 Price D Bell H Hinchliffe G 2009 A venom metalloproteinase from the parasitic wasp Eulophus pennicornis is toxic towards its host, tomato moth (Lacanobia oleracae) Insect Molecular Biology 18 195 202 19320760 Quicke DL 1997 Parasitic Wasps Chapman & Hall Ltd London Rivers DB Denlinger DL 1994 Redirection of metabolism in the flesh fly, Sarcophaga bullata, following envenomation by the ectoparasitoid Nasonia vitripennis and correlation of metabolic effects with the diapause status of the host Journal of Insect Physiology 40 207 215 Rivers DB Denlinger DL 1995a Fecundity and development of the ectoparasitic wasp Nasonia vitripennis are dependent on host quality Entomologia Experimentalis et Applicata 76 15 24 Rivers DB Denlinger DL 1995b Venom-induced alterations in fly lipid metabolism and its impact on larval development of the ectoparasitoid Nasonia vitripennis (Walker) (Hymenoptera: Pteromalidae) Journal of Invertebrate Pathology 66 104 110 Rivers D Ruggiero L Hayes M 2002a The ectoparasitic wasp Nasonia vitripennis (Walker)(Hymenoptera: Pteromalidae) differentially affects cells mediating the immune response of its flesh fly host, Sarcophaga bullata Parker (Diptera: Sarcophagidae) Journal of Insect Physiology 48 1053 1064 12770028 Rivers DB Rocco MM Frayha AR 2002b Venom from the ectoparasitic wasp Nasonia vitripennis increases Na+ influx and activates phospholipase C and phospholipase A2 dependent signal transduction pathways in cultured insect cells Toxicon 40 9 21 11602274 Rivers DB Keefer DA Ergin E Uçkan F 2011 Morphology and ultrastructure of brain tissue and fat body from the flesh fly, Sarcophaga bullata Parker (Diptera: Sarcophagidae), envenomated by the ectoparasitic Wasp Nasonia vitripennis (Walker) (Hymenoptera: Pteromalidae) Psyche: A Journal of Entomology 2011 10.1155/2011/520875 Robinson MD McCarthy DJ Smyth GK 2010 edgeR: a Bioconductor package for differential expression analysis of digital gene expression data Bioinformatics 26 139 140 19910308 Salvucci ME 2000 Sorbitol accumulation in whiteflies: evidence for a role in protecting proteins during heat stress Journal of Thermal Biology 25 353 361 10838174 Schmieder R Edwards R 2011 Quality control and preprocessing of metagenomic datasets Bioinformatics 27 863 864 21278185 Schulz MH Zerbino DR Vingron M Birney E 2012 Oases: robust de novo RNA-seq assembly across the dynamic range of expression levels Bioinformatics 28 1086 1092 22368243 Vincent B Kaeslin M Roth T 2010 The venom composition of the parasitic wasp Chelonus inanitus resolved by combined expressed sequence tags analysis and proteomic approach BMC Genomics 11 693 708 21138570 Werren JH Loehlin DW 2009a The parasitoid wasp Nasonia: an emerging model system with haploid male genetics Cold Spring Harbor protocols 2009 1 10 Werren JH Loehlin DW 2009b Rearing Sarcophaga bullata fly hosts for Nasonia (parasitoid wasp) Cold Spring Harbor protocols 10 1 4 Werren JH Richards S Desjardins CA 2010 Functional and evolutionary insights from the genomes of three parasitoid Nasonia species Science 327 343 348 20075255 Zettervall C-J Anderl I Williams MJ 2004 A directed screen for genes involved in Drosophila blood cell activation Proceedings of the National Academy of Sciences of the United States of America 101 14192 14197 15381778 Zhang Z Ye G-Y Cai J Hu C 2005 Comparative venom toxicity between Pteromalus puparum and Nasonia vitripennis (Hymenoptera: Pteromalidae) toward the hemocytes of their natural hosts, non-target insects and cultured insect cells Toxicon 46 337 349 16026808

PMC005xxxxxx/PMC5130411.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 0373226 3643 Exp Cell Res Exp. Cell Res. Experimental cell research 0014-4827 1090-2422 26431586 5130411 10.1016/j.yexcr.2015.09.017 NIHMS831085 Article PKCε as a novel promoter of skeletal muscle differentiation and regeneration Di Marcantonio D ˆ12 Galli D ˆ134 Carubbi C 1 Gobbi G 134 Queirolo V 1 Martini S 1 Merighi S 5 Vaccarezza M 67 Maffulli N 89 Sykes SM 2 Vitale M *134 Mirandola P 134 1 Department of Biomedical, Biotechnological and Translational Sciences (S.Bi.Bi.T.), University of Parma, Italy 2 Immune Cell Development and Host Defense, Research Institute of Fox Chase Cancer Center, Philadelphia, PA 3 Centre for Molecular and Translational Oncology (COMT), University of Parma, Italy 4 Sport and Exercise Medicine Center (SEM), University of Parma, Italy 5 Department of Medical Science, University of Ferrara, Italy 6 Department of Human Sciences, Society and Healt (HSSH), University of Cassino, FR, Italy 7 School of Biomedical Sciences, University of Queensland, Brisbane, QLD 4072 , Australia 8 Barts and The London School of Medicine and Dentistry, Queen Mary University of London, UK 9 Department of Musculoskeletal Disorders, University of Salerno School of Medicine and Surgery, Salerno, Italy * Corresponding author Prof. Marco Vitale, M.D., University of Parma, Dept. of Biomedical, Biotechnological and Translational Sciences (SBiBiT), Via Gramsci, 14 – 43100 Parma – Italy, tel. +39.0521.033034, fax +39.0521.033033, marco.vitale@unipr.it ˆ Both authors contributed equally to this work. 23 11 2016 30 9 2015 15 11 2015 30 11 2016 339 1 1019 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. Introduction Satellite cells are muscle resident stem cells and are responsible for muscle regeneration. In this study we investigate the involvement of PKCε during muscle stem cell differentiation in vitro and in vivo. Here, we describe the identification of a previously unrecognized role for the PKCε – HMGA1 signaling axis in myoblast differentiation and regeneration processes. Methods PKCε expression was modulated in the C2C12 cell line and primary murine satellite cells in vitro, as well as in an in vivo model of muscle regeneration. Immunohistochemistry and immunofluorescence, RT-PCR and shRNA silencing techniques were used to determine the role of PKCε and HMGA1 in myogenic differentiation. Results PKCε expression increases and subsequently re-localizes to the nucleus during skeletal muscle cell differentiation. In the nucleus, PKCε blocks Hmga1 expression to promote Myogenin and Mrf4 accumulation and myoblast formation. Following in vivo muscle injury, PKCε accumulates in regenerating, centrally-nucleated myofibers. Pharmacological inhibition of PKCε impairs the expression of two crucial markers of muscle differentiation, namely MyoD and Myogenin, during injury induced muscle regeneration. Conclusion This work identifies the PKCε – HMGA1 signaling axis as a positive regulator of skeletal muscle differentiation. PKCε HMGA1 C2C12 satellite cells skeletal muscle differentiation Introduction Adult skeletal muscle homeostasis as well as myofiber repair are maintained by a small subset of muscle stem/progenitor cells called Myosatellites or Satellite Cells (SCs). SCs reside between the sarcolemma and the basal membrane of skeletal muscle fibers and are able to give rise to additional SCs or differentiate into mature skeletal muscle cells to form new fibers [1, 2]. The members of the MyoD family (Myod, Myf5, Myogenin and Mrf4) are basic helix–loop–helix (bHLH) transcription factors that are critical molecular mediators of skeletal muscle differentiation [3]. Myod and Myf5 are considered promote the early stages of differentiation regulating skeletal muscle cell commitment, proliferation and cell cycle withdrawal of SCs [4], whereas Myogenin and Mrf4 mediate the processes of late muscle cell differentiation, promoting the formation and the final maturation of myotubes [5, 6]. High mobility group (HMG) proteins are non-histone chromatin associated proteins that indirectly modulate the transcription of their targets by altering higher order chromatin structure. HMGA1 is expressed in embryonic and undifferentiated cells, but is largely absent in adult organs [7]. HMGA1 down-regulation in C2C12 cell line is required to initiate the skeletal muscle differentiation program allowing the expression of the MyoD family myogenic factors [8]. However, little is known about the regulatory mechanisms that influence HMGA1 expression during myogenic differentiation. The ε isoform of the PKC family (PKCε) is a serine-threonine kinase that is expressed in a wide variety of tissues including the hematopoietic system, intestine, brain, skin, liver, adipose tissue, kidney as well as cardiac and skeletal muscle. In many of these, PKCε regulates tissue homeostasis by regulating cell death and differentiation [9-14]. It is known that the θ isoform of the PKC family promotes the fusion of myoblasts and regulates the expression of caveolin-3 and β1D integrin [15]. Of note, it has also been demonstrated that PKCε expression increases during insulin-induced myogenic differentiation of the C2C12 cells [16]. In this study we investigated the functional role of PKCε in skeletal muscle cell differentiation as well as a potential role of PKCε as an upstream suppressor of Hmga1. We found that inhibition of PKCε prevents myogenic differentiation of C2C12 and primary SCs, whereas its overexpression accelerates cell differentiation. In vivo, PKCε inhibition results in impaired muscle regeneration and reduced expression of Myogenin and Mrf4. Mechanistically, we show that PKCε down-regulates Hmga1 expression, which consequently leads to the increase expression of myogenic differentiation genes. Finally, we demonstrate PKCε inhibition obstructs the process of injury-induced muscle regeneration in vivo. Materials and methods Mice The experimental procedures were conducted according to the “Guide for the Care and Use of Laboratory Animals” (Directive 2010/63/EU of the European Parliament). All the procedures described in this study were also approved by the Local Animal Research Ethics Committee of Ferrara (C.E.A.S.A) and Parma. Cardiotoxin injury and immunohistochemistry Acute injury was induced by intramuscular injection of Cardiotoxin (10 μM) in the tibialis muscle of CD1 adult mice [17]. In the case of PKCε –active peptides treatment, εV1-2 or ψεRACK (100 nM) were injected together with cardiotoxin. To study the regenerative process, mice were euthanized for histological analysis 3 and 7 days after injury. Muscle samples were fixed with 4% paraformaldehyde and embedded in paraffin. Sections (4 μm) were blocked with goat serum and incubated with primary anti PKCε antibody (Novus Biological NBP1-30126). Detection was performed using Vectastain elite ABC kit (Vector Laboratories) and nuclei were counterstained with haematoxylin [18]. Cell cultures Mouse myoblast C2C12 cell line and primary SC were cultured in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with heat-inactivated 10% fetal bovine serum (FBS), 2mM glutamine and 1% antibiotics (Growth Medium, GM). Cells were maintained in a humidified 5% CO2 atmosphere at 37°C. When the cell cultures reached 80% confluence, GM was substituted with DMEM supplemented with 2% horse serum (Differentiation Medium, DM) to induce myogenic differentiation. Each experiment was performed in triplicate. Satellite cells isolation SCs were isolated from hindlimb muscles of 2 days old CD1 mice. Briefly, muscles were incubated with collagenase/dispase solution (Roche, Basel, Switzerland) 4 times for 15 minutes at 37°C in agitation. Cell suspension was filtered with 40 μm nylon cell strainer and processed with Feeder Removal Microbeads (Miltenyi Biotec, Bergisch Gladbach, Germany). This immunomagnetic separation kit allows depletion of mouse fibroblasts from muscle digestion and ensures higher levels of SC purity than the “pre-plating SC isolation” method [17]. The SC obtained were seeded at a density of 1.25 × 105/cm2 in collagen-coated culture dishes and grown in fibroblast-conditioned GM medium (fcGM). fcGM was obtained diluting (1:1 ratio) the filtered supernatant of primary cultures of mouse fibroblasts with fresh GM medium. RNA extraction and quantitative RT-PCR Total RNA was extracted using the RNeasy mini kit (Qiagen) according to the manufacturer’s instructions. 1 μg of total RNA was reverse transcribed using ImProm-II™ Reverse Transcription System (Promega, Fitchburg, WI) in a final volume of 20 μl. Quantitative real-time PCR assay of mouse differentiation myogenic markers was performed using Syber Green method. Myod primers: fw 5’-TTC TTC ACC ACA CCT CTG ACA -3’ rev 5’-GCC GTG AGA GTC GTC TTA ACT T -3’ Mrf4 (Myf6) primers: fw 5’ –GAG ATT CTG CGG AGT GCC AT -3’ rev 5’- TTC TTG CTT GGG TTT GTA GC-3’ Myogenin primers: fw 5’- ATC CAG TAC ATT GAG CGC CT-3’ rev 5’-GCA AAT GAT CTC CTG GGT TG -3’ Myf5 primers: fw 5’- TGA GGG AAC AGG TGG AGA AC -3’ rev 5’ – AGC TGG ACA CGG AGC TTT TA -3’ Pkcε (prkce) primers: fw 5’- ATG TGT GCA ATG GGC GCA AG -3’ rev 5’- CGA GAG ATC GAT GAT CAC GT -3’ Hmga1 primers: fw 5’-CAA GCA GCC TCC GGT GAG -3’ rev 5’- TGT GGT GAC TTT CCG GGT CTT G -3’ Mouse beta-glucoronidase (Gusb), known to be a good internal control to study mRNA expression in muscular derived cell lines [19] was used to normalize all results. Gusb primers: fw 5’ – CCG CTG AGA GTA ATC GGA AAC – 3’ rev 5’- TCT CGC AAA ATA AAG GCC G -3’ Polymerase chain reactions were made by StepOne Real-Time PCR System (Applied Biosystems) and GoTaq ® qPCR Master Mix (Promega). For each well, the 20 μl reaction medium contained: 10 μl of 2X GoTaq ® qPCR Master Mix (with SYBR Green), 100 nM each forward and reverse primer, 7.6 μl of RNase-free water and 2 μl cDNA template 1:5. The cycling conditions were: 95°C for 20s followed by 40 cycles of 95°C for 3s and 60°C for 30s. Real-Time RT-PCR products were confirmed by the analysis of melting curves. Immunofluorescence Immunofluorescence was performed as previously described [20]. Briefly, cells were grown in 48 wells dishes containing a cover slide. At the indicated time points, cells were washed in PBS and fixed with 4% paraformaldehyde in PBS for 10 minutes at room temperature and stored in PBS at 4°C. Samples were permeabilized 3 times with 1% BSA, 0.2% Triton X-100 in PBS for 5 minutes at room temperature. Then, cells were incubated in 10% goat serum in PBS for 1 hour at room temperature to saturate non-specific binding sites. Samples were incubated for 1.5 hours with primary antibody diluted 1:200 in 1% goat serum in PBS. PKCε and myosin were detected by anti-PKCε rabbit serum (Novus Biologicals, Littleton, CO NBP1-30126) and anti-Myosin Heavy Chain (MHC) monoclonal antibody (clone MF-20; Developmental Study Hybridoma Bank), respectively. Cells were washed in PBS and then incubated with secondary antibody (Alexa Fluor 488 Donkey anti-mouse IgG and Alexa Fluor 594 anti-rabbit Donkey IgG) 1:1000 for 1 hour at room temperature. Nuclei were counterstained with DAPI; fluorescence was observed with a Nikon Eclipse 80i (Tokyo, Japan) fluorescent microscope (Nikon Plan). Images were acquired by Nikon Camera DS-JMC and analysed by Nis element F2.30 (Nikon, Japan). Myogenic differentiation levels were analyzed by fusion index (number of nuclei in the myotubes/total number of nuclei). For each sample at least 500 nuclei were counted and reported values are means of 3 independent experiments ± standard deviation. Fusion index analysis is reported as percentage (0% = no detectable fusion event among MYOSIN+ cell). *p<0,05 Anova-Dunnett test vs control cells. Cellular fractions separation and Western Blot analysis 5×106 cells were treated with NE-PER Nuclear and Cytoplasmic Extraction Reagents (Pierce), used according to manifacturer’s protocol. For Western Blot analysis, samples were resuspended in lysis buffer (50 mM Tris-HCl, pH 7.4; 1% NP-40; 0.25% sodium deoxycholate; 150 mM NaCl; 1 mM EDTA; 1 mM phenylmethylsulfonyl fluoride; 1 mM Na3VO4; 1 mM NaF) and 30 μg of total proteins were loaded on 10% SDS-polyacrylamide gels. Nitrocellulose membranes were incubated with the specific primary antibody (dilutions and buffers were as indicated by manufacturer) anti-PKCε (Merck Millipore, Darmstadt, Germany 06-991), anti-HSP70 (Sigma-Aldrich, St. Louis, MO, H5147), anti- α-tubulin (Sigma-Aldrich, St. Louis, MO), anti-insulin receptor β chain (IRβ, (Cell Signaling, Danvers, MA, #3025), anti-Myogenin (Santa Cruz, Dallas, TE sc-12732), anti-myoD (Santa Cruz sc-32758), anti GAPDH (Merk Millipore MAB374) anti-HMGA1 (Abcam, Cambridge, UK ab4078), then washed and incubated with 1:5000 peroxidase-conjugated anti-rabbit or with 1:2000 peroxidase conjugated anti-mouse IgG (Pierce). Signals were revealed by ECL Supersignal West Pico Chemiluminescent Substrate detection system (Pierce). Cell transfection PKCε expression levels were up-regulated in C2C12 cells by the transfection of 3 μg of murine GFP-PKCε plasmid and of GFP-K522M mutated PKCε control plasmid (kindly provided by Prof. Peter Parker, Cancer Research Institute, UK) [21] using the Superfect Transfection reagent (Qiagen, Hilden, Germany). Hmga1 silencing was obtained by transfection of 100 nM specific siRNAs or control siRNA (Ambion, Austin, TX). The siRNAs from Ambion are identified by the following catalog numers: ID S67596 and ID S67598. In addition, PKCε activity was pharmacologically modulated by the εV1-2 (CEAVSLKPT) and ψεRACK (CHDAPIGYD) peptides, conjugated to TAT47-57 (CYGRKKRRQRRR) by a cysteine disulfide bound [22]. Briefly, εV1-2 is a specific PKCε inhibitor designed from the C2 region of PKCε protein that acts as a binding competitor between PKCε and its anchoring protein εRACK. Instead, ψεRACK is a PKCε allosteric activator derived from the C2 region sequence, implicated in auto inhibitory intramolecular interactions. Peptides are high specific for PKCε and they don’t interact with other PKC isozymes [23]. Peptides regulate both the enzymatic function and the localization of PKCε through the subcellular compartments. C2C12 cells and SC were incubated with DM and treated with 1μM of peptides every 24 hours for 48 or 72 hours. Short hairpin RNA (shRNA) cell infection In some experiments we also used shRNA gene silencing to obtain a complete shut-down of PKCε expression. In this case we used a pLKO.1 lentiviral vector encoding shRNA against mouse PKCε (Open-Biosystem, Thermo Scientific,Waltham, MA). As control (shRNACTRL), we used the MISSION pLKO.1-puro Non- Target shRNA Control Plasmid, containing a shRNA insert that does not target any known genes from any species (Sigma-Aldrich, St. Louis, MO). The shRNA expressing viruses were produced in 293TL cells according to standard protocols. Mouse proliferating C2C12 cell line was infected with Pkcε shRNA or CTRL shRNA and then cultured in the presence of puromycin (2 μg/ml) to select infected, puromycin-resistant cells. Statistical analysis Data sets were examined by analysis of variance (ANOVA) for comparisons between multiple groups and Dunnett’s test for comparing a control group to all other groups (when necessary). A P value of less than 0.05 was considered statistically significant. RESULTS PKCε expression, activation and localization during C2C12 and primary satellite cell differentiation To evaluate PKCε expression during myotube formation in vitro and ex vivo, C2C12 and SC cells, respectively, were cultured in low serum medium for one week. Quantitative real time PCR analyses at several time points during the differentiation process confirmed that the expression of the early myogenic differentiation markers (Myod and Myf5), progressively decreased during the differentiation of C2C12 and primary SCs. As previously described [24], the transcription factors of the middle and late phases of skeletal muscle differentiation, Myogenin and Mrf4 accumulated during myofibers formation (Figure S1 A-B). Both Pkcε mRNA and PKCε protein levels progressively increased as proliferating myoblasts transitioned into myotube formation (Figure 1 A-D). Immunofluorescence microscopy was then applied to evaluate the subcellular localization of PKCε protein during the differentiation of C2C12 cell cultures. In undifferentiated C2C12 cells, PKCε levels were low with prevalent peri-nuclear staining (Figure 1E). During the first 24 hours of skeletal muscle differentiation, PKCε is preferentially localized inside the nucleus (arrow heads, middle panels of Figure 1E). PKCε then increases in both in the nucleus and cytoplasm at 72 hours (Figure 1E). The expression of the late muscle cell differentiation marker myosin was not detected in undifferentiated C2C12 cells, but progressively accumulated in the cytoplasm of forming myotubes (Figure 1E). Consistent with the results obtained by immunofluorescence, cell fractionation of C2C12 cells revealed that the nuclear content of PKCε protein significantly increases 3 days after the induction of cell differentiation (Figure 1F-G). Interestingly, while PKCε is upregulated and activated (increase of phospho-PKCε levels), HMGA1 expression is concomitantly down-regulated (Figure 1F-G). PKCε stimulates in vitro C2C12 and satellite cells differentiation via Myogenin and Mrf4 modulation Given these data, we subsequently investigated how the induction of PKCε expression correlates with changes in the expression of myogenic genes during terminal muscle differentiation. To determine whether PKCε influences the expression of the myogenic transcription factors (Myod, Myf5, Myogenin and Mrf4), C2C12 cells were engineered to express either a wild type mouse PKCε-GFP fusion protein (PKCε-GFP) or a kinase-inactive fusion protein carrying a point mutation in the catalytic core of the enzyme (PKCεm-GFP). Transfection efficiency of both plasmids was comparable (40±3% for PKCε-GFP and 43±5% PKCεm-GFP; supplementary figure 2). Expression of PKCε-GFP, but not inactive PKCεm-GFP, significantly increased Mrf4 and Myogenin mRNA levels but didn’t significantly impact MyoD and Myf5 expression (Figure 2B). Similarly, C2C12 cells (Figure 2C) and primary SC cultures (Figure 2D) treated with ψεRACK PKCε activator showed increased Mrf4 and Myogenin mRNA expression levels, whereas the εV1-2 PKCε inhibitor yielded the opposite effect. Fusion index analysis was performed on C2C12 cells treated with the εV1-2 PKCε inhibitor or ψε RACK activator to assess the extent by which PKCε inhibition impacts differentiation (Figure 2E-M). C2C12 cells exposed to the εV1-2 PKCε inhibitor showed a significant decrease in fusion index (20±15% vs 50±10% of TAT treated cells, p<0.05 Anova-Dunnett test vs TAT treated cells), while cells treated with the ψε RACK activator showed a significant increase in fusion index (85±12% vs 50±10% in TAT treated cells, p<0.05). These results, in combination with those of gene expression modulation experiments [16], reinforce a critical non-redundant role of nuclear PKCε in myogenic differentiation. Hmga1 is down-modulated by PKCε during C2C12 cell differentiation Consistent with previous studies, we observed a progressive decrease of Hmga1 expression (Figure 3A) in terminally differentiating C2C12 cell cultures [8]. Therefore, a potential relationship was investigated between PKCε and HMGA1 in proliferating C2C12 cells over-expressing PKCε. Expression of PKCε-GFP, but not of the inactive mutated PKCεm-GFP correlated with decreased expression of Hmga1 mRNA and protein as well as an accumulation of Myogenin in undifferentiated cells (Figure 3B-D). HMGA1 immunoprecipitation studies revealed that, although at low levels, endogenous PKCε form a complex with HMGA1. Furthermore, by overexpressing recombinant PKCε we observed the catalytically active form, but not the kinase dead version, co-precipitates with HMGA1, suggesting that kinase activity of PKCε is required for this interaction (Figure 3E). To examine how combined inhibition of Pkcε and Hmga1 affect Mrf4 and Myog gene expression during cell differentiation, C2C12 were transfected with Pkcε-targeting shRNA, Hmga1-specific siRNA or both. Reducing Hmga1 expression results in a significant increase of Myogenin and Mrf4 steady-state mRNA levels, whereas Pkcε inhibition significantly reduces Myogenin and Mrf4 expression (Figure 4A). Blocking Pkcε expression significantly impairs myotube formation (Figure 4E-G), determining a significant reduction of fusion index (10±3% vs 43±10% of shCTRL/siCTRL treated cells, p<0.05 Anova-Dunnett test vs shCTRL/siCTRL treated cells). Inhibition of Hmga1 expression leads to an increase in myotube formation (Figure 4H-J) and consequently of fusion index (62±8% vs 43±10% of shCTRL/siCTRL treated cells, p<0.05 Anova-Dunnett test vs shCTRL/siCTRL treated cells). Combined inhibition of Pkcε and Hmga1 expression significantly increases the expression of muscular differentiation markers (Figure 4A), the number of myotubes (Figure 4K-M) and fusion index (85±7% vs 43±10% of shCTRL/siCTRL treated cells, p<0.05 Anova-Dunnett test vs shCTRL/siCTRL treated cells), indicating that Hmga1 is a down-stream target of PKCε in the regulation of muscle cell differentiation program. In vivo induction of Pkcε during muscle regeneration To extend these initial observations, the impact of modulating Pkcε expression skeletal muscle repair and regeneration in vivo was assessed. To induce muscle injury and stimulate repair mechanisms, Cardiotoxin (CTX) was injected into mouse tibialis muscles. Western blot analyses of bulk muscle tissue revealed that PKCε sharply increases at day 3 post-CTX injection and continues to increase for at least 7 days following injury (Figure 5A and 5B). Histo-pathological analysis showed that the up-regulation of PKCε expression is most prominent in the fibers located at the site of injury, including the new regenerating fibers (centrally-nucleated fibers) (Figure 5C). Mouse tibialis muscles were then injected with CTX in combination with the PKCε inhibitor peptide (εV1-2), the PKCε activator peptide (ψεRACK) or control. Administration of εV1-2 inhibitor significantly inhibits CTX-induced PKCε phosphorylation (Figure 5D) and leads to a significant decrease in the levels of Myogenin and MyoD (Figure 5D and 5E). The non-redundant role of PKCε on muscle regeneration in vivo was also observed by morphological analysis of peptide treated tibialis muscles (supplementary figure 3). DISCUSSION During muscle development, myoblasts fuse together to form muscle fibers. Once the muscle is built, postnatal muscle growth and regeneration is maintained by the subset of muscle stem/progenitor cells called satellite cells (SC). Recent studies have raised the possibility that PKC family members play a crucial role in muscle differentiation [15, 16]. In the context of myogenic differentiation of C2C12 cell line and SC primary cells, our present data show that PKCε, belonging to the novel group of the serine-threonine kinase C family, is activated and up-regulated during muscle stem cell differentiation. Interestingly, the active form of PKCε, phosphorylated on Serine 729, increases during differentiation and is preferentially located in the nucleus (Figure 2). Previous studies have shown that the nuclear translocation of PKCε occurs through F-Actin as a possible transporter of phospho-PKCε [25]. Our data seem to be different from what Gaboardi et al. previously described. In their article, using an insulin- induced model of C2C12, they demonstrated that PKCε is mostly localized in cytoplasm, nearby the Golgi membrane. The discrepancy with our data can be explained in part by the use of different protocols and reagents. In the nucleus, PKCε is able to mediate the phosphorylation of many targets and alter their activation, subcellular localization or degradation [26, 27]. We have observed that Hmga1 is a possible target of nuclear PKCε in muscle cell differentiation. HMGA proteins are non-histone architectural elements of chromatin that dynamically modulate DNA-linked processes. These proteins are expressed in embryonic stem cells and in proliferating cells but are not detectable in fully differentiated cells [28]. Li et al. demonstrated that Hmga2 is important for myoblast proliferation and early myogenesis [29]. Also the Hmga1 isoform is known to be involved in muscle differentiation. Notably, Brocher et al. [8] have shown that Hmga1 down-regulation during the early phases of myogenesis is important for inducing the expression of myogenic markers, MyoD and Myogenin. Less is known about the signaling pathway that is involved in Hmga1 regulation during myogenesis. Here, for the first time, we show that PKCε alters Hmga1 expression during in vitro and ex vivo skeletal muscle differentiation. Specifically, we have found that siRNA-mediated inhibition of Hmga1 leads to increased expression of Myogenin and Mrf4 mRNA. We have also observed that the levels of nuclear PKCε expression increase in the nucleus upon differentiation and that inhibition of PKCε diminishes Myogenin and Mrf4 expression as well as myotube formation. Of note, the inhibition of muscle cell differentiation generated by shRNA Pkcε silencing could be completely abrogated by the simultaneous inhibition of Hmga1 expression. As skeletal muscle cell differentiation needs Hmga1 shut down to progress, we suggest that the nuclear translocation of activated PKCε is critical for Hmga1 inhibition and SC differentiation. Our data together with Gogoi et al. observations [30] demonstrate that HMGA1, phosphorylated by PKCε, may reside longer in the heterochromatin preferentially interacting with positively charged histones. Since PKCε promotes myogenic differentiation in vitro and ex vivo, which is a crucial phase of skeletal muscle regeneration, we studied the involvement of this kinase in a model of CTX - induced muscle repair in mice. We found that PKCε is up-regulated 7 days after injury, preferentially localizing at regenerating centrally-nucleated fibers. To pursue a better understanding of the PKCε involvement in muscle regeneration, we injected (intra-muscular) CTX- treated animals with a specific PKCε inhibitor peptide (εV1-2) to block PKCε activation and translocation. The consistent decrease of both Myogenin and Myod expression upon PKCε inhibition supports that PKCε contributes to the muscle regeneration process in vivo. The PKCε activator peptide, ψεRACK, did not enhance PKCε phosphorylation or the expression of either Myod or Myogenin induced by CTX. We infer that this observation is likely due to PKCε activation reaching a plateau level in the injured muscle. Overall, this study provides the first evidence for a role of the PKCε-HMGA1 axis in skeletal muscle differentiation and regeneration. Supplementary Material We are grateful to Vincenzo Palermo and Luciana Cerasuolo for technical support. FUNDING This work was supported by FIRB-accordi di programma 2010 CUP D91J10000100001 to M.V. (IT-Ministry of the University and Scientific and Technological Research/Ministry of Education, University and Research, MIUR) and Regione Emilia-Romagna Area 1 – Strategic Program 2010-2012 code PRUa1RI-2012-006 to P.M. D.D.M.: PhD fellow was supported by Cariparma Foundation. Figure 1 PKCε expression and localization during C2C12 and primary SC differentiation Panel A and B: PCR Real Time analysis of Pkcε mRNA during C2C12 (panel A) and SC cultures (panel B) differentiation, respectively. Results are representative of three independent experiments; values are reported as fold increase of control cell cultures (0 days) ± standard deviation. *p<0.05 Anova-Dunnett test (vs undifferentiated cells). Panel C: Western Blot analysis of PKCε protein expression levels during C2C12 cell differentiation; HSP70 was used as housekeeping protein. Panel D: Densitometry analysis of PKCε protein levels. HSP70 was used for normalization. Panel E: Immunofluorescence analysis: blue signal from nuclei obtained by DAPI staining in control (ctrl), in 24 hours (24h) differentiated- and in 72 hours (72h) differentiated cells; PKCε staining red; MYOSIN staining green colour. Arrow heads indicate cells with strong PKCε nuclear staining. Scale bar corresponds to 10 μm. Panel F: Western blot analysis of nuclear (n) and cytoplasmic (c) extracts from undifferentiated (Ctrl) and 72h differentiated C2C12 cells (72h); membranes were probed with anti-PKCε, anti phospho-PKCε (pPKCε), anti-HMGA1, anti-Myogenin and anti-HSP70 antibodies. Anti-Insulin Receptor (IR) antibody was used to exclude nuclear contamination by the cytoplasmic fraction. Panel G: Densitometry analysis of the PKCε expression levels. The values, normalized with respect to HSP70, are the mean of three independent experiments ± standard deviations (n=3). *p<0.05 Anova-Dunnett test (vs control cells). Figure 2 PKCε promotes myogenic differentiation through Myogenin and Mrf4 mRNA expression Panel A: Quantitative Real Time-PCR for Pkcε mRNA expression in C2C12 cell cultures transfected with wild type Pkcε(PKCε-GFP) or mutated Pkcε (PKCεm-GFP) compared with not transfected cells (-). Panel B: Quantitative Real Time-PCR for MyoD, Myf5, Mrf4 and Myogenin mRNAs (myog) in C2C12 cells transfected with wild type Pkcε (PKCε-GFP) or mutated Pkcε (PKCεm-GFP). Panel C-D: Quantitative Real Time-PCR for Mrf4 and Myogenin mRNAs (MYOG) in C2C12 (panel C) and SC cultures (Panel D) treated with 1 μM of PKCε specific activator and inhibitor peptides (ψεRACK and εV1-2, respectively). Housekeeping Gusb was used as reference gene. Values are reported as means of 3 independent experiments ± standard deviation. *p<0.05 Anova-Dunnett test vs untreated cells. Panels E-M: MYOSIN and Hoechst staining of 48h differentiated C2C12 cultures treated with peptides. (E) Hoechst staining of TAT treated cells; (F) Myosin (MHC) immunofluorescence of TAT treated cells, (G) merge of panels E-F. (H) Hoechst staining of εV1-2 treated cells; (I) Myosin (MHC) immunofluorescence of εV1-2 treated cells; (J) merge of panels H-I. (K) Hoechst staining of ψεRACK treated C2C12 , (L) Myosin (MHC) immunofluorescence of ψεRACK treated cells; (M) merge of panels K-L. Arrow heads indicate myotubes. Scale bar in M (100 μm) is the same for all the panels. Figure 3 HMGA1 is a target of PKCε during C2C12 cell differentiation Panel A Western blot analysis of HMGA1 during C2C12 myogenic differentiation for 4 days. HSP70 was used for normalization. Panels B-C: Western blot analysis of PKCε, Myogenin, HMGA1, and HSP70 in undifferentiated C2C12 cell cultures treated with (+) vectors expressing wild type Pkcε (PKCε-GFP) or mutated Pkcε (PKCεm-GFP). A representative experiment of three replicates is shown. Panel C: Densitometry analysis of HMGA1 and Myogenin (Myog) protein expression in C2C12 cells transfected with wild type or mutated Pkcε. Values are means of 3 independent experiments ± standard deviation. HSP70 was used for normalization. *p<0.05 Anova-Dunnet test (vs untreated cells). Panel D: Quantitative Real Time PCR analysis of Hmga1 in C2C12 cell cultures transfected with wild type Pkcε (PKCε-GFP) or mutated Pkcε (PKCεm-GFP) compared with not transfected cells (-). Panel E:Immunoprecipitation of HMGA1 in not transfected C2C12 cells (-), in C2C12 cells overexpressing the with wild form of PKCε-GFP fusion protein or the kinase-dead PKCεm-GFP. The immunoprecipitate was blotted with PKCε or HMGA1 antibodies (upper blots). Input lysates were blotted with PKCε and α-TUBULIN antibodies. Figure 4 PKCε - HMGA1 axis promotes C2C12 cell differentiation Panel A: Quantitative Real Time-PCR for Mrf4 mRNA expression (mrf4), Myogenin (myog), Pkcε and Hmga1 in C2C12 cell cultures infected with PKCε specific shRNA (shε) or control shRNA (shCTRL). After selection with puromycin (2μg/ml), infected cells were transfected with Hmga1 specific siRNAs (siHMGA1) or control siRNA (siCTRL) and then induced to differentiate for 2 days. Values are means of 3 independent experiments ± standard deviation. *p<0,05 by Anova-Dunnett test of Mrf4 and Myogenin expression (vs control cell cultures), respectively. Panels B-M: MYOSIN and Hoechst staining of 48h differentiated C2C12 cultures after silencing of Pkcε (panels E-G and K-M) or Hmga1 (panels H-J and K-M). Panels B-D: control C2C12 cultures infected with control shRNA and, after puromycin selection, transfected with control siRNA (shCTRL siCTRL); (B) Hoechst staining, (C) Myosin immunofluorescence, (D) merge of B-C. Panels E-G: C2C12 cultures infected with Pkcε shRNA and, after puromycin selection, transfected with control siRNA (shε siCTRL); (E) Hoechst staining, (F) Myosin immunofluorescence, (G) merge of E-F. Panels H-J: C2C12 cultures infected with control shRNA and, after puromycin selection, transfected with Hmga1 siRNA (shCTRL siHMGA1); (H) Hoechst staining; (I) Myosin immunofluorescence; (J) merge of panels H-I. Panels K-M: C2C12 cultures infected with Pkcε shRNA and, after puromycin selection, transfected with Hmga1 siRNA (shε siHMGA1); (K) Hoechst staining; (L) Myosin immunofluorescence; (M) merge of K-L. Arrow heads indicate myotubes. Scale bar in M (100 μm) is the same for all the panels. Figure 5 PKCε is up-regulated during in vivo skeletal muscle regeneration Panel A: Western blot analysis of protein extracts from regenerating tibialis muscle at 3 and 7 days after cardiotoxin induced injury in CD1 adult mice. The blot was incubated with anti-PKCε, anti-Myogenin and anti-HSP70 antibodies. Panel B: Densitometry analysis of PKCε protein levels. Values, normalized by HSP70 expression levels, are mean of 3 independent experiments ± standard deviations (n=3). Panel C: Immunohistochemical detection of PKCε and haematoxilin/eosin (H/E) staining of serial muscle section of CD1 untreated adult mice (control) and treated with CTX (3 and 7 days). Centro-nucleated regenerating fibers expressing PKCε are indicated (arrow heads). Scale bar corresponds to 40 μm and it is the same for all panels. Panel D: p-PKCε, Myogenin and MYOD western blot analysis of protein extracts from regenerating tibialis muscles at 7 days after cardiotoxin (CTX), cardiotoxin with εV1-2 (CTX εV1-2) and cardiotoxin with ψεRACK (CTX ψεRACK) injection. GAPDH was used as loading control. Panel E: Densitometry analysis of p-PKCε, Myogenin and MYOD expression levels. The values, normalized respect to GAPDH, are mean of 3 independent experiments ± standard deviations. *p<;0.05 Anova-Dunnett test of PKCε expression vs untreated muscle; # p≤ 0.05 and § p≤0.03 Anova-Dunnett test (vs CTX treated muscle). Disclosures No conflicts of interest are declared by the author(s) 1 Mauro A Satellite cell of skeletal muscle fibers J Biophys Biochem Cytol 1961 9 493 495 13768451 2 Tajbakhsh S Skeletal muscle stem cells in developmental versus regenerative myogenesis J Intern Med 2009 266 372 389 19765181 3 Pownall ME Gustafsson MK Emerson CP Jr Myogenic regulatory factors and the specification of muscle progenitors in vertebrate embryos Annu Rev Cell Dev Biol 2002 18 747 783 12142270 4 Ishibashi J Perry RL Asakura A Rudnicki MA MyoD induces myogenic differentiation through cooperation of its NH2- and COOH-terminal regions J Cell Biol 2005 171 471 482 16275751 5 Kassar-Duchossoy L Gayraud-Morel B Gomès D Rocancourt D Buckingham M Shinin V Tajbakhsh S Mrf4 determines skeletal muscle identity in Myf5:Myod double-mutant mice Nature 2004 431 466 471 15386014 6 Venuti JM Morris JH Vivian JL Olson EN Klein WH Myogenin is required for late but not early aspects of myogenesis during mouse development J Cell Biol 1995 128 563 576 7532173 7 Ozturk N Singh I Mehta A Braun T Barreto G HMGA proteins as modulators of chromatin structure during transcriptional activation Front Cell Dev Biol 2014 3 6 2 5 10.3389/fcell.2014.00005 eCollection 2014. Review 25364713 8 Brocher J Vogel B Hock R HMGA1 down-regulation is crucial for chromatin composition and a gene expression profile permitting myogenic differentiation BMC Cell Biol 2010 11 64 10.1186/1471-2121-11-64 20701767 9 Newton PM Messing RO The substrates and binding partners of protein kinase C epsilon Biochem J 2010 427 189 196 20350291 10 Gobbi G Di Marcantonio D Micheloni C Carubbi C Galli D Vaccarezza M Bucci G Vitale M Mirandola P TRAIL up-regulation must be accompanied by a reciprocal PKCε down-regulation during differentiation of colonic epithelial cell: implications for colorectal cancer cell differentiation J Cell Physiol 2012 227 630 638 21465464 11 Mirandola P Gobbi G Ponti C Sponzilli I Cocco L Vitale M PKCepsilon controls protection against TRAIL in erythroid progenitors Blood 2006 107 508 513 16166586 12 Gobbi G Mirandola P Carubbi C Micheloni C Malinverno C Lunghi P Bonati A Vitale M Phorbol ester-induced PKCepsilon down-modulation sensitizes AML cells to TRAIL-induced apoptosis and cell differentiation Blood 2009 113 3080 3087 18988868 13 Gobbi G Mirandola P Sponzilli I Micheloni C Malinverno C Cocco L Vitale M Timing and expression level of protein kinase C epsilon regulate the megakaryocytic differentiation of human CD34 cells Stem Cells 2007 25 2322 2329 17569788 14 Gobbi G Mirandola P Carubbi C Masselli E Sykes SM Ferraro F Nouvenne A Thon JN Italiano JE Jr Vitale M Proplatelet generation in the mouse requires PKCε-dependent RhoA inhibition Blood 2013 122 1305 11 23838351 15 Madaro L Marrocco V Fiore P Aulino P Smeriglio P Adamo S Molinaro M Bouché M PKCθ signaling is required for myoblast fusion by regulating the expression of caveolin-3 and β1D integrin upstream focal adhesion kinase Mol Biol Cell 2011 22 1409 1419 21346196 16 Gaboardi GC Ramazzotti G Bavelloni A Piazzi M Fiume R Billi AM Matteucci A Faenza I Cocco L A role for PKCepsilon during C2C12 myogenic differentiation Cell Signal 2010 22 629 635 19954762 17 Ceccarelli G Benedetti L Galli D Prè G Silvani G Crosetto N Magenes G Cusella De Angelis MG Low-amplitude high frequency vibration down-regulates myostatin and atrogin-1 expression, two components of the atrophy pathway in muscle cells J Tissue Eng Regen Med 2014 8 396 406 10.1002/term.1533 22711460 18 Galli D Carubbi C Masselli E Corradi D Dei Cas A Nouvenne A Bucci G Arcari ML Mirandola P Vitale M Gobbi G PKCε is a negative regulator of PVAT-derived vessel formation Exp Cell Res 2015 1 15 330 277 86 10.1016/j.yexcr.2014.11.011 25433270 19 Nishimura M Nikawa T Kawano Y Nakayama M Ikeda M Effects of dimethyl sulfoxide and dexamethasone on mRNA expression of housekeeping genes in cultures of C2C12 myotubes Biochem Biophys Res Commun 2008 367 603 8 18191039 20 Galli D Gobbi G Carrubbi C Di Marcantonio D Benedetti L De Angelis MG Meschi T Vaccarezza M Sampaolesi M Mirandola P Vitale M The role of PKCε-dependent signaling for cardiac differentiation Histochem Cell Biol 2013 139 35 46 22936275 21 Ivaska J Whelan RD Watson R Parker PJ PKC epsilon controls the traffic of beta1 integrins in motile cells EMBO J 2002 21 3608 3619 12110574 22 Brandman R Disatnik MH Churchill E Mochly-Rosen D Peptides derived from the C2 domain of protein kinase C epsilon (epsilon PKC) modulate epsilon PKC activity and identify potential protein-protein interaction surfaces J Biol Chem 2007 282 4113 4123 17142835 23 Begley R Liron T Baryza J Mochly-Rosen D Biodistribution of intracellularly acting peptides conjugated reversibly to Tat Biochem Biophys Res Commun 2004 318 949 54 15147964 24 Ogawa M Mizofuchi H Kobayashi Y Tsuzuki G Yamamoto M Wada S Kamemura K Terminal differentiation program of skeletal myogenesis is negatively regulated by O-GlcNAc glycosylation Biochim Biophys Acta 2012 1820 24 32 22056510 25 Dasgupta S Bhattacharya S Maitra S Pal D Majumdar SS Datta A Bhattacharya S Mechanism of lipid induced insulin resistance: activated PKCε is a key regulator Biochim Biophys Acta 2011 1812 495 506 21236337 26 Gupta P Ho PC Huq MD Khan AA Tsai NP Wei LN PKCepsilon stimulated arginine methylation of RIP140 for its nuclear-cytoplasmic export in adipocyte differentiation PLoS One 2008 3 e2658 18628823 27 Dey D Bhattacharya A Roy S Bhattacharya S Fatty acid represses insulin receptor gene expression by impairing HMGA1 through protein kinase Cepsilon Biochem Biophys Res Commun 2007 357 474 9 17434141 28 Catez F Hock R Binding and interplay of HMG proteins on chromatin: Lessons from live cell imaging Biochim Biophys Acta 2010 1799 15 27 20123065 29 Li Z Gilbert JA Zhang Y Zhang M Qiu Q Ramanujan K Shavlakadze T Eash JK Scaramozza A Goddeeris MM Kirsch DG Campbell KP Brack AS Glass DJ An HMGA2-IGF2BP2 axis regulates myoblast proliferation and myogenesis Dev Cell 2012 23 1176 88 23177649 30 Gogoi B Chatterjee P Mukherjee S Buragohain AK Bhattacharya S Dasgupta S A polyphenol rescues lipid induced insulin resistance in skeletal muscle cells and adipocytes Biochem Biophys Res Commun 2014 452 382 8 25157809

PMC005xxxxxx/PMC5130614.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 8501884 5061 J Perinatol J Perinatol Journal of perinatology : official journal of the California Perinatal Association 0743-8346 1476-5543 27583398 5130614 10.1038/jp.2016.135 HHSPA806234 Article Pulse Oximetry Screening for Critical Congenital Heart Disease in Planned Out of Hospital Births and the Incidence of Critical Congenital Heart Disease in the Plain Community Miller Kathleen K MD 1 Vig Kara S MD 2 Goetz Elizabeth M MD MPH 1 Spicer Gretchen LM CPM 34 Yang Alyssa J MPH 5 Hokanson John S MD 1 1 Department of Pediatrics, University of Wisconsin School of Medicine and Public Health, Madison, WI 2 NorthShore University Health Systems, Skokie, IL 3 Wisconsin Guild of Midwives, a Division of the Seattle Indian Health Board, Seattle, WA 4 Wisconsin Department of Health Services, a Division of the Seattle Indian Health Board, Seattle, WA 5 Urban Indian Health Institute, a Division of the Seattle Indian Health Board, Seattle, WA Corresponding Author: John Smith Hokanson MD, jhokanson@wisc.edu, H6/516c, 600 Highland Avenue, Madison, WI 53792, Telephone: 608-263-8535, Fax: 608-265-8065 5 8 2016 1 9 2016 12 2016 01 6 2017 36 12 10881091 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. Objective This study evaluated pulse oximetry screening (POS) for critical congenital heart disease (CCHD) in planned out of hospital births with special attention to births in Plain communities (Amish, Mennonite and similar). Design Wisconsin out of hospital births in 2013 and 2014 were evaluated. Care providers were supplied with and trained in the use of pulse oximeters for CCHD screening. State records were reviewed to identify deaths and hospital admissions due to CCHD in this population. Results Detailed information on POS was available in 1,616 planned out of hospital births. 799 were from the Plain community. 1,584 babies (98%) passed their POS, 16 infants (1%) failed, and 16 (1%) were not screened. 5 infants from the Plain community had CCHD, 3 were detected by POS. Conclusion POS for CCHD can be successfully implemented outside the hospital setting and plays a particularly important role in communities with high rates of CCHD and where formal prenatal screening is uncommon. INTRODUCTION Infants with congenital heart disease may be missed by both prenatal detection and physical examination in the immediate newborn period.1, 2, 3, 4, 5 Critical Congenital Heart Disease (CCHD), or congenital heart diseases in which intervention is needed in infancy, is not uncommon6 and delays in diagnosis can lead to significant morbidity and mortality 7, 8, 9, 10 Pulse oximetry screening (POS) has been shown to decrease the rate of missed CCHD 11, 12, 13, 14 and decrease the associated mortality due to CCHD in hospital born infants 11. Limited information is available on the utility of POS to detect CCHD has not yet been demonstrated in the out of hospital (OOH) birth population.15 Many factors complicate the use and evaluation of POS in the OOH birth population. Definitions of CCHD in the literature are not uniform, and with any newborn screening test, the yield of POS is affected by the prenatal detection rate. Accepting the variable definitions of CCHD and variable prenatal detection rates, sensitivities between 49.06% - 62.07% and specificities of 99.16% - 99.82% 11, 12, 13 have been reported using the recommended protocol16. The reported positive predictive values (PPV) range from 13.33% - 35.90% and negative predictive values (NPV) range from 99.16 – 99.82%. In these studies, the false positive rate ranges from 0.18% - 0.81% and the overall failure rates from 0.22% - 0.97%.11, 12, 13 Universal POS for CCHD was recommended by the US Secretary of Health and Human Services in 2011 and is now considered the standard of care for hospital born infants. 16, 17 In 2013 the AAP also recommend POS for planned home births.18 Reasoning behind this recommendation includes less frequent use of prenatal diagnostic testing, more limited periods of postnatal observation, and higher rates of missed CCHD in the OOH birth population. 9 In addition, OOH births may be attended by a wide variety of care providers with a range of experiences and skills including physicians, licensed and unlicensed midwives, and community birth attendants. In Wisconsin, women from Plain communities (Amish, Mennonite, and similar backgrounds) frequently opt for home deliveries and account for a significant proportion of the OOH birth community. The risk of CCHD may be higher in the Plain population. Ellis Van Creveld Syndrome (EVCS) is substantially more common in the Lancaster County (PA) Amish19 than the general population, is associated with a high incidence of congenital heart disease19, 20, 21, 22, and might contribute to an increase in CCHD in the Amish community. This study sought to evaluate the use of POS in OOH births in Wisconsin and to evaluate the incidence of CCHD in this population with special attention to births in Plain communities. METHODS This study of Wisconsin OOH births was performed from January 2013 through December 2014. This study was part of a larger project to implement and assess POS screening for CCHD funded by Health Resources and Services Administration (HRSA) Demonstration Grant H46MC24057. A detailed explanation of the Wisconsin Screening Hearts in NEwborns (SHINE) Project has been previously reported.15 The functions of the Wisconsin SHINE project were reviewed by the University of Wisconsin Health Sciences Institutional Review Board and determined to be quality assurance measures and not human subject research. For the purposes of this study, OOH births included those at the family home, those taking place at birthing centers, and births that occurred at the homes of midwives or community birth attendants. Initially, the Wisconsin SHINE Project provided pulse oximeters and training to members of the Wisconsin Guild of Midwives. Enrollment of licensed midwives began in late 2012 and continued throughout the study. The project later expanded to include unlicensed midwives, Plain community birth attendants, and members of the mainstream health care system involved in OOH births. A total of 83 health care personnel were trained in the use of pulse oximetry, of whom 8 were Plain community birth attendants, 12 were public health nurses, 2 were unlicensed English midwives, and 1 was a physician. The remaining 60 were licensed members of the Wisconsin Guild of Midwives. A total of 73 pulse oximeters were deployed during the study. Participants offered POS to families on a voluntary basis. Recommended screening time was between 24 and 48 hours after birth, and oxygen saturation was measured in the right hand and either foot with a handheld pulse oximeter and reusable probe (Masimo, Irvine, CA). Pass / fail results were determined as per the two-site oximetry protocol described by Kemper et al.16 Participants reported screening results and clinical outcomes on a standardized questionnaire. The standardized questionnaire included timing of screening, pass / fail, number of attempts, and basic demographic data such as zip code and maternal age. Mothers were identified as being part of a Plain community or not, but further differentiation within the Plain communities was not recorded. As membership in a Plain community is not routinely recorded on other Wisconsin documents, this designation could only be determined for home births within the SHINE project. Members of the Plain community often refer to people outside their community as "English". We used the designation of "English" to identify those families known to be outside the Plain community. CCHD was defined as one of the twelve diagnoses mentioned in the 2009 AAP evaluation of POS 23 (hypoplastic left heart syndrome, pulmonary atresia, tetralogy of Fallot, total anomalous pulmonary venous return, transposition of the great arteries, tricuspid atresia, truncus arteriosus, coarctation of the aorta, double outlet right ventricle, Ebstein's anomaly, interrupted aortic arch, and single ventricle) . Information on infants with other forms of congenital heart disease was not systematically recorded and could not be fully determined from the data gathered. Infants who passed the POS required no further evaluation. A protocol was established for failed screening that included contacting a hotline that would respond to questions regarding the algorithm or data collection methods, and would provide consultation and clinical support for any infant failing the screen. Access to an on-call pediatric cardiologist was available to the participating midwives at all times. As part of the Wisconsin SHINE project, the charts of all patients under one year of age admitted to the American Family Children's Hospital (Madison) or the Children's Hospital of Wisconsin (Milwaukee) with 1 of the 12 CCHD diagnoses were reviewed in detail to determine the mechanism of diagnosis, if POS had been performed, the place of birth, and if the baby was a member of a Plain community. These are the only centers in Wisconsin which provide interventional catheterization and surgical treatment for CCHD. A prior analysis of Wisconsin births suggested that 13.6% of critically ill neonates would be transferred out of state for continuing care,24 primarily to Minnesota. State death records and hospital discharge records were also reviewed to identify any babies with CCHD that might have otherwise been missed. This information was combined with the information reported by participants to maximize ascertainment of infants with CCHD. Statistical analysis: Categorical data were summarized in terms of frequencies and percentages. Data measured on a continuous scale were summarized using means +/- standard deviations. Chi-square or Fisher's exact test was used to compare categorical subjects’ characteristics between cohorts (Plain community vs. Non-Plain community). The nonparametric Wilcoxon rank sum test was utilized to compare maternal age between cohorts. Sensitivity, specificity, positive predictive value (PPV) and negative predictive value for CCHD screening were calculated and reported along with the corresponding 95% confidence intervals. All P-values are two-sided and P<0.05 was used to determine statistical significance. Data analysis was conducted using SAS software (SAS Institute Inc., Cary NC) version 9.4. RESULTS Demographics According to the Wisconsin Department of Health Services, there were 130,756 births reported on blood cards in the state in 2013 and 2014. There were a total of 2,753 OOH births from 2013 – 2014, representing 2.1% percent of all births. The number of reported OOH births increased from 1,297 (1.93%) in 2013 to 1,456 (2.19%) in 2014. Detailed information on POS was available from the SHINE Project on 1,616/2,753 (58.7% percent) of 2013 and 2014 OOH births. Of the 1,616 infants, there were 842 boys and 774 girls. 799 were from the Plain community, 775 were English, and in 42 the baby's background was not reported. There were a number of differences between the Plain and English populations. Prenatal ultrasound was performed in 557 English infants (71.9%) but in only 250 (31.3%) of Plain infants (p <0.0001). Notably, many of the ultrasounds in the Plain community were limited to assessments for gestational age and fetal position with no intent to screen for congenital heart defects. The average maternal age in the English population was 30.9 +/- 5.0 years and 29.8 +/- 6.3 years in the Plain population (p= 0.0003). Plain infants were screened later than English infants. 229 Plain infants (28.7%) had POS at > 48 hours, compared to 42 (5.4%) of the English infants (p <0.0001). Age at screening was not reported for 26 Plain infants, 9 English infants, and 6 infants whose background was unknown. Screening was declined in 14/799 (1.8%) of Plain births and 2/775 (0.3%) English births (p=0.0069). As outreach to the Plain community increased, births to Plain families in the study exceeded those of English families. In 2013, Plain births represented 203/503 (40.4%) of births evaluated. This increased to 596/1113 (53.5%) in 2014 (Table 1). Screening Results Of the 1,616 babies, 1584 passed, 16 failed, and 16 weren't screened. The sensitivity of the screening for CCHD was 60% (95% CI: 23-88%), with a specificity of 99.2% (95% CI: 98.4-99.4%). The positive predictive value was 18.8% (95% CI: 7-43%) and negative predictive value was 99.9% (95% CI: 99.3-99.9%) (Table 2). There were significant differences in the results of screening between the Plain and English populations. 773/799 of the Plain infants (97%) passed, compared to 770/775 (99%) of the English infants (p= 0.0004). Of the 12 Plain infants who failed their POS, 3 were found to have CCHD. These infants were diagnosed with 1) type 1 tricuspid atresia, 2) type 2 tricuspid atresia with an interrupted aortic arch, and 3) double inlet left ventricle with transposition and coarctation. There were 2 false negatives in the Plain population, one infant had an isolated coarctation of the aorta and the other had a coarctation of the aorta and ventricular septal defect. All five infants with CCHD in the study population were Amish, none of which had EVCS. Two Plain infants with significant congenital heart disease were identified. One baby with EVCS and an unbalanced atrioventricular canal failed their pulse oximetry screening and one baby with heterotaxy and severe pulmonary valve stenosis passed their pulse oximetry screening. The POS results of these seven infants are given in Table 3. Of the 3 English infants who failed their POS, none had CCHD, but two had sepsis. In these two babies, failed POS prompted early diagnosis and treatment. One of the babies in whom Plain status was unknown failed, but did not have CCHD. Review of CCHD admissions, hospitalization and death records, identified no babies with CCHD in the 1,137 home births that were not part of the SHINE project. DISCUSSION Our study demonstrates that POS screening can successfully be implemented outside of a hospital setting, with 58.7% of all OOH births in Wisconsin participating in POS screening as part of this study from 2013 – 2014 despite a rolling enrollment through the study period. The sensitivity, specificity, PPV, and NPV of POS screening in our study are similar to those reported in hospital born infants (Table 4). 11, 12, 13 Both infants with coarctation of the aorta in this cohort passed their POS. Prior studies of hospital born babies have also shown low sensitivity for coarctation of the aorta ranging from 30-43%.11, 12, 13 In this study, there was a high burden of CCHD in the OOH birth community, which appears to be was borne primarily by the Plain community. Our ability to fully assess the burden of CCHD in the OOH and Plain clothes communities is limited by the nonuniform recruitment of midwives and other OOH providers and our inability to determine whether births outside the SHINE Project were from Plain or English families. This is the first large study of POS that includes Plain births. Although a higher incidence of congenital heart disease is often assumed in Amish and other Plain communities, there is no published literature on the incidence of CCHD in the Plain community. Although an increased incidence of congenital heart disease in Plain communities is often attributed in part to EVCS, none of the infants with CCHD in this study carried this diagnosis. As those forms of congenital heart disease beyond the twelve CCHD diagnoses were not systematically recorded, their incidence cannot be evaluated by this study. This study also demonstrated an increasing number of Plain births. This may be due in part to increased reporting of Plain births as a consequence of the increased outreach to the community. However, an increasing Plain population is consistent with anecdotal evidence and the experience of clinicians in the state of Wisconsin, suggesting that the Plain population may in fact be increasing. Pulse oximetry screening detects more infants in settings with a lower prenatal diagnosis rate. 11, 13 In the OOH birth population, prenatal ultrasounds were performed in 71.9% of English women and only 31.3% of Plain women. Thus, in this population of women with limited prenatal screening, POS becomes even more useful and clinically significant. POS may also be more palatable for patients who decline prenatal ultrasounds and other testing, as it is minimally invasive and inexpensive. There was wide variation in the reported time of POS. This was particularly true for the Plain community birth attendants, with a significant portion of screening performed more than 48 hours after birth (229 Plain infants versus 42 English infants). Ideally, screening should take place after 24 hours to minimize false positive results25. The preferred time for screening is between 24 and 48 hours of life, to maximize sensitivity while allowing early detection and intervention prior to symptoms. Delayed POS in Plain infants puts them at greater risk for clinical complications of undetected CCHD. We hypothesize that there is greater practice variation in Plain community birth attendants, resulting in greater variation in data collection in the Plain community. Additionally, some very conservative Plain families may be less likely to allow midwives and nurses from outside the community into their homes, and often decline to have postpartum visits on Sundays creating additional barriers for POS screening and less precisely timed screening. CONCLUSION Pulse oximetry screening is of particularly high value in screening for CCHD in high risk populations such as the Plain community and can be effectively introduced into the care of babies born outside a hospital setting. Supplementary Material 1 2 Acknowledgements This project was supported by the Health Resources and Services Administration (HRSA) of the U.S. Department of Health and Human Services (HHS) under grant number H46Mc24057, Critical Congenital Heart Disease Newborn Screening Demonstration Program. Alyssa Yang's participation was supported in part by an appointment to the Applied Epidemiology Fellowship Program administered by the Council of State and Territorial Epidemiologists (CSTE) and funded by the Centers for Disease Control and Prevention (CDC) 1U380T000143-01. This study was supported in part by an appointment to the Applied Epidemiology Fellowship Program administered by the Council of State and Territorial Epidemiologists (CSTE) and funded by the Centers for Disease Control and Prevention (CDC) Cooperative Agreement Number 1U380T000143-01. Plain English Not Specificed Total 2013 203 286 14 503 2014 596 489 28 1113 Total 799 775 42 1616 CCHD No CCHD Total Fail 3 13 16 PPV 18.8% Pass 2 1582 1584 NPV 99.9% Total 5 1595 1600* Sensitivity Specificity 60% 99.2% *16 Refused POS Measurements Age at POS Critical Congenital Heart Disease Type 1 Tricuspid Atresia 87/87 45 hours Type 2 Tricuspid Atresia, IAA 91/92, 90/91, 91/93 24 hours DILV, D-TGA, Coarctation 88/84 24 hours Coarctation 95/95 >48 hours Coarctation, VSD 96/96 >48 hours Significant Congenital Heart Disease Unbalanced Atrioventricular Canal 86/84 8 hours Heterotaxy, Severe Pulmonary Stenosis 96/96 >48 hours SHINE Zhao Ewer DeWahl US China UK Sweden Year 2016 2014 2011 2009 Sensitivity 60% 58.70% 49.06% 62.07% Specificity 99.18% 99.70% 99.16% 99.82% PPV 18.75% 35.90% 13.33% 20.69% NPV 99.87% 99.89% 99.86% 99.97% Failure Rate 1.00% 0.43% 0.97% 0.22% FP Rate 0.81% 0.25% 0.81% 0.18% The first draft of the manuscript was written by Kathleen Miller, MD, pediatrics resident at the University of Wisconsin. No honorarium, grant, or other form of payment was given to anyone to produce the manuscript. This project was made possible by the efforts of many contributors. The authors are grateful to the families who participated and to the Wisconsin Guild of Midwives for their support. Special thanks are extended to the following individuals: Marijke van Roojen President, Wisconsin Guild of Midwives,Midwifery Clinical Education Coordinator, Southwest Wisconsin Technical College Jody Belling Research Nurse Coordinator, Department of Pediatrics, University of Wisconsin School of Medicine and Public Health Jens C Eickhoff Senior Scientist, Department of Biostatistics and Medical Informatics, University of Wisconsin School of Medicine and Public Health Sharon Fleischfresser Medical Director, Children and Youth with Special Health Care Needs, Wisconsin Department of Health Services Elizabeth Oftedahl Epidemiologist, Children and Youth with Special Health Care Needs, Wisconsin Department of Health Services Michael Payne Web Developer/Analyst, Information Architecture / Public Health Information Network, Wisconsin Department of Public Health Office of Health Informatics Mei Baker Co-Director, Newborn Screening Laboratory, Wisconsin State Laboratory of Hygiene Constanza Bravo University of Wisconsin, Department of Geography WORKS CITED 1 van Velzen C Clur S Rijlaarsdam M Bax C Pajkrt E Heymans M Prenatal detection of congenital heart disease-results of a national screening programme. BJOG : an international journal of obstetrics and gynaecology 2015 2 Johnson LC Lieberman E O'Leary E Geggel RL Prenatal and newborn screening for critical congenital heart disease: findings from a nursery. Pediatrics 2014 134 5 916 922 25287457 3 Evans W Castillo W Rollins R Luna C Kip K Ludwick J Moving towards universal prenatal detection of critical congenital heart disease in southern nevada: a community-wide program. Pediatr Cardiol 2015 36 2 281 288 25103855 4 Levy DJ Pretorius DH Rothman A Gonzales M Rao C Nunes ME Improved prenatal detection of congenital heart disease in an integrated health care system. Pediatr Cardiol 2013 34 3 670 679 23117330 5 Friedberg MK Silverman NH Moon-Grady AJ Tong E Nourse J Sorenson B Prenatal detection of congenital heart disease. J Pediatr 2009 155 1 26 31 31 e21 19394031 6 Peterson C Ailes E Riehle-Colarusso T Oster ME Olney RS Cassell CH Late Detection of Critical Congenital Heart Disease Among US Infants: Estimation of the Potential Impact of Proposed Universal Screening Using Pulse Oximetry. JAMA pediatrics 2014 168 4 361 370 24493342 7 Chang RK Gurvitz M Rodriguez S Missed diagnosis of critical congenital heart disease. Arch Pediatr Adolesc Med 2008 162 10 969 974 18838650 8 Aamir T Kruse L Ezeakudo O Delayed diagnosis of critical congenital cardiovascular malformations (CCVM) and pulse oximetry screening of newborns. Acta Paediatr 2007 96 8 1146 1149 17590190 9 Ng B Hokanson J Missed congenital heart disease in neonates. Congenit Heart Dis 2010 5 3 292 296 20576049 10 Brown KL Ridout DA Hoskote A Verhulst L Ricci M Bull C Delayed diagnosis of congenital heart disease worsens preoperative condition and outcome of surgery in neonates. Heart 2006 92 9 1298 1302 16449514 11 de-Wahl Granelli A Wennergren M Sandberg K Mellander M Bejlum C Inganas L Impact of pulse oximetry screening on the detection of duct dependent congenital heart disease: a Swedish prospective screening study in 39,821 newborns. BMJ 2009 338 a3037 19131383 12 Ewer AK Middleton LJ Furmston AT Bhoyar A Daniels JP Thangaratinam S Pulse oximetry screening for congenital heart defects in newborn infants (PulseOx): a test accuracy study. Lancet 2011 378 9793 785 794 21820732 13 Zhao QM Ma XJ Ge XL Liu F Yan WL Wu L Pulse oximetry with clinical assessment to screen for congenital heart disease in neonates in China: a prospective study. Lancet 2014 14 Ailes EC Gilboa SM Honein MA Oster ME Estimated number of infants detected and missed by critical congenital heart defect screening. Pediatrics 2015 135 6 1000 1008 25963011 15 Lhost JJ Goetz EM Belling JD van Roojen WM Spicer G Hokanson JS Pulse Oximetry Screening for Critical Congenital Heart Disease in Planned Out-of-Hospital Births. J Pediatr 2014 16 Kemper AR Mahle WT Martin GR Cooley WC Kumar P Morrow WR Strategies for implementing screening for critical congenital heart disease. Pediatrics 2011 128 5 e1259 1267 21987707 17 Mahle WT Martin GR Beekman RH 3rd Morrow WR Section on C, Cardiac Surgery Executive C. Endorsement of Health and Human Services recommendation for pulse oximetry screening for critical congenital heart disease. Pediatrics 2012 129 1 190 192 22201143 18 Watterberg KL Committee on F Newborn. Policy statement on planned home birth: upholding the best interests of children and families. Pediatrics 2013 132 5 924 926 24144706 19 McKusick VA Egeland JA Eldridge R Krusen DE Dwarfism in the Amish I. The Ellis-Van Creveld Syndrome. Bulletin of the Johns Hopkins Hospital 1964 115 306 336 14217223 20 O'Connor MJ Rider NL Thomas Collins R Hanna BD Holmes Morton D Strauss KA Contemporary management of congenital malformations of the heart in infants with Ellis - van Creveld syndrome: a report of nine cases. Cardiology in the young 2011 21 2 145 152 21070693 21 Hills CB Kochilas L Schimmenti LA Moller JH Ellis-van Creveld syndrome and congenital heart defects: presentation of an additional 32 cases. Pediatr Cardiol 2011 32 7 977 982 21533779 22 O'Connor MJ Collins RT 2nd. Ellis-van Creveld syndrome and congenital heart defects: presentation of an additional 32 cases. Pediatr Cardiol 2012 33 4 491 discussion 491-492 22286269 23 Mahle WT Newburger JW Matherne GP Smith FC Hoke TR Koppel R Role of pulse oximetry in examining newborns for congenital heart disease: a scientific statement from the AHA and AAP. Pediatrics 2009 124 2 823 836 19581259 24 Beissel DJ GEM, Hokanson J.S. Pulse Oximetry Screening For Congenital Heart Disease in Wisconsin. Congenit Heart Dis 2011 6 5 521 522 25 Mahle WT Martin GR Beekman RH 3rd Morrow WR Rosenthal GL Snyder CS Endorsement of health and human services recommendation for pulse oximetry screening for critical congenital heart disease. Pediatrics 2012 129 1 190 192 22201143

PMC005xxxxxx/PMC5130882.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 8406899 7945 Vaccine Vaccine Vaccine 0264-410X 1873-2518 27317264 5130882 10.1016/j.vaccine.2016.02.015 HHSPA821075 Article Adventitious Agents and Live Viral Vectored Vaccines: Considerations for Archiving Samples of Biological Materials for Retrospective Analysis Klug Bettina a Robertson James S. b Condit Richard C. c Seligman Stephen J. d Laderoute Marian P. e Sheets Rebecca f Williamson Anna-Lise g Chapman Louisa h Carbery Baevin h Mac Lisa M. h Chen Robert T. h For The Brighton Collaboration Viral Vaccine Vector Safety Working Group a Divison Immunology Paul-Ehrlich-Institut, D-63225 Langen, Germany b Independent Adviser (formerly of National Institute for Biological Standards and Control, Potters Bar, EN6 3QG, UK) c Department of Molecular Genetics & Microbiology, University of Florida, Gainesville, FL 32610 d Department of Microbiology and Immunology, New York Medical College Valhalla, NY 10595, USA e Immune System Management Inc., Ottawa, Ontario, Canada, K1S 5R5 (formerly of Blood Safety Contribution Program, Public Health Agency of Canada, Ottawa, Ontario, Canada, K1A 0K9 f Independent Adviser (formerly of NIAID, NIH, Bethesda, MD 20893, USA) g Institute of Infectious Disease and Molecular Medicine, University of Cape Town and National Health Laboratory Service, Cape Town, South Africa h Corresponding author: Secretariat, brightoncollaborationv3swg@gmail.com 6 10 2016 15 6 2016 12 12 2016 12 12 2016 34 51 66176625 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 1. The need for archiving vaccine samples and other biological materials Vaccines are one of the most effective public health medicinal products with an excellent safety record. Well-planned and implemented immunization programs have profoundly reduced the morbidity and mortality of targeted diseases [1], such as the global eradication of smallpox [2] and the elimination of poliomyelitis [3] and measles [4] from many regions of the world. Since vaccines are usually administered to large populations of healthy people including children, frequently with the goal of near universal coverage (under legal mandate in some countries), their safety and quality are paramount for public health. As vaccines are produced using biological materials, there is a need to safeguard against potential contamination with adventitious agents. Adventitious agents are defined by the World Health Organization (WHO) as microorganisms that may have been unintentionally introduced into the manufacturing process of a biological medicinal product [5]: these include bacteria, fungi, mycoplasma/spiroplasma, mycobacteria, rickettsia, protozoa, parasites, transmissible spongiform encephalopathy (TSE) agents and viruses. Adventitious agents could be inadvertently introduced into a vaccine through starting materials used for production, such as cell substrates, porcine trypsin, bovine serum, or any other source materials of animal or human origin [6]. Therefore, extensive testing is recommended at various stages during vaccine manufacture to demonstrate the absence of adventitious agents [5]. Additionally, the incorporation of viral clearance steps in the manufacturing process, which evaluate the capability of the manufacturing production process to inactivate and/or remove potential viral contaminants [7] can aid in reducing the risk of adventitious agent contamination in a biological product; however, for live viral vaccines, aside from possible purification of the virus or vector, extensive adventitious agent clearance may not be feasible. Hence, the issue of unknown contamination risks of live or vectored vaccines requires more stringent safety oversight [5]. In the event that an adventitious agent is detected in a current vaccine, it is important to determine its origin, evaluate its potential for human infection, and discern which batches of vaccine may have been affected for notification and in order to take risk management action plans. To achieve this, it is necessary to have archived samples of the vaccine and ancillary components, ideally from developmental through to current batches, as well as samples of the biological materials used in the manufacture of the vaccine, since these are the most likely sources of an adventitious agent. Although currently recommended testing has a good record for demonstrating absence of adventitious agents in vaccines, there have been rare cases of adventitious agent detection in some licensed vaccines. A recent notable event was that of porcine circovirus 1 (PCV1) in a rotavirus vaccine [8–10]. Early episodes of contamination of biologicals (e.g., tetanus contamination of diphtheria anti-toxin) date back to the beginning of modern immunization and led to the establishment of regulatory oversight in the early 1900’s [11]. The discovery that early polio vaccine was contaminated by simian virus 40 (SV40) due to infection of rhesus monkeys resulted in a major manufacturing change in the cell substrate from primary rhesus monkey kidney cells to African Green monkey kidney cultures [12]. The detection of bacteriophage was detected in measles and polio vaccines, reverse transcriptase in measles and mumps vaccines, and the emergence of bovine spongiform encephalopathy (BSE) commonly known as “mad cow disease” in the 1980’s, and ultimately the human version variant Creutzfeldt-Jakob disease (vCJD) in the 1990’s, led to considerable regulatory deliberations, and also guidance on the use of bovine (and other) materials that could transmit transmissible spongiform encephalopathies (TSE’s) [12–15]. Viral vaccines are grown in cell cultures that may have been propagated in media containing bovine serum, and possibly used porcine trypsin for cell passage. Thus, in addition to archiving final released vaccine, there is also an argument for archiving the starting biological materials and of records that provide full traceability of biological materials used in vaccine manufacture. However, the issue of cell and serum archiving and their full traceability are not within the scope of this document at this point in time, and this paper will focus on the live recombinant viral vectored vaccine itself. Laboratory testing is used to demonstrate the absence of adventitious agents in the vaccine. In the event that contamination is found in a released vaccine after it has been marketed, samples obtained from vaccinees (e.g serum and PBMCs) may be used to evaluate whether the adventitious agent infected the vaccine recipient. Retrospective testing confirmed the presence of PCV1 DNA in Rotarix® since the initial stages of its development and in vaccine lots used in clinical studies conducted pre- and post-licensure [10]. Therefore, adventitious agents that fail detection using technologies available at the time a vaccine was originally produced and used, might at a later stage be detected by re-testing using emerging technology. In order for a new technology to be utilized to improve vaccine safety and detect past contamination events, samples of the vaccines and materials used in their production and samples from the vaccine recipients need to be collected and archived. Hitherto, the need for formal guidance on such vaccine sample archiving has been recognized but not fulfilled [15]. The Brighton Collaboration Viral Vector Vaccine Safety Working Group, formed in 2008 with voluntary representatives from academia, government and industry [16], has therefore summarized in this paper several prior major cases of vaccine contamination and provides points for consideration on sample archiving of live recombinant viral vector vaccines in humans. The Group recognizes that this document may be controversial, especially the cost implications, but feel it is important to stimulate the discussion on both the need for vaccine sample archiving and how this need might be met. While this document focuses on live viral vector vaccines, relevant past experience with traditional viral vaccines are discussed and the lessons learnt may be usefully applied to novel vaccines, especially those that are live attenuated. 2. Historical context: past to future History has shown that extensive testing for adventitious agents during manufacture of vaccines has prevented major contamination events and potential adverse clinical consequences. However, reports of product contamination have occurred periodically, mostly due to viruses present in biological reagents used for production (e.g. animal tissues or primary cell substrates, serum, or trypsin). The genomic and biotechnology revolution of the last decades has enabled the development, licensure, and production of many new vaccines and biologicals. The number of vaccine manufacturers who supply the global market has also been increasing, many of whom are from emerging economies [17]. While all vaccine manufacturers are regulated by their national health authorities, and those who supply UNICEF are pre-qualified by the WHO as meeting good manufacturing practice (GMP) standards [18], their capabilities differ and many need improved pharmacovigilance systems, such as standardization of safety reporting [19]. Since many if not most vaccines globally will likely continue to be made using biological reagents for the foreseeable future, the possibility of adventitious contamination cannot be totally excluded. Therefore, it is important to consider prospective sample archiving of vaccines and the use of new technologies and knowledge to test for contamination as they become available. To provide the context for and to better illustrate the need for this document, we have reviewed several notable contamination events and the resulting corrective regulatory actions. These events have also been reviewed in detail elsewhere. [12, 20, 21] 2.1 SV40 contamination of polio vaccine 2.1.1 Discovery of SV40 contamination in polio vaccine SV40 is a monkey polyoma virus that was discovered in 1960 and can induce tumors in rodents and transform human cells in culture [22]. The Salk inactivated polio vaccine (IPV), first licensed in 1955 in the U.S.A., was made in primary rhesus monkey kidney cells. It was already in wide use in 1961 when it was discovered that some of the vaccine lots were contaminated with SV40. At least 10–30 million persons were estimated to have been exposed to SV40-contaminated polio vaccine in the U.S.A. [23]. Testing of stored U.S. samples from vaccine lots produced in 1955 showed that the levels of SV40 were inconsistent across vaccine lots with some uncontaminated lots [24]. However, as samples of vaccine lots produced were not archived during 1955–1961, the period of likely SV40 contamination, no further testing was possible. Since the early 1960’s, polio vaccines have been tested for SV40 infectivity in cell cultures. In a retrospective UK study, PCR was used to examine archived samples of oral polio vaccines (OPV) dating from 1966 to the time of the study (1999), including all vaccines used in the UK since 1980, for the presence of SV40 sequences [25, 26]. Of 132 materials examined, 118 were negative on initial testing and fourteen gave reactions which on further examination were attributed either to cross contamination during handling in the laboratory at the National Institute for Biological Standards and Control (NIBSC), UK or to non-specific amplification; it was concluded that none of the samples contained SV40 sequences [25]. Some polio vaccines prepared from 1954 to 1961 were contaminated with infectious SV40. It has been assumed that all polio vaccines were SV40 free in the United States after 1961 and in other countries after 1962. Following a WHO requirement [27] that was prompted by the detection of SV40 in some human tumors, [28] a multilaboratory study was conducted to test SV40 polio vaccines prepared after 1961. Vaccine samples from 13 countries and the WHO seed were tested. All vaccines were SV40 free, except for vaccine from a major eastern European manufacturer whose inactivation procedure failed to completely inactivate SV40 in OPV vaccine seed stocks [29]. In Sweden, US-produced polio vaccine was used in 1957; but from 1958 on, only Swedish produced vaccine was used. Testing for SV40 began in 1961, including retrospectively of vaccines produced earlier, but there is doubt as to the validity of the negative results [25, 26]. Multiple epidemiologic studies have been conducted to assess the long term effects of SV40 contaminated vaccine in humans [15]. More recently, there was concern that although SV40 infection alone is unlikely to cause mesotheliomas and brain tumors in which SV40 genetic sequences had purportedly been detected, it may have acted as a cofactor in the pathogenesis of some tumors, with co-carcinogenicity between SV40 and asbestos being of particular concern [30]. However, in an extensive review by the US Institute of Medicine (IOM) in 2002, it was concluded that these studies were “sufficiently flawed” so there was insufficient evidence to determine whether SV40-contaminated polio vaccine caused cancer or not [15]. 2.1.2 Regulatory actions taken after SV40 contamination of poliovirus In 1989, the WHO developed guidelines that required monkeys to be free of SV40, a practice already implemented in many countries. Validated nucleic acid amplification tests are generally now used to determine that virus seed lots used to produce viral vaccines are specifically free of SV40, along with a tissue culture test in Vero cells [5, 31]. Worldwide, manufacture of the vaccine was changed to African green monkey kidney cells, since this species is generally free of SV40. Authorities worldwide require all licensed vaccines to fulfill general safety, sterility, and purity requirements [32]. The 2002 IOM report recommended that federal agencies develop a ‘Vaccine Contamination Prevention and Response Plan’ which would include “strategies for routine assessment of vaccine for possible contamination; notification of public health officials, health care providers and the public if contamination occurs; identification of recipients of contaminated vaccines; and surveillance and research to assess health outcomes associated with the contamination” [15]. It also recommended considering a program to store samples from each vaccine lot approved for release in order to make it possible to test for contaminants if new detection methods become available or safety questions arise well after the vaccine has been used. Currently, manufacturers are required to store samples of each released lot only until one year following the expiration of that lot [32, 33]. 2.2 Contamination of yellow fever vaccine 2.2.1 Avian retrovirus contamination of yellow fever vaccine Avian leucosis virus (ALV) is an exogenous retrovirus that causes leukemia in chickens by means of insertional activation of cellular oncogenes [34]. The yellow fever (YF) vaccine comprises the 17D attenuated strain and is propagated in chicken embryos by inoculation of 7 to 9 day old embryonated eggs with the vaccine strain. The 17D YF vaccine became the main means of protection for travelers and those in the military during World War II [35] and was received by over one third of the US Army [36]. ALV contamination of the YF vaccine was first discovered in 1966 and concern arose about the possible oncogenic risk among former military vaccinees [36]. Waters et al. conducted a retrospective case control study, examining record-documented YF vaccination history during World War II among representative sample of 2,659 veterans who died of various specific cancers between 1950–1954 or 1959–1963 and age-matched controls [36]. The study found no suggestion of association between the vaccine and cancers as classified, despite good statistical power. However, this study could only examine cancers with a latent period between 5 and 22 years, and failed to detect any elevated risk of hepatic neoplasia among vaccinees with prior history of serum hepatitis (see 2.3.2). More recently, YF vaccines produced by three manufacturers were all found to have endogenous avian retrovirus (EAV) particles and endogenous avian leucosis virus (ALV-E) particles, which originate from ancient retroviral sequences and from a nonpathogenic ALV, respectively, that exist as a normal part of the chicken genome (discussed in 2.4 below). The absence of evidence of infection with ALV-E or EAV in 43 YF vaccine recipients suggests a low risk, if any, for transmission of these viruses [37]. 2.2.2 Hepatitis B virus (HBV) contamination of yellow fever vaccine An epidemic of icteric hepatitis in 1942 affected approximately 330,000 U.S. Army personnel. This outbreak was linked to specific lots of YF 17D vaccine stabilized with human serum that retrospectively was found likely to have been contaminated by HBV [35, 38, 39]. The outbreak was controlled by shifting to a new serum-free YF vaccine. However, the link between the hepatitis and the YF vaccine was not proven until a 1985 study in which 597 veterans who had been in the army in 1942 were interviewed and serologically screened. They were categorized in three groups: the first group included patients who had jaundice after having received the vaccine, and 97% of them were positive for antibodies to HBV. The second group contained those who had received the vaccine but did not fall ill, 76% of whom had positive HBV antibodies. The third group consisted of persons who received a serum-free vaccine and did not have jaundice; 13% of them had positive antibodies to HBV, similar to the prevalence in the general US population [35, 38, 39]. Together these results suggested that the YF vaccine transmitted HBV. 2.2.3 Regulation resulting from ALV and HBV contamination of YF vaccine Extensive testing is recommended to assure vaccine safety; only a few cases of unexpected viruses have occurred but they highlight the importance of adventitious agent testing for all biological materials that are used for vaccine production. Although there is no evidence for human disease associated with ALV, all countries now use seed virus prepared in specific-pathogen free (SPF) eggs that are free from ALV as indicated by WHO [40, 41]. However, some permit the production of vaccine in embryonated chicken eggs that may contain ALV, but need justification due to cost and difficulty in procuring ALV-free eggs that would result in restricting availability of the YF vaccine. For this reason, the revised WHO Requirements for YF vaccine do not require ALV-free eggs. It should be noted that the WHO requirements regarding YF vaccines are not mandatory and approval for use is controlled by individual nations [38, 40, 42]. Accordingly, the vaccines, particularly with respect to their quality control, can vary. HBV contamination in the early lots of YF 17D vaccine due to pooled human serum that was used as a stabilizer resulted in the elimination of human serum from YF vaccines. 2.3 Endogenous avian retroviral particles in MMR vaccines In 1996, reverse transcriptase (RTase) activity, an enzyme typically associated with retroviruses, was detected in chicken cell-derived measles and mumps vaccines [43]. The RT activity was found to originate from the chicken embryonic fibroblasts used as a substrate for vaccine manufacture and was associated with virus-like particles containing endogenous retrovirus sequences (EAV) that are normally present in the host genome. Infectivity studies demonstrated these particles were non-infectious in a variety of human cell lines [42, 44, 45]. Although EAV and also endogenous avian leukosis virus (ALV-E) RNA sequences were reported in MMR vaccines, there was no evidence of transmissibility of ALV and EAV sequences to MMR recipients [40]. Pre- and post MMR vaccination samples from 33 children as well as samples from randomly selected blood donors were tested for ALV and EAV sequences. Despite the use of a highly sensitive PCR assay none of the samples tested were positive for either ALV or EAV sequences [40]. Various studies did not reveal any adverse effects of the presence of these sequences or of RTase activity in chicken cell derived vaccines and the WHO determined that the overall benefit/risk balance remains highly in favor of continued use of the vaccines [12, 46]. 2.4 PCV-1 contamination of vaccine Porcine circoviruses (PCVs) are small non-enveloped virus containing a single-strand circular DNA genome virus. Two antigenically and genomically distinct variants exist in the swine population worldwide: PCV1 is non-pathogenic for pigs; PCV2 has been associated with various porcine disease syndromes [47]. PCV contamination of a vaccine was first discovered by Victoria et al. [8], while experimenting with new methods for detecting adventitious viral contamination. Using metagenomics and a pan-microbial microarray (versus a more traditional method of viral species-specific PCR), a panel of eight live attenuated vaccines that included oral polio virus, rubella, measles, yellow fever, human herpesvirus 3 (HHV-3), rotavirus, and multivalent measles/mumps/rubella were analyzed. In one orally administered rotavirus vaccine the metagenomics study uncovered a complete porcine circovirus-1 (PCV1) genetic sequence. Follow-up studies indicated that the number of PCV1 viral particles present in the vaccine was about the same as the number of rotavirus vaccine particles [48]. The contaminant was subsequently easily detected by virus-specific PCR; this had never been previously applied, because this agent, not being of concern to the swine industry, was not specifically included in the testing recommended for porcine viruses [49] and in tests recommended for extraneous agents [50]. No other microbial genetic sequences were detected in the study, that had not been previously uncovered in any of the vaccines. In cell cultures, although PCV gene expression and replication takes place in human cells, the infection is non-productive [9]. Furthermore, PCR screening of a variety of different human cell lines, including human tumor cells, demonstrated that PCV1 was not generally prevalent in commercially-available cell lines [8]. Epidemiological data for humans show ambivalent results for serum antibody to PCV1 and current PCV1 knowledge is sparse and contradictory [51, 52]. 2.4.1 PCV-1 in Rotarix® The rotavirus vaccine contaminated with PCV1 described above was Rotarix®, an oral rotavirus vaccine manufactured by GlaxoSmithKline (GSK), first licensed in Europe in 2006 and US in 2008. Two doses of the vaccine are given to infants beginning at six weeks of age to protect against gastroenteritis due to rotavirus infection. The WHO estimates that rotaviruses are responsible for approximately 500,000 deaths each year, with more than 85% occurring in low-income countries in Africa and Asia. Upon being informed of the PCV1 contamination of Rotarix®, GSK rapidly initiated an investigation to confirm the source, nature and amount of PCV1 in the vaccine manufacturing process and to assess potential clinical implications of the finding. The investigation also considered their inactivated poliovirus (IPV)-containing vaccines, since poliovirus vaccine strains are propagated using the same cell line as the rotavirus vaccine strain. Results confirmed the presence of PCV1 DNA and low levels of PCV1 viral particles at all stages of the Rotarix® manufacturing process. PCV1 DNA was not detected in the IPV-containing vaccine manufacturing process beyond the purification stage. GSK subsequently notified regulatory health authorities about the discovery of PCV1 in Rotarix® and conducted additional studies confirming that PCV1 DNA was present in both the finished Rotarix® vaccine, in vaccine lots used in clinical studies, and in the source cell bank and master seed; the latter findings suggesting that the PCV1 contamination occurred during the early stages of vaccine development [10, 53]. The contamination was believed to have derived from the use of contaminated porcine trypsin in the development and manufacture of the vaccine. Rotarix® is widely used globally in both developed and less developed settings. At the time of discovery of PCV1 contamination, ~100,000 children had received the vaccine during clinical trials and ~68 million doses had been distributed worldwide. Therefore, due to the potential public health impact, regulatory agencies further examined the state of the contamination. 2.4.2 Regulatory Actions Taken for PCV1 Contamination of Rotarix® 2.4.2.1 European Union In the European Union, Rotarix® is available in all Member States, but is usually not part of their routine childhood vaccination schedules. After GSK notified the European Medicines Agency of the unexpected presence of PCV1 DNA in batches of Rotarix® in March 2010, its Committee for Medicinal Products for Human Use (CHMP) initiated a review. In view of the ubiquitous presence of the virus in food, the oral route of administration of the vaccine (mimicking the route of natural exposure), and the absence of both known pathogenicity and serious adverse reactions reported with the vaccine, the Committee concluded in March 2010 that the findings do not present a public health threat and vaccine usage should continue [52]. A formal review of Rotarix® was also initiated by the European Commission, which concluded that the vaccine continues to have a positive benefit-risk balance and that the presence of a small amount of PCV1 viral particles does not present a risk to public health. However, since PCV-1 should not be present in the vaccine, it was incumbent upon the manufacturer to propose measures of manufacturing the vaccine free of the virus, although such measures would take time to implement [53]. 2.4.2.2 United States- FDA An initial review of data on the presence of DNA from PCV1 in Rotarix® was performed in March 2010. The FDA similarly concluded that there was no evidence that the presence of PCV1 DNA in Rotarix® posed a safety risk and confirmed the excellent safety record of the vaccine [10]. Nevertheless, the FDA recommended that clinicians temporarily suspend use of vaccine until the Agency learned more. On 14 May 2010, after discussions in the FDA Vaccines and Related Biological Products Advisory Committee, the suspension of the use of Rotarix® was removed. The decision was based on an evaluation of information from laboratory results from the manufacturer and the FDA’s own laboratories, a thorough review of the scientific literature, and input from scientific and public health experts, including members of the FDA’s Vaccines and Related Biological Products Advisory Committee that convened on May 7, 2010 to discuss these vaccines. The Agency’s decision was further based on the strong safety records of the vaccine, the lack of evidence that PCV1 or PCV2 cause infection or disease in humans, and the substantial benefit of the vaccine in preventing death in some parts of the world and hospitalization for severe rotavirus disease in the United States. These benefits outweighed the theoretical risk posed by the presence of PCV1 in the vaccine [54]. Since the investigation into the PCV1 contamination of Rotarix® by GSK and federal agencies, PCV1 has continued to be researched and manufacturing procedures have been further developed. It was found that PCV1 could infect human hepatocellular carcinoma cells. Although the author emphasizes that the connection between this evidence and vaccine safety is unclear, it does demonstrate that a negative cell culture may not give the full scope of the contaminant’s capabilities [54]. The presence of PCV1 early in the vaccine production process has also triggered further research on contaminants in cell culture, and material used in the manufacturing process such as bovine serum and trypsin [55]. Furthermore, research is being done to improve the manufacturing procedure by creating a new quantitative tool to detect residual porcine DNA [56]. No PCV1 DNA was detected in a separate and widely used rotavirus vaccine, Rotateq™, manufactured by Merck, although sensitive assays detected small fragments of PCV2 genomic DNA. It was determined that these were of no consequence to the safety of the vaccine and no regulatory action was taken. 3. Existing guidelines to assure viral safety Strict measures are currently in place to assure the safety of vaccines as well as all other biological medicines. For example, the US Code of Federal Regulations defines product safety as “the relative freedom from harmful effect to persons affected, directly or indirectly, by a product when prudently administered, taking into consideration the character of the product in relation to the condition of the recipient at the time [32]”. The two critical components of safety are sterility, which is defined in 21CFR600.3(q) as “freedom from viable contaminating microorganisms”[32], and purity which is defined in 21CFR600.3(r) as the “relative freedom from extraneous matter in the finished product, whether or not harmful to the recipient or deleterious to the product to meet the requirements of 610.13” [55]. Vaccines are currently tested using a variety of assays to demonstrate safety and purity, including specific and general assays for detection of potential contaminants, and there may be a need to consider new technologies that become available for broad detection of unknown agents as well. Nevertheless, while materials and culture processes leading to medicinal products are tested to demonstrate the absence of adventitious agents, there might be occasional unintended introduction as demonstrated by the PCV1 situation. Reports indicate that adventitious agent contaminations are more frequently caused by bacteria or mycoplasmas, which are more easily detected, than by a virus. The safeguards against viral contamination include implementation of GMP, thorough testing or use of certified raw materials, viral safety evaluation at critical production stages (e.g. virus seeds and virus harvests) and validation of the viral clearance capacity (if any) of the downstream purification process [57]. National and international regulatory authorities provide guidelines on the manufacturing, standardization and quality control of medicines [32]. These are subject to continuous review and modification to reflect the current state of science and technology. However, for live viral vaccines, in-process adventitious agent inactivation steps are not part of the manufacturing process, since these steps would most likely compromise vaccine viability and immunogenicity. While inactivated vaccines include a vaccine virus inactivation step as part of the manufacturing process, the ability of that step to inactivate potential adventitious agents is often not evaluated, particularly for products that have been on the market for some time. For newer or investigational vaccines, an inactivation step(s) that assesses the ability to inactivate a variety of agents should be part of their manufacture. The safety of live viral vaccines has to be assured by direct testing of the vaccine and of materials used in its manufacture, and to use control cell cultures for demonstrating that batches of cells or eggs cultivated in parallel to those used in vaccine manufacture but not infected with the vaccine virus, show no signs of infection by other agents. It is important to include a risk assessment process in the overall viral control strategy used during the manufacture and testing of vaccines. The risk assessment is necessary to identify potential sources for entry of adventitious agents into the vaccine, and to develop a strategy to mitigate the risk of adventitious agent introduction. The risk assessment can be used to tailor the biosafety testing that is performed on raw materials, vaccine seeds, vaccine bulk materials and final product [58]. This is an evolving field and regulatory agencies are developing regulations regarding using new detection technologies to evaluate future vaccines for adventitious agents. 4. Methods for Developing Proposed Considerations A Brighton Collaboration Viral Vaccine Vector Safety Working Group was formed in 2008 with about 30 expert members. The group consists of persons with expertise in virology, regulation and vaccine safety, and meets via monthly conference calls. The development of this considerations paper was based upon literature review, a systematic review of current regulations from both Europe and United States and group consensus. Outside experts on sample archiving were invited to contribute as needed. 4.1 Lessons Learned from Past Contamination Events Contamination events have invoked considerable discussions in industry and regulatory agencies leading potentially to implementation of risk mitigation strategies or formulation of new recommendations [59]. However, the issue of comprehensive storage or archiving of vaccine samples so that the origin of any future contamination event can more easily be traced and corrective action taken has not been addressed. As the past examples of contamination events demonstrate, it is important to archive samples consistently for an extended period of time so as to allow future researchers to determine the extent and impact of any contamination. Some potential adverse events induced by an adventitious agent (e.g. cancer) can occur many years after vaccination and the period of archiving should reflect this scenario. The ALV contamination of the yellow fever vaccine was examined using available information on vaccinations, which came from a cohort that was not likely to be vulnerable to infection or to be immunocompromised, and so is not easily generalizable. With the HBV contamination of the yellow fever vaccine, there was a problem with obtaining historical samples and it was a challenge to use these samples due to the lack of guidance for sample archiving at the time they were prepared for storage. As occurred during the investigation of possible adverse events resulting from SV40 contamination of the polio vaccine, existing samples were not representative of the distribution of the vaccine and epidemiologic studies able to be performed with existing samples were flawed, preventing a concrete conclusion [15]. The existing samples were also precious, which led to problems of establishing acceptable protocols to extract DNA, and difficulties may have resulted in some initial cross contamination. The extended storage of vaccine samples would assist future researchers to identify contaminated vaccine lots, and so determine a more accurate relative risk for specific populations. The value of a centrally organized sample archive was illustrated during a relatively recent investigation of 1976–77 swine influenza vaccine to assess if the still unexplained elevated risk of Guillain-Barre syndrome (GBS) encountered with this vaccine was due to vaccine contamination by Campylobacter, a now known cause of GBS and endemic in poultry, from which eggs used for influenza vaccine production are sourced [15, 60]. By the time this hypothesis was formulated in 2006, however, some thirty years after vaccine production, there was extreme pessimism that vials of the original vaccine from different manufacturers and lots kept frozen throughout, could be found. Fortuitously, after considerable effort and a nationwide search, influenza researchers at Baylor University were found to have such an archive, thereby allowing this hypothesis to be tested, and ultimately rejected. These experiences highlight the need for the development and implementation of standard procedures for sample archiving, including guidance in the collection, preparation and storage of samples. 4.2 Potential safety concerns related to novel viral vaccines The development of some novel viral vaccines have necessitated the use of human tumorigenic and tumor-derived cell substrates, which could pose additional safety concerns related to the potential presence of unknown tumor viruses and latent viruses that may not be detected by the currently recommended assays [61]. Additionally, the use of large virus vectors can provide a target for endogenous retrovirus integration and amplification in the vector virus [62–65]. Therefore, advanced nucleic acid technologies with broad virus detection are being investigated for cell substrate characterization and may also be useful for characterization of the virus seed or products. 5. Avoidance of Adventitious Agents The production of live virus vaccines involves propagation of the vaccine virus in a suitable cell culture system, possible cell disruption for maximal yield of virus and, if necessary and if possible, purification of the virus. For biological products such as live virus vaccines, the introduction of an inactivation step(s) for adventitious agents as part of the downstream manufacturing process is not possible, since such a step is likely to compromise the immunogenicity of the vaccine virus. Thus the use of well characterized cell bank systems and qualified reagents for production is an even more important step to assure vaccine safety compared with their use for other vaccines or biological medicines. Progress has been made in the development of serum-free media for cell growth needed for the production of viral vaccines. However, the risk of introduction of adventitious agents through the use of other animal-derived substances such as trypsin during the production process remains. Use of gamma-irradiated or UV-treated reagents is also being considered in some cases when there is no adverse effect on the cell substrate. The risk of adventitious agents is reduced by current viral safety testing regulations and measures that recommend redundancy in testing using different assays and at different stages in manufacturing. Therefore, although the risk of adventitious agent introduction using primary cell substrates such as eggs and primary tissue cultures is higher than using a well characterized cell line, extensive and redundant testing provides confidence for their deployment during vaccine manufacture. Although complete elimination of animal derived reagents from the manufacturing procedure leads to a substantial reduction of the risk of contamination, the risk cannot be completely eliminated since animal-derived raw materials might be used in the production process of non-animal derived raw materials, such as enzymes to digest proteins to peptides and amino acids. Additionally, some cell substrates may not adapt to serum-free growth conditions. Further, it may be possible for viral contamination to arise from chemical reagents for growth medium preparation as illustrated by the minute virus of mice (MVM) contamination incident in the manufacture of a biopharmaceutical product [66]. On the other hand, advanced nucleic acid based technologies that have demonstrated success for detection and discovery of (new) adventitious agents such as virus microarrays, massively parallel or deep sequencing and broad range PCR combined with mass spectrometry could further contribute to the safety of biological products including vaccines. These new technologies still need to be validated for their intended use, determination of their performance parameters and how they can be applied to the safety of biological medicines. Efforts are ongoing to obtain data for scientific-decision making by regulators and industry regarding the use of the new technologies for evaluation of biological products. This was the focus of the 2013 PDA/FDA meeting on Advanced Technologies for Virus Detection in the Evaluation of Biologicals: Applications and Challenges [67]. Data was presented on the current use of the technologies for investigation of potential contaminants and characterization of cell substrates. Challenges for their routine use were identified plus ongoing group efforts were described. This meeting extended the discussions of the September 19, 2012 FDA Vaccines and Related Biological Products Advisory Committee (VRBPAC) on the use of human tumor cells for vaccine manufacture, which supported the use of the new technologies along with the currently recommended assays for detection of known and unknown viruses in novel cell substrates [68]. 6. Proposals for Archiving Vaccine Samples Comprehensive archiving samples of vaccine batches as well as the cell lines used for production would allow future retrospective analysis of vaccines by new (and presumably) improved technologies. In addition to the retention of physical samples, in order to investigate the impact of a contamination with an adventitious agent, a system of traceability for the used batches is proposed and should be in place. Retained samples from the seed lot and the cell bank, as well as of raw starting materials, would allow future scientists to determine the source of the contamination and who may have been exposed. In the conduct of clinical trials, samples of patients’ sera and peripheral blood mononuclear cells (PMBCs) taken prior to vaccination and at dedicated time points after vaccination should be stored in order to allow for investigation of the potential for human infection with any adventitious agent transmitted by the vaccine. 6.1 Type of Storage Vaccine samples should be frozen rapidly and stored below −70 °C to enhance retention of the viability of a live viral contaminant. For the purposes of future investigation of adventitious agents in cells, these should similarly be stored below −70 °C, although to retain long-term viability of cells, storage in the vapor phase of liquid nitrogen is required. Samples should be stored in suitable containers but preferably in the original containers to avoid any possibility of contamination being introduced during preparation for storage. A system should be in place for identifying and cataloguing stored samples. 6.2 Length of Sample Archiving Current US and EU regulations [32, 33] require manufacturers to retain a vaccine sample for one year post expiration of the vaccine (at the temperature that is indicated for the specific vaccine), vaccine ingredients that are used in the process, and 5% of each lot from the Phase I and II clinical trials for two years past the expiry date of the vaccine. However, in order to allow adequate retrospective testing for adventitious agents in vaccines in future years, past experience suggests that samples should be archived ideally for a minimum of 25 years. 6.3 Samples for storage For long-term archiving purposes it is proposed that for each batch of investigational or developmental (i.e. those used in pre-licensure studies) and commercial (i.e. licensed) vaccine, at least 10 ml of unformulated bulk and at least 10 vials/syringes of the final vaccine should be archived. Consideration has to be given to the quantity of vaccine likely to be required for analysis by any particular technique [32, 33]. This can be assessed for current technologies but is difficult to assess for future technologies; quite simply, the more, the better. Since regulatory laboratories are unlikely to have the resources to enable them to perform such archiving, this would have to be undertaken by the manufacturer of the vaccine. This however would probably require a change to the regulations and is unlikely to be achieved easily. This would not preclude a regulatory laboratory storing samples on an ad hoc basis and there may be room for negotiation between a government agency and the manufacturer as to where and by whom samples are archived. If a company were to dissolve, the company would be responsible for the transfer of the samples to a competent regulatory authority for archiving. 6.4 Financial Responsibility The purpose of this document is to provide technical considerations for guidance and not to determine financial responsibility for the cost of sample archiving. However, it is recognized that the financial burden for appropriate storage of samples above current regulatory requirements would be substantial, possibly even prohibitive. Despite this, the value of archiving material should not be underestimated and attempts should be made to establish a robust archiving system beyond that required by current regulations [32, 33]. Indeed, novel viral vectored vaccines that are “live” and have limited processing (i.e., no viral clearance steps) could be prioritized to follow the proposals provided herein as they could be riskiest to have a contamination. 6.5 Future Needs As noted earlier, there is a need for guidance on archiving of cells used to propagate virus vaccines and of records that provide full traceability of biological materials used in vaccine manufacture. Another key unresolved issue is the ability to track recipients of a contaminated vaccine accurately. Unlike efficacy, the safety of a vaccine usually cannot be measured directly; relative safety can usually only be inferred indirectly from the relative absence of multiple specific adverse events that have been measured. Either the discovery of an adventitious agent in a vaccine or the occurrence of adverse events in vaccinees can prompt an investigation of the vaccine. In fact, there is an ongoing root cause investigation for the recent identification of Mycoplasma hyorhinis in an investigational pox vector vaccine [69]. Adverse events get linked to specific vaccine exposures through epidemiological studies, and the possibility of contamination may be evaluated by laboratory testing with in vitro and in vivo studies, using conventional methods and new technologies, and through genetic sequencing. Epidemiological studies for determining relative risk are possible, however, only if there are records of who were exposed to the contaminated vaccine and who were not. While progress in developing computerized immunization information systems with tracking of vaccine manufacturer and lot number have been made in the U.S. [13], less progress has been made in the ability to track similar information in vaccinees in other countries [70]. Participation in voluntary centralized vaccination records in Canada has been made available through the launch of a phone app “ImmunizeCA app” in September 2014 [71], whilst in the USA, and effective June 10, 2015, applicants of biological products including vaccines are required to submit Lot distribution reports to the FDA according to amendments in 21CFR600.81 [72]. In addition to sample archiving, other aspects of the 2003 IOM recommendations for a “Vaccine Contamination Prevention and Response Plan’ remain undeveloped, such as “strategies for routine assessment of vaccine for possible contamination; notification of public health officials, health care providers, and the public if contamination occurs; identification of recipients of contaminated vaccines; and surveillance and research to assess health outcomes associated with the contamination”[15]. Given the large proportion of the human population exposed to vaccines annually, the large number of vaccine manufacturers and the diversity of their sourcing, the need for such a plan remains urgent. We wish to thank the following persons for their assistance in preparing this guidance document: 1) Other members of the V3SWG during the preparation of this document (Ken Berns, Jean-Louis Excler, Marc Gurwith, Michael Hendry, Najwa Khuri-Bulos, Dr. Lori D. Campbell of CDC for educating us on biobanking; 2) the following reference groups for their peer review: Disclaimer: The findings, opinions and assertions contained in this consensus document are those of the individual scientific professional members of this ad hoc Brighton Collaboration working group. They do not necessarily represent the official positions of each participant’s organization (e.g., government, university, or corporation). References 1 Centers for Disease Control and Prevention Impact of vaccines universally recommended for children--United States, 1990–1998 MMWR Morb Mortal Wkly Rep 1999 48 12 243 248 10220251 2 Fenner FHD Arita I Jeek Z Ladnyi ID Smallpox and its eradication. Geneva, Switzerland:World Health Organization 1988 Smallpox and its eradication 1988 Geneva, Switzerland World Health Organization 3 Centers for Disease Control and Prevention Progress toward interruption of wild poliovirus transmission--worldwide, January 2011-March 2012 MMWR Morb Mortal Wkly Rep 2012 61 19 353 357 22592275 4 Andrus JK Measles and rubella eradication in the Americas Vaccine 2011 29 Suppl 4 D91 D96 22185837 5 World Health Organization Annex 3. Recommendations for the evaluation of animal cell cultures as substrates for the manufacture of biological medicinal products and for the characterization of cell banks WHO Technical Report Series, No. 978 2010 6 Nims RW Detection of adventitious viruses in biologicals--a rare occurrence Dev Biol (Basel) 2006 123 153 164 discussion 183–97 16566443 7 The International Conference on Harmonisation (ICH) Guidance, Q5A: Viral Safety Evaluation of Biotechnology Products Derived from Cell Lines of Human or Animal Origin (63 FR 51074; September 24, 1998) 8 Victoria JG Viral nucleic acids in live-attenuated vaccines: detection of minority variants and an adventitious virus J Virol 2010 84 12 6033 6040 20375174 9 Hattermann K Infection studies on human cell lines with porcine circovirus type 1 and porcine circovirus type 2 Xenotransplantation 2004 11 3 284 294 15099209 10 Dubin G Investigation of a regulatory agency enquiry into potential porcine circovirus type 1 contamination of the human rotavirus vaccine, Rotarix: Approach and outcome Hum Vaccin Immunother 2013 9 11 11 Lilienfeld DE The first pharmacoepidemiologic investigations: national drug safety policy in the United States, 1901–1902 Perspect Biol Med 2008 51 2 188 198 18453724 12 Petricciani J Adventitious agents in viral vaccines: lessons learned from 4 case studies Biologicals 2014 42 5 223 236 25135887 13 Centers for Disease Control and Prevention Progress in immunization information systems - United States, 2011 MMWR Morb Mortal Wkly Rep 2013 62 48 51 23344697 14 U.S. Food and Drug Administration Transmissible Spongiform Encephalopathies Advirosry Committee Vaccines and Related Biological Producrs Advisory Committee 2000 15 Stratton KAD McCormick MC Immunization Safety Review: SV40 Contamination of Polio Vaccine and Cancer 2003 Washington D.C. Institute of Medicine of the National Academies 16 Chen RT The Brighton Collaboration Viral Vector Vaccines Safety Working Group (V3SWG) Vaccine 2015 33 1 73 75 25305565 17 Gilchrist SA Nanni A Lessons learned in shaping vaccine markets in low-income countries: a review of the vaccine market segment supported by the GAVI Alliance Health Policy Plan 2012 18 Milstien J Reaching international GMP standards for vaccine production: challenges for developing countries Expert Rev Vaccines 2009 8 5 559 566 19397413 19 Graham JE Capacity for a global vaccine safety system: the perspective of national regulatory authorities Vaccine 2012 30 33 4953 4959 22658930 20 Marcus-Sekura C Evaluation of the human host range of bovine and porcine viruses that may contaminate bovine serum and porcine trypsin used in the manufacture of biological products Biologicals 2011 39 6 359 369 22000165 21 Dodet B Viral safety and extraneous agents testing for veterinary vaccines Biologicals 2010 38 3 326 331 20503495 22 Atkin SJ Griffin BE Dilworth SM Polyoma virus and simian virus 40 as cancer models: history and perspectives Semin Cancer Biol 2009 19 4 211 217 19505648 23 Shah K Nathanson N Human exposure to SV40: review and comment Am J Epidemiol 1976 103 1 1 12 174424 24 Fraumeni JF Jr Ederer F Miller RW An evaluation of the carcinogenicity of simian virus 40 in man JAMA 1963 185 713 718 13959363 25 Sangar D Examination of poliovirus vaccine preparations for SV40 sequences Biologicals 1999 27 1 1 10 10441397 26 Simian Virus 40 (SV40): a Possible Human Polyomavirus. Symposium proceedings. Bethesda, Maryland, USA. January 27–28, 1997 Dev Biol Stand 1998 94 1 406 9867465 27 WHO Expert Committee on Biological Standardization. Forty-ninth report:1990 Requirements for poliomyelitis vaccine (oral) (Addendum 1998) World Health Organ Tech Rep Ser 2000 897 61 66 28 Gazdar AF Butel JS Carbone M SV40 and human tumours: myth, association or causality? Nat Rev Cancer 2002 2 12 957 964 12459734 29 Cutrone R Some oral poliovirus vaccines were contaminated with infectious SV40 after 1961 Cancer Res 2005 65 22 10273 10279 16288015 30 Qi F Simian virus 40 transformation, malignant mesothelioma and brain tumors Expert Rev Respir Med 2011 5 5 683 697 21955238 31 World Health Organization Recommendations for the production and control of poliomyelitis vaccine (inactivated) WHO Technical Report Series, No. 910 2002 32 US Food and Drug Administration Code of Federal Regulations Title 21 2011 33 EUROPEAN COMMISSION. EU Guidelines to Good Manufacturing Practice Medicinal Products for Human and Veterinary Use: Annex 19 Reference and Retention Samples. Brussels, 14 December 2005 34 Kalland KH Ke XS Oyan AM Tumour virology--history, status and future challenges APMIS 2009 117 5–6 382 399 19400863 35 Monath TP Yellow fever vaccine Expert Rev Vaccines 2005 4 4 553 574 16117712 36 Waters TD Yellow fever vaccination, avian leukosis virus, and cancer risk in man Science 1972 177 4043 76 77 4339427 37 Hussain AI Identification and characterization of avian retroviruses in chicken embryo-derived yellow fever vaccines: investigation of transmission to vaccine recipients J Virol 2003 77 2 1105 1111 12502826 38 Norman JE Mortality follow-up of the 1942 epidemic of hepatitis B in the U.S Army. Hepatology 1993 18 4 790 797 8406352 39 Seeff LB A serologic follow-up of the 1942 epidemic of post-vaccination hepatitis in the United States Army N Engl J Med 1987 316 16 965 970 2436048 40 Tsang SX Evidence of avian leukosis virus subgroup E and endogenous avian virus in measles and mumps vaccines derived from chicken cells: investigation of transmission to vaccine recipients J Virol 1999 73 7 5843 5851 10364336 41 Ferguson M WHO Working Group on Technical Specifications for Manufacture and Evaluation of Yellow Fever Vaccines, Geneva, Switzerland, 13–14 May 2009 Vaccine 2010 28 52 8236 8245 21055492 42 Robertson JS Nicolson C Riley A-M Bentley M Dunn G Corcoran T Schild GC Minor P Assessing the significance of the presence of RTase in live viral vaccines Biologicals 1997 25 403 414 9467036 43 Boni J Detection of reverse transcriptase activity in live attenuated virus vaccines Clin Diagn Virol 1996 5 1 43 53 15566860 44 Shahabuddin M Sears JF Khan AS No evidence of infectious retroviruses in measles virus vaccines produced in chicken embryo cell cultures J Clin Microbiol 2001 39 2 675 684 11158127 45 Khan AS The reverse transcriptase activity in cell-free medium of chicken embryo fibroblast cultures is not associated with a replication-competent retrovirus J Clin Virol 1998 11 1 7 18 9784139 46 Reverse transcriptase activity in chicken-cell derived vaccine Wkly Epidemiol Rec 1998 73 28 209 212 9677853 47 Meehan BM Characterization of novel circovirus DNAs associated with wasting syndromes in pigs J Gen Virol 1998 79 Pt 9 2171 2179 9747726 48 McClenahan SD Krause PR Uhlenhaut C Molecular and infectivity studies of porcine circovirus in vaccines Vaccine 2011 29 29–30 4745 4753 21569811 49 Administration, U.F.a.D. Code of Federal Regulations Title 9. 2012. 9 CRF 113.53 Requirements for ingredients of animal origin used for production of biologics 50 US Food and Drug Administration Code of Federal Regulations Title 19 CFR 113.47 Detection of extraneous viruses by the fluorescent antibody technique 51 Allan GM Some biological and physico-chemical properties of porcine circovirus Zentralbl Veterinarmed B 1994 41 1 17 26 7941842 52 Tischer I A very small porcine virus with circular single-stranded DNA Nature 1982 295 5844 64 66 7057875 53 Centers for Disease Control and Prevention Recommendation to Temporarily Suspend Usage of GlaxoSmithKline Rotarix (Rotavirus) Vaccine Health Alert Network 2010 54 Beach NM Productive infection of human hepatocellular carcinoma cells by porcine circovirus type 1 Vaccine 2011 29 43 7303 7306 21742002 55 Gagnieur L Unbiased analysis by high throughput sequencing of the viral diversity in fetal bovine serum and trypsin used in cell culture Biologicals 2014 56 Chang JT Quantitative detection of residual porcine host cell DNA by real-time PCR Biologicals 2014 42 2 74 78 24394374 57 Council of Europe 2.6.16. Tests for Etraneous Agents in Viral Vaccines for Human Use European Pharmacopoeia 5.0 2004 58 Pennathur S Application of Risk Assessments in the Design of the Overall Viral Control Strategy Used during the Manufacture and Testing of Live Virus Vaccines PDA J Pharm Sci Technol 2011 65 6 730 736 22294607 59 Agency EM Guideline on the use of porcine trypsin used in the manufacture of human biological medicinal products 2014 60 Nachamkin I Anti-ganglioside antibody induction by swine (A/NJ/1976/H1N1) and other influenza vaccines: insights into vaccine-associated Guillain-Barre syndrome J Infect Dis 2008 198 2 226 233 18522505 61 Food and Drug Administration and Center for Biologics Evaluation and Research U.S.D.o.H.a.H. Services Guidance for Industry Characterization and Qualification of Cell Substrates and Other Biological Materials Used in the Production of Viral Vaccines for Infectious Disease Indications 2010 62 Lerch RA Friesen PD The baculovirus-integrated retrotransposon TED encodes gag and pol proteins that assemble into viruslike particles with reverse transcriptase J Virol 1992 66 3 1590 1601 1371168 63 Davidson I Shkoda I Perk S Integration of the reticuloendotheliosis virus envelope gene into the poultry fowlpox virus genome is not universal J Gen Virol 2008 89 Pt 10 2456 2460 18796713 64 Kung HJ Retroviral insertion into herpesviruses: discovery and implications Dev Biol (Basel) 2001 106 223 228 discussion 228–9, 253–63 11761235 65 Isfort R Retrovirus insertion into herpesvirus in vitro and in vivo Proc Natl Acad Sci U S A 1992 89 3 991 995 1310544 66 Garnick RL Raw materials as a source of contamination in large-scale cell culture Dev Biol Stand 1998 93 21 29 9737373 67 Khan AS Vacante DA Introduction and workshop summary: advanced technologies for virus detection in the evaluation of biologicals-applications and challenges PDA J Pharm Sci Technol 2014 68 6 546 547 25475626 68 U.S. Food and Drug Administration 2012 Meeting Materials, Vaccines and Related Biological Products Advisory Committee 2012 69 Report., V.N. NYVAC-HIV vaccine applied in HVTN 092 and HVTN 096 clinical trials Vaccine News Daily 2015 70 Freitas DR Investigation of an outbreak of hypersensitivity-type reactions during the 2004 national measles-mumps-rubella vaccination campaign in Brazil Vaccine 2013 31 6 950 954 23246259 71 WHO Vaccine Safety Net Newsletter World Health Organization 72 USFDA Guidance, Compliance & Regulatory Information (Biologics) Lot Distribution Database: 600.81 Distribution Reports 1994 U.S. Food and Drug Administration

PMC005xxxxxx/PMC5131095.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 9211652 2196 Int Arch Allergy Immunol Int. Arch. Allergy Immunol. International archives of allergy and immunology 1018-2438 1423-0097 27820941 5131095 10.1159/000449249 NIHMS813003 Article Functions of Exosomes and Microbial Extracellular Vesicles in Allergy and Contact and Delayed-Type Hypersensitivity Nazimek Katarzyna 1* Bryniarski Krzysztof 1* Askenase Philip W. 2# 1 Department of Immunology, Jagiellonian University Medical College, Krakow, Poland 2 Section of Allergy and Clinical Immunology, Department of Internal Medicine, Yale University School of Medicine, New Haven, CT, 06520, USA # Corresponding author: Philip W. Askenase, Section of Allergy and Clinical Immunology, Department of Internal Medicine, Yale University School of Medicine, 333 Cedar St, New Haven, CT 06510, USA; philip.askenase@yale.edu; Phone: (001) 203-785-4170 Fax: (001) 203-785-3229 * Contributed equally 10 11 2016 8 11 2016 2016 08 11 2017 171 1 126 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. Extracellular vesicles, such as exosomes, are newly recognized intercellular conveyors of functional molecular mechanisms. Notably, they transfer RNAs and proteins between cells in general, that then can participate, as described herein, in the complex pathogenesis of allergic and related hypersensitivity responses and disease mechanisms. This review highlights this important new appreciation of the in vivo participation of such extracellular vesicles in the interactions between allergy-mediating cells, taking into account paracrine epigenetic exchanges mediated by surrounding stromal cells and the endocrine receipt of exosomes from distant cells via the circulation. Exosomes are natural ancient nanoparticles of life. They are made by all cells and in some form by all species down to fungi and bacteria, and are present in all fluids. Besides a new focus on their role in the transmission of genetic regulation, exosome transfer of allergens was recently shown to induce allergic inflammation. Importantly, regulatory and tolerogenic exosomes can potently inhibit allergy and hypersensitivity responses, usually acting non-specifically, but also can proceed in an antigen-specific manner due to coating of the exosome surface with antibodies. Deep analysis of processes mediated by exosomes should result in development of early diagnostic biomarkers, as well as allergen-specific, preventive and therapeutic strategies. These likely will significantly diminish the risks of current allergen specific parenteral desensitization procedures, and of the use of systemic immunosuppressive drugs. Since extracellular vesicles are physiological, they can be fashioned for specific delivery of therapeutic molecular instructions through easily tolerated, non-invasive routes, such as oral ingestion, nasal administration, and perhaps even inhalation. extracellular vesicles exosomes allergy asthma delayed-type hypersensitivity contact hypersensitivity contact dermatitis immunosuppression immunoregulation Introduction Exosomes are exciting newly recognized extracellular vesicles that transfer a variety of bioactive compounds, including RNA molecules, proteins and lipids, between nearby cells in a paracrine manner, such as at synapses. This has been exquisitely shown at the central immunological synapse between T cells and antigen (Ag) presenting cells (APC) [1]. Furthermore, locally secreted exosomes can then enter the vasculature to act systemically on distant cells in an endocrine manner. We have demonstrated this recently for splenic and lymph node-derived T suppressor (Ts) cell secreted suppressive exosomes regulating the interface of APC and effector T cells at the peripheral tissue site of an immune inflammatory response [2]. Morphologically, exosomes are spherical vesicles with a bilamellar membrane that are 50–200 nanometers in diameter depending on the source and activation state of the donor cell [2]. They are formed by budding from the wall of terminal endosomes to then accumulate at the intracellular periphery in the multivesicular body (MVB). When the MVB undergoes exocytosis, there is the release of the contained bunches of exosomes into the extracellular space [3,4]. Alternatively, microvesicles bud individually from the cell surface, are larger (200–1000 nanometers) and mediate similar mechanisms with great overlap in the characteristics of exosomes. Together they are called extracellular vesicles. Study of exosomes is a new field concerned with very large numbers of very tiny vesicles, for which there are yet no specially designed machines for their particular analysis. According to current experimental observations, there seems to be billions of exosomes per milliliter of blood, that is a thousand times greater than the number of white blood cells. Moreover, there is a great variability of vesicle types with many subtypes in exosome preparations [5]. There also is a significant variability in properties between individual exosomes, like their membrane content [6], their contained proteome and types of RNAs [7–9], with yet further variability due to the stage of maturity and activation of the cells generating the exosomes [2,10,11]. The proteome and RNAome of exosomes differ markedly from those from their donor cells, indicating active mechanisms of their production, sorting and loading as cargo. Exosomes are natural and physiological ancient nanoparticles of life, since they are made by all cells, are present in all body fluids and are made in some form by all species down to and through fungi [4] and bacteria [12,13]. Accordingly, bacteria are well known to produce at their surface the microbial extracellular vesicles (mEV), formerly termed outer membrane vesicles (OMV), that enables transfer of contents consisting of enzymes and toxins [14] to regulate other bacteria. Further, mEV, like exosomes, are now known to contain RNAs [15,16]. Importantly, mEV can crucially influence host allergic immune responses. They have the ability to drive clinical diseases such as in atopic dermatitis, via the mEV from skin surface microorganisms commonly associated with this allergic disease, such as staphylococci [17] and fungi [18]. Exosomes offer cell-free therapy without the transfer of possibly oncogenic DNA, compared to whole nucleated cells. However, DNA as fragments has been recently demonstrated in exosomes from cancer cells [19–21]. Thus, in some instances the functions of cells can be replaced by their produced exosomes. A prominent example is that mesenchymal stem cell healing functions often can be completely replaced by their exosomes [22–25], which in some cases was shown to be due to identified exosome miRNA cargo [26,27]. A very important difference from cells is that exosomes are resistant to harsh conditions, like a pH of 3–4 [28,29], or even pH = 1 [30] encountered in the gastrointestinal (GI) tract, as well as hyperosmolarity due to expression of the water channel called aquaporin-1 [31], and hyperoxia [32] or hypoxia [33], when the exosome donor cells are stimulated under those conditions. Therefore, extracellular vesicles, such as exosomes from various sources, play a newly recognized, important and diverse role in physiological processes and pathological conditions relevant to diseases, such as those of allergy and hypersensitivities. Exosome Transfer of Signaling Proteins and Genetically Active RNAs Pertinent to Allergy and Hypersensitivity There is a well-recognized role of exosomes in allergy [34]. The outstanding new property that exosomes bring to allergic and other responses is the transfer between cells of functional genetic material and signaling proteins. Transferred regulatory RNAs can then epigenetically alter gene function in the acceptor cells [35–41], while transferred signaling proteins, like wnt, can alter their intracellular functions [42–45]. Accordingly, such transfers of genetic information have been shown for immune [46–50], hypersensitivity [2], IgE [51] and allergic asthmatic [52] responses. Such intercellular communication and consequent regulation at the genetic level is unprecedented in human physiology and disease processes. Therefore, this greatly widens the possibilities of a more complete understanding and more precise interventions in allergy and related immune mediated diseases. As noted, there is a great variability between exosomes, even those produced by the same cell [6], in their proteome [53] and RNAome. Importantly, functional mRNA and miRNAs can be a small percentage of the total exosome RNA [2,54,55]. However, functions of many of the miRNAs still have to be delineated. The remaining RNA cargo is usually made up of other classes of RNA, including long non-coding RNAs of unknown significance, along with numerous ribosomal RNA fragments and tRNAs, also of unknown functions [2,54,55]. Therefore, exosome exchange of molecules between cells is a sophisticated and potentially quite heterogeneous process of great complexity, with diverse functional consequences for mediation and modulation of cellular functions in immunity, allergy and hypersensitivity. On a wider level, the potential for exosome transfer of functional RNAs and signaling proteins from neighboring and distant cells of other systems, that have been thought to be independent of allergy and immunity, now should be seen as likely interacting with or affecting the cells mediating allergic and immunological responses. For instance, exosomes derived from various cell types of the microenvironment influence the established immune cell-cell interactions in allergy and immunity, and thus are now considered as important in actual in vivo circumstances. Fig. 1 shows how this might apply to interactions of the various immune and stromal tissue cell populations in the airways of asthma patients. The intercellular transfer of RNAs mediating epigenetic changes and exchange of signaling molecules, including transcription factors [56], or their regulators [57], makes this a powerful new source for a fuller understanding of allergy and hypersensitivity. Therefore, this should provide new diagnostic opportunities and therapeutic maneuvers to potentially intervene in allergic and immunological disease processes at entirely new levels. Effects of Exosome Intercellular Interactions on Allergy and Immunity According to the above considerations, exosomes have recently been shown to exhibit great effects on immunological activities, as mediators involved in many stages of immune and inflammatory responses, such as induction, orchestration, elicitation, resolution and regulation. As an important example, T regulatory (Treg) Foxp3pos cells produce exosomes that suppress Th1 cell proliferation and cytokine production via gene silencing due to transferred Treg cell-derived miRNAs [58]. Further, thymic epithelial cells produce exosomes carrying tissue-restricted Ag that guide development of the Treg cells [59,60]. Signaling pathways also can be altered by either exosome transferred miRNA influencing translation by mRNA inhibition in the targeted cell, the transfer of mRNA to alter production of specific protein in the targeted cells, or exosome transfer of the proteins themselves. For instance, the signaling molecule wnt can be gained [44,45] or lost by cells via exosomes [42]. This protein can be expressed on the exosome surface and, thus, may potentially be transferred between cells in such a form [43]. Importantly, inflammation and immunity can also be influenced by transfer of exosome contained cytokines [61–64], their receptors [65], and signaling molecules [65], by stimulation of cytokine production via exosome-derived TLR agonists [61], or finally by cells activated by exosomes. This includes T cell vesicles activating mast cells [66,67] as well as exosome transfer of mRNA encoding mast cell cytokines [38]. Furthermore, endogenously induced, natural exosomes, or those modified ex vivo, can carry and transfer immunoregulation via miRNAs that might be altered for application in therapy of diverse allergic diseases, better control of dysregulated immune responses, as well as for influencing hypersensitivity mechanisms that underlie clinical disease processes. With the growing prevalence of allergies, hypersensitivities and diseases with immune mediated manifestations, this newly recognized function mediated by previously unknown exosome intercellular transfer of epigenetic regulation joins the currently unraveled genetic and molecular mechanisms underlying their complex pathogenesis. Many other aspects of these disorders remain to be clarified by further and deeper investigation and then translation to clinical understanding and new therapies, and these also seem to depend on exosome transfer of gene regulation and cell signaling. This current review particularly highlights the involvement of these processes in allergic and immunological diseases, that have recently have shown increasing prevalence and severity. A revealing example relevant to allergy concerns CD23pos (Fc-epsilonR-2pos) B cells that capture immune complexes consisting of IgE and allergen as Ag to then produce exosomes (B cell exosomes called “bexosomes”) carrying CD23, IgE and MHC class II, that are transferred to dendritic cells (DC) that in turn stimulate Ag-specific T cells [51]. This suggests that “bexosomes” can provide the essential transfer mechanism for IgE-Ag complexes from B cells to DC for subsequent activation of allergen-specific T cells. In this case, anti-IgE or rituximab (anti-CD20) therapy should block this Other examples of exosome-mediated intercellular influences in immunity, hypersensitivity and allergy, include: a. helper T cell augmentation of B cell production of HLApos exosomes [11], b. Ts [2] and Treg [58] cell-derived exosome inhibition of effector Th1 cells, c. thymic epithelial cell-derived, Ag carrying exosome mediated maturation of Treg cells [59,60], d. DC activation of other DC and also B cells [50], e. T cell activation of APC and DC [49], and, the opposite, f. DC-derived exosomes stimulating CD4pos T cells [68]. Further, DC can be guided by exosomes with MHC class II and co-stimulatory molecules present in human BALF [69], and by mast cell-derived exosomes inducing phenotypic and functional maturation of DC to elicit specific immune responses in vivo [70]. Exosome mediated intercellular interactions involving mast cells are particularly relevant to allergy. These include: T cell exosome activation of mast cells [66,67], and mast cell activation of T and B cells [71,72], mast cell exosome-mediated phenotypic and functional maturation of DC [61,62] and endothelial cells [73], or other mast cells or progenitor stem cells [74]. These effects can be induced via exosome transmitted cytokines [66,75] or mast cell exosome transfer of mRNA encoding the cytokines [37], and by affecting cytokine signaling [75] or triggering cytokine production by bronchial cells [76] or by airway smooth muscle cells [77]. Finally, mast cell exosomes induce phenotypic and functional maturation of DC enabling them to elicit specific immune responses in vivo [78], they activate endothelial cells to secrete clotting factors [73] and can activate T and B cells [79,80]. Taken together, such intercellular exchanges via exosomes occur in established processes dominant in classical allergic mechanisms, for instance in the release of traditional mast cell mediators, like histamine, bioactive peptides, leukotrienes and even cytokines. These traditional pro-allergic mediators affect surrounding cells in classical Type I immediate hypersensitivity. Additionally, target cells are likely modified by miRNAs and other elements transferred by the mast cell-derived exosomes to mediate epigenetic changes in these cells. Further, the releasing mast cells themselves are now seen as being affected by exosomes from surrounding and even distant cells. Of additional relevance to mechanisms of allergic asthma, these patients have increased numbers of airway exosomes expressing MHC and co-stimulatory molecules that may play a similar role as APC [69,81]. Further, among the increased numbers of exosomes in airways of mice with a model of asthma, the bronchoepithelial cell-derived exosomes stimulated by IL-13 induce activation of macrophages [77], and exosomes play a role in Th2 cell activation of auxiliary cells [77,78], acting via specific cell surface cytokine receptors, such as IL-4 receptor. This seems to influence triggering of established intracellular signaling pathways that regulate gene expression of Th2 responses. Eosinophil-derived exosomes are also suspected to play a role in asthma pathogenesis [78]. Thus, it now must be considered that established allergic mast cell activation for mediator release, as well as conventional intercellular cytokine dominated pathways, are proceeding in parallel with, or more likely interact with donation and receipt of a variety of intracellular communications arising locally or from distant cells via the blood. Simultaneously they are triggered by epigenetic effects of exosome-derived miRNAs and mRNAs, or proteins transferred between these cells [Fig. 1]. These newly recognized exosome influence described as an intercellular cloud of diverse exosomes from immune, inflammatory and tissue cells results in modulation of the established mechanisms of allergic and hypersensitivity responses and thus likely plays an important role allergic and immunological diseases. Mast Cell Exosomes Act in Allergen Ag Presentation It was previously shown that classical APC, like macrophages and particularly DC, pulsed with whole native Ag release immunogenic exosomes that have surface complexes of the Ag-peptides in MHC molecules. These exosomes can act as “mini-APC-like” substitutes, that can bind to T cell surface Ag-MHC-specific TCR receptors to induce T cell signaling for effector functions [79–81], such as activation of CD4pos T cells [68]. As above, sensitization for allergic atopic Type I responses involves allergen engulfment and then intracellular processing by APC, principally DC, but also macrophages. The subsequent cell surface Ag presentation of DC-digested Ag peptides derived from allergen proteins, to the effector Th2 lymphocytes stimulate their release of helper cytokines, like IL-4 and IL-13. These act on IgE positive B cells that produce allergen-specific IgE Ab and on asthma-associated tissue cells, like goblet cells, to produce eosinophil-recruiting chemokines. New findings pertinent to this review show that besides DC and macrophages, mast cells can function similarly as APC in Th2 responses, also by taking up allergens to possibly generate similar Ag-presenting exosomes. However, it remains possible that these are taken up by DC or B cells [82] to then truly present Ag [83]. Such transfer of mini-APC may thus account for transferred functions at a mast cell and DC synapses [1,84–86], and are connected with demonstrated ability of mast cell exosomes to induce maturation of DC that allows for their conventional Ag presentation to T cells [70]. IgE antibodies (Ab) may also play a role in mast cell function in Ag presentation. Thus, mast cell-released exosomes express FcεRI receptors that may be involved in the reuptake of IgE complexed with allergen, not only by mast cells but in humans also by conventional APC also expressing FcεRI, that in turn likely mediate allergen presentation [87]. At such newly described mast cell synapses formed with other cells, like recently shown at the central immunological synapse of T cells and APC or B cells [1], there is intercellular passage of RNAs and proteins from the released mast cell exosomes. Thus, during the effector phase of allergic responses, apart from classical allergen-specific IgE-mediated degranulation of mast cells to release diverse pro-allergy and inflammatory mediators, there is an additional layer of molecular mediator release due to exocytosis of MVB-derived exosomes that may transfer regulatory RNAs and proteins to neighboring cells [70,71]. Note that the first description of exosome-mediated intercellular exchange of functional RNA was shown in mast cell lines, with the appearance of donor cell proteins in the acceptor cells due to the transferred mRNA, and interestingly some of mRNAs were hardly expressed in the donor cells [38]. The release of Ag/MHC coated exosomes by traditional APC and mast cells, represents an entirely new aspect of Ag presentation. Not only does the process proceed at the immunological synapses between APC and effector T cells, but likely also during the interaction of APC exosomes with T cells, which can be envisioned as a cloud of “mini-Ag-presenting” exosomes around T cells. This would greatly change the stoichiometry of the APC phenomenon, since existence of such Ag-presenting exosome cloud would greatly enlarge this essential and central immune cell interaction. This therefore will not merely depend on the membrane to membrane co-localization of the T cell and APC for formation of the conventional synapse between the surfaces of rare DC acting as APC, with passing rare Ag-specific T cells. In an analogous fashion, it seems that there also is a potential cloud of effector T cell-derived exosomes expressing membrane TCR and CD3 that are capable of binding with Ag/MHC complexes on the cell membranes of neighboring Ag-presenting B cells to trigger their function [1], and conversely a mini αβTCR/CD3 T cell exosome cloud back-stimulating the B cell or DC. Thus, a molecular exosome RNA-mediated epigenetic effect of the T cells on the APC [36,37] or similarly on B cells [1], and, conversely, of the APC on the T cells [60], and even stimulated DC activating other DC, are newly recognized aspects of central immunological processes. Thereby, this opens possible new levels for understanding and treating allergic and hypersensitivity diseases. Therefore, this novel exosomal Ag presentation process can be exploited for new modes of allergen desensitization, that could perhaps be reduced from the use of native allergens to just treating patients with APC-derived exosomes, as is currently done for vaccination against infectious agents [88,89] and cancer [90–92]. Exosome Involvement in Allergic Rhinitis and Possibly in Non-Allergic Rhinitis Intriguingly, there is a relevant study that extends the concepts described above about APC exosome allergen-presentation. Cultured B cells from patients allergic to birch pollen, including those with allergic rhinitis, that have B cells expressing surface immunoglobulin receptors (BCR) of the same allergen specificity as their produced Ab, released Ag-specific exosomes with BCR on their surface. After in vitro loading with a birch pollen-derived allergenic peptide Bet v 1, these anti-birch pollen specific B cell exosomes were used as mini-APC to stimulate patient T cells. This new exosome mediated B and T cell collaboration caused in vitro T cell proliferation and secretion of the Th2-dependent cytokines IL-5 and IL-13, that likely could participate in the allergic rhinitis and asthma [93]. It was shown recently that there is great diversity of miRNA expression in tissues of nasal mucosal biopsies from patients with persistent and also non-persistent asthma in comparison to healthy volunteers [94]. In another relevant study, analysis of the miRNA profile in extracellular vesicles obtained from nasal mucus of patients with allergic rhinitis, when compared to healthy individuals, revealed differences, implying that nanovesicle-transmitted RNAs likely may also be involved in the pathogenesis of this most common Type I allergic disease [95]. There is an association of mRNA levels in nasal polyps with the release of eosinophil mediators, like RANTES and eotaxin [96], which raises the possibility of treating nasal polyps with intranasally administered exosomes containing RNA antagonist of the reverse sequence to the polyp-associated mRNAs or miRNAs. Finally, isolation of nasal mucosal and epithelial samples from patients with chronic obstruction due to atypical allergic rhinitis, for their further examination using immune gold electron microscopy, showed that the resulting exosomes carried relevant nasal allergens, like Staphylococcal enterotoxin B and house dust mite (HDM)-derived Derp1 Ag. These exosome mini-allergen-APC were claimed to induce naive CD3pos T cells to differentiate into CD8pos T cells. Further, on exposure to specific Ag, the CD8pos T cells released granzyme B and perforin, and more than 30% of the Ag-specific CD8pos T cells proliferated [97]. An interpretation of these data suggests that this could be an example of functioning in vivo allergen positive APC-derived exosomes that are able to stimulate T cells in the pathogenesis of chronic atypical rhinitis. Exosomes collected from human nasal lavage fluid were shown to induce the migration of innate immune cells, that may play an important role in the defense against pathogens and allergens. On the other hand, the decreased expression of antimicrobial proteins in nasal exosomes from patients with airway diseases likely contributes to an increased susceptibility to infections and to disease progression [98]. Thus, it could be postulated that similar clinical syndromes seeming to be independent of IgE Ab might be caused by pathogenic exosomes from local and/or distant cells, that perhaps act in an Ag-specific manner. Alternatively, the mechanism underlying these clinical symptoms could be due to a combined action of such mini-APC exosomes and very low levels of IgE Ag-specific antibodies that do not cause the elicitation of macroscopically positive skin tests or other in vitro tests. This recalls our demonstration in mice that very low amounts of IgE (1 ng per mouse equal to 2 pg per person) failed to elicit macroscopic immediate skin responses to allergen challenge, but still mediated microscopic Type I vasoactive responses that in fact can allow for local recruitment of hypersensitivity effector T cells into the tissues [99]. Another postulate involves cross kingdom considerations pertinent to exosomes; like action of food-derived miRNAs carried from the intestine by host exosomes to influence a human patient with allergy [100]. In this regard, it is postulated that extracellular vesicle-delivered enzymes or RNAs produced by the allergy-inducing plants, or even released from their pollens [101], or in allergy to house mites, where insects are known to release exosomes from their MVB [44,45], may play a role in some of these atypical clinical conditions [102]. Accordingly, exosomes derived from DC of patients exposed to cats, carrying the major peptide determinant of cat allergen (Fel d 1), recently were shown to potently stimulate cytokine production by Th2 lymphocytes in cat allergen sensitive individuals [103]. On the contrary, particular miRNAs from exosomes of nasal epithelial cells from this atypical allergic rhinitis microenvironment stimulate IL-10 production by monocytes to inhibit nasal allergy [104]. Thus, this sort of allergen induced intercellular communication might be responsible for the seemingly anomalous finding that, although cat dander is very sensitizing and produces strong symptoms in many atopic individuals, exposure to multiple cats curtails this powerful allergic sensitivity. This could be due to a “modified Th2 response;” characterized by an emerging role of antigen-specific Treg cells and associated IgG4 Ab responses, that prevent the specific allergic reaction in allergic individuals, and of clinical practice interest may constitute the results of clinical desensitization with high doses of cat allergen [105] as is true of desensitization with stinging insects [105a]. Additionally, tissue iNKT cells may be involved in such exosome communications, since exosomes express not only MHC but also can express the minor histocompatibility CD1d molecules known to bind glycolipid Ag, like alpha-galactosylceramide (alpha-gal-cer) the canonical glycolipid Ag activating semi-invariant αβTCR on iNKT cells [106]. Such mini-APC exosomes with surface CD1d may stimulate tissue iNKT cells to release allergy-promoting Th2 cytokines, like IL-4 and IL-13, after interaction with host, plant or mite glycolipids. As a result, these CD1d-dependent iNKT cells may induce allergic responses against glycolipid Ag with the release of IL-4, as we showed previously in contact hypersensitivity [107], or IL-13 reacting against polysaccharides from bacteria, as we showed in a murine model of pneumococcal pneumonia [108]; or the plant pollens [101], as a requirement for eventual effector Th1 cell participation. Very interestingly, plant allergens, that play a major role in atopic allergic rhinitis and asthma, have been known for a long time to be associated with their pollens, and recently were shown to be released from pollens in exosome-like vesicles called “pollensomes” [109]. Regarding such interactions in food allergy, extracellular MHC class II positive exosomes released by epithelial cells of the GI tract, that have mini-APC abilities, may play an important role in transmission of food allergens from the intestinal lumen to the local lymphatic system [110]. This intercellular exosome communication may thus allow for Ag-presentation by DC to specific T cells. Such a T cell response to the dietary Ag of these mini-APC exosomes may initiate and then aggravate the responses to food allergens. Distinctly, these conditions may be represented by patients with food allergy induced by iNKT cells reacting to the APC-derived exosomes pulsed with the plant Ag, and, therefore, producing clinically significant symptoms, but lacking the positive result of immediate skin tests to the food allergens. Moreover, exosomes were recently shown to transfer haptenated proteins to possibly activate allergy to beta-lactams in a model of clinical allergy to amoxicillin [111]. Host cell derived exosomes and cross kingdom allergens presented on mini-APC exosomes, should now be considered to possibly be involved in several other atopic disease imitators, that apparently are not due to classical IgE/mast cell mechanisms. Bacterial Extracellular Vesicles (mEV) Promote Atopic Dermatitis Remarkably, allergens can also be transmitted by extracellular vesicles originating from the microbiome, that can significantly drive clinical allergic diseases. As an example, Staphylococcus aureus is known to be strongly associated with the pathogenesis and progression of atopic dermatitis. Accordingly, anti-staphylococcal antibiotic therapy and/or daily immersion in Clorox mixed in a bath are routinely prescribed for patients with this disease. It has been known for a number of years that bacteria secrete mEV, that now are seen as related to exosomes. mEV are formed at the cell surface from budding off of a very thin tissue cell layer over the tough peptidoglycan outer calix that surrounds the bacterial cell membrane for maintaining shape and protecting against challenging environments and enemies. The mEV are subsequently extruded as vesicles containing signaling molecules for quorum sensing by other bacteria, leading to promotion of microbial growth and spreading. They also contain enzymes and toxins for fighting enemies, and importantly RNAs, as recently demonstrated [15,16]. These extracellular vesicles enable the bacteria to mediate long-distance delivery of toxic cargo with minimized dilution or degradation in transit to then optimally affect other organisms. Of clinical interest, these properties allow for the transmission of bacterial factors mediating invasion and virulence to act at the host-pathogen interface, and to participate in some allergic diseases, like atopic dermatitis [17,18]. Accordingly, application of mEV from Staphylococcus aureus to mouse skin, previously stimulated by local tape stripping and then Ag application to express a murine model of atopic dermatitis, results in augmented development of the characteristic eosinophil-rich inflamed lesions along with the increased local levels of Th2 cytokines [17]. Further, these staphylococcal mEV are able to stimulate dermal fibroblasts to release stromal-derived pro-allergic inflammatory mediators, like TSLP, IL-6, MIP-1a and eotaxin, and also to induce IgE responses to Ag, particularly found in the mEV, compared to the staphylococci [17]. In a related study, a large group of patients with atopic dermatitis was found to be reactive to Malassezia sympodialis fungal allergens, that were shown to be carried by mEV released by this yeast [18]. Like staphylococci, this yeast is considered as a part of the physiological cutaneous flora. These allergen-bearing fungal exosome-like vesicles isolated from the supernatant of the co-culture of Malassezia sympodialis and monocyte-derived dendritic cells collected from patients with atopic dermatitis, are able to stimulate the in vitro production of IL-4 by cultured peripheral blood mononuclear cells (PBMC) from these patients [18]. In an analogous manner, the mEV of the nasal microbiome may participate in allergic rhinitis to promote this most common Type I allergic disease. Accordingly, molecular evidence based on bacterial 16s-ribosomal sequencing shows a possible relationship between bacterial type and clinical outcome, and S. aureus dominance is being associated with development of chronic rhinosinusitis [112]. Taken together, these data suggest that infectious agent-derived mEV transport microbial allergens and likely deliver functional RNA and proteins that are involved in the pathogenesis of atopic allergic diseases. Helminthic Worm Exosome miRNAs Inhibit Host Th2 Immunity To Mediate Molecular Parasitism Related to the role of exosomes in allergy is recent work showing that the mucus oily surface secretion of an intestinal helminth, called “worm spit”, known to be associated with its pathogenicity, contains exosomes with pro-parasitic and anti-allergic miRNAs [113–115]. It is known that such worm-derived exosomes are taken up and internalized by cells of the parasitized host [114]. The recently uncovered mechanism of the molecular pathogenic role of these parasite miRNAs is to specifically bind to the 3’ untranslated region (UTR) of targeted specific mRNAs in the host, to antagonize them or cause their degradation. Thereby they alter gene translation epigenetically to then affect host cell function by decreasing production of targeted gene encoded proteins [113]. This provides a newly recognized exosome mediated molecular mechanism of parasitism; acting in this case by inhibiting translation of Th2-associated genes, like those encoding IL-33, which is an IL-1 related cytokine that normally enhances Th2 immune responses. This contributes to down regulation of innate Type II immune responses that are known to be associated with repelling worm infestation, and simultaneously promoting anti-worm Th2 responses [113]. Therefore, it has been proposed that extracellular vesicles of helminth origin, that are internalized by host immune cells, transfer inhibitory miRNAs for genetic down regulation of Th2 associated host-protective proteins. Clinically, the possible eventual use of these particular miRNAs, perhaps delivered in therapeutic exosomes, could be applied to treat Type I allergies in patients, employing natural and physiological mechanisms. Other related work also demonstrates the role of exosomes in the evolution of parasitism. It was shown that Leishmania constitutively secrete exosomes in the gut of the sand fly vector to become a part of the inoculum transferred by the fly into the host during the insect’s bite [116]. The co-transferred exosomes influence the infectious process by inducing host inflammation, particularly via IL-17a. The Leishmania-derived exosomes are an integral part of the parasite’s infectious life cycle, acting as newly recognized virulence factor vesicles that are associated with this vector-transmitted infection [117]. Prior work has demonstrated that Leishmania parasitism is augmented by production of exosomes modulating host innate and adaptive immune responses through their effects on monocytes and DC [117]. Conversely to helminth exosomes that block host Th2 responses, vesicles shed by Trypanosoma cruzi parasites increase parasitism in heart and generate an intense inflammatory response due to their induction of Th2 cell production of IL-4 and IL-10 [118]. Finally, it was shown that Trichomonas vaginalis-derived exosomes also deliver factors to the host that favor their infectivity [119]. Role of Exosomes and Contained miRNAs in Interactions Between Airway Tissue Cells and Infiltrating Asthma Pathogenic Immune and Inflammatory Cells As noted above, asthma can be augmented due to allergen presenting exosomes. The disease can progress in sensitized individuals due to chronic airway exposure to inhaled aerosol allergens. This consists of either limited seasonal contact with allergens of outdoor plants or year round contact with indoor allergens from house dust mites and the dander of mice and pets. Beyond the acute clinical bronchospastic phase, asthma is associated with “late phase”, eosinophil rich airway inflammation. These chronic manifestations can lead to airway tissue remodeling with fibrosis and airway muscle sensitization that induces easily provoked bronchoconstriction; all significantly impairing the life and comfort of patients. These phases of asthma are classically noted to be mediated by IgE-dependent, early immediate hypereactivity, and subsequent late phase inflammation resulting from IgE and Th2 cell-dependent mechanisms. This latter process, that is critical to the generation of chronic asthma, includes crucial interactions between local tissue cells of the airway microenvironment and the infiltrating immune inflammatory constituents. These interactions are now recognized to be mediated, at least in part, by released extracellular vesicles that transfer their miRNA and protein content between the involved cells [Fig. 1]. This likely includes exosomes exchanging information between the intrinsic lung tissue cells of the local airway microenvironment, and extracellular vesicles derived from infiltrating immune and inflammatory cells emigrating to the lungs from the bone marrow and lymphoid tissues. The local microenvironmental exosome sources include cells of airway smooth muscles, vasculature, and epithelium, as well as mast cells, goblet cells, DC and alveolar macrophages. We postulate that these wider than previously anticipated interactions produced by exosome mediated intercellular exchanges between cells of the tissue microenvironment and the infiltrating immune and inflammatory cells are important in the development of chronic airway remodeling. This is analogous to the newly expanded ideas of tissue stromal cell responses to cancer cells spreading metastasis [120,121]. Interestingly, this can depend on chronic exposure to tumor cell-derived exosomes delivering proteins or RNAs, that act on the tissue cells. It was demonstrated that exosomes from particular cancers act at a distance to fuse specifically with resident cells of the preferred organ to generate a pre-metastatic niche. This appears to be due to specific, Darwinian selected exosome surface integrins. These could be considered analogous to bar codes for specific recognition affinity of particular other cells [122]. By analogy, cells within an allergic tissue, such as in the late or especially remodeling phase of asthmatic airway inflammation, could generate a bronchial microenvironmental multicellular-derived cloud of exosomes with specific affinities for various local, or distant, targets driving the allergic tissue response [Fig. 1]. Thus, the local airway tissue cells are often the targets of immune and inflammatory cell-derived cytokines inducing production of chemokines and enzymes, like in chronic obstructive pulmonary disease (COPD) [123], and further often can be the targets of exosomes [81,124,125]. BALF Analysis Suggests Participation of Exosome Intercellular Exchanges in the Pathogenesis of Allergic Asthma and Perhaps its Late Sequela Exosomes derived from the BALF of healthy individuals differed from those of patients with even mild asthma, that expressed higher levels of exosome-associated markers, such as the tetraspanins CD63 and CD81, and the scavenger receptor CD36 [126]. Although there were no major differences between BALF exosomes from asthmatics before and after birch pollen allergen provocation, they contained enzymes for leukotriene biosynthesis that were able to promote LTC4 release, as well as synthesis and release of the neutrophil-recruiting IL-8. Importantly, a leukotriene receptor antagonist (montelukast) reduced exosome-induced secretion of this pro-inflammatory cytokine that behaves like a neutrophil chemokine [126]. Furthermore, in a mouse model of asthma, exosomes isolated from the BALF of asthmatic versus control mice had increased levels of exosome-associated proteins. The asthma featured enhanced secretion of exosomes by bronchial epithelial cells, but not macrophages. Stimulation with IL-13 to imitate the function of specific allergen-pulsed Th2 cells, which is common in asthma, enhanced asthma severity [76]. This resulted in the release of exosomes from the bronchial epithelial cells that induced proliferation and chemotaxis of macrophages. These effects were suppressed by treatment with the ceramide blocker GW4869 [76], an inhibitor of neutral sphingomyelinase-2 that is essential for exosome formation. Thus, it is known to inhibit exosome transfers [127–129], and indeed, it alleviated allergic airway inflammation in these asthmatic mice [76]. Interestingly, isolated wild type mast cells stimulated with Ag for typical mediator release were strongly inhibited in sphingomyelinase deficient mice [130]. However, a possible role of an activating exosome exchange between mast cell subtypes for mediator release was not yet investigated. Pertinent to possible exosome-derived miRNA regulation of asthma are experiments showing that allergic disease resulting from in vivo activation of TLR4 by house dust mite Ag leads to expression of an unique subset of miRNAs. Remarkably, selective blockade of miRNA-126 alone with specific antagonists inhibited the asthmatic phenotype. This resulted in diminished Th2 inflammation, airways hyperresponsiveness, eosinophil recruitment, and mucus hypersecretion [131]. The data suggest that targeting miRNA in the airways may lead to anti-inflammatory treatments for allergic asthma. In a follow up study using the ovalbumin (OVA)-induced asthma model, specific inhibition of prominently involved miRNA-155 with antagomiR treatment failed to alter the disease phenotype, leading to the conclusion that the level of a particular miRNA may not indicate its importance and that blocking is needed to ascertain the functional significance of a particular miRNA from exosomes [132]. Interestingly, there was the variable efficacy of the ant-miRNA treatment across different immune cell types. Thus, effective targeting of myeloid cells but not lymphocytes indicated the possible need for cell type specific targeting of such therapy. Taken together, these data show that exosomes produced by various cells in the airways have an important and previously unrecognized role in the pathogenesis of the acute early phase of atopic allergic asthma. Since exosome effects are often mediated by intercellular transfer of miRNAs, these findings raise the possibility of new therapies at an entirely new level; namely blocking epigenetic effects of miRNAs with appropriate polynucleotide antagonists, as we demonstrated in CHS [2,7]. Potential Role of Exosomes in Airway Remodeling During Late Phase of Asthma Inflammation Exosomes produced by eosinophils may possibly contribute to tissue remodeling during late phase of asthma [78,96,133–135]. Eosinophils collected from blood of asthmatic patients that were stimulated in vitro with cytokines, had increased release of tetraspanin CD63pos exosomes [135]. Potential augmenting of the number of eosinophil-derived exosomes can alter and even cause rapid necrosis of structural lung cells, that not only can exacerbate allergic asthma, but especially can lead to airway remodeling [133]. These effects may be larger than expected, since it is possible that some of the granules thought to be produced by eosinophils may be in fact exosomes. In an analogous manner bronchial epithelial cells seem to produce proinflammatory exosomes that activate macrophages in the inflammatory phase via the transfer of miRNA. Th2 cell-derived IL-13 stimulates airway epithelial cells to release exosomes that transfer Let-7 miRNA to activate macrophages in the lungs during asthmatic inflammation [136]. During the inflammation neutrophils may also contribute to subsequent remodeling. Here, the LTB4-induced neutrophil chemotaxis, their released exosomes contain LTB4 and its synthesizing enzymes [137], and can activate neighboring resting neutrophils in a LTB4 receptor-dependent manner [138]. Thus, the exosomal pool of LTB4 acts in an autocrine fashion to sensitize neutrophils towards a primary chemoattractant, and then in a paracrine fashion to mediate recruitment of neighboring neutrophils to augment inflammation. Therefore, this forms an exosome intercellular mediator cascade effect, augmenting inflammation pertinent to the late phase of asthma and perhaps contributing to the progress of remodeling [139]. Accordingly, use of drugs to inhibit exosome production might interfere with this deleterious augmenting cascade. Further, airway administration of LPS causes acute injury with recruitment of neutrophils into lung and BALF [140]. Finally, neutrophils stimulated with LPS produce exosomes that alter proliferative properties of airway smooth muscle cells [125]. In an analogous manner, LPS activates bronchial epithelial cells of asthmatic patients to release exosomes that contain overexpressed pro-angiogenic tissue factor stimulated by compressive stress, that possibly also contributes to airway remodeling during chronic asthma [124]. Role of Bacterial mEV of Dust in Airway Remodeling during Atopic Allergic Asthma Inhalation of allergens and other entities in dusts is a common exposure contributing to the pathogenesis of asthma in allergic individuals. Indoor dust contains not only house dust mite Ag as the dominant allergens, but also mEV components from Gram negative bacteria derived from the GI tract of humans and pets. Such an occurrence contributes to an expansion of the hygiene hypothesis. This proposes that our changed relationships to infectious agents are a part of processes that have led to a great increase in some diseases of Western Civilization, like allergies, and particularly asthma. In homes that are a part of dairy farms, the dust contains mEV LPS from Gram negative bacteria of the cattle. The LPS containing mEV are likely involved in an asthma-protective effect of chronic low dose exposure to the farm dust. This can happen in children growing up on a dairy farm near the animals, which protects them from allergic rhinitis and asthma [141]. The protective effect depends on mEV LPS induction of the ubiquitin-modifying enzyme A20 in lung epithelium [142]. Thus, the farming environment protects from allergy by modifying nanovesicle intercellular communication between barrier lung epithelial cells and then DC through the induction of A20 by LPS in mEV. Such a chronic exposure to a very low dose of mEV LPS in farm dust protects experimental mice from developing asthma in a model induced by house dust mite Ag. The mEV-protected mouse airway epithelial cells produced less cytokines to activate DC, thus suppressing Th2 reactivity to house dust mites. In contrast, repeated airway administration of LPS mEV isolated from indoor house dust to mice led to neutrophilic pulmonary inflammation with infiltration of Th1 and Th17 cells [143]. The vesicles were internalized by airway epithelial cells and alveolar macrophages. This process was blocked by treatment with polymyxin B antagonist of LPS in the mEV of Gram negative bacteria [143]. In this study, serum levels of IgG1 Ab reactive with the dust vesicles were significantly higher in atopic children with asthma, than in healthy children and those with rhinitis or dermatitis [143]. Thus, nanovesicle intercellular communication, like that mediated by mEV LPS in environmental dust, can be crucial in tipping the balance between development of Th2 versus Th17 dominated asthma [144]. Importantly, IL-17 is known to recruit and activate neutrophils and is an important mediator of chronic inflammation, so in the present context, it is another contributor to potential asthmatic airway remodeling. Interestingly, sphingosine-1-phosphate inhibits murine endotoxin-induced inflammatory lung injury. This likely is because of formation of exosomes and involves inhibitory G protein-coupled sphingosine-1-phosphate receptors [145] that regulate exosomal maturation in MVB [146]. As mentioned above, the release of exosomes depends on ceramide biosynthesis regulated by neutral sphingomyelinase-2 and thus can be inhibited by its inhibitor GW4869; both in vitro and in vivo [76,147], as already employed in a study on the role of exosomes in allergic airway inflammation in a murine model [76]. Yet another possibly was brought by the recent demonstration that mEV contain RNA [14,15], and can release this RNA [16] to possibly function or trigger TLR4 stimulation in mEV-induced airway remodeling in asthma. Accordingly, we predict that bacteria-derived mEV containing LPS and epigenetic effects of their delivered RNAs will be found to be an important causative agent for asthma remodeling. Of related interest, Staphylococcus aureus-derived mEV, that play a significant role in the pathogenesis of atopic dermatitis [17], also have been shown to induce neutrophilic pulmonary inflammation via both Th1 and Th17 cell dependent pathways [148,149]. This could be important in chronic phases of allergic asthma and in some forms of non-atopic asthma, with subsequent remodeling. Taken together, these findings suggest that exosomes derived from endogenous airway stromal cells and infiltrating inflammatory cells of late phase asthmatic inflammation, and or together with mEV, and possibly bacteria toxins, stimulating release of lung tissue-derived exosomes, can participate in the sensitization for airway hyperresponsiveness and tissue remodeling changes in allergen driven chronic asthma. Such a remodeling is particularly difficult to treat clinically, especially using inappropriate anti-mediator and anti-inflammatory agents aimed at earlier phases. Thus, the new realization of a potential role for exosome delivery of miRNAs in remodeling opens a new pathway that might possibly lead to entirely new and likely more effective therapies aimed at impairing the consequent epigenetic mechanisms of exosome transferred miRNAs. Also, documentation of participation of mEV, and possibly TLR agonists, in the evolution of remodeling in allergic asthma might lead to more vigorous use of antibiotics, perhaps administered to the lung topically in an aerosol, or specific bacterial vaccination, perhaps with mEV, to possibly impede this process, again by analogy with atopic dermatitis. Exosome RNA Biomarkers for Predicting and Following Allergic Asthma The use of miRNAs as biomarkers that are carried in extracellular vesicles from the blood [149], BALF [150] and airway epithelium [151] of asthmatic patients were recently described, and similarly in serum derived exosomes from mice with asthmatic airway inflammation [152]. In addition, this biomarker information of respiratory tract conditions was collected non-invasively, and very innovatively, by assaying the miRNAs of the exosomes from exhaled breath condensate from the patients. Samples were compared from patients with Type I allergic asthma, COPD, pulmonary tuberculosis, and from normal individuals [153,154]. The data showed potential distinctive alterations of miRNA profiles in exhaled exosomes from asthma patients; and very importantly, even those with mild or early disease. Thus, miRNA profiles of exosomes from BALF, airway exhalate and blood might serve as excellent asthma biomarkers, especially useful in possible early detection of components determined to be predictors of remodeling. If particular exosome miRNAs were known to be such predictors, then perhaps airway aerosol treatments with hybridizing polynucleotides of particular specific anti-miRNA sequence might prevent the process. It is established that extracellular miRNAs in the blood are carried in exosomes. However, some of the blood miRNAs are transported not by exosomes, but by chaperone hydrophobic proteins, like argonautes and lipoproteins, for similar protection from RNases [155–158]. We found that these non-exosomal miRNAs act to suppress in vivo immune inflammatory responses by associating with exosomes produced by neighboring cells to then target effector cells, and, like those in exosomes [2], were susceptible to anti-miR treatment [7]. In summary, the use of exosomes for development of useful clinical biomarkers that would aid early detection of susceptible individuals prone to develop more significant asthma and late tissue remodeling, and the use of entirely new modes of specific therapy based on inhibition of particular miRNAs with specific antagonists, would be revolutionary. Such new procedures may prevent the progression of airway remodeling from its current non-reversible state and possibly reduce the need for extensive chronic treatment with anti-inflammatory steroids and bronchodilators. Exosomes in Immune Tolerance and Suppression of Asthma Since miRNAs usually down regulate mRNA translation they often mediate suppression of immune responses. Therefore, exosomes delivering miRNAs play a particularly prominent role in immunoregulation, mostly in tolerance and suppression. As an example, BALF exosomes (called “tolerosomes”) obtained from donor mice tolerized to olive pollen peptide by repeated intranasal administration of high dose of Ag significantly suppressed allergic airway reactions when administered to mouse recipients with a model of asthma [159]. The tolerogenic exosomes decreased allergic eosinophil-rich bronchial tissue inflammation and reduced production of IgE Ab and Th2 cytokines. The observed suppression likely was Ag-non-specific, since mice pretreated by intranasal administration of BALF-derived exosomes from donors tolerized to the allergen-derived olive pollen peptide were protected from development of airway allergic inflammation after subsequent sensitization with another allergen, i.e. birch pollen peptide. However, the Ag cross-reactivity between both of the pollen Ag could possibly occur. In the analogous system of Ag high dose tolerance in a Th2-type response described above, treatment of recipient mice before sensitization with intestinal epithelial cell-derived tolerosomes found in serum of mice orally tolerized to OVA [160] induced oral tolerance. Further, as a possible example of a relationship between bacterial mEV and asthma tolerance, neonatal exposure to the enterotoxin superantigen of Staphylococcus aureus augmented subsequent induction of oral tolerance to OVA in a mouse model of asthma [161]. In contrast, exosomes from an epithelial cell line cultured in the presence of an OVA hydrolysate of peptides with IFNγ were not tolerogenic, but instead activated the humoral Ab immune response to OVA [162]. This suggested that such exosomes can tilt a tolerogenic versus immunogenic response by pre-incubation of donor cells with specific cytokines. Additionally, cytokines can be transported between cells by nanovesicles. In a neurological system, extracellular vesicles from neural stem cells were shown to be able to transfer IFNγ bound to its receptor, in order to activate Stat1 signaling in the targeted cells [65]. Exosomes of Breast Milk In Allergy and Immunity Breast milk contains a rich pool of exosomes carrying numerous immune-related miRNAs that are considered to be involved in the development of immunity and allergic disease responses [163,164]. Since miRNAs mediate inhibition of mRNA translation, they often are involved in immunosuppression, as is thought true for those present in exosomes of breast milk. By definition the miRNAs in these exosomes are protected from the GI tract RNases and other degradative conditions [165]. As mentioned, exosomes, as ancient particles of life, resist many harsh conditions that cells do not survive, such as the acidic pH of the stomach [28,29], even down to pH = 1 [30]. Thus, the miRNAs in breast milk exosomes can survive normal neonatal ingestion to then mediate immunoregulation in the neonate that is related to the microbiome being acquired and the possible development of allergies. It is established that colostrum delivered from mother to neonate just after birth is rich in IgA Ab protective against GI tract and skin infectious agents. This provides the infant with the mother’s prior humoral Ab experience. Colostrum is also known to contain exosomes [166,167] that can transfer different immune information, likely regarding cellular immunity that is delivered to the neonate during breast feeding. This is postulated to be related to cell-mediated responses aiding neonatal regulation of the microbiome that the child is acquiring and sharing with the mother. Accordingly, mother’s milk transfers various factors of immunosuppressive functions that likely protect the neonate from over reactivity to the great load of highly stimulatory foreign Ag of the microbiome they are acquiring. Furthermore, breast milk exosomes also seem to provide protection that is needed for control of unnecessary and deleterious allergic responses to foreign Ag. This likely pertains particularly to food Ag that are also being encountered for the first time. If not controlled, responses to new microbial Ag and food allergen exposures might lead to clinical hypersensitivity or allergy later in life. Functional analysis has revealed that the mother’s milk exosomes can inhibit T cell production of IL-2 and IFNγ following MHC stimulation [168]. In addition, incubation of the milk vesicle preparations with peripheral blood mononuclear cells (PBMC) increased the number of suppressive Foxp3pos Treg cells. In studies extending these findings, breast milk exosomes were shown to contain functional inhibitory TGF-beta [169], along with miRNA that can promote thymic Treg cell maturation [170], and that their immunosuppressive action likely inhibits innate immune functions of macrophages that may be acting as APC [171]. Further, mother’s milk exosomes contain miRNA-17, part of the miRNA-17–19 cluster, which is important for the activity of myeloid suppressor cells, as well as miRNA-181a, a modulator of TCR sensitivity to Ag. This acts partly through down regulation of phosphatases, which leads to elevated phosphorylation of intermediate signaling proteins and a consequent reduction of the TCR signaling threshold [172]. In addition, the miRNAs transferred by milk exosomes include miRNA-155 that is a distinctive regulator of T- and B-cell maturation and of the innate immune response [173]. miRNA-155 mediates inhibitory function in several other aspects of immune responses, including allergies. Such inhibition is in part due to the miRNA-155 control of the expression of the specific Treg transcription factor FoxP3, and consequent IL-4 signaling, together with immunoglobulin class switching to IgE and FcγRI expression [170]. These findings confirm that a dominant function of breast milk exosomes is to genetically control neonatal immune reactivity, inflammation and likely allergy via intercellular transfer of miRNAs focused on epigenetic modulation of cellular immunity relevant to the newly colonizing microbiota of the GI tract and exposure to food allergens. Therefore, isolation and characterization of inhibitory mother’s milk-derived exosome-derived miRNAs could then be developed into natural clinical candidate agents mediating completely new therapies for down regulation of atopic sensitization and Th2 effector immune responses associated with production of IL-4, IL-5 and IL-13. Accordingly, natural mother’s milk miRNAs may induce pivotal immunoregulatory and epigenetic modifications required for long-term central thymic Treg cell lineage commitment explaining the atopy-protective effect of mother’s milk. Since exosomes are poorly immunogenic and miRNAs are universal, this applies to raw allogeneic cow’s milk consumption as well [174]. Further, bovine milk components may also target APC in the neonate, since its exosomes have been shown to influence macrophages [171], and this might affect food allergy to milk too. These ideas offer a new option for the prevention of atopic diseases, for example by the addition of physiological amounts of miRNA-155-enriched exosomes into infant formula for mothers that are incapable of breastfeeding. In addition, bovine milk extracellular vesicles may show therapeutic effects since they were found to affect the course of mouse arthritis [175]. However, some studies with breast milk have pointed out that exosomes carrying allergen can in some circumstances augment rather than suppress immune responses, and thus may contribute to the development of allergy in infants. This may explain the higher incidence of neonatal allergy transmission from the mother than the father. Notably, the composition of exosomes in human breast milk differs greatly according to the mother’s sensitization status and lifestyle, that then can influence the risk of allergy development in breast milk-nourished children [174]. The recent rise in allergic diseases in infants of developed countries may in part be due to the reduced or different microbial exposure during early life and consequent alteration of gut microbiota-derived mEV [14], which is relevant to the “hygiene hypothesis”. Since the microbiota content of the infant’s GI tract likely plays a critical role in the maturation and development of the immune system, mEV from these “self microorganisms” may possibly contribute to the risk of allergic and immune diseases. Therefore, breast milk feeding may provide important exosomes that can either directly modify the immune response to and composition of the intestinal microflora or influence the host to reduce the development of allergic diseases. Further, mEV of the microbiome may be involved in its important role in the development of the neonatal immune system as well as aspects of the endocrine and nervous systems. Indeed, it was recently demonstrated that exosomes delivered in breast milk can promote the colonization of the newborn intestine with physiological microbial flora, that in turn seems to reduce the risk of allergy development and thus is crucially involved in the effects of the microbiome by the release of mEV or affection of intestinal production of host exosomes that transmit protective miRNAs [176]. Of further interest, it was recently proposed that polymorphisms in susceptibility genes in humans promote inflammatory bowel disease development causing impaired sensing of protective signals derived by the mEV, which confirms the significant role of microbiome-derived vesicles in the maintenance of immune homeostasis in digestive system [177]. The Role of Exosomes in Delayed-Type and Contact Hypersensitivity That Can Replace APC Function With Simpler “Mini-APC” vesicles Delayed-type hypersensitivity (DTH) reactions exemplify mechanistic steps underlying Th1-mediated immune diseases and also defenses against pathogenic intracellular bacteria, fungi, parasites and some viruses, that must include intercellular actions of exosomes [89]. Major examples of exosome involvement in cell-mediated immunity are represented by DC that were pulsed with native Ag of infectious agents [88,178], to then be used as non-cellular potent vaccines substituting for DC. Therefore, DC-derived nanovesicles competent to present peptide/MHC complexes, and thereby acting like parental DC, were used to induce immune responses against tumor cells [90–92], and also inhibition of inflammatory responses to LPS [179]. We are not yet aware of the use of allergen pulsed DC or other APC to produce Ag-presenting exosomes to desensitize against allergens, as is traditionally done by administration of ascending doses of soluble Ag in buffer for induction of tolerance. However, this can definitely be expected. Such a mini-APC approach should be far safer, since allergen peptides are being given in a form that should not readily trigger IgE/mast cell responses that are more directed at the specificity of IgE to conformational determinants of native allergens. Further, such a mini-APC vesicle method likely will be far more effective due to the direct facilitation of Ag presentation not requiring the processing of native allergen. Accordingly, exosomes derived from OVA-pulsed mouse bone marrow DC express OVA-peptide complexed in MHC. Administration of this exosome peptide-MHC complex leads to the polarization of immune response towards Th1 phenotype [82], and thus may be effective clinically in treating Th2-mediated allergic diseases. As noted above, exosome-like mEV can influence cell-mediated immune responses as agents of the microbiome that shape and modulate the immune response. This approach takes advantage of the fact that mEV can be manipulated for their immunogenic contents for utilization as potent pathogen-free vaccines for immunizing humans and animals against infectious agents. Accordingly, DC, that are pulsed with Toxoplasma gondii [180], release to the culture supernatant immunostimulatory exosomes that can replace DC. Additionally, Escherichia coli release mEV [181], which, after injection into mice, are readily taken up by APC to efficiently stimulate specific Th1-dependent immune response preventing bacterial infections, and likely allergy development as well. On the other hand, DC can be modified to produce immunosuppressive exosomes. Thus, administration of exosomes released by mouse DC genetically modified to express inhibitory Fas Ligand resulted in suppression of DTH elicited in mice immunized by intradermal injection of OVA or keyhole limpet hemocyanin (KLH) Ag emulsified with Freund’s complete adjuvant [182]. Similarly, exosomes derived from cultured DC stimulated with IL-4 [183] or IL-10 [184] inhibited DTH induced by KLH with adjuvant. This effect was found to be dependent on expression of the co-stimulatory molecule B7 (CD80/87) on the cytokine treated DC [184]. It was concluded that B7, but not PD-L1/L2, on the surface of IL-10-treated DC and their derived exosomes play a critical role in the observed immunosuppressive functions. In related in vivo studies, systemic injection of IL-10-containing DC-derived exosomes suppressed the onset of murine collagen induced arthritis (CIA) and reduced the severity of already established arthritis [185]. Taken together, these data suggest that DC secrete exosomes that can be tailored to suppress inflammatory and autoimmune responses. In further study, indoleamine 2,3-dioxygenase (IDO), that is a tryptophan-degrading enzyme important for immune regulation and tolerance maintenance, was expressed in DC and their exosomes inhibited DTH-effector T cells by depleting them of essential tryptophan and/or by producing toxic metabolites, as well as by generating Treg cells [186]. Accordingly, exosomes derived from these IDO-positive DC suppressed DTH and CIA. The suppressive effects were partially dependent on B7 co-stimulatory molecules. In addition, gene transfer of CTLA-4 immunoglobulin to DC resulted in induction of IDO in the DC releasing exosomes that were able to reduce inflammation in an IDO-dependent manner [186]. These results demonstrate that both types of IDO-expressing DC-derived exosomes are immunosuppressive and anti-inflammatory, and consequently are able to reverse established arthritis. Therefore, exosomes from IDO-positive DC may represent a novel therapy for rheumatoid arthritis, and highlight studies showing that altered exosomes from in vitro manipulated DC can significantly influence DTH and related clinical disease models, like CIA. As noted above, rats fed with OVA develop a tolerogenic activity in serum that can transfer tolerance to OVA and suppression when given at the time of induction of an immune response, and more over can abolish an established DTH response in the recipients, as well as their humoral IgG responses. The mechanism possibly involves inhibitory exosomes (“tolerosomes”) acting through CD25pos Treg cells and produced by GI epithelial cells from rats undergoing induction of oral tolerance to OVA. Further, this exosome induced tolerance seems to be MHC-restricted, and probably Ag non-specific, since tolerosomes from animals fed with OVA suppress DTH induced with OVA or with human serum albumin [187,188]. In contrast, exosomes from the culture supernatant of a tumor cell line manipulated genetically to express OVA, when injected at the time of DTH elicitation, were able to suppress mouse OVA-specific DTH induced as above, but not DTH induced by a KLH-adjuvant mixture, and thus acted in an Ag-specific manner [189]. As noted, tumor cell-derived exosomes usually contain tumor Ag and have been used as mini-APC to stimulate anti-tumor immune responses [90–92]. However, it is unclear if the tumor-derived exosomes can actually facilitate tumor immune evasion Ag specifically. Here, the tumor-derived OVA-expressing exosomes were internalized by CD11cpos DC and transported to the draining lymph nodes to likely induce TGF-β 1 and IL-4 producing Ts cells that seemed to modulate the APC to express Ag-specific inhibitory function [189]. Further, exosomes isolated from plasma of mice immunized to KLH, but not from naive mice nor OVA-immunized mice, could suppress KLH and not OVA-specific DTH, with the effect mediated by MHC class IIpos, Fas Ligandpos, CD11b pos, but CD11cneg plasma exosomes and, in part, dependent on the presence of Fas Ligand on the exosomes and Fas on the recipient mouse cells [190]. These results suggest that plasma exosomes likely produced by APC and expressing CD11b and MHC class II complexed with tumor cell-derived peptides can suppress immune response in a peptide Ag/MHC-specific manner partly through a Fas/FasL interactions. CD8pos Suppressor T Cell Exosomes Ag-Specifically Suppress CHS and DTH by Delivering Inhibitory miRNA-150 Current research of our laboratories aims to investigate the detailed mechanisms of Ag-specific suppression of mouse hapten-induced cutaneous contact hypersensitivity (CHS), as well as DTH that is induced by intradermal injection of protein Ag without an adjuvant. CHS is considered as a model that is relevant to human allergic contact dermatitis. Both are also considered to depend on mechanisms underlying cell mediated immune responses to proteins in a variety of diseases, such as asthma, atopic dermatitis and various autoimmune conditions. We have re-investigated previously observed suppression of CHS by an enigmatic Ag-specific factor secreted into culture supernatant by Ts lymphocytes harvested from mice tolerized by intravenous injection of high doses of Ag. This led to the discovery of suppressive exosomes carrying inhibitory miRNA-150 that are surface coated with Ab light chains responsible for their Ag-specificity [2,191]. Thus, we demonstrated that this enigmatic Ag-specific suppressor factor from mice rendered Ag-specifically tolerant by intravenous administration of a high dose of hapten Ag conjugated to autologous red blood cells, and subsequently immunized for CHS induced by the same hapten, was produced by CD8pos Ts cells as exosomes. These exosomes act by carrying and subsequently delivering inhibitory miRNA-150 Ag-specifically, due to their surface coating of Ab light chains, and interestingly not heavy chains nor whole Ab [2]. The suppressive exosomes could be isolated from either plasma of the tolerized animals or from culture supernatant of the Ts cells harvested from tolerized mouse spleen and peripheral lymph nodes [2,191]. Critically, inhibitory activity of the miRNA-150 was blocked by pre-incubation of the suppressive exosomes with an antagonist of miRNA-150, a polynucleotide of reverse sequence, i.e. anti-miR-150, compared to several controls [2]. The most important proofs were brought by experiments with miRNA-150−/− mice. Although normal appearing and able to be fully sensitized for elicitation of CHS, these animals could not be tolerized, and exosomes from these unsuccessfully tolerized mice were lacking suppressive function in vivo in CHS and in vitro in IL-2 receptivity of T cell line [2]. Also very crucially, in vitro supplementation of these non-suppressive exosomes from the miRNA-150−/− mice tolerized with high dose of hapten Ag, with synthetic miRNA-150 agonist rendered them suppressive [2,7]. Taken together, these findings show definitively that miRNA-150 delivered by the Ts cell-derived exosomes is the crucial mediator of the suppression and that surface Ab light chains are responsible for the Ag-specificity of their suppression. The CD8pos Ts cells producing the inhibitory exosomes are Foxp3neg [2], and thus are not conventional Treg cells, but instead originate from a separate lineage of regulatory cells that act to suppress CHS and DTH responses in vivo and in vitro [2]. A Special Small Subset of B Cells Produces the Ag-Specific Ab Light Chains That Coat the Suppressive Exosomes The Ts cell-derived exosomes are able to act as a preventive as well as therapeutic agent by inhibiting, respectively, sensitization and elicitation of CHS, and by alleviating of its symptoms [192]. However the Ts cell exosomes are not non-specifically pan-suppressive, since, while suppressing the Th1 effector cells mediating the classical late phase of CHS and DTH, they do not inhibit the required early phase mediated by B1a cells. These B1a cells are stimulated by IL-4 released rapidly after sensitization (by only18 minutes) by hepatic iNKT cells [107] that apparently are activated by glycolipid Ag swiftly released at the cutaneous sensitization site to then apparently migrate to the liver to preferentially target hepatic iNKT cells [193]. As stated above, miRNA-150 loaded exosomes act in a strictly Ag-specific manner. Definitive dual reciprocal testing showed that exosomes obtained from mice tolerized to the trinitrophenyl (TNP) hapten suppressed only TNP-induced CHS, but not oxazolone (OX)-induced CHS effector cells, and conversely, exosomes from mice tolerized to OX suppressed only CHS effector cells induced with the OX hapten, but not with TNP [2]. This unusual Ag-specificity has been unraveled by the discovery that the Ts cell-derived exosomes have a surface coating of hapten Ag-specific Ab light chains derived from B1a cells accompanying the CD8pos Ts cells. These B1a cells are activated by hapten-self complexes induced by application of the relevant hapten on mouse skin, which completes the tolerogenic procedure [2]. The strict Ag-specificity of the participating B1a cell subpopulation is due to activation induced deaminase (AID), a nuclear mutating enzyme, that acted on the DNA of these cells prior to immunization to cause mutations in the immunoglobulin V-region genes, thus encoding the Ag-specific combining sites of Ab heavy and light chains [108,194,195]. This compares to the usual strictly germ line V-region DNA sequences without mutations in the majority of typical B1a cells that are not acted on by AID, and produce so-called natural Ab. However, recent single cell sequencing of immunoglobulin V-regions of developing B1a cells has shown that many of these cells, more than expected by prior information, are acted on by AID [196]. However, the processes stimulating the occurrence of AID-dependent mutations before immunization require further investigation. They do not involve prior sensitizing interactions with the microbiome, as they occur in germ free animals. Thus, the presumed Ag activation, that stimulated this subpopulation of B1a cells, did not come from the standard external bacterial, viral or parasite-derived foreign Ag, but likely from unknown endogenous Ag [196]. Characteristically, the special B1a cells, having only a few immunoglobulin V-region mutations mediated by AID that account for their strict Ag-specificity, are able to respond within minutes of hapten-self primary skin immunization, documented to occur in separate systems by just 18 minutes [107], 30 minutes [108] or one hour [197] after the original skin sensitization. As a consequence, the exosome sensitizing Ab light chains with few mutations are of low affinity for Ag, compared to their associated isolated heavy chains or the parental whole Ab. However, arrangement as multiple adjacent on the exosome surface likely results in their greater Ag-avidity, sufficient to manifest the observed strict dual reciprocal Ag-specificity of the suppressive exosomes [2]. Altogether, the suppressive exosomes that inhibit allergic cutaneous CHS are an unusual and unique product of components from both T cells (the miRNA-150 containing exosomes) and B cells (the exosome coating Ab light chains). The crucial experiment determining the role of Ab light chains coating the suppressive exosomes attempted to induce typical Ag high dose tolerization in pan-immunoglobulin deficient JH−/− mice. These mice lack Ab and, therefore, Ab light chains to coat the exosomes. Thus, cutaneous immunization for CHS in JH−/− mice previously tolerized with high dose of Ag was not suppressed and their exosomes were not active [2]. Very importantly, suppression was restored by coating of these non-suppressive exosomes from tolerized JH−/− mice with monoclonal Ab light chains [2]. Of related translational clinical significance, in our prior work on immediate hypersensitivity, on models of CHS [198] and asthma [199], induced by cutaneous sensitization with hapten or protein [200], the rapid activation of the small subpopulation of Ag-specific special B1a cells induced early after immunization was observed in all of these instances. Their activation resulted in production of Ag-specific IgM Ab and the Ag-specific Ab light chains that coat the Ts cell-derived exosomes and that have the additional ability to sensitize cells; so far principally mast cells by surface coating for mediator release when Ag is added, in an analogous manner to IgE Ab [201]. Regarding the suppressor exosomes, Ag-specific Ab free light chains [2] are produced by previously activated B1a cells and have low numbers of AID-mediated V-region mutations [108,195], and bind to the surface of CD8pos Ts cell-derived exosomes in a manner that keeps their combining sites available and allows quite Ag-specific function [2]. Note that tests in Fc receptor deficient mice showed that these receptors are not involved in Ab light chain binding [201], as expected since they have no Fc portion. Instead, this binding may depend on surface lipid changes in the membrane of “activated” exosomes [2,7], or by other unknown processes that are induced by tolerization or immunization. In contrast, exosomes from spleen and lymph nodes of non-immunized mice do not have the ability to bind free Ab light chains. It is important to note that in related clinical work, it has been shown that various human leukocytes, like T cells, B cells and monocytes, can bind free Ab light chains [202,203], and it was shown that this depends on cell surface lipids [203,204]. Therefore, this subpopulation of B1a cells, stimulated by hapten contact skin application as the final step in the tolerance procedure or for induction of CHS, or the analogous maneuver of intradermal injection of OVA protein Ag to produce tolerance to this protein, is responsible for the generation of hapten or protein Ag-specific Ab for coating of Ts cell-derived “activated” exosomes in mice undergoing Ag high dose tolerization. Free miRNAs That Are Not In Exosomes Also Suppress Ag-Specifically by Final Association With Ag-Specific Exosomes In other studies we defined an alternate pathway of suppression mediated by miRNA-150 present in a mixture of miRNAs that are free of exosomes. This mixture of extracellular RNAs (exRNA) was obtained by phenol chloroform extraction of the exosomes derived from the Ag-tolerized mice. miRNAs free of exosomes are also protected from RNases, by complexing with a hydrophobic chaperone, like an argonaute protein [7,155–158]. Interestingly, miRNA-150 among the vast excess of other free RNAs and miRNAs in the exRNA mixture that are Ag-non-specific, nonetheless inhibited CHS effector cells Ag-specifically, as Ts cell derived exosomes [2]. This unusual finding of exRNAs acting Ag-specifically in the absence of exosomes was unraveled by demonstrating that the miRNAs became associated with Ag-specific exosomes produced by the Ag-specific B1a cells present in the mixture of CHS effector cells that made exosomes expressing surface Ab-like B cell Ag receptors (BCR). These mixed effector cell population also contained the Ag-specific CHS effector T cells and APC. These then were suppressed by the Ag-specific B1a cell-derived exosomes that had become associated with the miRNA [7]. These findings led to speculation that this represents an alternate pathway of Ag-specific targeting by freely circulating functional miRNA not contained in exosomes to regulate CHS effector T cells. In this case the miRNA among the exososme-free exRNA can be delivered to function in allergic CHS and other responses by its association with such B1a cell-derived, Ag-specific, originally non-suppressive exosomes. These exosomes were originally shown not to carry miRNA-150, but then were induced to gained it from the associating mixture of free exRNAs. Ag-specificity of the exosome-associated miRNA suppressive activity results from the presence of BCR on the B1a cell-derived exosome surface, allowing for their specific delivery to affect Ag-specific T cells in the CHS-effector cell mixture. The major miRNA fraction in the human circulation is free of exosomes, and instead is also protected from numerous RNases by complexing with chaperones, like argonautes and lipoproteins. This suggests that these miRNAs may similarly act by this alternative pathway of associating with or transfecting exosomes released by either cells neighboring the target cell population, or possibly even by the target cells themselves. Overall, this is the first demonstration that miRNA free of exosomes can be functional and gain Ag-specific cell targeting to affect genetic function. This alternate pathway appears to be the only natural mechanism yet described for passing functional free extracellular miRNA between cells. This pathway of free exRNA intercellular transfer contrasts with classical donor cell release of RNA-containing and cell-targeting exosomes. Since significant amounts of circulating exRNA are not in exosomes, these findings likely have important biological and immunological significance, and may be relevant to both pathogenesis and treatment of diseases. Determination That miRNA-150 Was Responsible for Exosome Suppression by Exploiting Their Ag-Specificity The discovery that coating of the suppressive exosomes with Ag-specific Ab light chains was responsible for their Ag-specificity then led to discovery of miRNA-150 as the crucial delivered inhibitory molecule. The Ag-specific Ab light chains were judged to enable Ag-specific binding of the suppressive exosomes. This allowed for their fractionation using Ag-affinity column chromatography [2]. The Ag-binding exosomes had all the suppressive activity [2]. Processing of these Ag-binding exosomes, versus the Ag non-binding exosomes, to mRNA, with further conversion to cDNA, and then DNA deep sequencing with subsequent bioinformatic analysis suggested consideration of miRNA-150 as the important inhibitory molecule in the Ag-specific suppressive exosomes. This was verified by blocking of exosome suppression with the anti-sense polynucleotide of miRNA-150 (anti-miR-150), compared to numerous controls. Final definitive confirmation was shown by experiments in Ag-tolerized miRNA-150−/− mice. The crucial experiment showed that association of the non-suppressive exosomes from the unsuccessfully tolerized miRNA-150−/− mice with miRNA-150, rendered them suppressive, in comparison to several controls as described above. This constituted a definitive proof that miRNA-150 is the crucial inhibitory molecule carried by the exosomes that is responsible for the suppression of allergic cutaneous CHS. APC Are the Target of the Ag-Specific Suppressive Exosomes Further experiments showed that the inhibited CHS effector T cells are not the direct target of the suppressive exosomes, but rather their companion Ag-presenting cells, represented in our initial experiments by Ag-carrying macrophages. These APC are altered to become suppressive by delivery of exosome miRNA-150 to then be able to Ag-specifically suppress companion Ag-specific CHS-effector T cells that are recognizing Ag/MHC complexes on the APC [205]. Finally, as confirmation of the above formulation, depletion of macrophages from the total CHS-effector cell population, prior to in vitro incubation with the suppressive exosomes, that precedes the adoptive cell transfer of CHS, abolishes the suppressive action of the Ts cell exosome-delivered miRNA-150 on the CHS effector cell mixture. This observation confirms that Ag-presenting macrophages are essential for transferring the subsequent suppressive signal of yet unknown nature [205]. In current follow-up studies employing DTH induced with OVA without an adjuvant, we have observed similar tolerance, allowing us to attempt to confirm the hypothesis that this mechanism requires the binding of OVA peptides on the APC surface by the Ab light chains on the suppressive exosomes. Accordingly, our preliminary evidence shows that the Ts cell exosomes inhibition of CHS involves the binding of Ag peptide determinants in the peptide/MHC complex on the surface of the APC by the Ag-specific Ab light chains on the exosome surface. This specific binding likely allows for internalization of the suppressive exosomes for delivery of their miRNA-150 into the Ag-presenting macrophages to tilt their gene function towards a suppressive pathway to inhibit the CHS effector T cells. Further preliminary data suggests that the Ag-presenting macrophages that received the Ts cell exosomes may suppress the CHS effector T cells via the release of secondary suppressive exosomes. These secondary vesicles as a part of suppressive exosome cascade finally target the CHS effector T lymphocytes, leading to inhibition of their function. Currently, the exact mechanism of the Ag-specific, miRNA-150-mediated suppression by the CD8pos Ts cell-derived exosomes affecting Ag-presenting macrophages to inhibit effector T cells is the focus of our research. In this regard, in a helper T cell system, pretreatment of Ag-primed macrophages with the Ts cell-derived, Ag-specific exosomes, as in the CHS suppression system above, results in the reduction of helper T cell dependent, macrophage-mediated induction of B cell Ab responses to corpuscular Ag [205]. Additionally, treatment of macrophages with suppressive exosomes causes the enhancement of the generation of reactive oxygen intermediates [206]. On the other hand, the mechanism of action of the suppressive exosome miRNA-150 cargo may depend on cytokine responsiveness of the final T cell targets. Accordingly, the suppressive exosomes or the mixed miRNA in their phenol chloroform extracts, or even miRNA-150 alone, inhibit the in vitro responsiveness of HT-2 cells, a T cell line, to IL-2 [2,7]. In contrast, exosomes derived from lymphoid cells of unsuccessfully tolerized miRNA-150−/− mice are inactive in this model and their suppressive activity is restored by addition of miRNA-150 itself [2]. These additional findings suggest that the mechanisms enumerated above, acting alone or together, may play a role in the effector T cell suppression begun by the CD8pos Ts cell-derived primary exosomes delivering inhibitory miRNA-150 [2,7]. Additional preliminary data show that we are able to similarly induce DTH by intradermal injection of soluble casein Ag from milk, and activate similar Ag-specific suppression mediated by exosomal miRNA-150. Note that clinical observations suggest that hypersensitivity to casein is an aspect of food allergy to milk products, that likely consists of a mixture of a Type I allergic response mediated by food Ag-specific IgE Ab, and an additional possible DTH-like component. Therefore, this new technology of established procedures for dissection of biological properties of functional miRNA carried by suppressive exosomes may enable us to examine and perhaps alter immunoregulatory pathways in food allergy to milk products in a new molecular way. Thus overall, our data have significant translational potential for greater understanding and potential construction of new therapies for diverse clinical allergic and immunological diseases based on the newly recognized, diverse and complex intercellular communications mediated by extracellular exosomes released by the involved cells, delivering RNAs and proteins to alter functions of acceptor cells involved in allergies and hypersensitivities. Concluding Remarks The pathogenesis of allergic and hypersensitivity responses relevant to human diseases appears to be significantly influenced by various exosome-dependent mechanisms. This principally involves exchange of functional RNA between cells that are nearby or at a distance. This leads us to the following conclusions. Firstly, appreciation of novel near or distant exosome release to mediate functional exchanges, along with established intercellular communication pathways via cytokines, provides new reasons to more strongly consider allergy and hypersensitivity responses in vivo in preference to in vitro; i.e. in vivo veritas. This leads to a preference for in vivo and ex vivo experiments in order to more fully appreciate the undoubtedly exosome mediated interactions between immune and allergy mediating cells, and the various stromal cells in their microenvironment, as well as other cells at a distance, that also are sending RNA genetic instructions via the circulation [Fig. 1]. Secondly, it seems fairly certain that there will be new types of diagnostic approaches to allergic and hypersensitivity diseases involving the analysis of the RNAome and proteome of extracellular vesicles, by which genetic instructions are delivered to other cells. These instructive RNAs either may be or may not be carried in exosomes in the peripheral circulation, or body fluids, like the BALF, and interestingly may even be present in the breath exhalate. Thirdly, understanding of allergic and hypersensitivity diseases will undoubtedly be better when integrating these newly recognized exosome intercellular exchange processes into prior findings and constructs. Thus, this new knowledge should lead to development of new preventive and therapeutic strategies beyond conventional modes of regulating and attenuating the immune response to specific allergens and Ag. Fourthly, our and other studies have determined that therapeutic exosomes are biologically active after delivery via different routes, that were not considered previously. Since exosomes are natural and physiological, they can resist hostile environments, can have long lasting effects and can penetrate tissue barriers that do not permit transit of cells. These possibilities include therapeutic approaches far more tolerable to the patient than routine allergy shots. Such potential exosome therapies involve essentially cell free, non-invasive approaches, with new routes of administration, such as oral, nasal or inhaled. For the first time there is a real prospect of being able to specifically alter cell functions genetically in a physiological manner. This would be a huge and significant advance with profound clinical consequences. This work was supported by NIH grants to P.W.A.: AI-59801, AI-07174, AI-76366-02, AI-1053786 and U54 DA036134, and by grants from the Polish National Scientific Centre number 2013/09/N/NZ6/00753 to K.N. and 2013/11/B/NZ6/02041 to K.B. Abbreviations Ab antibody Ag antigen APC antigen presenting cells BALF bronchoalveolar lavage fluid BCR B cell surface receptor for Ag CHS contact hypersensitivity DC dendritic cells DTH delayed-type hypersensitivity PBMC peripheral blood mononuclear cells KLH keyhole limpet hemocyanin mEV microbial extracellular vesicles TNP trinitrophenyl Ts T suppressor cells Fig. 1 Postulated allergic exosome cloud in the airway of an asthmatic patient Proposed details of a postulated exosome cloud in the airways of asthmatic patients and other relevant tissues like the nasal mucosa. The tissue cloud consists of diverse exosomes derived from various cell sources that have a variety of cargos. The exosomes and other related extracellular vesicles in this cloud are postulated to be at a concentration of about 1010 vesicles per milliliter in the interstitial fluids between the various cells, shown as tiny spheres of different colors expanded at lower right. The extracellular vesicles from donor cells are able to transfer miRNAs, other RNAs, proteins etc. to other cells via the fluid between them to potentially alter the functions of the acceptor cells epigenetically. These diverse intercellular transfers of genetic information can be mediated by exosome-derived miRNAs and is potentially able to alter the development, maturation, activation and importantly function of other cells of various types. Some of these exosomes of the local cloud likely leave this tissue site to enter another cloud in the draining lymph to affect distant cells, such as those in the bone marrow or the cells of other organs like immune cells in the spleen or lymph nodes, via entering another cloud in the systemic circulation. Those in the blood are a mixture from all or most of the cells in the body to potentially serve as accessible clinical markers of disease. The circulating exosomes are headed for all possible sites, including those from elsewhere to the bronchial airway in asthma, shown in the figure at the right, along with exosomes from the endothelium, (expanded at lower right) to enter this site to join the local cloud to potentially exert effects on its constituent cells, thus in an endocrine manner. Note that the local cellular interactions not only involve the usual cross talk between lymphoid T and B cells and antigen presenting DC (some in immunological synapse, mid central right of the figure) and macrophages, with other myeloid-derived cells like eosinophils and mast cells. Very importantly, these immune and myeloid cells interact via mutual released exosomes with other local cells of the microenvironment, like bronchial epithelium, smooth muscle cells, as well as fibroblasts and other cells of the stroma that lie between all of the above cells. Further note that constituent cells are sending out exosomes at base line for physiological interactions and then other “activtated” exosomes when the cells are stimulated like the mast cells (at the left) responding to inhaled allergen (top right) or C5a, or cells responding to microbial extracellular vesicles from bacteria in the airways (top left). These newly recognized intercellular genetic interactions mediated by diverse exosomes in an intercellular cloud are present in addition to the already known exchanges based on conventional cytokines and their receptors, along with “micro-mediators”, like histamine, leukotrienes and C5a. 1 Choudhuri K Llodrá J Roth EW Tsai J Gordo S Wucherpfennig KW Kam LC Stokes DL Dustin ML Polarized release of T-cell-receptor-enriched microvesicles at the immunological synapse Nature 2014 507 118 123 24487619 2 Bryniarski K Ptak W Jayakumar A Pullmann K Caplan MJ Chairoungdua A Lu J Adams BD Sikora E Nazimek K Marquez S Kleinstein SH Sangwung P Iwakiri Y Delgato E Redegeld F Blokhuis BR Wojcikowski J Daniel AW Groot Kormelink T Askenase PW Antigen-specific, antibody-coated, exosome-like nanovesicles deliver suppressor T-cell microRNA-150 to effector T cells to inhibit contact sensitivity J Allergy Clin Immunol 2013 132 170 181 e9 23727037 3 Raposo G Stoorvogel W Extracellular vesicles: exosomes, microvesicles, and friends J Cell Biol 2013 200 373 383 23420871 4 Théry C Exosomes: secreted vesicles and intercellular communications F1000 Biol Rep 2011 3 15 21876726 5 Bobrie A Colombo M Krumeich S Raposo G Théry C Diverse subpopulations of vesicles secreted by different intracellular mechanisms are present in exosome preparations obtained by differential ultracentrifugation J Extracell Vesicles 2012 1 18397 6 Smith ZJ Lee C Rojalin T Carney RP Hazari S Knudson A Lam K Saari H Ibañez EL Viitala T Laaksonen T Yliperttula M Wachsmann-Hogiu S Single exosome study reveals subpopulations distributed among cell lines with variability related to membrane content J Extracell Vesicles 2015 4 28533 26649679 7 Bryniarski K Ptak W Martin E Nazimek K Szczepanik M Sanak M Askenase PW Free extracellular miRNA functionally targets cells by transfecting exosomes from their companion cells PLoS One 2015 10 e0122991 25923429 8 Conde-Vancells J Rodriguez-Suarez E Embade N Gil D Matthiesen R Valle M Elortza F Lu SC Mato JM Falcon-Perez JM Characterization and Comprehensive Proteome Profiling of Exosomes Secreted by Hepatocytes J Proteome Res 2008 7 5157 5166 19367702 9 Ogawa Y Taketomi Y Murakami M Tsujimoto M Yanoshita R Small RNA transcriptomes of two types of exosomes in human whole saliva determined by next generation sequencing Biol Pharm Bull 2013 36 66 75 23302638 10 Arita S Baba E Shibata Y Niiro H Shimoda S Isobe T Kusaba H Nakano S Harada M B cell activation regulates exosomal HLA production Eur J Immunol 2008 38 1423 1434 18425730 11 Mfunyi CM Vaillancourt M Vitry J Nsimba Batomene TR Posvandzic A Lambert AA Gilbert C Exosome release following activation of the dendritic cell immunoreceptor: A potential role in HIV-1 pathogenesis Virology 2015 484 103 112 26092249 12 Kulp A Kuehn MJ Biological functions and biogenesis of secreted bacterial outer membrane vesicles Annu Rev Microbiol 2010 64 163 184 20825345 13 Schwechheimer C Kuehn MJ Outer-membrane vesicles from Gram-negative bacteria: biogenesis and functions Nat Rev Microbiol 2015 13 605 619 26373371 14 Tyrer PC Frizelle FA Keenan JI Escherichia coli-derived outer membrane vesicles are genotoxic to human enterocyte-like cells Infect Agent Cancer 2014 10 2 15 Ghosal A Upadhyaya BB Fritz JV Heintz-Buschart A Desai MS Yusuf D Huang D Baumuratov A Wang K Galas D Wilmes P The extracellular RNA complement of Escherichia coli Microbiologyopen 2015 10.1002/mbo3.235 16 Sjöström AE Sandblad L Uhlin BE Wai SN Membrane vesicle-mediated release of bacterial RNA Sci Rep 2015 5 15329 26483327 17 Hong SW Kim MR Lee EY Kim JH Kim YS Jeon SG Yang JM Lee BJ Pyun BY Gho YS Kim YK Extracellular vesicles derived from Staphylococcus aureus induce atopic dermatitis-like skin inflammation Allergy 2011 66 351 359 20831718 18 Gehrmann U Qazi KR Johansson C Hultenby K Karlsson M Lundeberg L Gabrielsson S Scheynius A Nanovesicles from Malassezia sympodialis and host exosomes induce cytokine responses - novel mechanisms for host-microbe interactions in atopic eczema PLoS One 2011 6 e21480 21799736 19 Kahlert C Melo SA Protopopov A Tang J Seth S Koch M Zhang J Weitz J Chin L Futreal A Kalluri R Identification of Double Stranded Genomic DNA Spanning all Chromosomes with Mutated KRAS and p53 DNA in the Serum Exosomes of Patients with Pancreatic Cancer J Biol Chem 2014 289 3869 3875 24398677 20 San Lucas FA Allenson K Bernard V Castillo J Kim D Ellis K Ehli EA Davies GE Petersen JL Li D Wolff R Katz M Varadhachary G Wistuba I Maitra A Alvarez H Minimally invasive genomic and transcriptomic profiling of visceral cancers by next-generation sequencing of circulating exosomes Ann Oncol 2015 10.1093/annonc/mdv604 21 Thakur BK Zhang H Becker A Matei I Huang Y Costa-Silva B Zheng Y Hoshino A Brazier H Xiang J Williams C Rodriguez-Barrueco R Silva JM Zhang W Hearn S Elemento O Paknejad N Manova-Todorova K Welte K Bromberg J Peinado H Lyden D Double-stranded DNA in exosomes: a novel biomarker in cancer detection Cell Res 2014 24 766 769 24710597 22 Camussi G Deregibus MC Cantaluppi V Role of stem-cell-derived microvesicles in the paracrine action of stem cells Biochem Soc Trans 2013 41 283 287 23356298 23 Camussi G Deregibus MC Tetta C Paracrine/endocrine mechanism of stem cells on kidney repair: role of microvesicle-mediated transfer of genetic information Curr Opin Nephrol Hypertens 2010 19 7 12 19823086 24 Rager TM Olson JK Zhou Y Wang Y Besner GE Exosomes secreted from bone marrow-derived mesenchymal stem cells protect the intestines from experimental necrotizing enterocolitis J Pediatr Surg 2016 51 10.1016/j.jpedsurg.2016.02.061 25 Xin H Li Y Cui Y Yang JJ Zhang ZG Chopp M Systemic administration of exosomes released from mesenchymal stromal cells promote functional recovery and neurovascular plasticity after stroke in rats J Cereb Blood Flow Metab 2013 33 1711 1715 23963371 26 Lou G Song X Yang F Wu S Wang J Chen Z Liu Y Exosomes derived from miR-122-modified adipose tissue-derived MSCs increase chemosensitivity of hepatocellular carcinoma J Hematol Oncol 2015 8 122 26514126 27 Xin H Li Y Liu Z Wang X Shang X Cui Y Zhang ZG Chopp M MiR-133b Promotes Neural Plasticity and Functional Recovery after Treatment of Stroke with Multipotent Mesenchymal Stromal Cells in Rats Via Transfer of Exosome-Enriched Extracellular Particles Stem Cells 2013 31 2737 2746 23630198 28 Ban JJ Lee M Im W Kim M Low pH increases the yield of exosome isolation Biochem Biophys Res Commun 2015 461 76 79 25849885 29 Parolini I Federici C Raggi C Lugini L Palleschi S De Milito A Coscia C Lessi E Logozzi M Molinari A Colone M Tatti M Sargiacomo M Fais S Microenvironmental pH is a key factor for exosome traffic in tumor cells J Biol Chem 2009 284 34211 34222 19801663 30 Kosaka N Izumi H Sekine K Ochiya T microRNA as a new immune-regulatory agent in breast milk Silence 2010 1 1 8 20226068 31 Blanc L Liu J Vidal M Chasis JA An X Mohandas N The water channel aquaporin-1 partitions into exosomes during reticulocyte maturation: implication for the regulation of cell volume Blood 2009 114 3928 3934 19724054 32 Hamzah M Braun R Eldridge M Exosomes secreted by mesenchymal stem cells prevent hyperoxia-induced lung injury in newborn rat Critical Care Med 2014 42 Supplement 1 A1531 33 King HW Michael MZ Gleadl JM Hypoxic enhancement of exosome release by breast cancer cells BMC Cancer 2012 12 421 22998595 34 Admyre C Telemo E Almqvist N Lotvall J Lahesmaa R Scheynius A Gabrielsson S Exosomes - nanovesicles with possible roles in allergic inflammation Allergy 2008 63 404 408 18315728 35 Camussi G Deregibus MC Bruno S Grange C Fonsato V Tetta C Exosome/microvesicle-mediated epigenetic reprogramming of cells Am J Cancer Res 2011 1 98 110 21969178 36 Ridder K Keller S Dams M Rupp AK Schlaudraff J Turco DD Starmann J Macas J Karpova D Devraj K Depboylu C Landfried B Arnold B Plate KH Höglinger G Sültmann H Altevogt P Momma S Extracellular vesicle-mediated transfer of genetic information between the hematopoietic system and the brain in response to inflammation PLoS Biol 2014 12 e1001874 24893313 37 Stoorvogel W Functional transfer of microRNA by exosomes Blood 2012 119 646 648 22262739 38 Valadi H Ekstrom K Bossios A Sjostrand M Lee JJ Lotvall JO Exosome-mediated transfer of mRNAs and microRNAs is a novel mechanism of genetic exchange between cells Nat Cell Biol 2007 9 654 672 17486113 39 van Balkom BW de Jong OG Smits M Brummelman J den Ouden K de Bree PM Eijndhoven MA Pegtel DM Stoorvogel W Würdinger T Verhaar MC Endothelial cells require miR-214 to secrete exosomes that suppress senescence and induce angiogenesis in human and mouse endothelial cells Blood 2013 121 3997 4006 23532734 40 Wang K Zhang S Weber J Baxter D Galas DJ Export of microRNAs and microRNA-protective protein by mammalian cells Nucleic Acids Res 2010 38 7248 7259 20615901 41 Xin H Li Y Buller B Katakowski M Zhang Y Wang X Shang X Zhang ZG Chopp M Exosome-mediated transfer of miR-133b from multipotent mesenchymal stromal cells to neural cells contributes to neurite outgrowth Stem Cells 2012 30 1556 1564 22605481 42 Chairoungdua A Smith DL Pochard P Hull M Caplan MJ Exosome release of β-catenin: a novel mechanism that antagonizes signaling J Cell Biol 2010 190 1079 1091 20837771 43 Gross JC Chaudhary V Bartscherer K Boutros M Active Wnt proteins are secreted on exosomes Nat Cell Biol 2012 14 1036 1045 22983114 44 Koles K Nunnari J Korkut C Barria R Brewer C Li Y Leszyk J Zhang B Budnik V Mechanism of evenness interrupted (evi)-exosome release at synaptic moutons J Biol Chem 2012 287 16820 16834 22437826 45 Korkut C Ataman B Ramachandran P Ashley J Barria R Gherbesi N Budnik V Trans-synaptic transmission of vesicular Wnt signals through Evi/Wntless Cell 2009 139 393 404 19837038 46 Agarwal A Fanelli G Letizia M Tung SL Boardman D Lechler R Lombardi G Smyth LA Regulatory T cell-derived exosomes: Possible therapeutic and diagnostic tools in transplantation Front Immunol 2014 5 555 25414702 47 Bobrie A Colombo M Raposo G Théry C Exosome secretion: molecular mechanisms and roles in immune responses Traffic 2011 12 1659 1668 21645191 48 Fernández-Messina L Gutiérrez-Vázquez C Rivas-García E Sánchez-Madrid F de la Fuente H Immunomodulatory role of microRNAs transferred by extracellular vesicles Biol Cell 2015 107 61 77 25564937 49 Mittelbrunn M Gutiérrez-Vázquez C Villarroya-Beltri C González S Sánchez-Cabo F González MÁ Bernad A Sánchez-Madrid F Unidirectional transfer of microRNA-loaded exosomes from T cells to antigen-presenting cells Nat Commun 2011 2 282 21505438 50 Montecalvo A Larregina AT Shufesky WJ Stolz DB Sullivan ML Karlsson JM Baty CJ Gibson GA Erdos G Wang Z Milosevic J Tkacheva OA Divito SJ Jordan R Lyons-Weiler J Watkins SC Morelli AE Mechanism of transfer of functional microRNAs between mouse dendritic cells via exosomes Blood 2012 119 756 766 22031862 51 Martin RK Badrooks KB Henningsson F Heyman B Conrad DH Antigen transfer from exosomes to dendritic cells as an explanation for the immune enhancement seen by IgE immune complexes PLoS One 2014 9 e110609 25330118 52 Fujita Y Yoshioka Y Ito S Araya J Kuwano K Ochiya T Intercellular communication by extracellular vesicles and their microRNAs in asthma Clin Ther 2014 36 873 881 24909737 53 Poliakov A Spilman M Dokland T Amling CL Mobley JA Structural Heterogeneity and Protein Composition of Exosome-Like Vesicles (Prostasomes) in Human Semen Prostate 2009 69 159 167 18819103 54 Jenjaroenpun P Kremenska Y Nair VM Kremenskoy M Joseph B Kurochkin IV Characterization of RNA in exosomes secreted by human breast cancer cell lines using next-generation sequencing Peer J 2013 1 e201 24255815 55 Nolte-’t Hoen ENM Buermans HPJ Waasdorp M Stoorvogel W Wauben MHM ’tHoen PAC Deep sequencing of RNA from immune cell-derived vesicles uncovers the selective incorporation of small non-coding RNA biotypes with potential regulatory functions Nucleic Acids Res 2012 40 9272 9285 22821563 56 Zhou H Cheruvanky A Hu X Matsumoto T Hiramatsu N Cho ME Berger A Leelahavanichkul A Doi K Chawla LS Illei GG Kopp JB Balow JE Austin HA III Yuen PST Star RA Urinary exosomal transcription factors, a new class of biomarkers for renal disease Kidney Int 2008 74 613 621 18509321 57 Ung TH Madsen HJ Hellwinkel JE Lencioni AM Graner MW Exosome proteomics reveals transcriptional regulator proteins with potential to mediate downstream pathways Cancer Sci 2014 105 1384 1392 25220623 58 Okoye IS Coomes SM Pelly VS Czieso S Papayannopoulos V Tolmachova T Seabra MC Wilson MS MicroRNA-containing T-regulatory-cell-derived exosomes suppress pathogenic T helper 1 cells Immunity 2014 41 89 103 25035954 59 Skogberg G Lundberg V Berglund M Gudmundsdottir J Telemo E Lindgren S Ekwall O Human thymic epithelial primary cells produce exosomes carrying tissue-restricted antigens Immunol Cell Biol 2015 93 727 734 25776846 60 Wang GJ Liu Y Qin A Shah SV Deng Z-b Xiang X Cheng Z Liu C Wang J Zhang L Grizzle WE Zhang HG Thymus Exosomes-Like Particles Induce Regulatory T Cells J Immunol 2008 181 5242 5248 18832678 61 Bretz NP Ridinger J Rupp A-K Rimbach K Keller S Rupp C Marme F Umansky L Umansky V Eigenbrod T Sammar M Altevogt P Body fluid exosomes promote secretion of inflammatory cytokines in monocytic cells via TLR signaling J Biol Chem 2013 288 36691 36702 24225954 62 Johnson K Chu J Huang M PowelL M Villinger F Bond V Innate cytokines associate with exosomes in plasma of HIV-1+ individuals (P6205) J Immunol 2013 190 Meeting Abstract Supplement 118.26 63 Konadu KA Chu J Huang MB Amancha PK Armstrong W Powell MD Villinger F Bond VC Association of Cytokines With Exosomes in the Plasma of HIV-1-Seropositive Individuals J Infect Dis 2015 211 1712 1716 25512626 64 Sampey GC Saifuddin M Schwab A Barclay R Punya S Chung M-C Hakami RM Asad Zadeh M Lepene B Klase ZA El-Hage N Young M Iordanskiy S Kashanchi F Exosomes from HIV-1-infected Cells Stimulate Production of Pro-inflammatory Cytokines through Trans-activating Response (TAR) RNA J Biol Chem 2016 291 1251 1266 26553869 65 Cossetti C Iraci N Mercer TR Leonardi T Alpi E Drago D Alfaro-Cervello C Saini HK Davis MP Schaeffer J Vega B Stefanini M Zhao CJ Muller W Garcia-Verdugo JM Mathivanan S Bachi A Enright AJ Mattick JS Pluchino S Extracellular Vesicles from Neural Stem Cells Transfer IFN-γ via Ifngr1 to Activate Stat1 Signaling in Target Cells Mol Cell 2014 56 193 204 25242146 66 Shefler I Pasmanik-Chor M Kidron D Mekori YA Hershko AY T cell derived microvesicles induce mast cell production of IL-24: Relevance to inflammatory skin diseases J Allergy Clin Immunol 2014 133 217 224 e3 23768573 67 Shefler I Salamon P Reshef T Mor A Mekori YA T Cell-Induced Mast Cell Activation: A Role for Microparticles Released from Activated T Cells J Immunol 2010 85 4206 4212 68 Thery C Duban L Segura E Veron P Lantz O Amigorena S Indirect activation of naive CD4+ T cells by dendritic cell-derived exosomes Nat Immunol 2002 3 1156 1162 12426563 69 Admyre C Grunewald J Thyberg J Gripenback S Tornling G Eklund A Scheynius A Gabrielsson S Exosomes with major histocompatibility complex class II and co-stimulatory molecules are present in human BAL fluid Eur Respir J 2003 22 578 583 14582906 70 Skokos D Botros HG Demeure C Morin J Peronet R Birkenmeier G Boudaly S Mécheri S Mast Cell-Derived Exosomes Induce Phenotypic and Functional Maturation of Dendritic Cells and Elicit Specific Immune Responses In Vivo J Immunol 2003 170 3037 3045 12626558 71 Skokos D Le Panse S Villa I Rousselle JC Peronet R David B Namane A Mécheri S Mast Cell-Dependent B and T Lymphocyte Activation Is Mediated by the Secretion of Immunologically Active Exosomes J Immunol 2001 166 868 876 11145662 72 Skokos D Le Panse S Villa I Rousselle JC Peronet R Namane A David B Mécheri S Nonspecific B and T cell-stimulatory activity mediated by mast cells is associated with exosomes Int Arch Allergy Immunol 2001 124 133 136 11306949 73 Al-Nedawi K Szemraj J Cierniewski CS Mast cell-derived exosomes activate endothelial cells to secrete plasminogen activator inhibitor type 1 Arterioscler Thromb Vasc Biol 2005 25 1744 1749 15920032 74 Ekström K Valadi H Sjöstrand M Malmhäll C Bossios A Eldh M Lötvall J Characterization of mRNA and microRNA in human mast cell-derived exosomes and their transfer to other mast cells and blood CD34 progenitor cells J Extracell Vesicles 2012 16 1 75 Wahlgren J de Karlson TL Glader P Telemo E Valadi H Activated Human T Cells Secrete Exosomes That Participate in IL-2 Mediated Immune Response Signaling PLoS One 2012 7 e49723 23166755 76 Kulshreshtha A Ahmad T Agrawal A Ghosh B Proinflammatory role of epithelial cell-derived exosomes in allergic airway inflammation J Allergy Clin Immunol 2012 131 1194 1203 e14 77 Xia YC Harris T Stewart AG Mackay GA Secreted factors from human mast cells trigger inflammatory cytokine production by human airway smooth muscle cells Int Arch Allergy Immunol 2013 160 75 85 22948287 78 Mazzeo C Canas JA Zafra MP Rojas Marco A Fernandez-Nieto M Sanz V Mittelbrunn M Izquierdo M Baixaulli F Sastre J Del Pozo V Exosome secretion by eosinophils: A possible role in asthma pathogenesis J Allergy Clin Immunol 2015 135 1603 1613 25617225 79 Clayton A Court J Navabi H Adams M Mason MD Hobot JA Newman GR Jasani B Analysis of antigen presenting cell derived exosomes, based on immuno-magnetic isolation and flow cytometry J Immunol Methods 2001 247 163 174 11150547 80 Raposo G Nijman HW Stoorvogel W Liejendekker R Harding CV Melief CJ Geuze HJ B lymphocytes secrete antigen-presenting vesicles J Exp Med 1996 183 1161 1172 8642258 81 Sobo-Vujanovic A Munich S Vujanovic NL Dendritic-cell exosomes cross-present Toll-like receptor-ligands and activate bystander dendritic cells Cell Immunol 2014 289 119 127 24759079 82 Qazi KR Gehrmann U Domange Jordo E Karlsson MC Gabrielsson S Antigen-loaded exosomes alone induce Th1-type memory through a B-cell-dependent mechanism Blood 2009 113 2673 2683 19176319 83 Vincent-Schneider H Stumptner-Cuvelette P Lankar D Pain S Raposo G Benaroch P Bonnerot C Exosomes bearing HLA DR1 molecules need dendritic cells to efficiently stimulate specific T cells Int Immunol 2002 14 713 722 12096030 84 Carroll-Portillo A Surviladze Z Cambi A Lidke DS Wilson BS Mast cell synapses and exosomes: membrane contacts for information exchange Front Immunol 2012 3 46 22566928 85 Gaudenzio N Espagnolle N Mars LT Liblau R Valitutti S Espinosa E Cell-cell cooperation at the T helper cell/mast cell immunological synapse Blood 2009 114 4979 4988 19805617 86 Ramachandran S Palanisamy V Horizontal transfer of RNAs: exosomes as mediators of intercellular communication Wiley Interdiscip Rev RNA 2012 3 286 293 22012863 87 Gong J Yang NS Croft M Weng IC Sun L Liu FT Chen SS The antigen presentation function of bone marrow-derived mast cells is spatiotemporally restricted to a subset expressing high levels of cell surface FcεRI and MHC II BMC Immunol 2010 11 34 20591187 88 Beauvillain C Juste MO Dion S Pierre J Dimier-Poisson I Exosomes are an effective vaccine against congenital toxoplasmosis in mice Vaccine 2009 27 1750 1757 19186199 89 Schorey JS Cheng Y Singh PP Smith VL Exosomes and other extracellular vesicles in host-pathogen interactions EMBO Rep 2015 16 24 43 25488940 90 Pitt JM Charrier M Viaud S André F Besse B Chaput N Zitvogel L Dendritic Cell-Derived Exosomes as Immunotherapies in the Fight against Cancer J Immunol 2014 193 1006 1011 25049431 91 Vermeij R Leffers N van der Burg SH Melief CJ Daemen T Nijman HW Immunological and Clinical Effects of Vaccines Targeting p53-Overexpressing Malignancies J Biomed Biotechnol 2011 2011 702146 21541192 92 Viaud S Théry C Ploix S Tursz T Lapierre V Lantz O Zitvogel L Chaput N Dendritic cell-derived exosomes for cancer immunotherapy: what’s next? Cancer Res 2010 70 1281 1285 20145139 93 Admyre C Bohle B Johansson SM Focke-Tejkl M Valenta R Scheynius A Gabrielsson S B cell-derived exosomes can present allergen peptides and activate allergen-specific T cells to proliferate and produce TH2-like cytokines J Allergy Clin Immunol 2007 120 1418 1424 17868797 94 Suojalehto H Lindström I Majuri ML Mitts C Karjalainen J Wolff H Alenius H Altered microRNA expression of nasal mucosa in long-term asthma and allergic rhinitis Int Arch Allergy Immunol 2014 163 168 178 24513959 95 Wu G Yang G Zhang R Xu G Zhang L Wen W Lu J Liu J Yu Y Altered microRNA Expression Profiles of Extracellular Vesicles in Nasal Mucus From Patients With Allergic Rhinitis Allergy Asthma Immunol Res 2015 7 449 457 26122505 96 Shin SH Park JY Jeon CH Choi JK Lee SH Quantitative analysis of eotaxin and RANTES messenger RNA in nasal polyps: association of tissue and nasal eosinophils Laryngoscope 2000 110 1353 1357 10942140 97 Qiu S Du Y Duan X Geng X Xie J Gao H Yang PC Cytotoxic T lymphocytes mediate chronic inflammation of the nasal mucosa of patients with atypical allergic rhinitis N Am J Med Sci 2011 3 378 383 22171246 98 Lässer C O’Neil SE Shelke GV Sihlbom C Hansson SF Gho YS Lundbäck B Lötvall J Exosomes in the nose induce immune cell trafficking and harbour an altered protein cargo in chronic airway inflammation J Transl Med 2016 14 181 27320496 99 Ptak W Geba GP Askenase PW Initiation of delayed-type hypersensitivity by low doses of monoclonal IgE antibody. Mediation by serotonin and inhibition by histamine J Immunol 1991 146 3929 3936 2033262 100 Zhang L Hou D Chen X Li D Zhu L Zhang Y Li J Bian Z Liang X Cai X Yin Y Wang C Zhang T Zhu D Zhang D Xu J Chen Q Ba Y Liu J Wang Q Chen J Wang J Wang M Zhang Q Zhang J Zen K Zhang CY Exogenous plant MIR168a specifically targets mammalian LDLRAP1: evidence of cross-kingdom regulation by microRNA Cell Res 2012 22 107 126 21931358 101 Prado N de Alché JD Casado-Vela J Mas S Villalba M Rodríguez R Batanero E Nanovesicles Are Secreted during Pollen Germination and Pollen Tube Growth: A Possible Role in Fertilization Mol Plant 2014 7 573 577 24177685 102 Wei T Hibino H Omura T Release of Rice dwarf virus from insect vector cells involves secretory exosomes derived from multivesicular bodies Commun Integr Biol 2009 2 324 326 19721879 103 Vallhov H Gutzeit C Hultenby K Valenta R Gronlund H Scheynius A Dendritic cell-derived exosomes carry the major cat allergen Fel d 1 and induce an allergic immune response Allergy 2015 70 1651 1655 26198793 104 Luo X Han M Liu J Wang Y Luo X Zheng J Wang S Liu Z Liu D Yang PC Li H Epithelial cell-derived micro RNA-146a generates interleukin-10-producing monocytes to inhibit nasal allergy Sci Rep 2015 5 15937 26526003 105 Platts-Mills TA Vaughan JW Blumenthal K Woodfolk JA Sporik RB Decreased prevalence of asthma among children with high exposure to cat allergen: relevance of the modified Th2 response Mediators Inflamm 2001 10 288 291 11817662 105a Akdis CA Akdis M Mechanisms of allergen-specific immunotherapy and immune tolerance to allergens World Allergy Organ J 2015 8 1 17 10.1186/s40413-015-0063-2 26023323 106 Gehrmann U Hiltbrunner S Georgoudaki AM Karlsson MC Näslund TI Gabrielsson S Synergistic induction of adaptive antitumor immunity by co-delivery of antigen with α-galactosylceramide on exosomes Cancer Res 2013 73 3865 3876 23658368 107 Campos RA Szczepanik M Itakura A Akahira-Azuma M Sidobre S Kronenberg M Askenase PW Cutaneous immunization rapidly activates liver invariant Vα14 NKT cells stimulating B-1 B cells to initiate T cell recruitment for elicitation of contact sensitivity J Exp Med 2003 198 1785 1796 14676294 108 Yamamoto N Kerfoot SM Hutchinson AT Dela Cruz CS Nakazawa N Szczepanik M Majewska-Szczepanik M Nazimek K Ohana N Bryniarski K Mori T Muramatsu M Kanemitsu K Askenase PW Expression of activation-induced cytidine deaminase enhances the clearance of pneumococcal pneumonia: evidence of a subpopulation of protective anti-pneumococcal B1a cells Immunology 2016 147 97113 109 Prado N De Linares C Sanz ML Gamboa P Villalba M Rodríguez R Batanero E Pollensomes as Natural Vehicles for Pollen Allergens J Immunol 2015 195 445 449 26041541 110 Gol J Van Niel G Lebreton C Lepelletier Y Candalh C Dugave C Heath JK Raposo G Cerf-Bensussan N Heyman M T84-intestinal epithelial exosomes bear MHC class II/peptide complexes potentiating antigen presentation by dendritic cells Gastroenterology 2007 132 1866 1876 17484880 111 Sánchez-Gómez FJ González-Morena JM Vida Y Pérez-Inestrosa E Blanca M Torres MJ Pérez-Sala D Amoxicillin haptenates intracellular proteins that can be transported in exosomes to target cells Allergy 2016 in press 10.1111/all.12958 112 Choi EB Hong SW Kim DK Jeon SG Kim KR Cho SH Gho YS Jee YK Kim YK Decreased diversity of nasal microbiota and their secreted extracellular vesicles in patients with chronic rhinosinusitis based on a metagenomic analysis Allergy 2014 69 517 526 24611950 113 Buck AH Coakley G Simbari F McSorley HJ Quintana JF Le Bihan T Kumar S Abreu-Goodger C Lear M Harcus Y Ceroni A Babayan SA Blaxter M Ivens A Maizels RM Exosomes secreted by nematode parasites transfer small RNAs to mammalian cells and modulate innate immunity Nat Commun 2014 5 5488 25421927 114 Marcilla A Trelis M Cortés A Sotillo J Cantalapiedra F Minguez MT Valero ML Sánchez del Pino MM Muñoz-Antoli C Toledo R Bernal D Extracellular Vesicles from Parasitic Helminths Contain Specific Excretory/Secretory Proteins and Are Internalized in Intestinal Host Cells PLoS One 2012 7 e45974 23029346 115 Siles-Lucas M Morchon R Simon F Manzano-Roman R Exosome-transported microRNAs of helminth origin: new tools for allergic and autoimmune diseases therapy? Parasite Immunol 2015 37 208 214 25712154 116 Atayde VD Aslan H Townsend S Hassani K Kamhawi S Olivier M Exosome secretion by the parasitic protozoan Leishmania within the sand fly midgut Cell Rep 2015 13 957 967 26565909 117 Silverman JM Clos J Horakova E Wang AY Wiesgigl M Kelly I Lynn MA McMaster WR Foster LJ Levings MK Reiner NE Leishmania exosomes modulate innate and adaptive immune responses through effects on monocytes and dendritic cells J Immunol 2010 185 5011 5022 20881185 118 Trocoli Torrecilhas AC Tonelli RR Pavanelli WR da Silva JS Schumacher RI de Souza W Silva ENC de Almeida Abrahamsohn I Colli W Manso Alves MJ Trypanosoma cruzi: parasite shed vesicles increase heart parasitism and generate an intense inflammatory response Microbes Infect 2009 11 29 39 19028594 119 Twu O de Miguel N Lustig G Stevens GC Vashisht AA Wohlschlegel JA Johnson PJ Trichomonas vaginalis Exosomes Deliver Cargo to Host Cells and Mediate Host:Parasite Interactions PLoS Pathog 2013 9 e1003482 23853596 120 Costa-Silva B Aiello NM Ocean AJ Singh S Zhang H Thakur BK Becker A Hoshino A Mark MT Molina H Xiang J Zhang T Theilen TM García-Santos G Williams C Ararso Y Huang Y Rodrigues G Shen TL Labori KJ Lothe IM Kure EH Hernandez J Doussot A Ebbesen SH Grandgenett PM Hollingsworth MA Jain M Mallya K Batra SK Jarnagin WR Schwartz RE Matei I Peinado H Stanger BZ Bromberg J Lyden D Pancreatic exosomes initiate pre-metastatic niche formation in the liver Nat Cell Biol 2015 17 816 826 25985394 121 Peinado H Alečković M Lavotshkin S Matei I Costa-Silva B Moreno-Bueno G Hergueta-Redondo M Williams C García-Santos G Ghajar C Nitadori-Hoshino A Hoffman C Badal K Garcia BA Callahan MK Yuan J Martins VR Skog J Kaplan RN Brady MS Wolchok JD Chapman PB Kang Y Bromberg J Lyden D Melanoma exosomes educate bone marrow progenitor cells toward a pro-metastatic phenotype through MET Nat Med 2012 18 883 891 22635005 122 Hoshino A Costa-Silva B Shen TL Rodrigues G Hashimoto A Tesic Mark M Molina H Kohsaka S Di Giannatale A Ceder S Singh S Williams C Soplop N Uryu K Pharmer L King T Bojmar L Davies AE Ararso Y Zhang T Zhang H Hernandez J Weiss JM Dumont-Cole VD Kramer K Wexler LH Narendran A Schwartz GK Healey JH Sandstrom P Labori KJ Kure EH Grandgenett PM Hollingsworth MA de Sousa M Kaur S Jain M Mallya K Batra SK Jarnagin WR Brady MS Fodstad O Muller V Pantel K Minn AJ Bissell MJ Garcia BA Kang Y Rajasekhar VK Ghajar CM Matei I Peinado H Bromberg J Lyden D Tumour exosome integrins determine organotropic metastasis Nature 2015 527 329 335 26524530 123 Chung KF Cytokines as targets in chronic obstructive pulmonary disease Curr Drug Targets 2006 7 675 681 16787167 124 Park JA Sharif AS Tschumperlin DJ Lau L Limbrey R Howarth P Drazen JM Tissue factor-bearing exosome secretion from human mechanically stimulated bronchial epithelial cells in vitro and in vivo J Allergy Clin Immunol 2012 130 1375 1383 22828416 125 Vargas A Lavoie J-P Neutrophil-Derived Exosomes: A New Mechanism Possibly Contributing To Airway Smooth Muscle Remodeling Am J Respir Crit Care Med 2015 191 A6475 126 Torregrosa Paredes P Esser J Admyre C Nord M Rahman QK Lukic A Radmark O Gronneberg R Grunewald J Eklund A Scheynius A Gabrielsson S Bronchoalveolar lavage fluid exosomes contribute to cytokine and leukotriene production in allergic asthma Allergy 2012 67 911 919 22620679 127 Kosaka N Iguchi H Hagiwara K Yoshioka Y Takeshita F Ochiya T Neutral Sphingomyelinase 2 (nSMase2)-dependent Exosomal Transfer of Angiogenic MicroRNAs Regulate Cancer Cell Metastasis J Biol Chem 2013 288 10849 10859 23439645 128 Singh R Pochampally R Watabe K Lu Z Mo YY Exosome-mediated transfer of miR-10b promotes cell invasion in breast cancer Mol Cancer 2014 13 256 25428807 129 Yuyama K Sun H Sakai S Mitsutake S Okada M Tahara H Furukawa J Fujitani N Shinohara Y Igarashi Y Decreased Amyloid-β Pathologies by Intracerebral Loading of Glycosphingolipid-enriched Exosomes in Alzheimer Model Mice J Biol Chem 2014 289 24488 24498 25037226 130 Yang W Schmid E Nurbaeva MK Szteyn K Leibrock C Yan J Schaller M Gulbins E Shumilina E Lang F Role of acid sphingomyelinase in the regulation of mast cell function Clin Exp Allergy 2014 44 79 90 24164338 131 Mattes J Collison A Plank M Phipps S Foster PS Antagonism of microRNA-126 suppresses the effector function of TH2 cells and the development of allergic airways disease Proc Natl Acad Sci USA 2009 4 18704 18709 132 Plank MW Maltby S Tay HL Stewart J Eyers F Hansbro PM Foster PS MicroRNA Expression Is Altered in an Ovalbumin-Induced Asthma Model and Targeting miR-155 with Antagomirs Reveals Cellular Specificity PLoS One 2015 10 e0144810 26693910 133 Canas JA Mazzeo C Guerra A Barranco P Quirce S Ruiz Hornillos J Izquierdo M Sastre J del Pozo V Exosomes from eosinophils of asthmatic patients produce functional alterations on structural lung cell EAACI Online Library 2015 98145 134 Del Pozo V Mazzeo C Rodriguez Marco A Fernandez-Nieto MDM Paz Zafra M Sanz V Sastre J Exosomes Secretion By Eosinophils: A Possible Role In Asthma Pathogenesis J Allergy Clin Immunol 2014 133 Supplement AB58 135 Mahmudi-Azer S Downey GP Moqbel R Translocation of the tetraspanin CD63 in association with human eosinophil mediator release Blood 2002 99 4039 4047 12010805 136 Kumar M Ahmad T Sharma A Mabalirajan U Kulshreshtha A Agrawal A Ghosh B Let-7 microRNA-mediated regulation of IL-13 and allergic airway inflammation J Allergy Clin Immunol 2011 128 1077 1085 e10 21616524 137 Esser J Gehrmann U D’Alexandri FL Hidalgo-Estevez AM Wheelock CE Scheynius A Gabrielsson S Radmark O Exosomes from human macrophages and dendritic cells contain enzymes for leukotriene biosynthesis and promote granulocyte migration J Allergy Clin Immunol 2010 126 1032 1040 20728205 138 Majumdar R Tavakoli Tameh A Parent CA Exosomes Mediate LTB4 Release during Neutrophil Chemotaxis Plos Biol 2016 14 e1002336 26741884 139 Sadik CD Luster AD Lipid-cytokine-chemokine cascades orchestrate leukocyte recruitment in inflammation J Leukoc Biol 2012 91 207 215 22058421 140 Reutershan J Basit A Galkina EV Ley K Sequential recruitment of neutrophils into lung and bronchoalveolar lavage fluid in LPS-induced acute lung injury Am J Physiol Lung Cell Mol Physiol 2005 289 L807 815 15951336 141 Peters M Kauth M Schwarze J Körner-Rettberg C Riedler J Nowak D Braun-Fahrländer C von Mutius E Bufe A Holst O Inhalation of stable dust extract prevents allergen induced airway inflammation and hyperresponsiveness Thorax 2006 61 134 139 16244088 142 Schuijs MJ Willart MA Vergote K Gras D Deswarte K Ege MJ Madeira FB Beyaert R van Loo G Bracher F von Mutius E Chanez P Lambrecht BN Hammad H Farm dust and endotoxin protect against allergy through A20 induction in lung epithelial cells Science 2015 349 1106 1110 26339029 143 Kim YS Choi EJ Lee WH Choi SJ Roh TY Park J Jee YK Zhu Z Koh YY Gho YS Kim YK Extracellular vesicles, especially derived from Gram-negative bacteria, in indoor dust induce neutrophilic pulmonary inflammation associated with both Th1 and Th17 cell responses Clin Exp Allergy 2013 43 443 454 23517040 144 Choy DF Hart KM Borthwick LA Shikotra A Nagarkar DR Siddiqui S Jia G Ohri CM Doran E Vannella KM Butler CA Hargadon B Sciurba JC Gieseck RL Thompson RW White S Abbas AR Jackman J Wu LC Egen JG Heaney LG Ramalingam TR Arron JR Wynn TA Bradding P TH2 and TH17 inflammatory pathways are reciprocally regulated in asthma Sci Transl Med 2015 7 301ra129 145 Peng X Hassoun PM Sammani S McVerry BJ Burne MJ Rabb H Pearse D Tuder RM Garcia JGN Protective Effects of Sphingosine 1-Phosphate in Murine Endotoxin-induced Inflammatory Lung Injury Am J Respir Crit Care Med 2004 169 1245 1251 15020292 146 Kajimoto T Okada T Miya S Zhang L Nakamura S Ongoing activation of sphingosine 1-phosphate receptors mediates maturation of exosomal multivesicular endosomes Nat Commun 2013 4 2712 24231649 147 Dinkins MB Dasgupta S Wang G Zhu G Bieberich E Exosome reduction in vivo is associated with lower amyloid plaque load in the 5XFAD mouse model of Alzheimer’s disease Neurobiol Aging 2014 35 1792 1800 24650793 148 Kim MR Hong SW Choi EB Lee WH Kim YS Jeon SG Jang MH Gho YS Kim YK Staphylococcus aureus-derived extracellular vesicles induce neutrophilic pulmonary inflammation via both Th1 and Th17 cell responses Allergy 2012 67 1271 1281 22913540 149 Rebane A Akdis CA MicroRNAs in allergy and asthma Curr Allergy Asthma Rep 2014 14 424 24504527 150 Levanen B Bhakta NR Torregrosa Paredes P Barbeau R Hiltbrunner S Pollack JL Skold CM Svartengren M Grunewald J Gabrielsson S Eklund A Larsson BM Woodruff PG Erle DJ Wheelock AM Altered microRNA profiles in bronchoalveolar lavage fluid exosomes in asthmatic patients J Allergy Clin Immunol 2013 131 894 903 23333113 151 Solberg OD Ostrin EJ Love MI Peng JC Bhakta NR Hou L Nguyen C Solon M Nguyen C Barczak AJ Zlock LT Blagev DP Finkbeiner WE Ansel KM Arron JR Erle DJ Woodruff PG Airway epithelial miRNA expression is altered in asthma Am J Respir Crit Care Med 2012 186 965 974 22955319 152 Maruoka S Gon Y Shikano S Shintani Y Koyama D Sekiyama T Hiranuma H Inoue T Takeshita I Tsuboi E Soda K Hashimoto S Exosomal MicroRNAs In The Serum Are Potential Real-Time Biomarkers For Allergic Inflammation In The Airway Of Mice ATS 2014 Meeting Abstracts A4211 153 Pinkerton M Chinchilli V Banta E Craig T August A Bascom R Cantorna M Harvill E Ishmael FT Differential expression of microRNAs in exhaled breath condensates of patients with asthma, patients with chronic obstructive pulmonary disease, and healthy adults J Allergy Clin Immunol 2013 132 217 219 23628339 154 Sinha A Yadav AK Chakraborty S Kabra SK Lodha R Kumar M Kulshreshtha A Sethi T Pandey R Malik G Laddha S Mukhopadhyay A Dash D Ghosh B Agrawal A Exosome-enclosed microRNAs in exhaled breath hold potential for biomarker discovery in patients with pulmonary diseases J Allergy Clin Immunol 2013 132 219 222 23683467 155 Arroyo JD Chevillet JR Kroh EM Ruf IK Pritchard CC Gibson DF Mitchell PS Bennett CF Pogosova-Agadjanyan EL Stirewalt DL Tait JF Tewari M Argonaute2 complexes carry a population of circulating microRNAs independent of vesicles in human plasma Proc Natl Acad Sci USA 2011 108 5003 5008 21383194 156 Kurchinovich A Burwinkel B Distinct AGO1 and AGO2 associated miRNA profiles in human cells and blood plasma RNA Biol 2012 9 1066 1075 22858679 157 Li L Zhu D Huang L Zhang J Bian Z Chen X Liu Y Zhang CY Zen K Argonaute 2 complexes selectively protect the circulating microRNAs in cell-secreted microvesicles PLoS One 2012 7 e46957 23077538 158 Turchinovich A Weiz L Langheinz A Burwinkel B Characterization of extracellular circulating microRNA Nucleic Acids Res 2011 39 7223 7233 21609964 159 Prado N Marazuela EG Segura E Fernandez-Garcia H Villalba M Thery C Rodriguez R Batanero E Exosomes from bronchoalveolar fluid of tolerized mice prevent allergic reaction J Immunol 2008 181 1519 1525 18606707 160 Karlsson M Lundin S Dahlgren U Kahu H Pettersson I Telemo E “Tolerosomes” are produced by intestinal epithelial cells Eur J Immunol 2001 31 2892 2900 11592064 161 Lonnqvist A Ostman S Almqvist N Hultkrantz S Telemo E Wold AE Rask C Neonatal exposure to staphylococcal superantigen improves induction of oral tolerance in a mouse model of airway allergy Eur J Immunol 2009 39 447 456 19130476 162 Van Niel G Mallegol J Bevilacqua C Candalh C Brugière S Tomaskovic-Crook E Heath JK Cerf-Bensussan N Heyman M Intestinal epithelial exosomes carry MHC class II/peptides able to inform the immune system in mice Gut 2003 52 1690 1697 14633944 163 Admyre C Johansson SM Qazi KR Filén JJ Lahesmaa R Norman M Neve EP Scheynius A Gabrielsson S Exosomes with Immune Modulatory Features Are Present in Human Breast Milk J Immunol 2007 179 1969 1978 17641064 164 Zhou Q Li M Wang X Li Q Wang T Zhu Q Zhou X Wang X Gao X Li X Immune-related MicroRNAs are Abundant in Breast Milk Exosomes Int J Biol Sci 2012 8 118 123 22211110 165 Izumi H Kosaka N Shimizu T Sekine K Ochiya T Takase M Bovine milk contains microRNA and messenger RNA that are stable under degradative conditions J Dairy Sci 2012 95 4831 4841 22916887 166 Gu Y Li M Wang T Liang Y Zhong Z Wang X Zhou Q Chen L Lang Q He Z Chen X Gong J Gao X Li X Lv X Lactation-Related MicroRNA Expression Profiles of Porcine Breast Milk Exosomes PLoS One 2012 7 e43691 22937080 167 Sun Q Chen X Yu J Zen K Zhang CY Li L Immune modulatory function of abundant immune-related microRNAs in microvesicles from bovine colostrum Protein Cell 2013 4 197 210 23483481 168 Parigi SM Eldh M Larssen P Gabrielsson S Villablanca EJ Breast Milk and Solid Food Shaping Intestinal Immunity Front Immunol 2015 6 415 26347740 169 Pieters BC Arntz OJ Bennink MB Broeren MG van Caam AP Koenders MI van Lent PL van den Berg WB de Vries M van der Kraan PM van de Loo FA Commercial Cow Milk Contains Physically Stable Extracellular Vesicles Expressing Immunoregulatory TGF-β PLoS One 2015 10 e0121123 25822997 170 Melnik BC John SM Schmitz G Milk: An exosomal microRNA transmitter promoting thymic regulatory T cell maturation preventing the development of atopy? J Transl Med 2014 12 43 24521175 171 Izumi H Tsuda M Sato Y Kosaka N Ochiya T Iwamoto H Namba K Takeda Y Bovine milk exosomes contain microRNA and mRNA and are taken up by human macrophages J Dairy Sci 2015 98 2920 2933 25726110 172 Li QJ Chau J Ebert PJ Sylvester G Min H Liu G Braich R Manoharan M Soutschek J Skare P Klein LO Davis MM Chen CZ miR-181a Is an Intrinsic Modulator of T Cell Sensitivity and Selection Cell 2007 129 147 161 17382377 173 Momen-Heravi F Bala S Bukong T Szabo G Exosome-mediated delivery of functionally active miRNA-155 inhibitor to macrophages Nanomedicine 2014 10 1517 1527 24685946 174 Torregrosa Paredes P Gutzeit C Johansson S Admyre C Stenius F Alm J Scheynius A Gabrielsson S Differences in exosome populations in human breast milk in relation to allergic sensitization and lifestyle Allergy 2014 69 463 471 24428462 175 Arntz OJ Pieters BC Oliveira MC Broeren MG Bennink MB de Vries M van Lent PL Koenders MI van den Berg WB van der Kraan PM van de Loo FA Oral administration of bovine milk derived extracellular vesicles attenuates arthritis in two mouse models Mol Nutr Food Res 2015 59 1701 1712 26047123 176 Kim JH Jeun EJ Hong CP Kim SH Jang MS Lee EJ Moon SJ Yun CH Im SH Jeong SG Park BY Kim KT Seoh JY Kim YK Oh SJ Ham JS Yang BG Jang MH Extracellular vesicle-derived protein from Bifidobacterium longum alleviates food allergy through mast cell suppression J Allergy Clin Immunol 2016 137 507 516 e8 26433560 177 Chu H Khosravi A Kusumawardhani IP Kwon AH Vasconcelos AC Cunha LD Mayer AE Shen Y Wu WL Kambal A Targan SR Xavier RJ Ernst PB Green DR McGovern DP Virgin HW Mazmanian SK Gene-microbiota interactions contribute to the pathogenesis of inflammatory bowel disease Science 2016 352 1116 1120 27230380 178 Cheng Y Schorey JS Exosomes carrying mycobacterial antigens can protect mice against Mycobacterium tuberculosis infection Eur J Immunol 2013 43 3279 3290 23943377 179 Alexander M Hu R Runtsch MC Kagele DA Mosbruger TL Tolmachova T Seabra MC Round JL Ward DM O’Connell RM Exosome-delivered microRNAs modulate the inflammatory response to endotoxin Nat Commun 2015 6 7321 26084661 180 Aline F Bout D Amigorena S Roingeard P Dimier-Poisson I Toxoplasma gondii antigen-pulsed-dendritic cell-derived exosomes induce a protective immune response against T. gondii infection Infect Immunol 2004 72 4127 4137 15213158 181 Kim OY Hong BS Park KS Yoon YJ Choi SJ Lee WH Roh TY Lötvall J Kim YK Gho YS Immunization with Escherichia coli Outer Membrane Vesicles Protects Bacteria-Induced Lethality via Th1 and Th17 Cell Responses J Immunol 2013 190 4092 4102 23514742 182 Kim SH Bianco N Menon R Lechman ER Shufesky WJ Morelli AE Robbins PD Exosomes derived from genetically modified DC expressing FasL are anti-inflammatory and immunosuppressive Mol Ther 2006 13 289 300 16275099 183 Kim SH Bianco NR Shufesky WJ Morelli AE Robbins PD Effective treatment of inflammatory disease models with exosomes derived from dendritic cells genetically modified to express IL-4 J Immunol 2007 179 2242 2249 17675485 184 Ruffner MA Kim SH Bianco NR Francisco LM Sharpe AH Robbins PD B7–1/2, but not PD-L1/2 molecules, are required on IL-10-treated tolerogenic DC and DC-derived exosomes for in vivo function Eur J Immunol 2009 39 3084 3090 19757438 185 Kim SH Lechman ER Bianco N Menon R Keravala A Nash J Mi Z Watkins SC Gambotto A Robbins PD Exosomes derived from IL-10-treated dendritic cells can suppress inflammation and collagen-induced arthritis J Immunol 2005 174 6440 6448 15879146 186 Bianco NR Kim SH Ruffner MA Robbins PD Therapeutic effect of exosomes from indoleamine 2,3-dioxygenase-positive dendritic cells in collagen-induced arthritis and delayed-type hypersensitivity disease models Arthritis Rheum 2009 60 380 389 19180475 187 Karlsson MR Kahu H Hanson LA Telemo E Dahlgren UI An established immune response against ovalbumin is suppressed by a transferable serum factor produced after ovalbumin feeding: a role of CD25+ regulatory cells Scand J Immunol 2002 55 470 477 11975758 188 Karlsson MR Kahu H Hanson LA Telemo E Dahlgren UI Tolerance and bystander suppression, with involvement of CD25-positive cells, is induced in rats receiving serum from ovalbumin-fed donors Immunology 2000 100 326 333 10929054 189 Yang C Kim SH Bianco NR Robbins PD Tumor-derived exosomes confer antigen-specific immunosuppression in a murine delayed-type hypersensitivity model PLoS One 2011 6 e22517 21829629 190 Kim SH Bianco NR Shufesky WJ Morelli AE Robbins PD MHC class II+ exosomes in plasma suppress inflammation in an antigen-specific and Fas ligand/Fas-dependent manner J Immunol 2007 179 2235 2241 17675484 191 Ptak W Nazimek K Askenase PW Bryniarski K From mysterious supernatant entity to miRNA-150 in antigen-specific exosomes: a history of hapten-specific T suppressor factor Arch Immunol Ther Exp 2015 63 345 356 192 Bryniarski K Nazimek K Martin E Ptak M Askenase PW Ptak W CD8+ T cell-derived exosomal miRNA-150 suppresses induction and effector phases of murine contact sensitivity as well as symptoms of active allergy Centr Eur J Immunol 2014 39 Supplement 1 20 193 Dey N Szczepanik M Lau K Majewska M Askenase P Stimulatory lipids accumulate in the mouse liver within 30 min of contact sensitization to facilitate the activation of naive iNKT cells in a CD1d-dependent fashion Scand J Immunol 2011 74 52 61 21352253 194 Askenase PW Bryniarski K Paliwal V Redegeld F Groot Kormelink T Kerfoot S Hutchinson AT van Loveren H Campos R Itakura A Majewska-Szczepanik M Yamamoto N Nazimek K Szczepanik M Ptak W A subset of AID-dependent B-1a cells initiates hypersensitivity and pneumococcal pneumonia resistance Ann N Y Acad Sci 2015 1362 200 214 26662721 195 Kerfoot S Szczepanik M Tung J Askenase P Identification of initiator B-cells, a novel subset of activation induced deaminase-dependent B-1-like cells that mediate initiation of contact sensitivity J Immunol 2008 181 1717 1727 18641308 196 Yang Y Wang C Yang Q Kantor AB Chu H Ghosn EE Qin G Mazmanian SK Han J Herzenberg LA Distinct mechanisms define murine B cell lineage immunoglobulin heavy chain (IgH) repertoires eLife 2015 4 e09083 26422511 197 Itakura A Szczepanik M Campos RA Paliwal V Majewska M Matsuda H Takatsu K Askenase PW An hour after immunization peritoneal B-1 cells are activated to migrate to lymphoid organs where within 1 day they produce IgM antibodies that initiate elicitation of contact sensitivity J Immunol 2005 175 7170 7178 16301620 198 Askenase PW Rosenstein RW Ptak W T cells produce an antigen binding factor with in vivo activity analogous to IgE antibody J Exp Med 1983 157 862 873 6187880 199 Geba GP Wegner CD Wolyniec WW Li Y Askenase PW Nonatopic asthma: in vivo airway hyperreactivity adoptively transferred to naive mice by THY-1(+) and B220(+) antigen-specific cells that lack surface expression of CD3 J Clin Invest 1997 100 629 638 9241124 200 Szczepanik M Akahira-Azuma M Bryniarski K Tsuji RF Kawikova I Ptak W Kiener C Campos RA Askenase PW B-1 B cells mediate required early T cell recruitment to elicit protein-induced delayed-type hypersensitivity J Immunol 2003 171 6225 6235 14634139 201 Redegeld FA van der Heijden MW Kool M Heijdra BM Garssen J Kraneveld AD Van Loveren H Roholl P Saito T Verbeek JS Claassens J Koster AS Nijkamp FP Immunoglobulin-free light chains elicit immediate hypersensitivity-like responses Nat Med 2002 8 694 701 12068287 202 Hutchinson AT Jones DR Raison RL The ability to interact with cell membranes suggests possible biological roles for free light chain Immunol Lett 2012 142 75 77 22138405 203 Hutchinson AT Ramsland PA Jones DR Agostino M Lund ME Jennings CV Bockhorni V Yuriev E Edmundson AB Raison RL Free Ig light chains interact with sphingomyelin and are found on the surface of myeloma plasma cells in an aggregated form J Immunol 2010 185 4179 4188 20817866 204 Hutchinson AT Malik A Berkahn MB Agostino M To J Tacchi JL Djordjevic SP Turnbull L Whitchurch CB Edmundson AB Jones DR Raison RL Ramsland PA Formation of assemblies on cell membranes by secreted proteins: molecular studies of free λ light chain aggregates found on the surface of myeloma cells Biochem J 2013 454 479 489 23822104 205 Nazimek K Ptak W Nowak B Ptak M Askenase PW Bryniarski K Macrophages play an essential role in antigen-specific immune suppression mediated by T CD8+ cell-derived exosomes Immunology 2015 146 23 32 25808106 206 Nazimek K Nowak B Marcinkiewicz J Ptak M Ptak W Bryniarski K Enhanced generation of reactive oxygen intermediates by suppressor T cell-derived exosome-treated macrophages Folia Med Cracov 2014 54 37 52 25556365

PMC005xxxxxx/PMC5131369.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 101573691 39703 Cell Rep Cell Rep Cell reports 2211-1247 27851966 5131369 10.1016/j.celrep.2016.10.068 NIHMS829661 Article Pancreatic Inflammation Redirects Acinar to Beta Cell Reprogramming Clayton Hannah W. 12 Osipovich Anna B. 23 Stancill Jennifer S. 12 Schneider Judsen D. 2 Vianna Pedro G. 2 Shanks Carolyn M. 2 Yuan Weiping 2 Gu Guoqiang 12 Manduchi Elisabetta 4 Stoeckert Christian J. Jr. 4 Magnuson Mark A. 123* 1 Department of Cell and Developmental Biology, Vanderbilt University, Nashville, TN, 37232, USA 2 Center for Stem Cell Biology, Vanderbilt University, Nashville, TN, 37232, USA 3 Department of Molecular Physiology and Biophysics, Vanderbilt University, Nashville, TN, 37232, USA 4 Institute for Biomedical Informatics and Department of Genetics, University of Pennsylvania School of Medicine, Philadelphia, PA, 19104, USA * Corresponding author: mark.magnuson@vanderbilt.edu 25 11 2016 15 11 2016 01 12 2016 17 8 20282041 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. Summary Using a transgenic mouse model to express MafA, Pdx1, and Neurog3 (3TF) in a pancreatic acinar cell- and doxycycline-dependent manner, we discovered that the outcome of transcription factor-mediated acinar to β-like cellular reprogramming is dependent on both the magnitude of 3TF expression and on reprogramming-induced inflammation. Overly robust 3TF expression causes acinar cell necrosis resulting in marked inflammation and acinar-to-ductal metaplasia. Generation of new β-like cells requires limiting reprogramming-induced inflammation, either by reducing 3TF expression or by eliminating macrophages. The new β-like cells were able to reverse streptozotocin-induced diabetes 6 days after inducing 3TF expression but failed to sustain their function after removal of the reprogramming factors. Graphical Abstract Introduction Reprogramming of pancreatic cells into new β-like cells represents a potential therapy for Type 1 diabetes (Bramswig et al., 2013; Dor et al., 2004; Li et al., 2014c; Thorel et al., 2010; Zhou et al., 2008). Pancreatic acinar cells are an appealing target for cellular reprogramming since they are abundant, derived from a common progenitor cell during pancreatic organogenesis (Gu et al., 2002), and exhibit significant transcriptional plasticity (Li et al., 2014c; Puri et al., 2015; Ziv et al., 2013). Towards this end, Zhou et al. reported that adenoviral-mediated expression of three pancreas-specific transcription factors MafA, Pdx1, and Neurog3 (3TF) in immunocompromised Rag1−/− mice results in the conversion of pancreatic acinar cells into new insulin-secreting β-like cells (Zhou et al., 2008). In addition, transient administration of epidermal growth factor and ciliary neurotrophic factor has also been reported to convert pancreatic acinar cells into new β-like cells (Baeyens et al., 2013). While the reports of acinar to β-cell (A→β) reprogramming appear promising, the effects of reprogramming on the microscopic anatomy, cellular function, and physiological function of the pancreas have not been explored but would be expected to be substantial due to the very marked physiological and histological differences between acinar and β-cells. In contrast to pancreatic β-cells, acinar cells produce copious amounts of proteases, lipases, and ribonucleases whose potentially auto-digestive abilities require sequestration mechanisms to prevent endogenous tissue damage (Logsdon and Ji, 2013). The exocrine pancreas protects itself from autodigestion through several mechanisms. First, many of the enzymes are secreted as inactive pro-enzymes, or zymogens, which only become active within the duodenum (Neurath and Walsh, 1976). Second, the proteolytic enzymes are co-secreted with a trypsin inhibitor that prevents premature activation of trypsinogen, which normally becomes activated in the small intestine and is responsible for activation of the other precursor digestive enzymes (Logsdon and Ji, 2013). Third, acinar-to-ductal metaplasia (ADM) occurs (Bockman et al., 1997; Liou et al., 2013; Pan et al., 2013) and has been suggested to limit autodigestion in the face of acinar cell injury (Puri et al., 2015). ADM, the conversion of acinar cells into a non-secretory duct-like cell, is characterized by the formation of duct-like complexes and fibrosis (Wang et al., 1995) in response to pancreatic inflammation. The mechanisms that initiate the inflammation are disputed. Some argue that it is due to intracellular activation of trypsinogen (Halangk et al., 2000; Szilagyi et al., 2001; Van Acker et al., 2002; Whitcomb et al., 1996) whereas others have suggested that it is due to calcium overload (Li et al., 2014a) and endoplasmic reticulum (ER) stress (Ji et al., 2003; Logsdon and Ji, 2013). In either case, ADM is characterized by aberrant expression of cytokeratins (Strobel et al., 2007), Pdx1, Sox9, and Onecut1 in pancreatic acinar cells (Rooman and Real, 2012). In order for an in vivo β-cell restorative therapy to become clinically feasible, a better understanding of the factors that modulate intercellular conversions and the physiological effects that such conversions may induce is required. Towards this end, we developed a diallelic transgene-based mouse model that expresses 3TF specifically in pancreatic acinar cells in a tetracycline-dependent manner. Such a model enables 3TF expression to be modulated in a manner that is unachievable using a virus-based expression system, thereby allowing us to examine the effects of both 3TF concentration and duration on generating new β-like cells. Our studies using this model indicate that the level of 3TF expression has a major influence, not only on reprogramming success but also on tissue response. Indeed, we found that robust 3TF expression causes acinar cell stress, marked inflammation, and ADM, and that attenuating reprogramming-induced inflammation, either by reducing 3TF expression or eliminating macrophages, results in the production of new β-like cells. Moreover, the duration of factor expression may also play a role in the reprogramming outcome since the ability of new β-like cells to improve glycemia was dependent on the dox-induced expression of 3TF, with removal of dox resulting in a worsening of glycemic control and reversion to a fully diabetic state within a few days. Results Design and validation of mouse alleles To explore the cellular dynamics of pancreatic A→β reprogramming, we developed a diallelic transgene-based mouse model that co-expresses both 3TF and mCherry specifically in pancreatic acinar cells in a tetracycline-inducible manner (Figure 1A). The first allele (Rosa263TF.mCherry) was made by replacing the coding sequences for Ptf1a with those for the reverse tetracycline Trans-Activator (rtTA). In the second allele (Ptf1artTA), a bi-directional Tet-operator cassette that in one direction expresses a 2A peptide-linked fusion gene of MafA, Pdx1, and Neurog3 and in the other direction the red fluorescent protein mCherry was inserted into a functionally disabled Rosa26/SetD5 gene locus (Chen et al., 2011). When adult mice containing both the Ptf1artTA and Rosa263TF.mCherry alleles were given 2.0 mg/ml doxycycline (dox) in their drinking water for 1 day, red fluorescence was observed in the pancreas but no other visceral organs (Figure 1B). Furthermore, mCherry expression in the pancreas was restricted to acinar cells and not observed in pancreatic ducts or endocrine cells, as expected due to the acinar cell-restricted expression of Ptf1a at this age (Figure 1C, S1A). Immunofluorescent staining for both mCherry and amylase indicated that 78% (±2.5%; n=3) of acinar cells expressed mCherry. Staining for mCherry and for each of the reprogramming factors showed that virtually all mCherry-positive (+) cells expressed the reprogramming factors (Figure 1D). To further validate the experimental model, we analyzed the function of the 2A peptide-cleaved transcription factors generated by the Rosa263TF.mCherry allele. First, immunoblot analysis showed PDX1, which is flanked by MAFA and NEUROG3 protein sequences in the 2A peptide-containing cassette, to be properly cleaved from the two surrounding proteins (Figure 1E). Second, analysis of protein function using reporter genes showed that each of the 2A peptide-modified proteins functioned in a normal manner, indistinguishable from that of their wild-type counterparts (Figure S1B–E). Third, a recombinant adenovirus containing the 3TF fusion gene, when injected together with a GFP-expressing virus into the pancreas of Rag1−/− mice, resulted in scattered insulin+/GFP+ co-expressing cells within the exocrine compartment of the pancreas, similar to those observed by Zhou et al. (Zhou et al., 2008) (Figure S1F, G). Together, these findings confirmed that the 2A peptide-modified MAFA, PDX1, and NEUROG3 made in response to dox-induction functioned normally. Lack of A→β reprogramming in transgenic mice after 3TF induction Since viral mediated expression of 3TF in the pancreas has previously been reported to cause A→β conversion in 10 days or less (Zhou et al., 2008), we treated our diallelic mice with dox for 1, 7, and 28 days and performed immunostaining for acinar and endocrine cell markers, including insulin. At 1 day, the 3TF-induced cells resembled acinar cells with all mCherry+ cells seen to express amylase, an acinar cell-specific marker. However, at 7 and 28 days, the expression of amylase in mCherry+ cells was greatly diminished or absent (Figure 2A, B). In addition, the mCherry+ cells were smaller and located in tubular-like cell clusters (Figure 2A). Chromogranin A, an endocrine cell marker that was absent in the 1 day sample, was expressed in nearly 100% of mCherry+ cells at 7 and 28 days (Figure 2A, B). Interestingly, despite the widespread expression of this endocrine marker in mCherry+ cells, we failed to observe any insulin, glucagon, somatostatin, or pancreatic polypeptide expression by immunofluorescence staining. However, after 7 days of dox treatment, we did observe that approximately 40% of mCherry+ cells expressed ghrelin, a hormone normally expressed in 1% or fewer adult pancreatic endocrine cells (Arnes et al., 2012). The portion of cells expressing ghrelin rose to nearly 60% in the 28 day sample (Figure 2A, B). RNA profiling of the 3TF-treated cells To better understand why new β-like cells were not observed and to corroborate the immunostaining results, we performed RNA-Seq on FACS-purified mCherry+ cells after 1 and 7 days of dox administration and compared their transcriptional profiles to FACS-purified uninduced acinar cells (Table S1). These datasets revealed that MafA, Pdx1, and Neurog3 mRNAs were all highly up-regulated after 1 day of dox treatment compared to uninduced acinar cells. Indeed, MafA increased from 0.67 ± 0.67 to 4069 ± 169, Pdx1 from 20 ± 4 to 7037 ± 471, and Neurog3 from 0 to 10931 ± 629 normalized counts (Table S1), strongly suggesting that the lack of insulin gene expression was not due to insufficient 3TF expression. Furthermore, inspection of the 7 day 3TF-induced RNA-seq dataset showed that while some genes that characterize either immature or mature β-cells, such as ChgA, Ghrl, Neurod1, and Insm1, were highly up-regulated compared to uninduced acinar cells, other genes that are normally present in β-cells such as Ins1, Ins2, Nkx6.1, Isl1, and Pax6 were not identified as significantly up-regulated in the 3TF-induced acinar cells (Figure S2A and Table S2, S3). In addition, many of the endocrine-specific genes that were up-regulated have been previously shown to be direct DNA binding targets for either Neurog3, Pdx1, or MafA. For instance, Neurod1, Nkx2.2, and Insm1 are targets of Neurog3 (Huang et al., 2000; Smith et al., 2003; Watada et al., 2003) and Pdx1 binds to the promoters of both Gck and Slc2a2 (Khoo et al., 2012; Watada et al., 1996). Importantly, we also noticed that many of the upregulated genes from the RNA-seq data set of 7 day dox-induced acinar cells were associated with inflammation (Figure S2B). Taken together, these findings suggest that while 3TF expression increased expression of several endocrine-specific genes, it did not cause acinar cells to thoroughly adopt a β-cell-like gene expression profile. Transgene-based expression of 3TF results in acinar-to-ductal metaplasia To confirm that inflammation-associated genes were expressed in response to 3TF-induction, we stained pancreata for inflammatory cells using CD45, a pan-leukocyte marker, and observed that nearly a third of the cells present in the pancreas were leukocytes (Figure 3A, B). Staining for F4/80, a macrophage marker, and CD3, a T-cell marker, revealed that both inflammatory cells were present with the majority being macrophages (Figure S3A, B). Finally, Masson’s trichrome staining revealed extensive fibrosis further indicating a potent inflammatory response (Figure S3C). Since pancreatic inflammation has been linked to metaplastic changes, we examined the histological appearance of 7 day-induced pancreatic tissue. After 7 days of 3TF expression, pancreata of the diallelic mice were smaller (Figure 3C) and characterized by the presence of many tubular complexes (Figure 3D), all of which are hallmarks of ADM (Jura et al., 2005; Parsa et al., 1985). Immunostaining for cytokeratin, a marker for pancreatic duct cells, further revealed that the 3TF-expressing mCherry+ cells had adopted duct-like characteristics (Figure 3E, F). After 2 days of dox, and prior to the onset of ADM, we observed acinar cell necrosis (Figure S3D). Since marked inflammation and metaplasia was not observed in our viral control experiments (Figure S1H, I) or previously reported by others when an adenovirus was used to introduce 3TF to the pancreas of normoglycemic mice (Cavelti-Weder et al., 2016; Zhou et al., 2008), we hypothesized that the overly robust 3TF expression that occurred using 2.0 mg/ml of dox was responsible for both the immune response and metaplastic changes. To determine why 3TF expression in our diallelic transgenic model was causing pancreatic inflammation, we considered the fact that pancreatic acinar cells are especially vulnerable to ER dysfunction owing to their high level of protein synthetic and secretory activity (Ji et al., 2003; Logsdon and Ji, 2013). In support of this notion, we noticed that 44 of 83 genes involved in the activation of the unfolded protein response (UPR) were upregulated after 7 days of 3TF expression (Figure S3E). In addition, expression of genes encoding voltage-gated and other Ca2+ channels (Table S4) were also markedly increased. In acinar cells, a rise in [Ca2+]i has been associated with ADM and is known to cause activation of inflammatory genes and the ER stress response (Sah et al., 2014). These findings are consistent with activation of the ER stress response in 3TF-expressing acinar cells, possibly by disrupting intracellular calcium homeostasis. Rag1 or adenovirus infection does not alter the reprogramming outcome Given that new β-like cells were observed only when 3TF was expressed by an adenoviral vector, but not from a transgene, we performed several experiments to exclude a role for two experimental variables that might have accounted for this divergent reprogramming outcome. The first experimental variable was the presence or absence of Rag1 (Zhou et al., 2008), a gene required for V(D)J recombination in B and T cells. Rag1 null mice were used in the adenoviral expression studies to prevent clearance of transcription factor-expressing viruses whereas the transgene-based expression studies utilized Rag1+/+ animals since viral clearance was not an issue. Since it is possible that the lack of mature B and T lymphocytes in Rag1−/− mice might modulate the inflammatory response and allow reprogramming to occur, we crossed the Rag1 null allele into our diallelic mice to derive Ptf1artTA/+; Rosa263TF.mCherry/+; Rag1−/− mice, then treated these mice with dox for 7 days. Pancreas immunostaining for CD3 showed that T cells were indeed absent (Figure S4A). However, similar to the Rag1+/+ mice, many F4/80+ cells were observed in the pancreas after 7 days of dox treatment (Figure S4B) and many cytokeratin+ tubular complexes were formed (Figure S4C,D) that expressed ghrelin (Figure S4F). Masson’s trichrome staining again revealed extensive fibrosis (Figure S4E) and, more importantly, there were no cells that co-expressed mCherry and insulin (Figure S4G). These findings showed that the presence or absence of Rag1 has no substantive role in determining the reprogramming outcome when we used 2.0 mg/ml dox to induce 3TF expression. The second experimental variable was adenoviral infection. Since it has been reported that viral infection and the presence of adenoviral regulatory proteins could improve reprogramming efficiency (Lee et al., 2012; Wang et al., 2007; Zaldumbide et al., 2012), we simultaneously induced 3TF expression for 7 days using 2.0 mg/ml dox while also injecting a GFP-expressing adenovirus directly into the pancreas (Figure S5A, B). Once again, and despite examining over 600 mCherry and GFP co-expressing cells, no cells were seen that also expressed insulin (Figure S5D). Instead, many of the mCherry and GFP co-expressing cells continued to express ghrelin (Figure S5C). These findings enabled us to exclude the host immune response to viral infection as having a major role in the divergent reprogramming outcome when using 2.0 mg/ml dox to induce 3TF expression. Reducing the level of 3TF expression lowers the coinciding immune response promoting A→β reprogramming Given that transgenic expression of 3TF induced widespread pancreatic inflammation, we next explored the effect of varying 3TF expression by treating the diallelic mice with a 10- and 100-fold lower concentration of dox for 7 days. Both groups of mice exhibited lower levels of mCherry fluorescence, consistent with 3TF expression being reduced (Figure 4A, B). We also observed that fewer acinar cells expressed mCherry as the concentration of dox was lowered. Indeed, a 10-fold lower concentration of dox to induced 3TF expression resulted in over a 60% reduction in the number of acinar cells expressing 3TF after 1 day of dox (78 ± 2.5% vs. 30 ± 9.9%; n=3). In addition, after 7 days of dox treatment, the pancreas in both groups was nearly normal in size (Figure 4A), cytokeratin staining was either reduced or absent (Figure S6A, B), and CD45 staining was reduced (Figure 4C, D). Most importantly, mice treated with the 10-fold lower concentration of dox (0.2 mg/ml) exhibited many mCherry+ cells that co-expressed low levels of insulin (Figure 4E, F). On the other hand, mice given the lowest concentration of dox (0.02 mg/ml) had no mCherry+/insulin+ co-expressing cells nor any mCherry+ cells that co-expressed chromogranin A or ghrelin (Figure S6C–F). Instead, the mCherry+ cells in these animals continued to express amylase at 7 days (Figure S6G–H), indicating that they had undergone very little, if any, A→β reprogramming within the one week experimental timeframe. Taken together, these findings indicate that A→β reprogramming depends on the magnitude of 3TF expression. When it is too low, no reprogramming occurs. Conversely, when it is too high, a potent inflammatory response occurs which diverts the reprogramming outcome to an ADM-like phenotype. Macrophage depletion permits A→β reprogramming We next sought to determine whether the inflammatory response might be responsible for the divergent reprogramming outcome. Since the majority of immune cells present after 7 days of 3TF expression were macrophages, we administered gadolinium chloride (GdCl3), a macrophage toxin (Jankov et al., 2001), both prior to and during dox treatment. The depletion of macrophages in the pancreas was confirmed by immunostaining for F4/80 (Figure 5A, B). Strikingly, the pancreata of mice treated with GdCl3 were nearly normal in both size and appearance (Figure 5C, D). Animals administered GdCl3 did not exhibit either the formation of tubular complexes or the diffuse cytokeratin-staining that characterized the diallelic mice not administered GdCl3 (Figure 5E–G and 6A) nor did they exhibit fibrosis (Figure 6B). More importantly, approximately 6% of the mCherry+ cells present after 7 days of dox treatment co-expressed insulin (Figure 6C, D). Cell counting after 7 days of treatment revealed approximately 650,000 new β-like cells per animal (Table S5). Interestingly cell counting also revealed a decrease in the number of cells expressing mCherry at 7 days of dox. To further confirm that macrophage depletion promotes 3TF-mediated A→β reprogramming, RT-qPCR was performed on FACS-sorted mCherry+ cells. Indeed, both Ins1 and Ins2 mRNA were increased after 7 days of dox in mice administered GdCl3 (Figure 6E). Contrary to immunofluorescence analysis, RT-qPCR also revealed a modest increase in Ins1 and Ins2 mRNA in FACS-sorted mCherry+ cells from mice that were only administered dox (Figure 6E). These discordant findings suggest that dox-only treated mice may produce Ins1 and Ins2 mRNA but not insulin protein, perhaps due to the inflammation, ER stress, and activation of the UPR, which is known to impair protein translation. These findings suggest that the overly robust expression of 3TF triggers a potent inflammatory response, mediated by macrophages, that prevents A→β reprogramming. To assess the function of the newly generated β-like cells, mice were rendered diabetic by administration of the β-cell toxin streptozotocin (STZ), treated with GdCl3 to attenuate inflammation, then robust 3TF expression was induced with 2 mg/ml dox. Within two days of dox treatment, the diabetic mice began to exhibit an improvement in blood glucose concentration, and at 6 days, their blood glucose concentrations were indistinguishable from untreated control animals (Figure 7A). Furthermore, by day 7 of dox, two mice died with low blood glucose levels (86 and 72 mg/dl). Removal of dox at day 7 was followed immediately by a worsening of glycemic control and reversion to a diabetic state within a few days. Interestingly, intraperitoneal glucose tolerance tests (GTT) performed after 7 days of dox treatment revealed two different patterns of response in the five mice treated. Two mice appeared to be glucose responsive whereas the other three mice, which were hypoglycemic at the beginning of GTT, lacked glucose-sensitive insulin secrection (Figure 7B). In any case, these findings indicate a sufficient number of new β-like cells were produced to rescue STZ-induced diabetes and that the production of insulin by these cells was dox-dependent. In addition, our results suggest the presence of at least two subgroups of reprogrammed β-like cells, one group that is glucose responsive and a second that may constitutively secrete insulin in a non-glucose responsive manner. Finally, we sought to determine whether the efficiency of 3TF-mediated A→β reprogramming could be increased by simultaneously lowering the concentration of dox and attenuating inflammation by depleting macrophages. Interestingly, while the dual treatment very clearly preserved pancreatic mass and histology (Figure S7A–C), prevented abnormal cytokeratin staining (Figure S7D–F), and further decreased tissue inflammation (Figure S7G–I), it did not increase the overall reprogramming efficiency (Figure S7J, K). Discussion Our findings indicate that the overly robust expression of 3TF in acinar cells induces pancreatic inflammation which blocks A→β reprogramming and results, instead, in the production of new duct-like cells (Figure 7C). Only when inflammation is attenuated, either by reducing the intensity of 3TF expression or by depleting macrophages, does the production of new β-like cells occur. We suggest that when the concentration of 3TF is too high, pancreatic acinar cell stress and damage occurs thereby causing cytokine release, macrophage infiltration, and ADM, which prevents A→β reprogramming. We further suggest that efficient reprogramming requires a coordinated series of events whereby acinar cells cease zymogen production, delaminate and migrate to the surrounding mesenchyme, then cluster into new vascularized islets. For these cellular changes to occur without causing cellular damage and inducing inflammation, acinar cells may need time to cease the production and secretion of tissue-digesting enzymes before delamination from the pancreatic duct. The pace at which reprogramming occurs is likely influenced by the concentration of the reprogramming factors. If 3TF expression is too high, and reprogramming occurs at a rate that exceeds the ability of the cell to undergo an orderly transition from one cell state (acinar) to another (endocrine), acinar cell damage may occur, preventing the formation of new β-like cells. Indeed, when we reduced the level of 3TF expression, the inflammatory response was attenuated and new β-like cells were produced. Conversely, if 3TF levels are too low, no reprogramming occurs, at least within a one week timeframe. Such a conclusion is consistent with previous studies that have shown the importance of factor levels in achieving successful reprogramming (Carey et al., 2011; Tonge et al., 2014). We also found that the presence of inflammatory macrophages within the pancreas greatly influences the outcome of 3TF-mediated acinar cell reprogramming. While depleting either T- or B-cells during reprogramming does not prevent ADM or allow for the production of new β-like cells, the depletion of macrophages, by the administration of GdCl3, prevented ADM, thereby enabling acinar cells to be reprogrammed into β-like cells. The mechanisms involved in the macrophage-dependent blockage of A→β reprogramming remain unclear, but it has been reported that macrophage-secreted cytokines mediate ADM through activation of NFκB and STAT3 (Liou et al., 2013). While our RNA-Seq data suggest that inflammation enhances signaling through NFκB, STAT3 signaling was recently shown to be required for cytokine-mediated A→β conversion (Baeyens et al., 2013). Thus, it is possible that NFκB and STAT3 signaling oppose each other, with STAT3 signaling promoting A→β reprogramming and NFκB signaling impairing reprogramming by causing ADM. We estimate that approximately 650,000 new β-like cells are produced, on average, in response to administration of 2 mg/ml dox and GdCl3. While this number of cells is about a third of the approximately 2 million β-cells in an average mouse pancreas (Dor et al., 2004), it is sufficient to transiently rescue STZ-induced diabetes. However, the ability of these new β-like cells to stably secrete insulin in a glucose-dependent manner is not established after only 6 days of 3TF treatment since the removal of dox at 7 days caused a quick reversion to a diabetic state. These findings are consistent with prior studies that used adenoviral delivery of the same three transcription factors in which two months were required for the reprogrammed acinar cells to adopt a DNA methylation and transcriptional profile similar to that of endogenous β-cells (Li et al., 2014b). Thus, our findings support the notion that an extended exposure to the three reprogramming factors is necessary for acinar cells to adopt both the epigenetic and transcriptional profile of an endogenous β-cell. Adenoviral delivery of the reprogramming factors has been reported to result in 40–50% of infected acinar cells being converted to new β-like cells (Li et al., 2014b). While we expected that the transgenic delivery of the factors would further improve reprogramming efficiency, we found the opposite with only 6% of 3TF-expressing cells expressing insulin after 7 days. This suggests that there are additional variables that distinguish adenoviral- and transgene-mediated reprogramming, such as the dynamics of 3TF-expression. While use of the Tet-On system has the distinct advantage of allowing us to simultaneously control the concentration and duration of 3TF expression, it does not allow stable expression of the reprogramming factors over an extended time. Use of the Ptf1a gene to drive expression of rtTA, while being a straightforward means of achieving acinar-cell specificity, has the limitation that Ptf1a expression is extinguished as acinar cells are converted into new β-cells. This limitation can only be overcome with a more complicated transgene design. Widespread metaplastic changes resulting from tissue inflammation have not been reported when adenoviruses are used to introduce the reprogramming factors into the pancreas of normoglycemic mice (Cavelti-Weder et al., 2016). We speculate that similar metaplastic conversions are not observed due to the low infection efficiency of the adenovirus, which results in fewer acinar cells expressing the reprogramming factors, thereby avoiding triggering widespread pancreatic inflammation and allowing for more of the virally-infected cells to be reprogrammed. However, despite the highly penetrant expression of 3TF in our transgene-based model (nearly 80% of acinar cells at 2 mg/ml dox), we could only achieve the visible reprogramming of 6% of the 3TF-induced cells. Even so, we were able to produce over twice the number of new insulin-positive cells (650,000 versus 245,000 ± 32,000) that were reported when using an adenovirus (Li et al., 2014c). This very marked difference in experimental outcome is likely due to our inability to fully suppress pancreatic inflammation when using a high dose of dox to express 3TF in the maximum number of acinar cells. In any case, the very marked differences in our experimental outcomes compared to those achieved using an adenovirus suggest that both the dynamics and secondary effects of A→β reprogramming are complex. However, by exploring the outcomes achieved with a transgene-based reprogramming model, we have obtained a better understanding of some of the many variables that will need to be understood to be able to safely and efficiently reprogram acinar cells in humans. Experimental Procedures Mouse lines and husbandry All animals were housed at Vanderbilt Institutional Animal Care Facility and experimental protocols were approved by Vanderbilt Institutional Animal Care and Use Committee. Mice were treated with doxycycline (dox) (Sigma) dissolved in a 5% sucrose solution beginning at 6 to 8 weeks of age. Dox was provided ad libitum in lieu of the normal water supply. Rag1−/− mice (Mombaerts et al., 1992) were purchased from Jackson Laboratory. Ptf1aYFP/+ (Burlison et al., 2008) and MIP.GFP mice (Hara et al., 2003) were genotyped as previously described. Generation of new mouse alleles and validation methods are described in Supplemental Information. Microscopy mCherry fluorescence intensity measurements in unfixed tissues were performed using a Leica MZ16 FA stereoscope at an exposure time of 39.5 milliseconds. For paraffin sections, pancreatic tissue was fixed with 20% formaldehyde and processed by the Vanderbilt Tissue Pathology Shared Resource (TPSR). H&E, Masson’s Trichrome Blue, CD3, F4/80, and Cytokeratin staining was performed by TPSR according to manufacturer’s directions. Immunofluorescence staining of frozen sections was performed as previously described (Burlison et al., 2008). See Supplemental Information for a complete list of antibodies. Images were acquired using a Zeiss Axioplan-II upright microscope or a LSM 710 META inverted confocal microscope then pseudo-colored using either ImageJ (NIH) or Zeiss LSM browser software. All images are representative of phenotypes observed in at least three different animals. Adenovirus construction and injection The AdV-CMV-3TF virus was made using pAd/CMV/V5-DEST (Invitrogen). High titer virus (6.5 × 1010 plaque-forming units (pfu)) was obtained by purification (Vector biolabs). Mice were subjected to laparotomy under general anesthesia (Ketamine/Xylazine). The splenic lobe of the dorsal pancreas of 8 week old Rag1−/− mice was injected with 100 μl of purified AdV-CMV-3TF (2×1010 pfu) and AdV-CMV-GFP (1×109 pfu) (Vector Biolabs) and animals were euthanized 7 days later. Macrophage depletion 6–8 week old mice were intravenously injected with either saline (control) or gadolinium chloride (GdCl3) (10 mg/kg, every 2 days for 1 week prior to dox treatment and then every 3rd day during dox treatment) and administered dox (2.0 or 0.2 mg/ml). Pancreata were harvested 7 days after dox. Only animals in which macrophages were reduced to less than 15% of the total DAPI-stained cells were used in the study. Physiological studies Adult mice were rendered diabetic with a single intraperitoneal injection of streptozotocin (180 mg per kg body weight dissolved in citrate buffer (pH 4.5)) after 4 hours fast. Mice with blood glucose levels >300 mg/dL were used for experiments. Glucose tolerance tests were performed by fasting animals overnight (16-hours) followed by an intraperitoneal injection of D-glucose (2 g per kg body weight). Blood glucose concentrations were measured using a BD Logic glucometer. RT-qPCR analysis cDNA was prepared from RNA using a high-capacity cDNA Archive Kit (Life Technologies) and amplified using real-time PCR with Power SYBR Green PCR master mix (Life Technologies) using gene specific primers (Table S6). Three experimental RNA replicates for each genotype were assayed. PCR was performed with an ABI 7900HT real-time PCR system (Life Technologies) and amplification data was analyzed using Sequence Detection System version 2.1 (Life Technologies) and Excel software (Microsoft). Hprt was used as an endogenous control for normalization and comparative Ct method was used to calculate relative fold expression by 2−Δ ΔCt. RNA-seq Analysis Three independent RNA isolates from each genotype were used for sequencing. RNA-sequencing methods were described previously (Choi et al., 2012). Single-end sequencing (110 bp) was performed on Illumina HiSeq2000 genome analyzer. Read alignment to the mouse genome (mm10) was performed using RNA-Seq Unified Mapper (RUM) (Grant et al., 2011). Genome alignment of sequencing data yielded 32–84 million uniquely mapped reads. Data was pre-processed with the PORT pipeline (https://github.com/itmat/Normalization; Kim et al. manuscript in preparation). Functional Annotation Clustering was performed using DAVID Bioinformatics Resources v.6.7 (Huang da et al., 2009). See Supplemental Information for detailed information on RNA isolation and preprocessing and differential expression analyses. Quantification of necrosis Necrosis in pancreatic acini was conducted as previously described (Liu et al., 2014; Yuan et al., 2012). Cell quantification Cells co-expressing specific genes was determined by manual counting. For each animal, over 500 cells were counted using ImageJ from five sections per animal that were separated by approximately 50 μM. All key experimental findings were observed in 3 or more animals. The total number of cells in the adult mouse pancreas was previously calculated (Dore et al., 1981) and used to determine the number of new β-cells produced. Statistical methods Statistical difference between two groups was assessed using student’s t-tests. *p < 0.05, **p < 0.001. All data represent mean ± SEM. Supplementary Material 1 2 3 4 5 6 These studies were supported by NIH grants DK72473 and DK89523 to MAM. We thank Haibo Jia for helping derive the Ptf1artTA allele, Lori Sussel, Ray MacDonald, Chris Wright, and Roland Stein for helpful suggestions or reagents, Jody Peters, Rama Gangula, and Laurel Grower for their skilled assistance with mouse husbandry, surgery, adenoviral construction, and genotyping, and Tiziana Sanavia for initial QC on the RNA-seq datasets. We thank the staff of the Vanderbilt Transgenic/ESC Shared Resource, Flow Cytometry Core, Translational Pathology Shared Resource, and VANTAGE for their help in deriving mice, staining slides, sorting cells, and performing RNA-Seq. The authors declare no conflict of interest. Figure 1 Design and validation of mouse alleles for the dox-dependent expression of 3TF in pancreatic acinar cells (A) Ptf1artTA and Rosa263TF.mCherry alleles were generated by recombinase-mediated cassette exchange. When interbred, the two alleles result in dox-inducible expression of MafA, Pdx1, Neurog3, and mCherry in a pancreatic acinar cell-specific manner. (B) mCherry expression was visible after administering 2.0 mg/ml of dox for 1 day (n=5). mCherry fluorescence was restricted to the pancreas (outlined) of dox treated mice and was not observed in other tissues. (C) Dox-inducible, acinar cell-specific expression of mCherry was confirmed with immunofluorescence analysis. Pancreas sections stained with insulin and mCherry showed that the two proteins were not co-localized. (D) Pancreas sections stained with antibodies against mCherry and either MafA, Pdx1, or Neurog3 showed co-expression of mCherry and 3TF after 1 day of dox. (E) Schematic of the transgene showing that the protein sequences for MAFA and NEUROG3 flank PDX1. To determine whether proper 2A mediated cleavage of 3TF was achieved, western blot against PDX1 was performed on pancreatic lysate from wild-type (WT) and 7 day induced, Ptf1artTA/+; Rosa263TF.mCherry/+ (3TF) mice. Arrows indicate both the 50 kDa PDX1 protein with the addition of the 2A peptide sequences and the endogenous PDX1 at 48 kDa. See also Figures S1 and Table S6. Figure 2 Temporal effects of 3TF overexpression in acinar cells (A) Animals were induced for 1, 7, or 28 days with dox. Immunofluorescence analysis revealed that amylase expression in mCherry+ cells decreased by 7 days and continued to decrease throughout the time course. Simultaneously, expression of chromogranin A and ghrelin began at 7 days and increased during the time-course. (B) Percentage of cells expressing amylase, chromogranin A and ghrelin among mCherry+ cells. Three mice per time point. Over 500 mCherry+ cells counted for each mouse. Data are represented as mean ± SEM. See also Figure S2 and Table S1–3. Figure 3 Transgene-mediated 3TF expression in pancreatic acinar cells causes a potent inflammatory response and ADM (A) Pancreatic immunofluorescence staining of CD45 from WT mice after 7 days of dox and 3TF mice after 1 and 7 days of dox. (B) Percentage of CD45+ cells among DAPI+ cells. Three mice per time point and over 1,000 DAPI+ cells counted for each mouse. Data are represented as mean ± SEM. (C) Representative pancreata and (D) hematoxylin and eosin (H&E) staining of WT and 3TF mice after 7 days of dox. Pancreas of 3TF mice is characterized by the presence of abundant tubular complexes (inset). (E) Pancreatic immunofluorescence staining of PanCK, a ductal marker, from WT mice after 7 days dox and 3TF mice after 1 and 7 days of dox. (F) Percentage of PanCK+ cells among mCherry+ cells. Three mice per time point and over 500 mCherry+ cells counted for each mouse. Data are represented as mean ± SEM. See also Figures S3–S5 and Table S4. Figure 4 Reducing 3TF expression attenuates inflammation and promotes A→β reprogramming (A) mCherry fluorescence in the pancreas of 3TF mice administered either 0.02, 0.2, or 2.0 mg/ml of dox for 7 days. (B) Quantification of fluorescence intensity per pancreas area. Data are represented as mean ± SEM. *p < 0.05, student’s t-test. (C) Pancreatic immunofluorescence CD45 staining of 3TF mice administered varying concentrations of dox at 7 days. (D) Percentage of CD45+ cells among DAPI+ cells. Three mice per time point and over 1,000 DAPI+ cells counted for each mouse. Data are represented as mean ± SEM. *p < 0.05, student’s t-test (E) Pancreatic immunofluorescence insulin staining of 3TF mice administered 0.2 mg/ml of dox for 7 days. mCherry+ cells that co-expressed insulin (arrows) were observed. (F) Percentage of insulin+ cells among mCherry+ cells. Three mice per time point and over 500 mCherry+ cells counted for each mouse. Data are represented as mean ± SEM. See also Figure S6. Figure 5 Macrophage depletion preserves pancreatic mass and architecture (A) Pancreatic immunofluorescence F4/80 staining of 3TF mice administered either saline or GdCl3 after 7 days dox. Arrows indicate F4/80+ cells. (B) Percentage of F4/80+ cells among DAPI+ cells at 7 days dox. Three mice per time point and over 1,000 DAPI+ cells counted for each mouse. Data are represented as mean ± SEM. **p < 0.001, student’s t-test. (C) Representative pancreata from 3TF mice given either saline or GdCl3 after 7 days dox. (D) Pancreatic weight per body weight of 3TF mice given either saline or GdCl3 and WT mice given GdCl3 after 7 days dox. Data are represented as mean ± SEM. **p < 0.001, student’s t-test. (E) Pancreatic immunofluorescence staining of PanCK of 3TF mice administered either saline or GdCl3 at 7 days dox. (F) Percentage of PanCK+ cells among mCherry+ cells at 7 days dox. Three mice per time point and over 500 mCherry+ cells counted for each mouse. Data are represented as mean ± SEM. **p < 0.001, student’s t-test. (G) Pancreatic PanCK staining of 3TF mice administered either saline or GdCl3 and WT mice administered GdCl3 at 7 days dox. Figure 6 Macrophage depletion promotes A→β reprogramming (A) Representative H&E staining of 3TF mice administered either saline or GdCl3 and WT mice administered GdCl3 at 7 days dox. Tubular complexes (inset) observed in pancreas of 3TF mice administered saline. (B) Representative Masson’s trichrome stain of 3TF mice administered either saline or GdCl3 and WT mice administered GdCl3 at 7 days dox. (C) Pancreatic immunofluorescence insulin staining of 3TF mice administered GdCl3 at 7 days dox. mCherry+ cells that co-expressed insulin were observed. (D) Percentage of Insulin+ cells among mCherry+ cells at 7 days of dox. Three mice per time point and over 1,000 mCherry+ cells counted. Data are represented as mean ± SEM. (E) RT-qPCR analysis of Ins1 and Ins2 expression in 3TF mice given either saline or GdCl3 after 7 days of dox. Fold change calculated against mRNA expression in uninduced acinar cells. Data are represented as mean ± SEM. **p < 0.001, student’s t-test. See also Figure S7 and Table S6. Figure 7 New β-like cells rescue STZ-induced diabetes (A) Mice were administered STZ on day 0 to induce diabetes and dox was administered for 7 days (day 6–13). Blood glucose was measured every other day. Statisical significance for the 3TF induced group was calculated against the 3TF control group, except were otherwise noted. (B) GTT was performed on day 13. ◆ black line: WT mice administered GdCl3 and dox but not STZ (WT control 1); ▲ black line: WT mice administered STZ, GdCl3, and dox (WT control 2); ■ red line: 3TF mice administered STZ and dox but not GdCl3 (3TF control); ● blue line: 3TF mice administered STZ, GdCl3, and dox (3TF induced). Of the five 3TF induced mice subjected to the GTT, two different patterns of response were observed: mice that were non-glucose responsive (▲ blue dotted line: 3TF induced subgroup 1) and mice that were glucose responsive (● blue dotted line: 3TF induced subgroup 2). (C) Model of divergent 3TF reprogramming. Adenoviral delivery of 3TF to the pancreatic acinar cells of Rag1−/− mice is relatively inefficient resulting in only a few pancreatic acinar cells expressing 3TF. Low levels of inflammation permit A→β reprogramming. Transgene expression of 3TF in the pancreatic acinar cells of Rag1+/+ mice is very efficient resulting in many acinar cells expressing 3TF and high levels of 3TF expression. Rapid reprogramming causes ER stress, a rise in [Ca2+]i, and cell death, triggering a potent inflammatory response that results in ADM, thereby blocking A→β reprogramming. See also Table S5. Accession number The accession number for the RNASeq data at ArrayExpress is E-MTAB-3921. Author Contributions Conceptualization, H.W.C., A.O., and M.A.M.; Methodology, H.W.C., A.O., J.D.S., W.Y., and E.M.; RNA-seq Analyses, E.M. and C.J.S.; Investigation, H.W.C., C.S., J.D.S., P.G.V., J.S.S., and W.Y.; Resources, G.G.; Writing-Original draft, H.W.C. and M.A.M.; Writing-Review & Editing, A.O., H.W.C, and M.A.M; Visualization, H.W.C., A.O., M.A.M, and C.J.S.; Project Administration, H.W.C. and M.A.M; Funding Acquisition, M.A.M. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. Arnes L Hill JT Gross S Magnuson MA Sussel L 2012 Ghrelin expression in the mouse pancreas defines a unique multipotent progenitor population PLoS One 7 e52026 23251675 Baeyens L Lemper M Leuckx G De Groef S Bonfanti P Stange G Shemer R Nord C Scheel DW Pan FC 2013 Transient cytokine treatment induces acinar cell reprogramming and regenerates functional beta cell mass in diabetic mice Nat Biotechnol Bockman DE Muller M Buchler M Friess H Beger HG 1997 Pathological changes in pancreatic ducts from patients with chronic pancreatitis Int J Pancreatol 21 119 126 9209953 Bramswig NC Everett LJ Schug J Dorrell C Liu C Luo Y Streeter PR Naji A Grompe M Kaestner KH 2013 Epigenomic plasticity enables human pancreatic alpha to beta cell reprogramming J Clin Invest 123 1275 1284 23434589 Burlison JS Long Q Fujitani Y Wright CV Magnuson MA 2008 Pdx-1 and Ptf1a concurrently determine fate specification of pancreatic multipotent progenitor cells Dev Biol 316 74 86 18294628 Carey BW Markoulaki S Hanna JH Faddah DA Buganim Y Kim J Ganz K Steine EJ Cassady JP Creyghton MP 2011 Reprogramming factor stoichiometry influences the epigenetic state and biological properties of induced pluripotent stem cells Cell Stem Cell 9 588 598 22136932 Cavelti-Weder C Li W Zumsteg A Stemann-Andersen M Zhang Y Yamada T Wang M Lu J Jermendy A Bee YM 2016 Hyperglycaemia attenuates in vivo reprogramming of pancreatic exocrine cells to beta cells in mice Diabetologia 59 522 532 26693711 Chen SX Osipovich AB Ustione A Potter LA Hipkens S Gangula R Yuan W Piston DW Magnuson MA 2011 Quantification of factors influencing fluorescent protein expression using RMCE to generate an allelic series in the ROSA26 locus in mice Dis Model Mech 4 537 547 21324933 Choi E Kraus MR Lemaire LA Yoshimoto M Vemula S Potter LA Manduchi E Stoeckert CJ Jr Grapin-Botton A Magnuson MA 2012 Dual lineage-specific expression of Sox17 during mouse embryogenesis Stem Cells 30 2297 2308 22865702 Dor Y Brown J Martinez OI Melton DA 2004 Adult pancreatic beta-cells are formed by self-duplication rather than stem-cell differentiation Nature 429 41 46 15129273 Dore BA Grogan WM Madge GE Webb SR 1981 Biphasic development of the postnatal mouse pancreas Biol Neonate 40 209 217 7032613 Grant GR Farkas MH Pizarro AD Lahens NF Schug J Brunk BP Stoeckert CJ Hogenesch JB Pierce EA 2011 Comparative analysis of RNA-Seq alignment algorithms and the RNA-Seq unified mapper (RUM) Bioinformatics 27 2518 2528 21775302 Gu G Dubauskaite J Melton DA 2002 Direct evidence for the pancreatic lineage: NGN3+ cells are islet progenitors and are distinct from duct progenitors Development 129 2447 2457 11973276 Halangk W Lerch MM Brandt-Nedelev B Roth W Ruthenbuerger M Reinheckel T Domschke W Lippert H Peters C Deussing J 2000 Role of cathepsin B in intracellular trypsinogen activation and the onset of acute pancreatitis J Clin Invest 106 773 781 10995788 Hara M Wang X Kawamura T Bindokas VP Dizon RF Alcoser SY Magnuson MA Bell GI 2003 Transgenic mice with green fluorescent protein-labeled pancreatic beta-cells Am J Physiol Endocrinol Metab 284 E177 183 12388130 Huang da W Sherman BT Lempicki RA 2009 Systematic and integrative analysis of large gene lists using DAVID bioinformatics resources Nat Protoc 4 44 57 19131956 Huang HP Liu M El-Hodiri HM Chu K Jamrich M Tsai MJ 2000 Regulation of the pancreatic islet-specific gene BETA2 (neuroD) by neurogenin 3 Mol Cell Biol 20 3292 3307 10757813 Jankov RP Luo X Belcastro R Copland I Frndova H Lye SJ Hoidal JR Post M Tanswell AK 2001 Gadolinium chloride inhibits pulmonary macrophage influx and prevents O(2)-induced pulmonary hypertension in the neonatal rat Pediatr Res 50 172 183 11477200 Ji B Chen XQ Misek DE Kuick R Hanash S Ernst S Najarian R Logsdon CD 2003 Pancreatic gene expression during the initiation of acute pancreatitis: identification of EGR-1 as a key regulator Physiol Genomics 14 59 72 12709512 Jura N Archer H Bar-Sagi D 2005 Chronic pancreatitis, pancreatic adenocarcinoma and the black box in-between Cell Res 15 72 77 15686632 Khoo C Yang J Weinrott SA Kaestner KH Naji A Schug J Stoffers DA 2012 Research resource: the pdx1 cistrome of pancreatic islets Mol Endocrinol 26 521 533 22322596 Lee J Sayed N Hunter A Au KF Wong WH Mocarski ES Pera RR Yakubov E Cooke JP 2012 Activation of innate immunity is required for efficient nuclear reprogramming Cell 151 547 558 23101625 Li J Zhou R Zhang J Li ZF 2014a Calcium signaling of pancreatic acinar cells in the pathogenesis of pancreatitis World J Gastroenterol 20 16146 16152 25473167 Li W Cavelti-Weder C Zhang Y Clement K Donovan S Gonzalez G Zhu J Stemann M Xu K Hashimoto T 2014b Long-term persistence and development of induced pancreatic beta cells generated by lineage conversion of acinar cells Nat Biotechnol 32 1223 1230 25402613 Li W Nakanishi M Zumsteg A Shear M Wright C Melton DA Zhou Q 2014c In vivo reprogramming of pancreatic acinar cells to three islet endocrine subtypes Elife 3 e01846 24714494 Liou GY Doppler H Necela B Krishna M Crawford HC Raimondo M Storz P 2013 Macrophage-secreted cytokines drive pancreatic acinar-to-ductal metaplasia through NF-kappaB and MMPs J Cell Biol 202 563 577 23918941 Liu Y Yuan J Tan T Jia W Lugea A Mareninova O Waldron RT Pandol SJ 2014 Genetic inhibition of protein kinase Cepsilon attenuates necrosis in experimental pancreatitis Am J Physiol Gastrointest Liver Physiol 307 G550 563 25035113 Logsdon CD Ji B 2013 The role of protein synthesis and digestive enzymes in acinar cell injury Nat Rev Gastroenterol Hepatol 10 362 370 23507798 Mombaerts P Iacomini J Johnson RS Herrup K Tonegawa S Papaioannou VE 1992 RAG-1-deficient mice have no mature B and T lymphocytes Cell 68 869 877 1547488 Neurath H Walsh KA 1976 Role of proteolytic enzymes in biological regulation (a review) Proc Natl Acad Sci U S A 73 3825 3832 1069267 Pan FC Bankaitis ED Boyer D Xu X Van de Casteele M Magnuson MA Heimberg H Wright CV 2013 Spatiotemporal patterns of multipotentiality in Ptf1a-expressing cells during pancreas organogenesis and injury-induced facultative restoration Development 140 751 764 23325761 Parsa I Longnecker DS Scarpelli DG Pour P Reddy JK Lefkowitz M 1985 Ductal metaplasia of human exocrine pancreas and its association with carcinoma Cancer Res 45 1285 1290 2982487 Puri S Folias AE Hebrok M 2015 Plasticity and dedifferentiation within the pancreas: development, homeostasis, and disease Cell Stem Cell 16 18 31 25465113 Rooman I Real FX 2012 Pancreatic ductal adenocarcinoma and acinar cells: a matter of differentiation and development? Gut 61 449 458 21730103 Sah RP Garg SK Dixit AK Dudeja V Dawra RK Saluja AK 2014 Endoplasmic reticulum stress is chronically activated in chronic pancreatitis J Biol Chem 289 27551 27561 25077966 Smith SB Gasa R Watada H Wang J Griffen SC German MS 2003 Neurogenin3 and hepatic nuclear factor 1 cooperate in activating pancreatic expression of Pax4 J Biol Chem 278 38254 38259 12837760 Strobel O Dor Y Alsina J Stirman A Lauwers G Trainor A Castillo CF Warshaw AL Thayer SP 2007 In vivo lineage tracing defines the role of acinar-to-ductal transdifferentiation in inflammatory ductal metaplasia Gastroenterology 133 1999 2009 18054571 Szilagyi L Kenesi E Katona G Kaslik G Juhasz G Graf L 2001 Comparative in vitro studies on native and recombinant human cationic trypsins. Cathepsin B is a possible pathological activator of trypsinogen in pancreatitis J Biol Chem 276 24574 24580 11312265 Thorel F Nepote V Avril I Kohno K Desgraz R Chera S Herrera PL 2010 Conversion of adult pancreatic alpha-cells to beta-cells after extreme beta-cell loss Nature 464 1149 1154 20364121 Tonge PD Corso AJ Monetti C Hussein SM Puri MC Michael IP Li M Lee DS Mar JC Cloonan N 2014 Divergent reprogramming routes lead to alternative stem-cell states Nature 516 192 197 25503232 Van Acker GJ Saluja AK Bhagat L Singh VP Song AM Steer ML 2002 Cathepsin B inhibition prevents trypsinogen activation and reduces pancreatitis severity Am J Physiol Gastrointest Liver Physiol 283 G794 800 12181196 Wang AY Ehrhardt A Xu H Kay MA 2007 Adenovirus transduction is required for the correction of diabetes using Pdx-1 or Neurogenin-3 in the liver Mol Ther 15 255 263 17235302 Wang RN Kloppel G Bouwens L 1995 Duct- to islet-cell differentiation and islet growth in the pancreas of duct-ligated adult rats Diabetologia 38 1405 1411 8786013 Watada H Kajimoto Y Umayahara Y Matsuoka T Kaneto H Fujitani Y Kamada T Kawamori R Yamasaki Y 1996 The human glucokinase gene beta-cell-type promoter: an essential role of insulin promoter factor 1/PDX-1 in its activation in HIT-T15 cells Diabetes 45 1478 1488 8866550 Watada H Scheel DW Leung J German MS 2003 Distinct gene expression programs function in progenitor and mature islet cells J Biol Chem 278 17130 17140 12604598 Whitcomb DC Gorry MC Preston RA Furey W Sossenheimer MJ Ulrich CD Martin SP Gates LK Jr Amann ST Toskes PP 1996 Hereditary pancreatitis is caused by a mutation in the cationic trypsinogen gene Nat Genet 14 141 145 8841182 Yuan J Liu Y Tan T Guha S Gukovsky I Gukovskaya A Pandol SJ 2012 Protein kinase d regulates cell death pathways in experimental pancreatitis Front Physiol 3 60 22470346 Zaldumbide A Carlotti F Goncalves MA Knaan-Shanzer S Cramer SJ Roep BO Wiertz EJ Hoeben RC 2012 Adenoviral vectors stimulate glucagon transcription in human mesenchymal stem cells expressing pancreatic transcription factors PLoS One 7 e48093 23110179 Zhou Q Brown J Kanarek A Rajagopal J Melton DA 2008 In vivo reprogramming of adult pancreatic exocrine cells to beta-cells Nature 455 627 632 18754011 Ziv O Glaser B Dor Y 2013 The plastic pancreas Dev Cell 26 3 7 23867225

PMC005xxxxxx/PMC5131527.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 8300331 2806 Cardiol Clin Cardiol Clin Cardiology clinics 0733-8651 1558-2264 27886793 5131527 10.1016/j.ccl.2016.08.010 NIHMS825276 Article Innovative Approaches to Hypertension Control in Low- and Middle-Income Countries Vedanthan Rajesh MD, MPH 1 Bernabe-Ortiz Antonio MD, MPH 2 Herasme Omarys I. MPH 1 Joshi Rohina MBBS, PhD, MPH 3 Lopez-Jaramillo Patricio MD, PhD 4 Thrift Amanda G. PhD 5 Webster Jacqui PhD 3 Webster Ruth PhD, MIPH 3 Yeates Karen MD, MPH 6 Gyamfi Joyce MS 7 Ieremia Merina PGDHS 8 Johnson Claire 3 Kamano Jemima H. MBChB, MMed 9 Lazo-Porras Maria MD 2 Limbani Felix MPH 10 Liu Peter MD 11 McCready Tara PhD, MBA 12 Miranda J. Jaime MD, PhD, MSc 2 Mohan Sailesh MD 13 Ogedegbe Olugbenga MD, MS, MPH 7 Oldenburg Brian PhD, MPsych 14 Ovbiagele Bruce MD, MSc 15 Owolabi Mayowa MBBS 16 Peiris David MBBS, PhD, MIPH 3 Ponce-Lucero Vilarmina 2 Praveen Devarsetty MBBS, MD, PhD 17 Pillay Arti PGDPH 18 Schwalm Jon-David MD, MSc 12 Tobe Sheldon W. MD, MScCH 19 Trieu Kathy MPH 3 Yusoff Khalid MBBS 20 Fuster Valentin MD, PhD 1 1 Icahn School of Medicine at Mount Sinai, New York, USA 2 CRONICAS Center of Excellence in Chronic Diseases, Universidad Peruana Cayetano Heredia, Lima, Peru 3 The George Institute for Global Health, University of Sydney, Sydney, Australia 4 Research Institute FOSCAL, Bucaramanga, Colombia 5 Monash University School of Clinical Sciences at Monash Health, Melbourne, Australia 6 Queens University School of Medicine, Ontario, Canada 7 New York University School of Medicine, New York, USA 8 Samoan Ministry of Health, Apia, Samoa 9 Moi University College of Health Sciences, Eldoret, Kenya 10 University of the Witwatersrand, Johannesburg, South Africa 11 University of Ottawa, Ontario, Canada 12 Population Health Research Institute, Hamilton, Canada 13 Public Health Foundation of India, New Delhi, India 14 University of Melbourne, School of Population and Global Health, Melbourne Australia 15 Medical University of South Carolina, Charleston, USA 16 University of Ibadan, Ibadan, Nigeria 17 The George Institute for Global Health, Hyderabad, India 18 Pacific Research Centre for the Prevention of Obesity and Non-Communicable Diseases at Fiji National University, Suva, Fiji 19 University of Toronto, Ontario, Canada 20 UCSI University, Kuala Lumpur, Malaysia Corresponding Author: Rajesh Vedanthan, MD, MPH, Icahn School of Medicine at Mount Sinai, One Gustave L. Levy Place, Box 1030, New York, NY 10029, rajesh.vedanthan@mssm.edu Authors Contact Information: Antonio Bernabe-Ortiz, CRONICAS, Center of Excellence in Chronic Diseases, Universidad Peruana Cayetano Heredia, Lima, Peru, antonio.bernabe@upch.pe Omarys I. Herasme, Icahn School of Medicine, Mount Sinai Hospital, New York, USA, omarys.herasme@mssm.edu Rohina Joshi, The George Institute for Global Health, University of Sydney, Sydney, Australia, rjoshi@georgeinstitute.org.au Patricio Lopez-Jaramillo, Research Institute FOSCAL, Bucaramanga, Colombia, jplopezj@gmail.com Amanda G. Thrift, Monash University, School of Clinical Sciences at Monash, Health, Melbourne, Australia, amanda.thrift@monash.edu Jacqui Webster, The George Institute for Global Health, University of Sydney, Sydney, Australia, jwebster@georgeinstitute.org.au Ruth Webster, The George Institute for Global Health, University of Sydney, Sydney, Australia, rwebster@georgeinstitute.org.au Karen Yeates, Queens University, School of Medicine, Ontario, Canada, yeatesk@queensu.ca Joyce Gyamfi, New York University, School of Medicine, New York, USA, joyce.gyamfi@nyumc.org Merina Ieremia, Samoan Ministry of Health, Apia, Samoa, merinaI@health.gov.ws Claire Johnson, The George Institute for Global Health, University of Sydney, Sydney, Australia, cjohnson@georgeinstitute.org.au Jemima H. Kamano, Moi University, School of Medicine, Eldoret, Kenya, shoine.hoine@gmail.com Maria Lazo-Porras, CRONICAS, Center of Excellence in Chronic Diseases, Universidad Peruana Cayetano Heredia, Lima, Peru, maria.lazo@upch.pe Felix Limbani, Centre for Health Policy, School of Public Health, University of the Witwatersrand, Johannesburg, South Africa, felix.limbani@wits.ac.za Peter Liu, University of Ottawa, Ontario, Canada, peter.liu@utoronto.ca Tara McCready, Population Health Research Institute, Hamilton, Canada, tara.mcCready@phri.ca J. Jaime Miranda, CRONICAS, Center of Excellence in Chronic Diseases, Universidad Peruana Cayetano Heredia, Lima, Peru, jaime.miranda@upch.pe Sailesh Mohan, Public Health Foundation of India, New Delhi, India, smohan@phfi.org Brian Oldenburg, University of Melbourne, School of Population and Global Health, Melbourne Australia, brian.oldenburg@unimelb.edu.au Olugbenga Ogedegbe, New York University, School of Medicine, New York, USA, olugbenga.ogedegbe@nyumc.org Bruce Ovbiagele, Medical University of South Carolina, Charleston, USA, ovibes@musc.edu Mayowa Owolabi, University of Ibadan, Ibadan, Nigeria, mayowaowolabi@yahoo.com David Peiris, The George Institute for Global Health, University of Sydney, Sydney, Australia, dpeiris@georgeinstitute.org Vilarmina Ponce-Lucero, CRONICAS, Center of Excellence in Chronic Diseases, Universidad Peruana Cayetano Heredia, Lima, Peru, vilarmina.ponce.l@upch.pe Devarsetty Praveen, The George Institute for Global Health, Hyderabad, India, dpraveen@georgeinstitute.org.in Arti Pillay, Fiji National University, Suva, Fiji, arti.pillay@fnu.ac.fj Jon-David Schwalm, Population Health Research Institute, Hamilton, Canada, schwalj@mcmaster.ca Sheldon W. Tobe, University of Toronto, Ontario, Canada, sheldon.tobe@sunnybrook.ca Kathy Trieu, The George Institute for Global Health, University of Sydney, Sydney, Australia, ktrieu@georgeinstitute.org.au Khalid Yusoff, UCSI University, Kuala Lumpur, Malaysia, khalid@ucsiuniversity.edu.my Valentin Fuster, Icahn School of Medicine, Mount Sinai Hospital, New York, USA, valentin.fuster@mssm.edu 28 10 2016 2 2017 01 2 2018 35 1 99115 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. Elevated blood pressure, a major risk factor for ischemic heart disease, heart failure, and stroke, is the leading global risk for mortality. Despite global efforts to combat hypertension, it continues to exert a significant health and economic burden on low- and middle-income country (LMIC) populations, thereby triggering the need to address the problem by way of novel approaches. The Global Alliance for Chronic Diseases has funded 15 research projects related to hypertension control in low-resource settings worldwide. These research projects have developed and evaluated several important innovative approaches to hypertension control, including: community engagement, salt reduction, salt substitution, task redistribution, mHealth, and fixed-dose combination therapies. In this paper, we briefly review the rationale for each of these innovative approaches, as well as summarize the experience of some of the research teams in these respective areas. Where relevant, we also draw upon the wider literature to illustrate how these approaches to hypertension control are being implemented in LMICs. The studies outlined in this report demonstrate innovative and practical methods of implementing for improving hypertension control in diverse environments and contexts worldwide. Hypertension Low- and middle-income countries Community engagement mHealth Task redistribution Salt reduction Salt substitution Polypill Introduction Cardiovascular disease (CVD) is the leading cause of mortality in the world, resulting in 17.3 million deaths annually, with 80% of these deaths occurring in low- and middle-income countries (LMICs).1 Elevated blood pressure, a major risk factor for ischemic heart disease, heart failure, and stroke,2 is the leading global risk for mortality.1 Despite global efforts to combat hypertension, treatment and control rates are very low in LMICs.3 Given the continued significant health and economic burden on LMIC populations, there is an urgent need to address the problem by way of novel approaches. Founded in 2009, the Global Alliance for Chronic Diseases (GACD), funds, coordinates, and facilitates global collaborations in implementation research, focusing on the prevention and treatment of chronic non-communicable diseases in LMICs and vulnerable populations in high-income countries.4 The first round of GACD-sponsored research projects focused on hypertension, and included 15 research teams from around the world.5 These research projects have involved the development and evaluation of several important innovative approaches to hypertension control, including: community engagement, salt reduction, salt substitution, task redistribution, mHealth, and fixed-dose combination therapies. In this paper, we briefly review the rationale for each of these innovative approaches, as well as summarize the experience of some of the GACD teams in these respective areas. Where relevant, we also draw upon the wider literature to illustrate how these approaches to hypertension control are being implemented in LMICs. Community Engagement Health care delivery and health systems often fail to meet the needs and expectations of those who need them.6, 7 Community engagement seeks to address this problem by optimizing the appropriateness and alignment of health care to the cultural, social, economic, and environmental setting.8, 9 It encompasses participation, mobilization, and empowerment (Figure 1).10 Participation refers to the active or passive engagement of the community in health services.10, 11 Mobilization furthers this engagement through facilitation by health professionals, while empowerment involves a capacity-building process to engage communities in planning, implementing and/or evaluating activities to achieve more sustainable health improvements.10, 11 Community engagement has shown promise in supporting interventions to improve health outcomes related to both HIV/AIDS as well as maternal and child health.12, 13 However, traditional methods for determining efficacy of community engagement are inadequate because there are significant challenges in teasing out the independent effects of the intervention vis a vis the process of community engagement itself. Four GACD projects described herein have been conducted in Tanzania, Kenya, Colombia, Malaysia, India, and Canada. The investigators of these GACD projects have adopted a diverse range of community engagement activities, targeted at both individuals and systems, in order to identify barriers and facilitators for the care of hypertension, and thereby tailor the intervention to the local context (Table 1). Prior to initiating each of these studies, investigators and research staff met with community leaders, health personnel, and other relevant community stakeholders, to facilitate entry to the communities and to appropriately contextualize their approaches. Components of community engagement included (1) individual interviews with diverse stakeholders; (2) focus group discussions (FGDs) with hypertension patients; (3) workshops with local community health workers (CHWs) and clinicians to refine the intervention and training materials, thus enhancing the capacity of CHWs to deliver the intervention by employing relevant and easy-to-use tools; (4) community social events and gatherings; and (5) mabaraza (singular baraza), traditional East African community gatherings, conducted among individuals with elevated blood pressure and CHWs to complement the purposive sampling inherent in FGDs.14 The baraza is a unique and novel qualitative research setting which has been used as a form of participatory action research, and allows organization of a diverse and heterogeneous large group of individuals.15 In Tanzania and Canada, the team used an adapted tool called I-RREACH: Intervention and Research Readiness Engagement and Assessment of Community Health Care.16 This tool was developed using a community-based consensus method, and is rooted in participatory principles, equalizing the importance of the knowledge and perspectives of researchers and community stakeholders while encouraging respectful dialogue. The I-RREACH tool is an engagement and assessment tool for improving the implementation readiness of researchers, organizations and communities in complex interventions, and consists of three phases: fact finding, stakeholder dialogue, and community member/patient dialogue. Another study being conducted in Canada, Malaysia, and Colombia leveraged non-medical community events for the purposes of screening, recruitment, intervention implementation, and follow-up. Using process evaluation, the GACD projects hope to add to our understanding of how community engagement can be used to support and strengthen programs aimed at improving hypertension control. Such an approach can be applied to more chronic diseases in low-resource settings worldwide. The need for this research is illustrated by work elsewhere. Although it may seem self-evident that a more participatory approach will improve the acceptability, and thus effectiveness of interventions, this is not fully supported by the evidence. Two projects conducted in Cape Town, South Africa, and El Paso, Texas, used community-based participatory research approaches to design an intervention to manage hypertension and diabetes.17, 18 Positive results included: 1) improved self-efficacy to manage hypertension, 2) greater improvements in health behaviours in the intervention group than in the control group,18 3) the development of culturally appropriate health education materials specifically developed for low-literacy populations,18 and 4) inclusion of learnings into local health sector planning for prevention and control of hypertension and diabetes.17 Although the authors stated that the materials were well received by participants in one study,18 no evidence for clinical success of community engagement was provided in either study.17, 18 Salt Reduction Evidence shows that a reduction in the consumption of sodium—found in table salt and naturally occurring foods such as milk, eggs, meat, and shellfish—decreases blood pressure in adults and diminishes the risk of CVD.19, 20 While there is controversy about the most appropriate target for sodium intake, higher sodium intake in general is associated with poorer outcomes.21 The World Health Organization (WHO) recommends a reduction in sodium intake to < 2 g/day in adults.22 In 2013, member states of the United Nations established a target to reduce the average population salt intake by 30% by 2025,23 and 75 countries now have strategies in place to achieve this target.24 The majority of these national programs are multifaceted and include initiatives such as industry engagement to lower salt content in foods, consumer education and awareness, establishing front-of-pack labelling schemes and nutrition standards for foods procured in public settings. Three of the GACD Hypertension programs have implemented innovative salt reduction programs to reduce blood pressure. The first step in any program is to measure existing consumption patterns. These projects measured salt intake using 24-hour urine excretion and tried to understand people’s knowledge and eating behaviors through community surveys. Average daily salt excretion at baseline varied from 7 gm (Samoa);25 11 gm (Fiji); 9.5 gm and 8.6 gm (Andhra Pradesh and Delhi/Haryana, respectively, India); to 12.6 gm (Shanxi, China). The information on diet was then used to inform the different intervention strategies. Based on the WHO’s framework for Creating an Enabling Environment for Salt reduction,26 the project in Fiji and Samoa used multi-faceted intervention programs to reduce salt in the food supply, while concurrently implementing media and community mobilization campaigns to increase awareness (Figure 2).27 A parallel project in Andhra Pradesh and Delhi/Haryana, India, used community surveys and stakeholder mapping and established a comprehensive food composition database (based on the George Institute’s leading FoodSwitch innovation for monitoring the food supply and identifying healthy choices).28 This information is being used to inform the development of a government-led salt reduction strategy for India. The Little Emperor project in China trained children to encourage their parents to reduce salt intake. Implemented in the northern province of Shanxi, the researchers taught the children about the harmful effects of a salty diet and asked them to share the messages with adults back home. Innovative children’s approaches including hiding the salt pot, making up rhymes or using their status as “Little Emperors” to refuse to eat unhealthy foods, led to a 26% reduction in participants’ salt intake in less than 4 months.29 More than 270 million people currently have hypertension in China; therefore, if applied nationally, such a strategy could have substantial health and potential economic benefits. Post-intervention monitoring in Fiji and Samoa is being finalized and has been supplemented through an in-depth process evaluation to better understand how the interventions have been implemented and potential barriers to effectiveness. Some of the challenges have included the changing political environment, difficulties of multi-sectoral action and limited experience in engaging the food industry. Mainstreaming the agendas with the Health Ministries in the different countries has been key to overcoming some of these problems. The lessons are being documented and will be disseminated widely through the WHO Collaborating Centre for Population Salt Reduction at the George Institute for Global Health, thus supporting rapid and effective translation of research into policy and practice. These and other studies will help to elucidate and clarify the relationship between sodium reduction and CVD. Salt substitution In addition to salt reduction, salt substitution is an innovative, non-pharmacological approach to reduce blood pressure. It involves the partial replacement of sodium chloride with any combination of other salt containing potassium, magnesium, or aluminum. A meta-analysis from six randomized controlled trials using different combinations of salt substitute in comparison to usual salt found, in pooled results, that a salt substitute reduced systolic blood pressure by −4.9 mm Hg (95% CI: −7.3, −2.5) and diastolic blood pressure −1.5 mm Hg (95% CI: −2.7, −0.3). However, in the subgroup analysis, the effect was significant only among individuals with hypertension.30 One of the GACD projects, conducted in Peru,31 is using a population-wide approach to test the effect on blood pressure of replacing regular salt by an iodine-fortified substitute containing 25% potassium chloride and 75% sodium chloride. This involves a pragmatic stepped wedge trial design, in which the intervention is progressively implemented at random in six villages. The study has been implemented in two phases (Figure 3). The first phase was exploratory and included: (a) formative in-depth interviews and FGDs; (b) a triangle taste test, which found that a salt with 25% of potassium chloride was indistinguishable from regular salt;32 and (c) the development of the social marketing campaign targeting primarily women responsible for cooking at their home, and focused on promoting consumption and adherence of participants to the potassium-enriched salt. The second phase involved implementation of the intervention. The salt substitute has progressively replaced the common salt used in households, relying heavily on the social marketing/branding campaign as well as educational entertainment delivered by trained community health workers. Salt replacement has been implemented at households, bakeries, community kitchens and restaurants in each village. Previous salt substitute strategies have focused on delivering the salt substitute product among participants with a diagnosis of hypertension, focusing almost exclusively on the hypertension status of the participant rather than on the product’s concept. For instance, the salt substitute used in other studies were no different between intervention and control arms (i.e. bags were identical in appearance; products were manufactured, packaged, and labeled by the same company).33–35 The novelty of the Peru study relies on the implementation mechanisms that were developed and put in place, at the community level, aiming to increase the uptake of the salt substitute product as well as ensuring its sustained used over time in populations irrespective of hypertension status. To date, acceptability of the salt substitute to participants has been successful with very low rates of adverse effects related to its use. The study is ongoing and the fourth wedge has been concluded, with expected outcomes in early 2017. If successful, this project’s implementation approach may serve as a model for other LMIC settings. Task Redistribution In most countries, primary care physicians are the main providers of healthcare for individuals with CVD. Unfortunately, most LMICs have an inadequate number of physicians, especially in rural and remote regions where a majority of the population reside.36, 37 According to the WHO Global Health Observatory, there are 0.3 physicians available for every 1000 population in low income countries, 1.2 physicians per 1000 population in lower-middle income countries, and 2.0 per 1000 population in upper-middle income countries.38 In response to this physician workforce shortage, appropriate strategies for task redistribution—from doctors to a team consisting of doctors and trained non-physician health workers (NPHWs)—have been developed and implemented, especially in the areas of maternal and child health needs,39, 40 and HIV/AIDS.41 Task redistribution describes a situation where a task normally performed by a physician is shared between physicians and other health workers with a different or lower level of education and training (Figure 4).42 Task redistribution may be aided by technology, clear guidelines, or close supervision by physicians, to help standardize the performance and interpretation of certain tasks, therefore allowing them to be performed by NPHWs.43 Systematic reviews on task-redistribution for CVD management,44, 45 indicate that not many studies have been conducted to test the effectiveness of task redistribution, and that further operational research, including detailed process evaluation, is required to understand the complexity, effectiveness, and cost-effectiveness of task-redistribution within different country contexts. Recent studies involving task-redistribution have shown that NPHWs can be effectively trained in the implementation of CVD prevention and management guidelines,46, 47 successfully screen individuals at high-risk of CVD,48, 49 provide lifestyle education and adherence to patients,50 and support patients with acute coronary syndrome.51 This approach has also been shown to be cost-effective for chronic disease care in the LMIC context.52, 53 While there are now some published studies concerning the effectiveness of task-redistribution, there remain large evidence gaps and obstacles regarding the translation of positive research findings into routine health care delivery in LMICs, while also ensuring quality of care, safety, and patient acceptability. These shortcomings notwithstanding, task redistribution for the prevention and control of hypertension and other chronic diseases presents a great opportunity that could increase access to care, reduce health care costs, free up physician time for other tasks, and increase system efficiency in the long-term. Eight of the GACD projects included a component of task redistribution for the detection and management of hypertension. These include the redistribution of tasks related to hypertension screening, referral to clinicians, providing lifestyle advice, and support for adherence to medications to NPHWs. All the studies supported NPHWs by training them for two to six days, followed by re-training where required.14, 54, 55 Some studies facilitated task redistribution by using mHealth technology,14, 56 whereby NPHWs used electronic decision support tools to screen individuals in the community and link them to hypertension care. Process and interim evaluations have identified that the main barriers to task-redistribution include resistance from other health professionals; increasing NPHW workload due to additional tasks; complexity of training materials; health system-related issues such as non-availability or non-functioning BP machines, poor drug supply, lack of physician availability for referral; regulatory restrictions including the inability to prescribe medications; and low remuneration of NPHWs.57 The key enablers included an increase in the enthusiasm and motivation of NPHWs to be trained and take on new roles, as well as a reduction in the physician workload leading to improved performance. All of these studies are currently in progress and will have effectiveness and cost-effectiveness results in the near future. mHealth mHealth is the use of mobile phones to improve and support health, and can be used for a variety of purposes to connect clinicians, other health workers including CHWs, and patients or patient caregivers (Figure 6). mHealth can be used to provide health education, promote behavior change, facilitate decision support in diagnosis and management of a wide variety of conditions, support diagnostic testing, or link medical records.58 Evidence for benefits of mHealth is widespread among a variety of high-income country settings, and further data are emerging on the use of mHealth in LMICs with respect to the impact on clinical outcomes, processes of care, health care costs and health related quality of life.59–61 There is great potential for the use of mHealth for hypertension management in LMICs as mobile phone ownership is high and growing rapidly, even among the poor.62 However, there still remain research gaps with a relatively limited number of studies in this area, particularly in hypertension. Five projects within the GACD research network have a mHealth component at their core, or in conjunction with other innovations, in order to address barriers within health systems and to optimize opportunities for the detection and management of hypertension. The projects are taking place in communities in rural Kenya,14 rural Tanzania, both urban and rural Colombia and Malaysia, rural and remote Aboriginal communities in Canada, Nigeria,63 and rural India. All of the projects are utilizing either a smartphone- or tablet-based tool designed for use by community health workers (CHWs) to improve hypertension care; facilitate improved identification, follow-up, and tracking of patients; promote adherence to medications; and improve education of patients and CHWs. All of the programs have a component of real-time decision support. In addition, the Nigerian and Tanzanian/Canada program also send educational, behaviour change communication messages via text message directly to patients’ mobile phones, while the India project uses interactive voice response messaging because text messaging was not acceptable in this setting. The Kenya and India projects embed educational messaging in novel audio-visual formats, so that CHWs can share these audio-visual materials to patients during home visits. Some unique features among the mHealth innovations and programs should be highlighted. The programs in Kenya and India use an open-source platform called Open Data Kit, which has been utilized successfully to provide clinical decision source tools for HIV care. The Indian program also provides access to a mobile device that allows primary care physicians in government health clinics to access the health information of participants screened by CHWs; the device offers decision support for those participants identified at high CVD risk. This feature is also a component of the Tanzania-Canada program, whereby health center nurses and clinical officers can access all BP measurements taken for a patient by CHWs. A substudy of the Tanzania project is also evaluating the effectiveness of a phone-based drug voucher program to ensure the authenticity of drug supply and adherence factors in hypertension control. The Nigerian program is targeting patients who have experienced a stroke, who are at high risk for another stroke. Across the programs there have been common challenges, which include both technical and human factors. Technical factors have included mobile network coverage and server issues. Human factors have included overcoming end-user challenges with the new technology, as well as implementation delays due to government approval processes, equipment procurement delays, misalignment of incentives, competing obligations, and excessive workload for the health providers who are utilizing these new systems. Polypill – Fixed-dose Combination Therapy Most patients with hypertension generally require blood pressure (BP) lowering medication from multiple classes to achieve adequate control.64 The need for titration of medication and addition of multiple classes of drug requires multiple physician visits and this in itself can lead to poor adherence to prescribed medication and poor attendance at scheduled visits.65 The requirement to take multiple medications in complex regimes also encourages poor adherence.66 For physicians, the need for repeated up-titrating or adding extra medications can lead to inertia and tacit acceptance of inadequate BP control.67, 68 Initiating anti-hypertensive treatment with dual combination therapy not only accelerates the time taken to achieve control but also attains a lower final target.69, 70 For the patient, improved adherence has also been demonstrated without worsening the side effect profile.71 Further benefits in BP control can also result from simplifying up-titration regimes.70 Use of multi-modal fixed-dose combination pills (FDCs)—also known as ‘polypills’–containing not only multiple low-dose blood pressure-lowering drugs, but also statins and aspirin, has the potential to reduce a person’s cardiovascular risk beyond that achieved by simply lowering their blood pressure, by addressing multiple risk factors concurrently in a single pill. Multiple large clinical trials have shown that use of ‘polypills’ in patients at high risk of CVD improves adherence to long-term medication with consequent improvements in cholesterol and blood pressure measurements, and are highly acceptable to patients and physicians alike (Figure 5).72, 73 The recently published HOPE-3 study utilized a polypill type strategy in patients at moderate CVD risk and showed a significant reduction in CV events in patients with hypertension.74 While reducing BP was a benefit only in those in the hypertensive range, lowering cholesterol had beneficial effects in reducing fatal and non-fatal cardiovascular events overall.75 Evidence is needed, however, on the implementation of such a strategy in real-life clinical contexts rather within the constraints of a highly regulated clinical trial. The GACD has funded two projects looking at whether use of FDCs will improve management of hypertension, and also overall CVD risk, in real-life clinical contexts in several LMICs. The TRIple Pill vs. Usual care Management for Patients with mild-to- moderate Hypertension (TRIUMPH) study,76 is a prospective, open, randomized controlled clinical trial (n=700) of a fixed dose combination blood pressure-lowering pill (“Triple Pill”)-based strategy compared to usual care among individuals with persistent mild-to-moderate hypertension on no or minimal drug therapy. The aim is to see whether early use of low dose FDC medications will result in faster and better control of blood pressure. The HOPE-4 study, being conducted in 50 urban and rural communities in Canada, Colombia, and Malaysia, is implementing and evaluating (compared to usual care) an evidence-based, contextually-appropriate program for CVD risk assessment, treatment and control involving simplified algorithms implemented by NPHWs, supported by e-health technologies; initiation of FDC of two antihypertensive drugs plus one statin; and use of treatment supporters to optimize long-term medication and lifestyle adherence. Both studies are ongoing with outcomes anticipated in the near future. The use of a simplified strategy utilizing early introduction of inexpensive generic FDC pills (or ‘polypills’) is an approach with important potential to impact on what are currently exceedingly poor blood pressure control rates in LMICs. If found to be effective, cost-effective, and acceptable, this approach has the potential to impact the cardiovascular health of significant numbers of individuals around the world. Discussion Elevated blood pressure is the leading global risk for mortality,1 and novel approaches for improving hypertension control are urgently required for LMICs. The GACD hypertension studies described here are beginning to disseminate outcomes, results, and lessons in relation to several different innovative approaches. In addition, they are well-poised to develop post-study knowledge translation strategies. Finally, the GACD researchers have the potential to engage policy makers, payers, and other stakeholders, to translate the findings of individual research studies into sustainable and scalable interventions. Each GACD-funded project has designed one or more innovative approaches to enable the implementation and evaluation of interventions within local contexts, in order to improve care without significant disruption to, and increased workload of, already over-burdened health workers and health care systems. All of the approaches described here have the potential to improve the cardiovascular health of populations in low-resource settings worldwide. Community engagement is a critical part of developing and introducing any new program, and increases the likelihood of successful uptake and implementation. Salt reduction and salt substitutes can reduce blood pressure and improve cardiovascular health, especially if combined with improved dietary intake of fresh fruits and vegetables. Task redistribution expands the reach of delegated medical acts, empowers and engages community members, improves health literacy of communities, and improves the efficiency of the existing pool of health care providers. mHealth can additionally provide decision support, remote medical record access, and novel educational interfaces, all of which can enhance care delivery in resource-limited settings. Finally, FDC pills have the potential to transform the landscape of medical management of hypertension and CVD. The studies outlined in this report demonstrate innovative and practical methods of implementing all of these strategies for hypertension control in diverse environments and contexts worldwide. The writing group would like to thank Gary Parker from the GACD Secretariat for invaluable logistical and administrative support, and Drs. Clara Chow, Pallab Maulik, and Martin McKee for critical review of the manuscript. They would also like to thank all members of the GACD Hypertension Research Program for their support and input throughout the preparation of this manuscript. Funding for the studies described and for manuscript submission was provided by the following GACD Hypertension Program funding agencies: Canadian Institutes of Health Research (Grant No. 120389); Grand Challenges Canada (Grant Nos. 0069-04, and 0070-04); International Development Research Centre; Canadian Stroke Network; Australian National Health and Medical Research Council (Grant Nos. ID 1040147, and 104018); the US National Institutes of Health (National Heart, Lung and Blood Institute and National Institute of Neurological Disorders and Stroke) (Grant Nos. U01 HL114200, U01 NS079179, and U01 HL114180); the United Kingdom Medical Research Council (Grant Nos. APP 1040179, APP 1041052, and J01 60201); the Malaysian Ministry of Higher Education (Long-term Research Grants Scheme); and the South African Medical Research Council. This report does not represent the official view of the National Institute of Neurological Disorders and Stroke, the National Institutes of Health, or any part of the US Federal Government. No official support or endorsement of this article by the National Institutes of Health is intended or should be inferred. Abbreviations BP Blood Pressure CHW Community Health Workers CVD Cardiovascular Disease FDC Fixed-Dose Combination Pills FGD Focus Group Discussions GACD Global Alliance for Chronic Diseases I-RREACH Intervention and Research Readiness Engagement and Assessment of Community Health Care LMIC Low- and Middle-Income Countries NPHW Non-physician Health Workers TRIUMPH TRIple Pill vs. Usual care Management for Patients with mild-to- moderate Hypertension WHO World Health Organization Figure 1 Community engagement activities undertaken within GACD Projects. Participation activities denote the least level of engagement while empowerment activities denote the greatest level of engagement. Figure 2 Framework for salt reduction strategies, including context, activities, outputs, and anticipated outcomes. Figure 3 Launching a salt substitute to reduce blood pressure at the population level in Peru, divided into two phases. Figure 4 The process of task-redistribution for the management of hypertension adapted from the WHO’s recommendations on task-shifting. From World Health Organization, PEPFAR, UNAIDS. Task shifting: rational redistribution of tasks among health workforce teams : global recommendations and guidelines. http://www.who.int/healthsystems/TTR-TaskShifting.pdf, 2016, with permission. Figure 5 Proportion of participants adherent to combination therapy at end of study in patients either with established CVD or at high calculated risk. Adherence is defined as taking antiplatelet, statin and ≥ 2 BP-lowering drugs at least 4 days of the last 7 at end of study in UMPIRE,78 Kanyini-GAP79 and IMPACT.80 Adherence in the FOCUS73 trial was defined as pill count between 80 and 110% at end of study plus a score of 20/20 on the Morisky-Green questionnaire. Data from references 73, 78, 79, 80. Figure 6 Schematic illustrating the potential for mHealth to connect clinicians, community health workers (CHWs), and patients. Blue arrows indicate direct interactions among individuals. Red arrows indicate interactions that are facilitated by mHealth. Table 1 Type and target group of community engagement activities undertaken within GACD projects, including timing of engagement and materials developed through each activity. Region Type Target Group Timing of Engagement Rationale for Activity Materials Developed INDIA Community Entry Community leaders Prior to the initiation of study activities within each cluster/community unit To gain entry into the community Protocol, specific aims, abstract Survey of community members Individuals Once at study initiation Length: 60–90 minutes To identify barriers to seeking health care and/or treatment Survey Community Focus Group Discussions Individuals with hypertension Up to 12 focus groups, each comprising up to 10 people Length: 60–90 minutes To identify barriers to seeking health care and/or treatment Structured guide for discussions In-depth interviews Health care providers 23 interviews with doctors, nurses, and CHWs To identify barriers to providing health care and/or treatment Structured guide for interviews Survey of medicines Public, private and other medicine outlets 20 public outlets 16 private outlets 2 other outlets selling medicines at subsidized rates to all patients To determine availability, affordability and acceptability of medications Structured list of essential medicines for audit Consultation via a planning day Local, and state government health officials, and local experts Once at a 4-hour planning session To ensure that the design of the intervention fit into the health system Final design of intervention Working group testing of intervention materials CHWs and local doctors Over 5 days, CHWs and doctors participated in a pilot training program To develop educational materials for training CHWs and to educate people with hypertension Educational materials for training CHWs and for people with hypertension Training CHWs 5 full days of training delivered by doctors and researchers To provide skills to CHWs to enable them to conduct a peer support group and educate people with hypertension Education materials for CHWs Community-based support group of people with hypertension Letter of support and encouragement from head of village (Sarpanch) 3-month intervention comprising 6 fortnightly education sessions delivered by CHWs, locally sourced expert advisers, health care providers, and researchers Self-management and education support group of people with hypertension Education materials for people with hypertension, including handouts Dissemination of study results Communities, local health providers, medicines outlets, Ministry of Health & Welfare, National health Mission At end of study To build capacity and sustainability Development of resources for use by heath care providers for assessing and treating hypertension KENYA Community Entry Community leaders, health personnel, community stakeholders Prior to the initiation of study activities within each cluster/community unit To gain entry into the community Protocol, specific aims, abstract, and PowerPoint summary Community Gatherings (Mabaraza) Community Six in total (until content saturation achieved) Length: 1–2 hours To identify the barriers and facilitators to linkage to care for hypertension and retention to care Structured discussion guides for mabaraza Focus Group Discussions Individuals with hypertension and CHWs 17 total (until content saturation achieved) Length: 1–2 hours To identify barriers to seeking and delivering health care and/or treatment Moderator guides Human-Centered Design Design team with diverse stakeholders; content validity testing with diverse stakeholders Occurrence: Approximately 10 design team meetings; nine content validity focus group discussions with patients, community health workers, and clinicians Length: 60 min Design of behavioral assessment and tailored communication strategy Final design of intervention TANZANIA & CANADA Community Entry Community leaders/stakeholders Prior to the initiation of study activities within each of the 2 selected communities To gain entry into the community and gauge interest Framework for development of the I-RREACH Tool Completion of 3 ‘consensus’ cycles Stakeholders and community-based researchers in Canada and Tanzania At project initiation moving forward over a 1 year period in 3 cycles To test theoretical frameworks regarding researcher’s practice-based knowledge, community readiness, Indigenous approaches to research, empowerment approaches Development of the I-RREACH Tool (insert Ref) Community Focus Group Discussions Individuals with hypertension and their families as well as local health care providers 3 focus groups were held in participating Indigenous communities in Canada and 1 in Tanzania of varying length with a total of 45 participants To identify major factors that may impact on the effectiveness of evidence-based educational SMS messages for people with hypertension and reduce health inequalities Content from focus groups informed the development of the SMS messages to be used for the intervention in each country Training CHWs and local health providers In the second year, CHWs and doctors participated in country specific training programs on hypertension and cardiovascular disease as well as use of the mHealth tools/equipment. In Tanzania there was also a pre-post evaluation of knowledge gained and an observed standardized clinical exam To prepare CHWs and health providers to provide educational support to their communities (people with hypertension and their families) Educational materials for training CHWs and health providers Training Local health providers (Tanzania only) 5 full days of training delivered by doctors and researchers in year 3 to evaluate the appropriateness of the treatment algorithm for management of hypertension (adapted from the existing Tanzanian hypertension guideline) To provide skills to health providers to enable them to manage hypertension effectively Treatment algorithm for hypertension that is specific to low-resource rural setting in sub-Saharan Africa Dissemination of study results Communities, local health providers, medicines outlets, Ministry of Health & Social Welfare, National health Mission Will occur at end of study To build capacity and sustainability Dissemination of resources for use by heath care providers for assessing and treating hypertension COLOMBIA, MALAYSIA, & CANADA Community social events or other non-clinical gatherings Non-physician health workers attend the community events NPHW attend events opportunistically with the permission of event organizers Posters explaining the NPHW attendance; curated collections of local government brochures regarding CV health and other available health services; personalized healthy lifestyle counselling based on WHO recommendations (Intervention-only). Key Points Elevated blood pressure is a major risk factor for cardiovascular disease, and it is the leading global risk for mortality. There is a need for novel approaches when addressing hypertension due to its growing health and economic burden on LMIC populations. The Global Alliance for Chronic Diseases sponsored 15 research projects focused on hypertension. These research projects have involved the development and evaluation of several important innovative approaches to hypertension control, including: community engagement, salt reduction, salt substitution, task redistribution, mHealth, and fixed-dose combination therapies. Synopsis Elevated blood pressure, a major risk factor for ischemic heart disease, heart failure, and stroke, is the leading global risk for mortality. Despite global efforts to combat hypertension, treatment and control rates are very low in LMICs. Given the continued significant health and economic burden on LMIC populations, there is an urgent need to address the problem by way of novel approaches. The Global Alliance for Chronic Diseases sponsored 15 research projects focused on hypertension, which have involved the development and evaluation of several important innovative approaches to hypertension control, including: community engagement, salt reduction, salt substitution, task redistribution, mHealth, and fixed-dose combination therapies. Disclosure Statement: The authors have nothing to disclose. Authors’ Contributions All authors were involved in the initial draft of this manuscript, made continual input as the drafts progressed, and approved the final draft for submission. The content within is solely the responsibility of the authors and does not necessarily represent the official views of the Global Alliance for Chronic Diseases funding agencies or affiliates. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. 1 Lim SS Vos T Flaxman AD A comparative risk assessment of burden of disease and injury attributable to 67 risk factors and risk factor clusters in 21 regions, 1990–2010: a systematic analysis for the Global Burden of Disease Study 2010 Lancet 2012 380 9859 2224 2260 23245609 2 Lewington S Clarke R Qizilbash N Age-specific relevance of usual blood pressure to vascular mortality: a meta-analysis of individual data for one million adults in 61 prospective studies Lancet 2002 360 9349 1903 1913 12493255 3 Chow CK Teo KK Rangarajan S Prevalence, awareness, treatment, and control of hypertension in rural and urban communities in high-, middle-, and low-income countries JAMA : the journal of the American Medical Association 2013 310 9 959 968 24002282 4 Disease GAfC http://www.gacd.org/about/history/whoweare Accessed 5/20/2016 5 Tobe W S, the Global Alliance for Chronic Diseases Hypertension Research Teams With the World Hypertension L The Global Alliance for Chronic Diseases Supports 15 Major Studies in Hypertension Prevention and Control in Low- and Middle-Income Countries The Journal of Clinical Hypertension 2016 n/a-n/a 6 Khatib R Schwalm JD Yusuf S Patient and healthcare provider barriers to hypertension awareness, treatment and follow up: a systematic review and meta-analysis of qualitative and quantitative studies PLoS One 2014 9 1 e84238 24454721 7 Maimaris W Paty J Perel P The influence of health systems on hypertension awareness, treatment, and control: a systematic literature review PLoS Med 2013 10 7 e1001490 23935461 8 Digiacomo M Abbott P Davison J Facilitating uptake of Aboriginal Adult Health Checks through community engagement and health promotion Quality in primary care 2010 18 1 57 64 9 National Institute for Health and Care Excellence Community engagement to improve health (downloaded from https://www.nice.org.uk/guidance/lgb16/resources/communityengagement-to-improve-health-60521149786309 14 May 2016). United Kingdom National Institute for Health and Care Excellence 2014 10 Rosato M Laverack G Grabman LH Community participation: lessons for maternal, newborn, and child health Lancet 2008 372 9642 962 971 18790319 11 Joint United Nations Programme on HIV/AIDS (UNAIDS) Promising practices in community engagement for elimination of new HIV infections among children by 2015 and keeping their mothers alive (dowbloaded from http://www.unaids.org/sites/default/files/media_asset/20120628_JC2281_PromisingPracticesCommunityEngagements_en_0.pdf 15 May 2016) Geneva, Switzerland UNAIDS 2012 12 Rifkin SB Examining the links between community participation and health outcomes: a review of the literature Health policy and planning 2014 29 Suppl 2 ii98 106 25274645 13 Marston C Renedo A McGowan CR Effects of community participation on improving uptake of skilled care for maternal and newborn health: a systematic review PloS one 2013 8 2 e55012 23390509 14 Vedanthan R Kamano JH Naanyu V Optimizing linkage and retention to hypertension care in rural Kenya (LARK hypertension study): study protocol for a randomized controlled trial Trials 2014 15 1 143 24767476 15 Naanyu V Vedanthan R Kamano JH Barriers Influencing Linkage to Hypertension Care in Kenya: Qualitative Analysis from the LARK Hypertension Study Journal of general internal medicine 2016 31 3 304 314 26728782 16 Maar M Yeates K Barron M I-RREACH: an engagement and assessment tool for improving implementation readiness of researchers, organizations and communities in complex interventions Implement Sci 2015 10 64 25935849 17 Bradley HA Puoane T Prevention of hypertension and diabetes in an urban setting in South Africa: participatory action research with community health workers Ethnicity & disease 2007 17 1 49 54 17274209 18 Balcazar HG Byrd TL Ortiz M A randomized community intervention to improve hypertension control among Mexican Americans: using the promotoras de salud community outreach model Journal of health care for the poor and underserved 2009 20 4 1079 1094 20168020 19 He FJ Li J Macgregor GA Effect of longer term modest salt reduction on blood pressure: Cochrane systematic review and meta-analysis of randomised trials Bmj 2013 346 f1325 23558162 20 Aburto NJ Ziolkovska A Hooper L Effect of lower sodium intake on health: systematic review and meta-analyses Bmj 2013 346 f1326 23558163 21 Reducing salt intake in populations: report of a WHO forum and technical meeting Geneva World Health Organization 2007 22 Guideline: Sodium intake for adults and children Geneva World Health Organization 2012 23 WHO Monitoring framework and targets for the prevention and control of NCDs Revised WHO discussion paper on the development of a comprehensive global monitoring framework, including indicators, and a set of voluntary global targets for the prevention and control of NCDs 7 25 2012 http://www.who.int/nmh/events/2012/discussion_paper3.pdf accessed February 17, 2013 24 Trieu K Neal B Hawkes C Salt Reduction Initiatives around the World - A Systematic Review of Progress towards the Global Target PloS one 2015 10 7 e0130247 26201031 25 Webster J Su’a SA Ieremia M Salt Intakes, Knowledge, and Behavior in Samoa: Monitoring Salt-Consumption Patterns Through the World Health Organization’s Surveillance of Noncommunicable Disease Risk Factors (STEPS) Journal of clinical hypertension 2016 26 The World Health Organization Creating an enabling environment for population-based salt reduction strategies World Health Organization 2011 27 Webster J Snowdon W Moodie M Cost-effectiveness of reducing salt intake in the Pacific Islands: protocol for a before and after intervention study BMC public health 2014 14 107 24495646 28 Dunford E Trevena H Goodsell C FoodSwitch: A Mobile Phone App to Enable Consumers to Make Healthier Food Choices and Crowdsourcing of National Food Composition Data JMIR mHealth and uHealth 2014 2 3 e37 25147135 29 He FJ Wu Y Feng XX School based education programme to reduce salt intake in children and their families (School-EduSalt): cluster randomised controlled trial Bmj 2015 350 h770 25788018 30 Peng YG Li W Wen XX Effects of salt substitutes on blood pressure: a meta-analysis of randomized controlled trials The American journal of clinical nutrition 2014 100 6 1448 1454 25411279 31 Bernabe-Ortiz A Diez-Canseco F Gilman RH Launching a salt substitute to reduce blood pressure at the population level: a cluster randomized stepped wedge trial in Peru Trials 2014 15 93 24667035 32 Saavedra-Garcia L Bernabe-Ortiz A Gilman RH Applying the Triangle Taste Test to Assess Differences between Low Sodium Salts and Common Salt: Evidence from Peru PloS one 2015 10 7 e0134700 26225848 33 Salt substitution: a low-cost strategy for blood pressure control among rural Chinese. A randomized, controlled trial J Hypertens 2007 25 10 2011 2018 17885542 34 Zhou X Liu JX Shi R Compound ion salt, a novel low-sodium salt substitute: from animal study to community-based population trial American journal of hypertension 2009 22 9 934 942 19661926 35 Zhao X Yin X Li X Using a low-sodium, high-potassium salt substitute to reduce blood pressure among Tibetans with high blood pressure: a patient-blinded randomized controlled trial PloS one 2014 9 10 e110131 25338053 36 World Health Organization World Health Report 2006: Working together for health Geneva WHO 2006 37 Ministry of Health and Family Welfare Rural Health Statistics Bulletin, March 2010 New Delhi Government of India 2010 38 World Health Organisation Density of physicians (total number per 1000 population, latest available year) Global Health Observatory Data http://www.who.int/gho/health_workforce/physicians_density_text/en/ Accessed 13.5.16, 2016 39 Deller B Tripathi V Stender S Task shifting in maternal and newborn health care: Key components from policy to implementation International Journal of Gynecology & Obstetrics 2015 130 Supplement 2 S25 S31 26115853 40 Dawson AJ Buchan J Duffield C Task shifting and sharing in maternal and reproductive health in low-income countries: a narrative synthesis of current evidence Health policy and planning 2013 41 World Health Organization Task shifting to tackle health worker shortages 2007 42 Lekoubou A Awah P Fezeu L Hypertension, Diabetes Melitus and task shifting and their management in Sub-saharan Africa Int J Environ Res Public Health 2010 7 353 363 20616978 43 60th WMA General Assembly WMA Resolution on Task Shifting from the Medical Profession New Delhi World Medical Association 2009 44 Joshi R Alim M Kengne AP Task Shifting for Non-Communicable Disease Management in Low and Middle Income Countries? A Systematic Review PloS one 2014 9 8 e103754 25121789 45 Ogedegbe G Gyamfi J Plange-Rhule J Task shifting interventions for cardiovascular risk reduction in low-income and middle-income countries: a systematic review of randomised controlled trials BMJ open 2014 4 10 e005983 46 Gaziano TA Abrahams-Gessel S Denman CA An assessment of community health workers’ ability to screen for cardiovascular disease risk with a simple, non-invasive risk assessment instrument in Bangladesh, Guatemala, Mexico, and South Africa: an observational study The Lancet Global Health 2015 3 9 e556 e563 26187361 47 Akinyemi RO Owolabi MO Adebayo PB Task-shifting training improves stroke knowledge among Nigerian non-neurologist health workers Journal of the neurological sciences 2015 359 1–2 112 116 26671098 48 Joshi R Chow C Raju PK The Rural Andhra Pradesh Cardiovascular Prevention Study JACC 2012 59 13 1188 1196 22440219 49 Kar SS Thakur JS Jain S Cardiovascular disease risk management in a primary health care setting of north India Indian heart journal 2008 60 1 19 25 19212017 50 Jafar TH Islam M Hatcher J Community based lifestyle intervention for blood pressure reduction in children and young adults in developing country: cluster randomised controlled trial Bmj 2010 340 51 Xavier D Gupta R Kamath D Community health worker-based intervention for adherence to drugs and lifestyle change after acute coronary syndrome: a multicentre, open, randomised controlled trial The Lancet Diabetes & Endocrinology 4 3 244 253 52 Gaziano T Abrahams-Gessel S Surka S Cardiovascular Disease Screening By Community Health Workers Can Be Cost-Effective In Low-Resource Countries Health affairs (Project Hope) 2015 34 9 1538 1545 26355056 53 Buttorff C Hock RS Weiss HA Economic evaluation of a task-shifting intervention for common mental disorders in India Bull World Health Organ 2012 90 813 821 23226893 54 Thorogood M Goudge J Bertram M The Nkateko health service trial to improve hypertension management in rural South Africa: study protocol for a randomised controlled trial Trials 2014 15 435 25380994 55 Ogedegbe G Plange-Rhule J Gyamfi J A cluster-randomized trial of task shifting and blood pressure control in Ghana: study protocol Implementation science : IS 2014 9 73 24923300 56 Praveen D Patel A McMahon S A multifaceted strategy using mobile technology to assist rural primary healthcare doctors and frontline health workers in cardiovascular disease risk management: protocol for the SMARTHealth India cluster randomised controlled trial Implementation science : IS 2013 8 137 24274431 57 Praveen D Patel A Raghu A SMARTHealth India: Development and Field Evaluation of a Mobile Clinical Decision Support System for Cardiovascular Diseases in Rural India JMIR mHealth and uHealth 2014 2 4 e54 25487047 58 Piette JD List J Rana GK Mobile Health Devices as Tools for Worldwide Cardiovascular Risk Reduction and Disease Management Circulation 2015 132 21 2012 2027 26596977 59 Peiris D Praveen D Johnson C Use of mHealth systems and tools for non-communicable diseases in low- and middle-income countries: a systematic review Journal of cardiovascular translational research 2014 7 8 677 691 25209729 60 Bloomfield GS Vedanthan R Vasudevan L Mobile health for non-communicable diseases in Sub-Saharan Africa: a systematic review of the literature and strategic framework for research Globalization and health 2014 10 49 24927745 61 Beratarrechea A Lee AG Willner JM The impact of mobile health interventions on chronic disease outcomes in developing countries: a systematic review Telemedicine journal and ehealth : the official journal of the American Telemedicine Association 2014 20 1 75 82 62 Okoro E Sholagberu H Kolo P Mobile phone ownership among Nigerians with diabetes African Health Sciences 2010 10 2 183 186 21326973 63 Owolabi MO Akinyemi RO Gebregziabher M Randomized controlled trial of a multipronged intervention to improve blood pressure control among stroke survivors in Nigeria International journal of stroke : official journal of the International Stroke Society 2014 9 8 1109 1116 25042605 64 Cushman WC Ford CE Einhorn PT Blood pressure control by drug group in the Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial (ALLHAT) Journal of clinical hypertension 2008 10 10 751 760 19090876 65 Johnston A Stafylas P Stergiou GS Effectiveness, safety and cost of drug substitution in hypertension British journal of clinical pharmacology 2010 70 3 320 334 20716230 66 Shaw E Anderson JG Maloney M Factors associated with noncompliance of patients taking antihypertensive medications Hosp Pharm 1995 30 3 201 203 10140764 67 Faria C Wenzel M Lee KW A narrative review of clinical inertia: focus on hypertension Journal of the American Society of Hypertension : JASH 2009 3 4 267 276 20409968 68 Okonofua EC Simpson KN Jesri A Therapeutic inertia is an impediment to achieving the Healthy People 2010 blood pressure control goals Hypertension 2006 47 3 345 351 16432045 69 Brown MJ McInnes GT Papst CC Aliskiren and the calcium channel blocker amlodipine combination as an initial treatment strategy for hypertension control (ACCELERATE): a randomised, parallel-group trial Lancet 2011 377 9762 312 320 21236483 70 Feldman RD Zou GY Vandervoort MK A Simplified Approach to the Treatment of Uncomplicated Hypertension: A Cluster Randomized, Controlled Trial Hypertension 2009 53 4 646 653 19237683 71 Gupta AK Arshad S Poulter NR Compliance, safety, and effectiveness of fixed-dose combinations of antihypertensive agents: a meta-analysis Hypertension 2010 55 2 399 407 20026768 72 Webster R Patel A Selak V Effectiveness of fixed dose combination medication (‘polypills’) compared with usual care in patients with cardiovascular disease or at high risk: A prospective, individual patient data meta-analysis of 3140 patients in six countries International journal of cardiology 2016 205 147 156 26736090 73 Castellano JM Sanz G Penalvo JL A polypill strategy to improve adherence: results from FOCUS (Fixed-dose Combination Drug for Secondary Cardiovascular Prevention) Project Journal of the American College of Cardiology 2014 74 Lonn EM Bosch J López-Jaramillo P Blood-Pressure Lowering in Intermediate-Risk Persons without Cardiovascular Disease New England Journal of Medicine 0(0):null 75 Yusuf S Lonn E Pais P Blood-Pressure and Cholesterol Lowering in Persons without Cardiovascular Disease The New England journal of medicine 2016 76 Salam A Webster R Singh K TRIple pill vs Usual care Management for Patients with mildto- moderate Hypertension (TRIUMPH): Study protocol American heart journal 2014 167 2 127 132 24439972 77 World Health Organization, PEPFAR, UNAIDS Task shifting: rational redistribution of tasks among health workforce teams : global recommendations and guidelines http://www.who.int/healthsystems/TTR-TaskShifting.pdf 2016 78 Thom S Poulter N Field J Effects of a fixed-dose combination strategy on adherence and risk factors in patients with or at high risk of cvd: The umpire randomized clinical trial JAMA 2013 310 9 918 929 24002278 79 Patel A Cass A Peiris D A pragmatic randomized trial of a polypill-based strategy to improve use of indicated preventive treatments in people at high cardiovascular disease risk European journal of preventive cardiology 2015 22 7 920 930 24676715 80 Selak V Elley CR Bullen C Effect of fixed dose combination treatment on adherence and risk factor control among patients at high risk of cardiovascular disease: randomised controlled trial in primary care Bmj 2014 348 g3318 24868083

PMC005xxxxxx/PMC5131535.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 101525337 37346 ACS Chem Neurosci ACS Chem Neurosci ACS chemical neuroscience 1948-7193 27184474 5131535 10.1021/acschemneuro.6b00111 NIHMS830095 Article ML418: The first selective, sub-micromolar pore blocker of Kir7.1 potassium channels Swale Daniel R. 1€ Kurata Haruto 23€ Kharade Sujay V. 1 Sheehan Jonathan 4 Raphemot Rene R. 12 Voigtritter Karl R. 23 Figueroa Eric 12 Meiler Jens 4 Blobaum Anna L. 23 Lindsley Craig W. 23 Hopkins Corey R. *23 Denton Jerod S. *12 1 Department of Anesthesiology, Vanderbilt University Medical Center, Nashville, TN 37232 2 Department of Pharmacology, Vanderbilt University Medical Center, Nashville, TN 37232 3 Vanderbilt Center for Neuroscience Drug Discovery and the Vanderbilt Specialized Chemistry Center for Accelerated Probe Development, Vanderbilt University Medical Center, Nashville, TN 37232 4 Department of Chemistry, Vanderbilt University, Nashville, TN 37232 * Corresponding Authors Jerod S. Denton, Ph.D., T4208 Medical Center North, 1161 21st Avenue South, Nashville, TN 37232, jerod.s.denton@vanderbilt.edu, T: 615-343-7385 * Corey R. Hopkins, Ph.D., Cool Springs Life Science Center, 393 Nichol Mill Lane, Franklin, TN 37067, corey.r.hopkins@vanderbilt.edu, T: 615-936-6892 € Authors contributed equally 16 11 2016 24 5 2016 20 7 2016 01 12 2016 7 7 10131023 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. The inward rectifier potassium (Kir) channel Kir7.1 (KCNJ13) has recently emerged as a key regulator of melanocortin signaling in the brain, electrolyte homeostasis in the eye, and uterine muscle contractility during pregnancy. The pharmacological tools available for exploring the physiology and therapeutic potential of Kir7.1 have been limited to relatively weak and non-selective small-molecule inhibitors. Here, we report the discovery in a fluorescence-based high-throughput screen of a novel Kir7.1 channel inhibitor, VU714. Site-directed mutagenesis of pore-lining amino acid residues identified Glutamate 149 and Alanine 150 as essential determinants of VU714 activity. Lead optimization with medicinal chemistry generated ML418, which exhibits sub-micromolar activity (IC50 = 310 nM) and superior selectivity over other Kir channels (at least 17-fold selective over Kir1.1, Kir2.1, Kir2.2, Kir2.3, Kir3.1/3.2, and Kir4.1) except for Kir6.2/SUR1 (equally potent). Evaluation in the EuroFins Lead Profiling panel of 64 GPCRs, ion-channels and transporters for off-target activity of ML418 revealed a relatively clean ancillary pharmacology. While ML418 exhibited low CLHEP in human microsomes which could be modulated with lipophilicity adjustments, it showed high CLHEP in rat microsomes regardless of lipophilicity. A subsequent in vivo PK study of ML418 by intraperitoneal (IP) administration (30 mg/kg dosage) revealed a suitable PK profile (Cmax = 0.20 µM and Tmax = 3 hours) and favorable CNS distribution (mouse brain:plasma Kp of 10.9 to support in vivo studies for in vivo studies. ML418, which represents the current state-of-the-art in Kir7.1 inhibitors, should be useful for exploring the physiology of Kir7.1 in vitro and in vivo. KCNJ13 thallium flux electrophysiology comparative modeling melanocortin signaling myometrium INTRODUCTION Inward rectifier potassium (Kir) channels play fundamental roles in diverse organ systems, and could in some cases represent unexploited drug targets for neurological, cardiovascular, endocrine, and muscle disorders 1, 2. Kir7.1, which is encoded by KCNJ13, one of sixteen genes comprising the Kir channel family, is expressed in the eye, brain, uterus, kidney, gut, and thyroid gland3–9. The genetic loss of Kir7.1 function in retinal pigmented epithelial cells of the eye leads to derangements in subretinal electrolyte homeostasis and cell degeneration underlying leber congenital amaurosis and snowflake vitreoretinopathy10–13. Using a panel of mostly non-specific inhibitors with differential activities against Kir channels, Ghamari-Langroudi and colleagues recently identified Kir7.1-like currents in neurons of the paraventricular nucleus (PVN) that are functionally coupled to the melanocortin-4 receptor (MC4R)14. Agonist binding to MC4R inhibits Kir7.1 activity, depolarizes the membrane potential, and increases neuronal firing, whereas competitive antagonist binding increases Kir7.1 activity and dampens neuronal excitability. This model suggests that Kir7.1 coupling to MC4R plays a key role in the central regulation of food intake and energy homeostasis by the PVN14. Kir7.1 expression in uterine muscles increases dramatically during pregnancy, thereby hyperpolarizing the membrane potential, inhibiting calcium signaling, and inducing uterine quiescence during fetal development15. Inhibiting Kir7.1 expression with microRNAs or inhibiting channel function with small-molecule inhibitors (i.e. VU59016, MRT0020076917) induces long-lasting uterine muscle contractions, lending support to the idea that Kir7.1 represents a novel drug target for augmentation of labor and treating post-partum hemorrhage17. The current pharmacological ‘toolkit’ for exploring the physiology and therapeutic potential of Kir7.1 is inadequate, prompting us to perform a high-throughput screen (HTS) and lead optimization campaign in order to identify more potent and selective inhibitors. Here, we report the development and in vitro characterization of ML418, a new state-of-the-art inhibitor of Kir7.1. RESULTS AND DISCUSSION Kir7.1-M125R Assay Development HTS was performed using a fluorescence-based assay that reports the inward movement of the potassium (K+) congener thallium (Tl+) through Kir7.1 channels expressed in the plasma membrane of T-REx-HEK293 cells. The higher conductance Kir7.1-M125R mutant was used as a surrogate to circumvent the weak Tl+ flux observed for wild type (WT) Kir7.1 (see ref. 18). As shown in Fig. 1, the assay reports robust Kir7.1-M125R–dependent Tl+ flux after induction with tetracycline (Fig. 1A), is DMSO tolerant up to 0.6% v/v (Fig. 1B; screening DMSO concentration = 0.1% DMSO v/v), and is sufficiently reproducible for HTS (Fig. 1C; mean ± SEM Z’= 0.67 ± 0.03; n = 3 plates on 3 separate days). Discovery and Characterization of VU714 From a pilot screen of 5,230 compounds in the Vanderbilt Institute of Chemical Biology (VICB) library, 11 putative Kir7.1-M125R inhibitors, comprising 5 distinct scaffolds, and with differing levels of selectivity over other Kir channels, were identified (data not shown). VU714 (Fig. 2A) was the most potent and selective inhibitor from the screen, and was therefore re-synthesized and confirmed from powder to be an authentic Kir7.1-M125R inhibitor. VU714 inhibited Kir7.1-M125R-mediated Tl+ flux in a dose-dependent manner with an IC50 of 5.6 µM (95% Confidence Interval [CI]: 1.9 µM - 17.5 µM) (Fig 2B–C). In “gold-standard” whole-cell voltage clamp experiments, the rate of Kir7.1-M125R inhibition by VU714 was concentration dependent (Fig. 2D), 10 µM VU714 fully inhibited outward and inward Kir7.1-M125R–mediated current (Fig. 2E), and the IC50 was 1.5 µM (CI: 1.3 µM - 1.7µM) (Fig. 2F). The 3.7-fold shift in IC50 determined with patch clamp electrophysiology, as compared with Tl+ flux, is consistent with previous observations of other Kir channel inhibitors 18–20. Quantitative Tl+ flux assays were utilized to evaluate the selectivity of VU714 for Kir7.1 over Kir1.1, Kir2.1, Kir2.2, Kir2.3, Kir3.1/3.2, Kir4.1, and Kir6.2/SUR1, as reported previously 16, 21, 22. The concentration-response curves (CRCs) shown in Fig. 3A revealed that VU714 is only moderately selective, and inhibits other Kir channels with a rank order potency of Kir7.1 (IC50 = 5.6 µM) > Kir4.1 (IC50 = 13 µM) > Kir1.1 (IC50 = 16 µM) > Kir6.2/SUR1 (IC50 = 30 µM) > Kir2.1, Kir2.2, Kir2.3, Kir3.1/3.2 (IC50 > 30 µM). Kir2.2,Kir2.3, and Kir3.1/3.2 CRCs have been excluded from Fig. 3A for clarity. VU714 Requires Pore-lining E149 and A150 Residues for Activity Kir channels are tetrameric proteins consisting of 8 membrane-spanning domains, an ion-conduction pore, and cytoplasmic domain. Most small-molecule inhibitors studied to date appear to block the conduction pathway by interacting with amino acid residues lining the transmembrane pore (reviewed in 1). We therefore performed scanning mutagenesis and voltage-clamp electrophysiology to test if VU714 interacts with pore-lining amino acids in Kir7.1. Residues that are predicted from homology modeling (Fig. 4A) to face the pore were mutated to the corresponding residues in Kir2.1 and Kir1.1, which are only weakly inhibited by (Kir1.1) or insensitive (Kir2.1) to VU714 (Fig. 3A). Mutations were first introduced into WT Kir7.1, and if they impaired channel activity, were retested in the M125R background due to its more robust functional expression and ability to rescue the activity of some mutants. For clarity, only non-functional mutants in the M125R background are indicated in Fig. 4B; however, these mutations were first tested in the WT background and found to be non-functional. The M125R mutation does not alter VU714 sensitivity (Fig. 4C). Mutations L146V, I160M, and A161S had no effect on Kir7.1 inhibition by 3 µM VU714, whereas the T153C/WT mutation significantly (P <0.05) increased channel inhibition. Mutation of E149 to the corresponding Asparagine residue in Kir1.1 (E149N) abolished Kir7.1 activity in both WT and M125R backgrounds. However, the more conservative mutation, E149Q/M125R, led to a partial, albeit significant (P<0.05), loss of sensitivity to VU714. The negative charge at position 149 appears to be important for VU714 activity since mutating E149 to Aspartate (E149D) had no effect on block. Mutation of the adjacent residue A150 to Serine (A150S) led to a partial loss of inhibition. As shown in Fig. 4C, the E149Q and A150S mutations significantly (P <0.0001) shifted the IC50 for VU714 from ~2 µM in the WT or M125R mutants to ~18 µM and ~7 µM, respectively. Mutation of both residues simultaneously led to an additive loss of VU714 sensitivity (Fig. 4B). Molecular modeling was used to visualize how VU714 might interact with Kir7.1 to induce channel block. Interactive ligand placement near E149 and A150, combined with energy minimization, revealed the best theoretical configuration of VU714 in the Kir7.1 channel pore (Fig. 4D–E). The limited volume of the pore in the putative binding region results in obstruction of the channel by VU714, just below the selectivity filter. This model suggests a straightforward steric mechanism for the observed conduction block by VU714. It is notable that this general region of pore is also involved in block of Kir1.1 by VU59123, Kir2.1 by ML13324, and Kir4.1 by fluoxetine25 (reviewed in ref. 1). Pentamidine26 and chloroquine27 require residues in the cytoplasmic pore for block; however, it is unlike that these residues participate in VU714 inhibition of Kir7.1 since the double mutation E149Q and A150S virtually eliminated block of the channel (Fig. 4B). VU714 Structure-Activity Relationships Medicinal chemistry was employed in an effort to identify VU714 analogs with improved potency and selectivity for Kir7.1. Our first priority was to reduce the high lipophilicity of VU714 (clogP = 5.83), which may be a liability for off-target selectivity, metabolism and toxicity. The initial library to explore SAR (Structure-Activity Relationship) in the right-hand portion of VU714 identified several equipotent moieties, including a simple methyl compound (1, IC50 = 4.8 µM, clogP = 4.41) without the benzyl moiety (Table 1). Subsequent libraries with a hydrophilic handle indicated that changing the piperidine ring system to a piperazine was not productive (amide, sulfonamide, urea and carbamate, 2 – 6), including the direct benzyl analog which resulted in a 4.5-fold loss of potency compared to VU714 (5, IC50 = 22.0 µM). However, attaching a hydrophilic handle (−NHR’) at the 4-position of the piperidine ring system endowed less lipophilic compounds with equal potency to the HTS hit compound, namely the amide compound (7, IC50 = 8.3 µM, clogP = 3.43) and the carbamate (11, IC50 = 1.7 µM, clogP = 3.65) which were generally more potent than the corresponding N-alkylated derivatives (12 – 16) (Table 1). Based on these results, we next evaluated a number of carbamate analogs in order to explore the SAR around this moiety (Table 2). While the simple methyl carbamate (17, inactive) and ethyl carbamate (18, IC50 = 9.8 µM) analogs were less active, all of the branched alkyl carbamates (19 – 22) were equipotent with the HTS hit compound. Among them, the isopropyl (19, IC50 = 1.3 µM, clogP = 3.25) and the tert-butyl carbamate (11, IC50 = 1.7 µM, clogP = 3.65) were the most active compounds. The benzyl compound (23), ring contracted compounds (24 and 25) and the spirocyclic compounds (26 and 27) were also tolerated, but less active than the HTS hit VU714. The next library was focused on right-hand 4-aminopiperidine amide analogs. The most active amides were aryl analogs in nature (28 – 40, 43) with the 3-chlorophenyl (32, IC50 = 3.1 µM, clogP = 4.34) being the most potent in this class. Branched alkyl amides (41 – 42) were also evaluated, but they were less active than the parent compound (Table 2). Next, we explored SAR around a left-hand quinoline moiety keeping the 4-methyl piperidine in the right-hand portion constant. Unfortunately, structural modifications of this portion of the molecule were not well tolerated, even minor changes such as changing the chlorine to fluorine and bromine atoms were less active. The only modification tolerated in our SAR exploration was an introduction of a methyl group at the 4-position of the quinoline ring (Fig. S1). The SAR data (Fig. S1) is in qualitative agreement with the docking model that illustrates a plausible binding mode (Fig. 4E). Specifically, position 2 of the quinoline ring faces towards the protein backbone. There is no space for additional bulk. This is in contrast to positions 3 and 4 of the quinoline ring where methyl substitution is tolerated. The Nitrogen atom of the quinoline points to a polar pocket that could provide space for an additional, hydrogen-bound water molecule. A Carbon atom could not engage in these favorable interactions. The chloride and hydroxyl substituents are in tight, polar pockets. No larger substituents will fit into these pockets. A quantitative analysis of the agreement is beyond the predictive power of the homology model used for the docking simulation. Discovery and Characterization of ML418 Having identified several VU714 analogs with lower lipophilicity and similar potencies in Tl+ flux assays, we next evaluated their activity toward Kir7.1 using voltage clamp electrophysiology and selectivity among Kir1.1, Kir2.1, Kir2.2, Kir2.3, Kir3.1/3.2, Kir4.1 and Kir6.2/SUR1 in Tl+ flux assays. The methyl piperidine analog (1) was similar in activity and selectivity to VU714 (Table 3). However, the carbamate (11 and 19) and amide (32) analogs exhibited superior inhibitory activity and selectivity. The isopropyl carbamate analog is the most potent in Kir7.1 activity (19 = ML418) in our library and is at least 23-fold selective against the Kir channels tested, with the exception of Kir6.2/SUR1 (Fig. 3B; Table 3). ML418 inhibits Kir7.1 dose-dependently (Fig. 3C) with an IC50 in patch clamp electrophysiology experiments of 310 nM (Fig. 3D; Table 3). While the tert-butyl carbamate (11) showed similar potency and selectivity to ML418, the amide analog (32) exhibited modest activity and greater than 10-fold selectivity over the Kir channels tested (Table 3). Although compounds 44 and 45 also show selectivity preference over Kir6.2/SUR1, and should be considered for use as an in vitro tool compound, the totality of the properties favors ML418 as the first selective Kir7.1 inhibitor. In a Lead Profiling Screen (Eurofins) of 64 potential off-targets, ML418 showed a relatively clean ancillary pharmacology having significant interactions (i.e. >50% radioligand displacement) with the L-type calcium channel, voltage-gated sodium channel, Dopamine D2S and D4.2 receptors, Sigma σ1 receptor, and Norepinephrine receptor (Table S1). DMPK Evaluation for Selected Analogs We evaluated selected analogs in our in vitro DMPK panel of assays, an assessment of intrinsic clearance (CLINT) and predicted hepatic clearance (CLHEP) in hepatic microsomes and protein binding in plasma (PPB) in multiple species. Interestingly, while the lipophilic HTS hit compound VU714 (cLogP = 5.83) was shown to have high intrinsic clearance (CLHEP (human) = 16.7 mL/min/kg and CLHEP (rat) =64.1 mL/min/kg) and low unbound fraction in plasma (%Fu (mouse) = 0.7, %Fu (rat) = 1.8 and %Fu (human) = 0.5) across species, more potent analogs than VU714 (32, 11 and 19 = ML418) exhibited lower intrinsic clearance in human hepatic microsomes (CLHEP (human) = 10.7, 7.9 and 3.6 mL/min/kg, respectively) and higher unbound fraction in mouse and rat plasma (%Fu (mouse) = 2.0, 5.6 and 11.4 and %Fu (rat) = 2.5, 3.6 and 8.8, respectively) as the cLogPs are reduced. While intrinsic clearance in rat hepatic microsomes of ML418 are slightly reduced (CLHEP (rat) =57.9 mL/min/kg) compared to VU714, subsequent PK study of ML418 by 30 mg/kg dosage of intraperitoneal administration revealed its characteristic PK profile (Fig. 5: Cmax = 0.20 µM and Tmax = 3 hours). It may be attributed to enterohepatic recirculation caused by phenol moiety. In a parallel CNS distribution study in mice, ML418 demonstrated excellent CNS penetration with a mouse brain:plasma ratio (Kp) of 10.9 (average of four mice, brain (323.9 ng/g):plasma (29.5 ng/mL)). Thus, ML418 possesses a DMPK profile suitable for both in vitro and in vivo studies aimed at modulating both peripheral and central Kir7.1 potassium channels. CONCLUSIONS ML418 is the fifth small-molecule inhibitor of Kir7.1 reported to date and currently represents the state-of-the-art for the field (Table 5). The first Kir7.1 inhibitor, VU590, was identified in a HTS of approximately 225,000 compounds for inhibitors of Kir1.1 (IC50 = 0.24 µM) and subsequently found to have weak off-target activity toward Kir7.1 (IC50 ~ 8 µM)16. Evaluation of eight VU590 analogs revealed a flat SAR against Kir7.117. Another compound, VU573 was identified in the same Kir1.1 HTS, but found to have superior activity toward Kir7.1 (IC50 = 4.9 µM) over Kir1.1 (IC50=19 µM) in Tl+ flux assays. VU573 also inhibits Kir2.3 and Kir3 channels with single-micromolar potency18. Lead optimization efforts failed to generate more potent or selective inhibitors of Kir7.1, but did generate ‘inactive’ analogs that can be useful for determining the specificity of VU573 in pharmacology experiments targeting Kir7.1 (e.g. see ref. 14). ML133 was identified in a HTS of more than 300,000 compounds for modulators of Kir2.1 (IC50 = 0.3 µM at pH 8.0; approximately equal activity toward Kir2.2, Kir2.3, and Kir2.6) and found to possess weak Kir7.1 activity (IC50=33µM)24. MRT00200769 was discovered in an electrophysiology-based screen of 7,087 compounds for Kir7.1 inhibitors, but, despite being one of the most potent Kir7.1 inhibitors identified, was found to exhibit preferential activity toward cardiac hERG K+ channels (IC50=0.3 µM) over Kir7.1 (IC50=1.3µM). MRT00200769 also suffers from flat SAR17. In the present study, a HTS of 5,230 compounds from the VICB library led to the discovery of the novel pore blocker VU714, which was optimized with medicinal chemistry to generate ML418. The salient features of ML418 include: 1) Kir7.1 IC50=0.3 µM; 2) at least 23-fold selectivity for Kir7.1 over Kir1.1, Kir2.1, Kir2.2, Kir2.3, Kir4.1 except for Kir6.2/SUR1; 3) clean ancillary pharmacology against 58 GPCRs, ion channels, and transporters, including hERG; 4) PK profile in rat (Cmax = 0.20 µM and Tmax = 3 hours at 30 mg/kg i.p. dosing) and highly CNS penetrant (mouse brain:plasma Kp of 10.9). Further investigations of ML418 in order to probe the detailed physiological functions of Kir7.1 will be reported in due course. METHODS Molecular biology The open reading frame (ORF) of human Kir7.1 was sub-cloned into pcDNA5/TO (Life Technologies) to enable tetracycline-regulated expression (see below). Mutations were introduced into the Kir7.1 ORF using a QuickChange II mutagenesis kit (Agilent Technologies) and sequenced to verify incorporation of the intended mutation. Tetracycline-inducible stable cell lines Stably transfected T-REx-HEK-293 cell lines expressing the following Kir channels were generated and maintained in culture as described previously: Kir1.128, Kir2.122, Kir4.121, Kir6.2/SUR122, Kir7.1-M125R18. Kir channel expression was induced by culturing cells overnight in media containing 1 µg/ml tetracycline. High-throughput screening (HTS) HTS for Kir7.1-M125R modulators was performed using a Tl+-flux reporter assay essentially as described previously16, 20, 21. The Kir7.1-M125R mutation increases the channel unitary conductance and enables robust Tl+ flux measurement that cannot be achieved with the WT channel18. T-REx-HEK-293-Kir7.1-M125R cells (20,000/well) were plated in clear-bottomed, black-walled, 384-well plates, and cultured overnight in Delbecco’s Modified Eagle’s Medium (DMEM) containing 10% dialyzed FBS and 1 µg/ml tetracycline to induce Kir7.1-M125R expression. The following day, the cells were incubated with dye-loading assay buffer (Hank’s Balanced Salt Solution, 20 mM HEPES, pH 7.3) containing 0.01% (w/v) Pluronic F-127 (Life Technologies) and 1.2 µM Thallos-AM (TEFlabs, Austin, TX) Tl+ reporter dye. The dye-loading solution was replaced after 1hr with 20 µL/well assay buffer. Test compounds from the VICB Library were dispensed into 384-well plates using an Echo555 liquid handler (Labcyte, Sunnyvale, CA) and diluted to a 2X concentration. Cell plates were transferred to a Hamamatsu Functional Drug Screening System 6000 (FDSS6000; Hamamatsu, Tokyo, Japan) where 20 µL/well test compounds were dispensed into wells to a nominal concentration of 10 µM. After a 20-min incubation period, baseline fluorescence (excitation 470 ± 20 nm, emission 540 ± 30 nm) was recorded at 1Hz for 10 sec before addition of thallium stimulus buffer containing (in mM): 125 NaHCO3, 1.8 CaSO4, 1 MgSO4, 5 glucose, and 1.8 Tl2SO4. Fluorescence data were collected for an additional 4 min at 1Hz. Data were analyzed essentially as described previously28, 29 using a combination of Microsoft Excel (Microsoft Corporation, Redmond, WA) with XLfit add-in (IDBS, Guildford, Surrey, UK) and GraphPad Prism™(GraphPad Software, San Diego, CA, USA). Raw data were opened in Microsoft Excel and each data point in a given trace was divided by the first data point from that trace followed by subtraction of data points from control traces that were generated in the presence of vehicle controls. The slope of the fluorescence increase beginning 5 sec after Tl+ addition and ending 20 sec after Tl+ addition was calculated. The data were then plotted in Prism software to generate CRCs. Potencies were calculated from fits to CRC data using a non-linear regression analysis. Transient transfections and whole-cell patch clamp electrophysiology Transfections and whole-cell patch clamp experiments were performed essentially as described previously22. Briefly, HEK-293T cells were transfected with plasmids encoding WT or mutant Kir7.1 and EGFP (transfection marker) using Lipofectamine LTX (Life Technologies). The next day, the cells were dissociated with trypsin, plated on poly-L-lysine-coated round glass coverslips, and allowed to recover for at least 1hr before beginning experiments. Coverslips were transferred to a small-volume recording chamber on the stage of an inverted fluorescence microscope and superfused with a control bath solution containing (in mM), 135 NaCl, 5 KCl, 2 CaCl2, 1 MgCl2, 5 glucose, 10 HEPES, pH 7.4. Electrodes were pulled with a Flaming-Brown P-1000 micropipette puller and had resistances between 2–3 MΩ when filled with the following solution (in mM): 135 mM KCl, 2 MgCl2, 1 EGTA, 10 HEPES, pH 7.3. The cells were voltage clamped at a holding potential of −75 mV and then stepped to −120 mV for 500 msec before ramping to 120 mV at a rate of 2.4 mv/msec. The cell potential was returned to −75 msec for 5 sec before initiating the step-ramp protocol again. Dose-response experiments were performed by superfusing cells with increasing doses of Kir7.1 inhibitor (e.g. ML418) followed by 4 mM BaCl2 to fully block Kir7.1. Cells exhibiting less than 90% block by BaCl2 were excluded from analysis. IC50 values were determined by fitting the Hill equation to CRCs using variable-slope, unconstrained, nonlinear regression analyses performed with GraphPad Prism (GraphPad Software, San Diego, CA, USA). All experiments yielded acceptable Hill slope (>0.8) and r2 (0.99) values. IC50 values are expressed as mean of n = 5 values. Mean IC50 values and 95% confidence limits were determined with GraphPad InStatTM (GraphPad Software, San Diego, CA, USA). The mean IC50 values were statistically analyzed using a one-way ANOVA measure with Dunnets multiple comparison test with significance being represented by P<0.05. Statistical analyses were performed using InStatTM (GraphPad Software, San Diego, CA, USA). Comparative homology modeling and in silico docking A comparative model of Kir7.1 was generated based on the 3.11 Angstrom resolution crystal structure Kir2.2 (PDB ID 3JYC)30. The sequence identity between Kir7.1 and Kir2.2 is 33.5% among the 343 residues aligned in each subunit of the tetramer. MODELLER version 9.9 [REF 2] was used to construct the model. The inhibitor VU0488714 was docked into the model of Kir7.1 using the Molecular Operating Environment (MOE) software package 31. Docking was performed by allowing flexibility of the ligand molecule and also the side chains of the protein model. A representative docking pose was selected from among the 29 lowest-energy results by choosing the best-scoring pose that features multiple interactions between the inhibitor and the experimentally-determined residues glutamate 149 and alanine 150. Chemical synthesis General. All NMR spectra were recorded on a 400 MHz AMX Bruker NMR spectrometer. 1H and 13C chemical shifts are reported in δ values in ppm downfield with the deuterated solvent as the internal standard. Data are reported as follows: chemical shift, multiplicity (s = singlet, d = doublet, t = triplet, q = quartet, b = broad, m = multiplet), integration, coupling constant (Hz). Low resolution mass spectra were obtained on an Agilent 6120 or 6150 with ESI source. Method A: MS parameters were as follows: fragmentor: 70, capillary voltage: 3000 V, nebulizer pressure: 30 psig, drying gas flow: 13 L/min, drying gas temperature: 350 °C. Samples were introduced via an Agilent 1290 UHPLC comprised of a G4220A binary pump, G4226A ALS, G1316C TCC, and G4212A DAD with ULD flow cell. UV absorption was generally observed at 215 nm and 254 nm with a 4 nm bandwidth. Column: Waters Acquity BEH C18, 1.0 × 50 mm, 1.7 um. Gradient conditions: 5% to 95% CH3CN in H2O (0.1% TFA) over 1.4 min, hold at 95% CH3CN for 0.1 min, 0.5 mL/min, 55 °C. Method B: MS parameters were as follows: fragmentor: 100, capillary voltage: 3000 V, nebulizer pressure: 40 psig, drying gas flow: 11 L/min, drying gas temperature: 350 °C. Samples were introduced via an Agilent 1200 HPLC comprised of a degasser, G1312A binary pump, G1367B HP-ALS, G1316A TCC, G1315D DAD, and a Varian 380 ELSD (if applicable). UV absorption was generally observed at 215 nm and 254 nm with a 4 nm bandwidth. Column: Thermo Accucore C18, 2.1 × 30 mm, 2.6 um. Gradient conditions: 7% to 95% CH3CN in H2O (0.1% TFA) over 1.6 min, hold at 95% CH3CN for 0.35 min, 1.5 mL/min, 45 °C. High resolution mass spectra were obtained on an Agilent 6540 UHD Q-TOF with ESI source. MS parameters were as follows: fragmentor: 150, capillary voltage: 3500 V, nebulizer pressure: 60 psig, drying gas flow: 13 L/min, drying gas temperature: 275 °C. Samples were introduced via an Agilent 1200 UHPLC comprised of a G4220A binary pump, G4226A ALS, G1316C TCC, and G4212A DAD with ULD flow cell. UV absorption was observed at 215 nm and 254 nm with a 4 nm bandwidth. Column: Agilent Zorbax Extend C18, 1.8 µm, 2.1 × 50 mm. Gradient conditions: 5% to 95% CH3CN in H2O (0.1% formic acid) over 1 min, hold at 95% CH3CN for 0.1 min, 0.5 mL/min, 40 °C. For compounds that were purified on a Gilson preparative reversed-phase HPLC, the system comprised of a 333 aqueous pump with solvent-selection valve, 334 organic pump, GX-271 or GX-281 liquid hander, two column switching valves, and a 155 UV detector. UV wavelength for fraction collection was user-defined, with absorbance at 254 nm always monitored. Method: Phenomenex Axia-packed Luna C18, 30 × 50 mm, 5 µm column. Mobile phase: CH3CN in H2O (0.1% TFA). Gradient conditions: 0.75 min equilibration, followed by user defined gradient (starting organic percentage, ending organic percentage, duration), hold at 95% CH3CN in H2O (0.1% TFA) for 1 min, 50 mL/min, 23 °C. Solvents for extraction, washing and chromatography were HPLC grade. All reagents were purchased from Aldrich Chemical Co. and were used without purification. Synthetic scheme and characterization of ML418 Synthetic scheme of ML418 is shown in Figure S2 as an example of general synthetic scheme. Experimental procedure for ML418 is described below. Specific synthetic schemes for each compound are also shown in supplemental information (Figures S3–6). 5-Chloro-8-hydroxyquinoline-7-carbaldehyde (49A) To a solution of 5-chloroquinolin-8-ol 48A (8.08 g, 45.0 mmol) in TFA (75 mL) was added hexamethylenetetramine (12.62 g, 90.0 mmol) at ambient temperature. After a resulting reddish solution was refluxed at 120 −130°C for 5 hours, 1 mol/L HCl-aq (200mL) was added to the reaction mixture at 0°C which was stirred at ambient temperature for 20 min. To the reaction mixture was added ethyl acetate (75 mL) and 5 mol/L NaOH-aq (140mL) at 0°C. The resulting precipitates were collected by filtration and washed with water to give a crude product (6.59 g) which was triturated with mixed solvent of EtOH (25 mL) and Et2O (20 mL) to yield 5-chloro-8-hydroxyquinoline-7-carbaldehyde 49A (3.480 g, 37% yield) as a beige powder. tert-Butyl (1-((5-chloro-8-hydroxyquinolin-7-yl)methyl)piperidin-4-yl)carbamate (11) To a suspension of 49A (818 mg, 3.94 mmol) in DCM (25 mL) was added 4-boc-aminopiperidine (1.58 g, 7.89 mmol) at ambient temperature. After a resulting greenish solution was stirred at ambient temperature for 1 hour, sodium triacetoxyborohydride (1.25 g, 5.91 mmol) was added to the reaction mixture which was stirred at ambient temperature for 18 hours. The reaction mixture was poured into NaHCO3-aq (150 mL) and it was extracted with DCM (1st: 125 mL, 2nd: 50 mL). Combined organic extracts were washed with (NaHCO3-aq + water + brine) and dried over MgSO4. The filtrate was evaporated under reduced pressure to give crude product (2.28 g) which was purified on silica gel chromatography (DCM/MeOH) after combined with another crude product (1.41 g) by same reaction conditions from 49A (451 mg, 2.17 mmol) to yield tert-butyl (1-((5-chloro-8-hydroxyquinolin-7-yl)methyl)piperidin-4-yl)carbamate 11 (2.05 g, 86% yield) as a pale yellow powder. 7-((4-Aminopiperidin-1-yl)methyl)-5-chloroquinolin-8-ol trihydrochloride (51) To a suspension of 11 (1.034 g, 2.64 mmol) in 1,4-dioxane (10 mL) was added HCl/1,4-dioxane (4 mol/L, 20 mL) at ambient temperature. After a reaction mixture was stirred for 24 hours, the precipitate was collected by filtration to yield 7-((4-aminopiperidin-1-yl)methyl)-5-chloroquinolin-8-ol trihydrochloride 51 (1.034 g, 98% yield) as a yellow powder. iso-Propyl (1-((5-chloro-8-hydroxyquinolin-7-yl)methyl)piperidin-4-yl)carbamate (19 = ML418) To a solution of 51 (1.015 g, 2.53 mmol) and DIPEA (1.76 mL, 10.1 mmol) in DCM (25 mL) was added a solution of iso-propyl chloroformate in toluene (2 mol/L, 1.30 mL, 2.66 mmol) at 0°C. After a resulting greenish solution was stirred at ambient temperature for 1 hour, it was poured into ice/NaHCO3-aq, which was extracted with DCM (x 2). Combined organic extracts were dried over MgSO4 and the filtrate was evaporated under reduced pressure. The residue was purified by Gilson HPLC separation system using (0.1% TFA in water)/CH3CN as an eluent to give crude product (c.a. 1.1 g) which was recrystallized from Et2O to yield desired product (633 mg 51% yield) as a TFA salt. To a solution of TFA salt of the desired product (1,236 g, 2.51 mmol) in DCM (37 mL) was added saturated NaHCO3-aq (12.5 mL) at ambient temperature. Organic phase was separated and aqueous phase was extracted with DCM (x 2). Combined organic phase was dried over MgSO4. The filtrate was evaporated under reduced pressure to yield iso-propyl (1-((5-chloro-8-hydroxyquinolin-7-yl)methyl)piperidin-4-yl)carbamate 19 (ML418) (915 mg, 96% yield) as a pale yellow powder. 1H NMR (400.1 MHz, DMSO-d6): 8.93 (dd, J = 4.2, 1.3 Hz, 1H), 8.47 (dd, J = 8.5, 1.3 Hz, 1H), 7.69 (dd, J = 8.5, 4.2 Hz, 1H), 7.61 (s, 1H), 7.02 (d, J = 7.7 Hz, 1H), 4.73 (sept, J = 6.2 Hz, 1H), 3.51-3.16 (br, 1H), 2.85-2.82 (m, 2H), 2.15-2.10 (m, 2H), 1.75-1.72 (m, 2H), 1.46-1.39 (m, 2H), 1.16 (d, J = 6.2 Hz, 6H). 13C NMR (100.6 MHz, DMSO-d6): 155.61, 151.63, 149.47, 139.31, 132.80, 128.72, 125.33, 123.09, 121.32, 118.58, 66.86, 56.62, 52.36, 47.98, 32.29, 22.58. LCMS: RT = 0.773 min, m/z = 378 [M + H]+. HRMS calc’d for: C19H24ClN3O3 [M+], 377.1506; found 377.1507. Funding This work was funded in part by the Vanderbilt Institute of Clinical and Translational Research pilot grant (D.R.S, J.S.D.), DK082884 (J.S.D.), and MLPCN (C.R.H, C.W.L.). Figure 1 Kir7.1 Tl+ flux assay used for HTS (A) Representative Thallos fluorescence traces recorded from T-REx-HEK-293-Kir7.1-M125R cells cultured overnight with (grey line) or without (black line) tetracycline. Thallium stimulus buffer was added to each well simultaneously as indicated with the arrow. (B) DMSO tolerance test indicating that DMSO has no effect on Kir7.1-M125R–mediated Tl+ flux as concentrations up to 1.3% (v/v). (C) Determination of assay reproducibility. Alternate wells of a 384-well plate were treated with DMSO (vehicle) or Kir7.1 inhibitor VU573 (30 µM) before initiating Tl+ flux. Mean fluorescence and 3 S.D. from the mean for each well population are indicated with a blue dashed line and solid black line, respectively. The mean ± SEM. Z’ for 3 plates assayed on 3 separate days was Z’ = 0.67 ± 0.03. Figure 2 Discovery and characterization of VU714 (A) Chemical structure of VU714. (B) Dose-dependent inhibition of Kir7.1-M125R–dependent Tl+ flux by VU714. Cells were pre-treated with the indicated concentrations of VU714 for 10 min before adding Tl+ stimulus buffer (arrow). (C) Mean ± SEM % control fluorescence recorded in the indicated concentrations of VU714 (n = 4). (D) Representative whole-cell patch clamp experiment showing timecourse of VU714-dependent inhibition of Kir7.1 current recorded at −120 mV. VU714 concentrations (in µM) are indicated at the top. Experiments were terminated by bath application of 2 mM barium (Ba). (E) Current-voltage plot showing inhibition of Kir7.1 by 10 µM VU714 or 2 mM Ba. (F) Mean ± SEM % Kir7.1 inhibition at −120 mV. IC50 values were derived by fitting CRC data with a 4-parameter logistical function. Figure 3 Analysis of VU714 and ML418 selectivity for Kir7.1 over other Kir channels (A) VU714 CRCs constructed for Kir7.1-M125R over Kir6.2/SUR1 (open diamonds), Kir1.1 (closed circles), Kir2.1 (closed squares), Kir4.1 (open squares) in Tl+ flux assays. Kir2.2,Kir2.3, Kir3.1/3.2 (IC50s >30 µM) have been excluded for clarity. Data are means ± SEM % control fluorescence (n = 4–10 per concentration). (B) ML418 CRCs constructed for the same channels in Tl+ flux assays. (C) Representative whole-cell patch clamp experiment showing dose-dependent inhibition of Kir7.1 current at −120 mV by the indicated concentration of ML418. The experiment was terminated by bath application of 2 mM Ba. (D) Comparison of CRCs for VU714 (grey line, data from Fig. 2) and ML418 determined in patch clamp electrophysiology experiments. Figure 4 Identification of pore-lining residues in Kir7.1 required for VU714 activity (A) Alignment of pore-lining M2 helices from human Kir7.1, Kir2.1, and Kir1.1, with predicted pore-facing residues indicated with arrowheads. (B) Effects of pore mutations on Kir7.1-WT or Kir7.1-M125R sensitivity to 3 µM VU714. Data are mean ± SEM % inhibition at −120 mV. * P <0.05 compared to respective control. N.F., not functional. (C) VU714 CRC for Kir7.1-WT (closed squares; IC50=1.4 µM ), Kir7.1-M125R (open squares; IC50=1.6 µM), Kir7.1-M125R–E148Q (open circles; IC50=18.1 µM), Kir7.1-WT-A150S (closed circles; IC50=6.9 µM). (D) Kir7.1 homology model showing low-energy pose of VU714 near residues E149 and A150. (E) Higher-magnification view (from white box in D) of VU714 near E149 and A150. Figure 5 Time course in vivo PK profile of ML418 ML418 was dosed at 30 mg/kg in 10% EtOH, 40% PEG 400, 50% saline vehicle. The dosing solution was administered by intraperitoneal injection and whole blood collections via the carotid artery were performed at 0.117, 0.25, 0.5, 1, 2, 4, 7, and 24 hours post dose. Table 1 SAR on attaching hydrophilic handles in a right-hand portion Entry R R’ Kir7.1 IC50 (µM)a 1 - 4.8 2 COPh Inactive 3 SO2Ph Inactive 4 CONHPh Inactive 5 CH2Ph 22.6 6 CO2But Inactive 7 COPh 8.3 8 SO2Ph 14.7 9 CONHPh 15.8 10 CH2Ph 12.4 11 CO2But 1.7 12 COPh Inactive 13 SO2Ph 12.2 14 CONHPh 23.6 15 CH2Ph 7.4 16 CO2But 4.1 a hKir7.1 IC50 reported as average from our Thallium flux assay, n = 3 Table 2 SAR on the carbamate and amide analogs Entry R R’ Kir7.1 IC50a (µM) Entry R R’ Kir7.1 IC50a (µM) 17 Me Inactive 7 Ph 8.3 18 Et 9.8 28 2-F-Ph 9.6 19 (ML418) i-Pr 1.3 29 3-F-Ph 5.7 11 t-Bu 1.7 30 4-F-Ph 12.1 20 i-Bu 4.1 31 2-Cl-Ph 15.8 21 Cyclopentyl 2.8 32 3-Cl-Ph 3.1 22 Cyclohexyl 4.2 33 4-Cl-Ph 3.9 23 CH2Ph 6.3 34 2-OMe-Ph 7.1 24 6.4 35 3-OMe-Ph 9.3 36 4-OMe-Ph 10.1 25 10.6 37 3-Me-Ph 4.2 38 3-CF3-Ph 6.1 26 7.7 39 3-CN-Ph 12.4 40 3-Br-Ph 4.0 41 t-Bu Inactive 27 6.1 42 Cyclohexyl 9.6 43 2-Naphthyl 12.3 a hKir7.1 IC50 reported as average from our Thallium flux assay; n = 3. Table 3 Potency in patch clamp assay and selectivity over related Kir channels in Tl+ assay for selected analogs. ND, not determined. Cmpd R Patch clamp Thallium flux Kir7.1 IC50 (µM) Kir7.1 IC50 (µM) Kir Channel Selectivity (Fold) 1.1 2.1 2.2 2.3 4.1 6.2 3.1/3.2 VU714 1.5 5.6 3 >5 2 5 >5 1 1.4 4.8 6 >7 3 ND 44 0.93 3.0 9 6 9 5 10 6 ND 45 0.59 2.3 8 >13 ND 19 (ML418) 0.31 1.3 >23 1 >23 11 0.47 1.7 >18 1 ND 32 0.86 3.1 >10 ND Table 4 In vitro DMPK profile for selected analogs Cmpd R PPB (%Fu) CLHEP (mL/min/kg) cLogPa Mouse Rat Human Rat Human VU714 0.7 1.8 0.5 64.1 16.7 5.83 19 (ML418) 11.4 8.8 1.0 57.9 3.6 3.25 11 5.6 3.6 1.1 56.2 7.9 3.65 32 2.0 2.5 1.6 60.1 10.7 4.34 a Calculated by ChemDraw Table 5 Comparison with known inhibitors of Kir7.1 Compound Structure Kir7.1 IC50 (µM) ROMK IC50 (µM) cLogPd pKad Tl+ EP Tl+ EP ML418 1.3 0.31 >30 Not Tested 3.25 5.8 VU590 (ref. 16) Not Tested 8 0.3 0.24 7.46 7.5 ML133 (ref. 24a) Not Tested 33 Not Tested >300 3.76 9.1 VU573 (ref. 18) 4.8 0.9 19 Not Tested 5.09 13.3 MRT00200769 (ref. 17b) Not Tested 1.3 (38)c Not Tested 4.47 8.8 a Kir2.1 IC50 = 0.29 µM b hERG IC50 = 0.3 µM c Calculated by ChemDraw SUPPORTING INFORMATION SAR, chemical synthesis of all analogs, in vitro pharmacology procedures, in vitro PK methods, in vivo PK methods. Author Contributions Conceived and designed the experiments: D.R.S., H.K., K.S.V, J.S., R.R.R, E.F., J.M., A.L.B., C.W.L., C.R.H., J.S.D. Performed data analysis: D.R.S., H.K., K.S.V., R.R.R., K.R.V., E.F., J.S., J.M., A.L.B., C.W.L., C.R.H., Contributed reagents/materials/analysis tools: D.R.S., H.K., J.S., J.M., C.W.L., Wrote manuscript: D.R.S., H.K., J.S., J.M., A.L.B., C.W.L., C.R.H., J.S.D. REFERENCES 1 Swale DR Kharade SV Denton JS Cardiac and renal inward rectifier potassium channel pharmacology: emerging tools for integrative physiology and therapeutics Curr Opin Pharmacol 2014 15 7 15 24721648 2 Denton JS Pao AC Maduke M Invited Review - Novel Diuretic Targets Am J Physiol Renal Physiol 2013 305 7 F931 F942 23863472 3 Ookata K Tojo A Suzuki Y Nakamura N Kimura K Wilcox CS Hirose S Localization of inward rectifier potassium channel Kir7.1 in the basolateral membrane of distal nephron and collecting duct J Am Soc Nephrol 2000 11 11 1987 1994 11053473 4 Nakamura N Suzuki Y Sakuta H Ookata K Kawahara K Hirose S Inwardly rectifying K+ channel Kir7.1 is highly expressed in thyroid follicular cells, intestinal epithelial cells and choroid plexus epithelial cells: implication for a functional coupling with Na+ K+-ATPase Biochem J 1999 342 Pt 2 329 336 10455019 5 Krapivinsky G Medina I Eng L Krapivinsky L Yang Y Clapham DE A novel inward rectifier K+ channel with unique pore properties Neuron 1998 20 5 995 1005 9620703 6 Yasuda K Shimura M Nakazawa T Sato H Tomita H Sugano E Tamai M Expression and functional properties of unique inward rectifier K+ channel Kir7.1 in the porcine iris and retinal pigment epithelium Curr Eye Res 2003 27 5 279 287 14562164 7 Fujita A Horio Y Higashi K Mouri T Hata F Takeguchi N Kurachi Y Specific localization of an inwardly rectifying K(+) channel, Kir4.1, at the apical membrane of rat gastric parietal cells; its possible involvement in K+ recycling for the H+-K+-pump J Physiol 2002 540 Pt 1 85 92 11927671 8 Kusaka S Inanobe A Fujita A Makino Y Tanemoto M Matsushita K Tano Y Kurachi Y Functional Kir7.1 channels localized at the root of apical processes in rat retinal pigment epithelium J Physiol 2001 531 Pt 1 27 36 11179389 9 Derst C Hirsch JR Preisig-Muller R Wischmeyer E Karschin A Doring F Thomzig A Veh RW Schlatter E Kummer W Daut J Cellular localization of the potassium channel Kir7.1 in guinea pig and human kidney Kidney Int 2001 59 6 2197 2205 11380822 10 Pattnaik BR Shahi PK Marino MJ Liu X York N Brar S Chiang J Pillers AM Traboulsi EI A Novel KCNJ13 Nonsense Mutation and Loss of Kir7.1 Channel Function Causes Leber Congenital Amaurosis (LCA16) Hum Mutat 2015 11 Pattnaik BR Tokarz S Asuma MP Schroeder T Sharma A Mitchell JC Edwards AO Pillers DA Snowflake vitreoretinal degeneration (SVD) mutation R162W provides new insights into Kir7.1 ion channel structure and function PLoS One 2013 8 8 e71744 23977131 12 Sergouniotis PI Davidson AE Mackay DS Li Z Yang X Plagnol V Moore AT Webster AR Recessive mutations in KCNJ13, encoding an inwardly rectifying potassium channel subunit, cause leber congenital amaurosis Am J Hum Genet 2011 89 1 183 190 21763485 13 Hejtmancik JF Jiao X Li A Sergeev YV Ding X Sharma AK Chan CC Medina I Edwards AO Mutations in KCNJ13 cause autosomal-dominant snowflake vitreoretinal degeneration Am J Hum Genet 2008 82 1 174 180 18179896 14 Ghamari-Langroudi M Digby GJ Sebag JA Millhauser GL Palomino R Matthews R Gillyard T Panaro BL Tough IR Cox HM Denton JS Cone RD G-protein-independent coupling of MC4R to Kir7.1 in hypothalamic neurons Nature 2015 520 7545 94 98 25600267 15 McCloskey C Rada C Bailey E McCavera S van den Berg HA Atia J Rand DA Shmygol A Chan YW Quenby S Brosens JJ Vatish M Zhang J Denton JS Taggart MJ Kettleborough C Tickle D Jerman J Wright P Dale T Kanumilli S Trezise DJ Thornton S Brown P Catalano R Lin N England SK Blanks AM The inwardly rectifying K+ channel Kir7.1 controls uterine excitability throughout pregnancy EMBO Mol Med 2014 6 9 1161 1174 25056913 16 Lewis LM Bhave G Chauder BA Banerjee S Lornsen KA Redha R Fallen K Lindsley CW Weaver CD Denton JS High-throughput screening reveals a small-molecule inhibitor of the renal outer medullary potassium channel and Kir7.1 Mol Pharmacol 2009 76 5 1094 1103 19706730 17 Wright PD Kanumilli S Tickle D Cartland J Bouloc N Dale T Tresize DJ McCloskey C McCavera S Blanks AM Kettleborough C Jerman JC A High-Throughput Electrophysiology Assay Identifies Inhibitors of the Inwardly Rectifying Potassium Channel Kir7.1 J Biomol Screen 2015 18 Raphemot R Lonergan DF Nguyen TT Utley T Lewis LM Kadakia R Weaver CD Gogliotti R Hopkins C Lindsley CW Denton JS Discovery, characterization, and structure-activity relationships of an inhibitor of inward rectifier potassium (Kir) channels with preference for Kir2.3, Kir3.x, and Kir7.1 Front Pharmacol 2011 2 75 22275899 19 Raphemot R Rouhier MF Hopkins CR Gogliotti RD Lovell KM Hine RM Ghosalkar D Longo A Beyenbach KW Denton JS Piermarini PM Eliciting renal failure in mosquitoes with a small-molecule inhibitor of inward-rectifying potassium channels PLoS One 2013 8 5 e64905 23734226 20 Raphemot R Rouhier MF Swale DR Days E Weaver CD Lovell KM Konkel LC Engers DW Bollinger SF Hopkins C Piermarini PM Denton JS Discovery and Characterization of a Potent and Selective Inhibitor of Aedes aegypti Inward Rectifier Potassium Channels PLoS One 2014 9 11 e110772 25375326 21 Raphemot R Kadakia RJ Olsen ML Banerjee S Days E Smith SS Weaver CD Denton JS Development and validation of fluorescence-based and automated patch clamp-based functional assays for the inward rectifier potassium channel Kir4.1 Assay Drug Dev Technol 2013 11 9–10 532 543 24266659 22 Raphemot R Swale DR Dadi PK Jacobson DA Cooper P Wojtovich AP Banerjee S Nichols C Denton JS Direct Activation of beta-cell KATP Channels with a Novel Xanthine Derivative Mol Pharmacol 2014 23 Swale DR Sheehan JH Banerjee S Husni AS Nguyen TT Meiler J Denton JS Computational and Functional Analyses of a Small-Molecule Binding Site in ROMK Biophys J 2015 108 5 1094 1103 25762321 24 Wang HR Wu M Yu H Long S Stevens A Engers DW Sackin H Daniels JS Dawson ES Hopkins CR Lindsley CW Li M McManus OB Selective inhibition of the Kir2 family of inward rectifier potassium channels by a small molecule probe: the discovery, SAR, and pharmacological characterization of ML133 ACS Chem Biol 2011 6 8 845 856 21615117 25 Furutani K Ohno Y Inanobe A Hibino H Kurachi Y Mutational and in silico analyses for antidepressant block of astroglial inward-rectifier Kir4.1 channel Mol Pharmacol 2009 75 6 1287 1295 19264848 26 de Boer TP Nalos L Stary A Kok B Houtman MJ Antoons G van Veen TA Beekman JD de Groot BL Opthof T Rook MB Vos MA van der Heyden MA The anti-protozoal drug pentamidine blocks Kir2.x-mediated inward rectifier current by entering the cytoplasmic pore region of the channel Br J Pharmacol 2010 159 7 1532 1541 20180941 27 Rodriguez-Menchaca AA Navarro-Polanco RA Ferrer-Villada T Rupp J Sachse FB Tristani-Firouzi M Sanchez-Chapula JA The molecular basis of chloroquine block of the inward rectifier Kir2.1 channel Proc Natl Acad Sci U S A 2008 105 4 1364 1368 18216262 28 Bhave G Chauder BA Liu W Dawson ES Kadakia R Nguyen TT Lewis LM Meiler J Weaver CD Satlin LM Lindsley CW Denton JS Development of a selective small-molecule inhibitor of Kir1.1, the renal outer medullary potassium channel Mol Pharmacol 2011 79 1 42 50 20926757 29 Raphemot R Weaver CD Denton JS High-throughput Screening for Small-molecule Modulators of Inward Rectifier Potassium Channels J Vis Exp 2013 71 30 Tao X Avalos JL Chen J MacKinnon R Crystal structure of the eukaryotic strong inward-rectifier K+ channel Kir2.2 at 3.1 A resolution Science 2009 326 5960 1668 1674 20019282 31 Chemical Computing Group, I Molecular Operating Environment (MOE) 2015 Canada Montreal, QC

PMC005xxxxxx/PMC5131541.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 9802571 20730 Mol Cell Mol. Cell Molecular cell 1097-2765 1097-4164 27840029 5131541 10.1016/j.molcel.2016.10.008 NIHMS827732 Article The Histone Modification Domain of Paf1 Complex Subunit Rtf1 Directly Stimulates H2B Ubiquitylation Through an Interaction with Rad6 Van Oss S. Branden 17 Shirra Margaret K. 17 Bataille Alain R. 27 Wier Adam D. 1 Yen Kuangyu 23 Vinayachandran Vinesh 2 Byeon In-Ja L. 4 Cucinotta Christine E. 1 Héroux Annie 5 Jeon Jongcheol 6 Kim Jaehoon 6 VanDemark Andrew P. 1 Pugh B. Franklin 2 Arndt Karen M. 1* 1 Department of Biological Sciences, University of Pittsburgh, Pittsburgh, PA 15260 2 Center for Eukaryotic Gene Regulation, Pennsylvania State University, University Park, PA, 16802 3 Department of Developmental Biology, Southern Medical University, Guangzhou, China 4 Department of Structural Biology, University of Pittsburgh School of Medicine, Pittsburgh, PA, 15260 5 Department of Biology, Brookhaven National Laboratory, Upton, NY 11973 6 Department of Biological Sciences, Korea Advanced Institute of Science and Technology, Daejeon 34141, South Korea 7 Co-first author * Lead Contact: arndt@pitt.edu 5 11 2016 10 11 2016 17 11 2016 17 11 2017 64 4 815825 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. Summary The five-subunit yeast Paf1 Complex (Paf1C) regulates all stages of transcription and is critical for the monoubiquitylation of histone H2B (H2Bub), a modification that broadly influences chromatin structure and eukaryotic transcription. Here we show that the histone modification domain (HMD) of Paf1C subunit Rtf1 directly interacts with the ubiquitin conjugase Rad6 and stimulates H2Bub independently of transcription. We present the crystal structure of the Rtf1 HMD and use site-specific, in vivo crosslinking to identify a conserved Rad6 interaction surface. Utilizing ChIP-exo analysis, we define the localization patterns of the H2Bub machinery at high resolution and demonstrate the importance of Paf1C in targeting the Rtf1 HMD, and thereby H2Bub, to its appropriate genomic locations. Finally, we observe HMD-dependent stimulation of H2Bub in a transcription-free, reconstituted in vitro system. Taken together, our results argue for an active role for Paf1C in promoting H2Bub and ensuring its proper localization in vivo. Graphical Abstract Introduction Eukaryotic transcription is regulated by dynamic changes in chromatin, which include conserved post-translational modifications of the core histones that comprise the protein component of the nucleosome. The consequences of these modifications have been the subject of intense study. One modification of particular interest is the monoubiquitylation (ub) of a lysine (K) residue on the C-terminal helix of histone H2B: K120 in H. sapiens and the orthologous K123 residue in S. cerevisiae. This modification is important for preventing changes in transcription that are associated with tumorigenesis (Shema et al., 2008), defects in differentiation of stem cells (Fuchs et al., 2012; Karpiuk et al., 2012), and improper development (Zhu et al., 2005). H2B is ubiquitylated co-transcriptionally and levels of H2Bub correlate with RNA polymerase II (Pol II) elongation rates (Fuchs et al., 2014; Kim et al., 2009). Despite its enrichment at regions of active transcription (Minsky et al., 2008), H2Bub is associated with both gene activation and repression (Batta et al., 2011; Mutiu et al., 2007; Shema et al., 2008; Zhang et al., 2005). Decreased nucleosome occupancy at transcribed genes in a yeast strain lacking H2B K123ub indicates that this modification promotes nucleosome stability in vivo, most likely in conjunction with the FACT histone chaperone complex (Batta et al., 2011; Fleming et al., 2008). Beyond any direct influence that the ubiquitin moiety exerts on chromatin structure, H2Bub is also required for di- and trimethylation (Me2/Me3) of H3 K4 and H3 K79, marks associated with active chromatin and required for proper silencing and establishment of heterochromatic regions (Krogan et al., 2002; Nguyen and Zhang, 2011; Smolle and Workman, 2013). The ubiquitin conjugase (E2) Rad6 is required for H2B K123ub in yeast, and Rad6 homologues function similarly in other organisms, including humans (Kim et al., 2009; Robzyk et al., 2000). Rad6 promiscuously mono- and polyubiquitylates free histones and nucleosomes in vitro (Haas et al., 1988; Kim and Roeder, 2009; Sung et al., 1988); however, in the cell, the ubiquitin ligase (E3) Bre1, as well as the Bre1-associated protein Lge1, are required to monoubiquitylate the target lysine on H2B (Hwang et al., 2003; Wood et al., 2003a). Removal of H2B K123ub is carried out by the ubiquitin-specific proteases Ubp8 and Ubp10, which target distinct cellular pools of the mark (Schulze et al., 2011). In addition to the enzymes that deposit or remove the mark, H2Bub is regulated by the nucleosome itself (Basnet et al., 2014; Cucinotta et al., 2015; Morgan et al., 2016; Zheng et al., 2010) and by additional factors. Prominent among these is the multi-functional Paf1C. Paf1C promotes H2Bub in vivo (Ng et al., 2003a; Wood et al., 2003b; Xiao et al., 2005), and this function is dependent on a small, conserved domain within the Rtf1 subunit, both in yeast (Piro et al., 2012; Warner et al., 2007) and in higher eukaryotes (Cao et al., 2015). This domain, termed the Rtf1 histone modification domain (hereafter, HMD), is required for H2B K123ub in S. cerevisiae (Tomson et al., 2011; Warner et al., 2007), and expression of the HMD is sufficient to restore global H2B K123ub in a yeast strain deleted of the endogenous RTF1 gene (Piro et al., 2012). Despite these advances and studies showing a role for Paf1C in recruiting various histone modifiers to genes (Chu et al., 2007; Krogan et al., 2003; Ng et al., 2003b; Xiao et al., 2005), the mechanism by which Paf1C facilitates H2Bub has remained unclear. Here we use a range of approaches to investigate this question. Our results support an active role for Paf1C in promoting H2Bub and provide mechanistic insight into a histone modification with important connections to gene expression, the establishment of genome-wide histone modification patterns, and human health. Results In Vivo Site-Specific Crosslinking Reveals a Direct Interaction Between the Rtf1 HMD and Rad6 While a requirement for Paf1C in promoting H2B K123ub has long been appreciated, mechanistic insight into the role of Paf1C in this process has been lacking. We previously demonstrated that a fragment of S. cerevisiae Rtf1 containing residues 63-152 (HMD63-152) restores bulk H2B K123ub in rtf1Δ strains, suggesting that Rtf1 stimulates H2B K123ub primarily through the activity of the HMD (Piro et al., 2012). In support of this, Rtf1 is the only Paf1C subunit strictly required for detection of this modification in vivo (Figure S1A). We thus focused on the mechanism by which the Rtf1 HMD promotes H2B K123ub. We hypothesized that the HMD might promote H2B K123ub through an interaction with another protein. Given the dynamic nature of protein ubiquitylation, we used crosslinking experiments to detect HMD-interacting proteins in vivo. We first identified functionally important residues within the HMD to inform the placement of a crosslinking-competent amino acid analog. Guided by structural data (see below) and sequence conservation (Figure S2), we performed extensive mutagenesis of RTF1 and assessed the effects of substitutions within the HMD on bulk H2B K123ub levels. Substitutions E104A and R110E, as well as the previously identified E104K substitution (Tomson et al., 2011), eliminated detectable H2B K123ub, while substitutions D91A, E100A, L107A, F108A, and R110A greatly reduced the mark (p<0.05) (Figures 1A and S1B). We observed a strong correlation between relative H2B K123ub and both H3 K4Me3 and H3 K79Me2/3 levels, while H2B K123ub was less well correlated with H3 K4Me2 (Figure S3). We also found that the substitutions within the HMD that diminish H2B K123ub resulted in derepression of a telomeric reporter gene, further confirming the functional importance of these residues (Figure S1C). Informed by these results, we employed an in vivo crosslinking strategy that makes use of the non-natural, photo-reactive phenylalanine analog p-benzoyl-L-phenylalanine (BPA). When grown in the presence of BPA, cells carrying a plasmid expressing an engineered aminoacyl-tRNA synthetase and a tRNA that recognizes the amber (UAG) codon will incorporate BPA into the elongating polypeptide chain by amber suppression (Chin et al., 2003). We created nine full-length HSV-tagged Rtf1 derivatives and five HSV-tagged HMD63-152 derivatives by introducing an amber codon at various positions within the HMD. The fourteen derivatives were expressed only when BPA was present, confirming its incorporation into the polypeptide chain (Figures 1B and 1C). BPA was substituted for residues that, when altered, did not impair the HMD’s ability to promote H2B K123ub but were near functionally important residues (Figure 1A). One exception was F108. Although the F108A substitution nearly abolishes H2B K123ub, we found that, as at other sites, replacement of F108 with BPA restored H2B K123ub-dependent histone modifications (Figure S1D). When exposed to UV light, proteins containing BPA form crosslinks with a short linker distance to the side chains of nearby amino acids (Dormán and Prestwich, 1994), leading to the identification of proteins likely to be in direct contact. Immunoblotting revealed multiple high molecular weight bands for each of the Rtf1 and HMD63-152 BPA derivatives (Figures 1B and 1C). These bands were BPA- and UV-dependent and were absent from cells carrying the corresponding wild-type RTF1 plasmid, indicating that they represent BPA crosslinking between the HMD and directly interacting proteins. Two likely candidates for proteins that interact with Rtf1 through the HMD were Rad6 and Bre1, the E2 and E3 for H2B K123ub. To test for an interaction with Rad6, we performed BPA crosslinking experiments in a strain where Rad6 is Myc-tagged. Probing for Myc revealed a high molecular weight band exclusively for Rtf1 and HMD63-152 derivatives in which BPA was incorporated in place of T105, F108, or Q112 (Figures 2A and 2B). This band was not detected in the wild-type controls or when the inactivating E104K substitution was introduced into the Rtf1 derivatives that contained BPA at positions 108 or 112 (Figure 2A; BPA derivative 105 was not tested). The specific detection of this band in both the full-length and HMD-only constructs under conditions of BPA crosslinking at the same three residues, and not in the E104K mutant context, strongly suggests that Rtf1 directly contacts Rad6 through the HMD and that this interaction is required for the establishment of H2B K123ub. Using a similar strategy, we did not reliably detect crosslinking to Bre1 with any of the 14 BPA-containing Rtf1 or HMD derivatives. However, our data do not exclude interactions between Bre1 or Rad6 with other Paf1C subunits or regions within Rtf1 outside the HMD. To test if the Rtf1-Rad6 interaction is dependent on Bre1, we repeated the experiment in a bre1Δ background. Interestingly, for the HMD63-152 BPA derivatives, the extent of crosslinking to Rad6 was unaffected by the absence of Bre1 (Figure 2C). For the full-length Rtf1 BPA derivatives, we observed a substantial reduction in the level of the crosslinked species, which may suggest that regions in Rtf1 outside the HMD impose a requirement for Bre1 in facilitating the interaction with Rad6 (Figure 2D). This interpretation is complicated somewhat by the partial reduction in Rtf1 protein levels observed in the bre1Δ background; however, Rad6 levels are unaffected by the absence of Bre1 (Figure 2D). While the physical interaction between the HMD and Rad6 does not seem to require Bre1, H2B K123ub and downstream modifications remain dependent on Bre1 in strains expressing HMD63-152 as their sole source of Rtf1 (Piro et al., 2012) (Figure S1E). Structural Analysis of the Rtf1 HMD Identifies a Conserved Interaction Surface for Rad6 Based on sequence conservation and domain mapping (Figure S2) and as a first step in obtaining structural data on the HMD, we performed NMR analysis on a larger HMD fragment containing residues 57–184. Preliminary assignments of this domain indicated a structured region containing residues 74-139, while residues 140–184 appeared poorly ordered (Figures S4A and S4B). We therefore asked whether a minimal 66-amino acid HMD construct, consisting only of residues 74-139 tagged with a Myc epitope and a nuclear localization sequence, was capable of promoting H2B K123ub. When expressed in an rtf1Δ background, HMD74-139 restored global H2B K123ub levels and dependent H3 methylation to the same extent as Rtf1 and HMD63-152 (Figures 3A and S5A). Remarkably, expression of either HMD74-139 or HMD63-152 in a quintuple deletion strain lacking all endogenous Paf1C genes also rescued H2B K123ub levels (Figure 3B, compare lanes 4 and 6 with lane 1) and, to a substantial extent, H3 K79Me2/3 levels (Figure S5B). H3 K4Me2 and H3 K4Me3 levels, however, were only weakly restored. Therefore, although other members of the complex are important for optimal H3 K4 and K79 methylation, a 66 amino acid domain of Rtf1 can bypass the requirement of all five subunits of Paf1C in supporting global H2B K123ub. Having confirmed the function of HMD74-139 in vivo, we used X-ray crystallography to determine its structure at high resolution. Crystals were obtained by incorporating strategic amino acid substitutions, initially a R124A/R126A/R128A triple substitution (3R) to reduce overall charge and side chain entropy. Since this mutant protein had greatly reduced activity in vivo (Figure S1B), we generated and crystallized a single substitution derivative, R126A. Importantly, this protein retained partial function in promoting H2B K123ub (Figure 1A). The structures were refined at 1.4 Å (3R) and 1.6 Å (R126A) resolution, respectively (Table 1). With the exception of the substituted residues, the structures of HMD74-139-R126A and 3R are nearly identical (r.m.s.d = 0.31 Å over 52 Cα atoms), and we confine our discussion to the R126A structure. The HMD is composed of two helices, helix A (residues 89–97) and helix B (residues 100–127), preceded by a loop region (residues 75–88) that interacts with both helices and completes the fold (Figure 3C). The R126A position makes numerous interactions with neighboring proteins within the crystal, explaining why this substitution facilitated crystallization (Figure S4C). Mapping of sequence conservation from an alignment of fungal Rtf1 proteins (Figure S2) onto the surface of the HMD revealed two highly conserved regions. One contains E100, R103, E104 and L107 while the second is formed by E83, G84, and K85 (Figure 3D). This result suggested two prominent features within the HMD that might be responsible for Rad6 binding and H2B K123ub activity. Locations of HMD residues that supported BPA-mediated crosslinking to Rad6 (T105, F108, and Q112) are found on one face of helix B and adjacent to the highly conserved surface containing E104 (Figure 3E). Residues T105, F108, and Q112 also play prominent roles in an antiparallel helical packing arrangement between the two HMD molecules within the asymmetric unit (Figure 3F). This packing buries 1,504 Å2 of surface area (Figure S4C) and contains significant hydrophobic and hydrophilic interactions. While it is unclear whether this packing is biologically relevant, our results demonstrate that this surface is functionally important, as numerous residues that are required to maintain proper levels of H2B K123ub and H3 methylation are contained within it (Figures 1A, S1B, and S3). Pol II, Spt4-Spt5, and Paf1C Have Distinct Assembly Positions Along Gene Bodies To investigate the global localization of proteins involved in the H2B K123ub pathway at near single-nucleotide resolution, we conducted ChIP-exo experiments (Rhee and Pugh, 2011). ChIP-exo allows peaks of factor enrichment to be assigned to subnucleosomal positions, such as individual histones or the nucleosomal dyad (Rhee et al., 2014). We first examined the five Paf1C subunits and plotted their averaged locations relative to the first nucleosome (+1) within gene bodies (Figure 4A). All five subunits displayed similar detailed binding patterns, which confirms their presence within the same complex and further cross-validates the use of epitope-tagged proteins. Crosslinking reached local maxima at nucleosome dyads starting at +2 nucleosomes and reached its apex at +3 nucleosomes (note that dyad locations are identifiable as the midpoint between adjacent pairs of H2B peaks). This same +2 and +3 region is where both Pol II and histones undergo transitions in modification states (Mayer et al., 2010; Rhee et al., 2014), and so may be linked in some way. The association of yeast Paf1C with transcribing Pol II is mediated in large part by a central Plus3 domain within Rtf1, which interacts directly with the phosphorylated C-terminal repeat domain of Spt5 within the Spt4-Spt5 complex (Liu et al., 2009; Mayekar et al., 2013; Warner et al., 2007; Wier et al., 2013; Zhou et al., 2009). We determined the genomic localization patterns of Spt4, Spt5, and Pol II (Rpb3 subunit) and compared these profiles to that of Rtf1, as a representative subunit of Paf1C (Figure 4B). As expected, Spt4 and Spt5 displayed similar binding patterns to each other. However, unlike Rtf1, Spt4-Spt5 crosslinking peaked at the same location as the promoter-distal or downstream (in the direction of transcription) H2B within the +1 nucleosome, as well as the promoter-proximal H2B position of the +2 nucleosome. As previously shown, the peak of Pol II crosslinking was just 5’ to the proximal H2B peak, where the transcription start site (TSS) resides, and also within the distal H2B peak of the +1 nucleosome (Rhee and Pugh, 2012b). All factors were strongly enriched at highly transcribed genes and depleted at lowly transcribed genes (Figure 4C and Figure S6A). The localized enrichment of Pol II at the TSS might reflect relatively slow initiation (compared to elongation), but the enrichment over the distal +1 H2B might also reflect an additional slow step in elongation. Whether this indicates physical obstruction by the nucleosome or is a manifestation of checkpoint events is unclear, as any slowing of Pol II would provide more opportunity to establish crosslinks. While it is plausible that Pol II crosslinks through H2B, the highly formaldehyde-reactive properties of single-stranded DNA in the Pol II active site make it the likely primary target. As it moves through the +1 nucleosome region, our data suggest that Pol II picks up Spt4-Spt5, and then further downstream Paf1C is acquired, at least from the perspective of crosslinking to the underlying DNA. Earlier association events could occur that are not evident by formaldehyde crosslinking to DNA. This result supports the previous notion of an ordered recruitment of the Pol II elongation machinery (Mayer et al., 2010), but provides substantially greater positional resolution. Rad6 and Bre1 Colocalize with H2B and Are Partially Dependent on Rtf1 for Their Occupancy at Highly Transcribed Genes Our BPA crosslinking results raised the possibility that Rtf1 might mediate the recruitment of Rad6 to genes or its retention there. We tested this by performing ChIP-exo on Rad6 and Bre1 in strains containing or lacking Rtf1. First, consistent with their central roles in catalyzing H2B K123ub, Rad6 and Bre1 are enriched at both the distal and proximal H2B within genic nucleosomes (Figure 4D). Second, deletion of RTF1 caused modest reductions in the levels of Rad6 and Bre1 across the average gene body (Figure 4D), but larger reductions at highly expressed genes, particularly downstream of the +1 nucleosome (Figure 4E). This agrees with our observation that Paf1C enrichment is also greatest in this region (Figure 4A) and expands upon an earlier study done on the GAL1 gene (Xiao et al., 2005). These results suggest that Rtf1 is important for normal Rad6-Bre1 occupancy in the bodies of active genes, particularly those with the highest levels of expression. Normal H2B K123ub Patterns Are Dependent on an Intact Paf1C Where Paf1C has loaded on a gene, the region should then become enriched for H2Bub. Indeed, the enrichment profile of H2B K123ub was similar to that of Paf1C, with relatively low levels of ubiquitylation at the +1 nucleosome, higher levels at the +2 nucleosome and highest levels at the +3 and +4 nucleosomes (Figure 5A). Consistent with a functional link between H2B K123ub and Pol II/Spt4-Spt5/Rtf1, genes with high levels of H2B K123ub also have high levels of these factors (Figure 5B and Figure S6B). Moreover, regions with the highest H2B K123ub levels also have the highest levels of Rad6, Bre1 and Rtf1 (Figure 5C). Using conventional ChIP, we showed that, when overexpressed in an rtf1Δ strain, HMD63-152 associates with chromatin (Piro et al., 2012). To identify the localization pattern of the HMD at high resolution and to ask if this pattern is dependent on its ability to promote H2B K123ub, we performed ChIP-exo analysis on HSV-HMD63-152 and its inactive E104K derivative. Strikingly, the isolated HMD colocalized with H2B, and this enrichment was eliminated by the E104K substitution, suggesting that chromatin association of the HMD and its ability to promote H2B K123ub are interdependent (Figure 5D). In contrast to the HMD, the localization of full-length Rtf1 was unaffected by the E104K substitution. This is consistent with evidence that domains within Rtf1, outside the HMD, mediate its attachment to Paf1C and the Pol II elongation machinery (Mayekar et al., 2013; Warner et al., 2007) (Figure 5E). The isolated HMD targeted H2B more broadly in gene bodies, rather than being enriched at the +2 nucleosome and nucleosomes further downstream like Rtf1 (compare Figures 5D and 5E). The enhanced enrichment of the HMD at the +1 nucleosome further suggests that its localization to chromatin is no longer governed by Paf1C interactions with Pol II and Spt5. Similar to the localization of the HMD itself, H2B K123ub enrichment in the strain expressing HSV-HMD63-152 revealed a 5’ shift, and in particular, greater enrichment at the +1 nucleosome (Figure 5F). Rtf1 occupancy is strongly biased toward transcribed loci (Figures 4C and 5G), whereas, consistent with the idea that HMD targeting is deregulated when it cannot associate with Paf1C, HMD occupancy is similar at transcribed and non-transcribed regions, with consequent effects on H2B K123ub (Figure 5G). Thus, while ChIP-exo revealed the ability of the HMD to localize to H2B, presumably via its interaction with Rad6, normal H2B K123ub patterns require the coupling of HMD activity to an intact Rtf1 protein in the context of Paf1C. A Recombinant HMD Protein Stimulates Bre1-Dependent H2Bub in a Transcription-Free In Vitro System Though H2B K123ub is undetectable in vivo in the absence of Rtf1 (Figure S1A), H2B can be ubiquitylated in vitro in an ATP-dependent reaction containing recombinant ubiquitylation factors and HeLa nucleosomes (Kim and Roeder, 2009). However, when added to this minimal system, yPaf1C failed to stimulate H2Bub and instead was somewhat inhibitory (Kim and Roeder, 2009). Given the ability of the HMD to promote H2B K123ub in vivo in the absence of Paf1C (Figure 3B), we examined the effect of adding purified HMD to a similar in vitro system using nucleosomes reconstituted from recombinant X. laevis histones. Neither recombinant HMD74-139 nor HMD57-152 affected H2Bub levels when added to the reaction (Figure S7). However, because our bioinformatic analysis identified a larger HMD-containing domain, which included Rtf1 residues 74–184 (Figure S2), we examined the effect of adding HMD74–184 to the reconstituted reaction. Interestingly, the larger Rtf1 fragment stimulated H2Bub by approximately 4-fold in a manner dependent on E104, a residue essential for stimulation of H2B K123ub in vivo (Figure 6A). While Rtf1 residues 140–184 are critical for HMD activity in vitro, HMD74-139 is sufficient to restore global H2B K123ub levels in vivo (Figures 3A and 3B). It is possible that residues 140–184 stabilize the active conformation of the HMD in vitro. Rad6 can ubiquitylate core histones in vitro without Bre1 or any other E3 present (Haas et al., 1988; Kim and Roeder, 2009; Sung et al., 1988). As detected with the antibody against human H2B K120ub, we observed H2Bub of Xenopus nucleosomes after 2 hr in the absence of the E3 (Figure 6B). We found that the ability of the HMD to stimulate H2Bub was dependent on Bre1 (Figure 6B). In fact, HMD74–184 had a modest inhibitory effect on Bre1-independent H2Bub by Rad6 (Figure 6B). These results indicate that the Rtf1 HMD plays a direct role in stimulating H2Bub catalysis in a manner dependent on its interaction with Rad6 and the presence of Bre1. Discussion Previous studies demonstrated a conserved requirement for Paf1C in promoting H2Bub in vivo, but the mechanism underlying this requirement has remained largely unexplored. In this report, we (1) reveal a direct interaction between the Rtf1 HMD and Rad6 through site-specific, in vivo crosslinking; (2) demonstrate that the HMD, consisting of as few as 66 amino acids, can promote H2B K123ub in the absence of all other Paf1C subunits; (3) report the crystal structure of the HMD and identify the interaction surface for Rad6; (4) map the genomic positional organization of Paf1C, Spt4, Spt5, the HMD, H2B K123ub, Rad6 and Bre1 at high resolution; and (5) demonstrate a stimulatory effect of the HMD on H2B ubiquitylation in a reconstituted system. Together, these findings indicate that Paf1C directs the activity of the HMD to active genes and thereby determines the patterning of transcription-linked H2B K123ub across the genome. Using BPA crosslinking, we detected a direct interaction between Rad6 and a conserved, surface of the HMD that contains residues required for proper H2B K123ub in vivo. This interaction is dependent on the Rtf1 E104 residue, an amino acid that is critical for Rtf1’s role in promoting H2B K123ub and is invariant across 73 fungal species. Our inability to reproducibly detect an HMD-Rad6 interaction with standard affinity purification-mass spectrometry methods or far western analysis with recombinant proteins (data not shown) suggests that the interaction is dynamic and/or requires the involvement of other factors. Using recombinant human proteins, a previous study demonstrated an interaction between hPaf1C and hRAD6A or hRAD6B that is mediated by the hBRE1 complex (Kim et al., 2009). While we were unable to convincingly detect crosslinking to Bre1 with any of our Rtf1 BPA derivatives, Bre1 may contact other Paf1C subunits and/or other regions within Rtf1. Alternatively, a weak interaction between Bre1 and the HMD may have been obscured by signal-to-noise issues intrinsic to BPA experiments. In agreement with the former, interactions between human Paf1C subunits and the hBRE1 complex have been detected (Hahn et al., 2012; Kim et al., 2009). In our BPA experiments, deletion of BRE1 did not detectably affect crosslinking between Rad6 and the HMD, when the HMD was expressed outside the context of Paf1C (Figure 2C). This observation suggests that Bre1 is not strictly essential for the HMD-Rad6 interaction. However, in experiments involving an intact Rtf1 (and hence Paf1C), we detected reduced crosslinking between Rad6 and the HMD (Figure 2D), suggesting that regions of Paf1C outside the HMD may regulate the HMD-Rad6 interaction and impart a requirement for Bre1. Independent of its effects on the Rad6-HMD interaction, Bre1 is indispensable for H2B K123ub in vivo, whether cells express the full Paf1C or just the HMD (Figure S1E). Positions within the HMD that crosslinked to Rad6 form a patch along the spine of helix B (Figure 3). Given the relatively small size of BPA, residues engaged in crosslinking are likely to be contained within or at the periphery of an interface between the HMD and Rad6 that may extend from E100 to Q115. Part of this HMD surface is highly conserved, especially E100, R103, and E104, and contains residues critical for H2B K123ub. In the crystal structure, helix B engages in an antiparallel coiled-coil interaction with the helix B of a second HMD molecule (Figure 3F), raising the possibility that Rad6 may interact similarly with helix B. The role of the other conserved patch within the HMD, containing residues E83, G84, and K85, remains unclear but may speak to a separate function. Substitution of E83 or K85 with alanine did not affect H2B K123ub levels, and introduction of BPA at these positions did not reveal crosslinks to Rad6 or Bre1 (data not shown), although these derivatives were competent to crosslink to other, still unidentified proteins. ChIP-exo analysis showed that the five yeast Paf1C subunits share a common localization pattern with significant enrichment at the +2 nucleosome and nucleosomes further 3’ (Figures 4A–C and 6C). This pattern, together with the correlation with gene expression levels, strongly supports an ordered recruitment model in which Pol II arrives first at the coding region, followed by Spt4-Spt5 and Paf1C, which recognizes Spt5 in its phosphorylated state (Mayekar et al., 2013; Mayer et al., 2010; Qiu et al., 2012; Wier et al., 2013). Similar to Paf1C, H2B K123ub levels increase in the 5’ to 3’ direction with peak enrichment at the +3 and +4 nucleosomes (Figures 5A and 6C). The similar patterns of Paf1C and H2B K123ub occupancy support the fundamental role Paf1C plays in establishing this modification. ChIP-exo analysis of the isolated HMD revealed local enrichment at H2B that was not detected in our analysis of full-length Rtf1 or other Paf1C subunits (Figure 5D). The localization pattern of the HMD can be explained by its interaction with Rad6, whose occupancy pattern, along with that of Bre1, also aligns with H2B. We propose that the association between H2B and Rtf1 is transient and therefore not captured by ChIP-exo analysis, as Paf1C is physically and likely kinetically coupled to Pol II. Moreover, unlike full-length Rtf1, the enrichment of the isolated HMD is shifted to include the +1 nucleosome, consistent with its co-localization with Rad6. Enrichment of H2B K123ub is also shifted to include the +1 nucleosome in these strains, supporting the central role of the HMD in stimulating H2B K123ub (Figures 5F and 6C). This correlation between HMD and H2B K123ub localization suggests that, in wild type cells, normal H2B K123ub patterning is governed by the localization of Rtf1, which contains both the HMD and a separate region that tethers it within Paf1C (Warner et al., 2007). The occupancy of Rad6 and Bre1 is only partially dependent on Rtf1 even at the most highly expressed genes (Figures 4D and 4E), yet H2B K123ub is undetectable in an rtf1Δ strain. This argues for a more direct role of Rtf1 in stimulating H2B K123ub catalysis. Earlier work did not reveal a stimulatory effect of Paf1C, either yeast or human forms, in catalyzing H2Bub in reconstituted systems containing HeLa nucleosomes (Kim et al., 2009; Kim and Roeder, 2009), while a requirement for Paf1C was observed in a transcription-coupled system using human factors (Kim et al., 2009). We show that an HMD-containing protein (HMD74-184) can stimulate Bre1-dependent H2Bub in a minimal, transcription-free in vitro system (Figure 6A and 6B). The inability of the E104K derivative to stimulate the reaction supports the functional significance of this result. Our ability to detect stimulation in a purified system may relate to our use of a recombinant substrate devoid of pre-existing marks, or may indicate the existence of regulatory domains within Paf1C that control the function of the HMD within Rtf1. Consistent with the latter, the HMD promotes normal levels of H2B K123ub when all other Paf1C subunits are absent, while expressing Rtf1 in this context only partially restores the mark (Figure 3B). An in vivo association between H2Bub and Pol II transcription is well established (Fuchs and Oren, 2014), and current evidence indicates that Pol II transcription acts as an upstream requirement for the modification (Fuchs et al., 2014). It is likely that Paf1C, through its role as an elongation factor, in part promotes H2B K123ub indirectly via stimulation of transcription (Kim et al., 2010; Tous et al., 2011). However, our data indicate the Rtf1 HMD can also stimulate H2Bub in the absence of transcription. We hypothesize that the HMD serves as a cofactor in the H2B ubiquitylation process by enhancing the ability of Bre1 to direct Rad6 to the appropriate lysine on H2B and/or by stimulating the catalytic ability of Rad6 in the presence of Bre1. This hypothesis does not exclude additional roles for Rtf1 or other Paf1C subunits in facilitating recruitment of Rad6-Bre1, stimulating H2Bub by an effect on elongation, or stabilizing Bre1 at the protein level (Wozniak and Strahl, 2014). Our results provide mechanistic insight into a crucial histone modification pathway. Given that the function of the HMD is conserved from yeast to humans, it will be interesting to investigate the extent to which the HMD is involved in the many processes regulated by Paf1C in higher eukaryotes. Experimental Procedures Materials An antibody against human H2B K120ub (Cell Signaling #5546; 1:1000 dilution) was used to detect H2B K123ub in yeast (Wozniak and Strahl, 2014). Lack of antibody reactivity in H2B K123R and bre1Δ mutant strains confirms specificity of the antibody for the appropriate lysine. Yeast strains, plasmids, and additional antibodies are described in Supplemental Information. Structure Determination Crystals of HMD74-139-3R and HMD74-139-R126A were grown and diffraction data collected to 1.4 and 1.6 Å resolution as described in the Supplemental Experimental Procedures. Phasing of HMD74-139-3R was achieved via SAD phasing of SeMET substituted crystals, while HMD74-139-R126A was solved by molecular replacement. In vivo BPA Crosslinking Cells were grown in the presence of 1 mM BPA (Bachem, F-2800) to early log phase and 10 OD600 units were harvested. Cells were resuspended in 1 mL ddH20, placed in the center of a petri dish, and exposed to UV irradiation (365 nm) for 10 min using a UVGL-55 handheld UV lamp (UVP) placed 2 cm above the cells. TCA extracts were prepared for western analysis (Supplemental Experimental Procedures). To confirm the function of the Rtf1 BPA derivatives, cells were grown and harvested without UV irradiation and extracts were prepared in RIPA buffer (Supplemental Experimental Procedures). In vitro H2Bub Assay The assay was adapted from (Kim and Roeder, 2009) but used recombinant X. laevis nucleosome core particles as a substrate. Amounts of individual proteins are described in Supplemental Experimental Procedures. ChIP-exo Analysis ChIP-exo experiments were carried out as described (Rhee and Pugh, 2012a). Briefly, formaldehyde-crosslinked and sonicated chromatin was immunoprecipitated with antibody-conjugated magnetic beads, followed by DNA polishing, A-tailing, Illumina adaptor ligation (ExA2), and on-beads digestion by lambda and recJ exonuclease. After elution of the single-stranded DNA, a primer was annealed to ExA2 and extended with phi29 DNA polymerase, then A-tailed. A second Illumina sequencing adaptor was ligated to exonuclease treated ends, and the products PCR-amplified and gel-purified. Sequencing was performed on an Illumina NextSeq500 system. Data were analyzed as described in Supplemental Experimental Procedures. Supplementary Material We thank Song Tan for recombinant nucleosomes, Steve Hahn and Linda Warfield for assistance with the BPA experiments, Stefan Brooks and Aubrey Lowen for technical assistance, and Nathan Clark for helpful discussions. We thank Mike McAlear and members of our laboratories for feedback on the manuscript. This work was supported by an NSF Graduate Research Fellowship to S.B.V.O. (DGE-1247842), Andrew Mellon Predoctoral Fellowships to C.E.C. and A.D.W., a University of Pittsburgh CRDF grant to A.P.V., NRF grants to J.K. (2012M3A9B4027956 and 2012M3A9C6049937), and NIH grants (GM052593 to K.M.A.; HG004160 to B.F.P.; P50GM082251 for funding I.-J.L.B.). B.F.P. has a financial interest in Peconic, LLC, which utilizes the ChIP-exo technology implemented in this study and could potentially benefit from the outcomes of this research. Figure 1 BPA Crosslinking Reveals In Vivo Protein Interactions with the Rtf1 HMD (A) Western blot analysis of an rtf1Δ strain transformed with plasmids expressing full-length HA-Rtf1 or derivatives with the indicated substitutions within the HMD. (B and C) Western blot analysis of an rtf1Δ strain transformed with two plasmids: (1) a high-copy plasmid expressing full-length, wild type (WT) HSV-Rtf1 (Panel B) or wild type HSV-HMD63-152 (Panel C), or their respective derivatives in which the indicated codon within the HMD has been replaced with the amber codon; and (2) a plasmid containing genes for a tRNA and tRNA synthetase that incorporate BPA into proteins by amber suppression (pLH157/HIS3). Cells were grown in the presence of BPA and exposed to UV light, as indicated. See also Figures S1 and S3. Figure 2 The HMD Directly Contacts Rad6 (A,B) Western blot analysis of an rtf1Δ RAD6-13XMyc strain transformed with pLH157/HIS3 (see Figure 1 legend) and plasmids expressing WT HSV-Rtf1 (Panel A) or HSV-HMD63-152 (Panel B), or their respective derivatives in which BPA is incorporated at the indicated position. Indicated plasmids also contain the inactivating E104K substitution. All samples were grown in the presence of BPA and exposed to UV light. (C,D) Western blot analysis was performed with samples processed as in (A) and (B), with the indicated HSV-HMD63-152 (Panel C) or HSV-Rtf1 (Panel D) derivatives expressed in an rtf1Δ RAD6-13XMyc strain or an rtf1Δ bre1Δ RAD6-13XMyc strain. Top row shows crosslinked species. Figure 3 A Conserved Surface Within the HMD Mediates H2B K123ub and the Interaction with Rad6 (A, B) Western blot analysis was performed on rtf1Δ and rtf1Δ paf1Δ ctr9Δ cdc73Δ leo1Δ strains transformed with empty vector, a low-copy plasmid expressing full-length HA-Rtf1, or high-copy plasmids expressing full-length Myc-Rtf1, wild type or mutant Myc-HMD63-152, or Myc-HMD74-139. An H2BK123R strain serves as a negative control. (C) Ribbon diagram of HMD74-139. Residues at which substitutions significantly diminished (yellow) or did not diminish (white) H2B K123ub are indicated. Dashes indicate hydrogen bonds. (D) Surface view of HMD74-139 with sequence identity among 73 fungal Rtf1 orthologs mapped onto the surface. Red indicates highest conservation. (E) Positions of residues that facilitate BPA-crosslinking to Rad6 (blue) cluster near E104 (magenta). Residues that failed to support Rad6 crosslinking are indicated in gray. (F) Residues that facilitate crosslinking to Rad6 interact with each other and to the other molecule of the HMD in the asymmetric unit. Molecule A (green) is shown with BPA crosslinking residues shown as blue sticks. E104 is colored magenta. Hydrogen bonding (yellow dashes) and van der Waals (blue dashes) interactions inferred from the structure are shown. See also Figures S2 and S4. Figure 4 ChIP-exo Analysis Indicates Ordered Recruitment for Early Steps in Elongation and Reveals a Role for Rtf1 in Promoting Rad6-Bre1 Occupancy (A–E) Protein localization patterns as determined by ChIP-exo are centered at the +1 nucleosome dyad of coding genes. The gray fill, showing H2B occupancy as determined by ChIP-exo in a wild-type strain, is not to scale and should be used for positioning comparison only. The first three genic nucleosomes are indicated. Experiments were normalized using total tag count; thus the Y-axis represents an arbitrary linear scale for each separate trace and the absolute magnitudes are not comparable between different experiments. All experiments employed strains with epitope-tagged, endogenously expressed proteins (see Strain Table). (C) Rtf1, Rpb3, Spt4, and Spt5 localization at highly and lowly expressed genes (top and bottom quintiles), based on transcription frequency. (D–E) Occupancy of endogenously expressed Rad6 and Bre1 in strains containing or lacking RTF1 is shown for all genes (Panel D) or for genes in the highest and lowest expression quintiles (Panel E). See also Figure S6. Figure 5 Paf1C Targets HMD Function to Its Appropriate Genomic Location (A–F) ChIP-exo data are presented as in Figure 4 and were normalized using total tag count unless otherwise indicated; thus the Y-axis represents an arbitrary linear scale for each separate trace and the absolute magnitudes are not comparable between different experiments. (A) Occupancy of H2B K123ub in a wild-type strain. H2B occupancy was determined in the same strain and is shown to scale. (B) Rtf1, Rpb3, Spt4, and Spt5 as in Figure 4B and 4C for genes with high and low H2B K123ub (top and bottom quintiles), based on H2B K123ub / bp values. (C) Occupancy of Rtf1, Rad6, and Bre1 for genes in the highest and lowest H2B K123ub quintiles. (D) Comparison of plasmid-expressed HSV-HMD63-152 and HSV-HMD63-152-E104K occupancy. To appreciate the difference in levels, datasets were normalized to the nucleosome free region. (E) Comparison of HA-Rtf1 and HA-Rtf1-E104K occupancy. Proteins were expressed from the endogenous RTF1 locus. (F) H2B K123ub localization in an rtf1Δ strain overexpressing either HSV-HMD63-152 (pHMD) or HSV-Rtf1 (pRtf1). H2B occupancy was determined in the same strains. Note that the H2B K123ub patterns are similar in strains containing a chromosomal copy of RTF1 (Figure 5A) or the high copy Rtf1 plasmid (pRtf1). (G) Box plot graphs of Rtf1 / HMD occupancy (left) and H2B K123ub / H2B ratio (right) in a wild-type strain, or in an rtf1Δ strain overexpressing either HSV-HMD63-152 or HSV-Rtf1 from a plasmid, for intergenic regions and transcription units of highly and lowly expressed genes. See also Figure S6. Figure 6 HMD74-184 Stimulates Bre1-dependent H2Bub In Vitro (A and B) Unless otherwise indicated, in vitro ubiquitylation reactions contained: recombinant X. laevis nucleosomes, FLAG-yBre1 (E3), FLAG-hE1, His-HA-pK-Ubiquitin, and yRad6 (E2). HMD74-184, or an equivalent volume of storage buffer, was added to the reactions. Reactions were incubated in 1X Reaction Buffer at 30 °C for the indicated times and analyzed by Western blotting. H2A levels function as a loading control. Relative H2Bub was determined by setting the signal for lane 3 (Panel A) or lane 2 (Panel B) to one. (C) Top: Model showing relative enrichment of Pol II, Spt4-Spt5, Rad6-Bre1, Paf1C, and H2B K123ub along active genes in a wild-type background. Bottom: Model showing localization of the Rtf1 HMD and H2B K123ub in an rtf1Δ background when the HMD is expressed as the sole source of Rtf1 and cannot interact with Paf1C. The localization of Spt4-Spt5 and Rad6-Bre1 in the HMD-only strain has not been directly measured but is inferred from data in Figure 4D and previous work that showed Spt4 recruitment is unaffected in an rtf1Δ strain (Qiu et al., 2006). See also Figure S7. Table 1 Data Collection and Refinement Statistics SeMet R124/126/128A R124/126/128A R126A PDB ID 5EMX 5E8B Data collection   Space group R32:H R32:H R32:H   Cell dimensions     a, b, c (Å) 94.3, 94.3, 76.4 94.3, 94.3, 76.8 93.9, 93.9, 75.3     α, β, γ (°) 90.0, 90.0, 120.0 90.0, 90.0, 120.0 90.0, 90.0, 120.0   Unique Reflections 13524 30553 16,294   Resolution (Å) 50.0 - 1.74 (1.77–1.74)a 50.0 - 1.32 (1.34–1.32) 50.0 - 1.62 (1.65–1.62)   Rmerge (%)b 11.6 (54.7) 5.1 (62.6) 5.7 (50.3)   I / σI 32.57 (2.40) 77.79 (1.30) 43.0(6.77)   Completeness (%) 100.0 (100.0) 98.3 (78.5) 99.6 (100.0)   Redundancy 10.2 (7.2) 13.7 (3.5) 20.1 (15.2) Refinement   Resolution (Å) 14.6–1.40 (1.45–1.40) 28.5–1.62 (1.68–1.62)   Rworkc / Rfreed (%) 14.2 / 17.8 (12.7 / 20.3) 15.1/19.3 (14.8/23.7)   Number of. atoms     Protein 922 912     Water 78 81   B-factors (Å2)     Protein 25.19 32.72     Solvent 38.61 44.63   R.m.s. deviations     Bond lengths (Å) 0.009 0.009     Bond angles (°) 1.004 0.867   Ramachandrian     Outliers (%) 0.00 0.00     Allowed (%) 3.7 3.92     Favored (%) 96.3 96.08 a Values in parentheses are for highest-resolution shell. b Rmerge = (|(ΣI - <I>)|)/(ΣI), where <I> is the average intensity of multiple measurements. c Rwork = Σhkl∥Fobs(hkl)∥ - Fcalc (hkl)∥/Σhkl|Fobs(hkl)|. d Rfree represents the cross-validation R factor for 5% of the reflections against which the model was not refined. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. Accession Number All sequencing data have been deposited in the GEO database under accession number GSE83348. PDB accession numbers are 5EMX and 5E8B for the HMD74-139 3R and R126A structures, respectively. Author contributions Conceptualization: K.M.A, B.F.P, A.P.V., S.B.V.O., M.K.S.; Investigation: S.B.V.O., M.K.S., A.R.B., A.D.W., V.V., I.-J.L.B., A.H., C.E.C.; Formal Analysis: A.R.B., K.Y., A.D.W., I.-J.L.B., A.P.V.; Resources: J.J., J.K.; Writing—original draft: S.B.V.O., K.M.A., B.F.P., A.P.V., A.R.B.; Writing—review & editing: M.K.S, , I.-J.L.B., J.K., A.D.W., C.E.C., K.Y.; Data curation: A.R.B., A.P.V.; Supervision and Funding Acquisition: K.M.A., B.F.P., A.P.V., J.K. References Basnet H Su XB Tan Y Meisenhelder J Merkurjev D Ohgi KA Hunter T Pillus L Rosenfeld MG Tyrosine phosphorylation of histone H2A by CK2 regulates transcriptional elongation Nature 2014 516 267 271 25252977 Batta K Zhang Z Yen K Goffman DB Pugh BF Genome-wide function of H2B ubiquitylation in promoter and genic regions Genes Dev 2011 25 2254 2265 22056671 Cao QF Yamamoto J Isobe T Tateno S Murase Y Chen Y Handa H Yamaguchi Y Characterization of the human transcription elongation factor Rtf1: evidence for nonoverlapping functions of Rtf1 and the Paf1 complex Mol. Cell. Biol 2015 35 3459 3470 26217014 Chin JW Cropp TA Anderson JC Mukherji M Zhang Z Schultz PG An expanded eukaryotic genetic code Science 2003 301 964 967 12920298 Chu Y Simic R Warner MH Arndt KM Prelich G Regulation of histone modification and cryptic transcription by the Bur1 and Paf1 complexes EMBO J 2007 26 4646 4656 17948059 Cucinotta CE Young AN Klucevsek KM Arndt KM The nucleosome acidic patch regulates the H2B K123 monoubiquitylation cascade and transcription elongation in Saccharomyces cerevisiae PLoS Genet 2015 11 e1005420 26241481 Dormán G Prestwich GD Benzophenone photophores in biochemistry Biochemistry 1994 33 5661 5673 8180191 Fleming AB Kao CF Hillyer C Pikaart M Osley MA H2B ubiquitylation plays a role in nucleosome dynamics during transcription elongation Mol. Cell 2008 31 57 66 18614047 Fuchs G Hollander D Voichek Y Ast G Oren M Cotranscriptional histone H2B monoubiquitylation is tightly coupled with RNA polymerase II elongation rate Genome Res 2014 24 1572 1583 25049226 Fuchs G Oren M Writing and reading H2B monoubiquitylation Biochim. Biophys. Acta 2014 1839 694 701 24412854 Fuchs G Shema E Vesterman R Kotler E Wolchinsky Z Wilder S Golomb L Pribluda A Zhang F Haj-Yahya M RNF20 and USP44 regulate stem cell differentiation by modulating H2B monoubiquitylation Mol. Cell 2012 46 662 673 22681888 Haas AL Bright PM Jackson VE Functional diversity among putative E2 isozymes in the mechanism of ubiquitin-histone ligation J. Biol. Chem 1988 263 13268 13275 2843516 Hahn MA Dickson KA Jackson S Clarkson A Gill AJ Marsh DJ The tumor suppressor CDC73 interacts with the ring finger proteins RNF20 and RNF40 and is required for the maintenance of histone 2B monoubiquitination Hum. Mol. Genet 2012 21 559 568 22021426 Hwang WW Venkatasubrahmanyam S Ianculescu AG Tong A Boone C Madhani HD A conserved RING finger protein required for histone H2B monoubiquitination and cell size control Mol. Cell 2003 11 261 266 12535538 Karpiuk O Najafova Z Kramer F Hennion M Galonska C König A Snaidero N Vogel T Shchebet A Begus-Nahrmann Y The histone H2B monoubiquitination regulatory pathway is required for differentiation of multipotent stem cells Mol. Cell 2012 46 705 713 22681891 Kim J Guermah M McGinty RK Lee JS Tang Z Milne TA Shilatifard A Muir TW Roeder RG RAD6-Mediated transcription-coupled H2B ubiquitylation directly stimulates H3K4 methylation in human cells Cell 2009 137 459 471 19410543 Kim J Guermah M Roeder RG The human PAF1 complex acts in chromatin transcription elongation both independently and cooperatively with SII/TFIIS Cell 2010 140 491 503 20178742 Kim J Roeder RG Direct Bre1-Paf1 complex interactions and RING finger-independent Bre1-Rad6 interactions mediate histone H2B ubiquitylation in yeast J. Biol. Chem 2009 284 20582 20592 19531475 Krogan NJ Dover J Khorrami S Greenblatt JF Schneider J Johnston M Shilatifard A COMPASS, a histone H3 (Lysine 4) methyltransferase required for telomeric silencing of gene expression J. Biol. Chem 2002 277 10753 10755 11805083 Krogan NJ Dover J Wood A Schneider J Heidt J Boateng MA Dean K Ryan OW Golshani A Johnston M The Paf1 complex is required for histone H3 methylation by COMPASS and Dot1p: linking transcriptional elongation to histone methylation Mol. Cell 2003 11 721 729 12667454 Liu Y Warfield L Zhang C Luo J Allen J Lang WH Ranish J Shokat KM Hahn S Phosphorylation of the transcription elongation factor Spt5 by yeast Bur1 kinase stimulates recruitment of the PAF complex Mol. Cell. Biol 2009 29 4852 4863 19581288 Mayekar MK Gardner RG Arndt KM The recruitment of the Saccharomyces cerevisiae Paf1 complex to active genes requires a domain of Rtf1 that directly interacts with the Spt4-Spt5 complex Mol. Cell. Biol 2013 33 3259 3273 23775116 Mayer A Lidschreiber M Siebert M Leike K Söding J Cramer P Uniform transitions of the general RNA polymerase II transcription complex Nat. Struct. Mol. Biol 2010 17 1272 1278 20818391 Minsky N Shema E Field Y Schuster M Segal E Oren M Monoubiquitinated H2B is associated with the transcribed region of highly expressed genes in human cells Nat. Cell Biol 2008 10 483 488 18344985 Morgan MT Haj-Yahya M Ringel AE Bandi P Brik A Wolberger C Structural basis for histone H2B deubiquitination by the SAGA DUB module Science 2016 351 725 728 26912860 Mutiu AI Hoke SM Genereaux J Liang G Brandl CJ The role of histone ubiquitylation and deubiquitylation in gene expression as determined by the analysis of an HTB1(K123R) Saccharomyces cerevisiae strain Mol. Genet. Genomics 2007 277 491 506 17447102 Ng HH Dole S Struhl K The Rtf1 component of the Paf1 transcriptional elongation complex is required for ubiquitination of histone H2B J. Biol. Chem 2003a 278 33625 33628 12876293 Ng HH Robert F Young RA Struhl K Targeted recruitment of Set1 histone methylase by elongating Pol II provides a localized mark and memory of recent transcriptional activity Mol. Cell 2003b 11 709 719 12667453 Nguyen AT Zhang Y The diverse functions of Dot1 and H3K79 methylation Genes Dev 2011 25 1345 1358 21724828 Piro AS Mayekar MK Warner MH Davis CP Arndt KM Small region of Rtf1 protein can substitute for complete Paf1 complex in facilitating global histone H2B ubiquitylation in yeast Proc. Natl. Acad. Sci. USA 2012 109 10837 10842 22699496 Qiu H Hu C Gaur NA Hinnebusch AG Pol II CTD kinases Bur1 and Kin28 promote Spt5 CTR-independent recruitment of Paf1 complex EMBO J 2012 31 3494 3505 22796944 Qiu H Hu C Wong CM Hinnebusch AG The Spt4p subunit of yeast DSIF stimulates association of the Paf1 complex with elongating RNA polymerase II Mol Cell Biol 2006 26 3135 3148 16581788 Rhee HS Bataille AR Zhang L Pugh BF Subnucleosomal structures and nucleosome asymmetry across a genome Cell 2014 159 1377 1388 25480300 Rhee HS Pugh BF Comprehensive genome-wide protein-DNA interactions detected at single-nucleotide resolution Cell 2011 147 1408 1419 22153082 Rhee HS Pugh BF ChIP-exo method for identifying genomic location of DNA-binding proteins with near-single-nucleotide accuracy Curr. Protoc. Mol. Biol. Chapter 2012a 21 Unit 21 24 Rhee HS Pugh BF Genome-wide structure and organization of eukaryotic pre-initiation complexes Nature 2012b 483 295 301 22258509 Robzyk K Recht J Osley MA Rad6-dependent ubiquitination of histone H2B in yeast Science 2000 287 501 504 10642555 Schulze JM Hentrich T Nakanishi S Gupta A Emberly E Shilatifard A Kobor MS Splitting the task: Ubp8 and Ubp10 deubiquitinate different cellular pools of H2BK123 Genes Dev 2011 25 2242 2247 22056669 Shema E Tirosh I Aylon Y Huang J Ye C Moskovits N Raver-Shapira N Minsky N Pirngruber J Tarcic G The histone H2B–specific ubiquitin ligase RNF20/hBRE1 acts as a putative tumor suppressor through selective regulation of gene expression Genes Dev 2008 22 2664 2676 18832071 Smolle M Workman JL Transcription-associated histone modifications and cryptic transcription Biochim. Biophys. Acta 2013 1829 84 97 22982198 Sung P Prakash S Prakash L The RAD6 protein of Saccharomyces cerevisiae polyubiquitinates histones, and its acidic domain mediates this activity Genes Dev 1988 2 1476 1485 2850263 Tomson BN Davis CP Warner MH Arndt KM Identification of a role for histone H2B ubiquitylation in noncoding RNA 3’-end formation through mutational analysis of Rtf1 in Saccharomyces cerevisiae Genetics 2011 188 273 289 21441211 Tous C Rondón AG Garcia-Rubio M Gonzalez-Aguilera C Luna R Aguilera A A novel assay identifies transcript elongation roles for the Nup84 complex and RNA processing factors EMBO J 2011 30 1953 1964 21478823 Warner MH Roinick KL Arndt KM Rtf1 is a multifunctional component of the Paf1 complex that regulates gene expression by directing cotranscriptional histone modification Mol. Cell. Biol 2007 27 6103 6115 17576814 Wier AD Mayekar MK Héroux A Arndt KM VanDemark AP Structural basis for Spt5-mediated recruitment of the Paf1 complex to chromatin Proc. Natl. Acad. Sci. USA 2013 110 17290 17295 24101474 Wood A Krogan NJ Dover J Schneider J Heidt J Boateng MA Dean K Golshani A Zhang Y Greenblatt JF Bre1, an E3 ubiquitin ligase required for recruitment and substrate selection of Rad6 at a promoter Mol. Cell 2003a 11 267 274 12535539 Wood A Schneider J Dover J Johnston M Shilatifard A The Paf1 complex is essential for histone monoubiquitination by the Rad6-Bre1 complex, which signals for histone methylation by COMPASS and Dot1p J. Biol. Chem 2003b 278 34739 34742 12876294 Wozniak GG Strahl BD Catalysis-dependent stabilization of Bre1 fine-tunes histone H2B ubiquitylation to regulate gene transcription Genes Dev 2014 28 1647 1652 25085417 Xiao T Kao CF Krogan NJ Sun ZW Greenblatt JF Osley MA Strahl BD Histone H2B ubiquitylation is associated with elongating RNA polymerase II Mol. Cell. Biol 2005 25 637 651 15632065 Zhang X Kolaczkowska A Devaux F Panwar SL Hallstrom TC Jacq C Moye-Rowley WS Transcriptional regulation by Lge1p requires a function independent of its role in histone H2B ubiquitination J. Biol. Chem 2005 280 2759 2770 15528187 Zheng S Wyrick JJ Reese JC Novel trans-tail regulation of H2B ubiquitylation and H3K4 methylation by the N terminus of histone H2A Mol. Cell. Biol 2010 30 3635 3645 20498280 Zhou K Kuo WH Fillingham J Greenblatt JF Control of transcriptional elongation and cotranscriptional histone modification by the yeast BUR kinase substrate Spt5 Proc. Natl. Acad. Sci. USA 2009 106 6956 6961 19365074 Zhu B Zheng Y Pham AD Mandal SS Erdjument-Bromage H Tempst P Reinberg D Monoubiquitination of human histone H2B: the factors involved and their roles in HOX gene regulation Mol. Cell 2005 20 601 611 16307923

PMC005xxxxxx/PMC5131564.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 101465400 34171 Sci Signal Sci Signal Science signaling 1945-0877 1937-9145 25650441 5131564 10.1126/scisignal.2005719 NIHMS695410 Article Cardiac hypertrophy induced by active Raf depends on Yorkie-mediated transcription Yu Lin Daniels Joseph P. Wu Huihui Wolf Matthew J. * Department of Medicine, Duke University Medical Center, Durham, NC 27710, USA * Corresponding author. matthew.j.wolf@dm.duke.edu 10 12 2015 3 2 2015 03 2 2015 01 12 2016 8 362 ra13ra13 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. Organ hypertrophy can result from enlargement of individual cells or from cell proliferation or both. Activating mutations in the serine-threonine kinase Raf cause cardiac hypertrophy and contribute to Noonan syndrome in humans. Cardiac-specific expression of activated Raf also causes hypertrophy in Drosophila melanogaster. We found that Yorkie (Yki), a transcriptional coactivator in the Hippo pathway that regulates organ size, is required for Raf-induced cardiac hypertrophy in flies. Although aberrant activation of Yki orthologs stimulates cardiac hyperplasia in mice, cardiac-specific expression of an activated mutant form of Yki in fruit flies caused cardiac hypertrophy without hyperplasia. Knockdown of Yki caused cardiac dilation without loss of cardiomyocytes and prevented Raf-induced cardiac hypertrophy. In flies, Yki-induced cardiac hypertrophy required the TEA domain–containing transcription factor Scalloped, and, in mammalian cells, expression of mouse RafL613V, an activated form of Raf with a Noonan syndrome mutation, increased Yki-induced Scalloped activity. Furthermore, overexpression of Tgi (a Tondu domain–containing Scalloped-binding corepressor) in the fly heart abrogated Yki- or Raf-induced cardiac hypertrophy. Thus, crosstalk between Raf and Yki occurs in the heart and can influence Raf-mediated cardiac hypertrophy. INTRODUCTION Pathophysiologic stimuli, such as pressure overload or heritable mutations in signaling molecules, can cause cardiac hypertrophy, characterized by an increase in cardiomyocyte size. Increased cardiomyocyte size causes the walls of the heart to thicken, decreasing the size of the cardiac chamber, and potentially restricting the heart’s ability to pump blood (1). Untreated, cardiac hypertrophy leads to heart failure, which can be fatal. Activation of receptor tyrosine kinases (RTKs) by extracellular growth factors and downstream signaling mediated by GTPase (guanosine triphosphatase), Ras, and the serine and threonine kinase Raf contribute to cardiac hypertrophy (2). Heritable mutations that activate the Ras-Raf pathway, collectively called rasopathies, are responsible for various syndromes associated with cardiac hypertrophy (3). For example, mutations in human RAF1 that lead to the substitution of valine for leucine at amino acid 613 are associated with Noonan syndrome, which is characterized in part by cardiac hypertrophy (3). Knock-in mice for the mutation encoding Raf1L613V have symptoms consistent with Noonan syndrome, including cardiac hypertrophy (4). Additionally, heterozygous Raf1L613V mice have enhanced mitogen-activated protein kinase kinase 1 (MEK1) and extracellular signal–regulated kinase 1 (ERK1) and ERK2 signaling, and treatment of these mice with MEK inhibitors attenuates many phenotypic abnormalities, including cardiac hypertrophy (4). The adult fly heart is a tubular structure with circumferentially oriented contractile fibers located along the dorsal aspect of the abdomen (5–7). Fifty-two pairs of cardiomyocytes, which can be identified by the expression of the gene tinC, comprise the embryonic heart. Although the larval heart undergoes major morphological changes to become the adult heart, the total number of cardiomyocytes remains constant (6, 8). Within each abdominal segment, denoted A1 to A6, the heart has eight tinC-expressing myocytes that are laterally paired and comprise the adult fly heart, which has a thickness of single-cell layer. Similar to mammals, cardiac-specific expression of activated human Raf in Drosophila decreases the size of cardiac lumens, increases heart wall thicknesses without inducing hyperplasia, and causes abnormalities in cardiomyocyte contractile fibers (9). Similar to pharmacological inhibition of MEK in mice, the cardiac-specific knockdown of MEK or ERK in flies rescues Raf-mediated hypertrophy (9). However, cardiac-specific expression of activated Drosophila ERK (ERKD334N), which promotes hyperplasia in noncardiac tissues, does not cause cardiomyocyte hypertrophy (9), suggesting that signals downstream from active Raf and MEK, in addition to ERK, are necessary to drive cardiomyocyte growth. The evolutionarily conserved Hippo signaling pathway also regulates cardiac growth (10–12). Hippo (MST1 and MST2 in mammals) is a serine and threonine kinase that is activated by low cell density or mechanical strain in epithelial tissues (13–15). In a complex with the scaffolding protein Salvador, Hippo phosphorylates the kinase Warts (LATS1 and LATS2 in mammals) (16, 17). Phosphorylated Warts phosphorylates the transcriptional coactivator Yorkie (Yki; YAP1 and TAZ in mammals), leading to the inhibition of Yorkie-induced transcription due to sequestration of phosphorylated Yorkie in the cytosol through binding to 14-3-3 proteins (18, 19). In the absence of phosphorylation, Yorkie binds to the TEA domain–containing DNA binding transcription factor Scalloped (Sd; TEADs in mammals) to promote the transcription of genes involved in cell growth and proliferation (20–22). In mice, genetic ablation of Salvador or transgenic expression of mutant YAP1 encoding a serine-to-alanine substitution that prevents phosphorylation by LATS (YAPS127A) in the mouse heart results in cardiomegaly due to increased cardiomyocyte proliferation during development (23–25). The mutation in YAPS127A prevents phosphorylation of this residue by Lats and cytosolic sequestration, thereby resulting in nuclear localization of YAP (24, 25). RTK signaling interacts with the Hippo pathway. Epidermal growth factor receptor (EGFR)–mediated signals occur through Raf to activate the Hippo and Yki signaling pathway in Drosophila eye discs, wing discs, and brain lobes (26). In mouse and human cells, Raf inhibits MST2 by preventing its dimerization and by recruiting a phosphatase that dephosphorylates key residues involved in activation of MST2. Both of these functions of Raf in mammals require binding to MST2 and are independent of the ERK pathway, including the kinase activity of Raf (28). We found that Raf stimulates Yki to promote cardiac hypertrophy in flies. Using optical coherence tomography (OCT) of cardiac chambers in awake, adult flies and histological analyses of heart wall thicknesses, we found that cardiac-specific expression of activated Yki caused cardiac hypertrophy, whereas cardiac-specific expression of RNA interference (RNAi) directed against yki caused cardiac dilation. Moreover, cardiac-specific knockdown of Yki prevented Raf-mediated cardiac hypertrophy. In mammalian cell culture, activated Raf enhanced the ability of activated Yki to stimulate Sd-dependent transcription, whereas genetic or pharmacological inhibition of the kinase activities of Raf, MEK, or ERK inhibited this effect. Cardiac-specific overexpression of a Tondu domain–containing, Sd-associated corepressor Tgi prevented cardiac hypertrophy in flies induced by expression of activated Raf or activated Yki. In addition, Tgi prevented the ability of activated Raf to augment Yki-induced, Sd-dependent transcription in cell culture. RESULTS Yki controls heart size in Drosophila We investigated the role of Yki in the maintenance of organ size in the fly heart. tinC is specifically expressed in the heart throughout development and in adult flies (29). The amino acid substitutions S168A, or S111A, S168A, and S250A in Yki are equivalent to S127A in murine YAP and prevent phosphorylation by Warts and thereby lead to constitutive activation of Yki (18). Therefore, we used the Gal4-UAS system to express yki transgenes (UAS-ykiS168A or UAS-ykiS111A,S168A,S250A) under the control of the enhancer tinC (tinC-GAL4). We also expressed green fluorescent protein (GFP) from the tinC enhancer (tinC-GFP) and OCT to visualize the beating heart in awake, adult flies. We measured the size of the ventricular diameter at the end of systole (ESD) and diastole (EDD) and calculated the fractional shortening, which is a measure of systolic function (30, 31). Compared to tinC-GFP; tinC-Gal4 flies, tinC-GFP; tinC-Gal4,UAS-ykiS168A (tinC>ykiS168A) flies had smaller EDDs and ESDs (Fig. 1A). Additionally, the heart walls were thicker in tinC>ykiS168A and tinC-GFP; tinC-Gal4,UAS-ykiS111A,S168A,S250A (tinC>ykiS111A,S168A,S250A) flies compared to tinC-GFP; tinC-Gal4 flies (Fig. 1, B and C). Thus, aberrant activation of Yki induces cardiac overgrowth. We also tested if Yki was required for organ size in the adult heart using a double-stranded RNAi transgene targeting yki (UAS-ykiRNAi). Because cardiac tissue is difficult to purify from the ventral longitudinal muscle and surrounding pericardial cells, we tested the ability of UAS-ykiRNAi to reduce the abundance of yki in the eyes of flies expressing Gal4 from an eye-specific driver (GMR-Gal4) (32). We found that yki expression was reduced by 70% in the eyes of adult GMR-Gal4;UAS-ykiRNAi flies compared to that in GMR-Gal4 flies (fig. S1). Cardiac-specific expression of ykiRNAi in tinC-GFP; tinC-GAL4,UAS-ykiRNAi (tinC>ykiRNAi) flies increased EDDs and ESDs and reduced the percentage of fractional shortening (Fig. 1, A and D), indicative of cardiac dilation. Moreover, the heart wall was thinner in tinC>ykiRNAi flies (Fig. 1, B and C), indicating that loss of Yki induces cardiac dilation. The timing of the expression of YAPS127A during the development of the mammalian heart results in different cardiac abnormalities, including abnormal hyperplasia (7–9). Therefore, to determine whether aberrant activation of Yki during development affected heart development in flies, we silenced ykiS168A expression in tinC>ykiS168A flies using a temperature-sensitive form of the Gal4-binding transcriptional corepressor, Gal80, driven by the tubulin promoter (tubulin-Gal80ts), which results in ubiquitous expression of Gal80ts (16). tubulin-Gal80ts; tinC>ykiS168A flies that were grown at the restrictive temperature (18°C) from egg deposition and then shifted to the permissive temperature (29°C) to induce ykiS168A expression at the wandering larvae stage, before morphogenesis of the pupal heart (8), had decreased EDDs compared to tubulin-Gal80ts; tinC-Gal4 flies temperature-shifted at the same age (fig. S2). In contrast, EDDs in tubulin-Gal80ts; tinC>ykiS168A flies shifted to the permissive temperature when adults did not differ from EDDs in tubulin-Gal80ts; tinC-Gal4 flies (fig. S2), suggesting that the ability of YkiS168A to induce cardiac overgrowth is limited to a developmental stage occurring before or during heart morphogenesis. Activation of Yki causes hyperplasia in many tissues, including the eye (10, 24, 25). Because a normal adult fly heart is single-cell thick (6, 7), the increased thickness of the heart walls in flies expressing activated Yki could be due to cardiomyocyte proliferation (hyperplasia), increased size of cardiomyocytes (hypertrophy), or a combination of cardiac hyperplasia and hypertrophy (fig. S3). Therefore, we investigated whether cardiac-specific expression of activated Yki increased the abundance of mRNAs associated with cell proliferation, including mRNAs encoding cyclin-dependent kinase 1 (CDK1) and proliferating cell nuclear antigen (PCNA) (33, 34). As a positive control, we examined the effect of Yki activation in the eye using GMR-Gal4;UAS-ykiS168A flies. We found that the expression of CDK1 and PCNA was increased in heads dissected from GMR-Gal4;UAS-ykiS168A flies relative to GMR-Gal4 flies (Fig. 2A). In contrast, the expression of CDK1 and PCNA was similar in dissected hearts of tinC>ykiS168A and tinC-Gal4 flies (Fig. 2A), suggesting that aberrant activation of Yki does not promote cell proliferation in the heart. We directly assessed whether changes in heart wall thickness in tinC>ykiS168A and tinC>ykiRNAi flies were due to changes in the number of cardiomyocytes. We used a transgene encoding red fluorescent protein (RFP) fused with a nuclear localization sequence (UAS-RFPnuc) in tinC-GFP; tinC-Gal4 flies (tinC>RFPnuc flies) to identify cardiomyocyte nuclei in the second and third abdominal segments (A2 and A3). Sixteen cardiomyocytes were present in tinC>RFPnuc flies, tinC>RFPnuc; tinC>ykiS111A,S168A,S250A flies, and tinC>RFPnuc; tinC>ykiRNAi flies (Fig. 2B), indicating that changes in heart wall thickness in flies with cardiac-specific expression of constitutively active Yki or in flies with cardiac-specific loss of Yki occur in the absence of changes in the numbers of cardiomyocytes. Cardiac hypertrophy in mammals, including humans who have longstanding hypertension, is associated with an increase in myocyte ploidy due to increased endoreplication, a process of genomic DNA synthesis without cytokinesis (35–37). We found that the average ploidy was increased in cardiomyocytes of tinC>ykiS168A flies compared to tinC-GAL4 flies (Fig. 2C), consistent with increased endoreplication. Loss of Yki in tinC>ykiRNAi flies did not affect average ploidy (Fig. 2C). Collectively, these results suggest that unlike in mammals, in the fly, Yki controls heart organ size through hypertrophy rather than hyperplasia. Sd is required for cardiac hypertrophy induced by activated Yki Several transcription factors interact with Yki, including Sd, Mothers against dpp (Mad), and homothorax (Hth) (20, 38–40). Therefore, we examined the effects of loss of Sd, Mad, or Hth on ykiS168A-induced cardiac hypertrophy. Expression of either of two independent RNAi constructs targeting Sd in the eye reduced Sd abundance by ~50% (fig. S1). Moreover, cardiac-specific expression of either Sd RNAi construct increased EDDs and decreased heart wall thickness in tinC>ykiS168A flies, but not in tinC-Gal4 flies (fig. S4). In contrast, RNAi constructs directed against Mad or hth, which reduced the abundance of Mad by ~40% or hth by ~70% when expressed in the eye (fig. S1), did not abrogate cardiac abnormalities in adult tinC>ykiS168A flies when expressed in the heart (fig. S4). Thus, Sd, but not Mad or Hth, is required for YkiS168A-induced cardiac hypertrophy. Yki is required for Raf-induced cardiac hypertrophy Similar to cardiac-specific expression of activated Yki, cardiac-specific expression of a transgene encoding constitutively active human Raf (hRafAct), which has a deletion of amino acids 2 to 334, causes cardiac hypertrophy in Drosophila (9). Therefore, we tested if Yki was required for RafAct-induced cardiac hypertrophy. Compared to tinC-GFP; tinC-Gal4;UAS-hRafAct (tinC>hRafAct) flies, tinC>hRafAct flies with UAS-ykiRNAi (tinC>hRafAct + ykiRNAi flies) had increased EDDs and ESDs and decreased heart wall thicknesses (Fig. 3, A to C). Moreover, EDDs and heart wall thicknesses in tinC>hRafAct + ykiRNAi flies were similar to those in tinC-Gal4 flies (Fig. 3, A to C). Cardiac-specific expression of hRafAct causes abnormal cardiomyocyte fiber morphology (9). tinC-GFP–positive cardiomyocytes in hearts dissected from tinC>hRafAct flies lacked well-defined circumferential fibers, whereas cardiomyocytes in tinC>hRafAct + ykiRNAi flies had partially rescued fiber abnormalities (Fig. 3D). In contrast, cardiac-specific expression of Sd RNAi did not rescue RafAct-induced cardiac hypertrophy in tinC>hRafAct flies (fig. S5). Thus, these data suggest that Yki, but not Sd, is at least partially required for cardiac defects induced by aberrant activation of Raf. We also tested if knockdown of Raf abrogated heart defects in tinC>ykiS168A flies. Eye-specific expression of either of two independent RNAi constructs targeting Raf decreased the abundance of Raf by ~35% (fig. S1). However, neither Raf RNAi construct reduced heart wall thicknesses in tinC>ykiS168A flies when expressed in the heart (fig. S6). In contrast, RNAi directed against ERK (41) partially reduced heart wall thicknesses in tinC>ykiS168A flies (fig. S6), suggesting that ERK, but not Raf, is required for the ability of activated Yki to induce cardiac hypertrophy. Raf signaling through MEK and ERK enhances Yki-induced Sd-dependent transcription in mammalian cells A point mutation in human RAF that results in constitutive activation of Raf occurs in patients with Noonan syndrome, who present with cardiac hypertrophy (3, 4). Therefore, we investigated whether Raf and Yki could coactivate Sd-dependent transcription in human cells using cultured human embryonic kidney (HEK) 293T cells. We evaluated Sd-dependent transcription using a luciferase-based reporter comprising three consensus Sd-binding motifs from the Drosophila serum response factor enhancer and a minimal promoter (Sd-reporter) (21). Expression of mouse RafL613V (mRafL613V), equivalent to the mutant human Raf found in patients with Noonan syndrome, did not activate the Sd-reporter (Fig. 4A). Moreover, expression of mRafL613V did not enhance the ability of Sd to activate the Sd-reporter (Fig. 4A). However, expression of mRafL613V increased Sd-reporter activity in cells coexpressing Sd and either wild-type Yki (YkiWT) or YkiS168A (Fig. 4A). Additionally, neither expression of wild-type mRaf (mRafWT) nor expression of mRafL613V with a mutation that abolishes catalytic activity (mRafK375M,L613V) (fig. S7A) (42) enhanced Sd-reporter activity in cells expressing Sd and YkiS168A (Fig. 4B). We also tested if MEK activity was required for activation of the Sd-reporter by Raf and Yki. The MEK inhibitor PD98059 inhibited phosphorylation of ERK in HEK293T cells expressing mRafL613V (fig. S7B). Moreover, PD98059 inhibited Sd-reporter activity in cells expressing mRafL613V and Sd or in cells expressing Sd, mRafL613V, and YkiWT (Fig. 4C). Inhibition of ERK using FR180204 (43), which partially reduced phosphorylation of the ERK target ELK (fig. S7C), partially inhibited activation of the Sd-reporter in HEK293T cells expressing Sd, mRafL613V, and YkiS168A (Fig. 4D). We also determined if the mammalian orthologs of Yki and Sd could be activated by Raf. We coexpressed mRafL613V with human TEAD3 or TEAD4 and human YAPS127A in HEK293T cells with the Sd-reporter or a connective tissue growth factor (CTGF)-luciferase–based reporter, which can be activated by TEADs and YAP (22). Expression of mRafL613V increased activation of the Sd- and CTGF-reporter in cells expressing YAPS127A and TEAD3 or TEAD4 (fig. S8). The Tondu domain–containing repressor Tgi inhibits RafAct- or YkiS168A-induced cardiac hypertrophy Expression of activated Raf enhanced Yki-induced activation of Sd-mediated transcription in cultured cells, whereas Yki, but not Sd, was required for Raf-induced cardiac hypertrophy in flies. Because Sd, in combination with Tgi but in the absence of Yki, represses the expression of target genes (12), we tested whether cardiac-specific overexpression of Tgi could inhibit Raf-induced cardiac hypertrophy. Expression of Tgi inhibited Sd-reporter activity in HEK293T cells expressing Sd and YkiS168A or Sd, YkiS168A, and mRafL613V (Fig. 5A). Moreover, cardiac-specific expression of Tgi in tinC-Gal4; UAS-Tgi flies increased EDDs and decreased heart wall thicknesses in tinC>ykiS168A flies and tinC>RafAct flies (Fig. 5, B to E). These findings suggest that there is an equilibrium between Yki and Sd complexes and Tgi and Sd complexes that may be mediated by Raf signaling, and that altering the balance of these complexes produces cardiac hypertrophy (Fig. 5F). DISCUSSION We found that the transcriptional coactivator Yki comprises a signal that contributes to RafAct-induced cardiac hypertrophy in flies. Cardiac-specific expression of activated Yki produced cardiac hypertrophy, whereas cardiac-specific knockdown of Yki inhibited RafAct-mediated cardiac hypertrophy. Knockdown of Raf did not inhibit hypertrophy induced by aberrant activation of Yki in the heart, suggesting that Yki acts downstream of Raf. Moreover, expression of activated Raf increased Yki-or YAP-induced activation of Sd- or TEAD-dependent transcription in human cultured cells. The evolutionarily conserved Hippo-Yki pathway controls organ size, including the heart, in several species (10, 44). The Yki ortholog YAP was initially identified in lysates of chicken embryonic fibroblasts and is present in various mammalian tissues, including the heart (45, 46). Genetic activation of YAP produces cardiac hyperplasia in mice (23–25). Cardiac-specific expression of activated YAP in mice during heart development causes abnormally thickened myocardia and increased immunostaining for phosphorylated histone-H3 in the heart, indicative of increased proliferation of cardiomyocytes (25). Moreover, the size of hearts in adult mice is not affected by abnormal activation of YAP during development, suggesting that there may be compensatory mechanisms to normalize heart size by reducing the size of cardiomyocytes (25). Expression of activated YAP in the hearts of postnatal mice increases the expression of genes involved in the cell cycle and stimulates proliferation of cardiomyocytes, and these effects require TEAD1 (24). Cardiac-specific genetic ablation of Yap1 in developing mice causes abnormally thin myocardia (24, 25), associated with fewer ventricular myocytes and embryonic lethality (25). Thus, these data suggest that YAP is important for cardiomyocyte proliferation during development. In flies, we found that aberrant activation of Yki increased cardiomyocyte cell size rather than cell number. Expression of YkiS168A in the fly eye or wing causes cell proliferation (10, 18, 19). However, cardiac-specific expression of YkiS168A caused hypertrophy without increasing the number of cardiomyocytes or the expression of genes associated with the cell cycle. Moreover, cardiac-specific expression of RNAi directed against yki resulted in enlargement of the heart chamber and thinning of the heart walls without changing the number of cardiomyocytes. Thus, Yki appears to control myocyte morphology rather than proliferation in the fly heart. One possible explanation for the difference between cardiomyocyte proliferation in mammals and hypertrophy in flies is that the cardiomyocytes in flies transition from cell division to endoreplication. Although speculative, signals that block in cytokinesis and/or induce endoreplication may exist, and the fly represents a potential model to identify these possibilities. Using temperature shift experiments to control YkiS168A transgene expression, we found that aberrant activation of Yki in larval flies, but not in adult flies, produced heart defects. One possible explanation for this result is that the cardiac-specific driver used in this study (tinC-Gal4) induced the expression of transgenes during a developmental stage after cardiomyocyte proliferation. Alternatively, fly cardiomyocytes may have different signals that control proliferation during embryonic and adult stages, similar to adult mammalian cardiomyocytes. Comparing the developmental and interspecies differences between the signals that govern cardiomyocyte growth and proliferation may reveal evolutionary changes that resulted in the differences in heart size and morphology across diverse phyla. Crosstalk between Raf and Yki occurs by several mechanisms (26, 27). The Drosophila ortholog of RASSF competes with Salvador for binding to Hippo, thereby reducing Hippo activity in wings and eye discs (47). In human cells in culture, Raf inhibits MST2 in a manner independent of Raf kinase activity and ERK activation by preventing MST2 dimerization and recruiting a phosphatase that removes activating phosphorylations on MST2 (27). We observed that the kinase activities of Raf and MEK were required for the ability of mutant active Raf to enhance Yki-induced activation of an Sd-reporter in cell culture. These findings suggest a mechanism of crosstalk between catalytically active Raf-MEK signaling and the Yki- and Sd-containing transcription complex. Pharmacological inhibition of MEK ameliorates Noonan syndrome–like phenotypes in heterozygous RafL613V knock-in mice (4). Whether YAP-TEAD signaling is involved downstream of Raf and MEK in causing the phenotypes in this mouse model remains to be investigated. Pharmacological inhibition of ERK using FR180204 partially reduced the ability of RafL613V to augment Yki- and Sd-induced Sd-reporter activity in cultured cells. Exposing cells to FR180204 inhibited phosphorylation of the ERK–target ELK, but we cannot rule out the possibility of off-target effects by FR180204 on other kinases. Cardiac-specific expression of RNAi targeting ERK partially rescued YkiS168A-induced cardiac hypertrophy in flies. We previously found that cardiac-specific expression of ERK RNAi completely rescues RafAct-induced cardiac hypertrophy (9). Thus, ERK may be involved in the crosstalk between Raf and Yki in the heart, but further studies are required to determine the potential involvement of other kinases. Raf signaling through Yki and Sd in the fly heart is probably more complex than a simple linear model (fig. S9). Flies with cardiac-specific knockdown of Yki, but not Sd, had dilated hearts. However, cardiac-specific knockdown of Sd rescued cardiac hypertrophy induced by expression of YkiS168A in the heart. Cardiac-specific knockdown of Yki, but not Sd, inhibited RafAct-induced cardiac hypertrophy. Similar observations pertaining to Yki and Sd are reported for phenotypes in the fly eye, ovarian follicle cells, and wing discs, and led to the identification of Tgi, which binds Sd and is required for Sd-mediated transcriptional repression in the absence of coactivators, such as Yki (12). There are four orthologs of Tgi in mammals, including humans, named Vestigial-like (Vgll) proteins (48, 49). Vgll4 decreases the activity of TEAD1 and counteracts activation of gene expression in cardiac myocytes by activation of a1-adrenergic receptors (50). In other tissues, including lung and gastric tissues, Vgll proteins, including Vgll4, function as tumor suppressors by inhibiting YAP- and TEAD-based transcriptional activation (51, 52). We observed that cardiac-specific overexpression of Tgi inhibited the ability of RafL613V to enhance Yki-induced activation of an Sd-reporter in cell culture and the ability of cardiac-specific expression of RafAct or YkiS168A to induce cardiac hypertrophy in flies. Thus, Tgi represses Yki-induced Sd activity and inhibits crosstalk between Raf and Yki and Sd (Fig. 5F). In addition, Tgi may act through transcription factors other than Sd to attenuate Raf-induced cardiac hypertrophy. Moreover, Vgll proteins may be involved in Raf-induced cardiac hypertrophy in mammals, including humans with Noonan syndrome. MATERIALS AND METHODS Fly stocks The p{tinC-GFP}; p{tinC- Gal4} stocks were derived from p{tinC-GFP} and p{tinC-Gal4} stocks as previously described (30, 53). p{tinC-GFP}; p{UAS-RFPNUC}, p{tinC-Gal4} was homologously recombined using p{tinC-GFP}; p{tinC-Gal4} and p{UAS-RFPNUC} transgenic stocks. p{UAS-Tgi} was obtained from D. Pan (12). All other fly stocks, including p{UAS-ykiS168A.V5} (stock #28818), p{UAS-ykiS111A,S168A,S250A.V5} (stock #28817), p{UAS-ykiRNAi} (stock #31965), p{UAS-hRafAct} (stock #2074), p{UAS-sdRNAi} (stocks #29352 and #35481), p{UAS-RafRNAi} (stocks #31038 and #31596), p{UAS-MADRNAi} (stocks #31315, #31316, and #35648), p{UAS-HthRNAi} (stocks #27655 and #34637), and p{UAS-ERKRNAi} (stock #34855), were obtained from the Bloomington Stock Center or the Transgenic RNAi Project (TRiP) at Harvard Medical School (www.flyrnai.org). The p{UAS-RafAct} stock corresponds to the truncated version of human Raf-1 (D2–334) (54). P{tubP-Gal80ts}10; p{tinc-Gal4} was generated as previously described (55). All fly stocks were maintained on standard cornmeal-yeast protein medium at room temperature (56). Plasmids The Sd luciferase reporter (3xSd2-Luc) was provided by J. Jiang (University of Texas Southwestern) (21). The CTGF-luciferase reporter was generated using CTGF promoter sequence described by Zhao et al. (22) cloned into pGL3-Basic Vector. cDNAs (complementary DNAs) corresponding to yki (clone LD21311) and Sd (clone IP16090) were obtained from the Drosophila Genomics Resource Center. cDNAs corresponding to YAPS127A, TEAD3, and TEAD4 were obtained from Addgene (15, 22). TEAD3 and TEAD4 were tagged on the C terminus with the hemagglutinin (HA) epitope, wild-type yki and ykiS168A were tagged on the N terminus with a V5 epitope, and Sd was tagged on the C terminus with an HA epitope by PCR and subcloned into pCDNA3.1. cDNA encoding mouse wild-type Raf was amplified by reverse transcription PCR (RT-PCR) from C57B6 mouse heart RNA, tagged on a C-terminal Flag epitope, and sub-cloned into pCDNA3.1. Mouse RafL613V and RafK375M,L613V were generated by PCR mutagenesis and subcloned into pCDNA3.1. Fly Tgi was amplified by RT-PCR from total RNA from w1118 flies, tagged on the N terminus with a Myc epitope, and subcloned into pcDNA3.1. All plasmids were sequenced across the entire open reading frame for validation. OCT measurement of cardiac function in adult Drosophila Cardiac function in Drosophila was measured using a custom-built OCT microscopy system (Bioptigen Inc.) as previously described (9, 30, 53). EDD and ESD were determined from three consecutive heartbeats. Fractional shortening was calculated as [EDD − ESD]/EDD × 100. Histological analysis Fly heart wall thicknesses were measured as previously described (31, 53). Briefly, adult female flies of 2 to 7 days after eclosion were collected and fixed in Telly’s fixation buffer (60% ethanol, 3.33% formalin, 4% glacial acetic acid) for at least 1 week at 4°C. Flies were embedded in paraffin, serially sectioned at 8-μm thicknesses, and stained with hematoxylin and eosin. Hearts were imaged using a Leica DM2500 microscope equipped with a Leica DFC310 FX digital camera. Wall thickness was calculated by measuring the cardiac chamber wall width along the middorsal, midventral, left lateral, and right lateral wall in three serial sections per fly. Evaluation of adult cardiac morphology and ploidy Adult Drosophila corresponding to the F1 offspring of p{tinC-GFP}; p{tinC-Gal4} stocks crossed to specific p{UAS-transgenes} or w1118 (controls) were collected at 2 to 3 days of age, post-eclosion, to examine adult cardiac morphology as previously described (9, 57). Dissected specimens were stained with an antibody against GFP (1:500) (Invitrogen) and a secondary antibody conjugated to Alexa Fluor 488 (1:500) (Invitrogen). Labeled heart preparations were imaged using a Zeiss LSM 510 confocal microscope, and 0.4-μm Z-stack images were collected. Cardiomyocyte ploidy was measured as previously described (9, 58). Briefly, adult Drosophila corresponding to the F1 offspring of p{tinC-GFP}; p{tinC-Gal4} stocks crossed to specific p{UAS-transgenes} or w1118 (controls) were collected at 2 to 3 days after eclosion, and hearts and testis were dissected. Testis are haploid and therefore were used as internal controls. Chromosome number was determined by quantification of TO-PRO-3 staining of cardiac nuclei compared to haploid testis nuclei. Ploidy was expressed as the C value, where a C value of 1 refers to the amount of DNA contained within a haploid nucleus. Quantitative RT-PCR Total RNA samples from 20 to 30 dissected adult fly hearts or 15 to 25 fly heads per experiment were extracted using RNA-Bee (Tel-Test “B”). Two micrograms of RNA was used for generation of cDNA using SuperScript II reverse transcriptase (Invitrogen Inc.). Applied Biosystems TaqMan Gene Expression Assays were used to perform quantitative (real-time) qRT-PCR (table S1) with ribosomal protein L32 (Rpl32): Dm 02151827-g1 as an endogenous control. The following reaction components were used for each probe: 2 μl of cDNA, 12.5 μl of 2X TaqMan Universal PCR Master Mix without AmpErase (Applied Biosystems Inc.), 1.25 μl of probe, and 9.25 μl of water in a 25-μl total volume. Reactions were amplified and analyzed in triplicate using a Bio-Rad CFX96 Real-Time System. PCR conditions were as follows: step 1: 95°C for 10 min; step 2: 40 cycles of 95°C for 15 s followed by 60°C for 1 min. Expression relative to Rpl32 was calculated using 2−ΔΔCt, and levels were normalized to baseline. Cell culture, luciferase assays, and Western blots HEK293T cells were obtained from the American Type Culture Collection and cultured in Dulbecco’s modified Eagle’s medium supplemented with 10% fetal bovine serum and 1% penicillin-streptomycin. Cells were plated in triplicate wells in 24-well plates and transfected with FuGENE HD (Promega Inc.) and the following amounts of plasmid DNA: 100 ng of Sd-reporter or CTGF-reporter; 20 ng of Renilla luciferase (pBIND, Promega); 200 ng of YkiWT, YkiS168A, or YAPS127A; 200 ng of Sd, hTEAD3, or hTEAD4; 200 ng of Tgi; and 800 ng of mRafWT, mRafL613V, or mRafK375M,L613V, or pcDNA3.1 to make the total amounts of transfected DNA the same for each experiment. Cells were harvested after 48 hours of transfection, and luciferase activity was measured using the Dual-Luciferase Reporter Assay System (Promega) following the manufacturer’s protocol using a BMG Labtech NOVOstar plate reader. DMSO, PD98059 (10 mM; Santa Cruz Biotechnology), FR180204 (2, 5, or 10 μM; Santa Cruz Biotechnology), or DMSO was added to the transfected cells 24 hours before harvest. For Western blots, HEK293T cells were seeded in six-well plates and transfected with 2 μg of mRafWT, mRafL613V, mRafK375M,L613V, or pcDNA3.1 (2 μg each) and incubated for 48 hours. Cells were harvested and lysed in modified radioimmunoprecipitation assay (RIPA) buffer [137 mM NaCl, 50 mM tris-HCl (pH 8.0), 5 mM EDTA, 1% NP-40, 0.5% (w/v) Na deoxycholate, 10% glycerol] containing phosphatase inhibitors (Roche, #04906837001) and protease inhibitors (Roche, #11836170001) for 3 hours at 4°C. Samples containing 100 μg of total protein were resolved on SDS–polyacrylamide gels before transfer to Immobilon PVDF (polyvinylidene difluoride) Transfer Membranes (Millipore). The following antibodies were used: rabbit phosphorylated ERK (Thr202 and Tyr204) (Cell Signaling Technology, #9101) (1:1000), rabbit ERK (Millipore, # 06-182) (1:3000), rabbit ELK (Cell Signaling Technology, # 9182S) (1:1000), rabbit phosphorylated ELK (Ser383) (Cell Signaling Technology, # 9181S) (1:1000), and rabbit Flag (Sigma, #F7425) (1:1000). Statistical analysis GraphPad Prism (GraphPad Software Inc.) software was used for all statistical analyses. Supplementary Material supplemental Fig. S1. Validation of UAS-RNAi lines. Fig. S2. YkiS168A-induced cardiac hypertrophy occurs during development. Fig. S3. Schematic of the adult fly heart and cellular mechanisms of increasing heart wall thicknesses. Fig. S4. YkiS168A-induced cardiac hypertrophy requires the transcription factor Sd, but not MAD or Hth. Fig. S5. Knockdown of Sd does not inhibit RafAct-induced cardiac hypertrophy. Fig. S6. YkiS168A-induced cardiac hypertrophy is not inhibited by knockdown of Raf but is partially inhibited by knockdown of ERK. Fig. S7. Validation of genetic or pharmacological inhibition of Raf, MEK, and ERK. Fig. S8. Activated Raf enhances YAP-induced activation of TEAD-dependent transcription. Fig. S9. Models of RafAct- and YkiS168A-induced cardiac hypertrophy. Table S1. List of qPCR probes used in the studies. We thank D. Pan for providing UAS-Tgi stocks; K. Guan for providing plasmids encoding YAPS127A, TEAD3, and TEAD4; and J. Jiang for providing the Sd-reporter. We also thank H. Rockman for helpful discussions. Funding: This work was supported by NIH R01 HL116581 (to M.J.W.). Fig. 1 Expression of activated Yki causes cardiac hypertrophy, and knockdown of Yki causes cardiac dilation (A) Representative longitudinal and transverse B- and M-mode OCT images of hearts of flies with the indicated genotypes. Arrows denote cardiac chambers, and arrowheads denote EDDs and ESDs. (B) Representative transverse sections of hearts in the A1 segment from flies with the indicated genotypes. Scale bar, 50 μm. (C) Quantification of heart wall thicknesses from images similar to those shown in (B). Data are means ± SEM for n = 4 to 7 flies per genotype. *P < 0.05 compared to control. Analysis of variance (ANOVA) with Tukey’s post hoc test. (D) Quantification of EDDs, ESDs, and fractional shortening from images similar to those shown in (A). Data are means ± SEM for n = 8 flies per genotype. *P < 0.05 for indicated cardiac parameter for ykiRNAi compared to control. Two-sample t test. Fig. 2 Expression of activated Yki induces cardiac hypertrophy not hyperplasia (A) Graph of the relative abundance of CDK1 and PCNA mRNA measured by quantitative polymerase chain reaction (qPCR) in dissected hearts of flies with cardiac-specific (tinC) or in dissected heads of flies with eye-specific (GMR) expression of the indicated transgenes. Data are means ± SEM for n = 3 to 5 experiments per genotype normalized to the negative control genotype. *P < 0.05 for GMR>ykiS168A compared to GMR-Gal4. Two-sample t test. (B) Images of cardiomyocytes in the A2 and A3 segments for tinC-GFP; tinC>RFPnuc flies of the indicated genotypes. All transgenes were heterozygous. Images are representative of 10 flies per genotype. Scale bar, 200 μm. (C) Distribution of cardiomyocyte ploidy (C value) of hearts of flies with the indicated genotypes. The median C value is shown for each genotype in parenthesis. n = 18 to 60 cardiomyocytes per genotype. *P < 0.0001 by Wilcoxon signed rank test. Fig. 3 Knockdown of Yki inhibits RafAct-induced cardiac hypertrophy (A and B) Quantification of EDDs (A) or ESDs (B) from M-mode OCT of hearts of flies with the indicated genotypes. Data are means ± SEM for n = 16 to 29 flies per genotype. *P < 0.05 compared to control. #P < 0.05 compared to tinC>hRafAct. ANOVA with Tukey’s post hoc test. (C) Quantification of heart wall thicknesses of hearts of flies with the indicated genotypes (n = 4 to 7 per group). *P < 0.05 compared to control. #P < 0.05 compared to tinC>hRafAct. ANOVA with Tukey’s post hoc test. (D) Representative confocal microscopy with Z-stack reconstruction showing cardiac fiber morphology of hearts of flies with the indicated genotypes. All flies were examined in a tinC-GFP heterozygous background. tinC-GFP–positive cardiomyocytes were immunolabeled for GFP, din. Arrow denotes cardiomyocyte fibers. Images are representative of three to five flies. Fig. 4 Activated Raf enhances Yki-induced Sd activity in a manner that depends on the kinase activities of Raf, MEK, and ERK (A to D) Quantification of luciferase activity from an Sd-reporter in HEK293T cells transfected with the indicated plasmids. Data are means ± SEM for n = 3 experiments per group. (A) *P < 0.05 compared to control. #P < 0.05 compared to Sd + ykiWT. ΔP < 0.05 compared to Sd + ykiS168A. ANOVA with Tukey’s post hoc test. (B) *P < 0.05 compared to control. #P < 0.05 compared to Sd + YkiS168A. ANOVA with Tukey’s post hoc test. (C) Transfected cells were exposed to vehicle [dimethyl sulfoxide (DMSO)] or the MEK inhibitor PD98059 (10 mM) for 24 hours before lysis. *P < 0.05 compared to control. #P < 0.05 compared to Sd + YkiWT. ΔP < 0.05 compared to Sd + YkiWT + mRafL613V exposed to vehicle. ANOVA with Tukey’s post hoc test. (D) Transfected cells were exposed to DMSO or FR1800204 for 24 hours before lysis. Data are means ± SEM for n = 3 experiments per group. *P < 0.05 compared to control. #P < 0.05 compared to Sd + YkiS168A. ΔP < 0.05 compared to Sd + YkiS168A + mRafL613V + DMSO. ANOVA with Tukey’s post hoc test. Fig. 5 Overexpression of Tgi inhibits the ability of activated Raf to enhance YkiS186A-induced Sd activity in cultured cells and YkiS186A- or RafAct-induced cardiac hypertrophy in flies (A) Quantification of luciferase activity from an Sd-reporter in HEK293T cells transfected with the indicated plasmids. Data are means ± SEM for n = 3 experiments per group. *P < 0.05 compared to control. #P < 0.05 compared to Sd + YkiS168A. ΔP < 0.05 compared to Sd + YkiS168A + mRafL613V. ANOVA with Tukey’s post hoc test. (B and C) Representative M-mode OCT (B) and quantification (C) of EDDs of hearts of flies with the indicated genotypes. Data are means ± SEM for n = 15 to 32 per group. *P < 0.05 compared to control. #P < 0.05 compared to tinC>ykiS168A. ΔP < 0.05 compared to tinC>hRafAct. ANOVA with Tukey’s post hoc test. (D and E) Representative histology (D) and quantification (E) of heart wall thicknesses of hearts of flies with the indicated genotypes. n = 4 to 6 per genotype. *P < 0.05 compared to control. #P < 0.05 compared to tinC>ykiS168A. ΔP < 0.05 compared to tinC>hRafAct. Scale bars, 50 μm. (F) Model of RafAct- and Yki-induced cardiac hypertrophy. Author contributions: L.Y. and H.W. performed luciferase activity assays and Western blotting. L.Y. and J.P.D. performed histology experiments. L.Y. performed qPCR, confocal microscopy, and ploidy experiments. J.P.D. and M.J.W. performed OCT. L.Y., J.P.D., and M.J.W. designed fly crosses. H.W. and M.J.W. cloned and generated reagents. L.Y., J.P.D., and M.J.W. interpreted data and wrote the manuscript. Competing interests: All authors declare that they have no competing interests. Data and materials availability: All data and materials are available by request. SUPPLEMENTARY MATERIALS www.sciencesignaling.org/cgi/content/full/8/362/ra13/DC1 REFERENCES AND NOTES 1 Braunwald E Bonow RO Braunwald’s Heart Disease: A Textbook of Cardiovascular Medicine Elsevier/Saunders Philadelphia, PA 9 2012 2 Heineke J Molkentin JD Regulation of cardiac hypertrophy by intracellular signalling pathways Nat Rev Mol Cell Biol 7 589 600 2006 16936699 3 Pandit B Sarkozy A Pennacchio LA Carta C Oishi K Martinelli S Pogna EA Schackwitz W Ustaszewska A Landstrom A Bos JM Ommen SR Esposito G Lepri F Faul C Mundel P López Siguero JP Tenconi R Selicorni A Rossi C Mazzanti L Torrente I Marino B Digilio MC Zampino G Ackerman MJ Dallapiccola B Tartaglia M Gelb BD Gain-of-function RAF1 mutations cause Noonan and LEOPARD syndromes with hypertrophic cardiomyopathy Nat Genet 39 1007 1012 2007 17603483 4 Wu X Simpson J Hong JH Kim KH Thavarajah NK Backx PH Neel BG Araki T MEK-ERK pathway modulation ameliorates disease phenotypes in a mouse model of Noonan syndrome associated with the Raf1L613V mutation J Clin Invest 121 1009 1025 2011 21339642 5 Shah AP Nongthomba U Kelly Tanaka KK Denton ML Meadows SM Bancroft N Molina MR Cripps RM Cardiac remodeling in Drosophila arises from changes in actin gene expression and from a contribution of lymph gland-like cells to the heart musculature Mech Dev 128 222 233 2011 21237266 6 Wasserthal LT Drosophila flies combine periodic heartbeat reversal with a circulation in the anterior body mediated by a newly discovered anterior pair of ostial valves and ‘venous’ channels J Exp Biol 210 3707 3719 2007 17951411 7 Wolf MJ Modeling dilated cardiomyopathies in Drosophila Trends Cardiovasc Med 22 55 61 2012 22863366 8 Monier B Astier M Sémériva M Perrin L Steroid-dependent modification of Hox function drives myocyte reprogramming in the Drosophila heart Development 132 5283 5293 2005 16284119 9 Yu L Daniels J Glaser AE Wolf MJ Raf-mediated cardiac hypertrophy in adult Drosophila Dis Model Mech 6 964 976 2013 23580199 10 Dong J Feldmann G Huang J Wu S Zhang N Comerford SA Gayyed MF Anders RA Maitra A Pan D Elucidation of a universal size-control mechanism in Drosophila and mammals Cell 130 1120 1133 2007 17889654 11 Huang J Wu S Barrera J Matthews K Pan D The Hippo signaling pathway coordinately regulates cell proliferation and apoptosis by inactivating Yorkie, the Drosophila homolog of YAP Cell 122 421 434 2005 16096061 12 Koontz LM Liu-Chittenden Y Yin F Zheng Y Yu J Huang B Chen Q Wu S Pan D The Hippo effector Yorkie controls normal tissue growth by antagonizing Scalloped-mediated default repression Dev Cell 25 388 401 2013 23725764 13 Codelia VA Sun G Irvine KD Regulation of YAP by mechanical strain through Jnk and Hippo signaling Curr Biol 24 2012 2017 2014 25127217 14 Mori M Triboulet R Mohseni M Schlegelmilch K Shrestha K Camargo FD Gregory RI Hippo signaling regulates microprocessor and links cell-density-dependent miRNA biogenesis to cancer Cell 156 893 906 2014 24581491 15 Zhao B Wei X Li W Udan RS Yang Q Kim J Xie J Ikenoue T Yu J Li L Zheng P Ye K Chinnaiyan A Halder G Lai ZC Guan KL Inactivation of YAP oncoprotein by the Hippo pathway is involved in cell contact inhibition and tissue growth control Genes Dev 21 2747 2761 2007 17974916 16 Tapon N Harvey KF Bell DW Wahrer DC Schiripo TA Haber D Hariharan IK salvador promotes both cell cycle exit and apoptosis in Drosophila and is mutated in human cancer cell lines Cell 110 467 478 2002 12202036 17 Wu S Huang J Dong J Pan D hippo encodes a Ste-20 family protein kinase that restricts cell proliferation and promotes apoptosis in conjunction with salvador and warts Cell 114 445 456 2003 12941273 18 Oh H Irvine KD In vivo analysis of Yorkie phosphorylation sites Oncogene 28 1916 1927 2009 19330023 19 Ren F Zhang L Jiang J Hippo signaling regulates Yorkie nuclear localization and activity through 14-3-3 dependent and independent mechanisms Dev Biol 337 303 312 2010 19900439 20 Wu S Liu Y Zheng Y Dong J Pan D The TEAD/TEF family protein Scalloped mediates transcriptional output of the Hippo growth-regulatory pathway Dev Cell 14 388 398 2008 18258486 21 Zhang L Ren F Zhang Q Chen Y Wang B Jiang J The TEAD/TEF family of transcription factor Scalloped mediates Hippo signaling in organ size control Dev Cell 14 377 387 2008 18258485 22 Zhao B Ye X Yu J Li L Li W Li S Yu J Lin JD Wang CY Chinnaiyan AM Lai ZC Guan KL TEAD mediates YAP-dependent gene induction and growth control Genes Dev 22 1962 1971 2008 18579750 23 Heallen T Zhang M Wang J Bonilla-Claudio M Klysik E Johnson RL Martin JF Hippo pathway inhibits Wnt signaling to restrain cardiomyocyte proliferation and heart size Science 332 458 461 2011 21512031 24 von Gise A Lin Z Schlegelmilch K Honor LB Pan GM Buck JN Ma Q Ishiwata T Zhou B Camargo FD Pu WT YAP1, the nuclear target of Hippo signaling, stimulates heart growth through cardiomyocyte proliferation but not hypertrophy Proc Natl Acad Sci USA 109 2394 2399 2012 22308401 25 Xin M Kim Y Sutherland LB Qi X McAnally J Schwartz RJ Richardson JA Bassel-Duby R Olson EN Regulation of insulin-like growth factor signaling by Yap governs cardiomyocyte proliferation and embryonic heart size Sci Signal 4 ra70 2011 22028467 26 Reddy BV Irvine KD Regulation of Hippo signaling by EGFR-MAPK signaling through Ajuba family proteins Dev Cell 24 459 471 2013 23484853 27 O’Neill E Kolch W Taming the Hippo: Raf-1 controls apoptosis by suppressing MST2/Hippo Cell Cycle 4 365 367 2005 15701972 28 O’Neill E Rushworth L Baccarini M Kolch W Role of the kinase MST2 in suppression of apoptosis by the proto-oncogene product Raf-1 Science 306 2267 2270 2004 15618521 29 Lo PC Frasch M A role for the COUP-TF-related gene seven-up in the diversification of cardioblast identities in the dorsal vessel of Drosophila Mech Dev 104 49 60 2001 11404079 30 Wolf MJ Amrein H Izatt JA Choma MA Reedy MC Rockman HA Drosophila as a model for the identification of genes causing adult human heart disease Proc Natl Acad Sci USA 103 1394 1399 2006 16432241 31 Yu L Daniels JP Wolf MF Manual of Research Techniques in Cardiovascular Medicine Ardehali H Bolli R Losordo DW Wiley Blackwell Chichester, West Sussex, UK 2014 432 439 32 Freeman M Reiterative use of the EGF receptor triggers differentiation of all cell types in the Drosophila eye Cell 87 651 660 1996 8929534 33 Landberg G Roos G Antibodies to proliferating cell nuclear antigen as S-phase probes in flow cytometric cell cycle analysis Cancer Res 51 4570 4574 1991 1678682 34 Rhind N Russell P Signaling pathways that regulate cell division Cold Spring Harb Perspect Biol 4 a005942 2012 23028116 35 Adler CP Friedburg H Myocardial DNA content, ploidy level and cell number in geriatric hearts: Post-mortem examinations of human myocardium in old age J Mol Cell Cardiol 18 39 53 1986 3950970 36 Laflamme MA Murry CE Heart regeneration Nature 473 326 335 2011 21593865 37 Sandritter W Scomazzoni G Deoxyribonucleic acid content (Feulgen photometry) and dry weight (interference microscopy) of normal and hypertrophic heart muscle fibers Nature 202 100 101 1964 14166707 38 Oh H Irvine KD Cooperative regulation of growth by Yorkie and Mad through bantam Dev Cell 20 109 122 2011 21238929 39 Peng HW Slattery M Mann RS Transcription factor choice in the Hippo signaling pathway: homothorax and yorkie regulation of the microRNA bantam in the progenitor domain of the Drosophila eye imaginal disc Genes Dev 23 2307 2319 2009 19762509 40 Zhang T Zhou Q Pignoni F Yki/YAP, Sd/TEAD and Hth/MEIS control tissue specification in the Drosophila eye disc epithelium PLOS One 6 e22278 41 Nir R Grossman R Paroush Z Volk T Phosphorylation of the Drosophila melanogaster RNA–binding protein HOW by MAPK/ERK enhances its dimerization and activity PLOS Genet 8 e1002632 2012 22479211 42 Dent P Reardon DB Morrison DK Sturgill TW Regulation of Raf-1 and Raf-1 mutants by Ras-dependent and Ras-independent mechanisms in vitro Mol Cell Biol 15 4125 4135 1995 7623807 43 Ohori M Kinoshita T Okubo M Sato K Yamazaki A Arakawa H Nishimura S Inamura N Nakajima H Neya M Miyake H Fujii T Identification of a selective ERK inhibitor and structural determination of the inhibitor–ERK2 complex Biochem Biophys Res Commun 336 357 363 2005 16139248 44 Pan D The Hippo signaling pathway in development and cancer Dev Cell 19 491 505 2010 20951342 45 Sudol M Hanafusa H Cellular proteins homologous to the viral yes gene product Mol Cell Biol 6 2839 2846 1986 3491292 46 Zhao YH Krueger JG Sudol M Expression of cellular-yes protein in mammalian tissues Oncogene 5 1629 1635 1990 2267131 47 Polesello C Huelsmann S Brown NH Tapon N The Drosophila RASSF homolog antagonizes the Hippo pathway Curr Biol 16 2459 2465 2006 17174922 48 Faucheux C Naye F Tréguer K Fédou S Thiébaud P Théze N Vestigial like gene family expression in Xenopus: Common and divergent features with other vertebrates Int J Dev Biol 54 1375 1382 2010 20712000 49 Pobbati AV Hong W Emerging roles of TEAD transcription factors and its coactivators in cancers Cancer Biol Ther 14 390 398 2013 23380592 50 Chen HH Mullett SJ Stewart AF Vgl-4, a novel member of the vestigial-like family of transcription cofactors, regulates α1-adrenergic activation of gene expression in cardiac myocytes J Biol Chem 279 30800 30806 2004 15140898 51 Jiao S Wang H Shi Z Dong A Zhang W Song X He F Wang Y Zhang Z Wang W Wang X Guo T Li P Zhao Y Ji H Zhang L Zhou Z A peptide mimicking VGLL4 function acts as a YAP antagonist therapy against gastric cancer Cancer Cell 25 166 180 2014 24525233 52 Zhang W Gao Y Li P Shi Z Guo T Li F Han X Feng Y Zheng C Wang Z Li F Chen H Zhou Z Zhang L Ji H VGLL4 functions as a new tumor suppressor in lung cancer by negatively regulating the YAP-TEAD transcriptional complex Cell Res 24 331 343 2014 24458094 53 Yu L Lee T Lin N Wolf MJ Affecting Rhomboid-3 function causes a dilated heart in adult Drosophila PLOS Genet 6 e1000969 2010 20523889 54 Brand AH Perrimon N Targeted gene expression as a means of altering cell fates and generating dominant phenotypes Development 118 401 415 1993 8223268 55 Kim IM Wolf MJ Serial examination of an inducible and reversible dilated cardiomyopathy in individual adult Drosophila PLOS One 4 e7132 2009 19771157 56 Ashburner M Golic KG Hawley RS Drosophila: A Laboratory Handbook Cold Spring Harbor Laboratory Press Cold Spring Harbor, NY 2 2005 57 Alayari NN Vogler G Taghli-Lamallem O Ocorr K Bodmer R Cammarato A Fluorescent labeling of Drosophila heart structures J Vis Exp 13 1423 2009 58 Fox DT Gall JG Spradling AC Error-prone polyploid mitosis during normal Drosophila development Genes Dev 24 2294 2302 2010 20952538

PMC005xxxxxx/PMC5131644.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 0372762 3389 Dev Biol Dev. Biol. Developmental biology 0012-1606 1095-564X 27521456 5131644 10.1016/j.ydbio.2016.08.003 NIHMS821470 Article Small RNA in situ hybridization in Caenorhabditis elegans, combined with RNA-seq, identifies germline-enriched microRNAs☆ McEwen Tamara J. a2 Yao Qiuming b Yun Sijung c Lee Chin-Yung d Bennett Karen L. a*3 a Molecular Microbiology and Immunology Department, University of Missouri School of Medicine, Columbia, MO 65212, USA b Department of Computer Science, Bond Life Science Center, University of Missouri, Columbia, MO 65211, USA c Yotta Biomed, LLC, 4835 Cordell Ave #1117, Bethesda, MD 20814, USA d The Seydoux Laboratory, Molecular Biology and Genetics Department, Johns Hopkins School of Medicine, Baltimore, MD 21205, USA 2 Current address: Natural Sciences Division, Southwestern College, Winfield, KS 67156, USA. 3 Professor Emerita, University of Missouri; Visiting Scientist, Seydoux Laboratory, USA. * Corresponding author. bennettk@missouri.edu (K.L. Bennett). 14 11 2016 10 8 2016 15 10 2016 01 12 2016 418 2 248257 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. Over four hundred different microRNAs (miRNAs) have been identified in the genome of the model organism the nematode Caenorhabditis elegans. As the germline is dedicated to the preservation of each species, and almost half of all the cells in an adult nematode are germline, it is likely that regulatory miRNAs are important for germline development and maintenance. In C. elegans the miR35 family has strong maternal effects, contributing to normal embryogenesis and to adult fecundity. To determine whether any particular miRNAs are greatly enriched in the C. elegans germline we used RNA-seq to compare the miRNA populations in several germline-defective strains of adult C. elegans worms, including glp-4(germline proliferation-4), glh-1(germline helicase-1) and dcr-1(dicer-1). Statistical analyses of RNA-seq comparisons identified 13 miRNAs that are germline-enriched, including seven members of the well-studied miR35 family that were reduced as much as 1000-fold in TaqMan qRT PCR miRNA assays. Along with the miR35s, six others: miR-56 (a member of the miR51 family),−70, −244, −260 , −788 and −4813, none of which previously considered as such, were also identified by RNA-seq as germline-enriched candidates. We went on to develop a successful miRNA in situ hybridization protocol for C. elegans, revealing miR35s specifically concentrate during oogenesis in the pachytene region of the gonad, and persist throughout early embryogenesis, while in adult animals neither let-7 nor miR-228 has a germline-bias. glp-4 glh-1 dcr-1 miR in situ protocol miR35 family let-7 miR-228 1. Introduction Although microRNAs (miRNAs) were first described in C. elegans more than twenty years ago with the report of the lin-4 small, temporal(st), non-coding(nc) RNA (Lee et al., 1993), the miRNA field exploded after the report of let-7, a second C. elegans miRNA affecting developmental timing, and one conserved in many other organisms, including humans (Pasquinelli et al., 2000). miRNAs, usually 21–22nts long, are endogenous to most plants and animals and are cleaved in the cytoplasm from hairpin precursor RNAs by the endo-ribonuclease Dicer. Mature miRNAs regulate their mRNA targets through antisense complementarity (often with mismatches), binding mRNAs in their UTRs (untranslated regions) or coding regions. miRNA binding causes post-transcriptional regulation of their targets, usually resulting in lower mRNA and/or protein levels. For more than a decade many studies have implicated miRNAs as regulators of development and as markers of disease. While lin-4 and let-7, and the C. elegans miRNA lsy-6 (Johnston and Hobert, 2003), were each discovered by classical genetics and display visible mutant phenotypes, targeted deletion of over 80 individual miRNAs or closely linked miRNA families revealed only a few striking abnormalities. Exceptions include deletions in two families, the miR51s or miR35s that each result in embryonic lethality (Miska et al., 2007; Alvarez-Saavedra and Horvitz, 2010). The highly-conserved miR51 family, consisting of miRs-51-56 (miR-100s in humans), is likely the oldest group of animal miRNAs, found in annelids, nematodes, flies and humans. This family is critical for pharyngeal development in the C. elegans embryo (Shaw et al., 2010). The miR35 family consists of eight miRNAs, each containing the same CACCGGC “seed” in their 5′ ends; miR35s are located in two clusters, with miR-35-41 clustered together and miR-42 350 kb away. Microarray and northern analyses revealed the miR35 family is strongly expressed in the oogenic germline (Miska et al., 2007), while additional studies revealed loss of this family results in embryonic death, male sterility and severe reductions in brood size (Alvarez-Saavedra and Horvitz, 2010; McJunkin and Ambros, 2014). Thus, the C. elegans miR35s are important both for embryonic development and to produce a normal adult germline. In mammals, several recent studies have also considered roles for miRNAs in embryonic and germline development. In humans, the enhancement of somatic fates by the let-7 miRNA family works in opposition to miR-372, which promotes human germ cell fates in embryonic stem cells (Tran et al., 2016), while in mice, sperm, as well as oocytes, transmit miRNAs and endo-siRNAs critical to the developing embryo (Yuan et al., 2016). Despite elegant studies of miRNAs in C. elegans identifying miRNAs that are differentially expressed during development (Kato et al., 2009) or with ageing (Kato et al., 2011; Lucanic et al., 2013), no studies targeting germline miRNAs have been reported. These studies were likely hindered by the inherent germline silencing of C. elegans transgenes, resulting in an inability of miRNA transgenes to express in the germline (Kelly et al., 1999; Martinez et al., 2008; Pierce et al., 2008; Shaw et al., 2010) and because in situ hybridizations for miRNAs in C. elegans had not yet been published, with the exception of miR-57, the most abundant miRNA in our RNA-seq studies, Supplemental Table S1, localizing to the posterior of a late-stage embryo (Zhao et al., 2010). Therefore, whether many miRNAs show a germline-bias was unknown. Thus the goals of this study were to identify germline-enriched miRNA candidates and to develop methods to visualize C. elegans germline miRNAs by in situ hybridization. To do so, we carried out RNA-seq, followed by statistical analyses, utilizing the rather unique advantage available in C. elegans of multiple germline-minus and germline-defective mutant strains. With statistical comparisons using worms with and without germlines, 13 miRNAs were identified as germline-enriched. These include seven miR35 family members, along with a single member of the miR51 family, and five additional, individual miRNAs. To visualize the localization pattern of the miR35 family, we also developed a successful in situ hybridization protocol and applied it to several other C. elegans miRNAs, including miR-228 and let-7. 2. Materials and methods 2.1. Strains Worm strains used for these studies include the following: wild type (N2 variety Bristol); VC178 [glh-1(gk100) I]; SS104 [glp-4(bn2) I]; and PD8753 [dcr-1(ok247) III/ghT2[bli-4(e937) let-?(q782) qIs48] (I;III)] (the ghT2, GFP-tagged chromosome, allows for selecting non-green homozygous dcr-1 worms); and MT14119 [miR-35-41(nDf50) II] (This latter mutant carries a deletion of miRs-35-41, eliminating the polycistronic miR-35-41 transcript (Alvarez-Saavedra and Horvitz, 2010)). All strains used in this study were maintained as described in (Brenner, 1974) and were grown at the optimal growth temperature of 20° unless otherwise noted. In this report strains will be referred to as follows: VC178 as glh-1; SS104 as glp-4; PD8753 as dcr-1, and MT14119 as miR-35-41. 2.2. Harvesting worms Wild type (N2) and glh-1 worms were placed at the first larval stage (L1) on NGM agar plates seeded with OP50 E. coli, cultured at 20° or 26° and harvested as young adults 1 day beyond the fourth larval stage (L4), when mature oocytes and embryos are first detected. N2s were harvested off approximately one hundred 15 × 100 mm plates and glh-1 animals were harvested off approximately two hundred 15 × 60 mm plates by washing with M9 buffer (22 mM KH2PO4, 22 mM Na2HPO4, 86 mM NaCl). glp-4 worms were cultured in liquid S media (10% S basal (0.1 M NaCl, 0.05 M potassium phosphate, 5 μg/ml cholesterol, 1 mM potassium citrate, 1 × trace metals solution [5 mM EDTA, 2.5 M FeSO4, 1 mM MnCl2, 1 mM ZnSO4, 0.3 mM MgSO4, 0.3 mM CaCl2, 0.03 mM MgSO4]) containing 7% χ1666 E. coli and incubated on an orbital shaker at 15° for ~6 days until the majority of worms in the culture were young adults full of embryos. The eggs in glp-4 adults were harvested by treatment with hypochlorite and grown until the adult stage at 26°. glp-4 26° adults were visually inspected for clear gonads, which indicates absence of a germline. All worms from a particular strain were combined and bacteria were cleared from the worms’ cuticles and intestines by shaking in M9 for 45 min. This was followed by a wash in 1 × lysis buffer (50 mM HEPES pH 7.5, 1.0 mM EGTA, 3.0 mM MgCl2, 100 mM KCl, 10% glycerol), with slurries of 1:1 volume of lysis buffer/worms stored in aliquots at −80°. Approximately 10,000 dcr-1(ok427) homozygous (non-green) worms were hand-picked and frozen in M9 and 50 × protease inhibitor (Complete™, mini, EDTA-free (Roche)). Multiple aliquots of dcr-1 homozygous worms, collected over a period of time, were combined and RNA was isolated. 2.3. RNA isolation Frozen worm lysate was thawed on ice and re-suspended in ~3 volumes of TRI-Reagent (Ambion, Inc). The TRI-Reagent/worm slurry was ground to a fine powder in liquid nitrogen and sonicated. After clearing insoluble debris, lysates were treated with 0.2 volume of chloroform, and the aqueous phase precipitated with equal volumes of isopropanol at −20°. RNA pellets were concentrated by centrifugation, and were stored in 10 mM Tris–HCl, pH7.5. RNA concentrations were determined with a Nanodrop ND1000 spectrophotometer; yields ranged from 2 to 5 μg/μL. 2.4. Small RNA enrichment Small RNA fractions were enriched from glp-4, glh-1, N2, and dcr-1 total RNA samples using the mirVana™ (Ambion, Inc) small RNA isolation protocol, with slight modifications as in (Gu et al., 2009). 5 μg aliquots were resolved on 2% agarose, MOPS-formaldehyde gels to visualize the RNAs prior to cDNA library construction and TaqMan® miRNA Assays. 2.5. Small RNA library construction and deep sequencing The construction of cDNA libraries from small RNA templates and subsequent sequencing on the Illumina HiSeq-2000 was carried out by University of Missouri DNA Core Facility staff. Libraries were constructed from small RNAs from glp-4 and glh-1 mutant and N2 worms grown at 15°, 20° or 26°, and dcr-1 mutants grown at 15°. For each strain and each tested temperature two cDNA libraries, each representing a unique biological replicate, were prepared with the Illumina® TruSeq™ Small RNA Sample Preparation kit as per the manufacturer's protocol. Each library was made from 2.5 μg of mir-Vana™-enriched small RNA and was individually identified with a short specific linker/adapter. The quality of construction was determined with an Agilent BioAnalyzer 2100 Chip. A plot illustrating this assay is shown in Supplemental Fig. S1. The cDNA libraries were sequenced using Illumina's HiSeq2000. 2.6. Bioinformatics Sequencing data are available at NCBI in the Bioproject database with accession number PRJNA322183 (http://www.ncbi.nlm.nih.gov/bioproject/PRJNA322183). Data were generated by Illumina sequencing, with reads cleaned to remove low quality bases and trimmed to remove the adaptors. All reads were aligned against the annotated, non-coding RNA transcripts downloaded from Wormbase ftp://ftp.wormbase.org/pub/wormbase/releases/WS242/species/c_elegans/PRJNA13758/c_elegans. PRJNA13758. WS242.ncrna_transcripts.fa.gz, which contains annotations for 223 miRNAs, 345 snoRNAs, 126 snRNA, 176 lincRNAs, 7980 ncRNAs, 15,365 piRNAs and other RNAs, including tRNAs, and rRNAs. The alignment from the above processed short reads to the non-coding transcripts was conducted using BWA, Burrows-Wheeler alignment (Li and Durbin, 2010). To obtain the counts for each miRNA, the HTSeq protocol (Anders et al., 2014) was applied to the mapping results, which are found in Supplemental Table S1. Statistical tests comparing differential abundance levels between the miRNAs of various populations were carried out using DESeq2, a statistical analysis software available as an R/Bioconductor package (Love et al., 2014). DESeq2 software was used because it reports consistent results even for experiments with small number of replicates (Seyednasrollah et al., 2015). For our sequencing results DESeq2 provided much more conservative estimates of statistical significance than did the software package edgeR (Robinson et al., 2010). The adjusted p-value (q-value) of <0.05 was used as the cutoff for differential expression with statistical significance. q-values are adjusted p-values, used with the Benjamini-Hochberg procedure (Benjamini and Hochberg, 1995). http://www.stat.purdue.edu/~doerge/BIOINFORM.D/FALL06/Benjamini%20and%20Y%20FDR.pdf. 2.7. TaqMan® miRNA assays qRT PCR was used to verify the RNAseq results for several candidates. Levels of miR-38, −40, −228, and miR-244 were analyzed in glp-4, glh-1, dcr-1 and wild type worms with TaqMan® MicroRNA Assays (Applied Biosystems). Quantitative PCR (qPCR) was performed in most cases with three independent biological samples, using triplicate technical replicates of each sample. Template cDNA specific for each miRNA was generated with the TaqMan® MicroRNA Reverse Transcription (RT) Kit as per the manufacturer's protocol. 10 ng of glp-4 (15° and 26°), glh-1 (26°), N2 (26° and 20°) or dcr-1 (15°) mirVana™-enriched small RNA were used as the starting material for each 15 μL RT reaction. Products from each RT reaction were used as template for qPCR using TaqMan® Universal PCR Master Mix II (no UNG), with primers supplied by the manufacturer on a 7900HT Fast Real-Time PCR System (Applied Biosystems). Cycling conditions were: 95° for 10 min, and 40 cycles of 95° for 15 s and 60° for 1 min. Relative expression levels were determined according to (Pfaffl, 2001) and the delta-delta CT method (Schmittgen and Livak, 2008), using U18, a small nucleolar RNA transcript, as an endogenous control. Statistical analyses for these qPCRs were performed using T-tests, with the Relative Expression Software Tool (REST) 2009 (www.gene-quantification.de/rest.html). 2.8. in situ hybridizations in situ hybridizations were carried out using modifications of the protocols reported by Seydoux and Fire (1994) for detecting messenger RNAs in C. elegans and by Pena et al. (2009) for localizing miRNAs in mouse tissue sections. Because the protocol is a new combination of others, it will be described in some detail here. C. elegans wild type and mutant worms were grown at 15°, 20° or 26 °C either to a young adult stage, when worms first begin producing embryos (~1 day beyond the L4, fourth larval stage), or to a later adult stage, at 2–3 days beyond L4, when worms are fully gravid and the plates contain many embryos and newly-hatched larvae. Worms were placed in M9 buffer containing 1 mM levamisole, an anesthetic, and splayed onto poly-lysine-treated slides to provide access for the miRNA probes to the internal tissues. The slides were placed on dry ice to freeze-crack the adult tissues and embryos, using the light pressure of a coverslip. Slides were stored in 100% methanol at −20° for as little as 10 min, or for up to two weeks, before further fixation, hybridization and probe detection. After an additional 5 min in 100% methanol at RT, slides were rehydrated through an alcohol series of 90%, 70% and 50% methanol in TBS (150 mM NaCl, 50 mM Tris-HCl, pH7.5) for 1 min each, and were washed twice in TBS. All subsequent washes were in 1x TBS or TBS-T (TBS plus 0.1% Tween-20) for 5 min, all solutions were made with DEPC (diethylpyrocarbonate)-treated water and incubations were at RT unless noted otherwise. Throughout the fixation process TBS was used instead of PBS to avoid adding phosphate to the tissue, which would interfere with the EDC treatments to follow. The fixed tissue was then treated with 5 μg/ml, a mild dose of proteinase K (Roche) in 1x TBS for 15–20 min at RT, followed by exposure to 2 mg/ml glycine in TBS, and two TBS washes. An additional fixation for 20 min in 3.2% formaldehyde in a buffer of 80 mM HEPES, pH6.9, 16 mM MgSO4, and 8 mM EGTA in TBS was followed by repeating the above glycine and wash steps. After the formaldehyde fixation, we followed the protocol of Pena et al., (2009) (Supplementary methods; Nat. Methods: doi.1038/nmeth.1294) using EDC (1-ethyl-3-(3-dimethylaminopropyl) carbodiimide), which significantly increases the levels of miRNAs remaining in the tissues by creating phosphoramidate linkages on the 5′ phosphates of the small RNAs. To first remove extraneous residual phosphates, slides were incubated twice, for 10 min, in 0.13 M 1-methylimidazole, 300 mM NaCl, pH8.0. Thereafter EDC was added to a concentration of 0.16 M into fresh 1-methylimidazole solution, placed on each slide and incubated for 2 h, followed by a third glycine treatment and two TBS washes. While the use of EDC was critical to our detection of a colorimetric signal, rather than storing EDC under argon, we purchased fresh aliquots of anhydrous EDC (Life Technologies) and brought this reagent up in the imidazole solution when needed for the in situ protocol. We also did not retain the harsh step of blocking all endogenous enzymatic activity with 0.1 M triethanolamine and 0.5% acetic anhydride that followed the EDC treatment (Pena et al., 2009). After the EDC fixation, the slides were pre-hybridized for 2 h in a standard hybridization solution of 50% formamide, 5 × SSC, 5 × Denhardt's solution, 250 μg/ml yeast tRNA, 500 μg/ml herring DNA, 2% Blocking Reagent (Roche), and 0.1% CHAPS and 0.5% Tween-20 detergents. The miRNA probes tested were LNAs, locked nucleic acids, which are nuclease resistant; the LNAs we used were complementary to miR-35, miR-38 and miR-40, let-7 and miR-228. Controls included an LNA corresponding to the C. elegans pre-SL-1, the nuclear precursor of spliced leader-1, a small ncRNA involved in trans-splicing (Krause and Hirsh, 1987), as a positive control, and as a negative control, a scrambled LNA that does not correspond to any known C. elegans miRNA sequence; the LNAs were pre-labeled on both 5′ and 3′ ends with Digoxigenin (Exiqon). The MT14119 miR-35–41 deletion strain was used as an additional negative control. Sequences of the LNA probes are found in Supplemental Table S2. Most probes were used for hybridization at a final concentration of 25 nM; however, when combined, the miR-35, −38 and −40 LNAs were each at 8 nM, while miR-228 was at 50 nM. Probes were briefly heat denatured at 68° before added to the hybridization solution. Slides were covered with parafilm and hybridized in a humidified chamber for 13–16 h at 48 °C, a temperature > 20° below the Tm of any of the LNAs used (57° was also successful). After hybridization, 5x SSC was used to float off the parafilm and slides were washed twice for 30 min each in 1 × SSC; 50% formamide, 0.1% Tween-20 at 48°, then washed at RT in 0.2 × SSC for 15 min and finally in TBS-T. This was followed by blocking for endogenous peroxidase activity, with 3% hydrogen peroxide in TBS-T for 30 min, followed by three 1 min TBS-T washes. To detect the hybridized probes, a antibody/amplification/antibody process was used. Slides were first blocked for 1 h at RT in 0.5% Blocking Reagent (Roche), with 10% heat-inactivated goat serum in TBS-T, followed by the addition of an anti-DIG-FAB peroxidase (POD, Roche) antibody diluted 1:500 in the above blocking solution and incubated for 1h at RT. Amplification of the peroxidase conjugate on the anti-DIG antibody was then achieved with the addition of the tyramide reagent that catalyzes deposits of dinityrophenyl (DNP) (TSA kit, PerkinElmer Renaissance® TSA™Plus DNP system NEL 747A). First, slides were washed 3 times in TNT buffer (0.1 M Tris-HCL, pH7.5, 0.15 M NaCl, 0.1% Tween-20) and then the tryamide solution was applied and incubated for 30 min, after which slides were washed 3 times in maleate buffer (0.09 M maleic acid, 0.175 M NaOH, 1 M NaCl, and 0.5% Tween-20, pH7.5). The DNP labels were subsequently detected by a second antibody reaction, with an anti-DNP antibody conjugated to Alkaline Phosphatase (AP) diluted 1:500 in the maleate buffer, supplemented with 10% blocking reagent (Roche); this reaction went for 40 min. Subsequent washes were twice in maleate buffer and four times in TMN buffer (0.1 M Tris-base, pH9.5, 0.05 M MgCl2, 0.5 M NaCl, 0.05% Tween-20, and 2 mM tetramisole hybrochloride). These steps were as per manufacturer's directions. The detection of signal followed using a colorimetric reaction with a BCIP/NBT solution (Roche Ready Mix tablets) that was allowed to proceed for 30 min, to allow an initial deposition of blue pigment, after which slides were incubated in KTBT (50 mM Tris-HCl, pH7.5, 150 mM NaCl, 10 mM KCl, with 1% Tween-20, a notably high detergent concentration) at 4° for 1–14 days and a second color reaction was carried out for an additional 30 min or more, until a strong signal was seen or the background became visible. This two-step color reaction, with KTBT buffer, was adapted from the GEISHA, chick protocols (Darnell et al., 2006) and made for a clearer, stronger signal. After completing the color reaction, slides were washed three times in a penultimate wash of 0.01 M Tris-HCL, pH7.5, 0.5 M NaCl. 5 mM EDTA, 0.05% Tween-20, followed by a 10 min fixation in a 4% aqueous paraformaldehyde solution, ending with a wash in water. Cover slips were applied using Vectashield mounting medium with DAPI (Vector Laboratories) and sealed with clear nail polish. Images were taken on a Zeiss Auxio Z1 microscope with SlideBox software and processed with Photoshop and Illustrator software. While the protocol used for these experiments was successful in repeated sets of experiments, using a single miR35- member probe was not effective. Future optimization of these in situ techniques for C. elegans may increase the sensitivity of miRNA detection. 3. Results 3.1. Comparing the small RNA transcriptomes of germline-defective C. elegans To determine if any C. elegans miRNAs are enriched in the germline, we conducted deep sequencing on 14 cDNA libraries prepared from small RNA populations isolated from wild-type (N2) worms, grown at 20 °C and 26 °C, as well as from three germline-defective strains: glp-4(bn2) at 15 °C and 26 °C, glh-1(gk100) at 20 °C and 26 °C, and dcr-1(ok247) homozygous worms raised at 15 °C used as a control. Each library was constructed from a unique biological replicate. glp-4(bn2) is a mis-sense, partial loss of function mutation in vars-2 (Rastogi et al., 2015), a gene coding for valine aminoacyl tRNA synthetase whose mRNA is highly enriched in the germline. glp-4(bn2) worms are fertile when grown at 15 °C, and are sterile when grown at the restrictive temperatures of 25–26 °C, with each hermaphrodite producing only ~12 germ cells compared to ~2000 in wild type (Beanan and Strome, 1992). glh1(gk100) is a non-null deletion allele in glh-1, a germline-specific RNA helicase component of P granules (Gruidl et al., 1996; Spike et al., 2008; Strome and Wood, 1983; Sheth et al., 2010). The truncated glh-1(gk100) product is expressed at ~10% of wild-type levels. glh-1(gk100) animals are fertile at 20 °C and 100% sterile at 26 °C, with under-proliferated germlines 1/5th-1/2 the normal size (Spike et al., 2008). The glh-1 mutant was chosen for this study in part because GLH-1 is a binding partner of Dicer and therefore GLH-1 might have a role in processing miRNAs or other Dicer-dependent small RNAs (Beshore et al., 2011). The dcr-1(ok247) strain is a null mutation in the gene coding for the Dicer protein. dcr-1(ok247)hermaphrodites have fully developed germlines, but produce abnormal, endo-replicating germ cells (Ketting et al., 2001; Knight and Bass, 2001). The dcr-1(ok247) strain is balanced with a GFP-tagged chromosome and non-green, sterile, dcr-1homozygous adults were hand-picked to generate small RNAs for libraries and qRT-PCRs. All libraries were constructed using a 5′ ligation-dependent protocol that is specific for mono-phosphorylated species, which include micro-, piwi(pi)- and 26G RNAs; each library was individually tagged. Deep sequencing of the multiple cDNA libraries generated from small RNA templates produced a total of 107,728,496 raw reads, 21,045,231 of which uniquely aligned to annotated, transcribed loci in the C. elegans genome. Of the uniquely aligned sequences, 5,064,689 reads mapped to 93 of the 223 annotated miRNA genes in WormBase. ftp://ftp.wormbase.org/pub/wormbase/releases/WS242/species/c_elegans/PRJNA13758/c_elegans.PRJNA13758.WS242.ncrna_transcripts.fa.gz. In addition 62,364 reads mapped to 96 pre-miRNA genes, and 486,561 reads mapped to piRNAs. The reads mapping to other small RNAs are illustrated in Supplemental Fig. S2. Through use of the analysis pipeline of BWA and HTseq protocols (Li and Durbin, 2010; Anders et al., 2014), the raw mature miRNAs reads detected in all 14 libraries were determined; however, mature miRNA reads in the two dcr-1 libraries were rare and are not included with the other 12 populations seen in Supplemental Table S1. Subsequently, the DESeq2 software program was used to examine the differentially-abundant miRNAs based directly on the read counts (Love et al., 2014). Both biological and technical variations were accommodated. We considered miRNAs to be differentially expressed using the q-value of <0.05 as the cutoff for statistical significance. 3.2. Several miRNAs are significantly reduced when glp-4 worms lack a germline Table 1A compares miRNA levels in glp-4 mutants grown at 15° (fertile) versus 26° (sterile). Among the 93 miRNAs detected by our RNA-seq, we identified 21 miRNAs whose relative levels significantly decrease, and eight whose relative levels increase in sterile glp-4 worms compared to fertile ones (Table 1A, Fig. 1A and Supplemental Table S3A). Because glp-4 mutants raised at 26° consist mainly of somatic tissues, the former category likely correspond to miRNAs expressed preferentially in the germline, whereas the latter are likely to correspond to miRNAs expressed preferentially in the soma. Since higher temperatures are known to adversely affect the fertility of wild type animals (Spike et al., 2008), we compared N2 worms grown at the optimal growth temperature of 20° to the more restrictive temperature, 26° and found no statistically significant differences among the 96 miRs detected by our RNAseq and the subsequent DESeq1 analysis, Supplemental Table S4. We then went on to compare glp-4 mutants grown at 26° to wild-type worms grown at 26° (Table 1B, and Fig. 1B). This comparison yielded 26 significantly mis-regulated mRNAs, including 17 down-regulated and nine up-regulated miRNAs. The gene overlap between the two glp-4 comparisons (Table 1A-B and Fig. 1A-B) was 77% (13/17) for the down-regulated category and 56% (5/9) for the up-regulated category. Among the former, are 13 miRNAs including all seven miR35 family members and six other unrelated miRNAs, miR-56, −70, −244, −260, −788 and −4813, which we consider the “best candidates” for being germline-enriched. With regard to the down-regulated miRNAs that don't overlap in the two comparisons, we predict the four miRNAs only reduced in the comparison of glp-4 26° versus wild type animals grown at 26° may reflect specific effects of high temperature on the defective valine tRNA synthetase in the glp-4(bn2) animals, as three of these four miRNAs (miR- 43,−61 and 1829.3) are not reduced in wild type 26° animals, Supplemental Table S4. One of the original aims of this research was to determine whether GLH-1, a constitutive P-granule component and a binding partner of Dicer (Beshore et al., 2011) affects germline miRNA levels. Therefore, we compared sterile glh-1 mutants grown at 26° to fertile N2 worms grown at 26°. However, RNA-seq with sterile glh-1 worms did not indicate all 13 of the germline-enriched miRNAs are statistically reduced, Table 1C. Four of the five reduced miRNAs, miR-56, −244, −788 and −4813, are among the germline-enriched miRNAs identified in the glp-4 analyses presented above, Table 1A-B. While these findings are consistent with sterile glh-1 mutants having tiny germlines and fewer germ cells (Spike et al., 2008), seven other germline candidates were reduced, but not significantly, and two showed no reductions, Supplemental Table S3C. We also examined the levels of precursor miRNAs in glh-1 mutants and did not find them obviously higher than those of N2 worms, in contrast to the accumulation of precursor miRNAs in dcr-1 worms, Supplemental Table S5. Therefore, our RNA-seq analysis is inconclusive as to whether this glh-1 strain affects the Dicer-dependent processing of all germline miRNAs. Perhaps too few replicates were tested or the limited response is due to this strongest glh-1 mutant not being a null. The GLH-1/DCR-1 germline complex might also function with the C. elegans 22Gs, endo-siRNAs that are critical in the germline for genome integrity and to repress selected protein coding sequences (Gu et al. 2009). In mice oocytes endo-siRNAs are essential for meiotic maturation (Stein et al. 2015). While the 22Gs are, like miRNAs, Dicer-dependent, they are not mono-phosphorylated, and therefore are not in our libraries. Thus, the basis of the DCR-1/GLH-1 relationship (Beshore et al., 2011) remains to be determined. 3.3. Verification of the RNA-seq results using qRT-PCR To independently verify some of the changes observed by RNA-seq, we carried out qRT-PCR for four miRNAs, three predicted to be preferentially expressed in the germline based on the RNA-seq data and one, miR-228, predicted to be expressed both in the soma (Pierce et al., 2008) and germline (Table 1A–C). As many as nine replicate TaqMan qRT-PCRs compared: (i). glp-4 26° levels against glp-4 15° worms, (ii). glh-1 worms against N2s, both grown at 26°, and as a control (iii). dcr-1, raised at 15°, vs. fertile N2 adults grown at 20°, Fig. 1C and Table 2. Quantification of the qRT-PCR results for glp-4 confirmed the RNA-seq results, panel 1. We observed the most dramatic changes for the two miR35 family members tested, with reductions as high as 1000-fold for miR-35 in glp-4 sterile mutants, Fig. 1C, panel 1 and Table 2. When fertile N2 worms were tested against sterile glh-1 animals, all four probes were significantly reduced in the qRT PCR assays, Fig. 1C, panel 2 and Table 2. Perhaps unexpectedly, both miR-38 and −40, which were not significantly reduced in the DESeq2 analysis of sterile glh-1 worms, Supplemental Table S3C, were significantly reduced with the increased numbers of biological and technical replicates used for qRT PCR, Fig. 1C, panel 2, and Table 2, further supporting the germline-enrichment of the miR35s. As expected, the dcr-1 RNA showed significant reductions for all miRNAs tested, Fig. 1C, panel 3 and Table 2. However, the markedly greater fold changes for the miR35 probes in the germcell-less glp-4 animals than seen for dcr-1 null mutants might be unexpected, Table 2. We note that dcr-1 animals have large, albeit abnormal, germlines, and therefore, could retain higher levels of stable, maternal miRNAs than the germline-minus glp-4 animals. Therefore, we propose this result is likely due to the perdurance of maternal miRNAs in the dcr-1 mutant, a suggestion also made by Drake et al. (2014) when studying Dicer-1 and observing “only modest differences in bulk miRNA levels between WT and dcr-1(0) animals”. 3.4. in situs indicate miR35s are highly enriched in the adult germline To visualize the localization of miRNAs in C. elegans, we developed an in situ hybridization protocol for miRNAs based on techniques reported for chick and mouse miRNAs (Darnell et al., 2006; Pena et al., 2009) as detailed in Section 2. Hybridizations were carried out on adult C. elegans worms that had been splayed open to release their gonads, partially removing the germline (red arrows), as well as the intestinal tissue (black arrows), away from the body wall cavity, Fig. 2. A cartoon of the two symmetric gonad arms as they appear in an intact adult worm illustrates the normal developing germline, Fig. 2A. In Fig. 2B, strong hybridization is seen in the meiotic pachytene region of the gonad where oogenic and maternal RNAs are abundantly transcribed (red arrow), when a combined LNA probe complementary to the mature miRNAs corresponding to miR-35, −38 and −40 was used. In Fig. 2C, the miR35s are also detected with the combined miR35 probe in the oocytes and embryos (white arrow) of this older adult. A five-fold extension of the color reaction time using two probes, miR-38 and −40, revealed staining throughout both completely-removed gonad arms, beginning in the distal region (red arrow) in this young adult animal, Fig. 2D. Even with this extended exposure, no signal was detected in the adjacent, isolated intestinal tissue (or in other somatic tissues), Fig. 2D, despite previous studies with a miR35 promoter-driven GFP transgene indicating GFP expression in the vulva, several anterior neurons and the rectum (Martinez et al., 2008). This difference in findings may reflect a lack of sensitivity of the current in situ protocol or a failure to penetrate the hard cuticle of the adult body wall with the relatively mild treatments used for fixation, Section 2. A DAPI image of Fig. 2D is seen in Fig. 2E. As negative controls, the mutant strain MT14119 miR-35-41(nDf50) was tested, Fig. 2F, with only background hybridization detected with the combined miR35 probe, while a scrambled LNA also produced no signal, Fig. 2G. Both negative controls were carried out for ~1 h, the same exposure used for Figs. 2B-C, and 2H-I, and for the positive SL-1 probe, which revealed concentrated nuclear staining for the SL-1 precursor, Supplemental Fig. S3, in contrast to the uniform, cytoplasmic staining of the mature miRNAs seen here, Fig. 2B-D and H-I. In addition, a let-7 LNA probe produced no signal in N2 adult tissues or wild type embryos (not shown), an expected result based on the rarity of let-7 in our libraries, Supplemental Table S1. However, surprisingly, we did detect let-7 hybridization in the abnormal embryos of the miR-35-41 deletion strain, Fig. 2H; this staining appears to reveal inappropriate embryonic expression of let-7 when the miR35s are missing. This could relate to findings in (Akay et al., 2013) that both miR35 and let7 families genetically interact with GLD-1, a well-studied RNA binding protein. While Brenner and Schedl (2016) present convincing evidence that GLD-1 is not a target of the miR35s, we suggest GLD-1 may indirectly regulate these two miRNA families and that when one miRNA family is missing, the other increases in expression. In Fig. 2I, in contrast to the exclusive germline location of the miR35s seen by in situ hybridizations in Fig. 2B-D, we detected strong signals in both the intestine and in the distal germline with a miR-228 probe. As the C. elegans miR-228 (miR-183 in mammals) was previously seen in multiple somatic tissues by GFP expression (Pierce et al., 2008) and based on our RNA-seq and qRT-PCR results, this dual localization was not unexpected. In summary, these in situ hybridizations are consistent with our RNA-seq results, Table 1A-B, and provide the first visualization of the miR35 microRNAs as highly enriched the adult germline, concentrating in the pachytene region and in maturing oocytes; they also reveal expected localizations for miR-228 and both expected and unexpected patterns for let-7. 4. Discussion We have compared the miRNA transcriptomes of wild-type and germline-defective worms and have identified 13 miRNAs that are preferentially expressed in the germline. These include seven of the eight members of the miR35 family, which appear highly enriched in germ cells. In using our newly developed in situ hybridization protocol to detect miRNAs in C. elegans tissues, the miR-35-41 miRNAs appear exclusive to the germline of adult hermaphrodites. These miRNAs are first detected during oogenesis and accumulate at high levels in oocytes and early embryos. Our findings are consistent with the genetic findings of (Alvarez-Saavedra and Horvitz, 2010) who determined that deletion of seven of the eight miR35 family members causes temperature-sensitive, maternal-effect and zygotic lethality, resulting in embryonic or L1 larval arrest, while loss of all members is lethal at all temperatures. Since many miRNAs have been individually deleted (Miska, et al., 2007), the phenotypes resulting from the loss of some of the germline miRNAs identified here have already have been reported. As described above, deletion of all miR35 family members results in embryonic lethality; however adding back maternal expression of the miR35–41 gene cluster or of each individual miR35 member, results in normal embryonic development (Alvarez-Saavedra and Horvitz, 2010), implying redundancy of the miR35s and a vital role for this family in embryonic development. In addition, McJunkin and Ambros (2014) have demonstrated that maternal, along with zygotic, miR35s act early in development to insure maximal fertility of the adult progeny, while Brenner and Schedl (2016) quantified the marked reduction in proliferation in the C. elegans germline with loss of all miR35 family members. In the miR51 family, miR-56 is the only member we identified as germline-enriched. Deletion of all six miR51 genes results in dead embryos, with any individual miR51 member able to rescue the embryonic lethality (Shaw et al., 2010); however, a specific role for miR-56 in the adult germline had not been reported. There are no deletions reported for miR-788 and −4813; these two more recently-discovered miRNAs were not a part of the large deletion study by the Horvitz group (Miska et al., 2007). There are individual deletions of miR-70, −244 and −260, which revealed no gross abnormalities, supporting the conclusions of others, including (Miska et al., 2007; Alvarez-Saavedra and Horvitz, 2010) that the normal regulatory functions of most individual miRNAs are likely subtle. And while studies in mammals have demonstrated that most germline miRNAs lack function in mouse oocytes or during early embryonic development (Suh et al., 2010), similar studies are incomplete in C. elegans. In addition, recent findings in mammals suggest that individual miRNAs have decisive roles in promoting germline development, as seen for human miR-372 (Tran et al., 2016). Therefore, in addition to the established maternal contributions of the miR35 and miR51 families, we predict additional individual and opposing C. elegans miRNA teams also contribute to germline development. Nine of the 13 germline miRNA candidates identified here are among the 50+ miRNAs that change in abundance as hermaphrodites age and cease reproduction (Kato et al., 2011; Lucanic et al., 2013). In particular, six members of the miR35 family dramatically increase in abundance in aged, post-reproductive worms, while miR-70, −244, and −788 decrease. The Lucanic studies also used the germline-minus strain glp-1(e2141) to further verify the age-dependent differences in levels of miR-35, −37 and −244; however, whether these germline miRNAs play a role in aging has not yet been established. Interestingly, only one of the germline miRNAs, miR-70, was reported as a miRNA that is abundant in embryos and increases in abundance during larval development, while miR-35, −37 and −40 decrease dramatically following embryogenesis and remain at low levels until adulthood (Kato et al., 2009). Perhaps miR-70 is involved in the establishment of the germline that begins in embryos and progresses in larvae, while the other 12 germline miRNAs function as maternal (or paternal) transcripts critical to embryogenesis, or in maintenance of the adult germline. In summary, RNA-seq and the differential statistical comparisons in this report have identified 13 miRNAs that are preferentially expressed in the germline of adult hermaphrodites. These studies have further parsed the relationships of individual miRNAs in the miR-35 and −51 families and have taken the C. elegans miRNA field beyond these two very interesting families, with striking embryonic phenotypes and germline connections, to reveal five additional, germline-enriched miRNAs. We anticipate future studies will reveal these five newly-identified miRNAs also play important roles in development, either individually or in concert. Supplementary Material 1 8 9 10 11 2 3 4 5 6 7 Acknowledgments We thank Bill Spollen, bioinformatics specialist at MU, who advised QY and TJM in their bioinformatics analyses and Dr. Darryl Conte, UMass, for technical advice regarding small RNA isolation. KLB thanks Geraldine Seydoux and Depika Calidas for critical manuscript review and the Seydoux lab members, especially Alex Paix for in situ advice and Dominique Rasoloson for technical help, and Sally Moody and her lab at George Washington U, who also hosted KLB. This work was supported by NSF funding IOS-0819712 (KLB), a University of Missouri Life Sciences Fellowship (TJM), and NIH grant NICHD90042240 to G. Seydoux for partial support of KLB. Fig. 1 Multiple miRNAs are affected by loss of the germline A The miRNAs that significantly change when sterile glp-4 worms grown at 26° are compared to fertile glp-4 worms grown at the permissive temperature of 15° on this volcano plot produced with Prism software. All 93 miRNAs detected in our adult populations are plotted here (black dots). The 21 miRNAs that decrease when the germline is missing are in the left quadrant; seven members of the miR-35 family are indicated in red. In addition, eight miRNAs, seen in the right quadrant, significantly increase in the glp-4 miRNA population when the germline is missing. The 64 miRNAs that showed no change, or changes not meeting the criteria for significance, are clustered in the center of the A plot below the dotted horizontal cutoff line which indicates a FDR(False Discovery Rate), using the q-value of <0.05 for significance. The data plotted in figure is presented in Table 1A–C, for the significantly changing miRNAs and in Supplemental Table S3A–C for all 93 miRNAs detected. B This volcano plot, similar to that in 1A, compares abundance of glp-4 26° miRNAs with those of N2 worms grown at the same temperature. C qRT-PCR verifies the relative abundance of miR-38, miR-40, miR-228 and miR-244. Levels of miR-38,-40,-228 and −244 were assayed using both biological and technical triplicates by qRT-PCR comparing small RNAs from glp-4 26° sterile worms to those of glp-4 15° fertile worms (panel 1); glh-1 26° sterile mutants compared to N2s at 26° (panel 2); and homozygous, sterile dcr-1 15° animals vs. N2s at 20° (panel 3). U18 was used as the endogenous control for each condition and strain tested. The statistical significance of differences in relative levels of miRNAs was determined using the Relative Expression Software Tool (REST) 2009 (www.gene-quantification.de/rest.html). The brackets with an asterisk in C indicate the p-value is statistically significant (p-value <0.0001). Numerical results are given in Table 2. Fig. 2 In situ hybridizations indicate miR35 miRNAs are exclusively detected in the adult germline, while miR-228 is present in both germline and somatic tissue A provides an illustration of the C. elegans adult bi-lobed gonad from (Smith et al., 2002) revealing the progression of germline development from the mitotic region in the distal gonad (D) beginning at the distal tip cell (DTC), with germ cells transitioning into meiosis (trans. zone) with oogenesis (pachytene and diakinesis) producing mature oocytes (O) in the proximal gonad (P) which contains mature sperm (S) in the spermatheca. After fertilization the embryos (E) develop and are later released, at about the 40-cell stage, through the vulva (V). In B-I and in Supplemental Fig. S3A several C. elegans miRNAs were localized by in situ hybridization to adult worm tissues using LNA probes complementary to miRs-35,−38 and −40, miR-228, a Scrambled LNA (a negative control), let-7, and to an LNA complementary to pre-SL-1 (a positive control). In each case, unless otherwise indicated, the total concentration of individual or combined probes did not exceed 25 nM and all worms were wild type (N2). All images were taken at the same 100X magnification.B is a whole hermaphrodite adult splayed open to expose the intestine and the germline. The worm was hybridized with a combination of miR-35, −38,−40 probes, revealing a signal concentrated in the meiotic pachytene region (the area of bend in the gonad, red arrow), with no signal seen in the gut (black arrow). In all images in this figure and Supplemental Fig. S3C the germline is marked with red, somatic tissue with black, and embryos with white arrows. Unless otherwise indicated, worms were young adults 1–2 days beyond the L4 stage (>L4) and grown at 15°, because the lower growth temperature extends the lifespan and a “young” adult worm state (Kato et al., 2011). The color reactions were carried out in two steps for a total of ~60 min, unless otherwise indicated. C shows an older adult (2–3 days >L4) also hybridized with the three miR-35 family LNAs and again the signal is concentrated in the mid-section of the germline tissue, in the meiotic, pachytene region (red arrow). In this image the combined miR-35s probe also hybridizes throughout all the cells in the extruded embryos (white arrows) of this older adult. D illustrates the use of two miR35 family probes, miR-38 and −40, on the two gonad arms and an intestinal tract that were completely removed from a young N2 adult grown at 26°. Here the final color reaction was allowed to proceed for 5.5 h. It is clear that the positive signal remains concentrated in the germline, although it is now throughout the two separated arms of the bi-lobed gonad (red arrows) and remains absent from the somatic intestine (black arrow). An asterisk marks the extruded pharynx. This longer exposure reveals that miR35 miRNAs, while abundant and concentrated at the pachytene stage of meiosis, are also present at lower levels throughout the germline, including the proliferative distal region (red arrows). E is the DAPI image of panel D. F shows a MT14119 miR-35-41 worm raised at 15°, splayed and hybridized with the miR-35, −38, and −40 combination. While this worm is negative, in some cases the deletion mutant shows weak hybridization, perhaps due the match of the LNAs to the “seed” sequence of miR-42, a family member not deleted in this strain. In G a Scrambled LNA, a second negative control, was used for the hybridization; this probe consistently produced no signal. H shows the let-7 LNA on an MT14119 miR-35-41 older adult grown at 26°. As with N2 adult worms (not shown), the let-7 LNA does not hybridize to adult tissue. At this non-permissive temperature the miR-35-41 deletion strain retains aberrantly developing embryos and it appears the let-7 probe uniformly hybridizes throughout these embryos (white arrows). I shows an in situ hybridization of miR-228 on an N2 worm grown at 26°. miR-228 localizes in the adult germline, especially at the distal region of the gonad (red arrow) and is also detected in the somatic gut (black arrow). Table 1 Statistical analyses of sterile germline-defective mutants compared to fertile worms These tables show the miRNAs that significantly change in each comparison. A The DESeq2 analysis of glp-415° miRNAs compared to glp-4 26°. B DESeq2 calls of significance for glp-4 26° vs. N2 26° miRNAs. C The DESeq2 analysis for glh-1 26° vs. N2 26° miRNAs. To view the results of statistical analyses for all detectable miRNAs see Supplemental Table S3A–C. log2 fold change Adj. P-value A. glp-4 (bn2) 26C vs 15C miR-4813 −4.89605 9.30E–08 miR-260 −6.36403 1.21E–07 miR-39 −6.26213 2.89E–07 miR-38 −7.40438 1.29E–06 miR-248 2.35392 2.35E–06 miR-35 −6.45842 0.0001017 miR-37 −6.21010 0.000252791 miR-70 −1.64195 0.000267552 miR-238 1.41267 0.001948051 miR-239.1 2.16398 0.003155295 miR-2212 −1.83031 0.004223912 miR-262 −4.15019 0.004887495 miR-240 −1.62272 0.004995897 miR-2214 1.23544 0.00714549 miR-244 −1.09378 0.00714549 miR-42 −2.73965 0.00714549 miR-73 −1.53514 0.00714549 miR-1830 −3.47870 0.00930712 miR-56 −1.24676 0.00930712 miR-1022 −1.38437 0.009435131 miR-53 1.21911 0.014087703 miR-40 −3.66994 0.015520947 miR-36 −4.62137 0.015631642 miR-60 1.63146 0.018428118 miR-4931 −4.38697 0.024686066 miR-71 1.08388 0.024686066 miR-788 −1.80748 0.02719458 miR-4916 1.80619 0.045959487 miR-1832.1 −2.46817 0.046651332 B.glp-4 (bn2) 26C vs N2 26C miR-248 2.95717 3.85E–12 miR-38 −6.37337 1.95E–11 miR-4813 −4.59489 2.43E–10 miR-70 −2.20808 4.83E–10 miR-39 −5.25596 3.82E–09 miR-37 −5.55568 2.81E–08 miR-35 −5.48695 4.64E–08 miR-56 −1.79510 3.25E–07 miR-788 −2.79181 1.22E–06 miR-1829.3 −2.14947 0.00026497 miR-249 1.68721 0.00029823 miR-260 −3.76763 0.00029823 miR-71 1.35860 0.00035472 miR-239.1 1.96080 0.0006414 miR-238 1.15484 0.00139564 miR-40 −3.35445 0.00226176 miR-42 −2.38211 0.00835347 miR-53 0.88166 0.00835347 miR-58 0.92226 0.00835347 miR-244 −0.85106 0.01037996 miR-43 −1.86624 0.01037996 miR-36 −3.16671 0.01530756 miR-72 0.80125 0.02515828 miR-65 −0.70855 0.03634389 miR-61 −2.13195 0.03755028 miR-60 1.19401 0.04133958 C. glh-1 (gk100) 26C vs N2 26C miR-244 −1.32081 1.44E–06 miR-4813 −2.32621 1.24E–05 miR-72 0.95901 0.00615274 miR-58 0.91224 0.01032463 miR-237 −0.91340 0.01035653 miR-56 −0.96261 0.0113635 miR-64 0.69839 0.0113635 miR-788 −1.56113 0.0113635 miR-51 1.08469 0.01302394 miR-85 0.68831 0.01302394 Table 2 Relative abundance of mature miRNAs in TaqMan® miRNA assays. qRT-PCR was utilized to assay miR-38, miR-40, miR-228, and miR-244 levels in germline-minus glp-4 26° compared to fertile glp-4 animals @15°, and in glh-1 worms compared to N2s, both grown at 26°. Levels of these miRNAs were also measured in dcr-1 null mutants, where they were expected to decrease. U18 was used as the endogenous control for each condition and strain tested. Statistical analysis with REST software (REST 2009) indicated that all three germline miRNAs tested were significantly decreased in comparisons at 26° of glp-4 and glh-1 worms, and in the 15° dcr-1 mutant. glp-4 (bn2)at 26C vs glp-4 at 15C miRNA Abundance Std. Error P-value Result miR-38 0.001 0–0.006 < 0.0001 1000-fold down miR-40 0.003 0.001–0.009 < 0.0001 333.3-fold down miR-228 0.377 0.190–0.838 < 0.0001 2.7-fold down miR-244 0.242 0.093–0.601 < 0.0001 4.1-fold down glh-1(gk100) at 26C vs N2 at 26C miRNA Abundance Std. Error P-value Result miR-38 0.284 0.165–0.517 < 0.0001 3.5-fold down miR-40 0.135 0.012–0.576 < 0.0001 8.0-fold down miR-228 0.319 0.128–0.598 < 0.0001 3.1-fold down miR-244 0.173 0.062–0.326 < 0.0001 5.8-fold down dcr-1(ok247) at 15C vs N2C miRNA Abundance Std. Error P-value Result miR-38 0.065 0.045–0.114 < 0.0001 15.4-fold down miR-40 0.018 0.010–0.027 < 0.0001 55.6-fold down miR-228 0.066 0.043–0.106 < 0.0001 15.2-fold down miR-244 0.017 0.007–0.042 < 0.0001 58.8-fold down ☆ Note added in proof: The Kohara group has also developed a C. elegans miRNA in situ hybridization protocol in a methods paper, Andachi and Kohara (2016) RNA July 22:1099-1106. Appendix A. Supplementary material Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j.ydbio.2016.08.003. References Akay A Craig A Lehrbach N Larance M Pourkarimi E Wright JE Lamond A Miska E Gartner A RNA-binding protein GLD-1/quaking genetically interacts with the mir-35 and the let-7 miRNA pathways in Caenorhabditis elegans. Open Biol 2013 3 130151 24258276 Alvarez-Saavedra E Horvitz HR ‘Many families of C. elegans microRNAs are not essential for development or viability’. Curr. Biol 2010 20 367 373 20096582 Anders S Pyl PT Huber W HTSeq, a Python framework to work with high-throughput sequencing data. Bioinformatics 2014 http://dx.doi.org/10.1096/bioinformatics/btu638. epub 2014Sept 23 Beanan MJ Strome S Characterization of a germ-line proliferation mutation in C. elegans. Development 1992 116 3 755 766 1289064 Benjamini Y Hochberg Y Controlling the false discovery rate: a practical and powerful approach to multiple testing. J. R. Stat. Soc. Ser. B 1995 57 1 289 300 Beshore EL McEwen TJ Jud MC Marshall JK Schisa JA Bennett KL ‘C. elegans Dicer interacts with the P-granule component GLH-1 and both regulate germline RNPs’. Dev. Biol 2011 350 2 370 381 21146518 Brenner JL Schedl T Germline stem cell determination entails regional control of cell fate regulator GLD-1 in Caenorhabditis elegans. Genetics 2016 202 1085 1103 26757772 Brenner S The genetics of Caenorhabditis elegans. Genetics 1974 77 1 71 94 4366476 Darnell DK Kaur S Stanslaw S Konieczka JH Yatskievych TA Antin PB ‘MicroRNA expression during chick embryo development’. Dev. Dyn 2006 235 11 3156 3165 17013880 Drake M Furuta T Suen KM Gonzaalez G Liu B Kalia A Ladbury JE Fire A,Z Skeath JB Arur S A requirement for ERK-dependent Dicer phosphorylation in coordinating oocyte-to-embryo transition in C. elegans. Dev. Cell 2014 31 614 628 25490268 Gruidl ME Smith PA Kuznicki KA McCrone JS Kirchner J Roussell DJ Strome S Bennett KL ‘Multiple potential germline helicases are components of the germline-line specific P granules of Caenorhabditis’. Proc. Natl. Acad. Sci. USA 1996 93 24 13837 13842 8943022 Gu W Shirayama M Conte D Jr. Vasale J Batista PJ Claycomb JM Moresco JJ Youngman EM Keys J Stoltz MJ Chen CC Chaves DA Duan S Kasschau KD Fahlgren N Yates JR 3rd Mirani S Carrington JC Mello CC Distinct Argonaute-mediated 22G-RNA pathways direct genome surveillance in the C. elegans germline. Mol. Cell 2009 36 2 231 244 19800275 Johnston RJ Hobert O ‘A microRNA controlling left/right neuronal asymmetry in C. elegans’. Nature 2003 426 6968 845 849 14685240 Kato M de Lencastre A Pincus Z Slack FJ ‘Dynamic expression of small non-coding RNAs, including novel microRNAs and piRNAs/21U-RNAs, during Caenorhabditis elegans development’. Genome Biol 2009 10 5 R54 19460142 Kato M Chen X Inukai S Zhao H Slack FJ ‘Age-associated changes in expression of small, noncoding RNAs, including microRNAs, in C. elegans’. RNA 2011 17 1804 1820 21810936 Kelly WG Xu S Montgomery MK Fire A ‘Distinct requirements for somatic and germline expression of a generally expressed Caernorhabditis elegans gene’. Genetics 1999 146 1 227 238 Ketting RF Fischer SE Bernstein E Sijen T Hannon GJ Plasterk RH ‘Dicer functions in RNA interference and in synthesis of small RNA involved in developmental timing in C. elegans. Genes Dev 2001 15 20 265 269 Knight SW Bass BL ‘A role for the RNase III enzyme DCR-1 in RNA interference and germ line development in Caenorhabditis elegans’. Science 2001 293 5538 2269 2271 11486053 Krause M Hirsh D ‘A trans-spliced leader sequence on actin mRNA in C. elegans’. Cell 1987 49 6 753 761 3581169 Lee RC Feinbaum RL Ambros V ‘The C. elegans heterochronic gene lin-4 encodes small RNAs with antisense complementarity to LIN-14’. Cell 1993 75 843 854 8252621 Li H Durbin R Fast and accurate long-read alignment with Burrows-Wheller transform. Bioinformatics 2010 26 5 589 595 http://dx.doi.org/10.1093/bioinformatics/btp698/ Epub2010 Jan15 20080505 Love ML Huber W Anders S ‘Moderate estimate of fold change and dispersion for RNA-seq data with DESeq2’. Genome Biol 2014 15 12 550 571 25516281 Lucanic M Graham J Scott G Bhaumik D Benz CC Hubbard A Lithgow GJ Melov S ‘Age-related micro-RNA abundance in individual C. elegans’. Aging 2013 5 6 394 411 23793570 Martinez NJ Ow MC Reece-Hoyes JS Barrasa MI Ambros VR Walhout AJM ‘Genome-scale spatiotemporal analysis of Caenorhabditis elegans microRNA promoter activity’. GenomeRes 2008 18 2005 2015 McJunkin K Ambros V The embryonic miR-35 family of microRNAs promotes multiple aspects of fecundity in Caenorhabditis elegans. G3 2014 4 1747 1754 25053708 Miska EA Alvarez-Saavedra E Abbott AL Lau NC Hellman AB McGonagle SM Bartel DP Ambros VR Horvitz HR ‘Most Caenorhabditis elegans microRNAs are individually not essential for development or viability’. PLoS Genet 2007 3 e215 18085825 Pasquinelli AE Reinhart BJ Slack F Martindale MQ Kuroda MI Maller B Hayward DC Ball EE Degnan B Müller P Spring J Srinivasan A Fishman M Finnerty J Corbo J Levine M Leahy P Davidson E Ruvkun G ‘Conservation of the sequence and temporal expression of let-7 heterochronic regulatory RNA’. Nature 2000 408 86 89 11081512 Pena JTG Sohn-Lee C Rouhanifard SH Ludwig J Hafner M Milhailovic A Lim C Holoch D Berninger P Zavolan M Tuschl T ‘miRNA in situ hybridization in formaldehyde and EDC-fixed tissues. Nat. Methods 2009 6 2 139 141 19137005 Pfaffl MW ‘A new mathematical model for relative quantification in real-time RT-PCR. Nucleic Acids Res 2001 29 9 e45 11328886 Pierce ML Weston MD Fritzsch B Gabel HW Ruvkun G Soukup GA ‘MicroRNA-183 family conservation and ciliated neurosensory organ expression. Evol. Dev 2008 10 11 106 113 18184361 Rastogi S Borgo B Pazdernik N Fox P Mardis ER Kohara Y Havranek J Schedl T Caenorhabditis elegans glp-4 encodes a valyl aminoacyl tRNA synthetase. G3 2015 5 12 719 728 http://dx.doi.org/10.1534/g3.115.021899 25740935 Robinson MD McCarthy DJN Smyth GR ‘edge R: a Bioconductor package for differential expression analysis of digital gene expression data. Bioinformatics 2010 26 1 139 140 doi:10.1093 bioinformatics/btp616 epub 2009 Nov11 19910308 Schmittgen TD Livak KJ ‘Analyzing real-time PCR data by the comparative C(T) method’. Nat. Protoc 2008 3 6 1101 1108 18546601 Seydoux G Fire A Whole mount in situ hybridization for the detection of RNA in Caenorhabditis elegans embryos. Development 1994 120 10 2823 2834 7607073 Seyednasrollah F Larho A Elo LL ‘Comparison of software packages for detecting differential expression in RNA-seq stuies’. Brief. Bioinform 2015 16 1 59 70 http://dx.doi.org/10.1093/bib/bbt086 24300110 Shaw WR Armissen J Lehrbach NJ Miska EA ‘The conserved miR-51 microRNA family is redundantly required for embryonic development and pharynx attachment in C. elegans’. Genetics 2010 185 3 897 905 20421599 Sheth U Pitt J Dennis S Priess JR ‘Perinuclear P granules are the principal sites of mRNA export in adult C. elegans germ cells’. Development 2010 137 8 1300 1314 Smith PA Leung-Chiu WM Montgomery R Osborn A Kuznicki KA Gressman-Coberly E Mutapcic L Bennett KL The GLH proteins, Caenorhabditis elegans P granule components, associate with CSN-5 and KGB-1, proteins necessary for fertility, and with ZYX-1, a predicted cytoskeletal protein’. Dev. Biol 2002 251 333 347 12435362 Spike C Meyer N Racen E Orsborn A Kirchner J Kuznicki K Yee C Bennett KL Strome S ‘Genetic analysis of the Caenorhabditis elegans GLH family of P-granule proteins’. Genetics 2008 178 4 1973 1987 18430929 Stein P Rozhkov NV Li F Cardenas FL Davydenko O Vandiver LE Gregory BD Hannon GJ Schultz RM Essential role for endogenous siRNAs during meiosis in mouse oocytes’. PLoS Genet 2015 11 4 e1005129 http://dx.doi.org/10.1371/journal.pgen.1005013 25860788 Strome S Wood WB Generation of asymmetry and segregation of germline granules in early C. elegans. Cell 1983 35 15 25 6684994 Suh N Baehner L Moltzahn R Melton C Shenoy A MicroRNA function is globally suppressed in mouse oocytes and early embryos. Curr. Biol 2010 20 271 277 http://dx.doi.org/10.1016/cub.2009.12.044. pmid:201116247 20116247 Tran ND Kissner M Subramanyam D Parchem RJ Laird DJ Blelloch RH A miR-372/let-7 axis regulates human germ versus somatic cell fates. Stem Cells 2016 http://dx.doi.org/10.1002/stem.2378 (Epub ahead of print) Yuan S Schuster A Tang C Yu T Ortongero N Bao J Zheng H Yan V ‘Sperm-borne miRNAs and endo-siRNAs are important for fertilization and preimplantation embryonic development’. Development 2016 143 4 636 647 Zhao Z Boyle T Liu Z Murray JI Wood WB Waterston RH ‘A negative regulatory loop between microRNA and Hox gene controls posterior identities in Caenorhabditis elegans’. PLoS Genet 2010 6 9 e1001089 20824072

PMC005xxxxxx/PMC5131699.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 101230445 32622 Expert Rev Med Devices Expert Rev Med Devices Expert review of medical devices 1743-4440 1745-2422 27112213 5131699 10.1080/17434440.2016.1174572 NIHMS828769 Article A review of the progression and future implications of brain-computer interface therapies for restoration of distal upper extremity motor function after stroke Remsik Alexander Young Brittany Vermilyea Rebecca Kiekoefer Laura Abrams Jessica Elmore Samantha Evander Schultz Paige Nair Veena Edwards Dorothy Williams Justin Prabhakaran Vivek Department of Radiology Clinical Science Center, University of Wisconsin Madison School of Medicine and Public Health Ringgold Standard, Institution, Madison, WI, USA 17 11 2016 5 2016 01 5 2017 13 5 445454 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. Stroke is a leading cause of acquired disability resulting in distal upper extremity functional motor impairment. Stroke mortality rates continue to decline with advances in healthcare and medical technology. This has led to an increased demand for advanced, personalized rehabilitation. Survivors often experience some level of spontaneous recovery shortly after their stroke event; yet reach a functional plateau after which there is exiguous motor recovery. Nevertheless, studies have demonstrated the potential for recovery beyond this plateau. Non-traditional neurorehabilitation techniques, such as those incorporating the brain-computer interface (BCI), are being investigated for rehabilitation. BCIs may offer a gateway to the brain’s plasticity and revolutionize how humans interact with the world. Non-invasive BCIs work by closing the proprioceptive feedback loop with real-time, multi-sensory feedback allowing for volitional modulation of brain signals to assist hand function. BCI technology potentially promotes neuroplasticity and Hebbian-based motor recovery by rewarding cortical activity associated with sensory-motor rhythms through use with a variety of self-guided and assistive modalities. Stroke background Stroke, resulting from the cessation of blood flow to cortex as a result of clotting (ischemic stroke) or bleeding (hemorrhagic stroke), is a serious medical emergency that can result in death or substantial neural damage and remains a leading contributor to acquired disability in the United States [1–3]. More than 795,000 individuals are affected by stroke annually [1] and approximately 610,000 [2] are first attacks. The total cost of stroke to the United States is estimated at $34 billion (USD) per year [2]. Despite technological advancements in health care and increased preventative measures such as education and public emphasis on healthy living practices, the incidence of stroke is anticipated to rise annually, increasing by 20% as early as 2030 [1]. Demographic factors will likely contribute to this trend with an increasing elderly population in America [4]. However, stroke rates have declined significantly in persons 60 years and older, but largely persist in adults aged 45–59 years of age [2]. Mild-to-severe physical, cognitive, and affective impairments are common consequences of stroke insult, and more than half of stroke survivors experience some level of lasting hemiparesis or hemiplegia [5,6]. To meet this increasing public health challenge, it is imperative that effective rehabilitations and treatment methodologies are developed to address each stage of post stroke recovery: chronic and nonchronic (acute and subacute), and adapted to each level of severity of impairment: mild, moderate, and severe. Stroke often results in a multiplex of motor, sensory, cognitive, and other impairments resulting from damage to neural tissues. Therefore, a successful rehabilitation therapy must increase functional capacities by promoting gradual adaptation of the brain’s remaining neuronal connections [7]. Effective rehabilitation and treatment methodologies would aim to increase the quality of life for individuals after stroke. Brain–computer interface (BCI) therapy offers a unique, multimodal, and multisensory platform for rehabilitative therapy after stroke and will thus be the focus of this review. BCI principles The human brain is largely accepted to be a plastic, appetitive, hedonistically driven feedback and feedforward circuit, especially susceptible to punishment or reinforcement and scheduled reward. Even though stroke survivors often present with damaged cortex or disrupted motor connection integrity, noninvasive electroencephalogram (EEG)-based BCIs are still capable of detecting significant and reproducible change. Potentially relying on the ability of residual motor neurons to fire and facilitate device control, BCIs help to train persisting cortical connections to execute motor output of the hand [4,7,8]. BCI therapies take advantage of the brain’s ability to associate novel and independent stimuli and the goal-directed nature of motor execution to create an environment in which motor skills can be trained, performed, and reinforced. Learning results from a change in behavior dependent on punishing or rewarding experiences. BCIs capitalize on the brain’s natural instinct to discern adaptive from maladaptive, or unsuccessful, strategies over time [7,9,10], due to the scheduled reinforcement provided by the BCI task. EEG-based BCIs function by establishing a closed-loop neural interface. A BCI uses raw functional cortical activity recorded by the EEG and translates it into a classified device command designed to circumvent or aid neuromuscular efferents potentially compromised by stroke [4]. Signal input is amplified and processed by a regression model [11–15] that extracts particular amplitude changes or features and accounts for signal noise. Specific features of the signal recorded from the subject’s scalp are selected against this noise, and information from the input data is selected and classified using an algorithm and parameters specified in the programming of the BCI device. With this real-time processing, the reduced representation of brain activity is effectively translated into an output or feedback modality, often one that allows a desired task to be performed more easily. EEG electrodes placed over the sensorimotor regions provide the most localized and reliable functional cortical activation changes relevant to the hand’s motor function and are most often selected to control the external device. When considering BCI use with additional modalities, lesion location and integrity of remaining neural pathways may limit the potential for normal function as well as the potential for recovery. EEG-based BCI is one of the most studied and popular BCI systems currently on the market. EEG-based BCI is most commonly used because it is cost-effective, noninvasive, portable, and has shown to be effective in improving motor function post-stroke [16,17]. Multiple studies have observed significant increases in the Fugl-Meyer Assessment (FMA) of Motor Recovery after Stroke and the Action Research Arm Test (ARAT) following EEG-based BCI therapy [11,18,19]. Treatment time in these studies varied greatly, suggesting that a minimal amount of EEG-BCI could create noticeable results in participants. While EEG-BCI appears to create noticeable improvements in upper extremity (UE) motor function in participants, learned nonuse is a phenomena which presents in stroke patients that offers an example of how environmental conditions might limit or allow for motor recovery. It is plausible that learning to associate movement intention with the successful completion of a BCI task or behavioral output is a possible means for recruitment of vestigial neuronal pathways preserved after insult or disease. Such pathways are very small remnants of neuronal pathways that were once much larger and are the compromised neurological and neuroanatomical profiles of typical BCI users that may exist following the neurological trauma which resulted in the user’s distal motor impairment. The possible recruitment of these vestigial pathways has the potential for the restoration of functional motor capacities, selection of letters from an array [2,20], or movement of a virtual cursor [7,12,21–26]. Furthermore, BCIs provide real-time feedback to the user and reward consistent production of neural features concordant with hand motor function. Therefore, apparent changes in functional cortical activation patterns may persist after therapy when attempting tasks similar to those trained with BCI therapy [7,11,27,28]. This theoretical knowledge supports the possibility of inducing lasting brain changes through BCI system and regimen. The necessary functional connectivity changes induced in stroke patients with lasting recovery of hemiparesis remain unclear, though mechanisms and strategies have been proposed [4]. Learning mechanisms Therapies that incorporate BCI devices can be explained by conventional learning theories and replication of the BCI-driven motor learning outcomes observed in the healthy brain [7,12,29]. Reinforcement of motor behaviors is a key mechanism evident in the training and use of BCI therapies. Simple Pavlovian conditioning, or learning a new behavior through association and reinforcement, is a primary mechanism for developing associations and expectations common to us as humans. This Pavlovian conditioning allows a BCI user to fully integrate themselves with the neural interface. The basic theoretical design of BCI paradigms operate on the principle that targeted functional cortical activation of the motor areas should result in task completion or facilitated motor output, and insufficient activation should not produce any significant change in the BCI’s behavior. Sufficient and targeted functional cortical activation is programmed to generate a successful manipulation of the BCI paradigm and activation of an assistive or augmentative device. The intrinsic reward of success is expected to guide motivation and behavior, the same way a developing brain might learn to interact with a novel environment to manipulate or engage with what is contained within that environment. Furthermore, motivation is important for motor reeducation as it assists in the recruitment of functional and residual neural pathways. Pairing a stimulus with reinforcement or punishment is integral to human learning. The most efficient known mechanism to facilitate such learning is persistent and repeated transmission between pre- and postsynaptic cells. ‘BC I-induced Hebbian neural recovery’ [4,19,24,30] posits that the amount of reinforcement and the timing or schedule of reinforcement can significantly impact the efficiency and specificity of learning [31]. This basic mechanism of synaptic plasticity can be assumed to operate in stroke afflicted brains similar to the way it operates in the healthy brain [32]. In fact, neuroimaging studies using functional magnetic resonance imaging (fMRI) have shown increased cortical activation in areas damaged by stroke following BCI therapy and training [4,7,12,31,33]. Specifically, Hebbian learning may facilitate rehabilitation in stroke survivors by retraining or recreating synaptic connections necessary for the functional cortical activity essential for smooth and controlled motor output [4,7,12,19,24,30,31,33,34]. This mechanism may offer predictive indications about the relative likelihood of extinction or retention of the newly learned behavior [16,27,31,33–36]. Because motor output is cued by the BCI paradigm, facilitative therapy can be administered as soon as the necessary cortical activity is recognized by the BCI classifier. Proprioceptive feedback and stimulation are associated with eliciting UE motor and hand function. It is understood that the reward of these ‘targeted activations’ acts to improve the likelihood of functional cortical activation, BCI task completion, and subsequent reinforcement provided by the task’s parameters. Presumably, even in trials where little or no motion is realized or facilitated, individuals might experience recovery of functional cortical activity or augmentation of existing functionality, attributable to BCI system therapies. If a reward does not present itself immediately (i.e. hitting the virtual target), the participant may still experience a positive effect because of the sensory input to their hand. This rehabilitation regimen serves the purpose of real-time reinforcement of satisfactory cortical oscillations for motor output and assisted completion of a rehabilitative task to expedite and focus the user’s motor learning as a means of maximizing the therapeutic and rehabilitative effect of the therapy session. The better a user’s movement intention is paired with assistive therapy, as executed by the BCI, the more likely cortical plasticity changes will occur, according to the Hebbian, ‘fire together, wire together’ rule [4]. As stated, this is particularly evident in BCI-based therapies incorporating functional electrical stimulation (FES) or other modalities for rehabilitation of motor function. When a feature signal is detected over the sensorimotor region of the cortex, stimulation of the distal hand muscles facilitates physical muscle contraction [12,24,37–40]. The appropriate moment for administration of facilitative therapy can be inferred by the BCI and administered in an iterative process because motor output is cued by the user’s direct neural input to the BCI paradigm. Pairing of stimulation with activation, as inferred by the BCI, acts to close the feedback loop of a normal motor program. A normal motor program includes planning of motor intent, initiating movement, monitoring progress, and ending with the recognition of successful execution. As indicated by the literature reviewed herein, such targeted paradigms result in recovery when administered to patients presenting with hemiparesis [19,24,30]. BCI technology for rehabilitation BCI technology is well suited for neural rehabilitation post stroke as it utilizes the user’s direct neural input for the purpose of manipulating a peripheral component, such as a user’s hand. Similar to the way the central nervous system operates the hand through physiological circuitry of the peripheral nervous system, the BCI executes similar actions via a device command. Noninvasive ‘hijacking’ of the brain’s residual functional connections by a BCI may be used to support the recovery of functional capacities in the brain such as voluntary motor function through goal-directed practice and training. Such connections have the potential to increase voluntary motor function, as BCI invokes the same neural mechanisms that control volitional movement of a hand. By encouraging motor-related functional cortical activity for the completion of a defined rehabilitation task, BCIs invoke the same neural mechanisms that control volitional movement of a hand. Utilizing BCI to mediate the user’s intention to move their hand and subsequent output acts to close the neural feedback loop that is potentially compromised after the stroke insult. BCI-mediated therapy can recruit the brain’s natural explicit and procedural learning mechanisms along with memory mechanisms to enhance functional capacity of a hand. BCIs can be used to train a patient to maximize the potential recovery of functional motor movement in their paretic hand. BCI can be coadministered with established interventions or paired with more novel, research-based, home-written, or home-brewed tasks. BCI therapies can incorporate various therapeutic interventions in conjunction with traditional EEG-based BCIs or as adjuvants including: virtual reality [20,22,28,35], constraint-induced movement [9,11,41,42], robot-assisted movement therapy [8,30,43–45], and FES [5,12,24,39]. BCI therapy tasks are designed to recruit multiple sensory systems, such as visual and tactile. For example, the combination of a visual display along with the tongue display unit (TDU) provides tactile and visual input, making it an immersive multisensory intervention [46]. They provide an environment that reinforces successful motor intention and output by priming or cueing a motor intention, which encourages formation and execution of a motor plan. Implementing relevant behavioral outcome assessments, such as ARAT, 9-Hole Peg Test, FMA, and others, can establish the efficacy of BCI therapies on functional outcomes and quality of life [7,22,27]. Additionally, future studies utilizing pre- and post-therapy functional neuroimaging scans, fMRI and DTI for example, are ideal for a more comprehensive understanding of the cortical plasticity mechanisms by which BCI therapies affect change in recovering stroke patients. Neuroimaging measures are principal for distinguishing post stroke changes, such as those in functional activation and connectivity, gray matter and white matter structural integrity, and perfusion. In this review, various BCI treatment modalities are considered and compared with respect to restoration of UE motor impairment as a result of stroke. Efficacy in stroke rehab Various studies have demonstrated clinical efficacy and support for use of BCI in stroke rehabilitation (see Table 1). Early clinical BCI studies showed evidence for the feasibility and potential efficacy of BCI combined with FES and motor imagery (MI)-based BCI in combination with physiotherapy and robotic orthoses for motor recovery post stroke [22,38,43,44,47,48]. Because BCIs using FES and other modalities close the neural feedback loop, proper neural signals can be reinforced through the assisted execution of the user’s motor plan. Recent studies have found similar outcomes regarding BCI effectiveness as the earlier BCI studies. The effectiveness of BCI-mediated interventions was reviewed as a progression from self-guided movement (MI-based BCI) to assisted movement (BCI-FES and BCI with orthoses). MI utilizes self-guided movement (i.e. the user’s imagination) of their paretic limb. MI has yielded promising results when used in conjunction with BCI. MI-based BCIs act to facilitate rehabilitation of UE motor function through the pairing of MI with completion of a simple motor task [13,22]. In many BCI therapies, MI-based BCI is used as a training condition or as a means for setting up control signals and classifiers for subsequent therapy conditions or administrations [10,16,17,32,38,48]. A review by Teo and Chew suggests that combining MI with BCI is a feasible intervention for improving hand motor function in the post stroke population, particularly with those in the chronic phase of stroke with mild-to-moderate severity [49]. While recent evidence supports the potential efficacy of MI based BCI in use with mild-to-moderate stroke for distal UE motor function, there are mixed findings on the effectiveness of MI combined with BCI for improving distal UE motor function in individuals with severe stroke, when used in solidarity. It is suggested that those with severe stroke may need more assistance to produce functional UE movement. A recent study by Ang et al. examined the efficacy of a MI-BCI system combined with a Haptic-Knob robotic hand interface (BCI-HK) in restoring UE motor function for individuals with chronic stroke and moderate to severe UE impairment. The outcomes indicated significant motor function gains in FMA of Motor Recovery after Stroke (FMMA) scores for both the proximal and distal UE, suggesting BCI-HK therapy is effective for improving UE motor impairment in individuals with chronic stroke when combined with therapist-assisted UE mobilization [50]. In addition, Ang et al. examined the efficacy of a MI-BCI system combined with shoulder–elbow robotic feedback (BCI-Manus) in restoring UE motor function for individuals with chronic stroke. The findings indicated significant motor function gains in FMMA scores, suggesting BCI-Manus therapy is effective for improving UE motor impairment in individuals with severe, chronic stroke [8]. Though this study was specific to proximal UE rehabilitation, systems for robotically guided distal UE movement may be similarly effective for severe stroke distal UE rehabilitation. An earlier study by Mihara et al. found greater functional gains on the hand/finger subscale of the FMA following six sessions of mental practice with near-infrared spectroscopy (NIRS)-mediated MI of the distal UE in addition to standard rehabilitation. A significant effect of neurofeedback was also found in those severely impaired, suggesting MI plus NIRS could be useful for a wide variety of stroke impairment [51]. A study by Naseer and Hong found enhanced performance of MI wrist classifiers in developing a BCI, demonstrating the feasibility of a functional NIRS (fNIRS)-based BCI [52]. These studies suggest that fNIRS-mediated MI-BCI therapies in adjuvant to standard rehabilitation may be more effective in addressing severe distal UE motor impairment than MI-based BCI alone. In addition to self-guided movement, BCI can be paired with assistive movement such as FES and orthoses. FES utilizes real-time feedback of BCI signal input to selectively administer therapeutic feedback responses only when the correct brain signals are detected [7,12,24,39,40,53,54]. The FES is cued by the BCI in response to its recognition of classified cortical activation features [7,24,39]. Further evidence for the effectiveness of BCI-FES in improving distal UE motor function has emerged recently. Biasiucci et al. found improvements in finger extension in the FMA after 10 sessions of FES –controlled BCI for individuals in the chronic phase of stroke [39]. Another study using FES -controlled BCI along with tongue stimulation found clinically significant improvements on behavioral measures including ARAT and the Stroke Impact Scale hand function domain after 18–30 h of treatment [37]. However, one requirement for FES to be effective is the user’s capacity for residual movements, which is not always plausible for individuals with severe motor impairments [55]. Due to this limitation, the use of hand orthotics is becoming increasingly used in conjunction with BCI-FES interventions. Orthotics combined with BCI are showing promising results for distal UE function. Shindo et al. used orthotics with BCI to elicit finger extension resulting in increased function and decreased spasticity. These findings also corresponded to increased excitability in the brain, which reflect the BCI principles of impacting brain plasticity through learning mechanisms to elicit improved motor function [17]. Similarly, Ramos-Murguialday et al. used hand and arm orthoses in severe stroke patients and found clinically significant improvements in FMA scores that correlated with plasticity changes in the brain [56]. Another study comparing different feedback types found that somatosensory feedback through an orthosis was more effective at improving finger motor function than animated visual feedback on a computer screen for individuals with chronic stroke. Although both groups had enhanced brain activation following BCI treatments, the difference in functional UE motor scores suggest that certain types of sensory feedback may better facilitate motor reorganization in the brain [57]. While BCI interventions have shown to be effective with both self-guided movement (i.e. MI) and assistive movements (i.e. FES and orthotics), some studies have found that BCI interventions have not shown distinct improvements as effectively within the stroke population. Ang et al. found that a 2-week BCI-MI intervention with transcranial direct current stimulation did not elicit significant motor improvements within a stroke population that included both acute and chronic survivors. However, although not significant findings, the researchers found that participants had better accuracy in the BCI-MI task and increases in laterality coefficients, reflecting cortical activation related to motor planning [58]. The short length of the study may be the reason behind the lack of physical motor improvements, but these findings show promise for future directions BCI research may take due to changes in brain activation. Other studies have raised the question of pre requirements and individual characteristics needed for BCI to be effective. A review by Ahn and Jun explored the idea of ‘BCI-illiteracy,’ referring to certain users that have more difficulty in controlling the BCI system due to variabilities in physiological and psychological characteristics. One such finding suggests that individuals that are successful BCI performers are better at recruiting MI-related brain networks and those that are ‘BCI-illiterate’ have less developed networks to recruit from. Other variables contributing to BCI performance success include motivation and the use of tactile and visual feedback modalities. Together, these elements may contribute to a user’s level of concentration, which may lead to greater BCI success [59]. Collectively, current research demonstrates that BCI has the potential to harness the reserve of recovery potential left after stroke insult [60]. A powerful platform for motor recovery post-stroke has been created through a combination of individual learning mechanisms, BCI principles, and various modalities such as FES and MI. Limitations Several themes emerged in review of previous studies on BCI therapies. Several studies studied the chronic stroke population, disregarding the mild-to-moderate stroke population. This has implications for the potential for recovery. In addition, several studies investigated BCI as an adjunct to traditional physiotherapy. When studying BCI in conjunction with traditional physiotherapy, it is difficult to discern the outcomes as resulting from BCI alone, traditional physiotherapy alone, or a combination of the two treatments. While the implications of BCI appear promising, one major limitation of BCI is underpowered studies due to small sample sizes [4,9,22,25,26,34,38,39,61,62]. Additional studies with larger and more heterogeneous samples are required to determine if adults’ post-stroke can successfully use BCI for motor learning and in particular using Hebbian-type learning [4,9,22,26,34,38,61,62]. More rigorous RCTs are required for BCI to have increased impact and power on the field of stroke neurorehabilitation. While BCI technology appears impactful on improving motor control for individuals post-stroke, there are more methodological studies on healthy individuals than individuals with stroke. In addition, many BCI methods studies focus on upper limbs and few targeting the lower limbs [63]. Therefore, the efficacy of BCI treatment should be examined in future studies. Recent studies have found that structured, task-oriented rehabilitation programs and EEG-biofeedback systems (neurofeedback) do not significantly improve UE motor function or recovery in adults’ post-stroke compared to conventional rehabilitation [64,65]. In addition, the efficacy of BCI combined with conventional therapy is unclear due to difficulty discriminating between outcomes due to BCI and outcomes due to conventional therapy [63]. Determining the optimal treatment dosage is another area that will require future BCI research. Currently, recommendations for treatment dosage vary greatly across studies. Previous studies have implemented treatment across 11–22 sessions [7,12,31,37,47], while others have found successful outcomes from fewer than 10 sessions [38]. More research on the frequency of BCI treatment is needed to determine the optimal amount to best promote UE motor rehabilitation. Additionally, these variations in study design have created inconclusive findings on the lasting effects of BCI therapy, in the absence of frequent motor rehabilitation [47]. Further, the generalizability of findings are inconclusive based on low participant numbers and varied dosing and study designs [11,62]. Expert commentary BCI systems are likely advantageous over standard stroke interventions due to their ability to engage multiple learning modes. Based in Pavlovian conditioning and facilitating Hebbian learning mechanisms, BCI therapies use the goal-directed nature of motor execution and the brain’s ability to associate novel and independent stimuli to create an environment in which motor skills can be trained, performed, and reinforced. In addition, BCI technology seems well suited for neural rehabilitation post-stroke as it utilizes the user’s direct neural input for the purpose of manipulating a peripheral component, creating a closed-loop feedback system between the CNS and PNS. Noninvasive reading of the brain’s residual CNS activity, commonly through an EEG cap, paired with external sensory (visual, tactile) input to the PNS may be used to support the recovery of functional capacities in the brain such as voluntary motor function. BCI can be coadministered with established intervention therapies as well as more novel tasks, including biofeedback or constraint-induced movement therapy. To date, research suggests various modalities used in conjunction with a BCI system are effective for UE stroke rehabilitation, including coadministration with FES, TDU, and MI. Certain studies are also beginning to incorporate neuroimaging systems such as fMRI. Currently, laterality index and activation maps are some of the most commonly studied fMRI data. BCI may be used to support the recovery of functional capacities in the brain such as voluntary UE and hand motor function through goal-directed practice and training, which in turn in is thought to improve quality of life. Ongoing research to evaluate the effectiveness of BCI-based stroke rehabilitation for hand therapy is currently in progress and must be a priority now and in the future. There is a demand for larger randomized control trials and clinical-based trials with BCI-based interventions. There is room for potential use of BCI-based interventions on a smaller, more personalized scale as technology improves and allows for increased portability and compatibility. This may lead to the use of BCI interventions in clinics or homes in the near future. With increased accessibility to these therapies, a more neurologically diverse population may have the opportunity to experience an increased quality of life from the outcomes of BCI interventions. Five-year view Ongoing research to evaluate the effectiveness of BCI-based stroke rehabilitation for hand therapy continues in earnest. Thus, it is important to direct development toward necessary improvements in BCI methodologies to address efficacy, reproducibility, and the identification of the specific mechanism of therapeutic action. Improvements and revisions, including relevant behavioral outcome measures and pre- and post-therapy fMRI, in future investigations are essential to enhancing device performance and improving the rehabilitative impact of each therapy session. Future BCI intervention studies will need to incorporate stroke patients from a wide range of neurological and demographic profiles. In addition, future research must focus on BCI methodology with use of subjects with stroke, rather than solely healthy individuals. Future research populations must differ in chronicity, severity, lesion location, as well as number of stroke insults to better formulate generalizable trends and outcomes. Further addressing different subpopulations could elicit clarification in the relationship between BCI therapy and the nature of neuroplasticity, which could allow for eventual tailoring of individual neurorehabilitation programs for stroke patients to realize maximal recovery. In the future, these paradigms may include those with different neurological impairments such as quadriplegia, which often manifest as exclusion criteria for many prior BCI studies [29,31,66]. Severely impaired patients previously unable to engage will experience continually improving interpersonal interactions. Moreover, it may be possible to improve BCI device design through further comparisons between spatial and frequency patterns of neural activity derived from task performance among different subpopulations. Future research exploring the multiple impacts of BCI on a wider range of severity will explore the different impacts BCI therapy has on individuals with varying levels of severity. For example, BCI may have different impacts on individuals with mild motor impairments compared to individuals with more severe motor impairments. To date, an increasing number of studies successfully explore various modalities with which to use a BCI system as a neurorehabilitation device for stroke. Such modalities, such as fMRI, EEG, EMG-triggered orthotics, and FES, aim to increase function and quality of life for individuals unable to engage in tasks due to CNS damage. Several examples demonstrate the flexibility of BCIs as a dynamic therapy. For example, a case study discussed by Reiss et al. [9] demonstrated how noninvasive BCI allowed individuals experiencing locked-in syndrome, due to a brainstem stroke, to produce messages. This important innovative function of BCI provided autonomy for these survivors, although they often required assistants from a health-care provider and at least some level of persisting ocular motor or musculoskeletal control. Such examples demonstrate the flexibility of BCIs as a dynamic therapy for stroke survivors, a patient population often presenting with a multiplex of impairments and comorbidities. Neuroimaging measures, such as NIRS, fMRI, and real-time fMRI, hold promise for maximizing BCI treatments by measuring and representing the rehabilitative capacities of BCI therapy on a patient-to-patient basis. Neuroimaging of functional brain organization will be used by therapists to make personalized adjustments to the BCIs in order to maximize the therapeutic effect of treatment. Simultaneous fMRI-EEG technology may be incorporated with BCIs in the future to image brain changes in an increased temporally relevant manner. This proposed methodology takes advantage of the spatial resolution of fMRI with the temporal resolution of EEG. Simultaneous EEG-fMRI should be studied further to establish validity and to reduce its cost. Advancements in neuroimaging measures are still forthcoming and their incorporation with emerging technologies such as BCI may provide a dynamic approach for establishing the ‘best fit’ therapy for a patient. Furthermore, future studies investigating EEG-fMRI must investigate if the neural changes centrally coexist with functional outcomes peripherally. Future studies must investigate the interaction between neural changes, functional outcomes, and self-efficacy in adults’ post-stroke after using BCI. A recent study found those with hand paresis after stroke improved more when they had combined EMG-triggered electrical stimulation with task-oriented training [67]. While EMG-triggered electrical stimulation is not driven by a traditional BCI, there is an emphasis on the potential for recovery when combining various modalities. The combination of BCI therapies with a wearable ‘smart glove’ orthotic is of particular interest. ‘Smart gloves’ are wearable biofeedback devices that record kinematic data associated with movements of the hand, wrist, and digits and can be used along with virtual reality to retrain motor patterns. Such devices and therapeutic combinations have been found to decrease levels of motor impairment and increase quality of life [68]. Administration of such a BCI-FES therapy may allow for rehabilitation of the upper and lower extremities as well as improve a patient’s posture and gait [24]. Coordinated by a BCI-FES system, the user is provided with a tactile, proprioceptive sense of their movements. It is also possible that impaired muscle groups will receive focused and timed electrical stimulation to facilitate contraction of dynamic muscle groups to perform complex motor tasks. Synchronized by a BCI, such FES pulses may be used to assist abated movement or augment normal contraction. Recent evidence, also reviewed herein, suggest BCI-FES can potentially induce neural improvements associated with motor function. Through this combination, BCI has future implications in motor recovery and muscular reeducation for those neurological deficits caused by stroke, as well as other neurological conditions. When considering what devices to combine with BCI, initially, it is important to consider feasibility and cost, as individuals are less likely to use treatments that are expensive or cumbersome to operate [69]. As research in the field progresses, the authors are hopeful that BCI therapy will become available as an affordable and effective in-home treatment. The authors suggest that development of a portable BCI system for in-home use will primarily consist of a laptop computer or tablet-like device containing appropriate software connected to a 16-channel (or fewer) EEG system with a stable, robust, and durable electrode array and amplifier. Human–device interaction might be realized as an augmentative implementation of BCIs as BCI-system software advance over time to become more standardized and user friendly. It may soon be feasible for stroke survivors to dial phone numbers, answer correspondences, or even operate home appliances with the help of a BCI. To ensure reliability and safety in the usage of in-home treatments, it is initially suggested that health-care professionals and eventually caretakers, such as family members or hospice staff, are trained to oversee the BCI therapy operations and to ensure proper regulatory compliance is adhered to. BCI technology provides patients with a range of impairments and across a spectrum of stroke chronicity the opportunity to benefit from a professionally prescribed, clinically designed, and individualized neurorehabilitation regimen. The authors believe that the most significant advancements over the 5 years following this review will be the transfer of noninvasive BCI intervention therapies from clinics and research labs to in-home and increasingly personalized treatments. These efforts are essential for continuing the advancement of BCI technologies toward improving the quality of life and sense of autonomy for stroke survivors in their daily living activities. This review suggests rich promise for the future of BCI technology as a treatment modality for distal UE motor impairment following stroke insult. Table 1 Reviewed studies of efficacy of BCI in stroke rehabilitation. Study Stroke chonicity N (sex) System and regimen Did therapy result in noted improvements? Are BCI therapy results statistically significant at exit? Outcomes and behavioral outcome measures of interest Ang et al. (2015) Chronic N = 19 10 sessions of tDCS or sham before 1 h of MI-BCI with robotic feedback for 2 weeks (sham-controlled, RCT) No No UE FMMA Ang et al. (2014) Chronic N = 21 M = 14 F = 7 3-arm RCT; MI-BCI with HK, HK, and standard therapy; 18 sessions of 1 h interventions Yes, BCI-HK group Yes, BCI-HK group FMMA Biasiucci et al. (2013) Chronic N = 4 At least 10 sessions of FES controlled by BCI over a period of 2 months Yes N/A FMA Broetz et al. (2010) Chronic N = 1 M = 1 F = 0 1 h physiotherapy with each BCI session Yes No FMA, Wolf motor function test, modified Ashworth scale Buch et al. (2008) Chronic N = 8 13–22 sessions MEG-BCI with hand orthosis over 3–8 weeks Yes No In-house measure of success (>50% target ‘hit’) Bundy et al. (2012) Chronic N = 4 Subjects performed between 85 and 246 control trials of BCI Yes Yes ARAT Caria et al. (2011) Chronic N = 1 M = 1 F = 0 After 2–4 weeks MEG-BCI, case study had EEG-BCI and 4-week periods of EEG-BCI 3 & 9 months later. 1 h physiotherapy w/each BCI session Yes Yes FMA; Wolf motor function test; modified Ashworth scale, and goal attainment score Daly et al. (2009) Chronic N = 1 M = 0 F = 1 Nine 45-min BCI-FES sessions over 3 weeks. Also weekly 1.6 h of non-BCI FES therapy Yes Yes Isolated movement index finger extension Liu et al. (2012) Sub acute and chronic N = 314 12–20 weekly or twice-weekly 1 h sessions using EEG-BCI triggering a hand orthosis for finger extension over 4–7 months Yes Yes, significance was found in FMA and ARAT, but no significance indicated in MAL FMA, ARAT, and MAL-14 Mihara et al. (2013) Chronic N = 20 6 sessions of NIRS-guided BCI with mental practice with MI + standard rehabilitation Yes Yes FMA and ARAT Murlidharan et al. (2011) Chronic N = 4 1 session a week for 4 weeks of BCI Yes Yes FMA Ono et al. (2014) Chronic N = 12 Visual feedback and somatosensory feedback groups. Each group received 12–20 sessions of 1 h length Yes, motor improvements in somatosensory group N/A SIAS for finger function Prasad et al. (2010) Chronic N = 5 M = 4 F = 1 2 treatment sessions each week of BCI-MI + PP for a total of 6 week Yes Yes Motricity index ARAT, 9-Hole Peg Test grip strength; fatigue and mood and qualitative feedback Ramos-Murguialday et al. (2013) Chronic N = 32 17.8 ± 1.4 days of training with BCI with an orthotic Yes Yes FMA Rayegani et al. (2013) N/A N = 30 10 sessions of conventional OT, in addition to either EMG-biofeedback therapy or neurofeedback therapy Yes Voluntary contraction of abductor pollicis brevis increased significantly after EMG-biofeedback therapy Jebsen hand function test Shin do et al. (2011) Chronic N = 8 M = 8 F = 0 12–20 weekly or twice-weekly 1 h sessions over 4–7 months using EEG-BCI triggering a hand orthosis for finger extension Yes Yes SIAS; MAL amount of use; modified Ashworth scale; resting motor threshold Young et al. (2014) Chronic N = 11 M = 8 F = 3 At least nine- and up to fifteen 2-h sessions of interventional BCI + FES + TS Yes Yes Stroke impact scale; ARAT; 9-Hole Peg test; laterality index N: Number; tDCS: transcranial direct current stimulation; MI: motor imagery; BCI: brain–computer interface; RCT: randomized control trial; UE: upper extremity; FMMA: Fugl-Meyer Motor Assessment; M: male; F: female; HK: Haptic-Knob robotic arm; FES: functional electrical stimulation; MEG: magnetoencephalograms; EEG: electroencephalogram; FMA Fugl-Meyer Assessment; ARAT: Action Research Arm Test; MAL: Motor Activity Log; NIRS: near-infrared spectroscopy; SIAS: Stroke Impairment Assessment Set; PP: physiotherapy; EMG: electromyography; TS: tongue stimulation. Key issues Despite advances in medical practice and technology, stroke incidence remains a leading cause of major disability and death worldwide. Though stroke mortality is declining, stroke incidence is increasing and so too are the health care costs associated with treatment and patient rehabilitation. BCIs offer a non-invasive, closed-loop neural interface with an option to include a prosthetic device, robot, or other machine to further facilitate rehabilitation therapy. BCI-mediated therapy offers the personalizable and adaptable therapy platform required of a modern restorative physiotherapy. BCIs operate by direct integration of the brain and an external computer or other devices for the purpose of rehabilitation of hand and other upper extremity motor impairments. BCI’s method of action is to train controlled neuromodulation via psychophysiological integration of motivation, learning mechanisms, motor plan rehearsal, proprioceptive feedback, and reward systems in order to restore and orchestrate intentional movement of hemiparetic distal extremities. BCI holds great promise as a future cost-effective, in-home, adaptive, augmentative medical device platform for stroke survivors. BCI and its possible adjuvants provide a rich suite of personalizable rehabilitations capable of incorporating existing physiotherapies as well as future therapies. Several BCI methodological studies were conducted on healthy subjects, which posits the needs for more methodological studies conducted on individuals post-stroke for future clinical studies Declaration of interest V. Prabhakaran and J. Williams have a pending US patent on the closedloop neurofeedback used for BCI-facilitated intervention, application number 12/715090. The patent was filed jointly by both V. Prabhakaran and J. 685 Williams. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed. 1 Go AS Mozaffarian D Roger VL Heart disease and stroke statistics–2013 update: a report from the American Heart Association Circulation 2013 127 e6 e245 10.1161/AQ2CIR.0b013e31828124ad 23239837 2 Mozaffarian D Benjamin E Go A Heart disease and stroke statistics–2015 update: a report from the American Heart Association Circulation 2015 131 e29 e322 25520374 3 NINDS stroke information page Stroke information page National Institute of Neurological Disorders and Stroke (NINDS) Available from: http://www.ninds.nih.gov/disorders/stroke/stroke.htm 4 Soekadar S Birbaumer N Slutzky MW Brain-machine interfaces in neurorehabilitation of stroke Neurobiol Dis 2014 cited 2014 Dec 7 10.1016/j.nbd.2014.11.025 0969-9961 5 Kelly-Hayes M Beiser A Kase C The influence of gender and age on disability following ischemic stroke: the Framingham study J Stroke Cerebrovasc Dis 2003 12 119 126 705 17903915 6 Soekadar SR Birbaumer N Cohen LG Brain-computer-interfaces in the rehabilitation of stroke and neurotrauma Psychology 2011 48 6 1 19 7 Young B Nigogosyan Z Remsik A Changes in functional connectivity correlate with behavioral gains in stroke patients after therapy using a brain-computer interface device Front Neuroeng 2014 7 25 25071547 8 Ang K Chua K Phua K A randomized controlled trial of EEGbased motor imagery brain-computer interface robotic rehabilitation for stroke Clin EEG Neurosci 2014 9 Reiss AP Wolf SL Hammel EA Constraint-induced movement therapy (CIMT): current perspectives and future directions Stroke Res Treat 2012 2012 159391 10.1155/2012/159391 22577601 10 Várkuti B Guan C Pan Y Resting state changes in functional connectivity correlate with movement recovery for BCI and robot-assisted upper-extremity training after stroke Neurorehabil Neural Repair 2013 27 53 62 22645108 11 Bundy D Wronkiewicz M Sharma M Using ipsilateral motor signals in the unaffected cerebral hemisphere as a signal platform for brain-computer interfaces in hemiplegic stroke survivors J Neural Eng 2012 9 12 12 Young B Nigogosyan Z Walton L Changes in functional brain organization and behavioral correlations after rehabilitative therapy using a brain-computer interface Front Neuroeng 2014 7 26 25076886 13 Kaiser V Kreilinger A Müller-Putz GR First steps toward a motor imagery based stroke BCI: new strategy to set up a classifier Front Neurosci 2011 5 10 21390289 14 Kasuga S Matsushika Y Kasashima-Shindo Y Transcranial direct current stimulation enhances mu rhythm desynchronization during motor imagery that depends on handedness Laterality 2015 1 16 15 Kreilinger A Kaiser V Breitwieser C Switching between manual control and brain-computer interface using long term and short term quality measures Front Neurosci 2012 6 11 22347157 16 Caria A Weber C Brötz D Chronic stroke recovery after combined BCI training and physiotherapy: a case report Psychophysiology 2011 48 578 582 20718931 17 Shindo K Kawashima K Ushiba J Effects of neurofeedback training with an electroencephalogram-based brain-computer interface for hand paralysis in patients with chronic stroke: a preliminary case series study J Rehabil Med 2011 43 951 957 10.2340/16501977-0859 21947184 18 Liu M Fujiwara T Shindo K Newer challenges to restore hemiparetic upper extremity after stroke: HANDS therapy and BMI neurorehabilitation Hong Kong Physiother J 2012 30 83 92 10.1016/j.hkpj.2012.05.001 19 Muralidharan A Chae J Taylor DM Extracting attempted hand movements from EEGs in people with complete hand paralysis following stroke Front Neurosci 2011 5 7 20 Schalk G Mellinger J Practical guide to brain-computer interfacing with BCI2000: general-purpose software for brain-computer interface research, data acquisition, stimulus presentation, and brain monitoring 2010 1 260 21 Orihuela-Espina F Del Castillo I Palafox L Neural reorganization accompanying upper limb motor rehabilitation from stroke with virtual reality-based gesture therapy Top Stroke Rehabil 2013 20 197 209 23841967 22 Prasad G Herman P Coyle D Applying a brain-computer interface to support motor imagery practice in people with stroke for upper limb recovery: a feasibility study J Neuroeng Rehabil 2010 7 17 20423488 23 Carter AR Connor LT Dromerick AW Rehabilitation after stroke: current state of the science Curr Neurol Neurosci Rep 2010 10 158 166 20425030 24 Do AH Wang PT King CE Brain-computer interface controlled functional electrical stimulation system for ankle movement J Neuroeng Rehabil 2011 8 14 21429200 25 Shelton FN Reding MJ Effect of lesion location on upper limb motor recovery after stroke Stroke 2001 32 107 112 11136923 26 Cho W Vidaurre C Hoffmann U Afferent and efferent activity control in the design of brain computer interfaces for motor rehabilitation 33rd Annual International Conference of the IEEE Engineering-in-Medicine-and-Biology-Society (EMBS) Boston, MA IEEE 2011 7310 7315 27 Wu C-Y Huang P-C Chen Y-T Effects of mirror therapy on motor and sensory recovery in chronic stroke: a randomized controlled trial Arch Phys Med Rehabil 2013 94 1023 1030 23419791 28 Nair V 2013 Conference: radiological society of North America 2013 scientific assembly and annual meeting 29 Wolf SL Winstein CJ Miller JP The EXCITE trial: retention of improved upper extremity function among stroke survivors receiving CI movement therapy Lancet Neurol 2008 7 1 33 40 10.1016/S1474-4422(07)70294-6 18077218 30 Gomez-Rodriguez M Grosse-Wentrup M Hill J Towards brain-robot interfaces in stroke rehabilitation IEEE International Conference on Rehabilitation Robotics (ICORR)/International Neurorehabilitation Symposium (INRS)/International Conference on Virtual Rehabilitation (ICVR) Zurich IEEE 2011 31 Young B Nigogosyan Z Walton L Dose-response relation-795 ships using brain-computer interface technology impact stroke rehabilitation Front Hum Neurosci 2015 9 361 26157378 32 Felton EA Williams JC Vanderheiden GC Mental workload during brain-computer interface training Ergonomics 2012 55 526 537 22506483 33 Stoeckel L Garrison K Ghosh S Optimizing real time fMRI 800 neurofeedback for therapeutic discovery and development Neuroimage Clin 2014 5 245 255 25161891 34 Song J Young B Nigogosyan Z Characterizing relationships of DTI, fMRI, and motor recovery in stroke rehabilitation utilizing brain-computer interface technology Front Neuroeng 2014 7 31 25120466 35 Miralles F Vargiu E Dauwalder S Brain computer interface on track to home Sci World J 2015 2015 1 17 10.1155/2015/623896 36 Sellers EW Vaughan TM Wolpaw JR A brain-computer interface for long-term independent home use Amyotroph Lateral Scler 2010 11 5 449 455 20583947 37 Young B Nigogosyan Z Nair V Case report: post-stroke interventional BCI rehabilitation in an individual with preexisting sensorineural disability Front Neuroeng 2014 7 18 25009491 38 Daly J Cheng R Rogers J Feasibility of a new application of noninvasive brain computer interface (BCI): a case study of training for recovery of volitional motor control after stroke J Neurol Phys Ther 2009 33 203 211 20208465 39 Takahashi M Takeda K Otaka Y Event related desynchronization-modulated functional electrical stimulation system for stroke rehabilitation: a feasibility study J Neuroeng Rehabil 2012 9 6 22304989 40 Biasiucci A Leeb R Al-Khodairy A 2013 Motor recovery after stroke by means of BCI-guided functional electrical stimulationat Proceedings of the Fifth International Brain-Computer Interface Meeting 2013 10.3217/978-3-85125-260-6-49 41 Dromerick AW Lang CE Birkenmeier RL Very early constraintinduced movement during stroke rehabilitation (VECTORS) A single-center RCT Neurology 2009 73 195 201 19458319 42 Lang KC Thompson PA Wolf SL The EXCITE trial: reacquiring upper-extremity task performance with early versus late delivery of constraint therapy Neurorehabil Neural Repair 2013 27 654 663 23542218 43 Ang K Guan C Chua K A clinical study of motor imagerybased brain-computer interface for upper limb robotic rehabilitation Annual International Conference of the IEEE-Engineering-in-Medicine-and-Biology-Society Minneapolis (MN) IEEE 2009 AQ10 5981 5984 44 Ang K Guan C Chua K Clinical study of neurorehabilitation in stroke using EEG-based motor imagery brain-computer interface with robotic feedback 32nd Annual International Conference of the IEEE Engineering-in-Medicine-and-Biology-Society (EMBC 10) Buenos Aires IEEE 2010 5549 5552.48 45 Lo A Guarino P Richards L Robot-assisted therapy for longterm upper-limb impairment after stroke N Engl J Med 2010 362 1772 1783 20400552 46 Wilson JA Walton LM Tyler M Lingual electrotactile stimulation as an alternative sensory feedback pathway for brain-computer interface applications J Neural Eng 2012 9 045007 22832032 47 Buch E Weber C Cohen L Think to move: a neuromagnetic brain-computer interface (BCI) system for chronic stroke Stroke 2008 39 910 917 18258825 48 Broetz D Braun C Weber C Combination of brain-computer interface training and goal-directed physical therapy in chronic stroke: a case report Neurorehabil Neural Repair 2010 24 674 679 20519741 49 Teo W-P Chew E Is motor-imagery brain-computer interface feasible in stroke rehabilitation? A narrative review Pm R 2014 6 8 723 728 24429072 50 Ang KK Guan C Phua KS Brain-computer interface-based robotic end effector system for wrist and hand rehabilitation: results of a three-armed randomized controlled trial for chronic stroke Front Neuroeng 2014 7 30 25120465 51 Mihara M Hattori N Hatakenaka M Near-infrared spectroscopy–mediated neurofeedback enhances efficacy of motor imagery–based training in poststroke victims: a pilot study Stroke 2013 44 4 1091 1098 23404723 52 Naseer N Hong K-S Classification of functional near-infrared spectroscopy signals corresponding to the right-and left-wrist motor imagery for development of a brain-computer interface Neurosci Lett 2013 553 84 89 10.1016/j.neulet.2013.08.021 23973334 53 Kaiser V Daly I Pichiorri F Relationship between electrical brain responses to motor imagery and motor impairment in stroke Stroke 2012 43 2735 2740 22895995 54 Tan H Kong K Shee C Post-acute stroke patients use braincomputer interface to activate electrical stimulation 32nd Annual International Conference of the IEEE Engineering-in-Medicine-and-Biology-Society (EMBC 10) Buenos Aires IEEE 2010 4234 4237 875AQ12 55 Nicolas-Alonso LF Gomez-Gil J Brain computer interfaces, a review Sensors 2012 12 2 1211 1279 22438708 56 Ramos-Murguialday A Broetz D Rea M Brain–machine interface in chronic stroke rehabilitation: a controlled study Ann Neurol 2013 74 1 100 108 23494615 57 Ono T Shindo K Kawashima K Brain-computer interface with somatosensory feedback improves functional recovery from severe hemiplegia due to chronic stroke Front Neuroeng 2014 7 24782757 58 Ang KK Guan C Phua KS Facilitating effects of transcranial direct current stimulation on motor imagery brain-computer interface with robotic feedback for stroke rehabilitation Arch Phys Med Rehabil 2015 96 3, Supplement S79 S87 25721551 59 Ahn M Jun SC Performance variation in motor imagery brain–computer interface: a brief review J Neurosci Methods 2015 243 0 103 110 25668430 60 Cramer SC Brain repair after stroke N Engl J Med 2010 362 1827 1829 10.1056/NEJMe1003399 20400553 61 Gomez-Rodriguez M Peters J Hill J Closing the sensorimotor loop: haptic feedback facilitates decoding of motor imagery J Neural Eng 2011 8 12 62 Ang K Guan C Chua K A large clinical study on the ability of stroke patients to use an EEG-based motor imagery brain-computer interface Clin EEG Neurosci 2011 42 253 258 22208123 63 Ang KK Guan C Brain-computer interface in stroke rehabilitation J Comput Sci Eng 2013 7 2 139 146 64 Winstein CJ Wolf SL Dromerick AW Effect of a task-oriented rehabilitation program on upper extremity recovery following motor stroke: the ICARE randomized clinical trial JAMA 2016 315 6 571 581 10.1001/jama.2016.0276 26864411 65 Rayegani S Raeissadat S Sedighipour L Effect of neurofeed-back and electromyographic-biofeedback therapy on improving hand function in stroke patients Top Stroke Rehabil 2014 21 2 137 151 10.1310/tsr2102-137 24710974 66 Song J Nair V Young B DTI measures track and predict motor function outcomes in stroke rehabilitation utilizing BCI technology Front Hum Neurosci 2015 9 195 25964753 67 Kim S Park J Jung M Effects of task-oriented training as an added treatment to electromyogram-triggered neuromuscular stimulation on upper extremity function in chronic stroke patients Occup Ther Int 2016 10.1002/oti.1421 68 Shin J Kim M Lee J Effects of virtual reality-based rehabilitation on distal upper extremity function and health-related quality of life: a single-blinded, randomized controlled trial J Neuroeng Rehabil 2016 13 1 10.1186/s12984-016-0125-x 69 Lakshminarayanan K Wang F Webster JG Feasibility and usability of a wearable orthotic for stroke survivors with hand impairment Disabil Rehabil Assist Technol 2016 1 9 10.3109/17483107.2015.1111945

PMC005xxxxxx/PMC5131718.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 9423872 2548 Free Radic Res Free Radic. Res. Free radical research 1071-5762 1029-2470 23937589 5131718 10.3109/10715762.2013.833331 NIHMS828681 Article Redox Signaling Mediated by the Gut Microbiota Jones Rheinallt M. 1 Neish Andrew S. 2 1 Department of Pediatrics, Emory University School of Medicine, Whitehead Biomedical Research Building, 615 Michaels St, Room 105-L, Atlanta GA, 30322 2 Department of Pathology and Laboratory Medicine, Emory University School of Medicine, Whitehead Biomedical Research Building, 615 Michaels St, Room 105-L, Atlanta GA, 30322 (p) 404-727-8545, (f) 404-727-8538, aneish@emory.edu 24 11 2016 4 10 2013 11 2013 01 12 2016 47 11 950957 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. The microbiota that inhabits the mammalian intestine can influence a range of physiological functions, including the modulation of immune responses, enhancement epithelial barrier function, and the stimulation of cell proliferation. While the mechanisms by which commensal prokaryotes stimulate immune signaling networks are well-characterized, less is known about the mechanistic control over homeostatic pathways within tissues. Recent reports by our research group have demonstrated that contact between the gut epithelia and some groups of enteric commensal bacteria prompts the rapid generation of reactive oxygen species (ROS) within host cells. Whereas the bacterial-induced production of ROS in phagocytes in response to ligand binding to Formyl Peptide Receptors (FPRs) and ensuing activation of NADPH oxidase 2 (Nox2) is a well-defined mechanism, ROS generated by other cell types such as intestinal epithelia in response to microbial signals via FPRs and the NADPH oxidase 1 (Nox1) is less appreciated. Importantly, enzymatically generated ROS have been shown to function as second messengers in many signal transduction pathways via the transient oxidative activity on sensor proteins bearing oxidant-sensitive thiol groups. Examples of redox sensitive proteins include tyrosine phosphatases that serve as regulators of MAPK pathways, focal adhesion kinase, as well as components involved NF-kB activation. Here, we review the leading edge discoveries gleaned from investigations that focus on microbial-induced generation of ROS and their functional effects on host physiology. These studies identify the functional molecular elements and mechanistic events that mediate the established effects of the normal microbiota on intestinal physiology. Graphical Abstract Probiotics Nox enzymes Reactive oxygen species Formyl peptide receptors The intestinal physiology and eukaryotic-prokaryotic interactions Symbiotic host-microbe communications has evolved in virtually every metazoan, with the human gut microbial population an example of increasingly documented medical significance [1]. In utero, the mammalian gut is sterile. The microbial colonization progression begins during birth, culminating in a diverse and stable community, though typically, microbial composition between individuals varies [2]. Recent advancements in high-throughput sequencing and bioinformatic methodologies have profoundly enhanced our knowledge of the diversity of the microbial population within the gut, revealing that the bulk of the microbial population are represented by Bacteroidetes and Firmicutes [3–6]. Particular taxa of the gut microbial population may be free-living in the luminal fecal stream thereby occupying a planktonic niche, or may be adherent to the gut mucous layer or to epithelial cells of the mucosa. Microbes in the intestine thrive in the nutrient rich environment, with certain taxa contributing favorable influences to the host that include, but not limited to the production of short chain fatty acids and enhanced energy extraction from foodstuffs, competitive exclusion of pathogenic microorganisms, and priming of innate and adaptive immune system responses, and influences on bone homeostasis [6, 7]. In addition, investigations utilizing germ-free mice have established a function for the gut microbiome and metabolic regulation [8]. The gut microbiome has also been shown to positively influence homeostasis of the intestinal mucosa by enhancing barrier function, as well as epithelial cell proliferation and survival [9–15]. For instance, villi of the small intestine of the germ-free mice have impaired angiogenesis [16] and have slower turnover rates of epithelial cells [17]. Mono-association of germ-free mice with a gut symbiont (Bacteriodes thetaiotaomicron) evoked a vigorous host transcriptional response, showing that the host can actively sense, perceive, and respond to the presence of commensal of symbiotic within the lumen [18]. These data demonstrate the existence of an active dynamic association between host cell and microbes inhabiting the gut. Nevertheless, it is also illustrates that anomalies in the quality and diversity of the gut microbiome (“dysbiosis”) may be sufficient by itself to aggravate intestinal inflammation as seen in Inflammatory bowel diseases (IBD). In addition, changes in the diversity of the gut microbiome have been linked with infectious disease such as pseudomembranous colitis, in systemic immune disorders such as multiple sclerosis, in allergic diseases such as celiac disease and asthma, and in the onset of diabetes and obesity in metabolic syndromes in adults [19–21]. Over the past few years, investigations have focused on exploiting the positive influences of certain taxa of the gut microbiome by supplementing the indigenous microbiome with purified cultures of symbiotic bacteria. This tactic, called ‘probiotics’, has been successful in suppressing inflammation, strengthening barrier function, facilitating tissue repair in response to injury in the intestine, thus offering a therapeutic approach to ameliorate disorders of the intestinal tract [22]. Altogether, mounting evidence have empirically demonstrated that the gut microbiome favorably influences intestinal physiology. However, there is a gap in our knowledge pertaining to an understanding of how the intestinal cells sense symbiotic bacteria, and mechanistically modulate gut physiology. This manuscript will review recent discoveries that have identified a redox based response within cells, that is emerging as a critically conserved element of host cell and symbiotic microbe interactions. Epithelial perception and monitoring of the microbiota The intestinal mucosa encompasses the outward facing epithelial cells, the structural components of the underlying lamina propria, and the immune cells residing in sub-epithelial compartments, which together form a functional barrier preventing the luminal contents from entering systemic compartments. The gut luminal contents are physically separated from the host interior by a thin layer of mucus that overlays a monolayer of columnar epithelial cells. Within epithelial cells are apical surface factors that function in the uptake of nutrients while at the same time, existing in intimate contact with the luminal contents. This active process exists against a background of the intestinal epithelium continually renewing itself in a progression involving asymmetrical proliferation of stem cells, ensuing differentiation, migration and eventual programed apoptosis and shedding at the villous apex - a homeostatic cycle that occurs over 5–7 days in humans. The intestine may become damaged following exposure to a pathogenic and/or immunologic insults. While overcoming and resolving damage to tissues, the gut mucosa must also maintain the beneficial relationship with the symbiotic taxa of the gut microbiota [23]. To succeed in this management, cells of the intestinal mucosa have dedicated sentinel elements for monitoring bacteria. Examples of such sentinel elements are Toll-like receptors (TLRs) and related Nod-like receptors (NLRs), collectively termed “pattern recognition receptors” (PRRs), that bind to motifs known as “microbe associated molecular patterns” (MAMPs) that are ubiquitously present across the bacterial phylogenetic domain [24]. Sensing of MAMPs by TLRs activate innate immune signaling cascades such as the MAPK and NF-kB pathways [25–28], which are by and large considered pro-inflammatory, although at lower ‘tonic’ levels of activation, have been connected to mechanisms of normal gut homeostasis [29, 30]. A distinct group of PRRs are the formylated peptide receptors (FPR), which are G-protein-linked seven membrane pass receptors originally discovered on the surface of professional phagocytes. There are three structurally-related FPRs in humans, known as FPR1, FPR2 and FPR3 [31]. Functionally, FPRs sense and bind to peptides that harbor a bacterial specific N-formyl group of which an example is N-formyl methionyl-leucyl-phenylalanine (fMLF). FPR1 binds to fMLF with particularly high-affinity with an ED50 in the nanomolar range, whereas FPR2 binds with lower affinity in the micromolar range. Furthermore, agonists generated by eukaryotic cells such as AnxA1, LXA4 and SAA, or mitochondrially derived translation products can also stimulate FPRs [32]. The importance of these receptors is highlighted by the observation that FPR1-deficient mice have augmented susceptibility to pathogens, suggesting that FPR1 is functions in supporting processes of acute inflammation [33]. In addition, mitochondrial derived formyl peptides are recognized as being the source of the agonists in non-infectious tissue injury (“sterile inflammation”). Here, mitochondria, which are prokaryotic endosymbionts, and as such, retain the prokaryotic translation machinery, including the bacterial-specific use of N-formyl-methionine capping of nascent transcripts. This means that formylated peptides are present inside cells within mitochondria. In a pathological context such as tissue necrosis (not apoptosis), cellular contents are released extracellularly and can be perceived by other cells and phagocytes as “danger associated molecular patterns” or DAMPS [34]. Once the agonist binds, FPRs are phosphorylated and glycosylated, which initiates interaction with pertussis toxin-sensitive Gi proteins [35–38]. Ensuring cell signaling cascades involves MAPK and phosphatidylinositol 3-kinase (PI3K) pathway activation, which together with small GTPases initiate phagocytic functions such as stimulation of actin dynamics and chemotaxis, and the activation of ROS generation by NADPH oxidase enzyme (Nox2) in a process known as the respiratory burst [39–41]. Indeed, it was the known function of agonist binding to FPRs eventuating in ROS generation in phagocytes that proved to be the rationale and the springboard for assessing the function of generated ROS in epithelial cells following in host-symbiotic bacterial interactions. This notion was substantiated with the discovery of functional FPRs located on the apical surface of the intestinal epithelial cells, suggesting that FPRs function in an similar physiological process in the intestinal mucosa [42]. Indeed, in epithelial cells, it was shown that MAPK pathway signaling is activated by fMLF by an FPR-dependent mechanism, and that fMLF binding to FPR also induced the generation of ROS in the epithelial cells [43]. The generation of physiological levels of reactive oxygen species (ROS) Reactive oxygen species (ROS) generated at high levels by professional phagocytes mediate the capacity of these cells to kill bacteria. Agonist binding to FPRs in phagocytes initiate the respiratory burst, which involves extensive enzymatic production of superoxide within the vacuole harboring phagocytosed microbe. ROS generation in phagocytes is catalyzed by a multi component and membrane-bound NADPH oxidase enzyme called Nox2 (formerly designated gp120phox). Nox2 is basally a dimer of gp91phox and gp22phox [44], and considering the potential harmfulness of elevated superoxide levels to surrounding tissue, explicably, ROS generation is firmly controlled by the consecutive recruitment of separate subunits of the Nox2 enzyme. The function of Nox2 in vivo is demonstrated by the discovery that Nox2 null mice develop chronic granulomatous disease (CGD), which is a disorder in patients where phagocytes are unable to produce ROS and are thus highly susceptible to repeated pyogenic infections. Nox2/gp120phox was first protein of the NADPH oxidase group or “Nox’es” to be discovered. Importantly, Nox enzymes are also functional in non-phagocytic cell types and tissues, with the discovery that Nox1 and Duox2 are strongly expressed in enterocytes, which are in close proximity to the gut luminal bacterial content being relevant to this review [45–48]. Generally, the Nox enzymes functional within non-phagocytic cells have analogous, but not identical subunit regulation and assembly to Nox2 in phagocytes. For example, Nox1 requires Rac-GTPase-initiated cascades for its enzymatic activity, whereas Duox2 activity is calcium dependent. Nox1 catalyzed generation of ROS within epithelial cells is hypothesized to function in regulating several signal transduction pathways, which was detected following the stimulation of Nox1 function by growth factors, hormones, and cytokines in a wide range of cells and tissues [45, 47]. Indeed, Nox enzymes orthologs catalyze ROS production in a breadth of multi cellular organisms [49–52]. In plants, ROS generated Nox enzymes regulates the transition from proliferation to differentiation in root tips [53]. In flies, Duox-generated ROS in gut epithelia functions in the control of the diversity of the intestinal microbiome [54–56], which implies a role for ROS generation in epithelial cells in host cell and microbe interactions. However, by and large, the functions of Duox in the epithelia are less studied, but have been implicated more in anti-microbial functions that in signaling events. Pertaining to the catalytic activity of Nox1 and Duox2, both enzymes are involved in trans-membrane generation of superoxide generation which is generally thought to be rapidly dismutated to H2O2. Thereafter, H2O2 may be transported back across the plasma membrane, likely via aquaporins, for cytoplasmic signaling functions [57–59]. Redox signaling and the oxidation of reactive cysteine residues Non-radical ROS, such as H2O2, that are generated by Nox enzymes in non-phagocytic tissues, are well documented as functioning in a variety of signaling pathways within a variety of cell types [60]. The diverse biological outcomes induced by ROS depends on the specific species generated, the as well as the duration, and the subcellular locations of generation [47, 52]. Because they are extremely short-lived, ROS have slight functional radii, which also means they can precisely discriminate their influence. Indeed, some receptors physically interact with Nox enzymes, apparently to bound the influences of the generated ROS to the vicinity of target proteins [61]. The mechanism through which ROS modulate cell signaling is through their capacity to oxidize certain reactive cysteine residues within enzymes controlling the activation of cell signaling pathways [62–64]. Proteins that harbor reactive cysteine residues can therefore function redox sensors and transducers of signaling initiated by elevated ROS. Moreover, reversible oxidation of cysteine residues allows for graded response to intracellular ROS concentrations, meaning that the cell can sense and respond to fluctuating levels of ROS. At the molecular level, the vast majority of cysteine residues are protonated at physiological pH (Cys-SH) (pKa ~8.5) and are unreactive at physiological pH. It is only a few cysteine residues that have ROS-sensitive properties, which due to them being charged at a low pKa, and thus extant as thiolate anions (Cys-S-) at physiological pH. Low pKa cysteine residues are highly reactive and are easily oxidized by ROS such as H2O2 [64]. Mass spectrometry analysis has revealed that redox-sensitive thiolates are present in a limited subset of enzymes. Examples include protein tyrosine phosphatases (PTP) [65], the lipid phosphatase (PTEN) [62, 66], MAP kinase phosphatases (MAPKP) such as DUSP3 [63], and low-molecular-weight protein tyrosine phosphatases (LMW-PTP) [67], in enzymes involved in the sumoylation and neddylation reactions, and well as in oxidant sensors such as Keap1, which control of overall redox balance of the cell. (Figure 1). Symbiotic bacterial-induced ROS generation in intestinal epithelial cells Symbiotic bacterial-induced ROS generated through catalytic activities of Nox enzymes have been detected in numerous forms of multicellular organisms, extending from social amoebae, to florae, and to humans. Thus it is thought that ROS signaling represents an evolutionary ancient form of host cell and microbe cross-talk [49, 54, 68, 69]. Of late, our research group and others reported that contact of certain symbiotic taxa with host epithelial cells can induce the generation of rapid, non-pathogenic levels of ROS, in a process that requires the catalytic activity of Nox1 in host cells. Furthermore, contact of intestinal epithelial cells with bacteria of the lactobacilli taxon induced increased oxidation of proteins that function as soluble redox sinks, including thioredoxin and glutathione, and result in the induction of the transcription of redox-stimulated modulons such as the Nrf2 pathway activity. These observations point to an active and dynamic response to increased levels of cellular ROS. Furthermore, specific bacterial contact with epithelial cells have regulatory effects on cytoskeletal dynamics and on host immune activity [14, 43, 70]. Of these taxa, diverse commensal bacteria elicit distinctly varying capacities of inducing cellular ROS generation, with lactobacilli particularly powerful inducers of ROS generation, although the majority of bacteria assayed exhibited some degree of ability to induce ROS generation in host epithelial cells. Lactobacilli have been shown to harbor membrane components or secreted factors that trigger cellular responses. For example, it has been demonstrated that soluble factors produced by Lactobacillus rhamnosus GG mediate beneficial effects in in vivo injury models [71]. In addition, lactobacilli that stimulate ROS may also have enhanced ability to penetrate the mucus layer, or have enhanced adhesive properties to come into contact with cellular receptors such as FPRs or TLRs. While levels of ROS produced by the action of non-phagocyte NADPH oxidases are indeed orders of magnitude less than the output of Nox2 in phagocytes, ROS production is not so vanishingly small that it is beyond the sensitivity of current biochemical assays. For example, ROS produced can be visualized by a variety of redox sensitive dyes in immune fluorescence. In our hands, the most faithful reporters of ROS generation in tissues are the ROSstarTM hydrocyanine probes [72]. These probes are specific for oxygen radicals, superoxide, and hydroxyl radicals and detect intracellular ROS with high sensitivity and specificity. The probes are cell-permeable and initially non-fluorescent. The probes become fluorescent upon oxidation. The probes are used to detect oxidative stress by fluorescence microscopy, flow cytometry, and can be quantified by microplate fluorometry. In addition, images captured by fluorescence microscope may be were quantified using image densitometry software. Physiological outcomes of symbiotic bacteria-induced redox signaling Effects on inflammatory signaling: Extensive reports have described the suppressive activity of lactobacilli on host inflammatory signaling pathways, having the net effect of dampening the ensuing innate immunity [71, 73–75]. For example, intestinal microbes can modulate gut inflammation, and probably additional cellular processes by modulating the ubiquitin-proteosome pathway [76]. Contact of lactobacilli with epithelial cells has the effect of inhibiting IκB ubiquitination by interfering with the IκB ubiquitin ligase, SCFβTrCP (Skp1, Cdc53/Cullin, F box receptor), thereby blocking activation of the pro-inflammatory NF-κB pathway [77–79]. Specifically, for activation of the SCFβTrCPcomplex, a covalent modification involving neddylation of the cullin-1 (Cul-1) regulatory subunit must occur. Oxidative signaling negatively regulates neddylation by the reversible inactivation of Ubc12, which is a redox sensitive Nedd8 ligase [70]. Contact of epithelial cells with lactobacilli induces rapid loss of the Nedd8 modification, thereby inhibiting SCF ubiquitin ligase function, and the suppression of NF-κB pathway activity [78]. Together, ROS generation in the gut epithelium modulates SCF ligase activity, and the NF-κB pathway by controlling the balance between neddylated and un-neddylated Cul-1. Interestingly, other pathways are known to be regulated by E3-SCFβTrCP, such as β-catenin, Snail, Twist and Hedgehog [80], pointing to further molecular influences that lactobacilli-induced generation of ROS could regulate various facets of intestinal physiology. Effects on epithelial cell motility. As stated previously, germ-free mice exhibit impaired gut physiology including proliferation and wound healing rates of the epithelial layer, showing that the enteric luminal contents have potent influences on host cells. Furthermore, germ-free mice have impaired rates of epithelial cell migration, which is a process that is regulated by the exquisitely coordinated restructuring of the actin cytoskeleton at the advancing edge of the cell to specialized signaling nidus points called focal adhesions (FA) of the extracellular matrix. The dynamics of the FA assembly is controlled by an enzyme called focal adhesion kinase (FAK), which is a 125 kDa protein that is held in its inactive state under the dephosphorylating influences of the redox-sensitive tyrosine phosphatases LMW-PTPase and SHP-2. However, interactions between growth factors and integrins at the basement membrane triggers Nox1 catalyzed ROS production which oxidative inactivates PTPase, resulting in releasing FAK from its dephosphorylated state thereby initiating FA turnover and cell motility [81]. Concerning the influence of the gut microbe population on cell motility, we have shown that contact between intestinal epithelia with symbiotic bacteria such as some lactobacilli strains induces generation of ROS in these cells, particularly at the edges of in vivo colonic biopsy wounds, and at the leading edge of the migrating cells. Mechanistically, the ROS generated in response to cell contact with lactobacilli reversibly oxidizes low pKa cysteines within LMW-PTP and SHP-2, thereby activating FAK activity and FA dynamic assembly and disassembly at the migrating edge of the monolayer. These molecular mechanisms induce augmented epithelial migration rates as demonstrated in improved wound sealing of in vitro and in vivo injury models [66]. Additionally, we showed that fMLF triggers FPR-dependent ROS generation in epithelial cells. Importantly, in either FPR or Nox1 null mice, lactobacilli-induced ROS generation, ERK and FAK phosphorylation, and improved wound healing are abolished. Consequently, these are molecular evidence for events that following ROS production associated with FPR/Nox1 dependent ROS generation following lactobacilli-epithelial contact, which eventuate in faster epithelial motility. It was also demonstrated that ROS generated by Nox are necessary for the function of invadopodia, which are actin-rich membrane protrusions in cells. The study proposed that the invadopodia protein Tks5 is a part of the Nox complex, and showed that depletion of Tks5 levels reduces total ROS amounts in cells [82]. These data demonstrate a further mechanism by which the microbiota positively influences physiological processes within the gut mucosa and may also point to possible mechanisms by which probiotics exert their beneficial influence on epithelial barrier integrity. Effects on epithelial growth and differentiation: The intestinal gut epithelium is the most dynamically renewing tissue in the adult body. The epithelium is a three dimensional architecture of invaginated crypts and projecting villi. At the base of the crypts reside multipotent stem cell niches that are the source of epithelial renewal and homeostasis. Stem cell undergo asymmetric division where the cell targeted for a differentiated fate migrates to a transient amplifying compartment where further signaling events eventuate in the specification of the cells as they absorptive enterocytes, mucus secreting goblet cells, or neuroendocrine epithelial cells. Germ-free mice have been shown to have slower rates of cell migration up the crypt-villus axis, while studies in germ-free Drosophila corroborate distinctly curbed proliferation rates of epithelial precursor cells, pointing to a role for the luminal microbiota in controlling the intestinal epithelial development and homeostasis [83, 84]. Importantly, ROS act as mediators of cell proliferation and differentiation in wide variety of unrelated systems such as in the growing plant root hair [53], to Drosophila hematopoiesis [85]. Our research group demonstrated lactobacilli-induced ROS generation leads to the activation of the pro-proliferation and developmental ERK pathway by a mechanism involving the redox inactivation of the ERK phosphatase DUSP6 [14, 43, 66]. Furthermore, Nox1 generation of ROS has been reported to regulate Wnt and Notch1 pathway signaling in the colon [86]. Indeed, the function of Nox1 as pivotal factor in cell fate through Wnt/beta-catenin and Notch1 signaling pathways was shown in Nox1-deficient mice which had elevated numbers of goblet cells as a result of PI3K/AKT/Wnt/beta-catenin and Notch1 signaling repression [86]. By extension, lactobacilli-induced and Nox dependent redox signaling may also function in gut development by similar mechanisms. Cytoprotection by lactobacilli-activation of Keap1/Nrf2/ARE signaling Another example of host signaling circuitry that responds to ROS generation in the cytoplasm of cells is the Keap1/Nrf2/ARE signaling module. Nrf2 (NF-E2-Related Factor 2) and its antagonist Keap1 (Kelch-like ECH-Associated Protein 1) are essential factors that mediate cytoprotection in response to xenobiotics [87]. The Keap1/Nrf2/ARE module is evolutionarily conserved across kingdoms with examples in Caenorhabditis elegans [88], D. melanogaster [89], zebrafish [90], and in mice [91]. Nrf2 activity is regulated by the binding action of its agonist and inhibitor, Keap1 [92]. Under non-stimulated situations of low cellular ROS levels, Keap1 attaches to Nrf2 and directing Nrf2 to Cullin-dependent E3 ubiquitin ligase proteosomal degradation. However, electrophilic stress in the cytoplasm, typically in situations of elevated ROS levels, results in the oxidation of low pKa cysteine residues within Keap1, thereby causing a conformational change in Keap1 structure that results in its disassociation from Nrf2. Nrf2 is therefore free to translocate the nuclear membrane where it associates with a conserved DNA sequence known as antioxidant response elements (ARE) which are situated in the promoters of a battery of cytoprotective factor genes [93]. Investigation of the result of bacterial-induced ROS generation on Nrf2 pathway signaling discovered that lactobacilli-induced, and Nox1 catalyzed generation of ROS triggered the transcription of Nrf2-responsive cytoprotective elements, and resulted in organismal cytoprotection against oxidative stress in Drosophila, and against radiological insult in mice [94]. Thus, the Nrf2 signaling pathway is yet another mechanism whereby host cells perceive and respond to bacterial contact, this time to activate cytoprotection within cells (Figure 2). Because we established that lactobacilli induce Nrf2 signaling, this unlocks the prospect of recognizing mechanisms by which probiotic bacteria may elicit beneficial effects on disease states that involve Nrf2 pathway signaling. As stated previously, Nrf2 signaling has been widely investigated in relation to cytoprotection from xenobiotic stresses, whose presence in the cell induces the basal Nrf2 transcriptional regulon of several hundred genes [95]. Importantly, studies into Nrf2 function exposed a plethora of cellular processes other than cytoprotection which are regulated by its signaling. These include diabetes [96], cancer cell growth and chemoresistance [97–99], neurodegenerative diseases [100], redox homeostasis in the aging heart [101], as well as oxidative stress and inflammatory pathways [102]. Furthermore, ROS are generated as by products during inflammation in the gut epithelia, primarily due to respiratory burst activity from phagocytes described earlier in this review. At the site of injury, Nrf2-responsive elements defend stem cell populations and promote restitutive cell proliferation and migration [103]. Indeed, each of the listed examples of cellular processes that are influenced Nrf2, are by extension potentially modulated by lactobacilli (probiotic) stimulation of Nrf2 pathway that would augment cytoprotective and reparative responses. Future Perspectives Experimental evidence generated by our research group demonstrate that intestinal epithelial cells generation ROS in response to contact with symbiotic bacteria, by mechanisms involving receptors and enzymatic processes similar to those evolved in phagocytic cells to induce microbial death. Evidence from the Drosophila model point to the notion that ROS generation in the gut epithelia may represent an evolutionary conserved response to microbes [55]. In mice, ROS generated in epithelial cells in response to lactobacilli undoubtedly functions in a signaling cell signaling events through reversible redox inactivation of regulatory proteins [52, 65]. Importantly, leading edge proteomic methodology can be employed as a screening system to identify microbial-specific, oxidant-sensitive regulatory proteins [104]. Identification of the function of identified reactive cysteine residues within redox-sensitive proteins in vivo will be challenging future work. The strength and duration of lactobacilli-ROS-mediated signaling may vary with quantitative and qualitative variations in bacterial populations, for example, as would occur as a result of antibiotic administration, probiotic use, dietary changes, or the acquisition of the microbiota following birth. The population density of the gut microbiota differs by several orders of magnitude along the course of the digestive tract, with the peak density and numbers occurring in the cecum. The significance of this is that various regions of the intestine are likely to undergo dissimilar levels of bacterial-induced ROS. Furthermore, the fact that specific bacterial taxa such as lactobacilli exhibit potent ROS-inducing activity ties with the idea that qualitative changes in this taxa can negatively influence host biology. Conceivably, the relative amounts of a particular microbial taxa in the intestinal lumen might result in specific physiological outcomes at the organism level. A comprehension of the association amongst microbes and cellular ROS generation will contribute towards for describing the “eubiotic” and the “dysbiotic” population diversity that lead to positive health or inflammatory disease, and certainly has implications for the characterization of new types of probiotics. In addition, lasting cellular adaptation to the intimate bacterial company, as occurs in the colon, or short term contact, as occurs in the small intestine may also have differential outcomes on redox biology. Future studies should also consider the influence of tissue oxygen partial pressure (pO2) at the interface of the epithelium and lumen, and its potential modulating influence on ROS/Nox signaling. A non-invasive method of determining tissue pO2 by electron paramagnetic resonance (EPR) oximetry showed oxygen levels of less than 2% in the intestinal lumen, 8% in the small intestinal wall, and 3 % at the villus tip [105]. These conditions of physiological oxygenation were termed physioxia. However, we detected similar qualitative ROS generation along the crypt villus axis, including the apex, where tissue pO2 is lowest [13]. Very little is known about how pO2 in the gut influences Nox activity, although reports using in vitro cultured cells claim that mRNAs of nox enzymes are sensitive to oxygen tension [106]. Nevertheless, our research group did described mucosal tissue oxygen depletion in the context of active neutrophilic inflammation, presumably due to oxygen consumption by emigrated phagocytes [11]. It is possible that, in this context, oxygen depletion my limit ROS production. Another subject of intense investigation in our laboratory at the moment is to determine the duration and persistence of lactobacilli-induced ROS generation. Our current hypothesis holds that this type of generation is acute and highly localized to sub-cellular organelles, and to specific cell types. The influence of ROS in the cytoplasm is undoubtedly short-lived due the presence of redox sinks such as glutathione-s-transferase (GSH) in the cytoplasm, and the activity of redox sensitive transcriptional pathways such as Nrf2, which serve to upregulate antioxidant gene products. In addition, Nox enzyme activity is presumably deactivated by currently unknown mechanisms. In all, we hypothesize that ROS signaling appears to be induced within minutes, with consequent modulation of protein activity and transcriptional responses in several hours, followed by a re-establishment of basal conditions. This pulsatile nature of ROS kinetics are also consistent with feeding cycles. Another variable to consider is the region of the gastrointestinal tract that ROS signaling is strongest. Areas of the gut, such as the upper small intestine, have vastly less microbial population than the colon. We have also demonstrated the importance of the mucus layer for limiting ROS activation, and that the loss of mucus seen in injury is associated with markedly increased ROS production [11]. Thus, future investigations must focus and consider the effects of local micro- and macro- environmental, temporal, bacterial taxa dependent, and disease related influences on microbial mediated ROS generation. In conclusion, microbial contact-induced epithelial ROS generation is an extremely conserved phenomenon across phyla with several known, and expected physiological consequences. This mechanism is an universal and non-discriminating means by which bacterial communities can effect a variety of signaling and homeostatic processes in the host [79]. It is anticipated that a comprehensive understanding of this association will improve our knowledge pertaining to the function of the gut microbiota, and its dysregulation on gut physiology. Figure 1 Host signaling events controlled by symbiotic bacterial-induced cellular ROS generation. Symbiotic bacteria residing in the gut lumen stimulate intestinal tissue homeostatic events via the reversible activation of cellular redox signaling processes. The gut microbiota generate formylated peptides that are sensed by formyl peptide receptors (FPRs) situated on the apical surface of epithelial cells. Receptor binding initiates a singling cascade that trigger the NADPH oxidase Nox1 to catalyze localized ROS generation, which then oxidizes critical cysteine residues and the regulatory influence of redox sensor proteins including the Nedd8 ligase, Ubc12, DUSP3, and LMW-PTPase. Ensuing signaling processes influence gut physiology by including stem cell proliferation, epithelial cell motility, and dampening of innate immune responses. Figure 2 Activation of the redox-sensitive Nrf2 cytoprotective pathway by lactobacilli-induced generation of cellular ROS. Under homeostatic conditions, Keap1 binds to Nrf2 and inhibits its nuclear translocation. Generation of lactobacilli-induced ROS catalyzed by Nox1 oxidizes cysteine residues within Keap1, resulting in conformation change and release of binding from Nrf2, allowing it to translocate into the nucleus. Nrf2 then binds to an anti-oxidant response DNA promoter element and induces the transcription of a battery of Nrf2-responsive genes including a plethora of cytoprotective factors. The Nrf2 responsive gene products protect macromolecules from oxidative damage thereby promoting cell survival and preserving tissue physiological integrity. Highlights Bacterial contact with host epithelial cells can result in the enzymatic generation of reactive oxygen species and consequent redox signaling. ROS is produced by mechanisms similar to the oxidant burst in phagocytes, involving formyl peptide receptors and NAPDH oxidases. Distinct bacterial taxa have very different abilities to stimulate redox signaling Redox signaling is involved in immune regulation, cytoprotection and control of cellular motility and proliferation. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. 1 Neish AS Microbes in gastrointestinal health and disease Gastroenterology 2009 136 1 65 80 19026645 2 Eckburg PB Bik EM Bernstein CN Purdom E Dethlefsen L Sargent M Gill SR Nelson KE Relman DA Diversity of the human intestinal microbial flora Science 2005 308 5728 1635 8 15831718 3 Xu J Mahowald MA Ley RE Lozupone CA Hamady M Martens EC Henrissat B Coutinho PM Minx P Latreille P Cordum H Van Brunt A Kim K Fulton RS Fulton LA Clifton SW Wilson RK Knight RD Gordon JI Evolution of Symbiotic Bacteria in the Distal Human Intestine PLoS Biol 2007 5 7 e156 17579514 4 Dethlefsen L McFall-Ngai M Relman DA An ecological and evolutionary perspective on human-microbe mutualism and disease Nature 2007 449 7164 811 818 17943117 5 Gill SR Pop M Deboy RT Eckburg PB Turnbaugh PJ Samuel BS Gordon JI Relman DA Fraser-Liggett CM Nelson KE Metagenomic analysis of the human distal gut microbiome Science 2006 312 5778 1355 9 16741115 6 Backhed F Ley RE Sonnenburg JL Peterson DA Gordon JI Host-Bacterial Mutualism in the Human Intestine Science 2005 307 5717 1915 1920 15790844 7 Li JY Chassaing B Tyagi AM Vaccaro C Luo T Adams J Darby TM Weitzmann MN Mulle JG Gewirtz AT Jones RM Pacifici R Sex steroid deficiency-associated bone loss is microbiota dependent and prevented by probiotics J Clin Invest 2016 126 6 2049 63 27111232 8 Smith K McCoy KD Macpherson AJ Use of axenic animals in studying the adaptation of mammals to their commensal intestinal microbiota Seminars in Immunology 2007 19 2 59 69 17118672 9 Ismail AS Hooper LV Epithelial cells and their neighbors. IV. Bacterial contributions to intestinal epithelial barrier integrity Am J Physiol Gastrointest Liver Physiol 2005 289 5 G779 84 16227525 10 Madsen K Cornish A Soper P McKaigney C Jijon H Yachimec C Doyle J Jewell L De Simone C Probiotic bacteria enhance murine and human intestinal epithelial barrier function Gastroenterology 2001 121 3 580 91 11522742 11 Alam A Leoni G Quiros M Wu H Desai C Nishio H Jones RM Nusrat A Neish AS The microenvironment of injured murine gut elicits a local pro-restitutive microbiota Nat Microbiol 2016 1 15021 27571978 12 Alam A Leoni G Wentworth CC Kwal JM Wu H Ardita CS Swanson PA Lambeth JD Jones RM Nusrat A Neish AS Redox signaling regulates commensal-mediated mucosal homeostasis and restitution and requires formyl peptide receptor 1 Mucosal Immunol 2014 7 3 645 55 24192910 13 Jones RM Luo L Ardita CS Richardson AN Kwon YM Mercante JW Alam A Gates CL Wu H Swanson PA Lambeth JD Denning PW Neish AS Symbiotic lactobacilli stimulate gut epithelial proliferation via Nox-mediated generation of reactive oxygen species EMBO J 2013 32 23 3017 28 24141879 14 Wentworth CC Alam A Jones RM Nusrat A Neish AS Enteric commensal bacteria induce extracellular signal-regulated kinase pathway signaling via formyl peptide receptor-dependent redox modulation of dual specific phosphatase 3 J Biol Chem 2011 286 44 38448 55 21921027 15 Jones RM The Influence of the Gut Microbiota on Host Physiology: In Pursuit of Mechanisms Yale J Biol Med 2016 89 3 285 297 27698613 16 Stappenbeck TS Hooper LV Gordon JI Developmental regulation of intestinal angiogenesis by indigenous microbes via Paneth cells Proceedings of the National Academy of Sciences 2002 99 24 15451 15455 17 Pull SL Doherty JM Mills JC Gordon JI Stappenbeck TS Activated macrophages are an adaptive element of the colonic epithelial progenitor niche necessary for regenerative responses to injury Proceedings of the National Academy of Sciences 2005 102 1 99 104 18 Hooper LV Gordon JI Commensal Host-Bacterial Relationships in the Gut Science 2001 292 5519 1115 1118 11352068 19 Sartor RB Microbial Influences in Inflammatory Bowel Diseases Gastroenterology 2008 134 2 577 594 18242222 20 Noverr MC Huffnagle GB Does the microbiota regulate immune responses outside the gut? Trends Microbiol 2004 12 12 562 8 15539116 21 Wills-Karp M Santeliz J Karp CL The germless theory of allergic disease: revisiting the hygiene hypothesis Nat Rev Immunol 2001 1 1 69 75 11905816 22 Hord NG Eukaryotic-Microbiota Crosstalk: Potential Mechanisms for Health Benefits of Prebiotics and Probiotics Annu Rev Nutr 2008 28 15 1 17 18598136 23 Sansonetti PJ War and peace at mucosal surfaces Nat Rev Immunol 2004 4 12 953 64 15573130 24 Jones RM Neish AS Recognition of bacterial pathogens and mucosal immunity Cell Microbiol 2011 13 5 670 6 21352463 25 Ishii KJ Koyama S Nakagawa A Coban C Akira S Host innate immune receptors and beyond: making sense of microbial infections Cell Host Microbe 2008 3 6 352 63 18541212 26 Jones RM Sloane VM Wu H Luo L Kumar A Kumar MV Gewirtz AT Neish AS Flagellin administration protects gut mucosal tissue from irradiation-induced apoptosis via MKP-7 activity Gut 2011 60 5 648 57 21199832 27 Zeng H Wu H Sloane V Jones R Yu Y Lin P Gewirtz AT Neish AS Flagellin/TLR5 responses in epithelia reveal intertwined activation of inflammatory and apoptotic pathways Am J Physiol Gastrointest Liver Physiol 2006 290 1 G96 G108 16179598 28 Vijay-Kumar M Wu H Jones R Grant G Babbin B King TP Kelly D Gewirtz AT Neish AS Flagellin suppresses epithelial apoptosis and limits disease during enteric infection Am J Pathol 2006 169 5 1686 700 17071592 29 Rakoff-Nahoum S Paglino J Eslami-Varzaneh F Edberg S Medzhitov R Recognition of Commensal Microflora by Toll-Like Receptors Is Required for Intestinal Homeostasis Cell 2004 118 2 229 241 15260992 30 Lee J Mo JH Katakura K Alkalay I Rucker AN Liu YT Lee HK Shen C Cojocaru G Shenouda S Kagnoff M Eckmann L Ben-Neriah Y Raz E Maintenance of colonic homeostasis by distinctive apical TLR9 signalling in intestinal epithelial cells Nat Cell Biol 2006 8 12 1327 36 17128265 31 Migeotte I Communi D Parmentier M Formyl peptide receptors: a promiscuous subfamily of G protein-coupled receptors controlling immune responses Cytokine Growth Factor Rev 2006 17 6 501 19 17084101 32 Le Y Murphy PM Wang JM Formyl-peptide receptors revisited Trends Immunol 2002 23 11 541 8 12401407 33 Gao JL Lee EJ Murphy PM Impaired antibacterial host defense in mice lacking the N-formylpeptide receptor J Exp Med 1999 189 4 657 62 9989980 34 Zhang Q Raoof M Chen Y Sumi Y Sursal T Junger W Brohi K Itagaki K Hauser CJ Circulating mitochondrial DAMPs cause inflammatory responses to injury Nature 2010 464 7285 104 7 20203610 35 Ali H Richardson RM Tomhave ED Didsbury JR Snyderman R Differences in phosphorylation of formylpeptide and C5a chemoattractant receptors correlate with differences in desensitization J Biol Chem 1993 268 32 24247 54 8226971 36 Tardif M Mery L Brouchon L Boulay F Agonist-dependent phosphorylation of N-formylpeptide and activation peptide from the fifth component of C (C5a) chemoattractant receptors in differentiated HL60 cells J Immunol 1993 150 8 Pt 1 3534 45 8468487 37 Bokoch GM Gilman AG Inhibition of receptor-mediated release of arachidonic acid by pertussis toxin Cell 1984 39 2 Pt 1 301 8 6094010 38 Lad PM Olson CV Smiley PA Association of the N-formyl-Met-Leu-Phe receptor in human neutrophils with a GTP-binding protein sensitive to pertussis toxin Proc Natl Acad Sci U S A 1985 82 3 869 73 2983319 39 Jesaitis AJ Naemura JR Sklar LA Cochrane CG Painter RG Rapid modulation of N-formyl chemotactic peptide receptors on the surface of human granulocytes: formation of high-affinity ligand-receptor complexes in transient association with cytoskeleton J Cell Biol 1984 98 4 1378 87 6325470 40 Sklar LA Hyslop PA Oades ZG Omann GM Jesaitis AJ Painter RG Cochrane CG Signal transduction and ligand-receptor dynamics in the human neutrophil. Transient responses and occupancy-response relations at the formyl peptide receptor J Biol Chem 1985 260 21 11461 7 2995337 41 Jesaitis AJ Tolley JO Painter RG Sklar LA Cochrane CG Membrane-cytoskeleton interactions and the regulation of chemotactic peptide-induced activation of human granulocytes: the effects of dihydrocytochalasin B J Cell Biochem 1985 27 3 241 53 3838753 42 Babbin BA Jesaitis AJ Ivanov AI Kelly D Laukoetter M Nava P Parkos CA Nusrat A Formyl peptide receptor-1 activation enhances intestinal epithelial cell restitution through phosphatidylinositol 3-kinase-dependent activation of Rac1 and Cdc42 J Immunol 2007 179 12 8112 21 18056353 43 Wentworth CC Jones RM Kwon YM Nusrat A Neish AS Commensal-epithelial signaling mediated via formyl peptide receptors Am J Pathol 2010 177 6 2782 90 21037077 44 Cheng G Lambeth JD NOXO1, Regulation of Lipid Binding, Localization, and Activation of Nox1 by the Phox Homology (PX) Domain J Biol Chem 2004 279 6 4737 42 14617635 45 Lambeth JD NOX enzymes and the biology of reactive oxygen Nat Rev Immunol 2004 4 3 181 9 15039755 46 Bedard K Krause KH The NOX family of ROS-generating NADPH oxidases: physiology and pathophysiology Physiol Rev 2007 87 1 245 313 17237347 47 Ogier-Denis E Mkaddem SB Vandewalle A NOX enzymes and Toll-like receptor signaling Semin Immunopathol 2008 48 Ardita CS Mercante JW Kwon YM Luo L Crawford ME Powell DN Jones RM Neish AS Epithelial adhesion mediated by pilin SpaC is required for Lactobacillus rhamnosus GG-induced cellular responses Appl Environ Microbiol 2014 80 16 5068 77 24928883 49 Kotchoni SO Gachomo EW The reactive oxygen species network pathways:an essential prerequisite for perception of pathogen attack and the acquired disease resistance in plants J Biosci 2006 31 3 389 404 17006022 50 Pauly N Pucciariello C Mandon K Innocenti G Jamet A Baudouin E Herouart D Frendo P Puppo A Reactive oxygen and nitrogen species and glutathione: key players in the legume-Rhizobium symbiosis J Exp Bot 2006 57 8 1769 1776 16698817 51 Tanaka A Christensen MJ Takemoto D Park P Scott B Reactive Oxygen Species Play a Role in Regulating a Fungus-Perennial Ryegrass Mutualistic Interaction Plant Cell 2006 18 4 1052 1066 16517760 52 Terada LS Specificity in reactive oxidant signaling: think globally, act locally J Cell Biol 2006 174 5 615 23 16923830 53 Tsukagoshi H Busch W Benfey PN Transcriptional regulation of ROS controls transition from proliferation to differentiation in the root Cell 2010 143 4 606 16 21074051 54 Ha EM Oh CT Ryu JH Bae YS Kang SW Jang IH Brey PT Lee WJ An antioxidant system required for host protection against gut infection in Drosophila Dev Cell 2005 8 1 125 32 15621536 55 Ha E-M Oh C-T Bae YS Lee W-J A Direct Role for Dual Oxidase in Drosophila Gut Immunity Science 2005 310 5749 847 850 16272120 56 Luo L Reedy AR Jones RM Detecting Reactive Oxygen Species Generation and Stem Cell Proliferation in the Drosophila Intestine Methods Mol Biol 2016 1422 103 13 27246026 57 Jang JY Rhee JY Chung GC Kang H Aquaporin as a membrane transporter of hydrogen peroxide in plant response to stresses Plant Signal Behav 2012 7 9 1180 1 22899075 58 Velde AAT Pronk I de Kort F Stokkers PCF Glutathione peroxidase 2 and aquaporin 8 as new markers for colonic inflammation in experimental colitis and inflammatory bowel diseases: an important role for H(2)O(2)? European Journal of Gastroenterology & Hepatology 2008 20 6 555 560 18467915 59 Laforenza U Water channel proteins in the gastrointestinal tract Molecular Aspects of Medicine 2012 33 5–6 642 650 22465691 60 Hernández-García D Wood CD Castro-Obregón S Covarrubias L Reactive oxygen species: A radical role in development? Free Radical Biology and Medicine 2010 49 2 130 143 20353819 61 Karrasch T Kim JS Muhlbauer M Magness ST Jobin C Gnotobiotic IL-10−/−;NF-kappa B(EGFP) mice reveal the critical role of TLR/NF-kappa B signaling in commensal bacteria-induced colitis J Immunol 2007 178 10 6522 32 17475882 62 Barford D The role of cysteine residues as redox-sensitive regulatory switches Curr Opin Struct Biol 2004 14 6 679 86 15582391 63 Kamata H Honda S Maeda S Chang L Hirata H Karin M Reactive oxygen species promote TNFalpha-induced death and sustained JNK activation by inhibiting MAP kinase phosphatases Cell 2005 120 5 649 61 15766528 64 Rhee SG Kang SW Jeong W Chang TS Yang KS Woo HA Intracellular messenger function of hydrogen peroxide and its regulation by peroxiredoxins Curr Opin Cell Biol 2005 17 2 183 9 15780595 65 Chiarugi P Buricchi F Protein tyrosine phosphorylation and reversible oxidation: two crosstalking posttranslation modifications Antioxid Redox Signal 2007 9 1 1 24 17115885 66 Swanson PA Kumar A Samarin S Vijay-Kumar M Kundu K Murthy N Hansen J Nusrat A Neish AS Enteric commensal bacteria potentiate epithelial restitution via reactive oxygen species-mediated inactivation of focal adhesion kinase phosphatases Proceedings of the National Academy of Sciences 2011 108 21 8803 8808 67 Tonks NK Redox redux: revisiting PTPs and the control of cell signaling Cell 2005 121 5 667 70 15935753 68 Neish AS Jones RM Redox signaling mediates symbiosis between the gut microbiota and the intestine Gut Microbes 2014 5 2 250 3 24637602 69 Jones RM Mercante JW Neish AS Reactive oxygen production induced by the gut microbiota: pharmacotherapeutic implications Curr Med Chem 2012 19 10 1519 29 22360484 70 Kumar A Wu H Collier-Hyams LS Hansen JM Li T Yamoah K Pan ZQ Jones DP Neish AS Commensal bacteria modulate cullin-dependent signaling via generation of reactive oxygen species Embo J 2007 26 21 4457 66 17914462 71 Yan F Cao H Cover TL Whitehead R Washington MK Polk DB Soluble Proteins Produced by Probiotic Bacteria Regulate Intestinal Epithelial Cell Survival and Growth Gastroenterology 2007 132 2 562 575 17258729 72 Kundu K Knight SF Willett N Lee S Taylor WR Murthy N Hydrocyanines: A Class of Fluorescent Sensors That Can Image Reactive Oxygen Species in Cell Culture, Tissue, and In Vivo Angewandte Chemie-International Edition 2009 48 2 299 303 73 Menard S Candalh C Bambou JC Terpend K Cerf-Bensussan N Heyman M Lactic acid bacteria secrete metabolites retaining anti-inflammatory properties after intestinal transport Gut 2004 53 6 821 8 15138208 74 Pena JA Versalovic J Lactobacillus rhamnosus GG decreases TNF-alpha production in lipopolysaccharide-activated murine macrophages by a contact-independent mechanism Cell Microbiol 2003 5 4 277 85 12675685 75 Madsen KL Doyle JS Jewell LD Tavernini MM Fedorak RN Lactobacillus species prevents colitis in interleukin 10 gene-deficient mice Gastroenterology 1999 116 5 1107 14 10220502 76 Petrof EO Kojima K Ropeleski MJ Musch MW Tao Y De Simone C Chang EB Probiotics inhibit nuclear factor-kappaB and induce heat shock proteins in colonic epithelial cells through proteasome inhibition Gastroenterology 2004 127 5 1474 87 15521016 77 Neish AS Gewirtz AT Zeng H Young AN Hobert ME Karmali V Rao AS Madara JL Prokaryotic Regulation of Epithelial Responses by Inhibition of Ikappa B-alpha Ubiquitination Science 2000 289 5484 1560 1563 10968793 78 Collier-Hyams LS Sloane V Batten BC Neish AS Cutting Edge: Bacterial Modulation of Epithelial Signaling via Changes in Neddylation of Cullin-1 J Immunol 2005 175 7 4194 4198 16177058 79 Lee WJ Bacterial-modulated signaling pathways in gut homeostasis Sci Signal 2008 1 21 pe24 18506033 80 Zhou BP Hung MC Wnt, hedgehog and snail: sister pathways that control by GSK-3beta and beta-Trcp in the regulation of metastasis Cell Cycle 2005 4 6 772 6 15917668 81 Chiarugi P Pani G Giannoni E Taddei L Colavitti R Raugei G Symons M Borrello S Galeotti T Ramponi G Reactive oxygen species as essential mediators of cell adhesion: the oxidative inhibition of a FAK tyrosine phosphatase is required for cell adhesion J Cell Biol 2003 161 5 933 44 12796479 82 Diaz B Shani G Pass I Anderson D Quintavalle M Courtneidge SA Tks5-Dependent, Nox-Mediated Generation of Reactive Oxygen Species Is Necessary for Invadopodia Formation Science Signaling 2009 2 88 83 Buchon N Broderick NA Chakrabarti S Lemaitre B Invasive and indigenous microbiota impact intestinal stem cell activity through multiple pathways in Drosophila Genes Dev 2009 23 19 2333 44 19797770 84 Buchon N Broderick NA Poidevin M Pradervand S Lemaitre B Drosophila intestinal response to bacterial infection: activation of host defense and stem cell proliferation Cell Host Microbe 2009 5 2 200 11 19218090 85 Owusu-Ansah E Banerjee U Reactive oxygen species prime Drosophila haematopoietic progenitors for differentiation Nature 2009 461 7263 537 41 19727075 86 Coant N Ben Mkaddem S Pedruzzi E Guichard C Treton X Ducroc R Freund JN Cazals-Hatem D Bouhnik Y Woerther PL Skurnik D Grodet A Fay M Biard D Lesuffleur T Deffert C Moreau R Groyer A Krause KH Daniel F Ogier-Denis E NADPH oxidase 1 modulates WNT and NOTCH1 signaling to control the fate of proliferative progenitor cells in the colon Mol Cell Biol 30 11 2636 50 87 Venugopal R Jaiswal AÄ Nrf1 and Nrf2 positively and c-Fos and Fra1 negatively regulate the human antioxidant response element-mediated expression of NAD(P)H:quinone oxidoreductase1‚Äâgene Proceedings of the National Academy of Sciences 1996 93 25 14960 14965 88 An JH Blackwell TK SKN-1 links C. elegans mesendodermal specification to a conserved oxidative stress response Genes & Development 2003 17 15 1882 1893 12869585 89 Sykiotis GP Bohmann D Keap1/Nrf2 Signaling Regulates Oxidative Stress Tolerance and Lifespan in Drosophila Developmental Cell 2008 14 1 76 85 18194654 90 Kobayashi M Itoh K Suzuki T Osanai H Nishikawa K Katoh Y Takagi Y Yamamoto M Identification of the interactive interface and phylogenic conservation of the Nrf2-Keap1 system Genes to Cells 2002 7 8 807 820 12167159 91 Itoh K Chiba T Takahashi S Ishii T Igarashi K Katoh Y Oyake T Hayashi N Satoh K Hatayama I Yamamoto M Nabeshima Y-i An Nrf2/Small Maf Heterodimer Mediates the Induction of Phase II Detoxifying Enzyme Genes through Antioxidant Response Elements Biochemical and Biophysical Research Communications 1997 236 2 313 322 9240432 92 Motohashi H O’Connor T Katsuoka F Engel JD Yamamoto M Integration and diversity of the regulatory network composed of Maf and CNC families of transcription factors Gene 2002 294 1–2 1 12 12234662 93 Kobayashi A Kang M-I Okawa H Ohtsuji M Zenke Y Chiba T Igarashi K Yamamoto M Oxidative Stress Sensor Keap1 Functions as an Adaptor for Cul3-Based E3 Ligase To Regulate Proteasomal Degradation of Nrf2 Mol Cell Biol 2004 24 16 7130 7139 15282312 94 Jones RM Desai C Darby TM Luo L Wolfarth AA Scharer CD Ardita CS Reedy AR Keebaugh ES Neish AS Lactobacilli Modulate Epithelial Cytoprotection through the Nrf2 Pathway Cell Rep 2015 12 8 1217 25 26279578 95 Malhotra D Portales-Casamar E Singh A Srivastava S Arenillas D Happel C Shyr C Wakabayashi N Kensler TW Wasserman WW Biswal S Global mapping of binding sites for Nrf2 identifies novel targets in cell survival response through ChIP-Seq profiling and network analysis Nucleic Acids Research 2010 38 17 5718 5734 20460467 96 Jimenez-Osorio AS Gonzalez-Reyes S Pedraza-Chaverri J Natural Nrf2 activators in diabetes Clin Chim Acta 2015 448 182 92 26165427 97 Furfaro AL Traverso N Domenicotti C Piras S Moretta L Marinari UM Pronzato MA Nitti M The Nrf2/HO-1 Axis in Cancer Cell Growth and Chemoresistance Oxid Med Cell Longev 2016 2016 1958174 26697129 98 Ryoo IG Lee SH Kwak MK Redox Modulating NRF2: A Potential Mediator of Cancer Stem Cell Resistance Oxid Med Cell Longev 2016 2016 2428153 26682001 99 Leinonen HM Kansanen E Polonen P Heinaniemi M Levonen AL Dysregulation of the Keap1-Nrf2 pathway in cancer Biochem Soc Trans 2015 43 4 645 9 26551706 100 Esteras N Dinkova-Kostova AT Abramov AY Nrf2 activation in the treatment of neurodegenerative diseases: a focus on its role in mitochondrial bioenergetics and function Biol Chem 2016 101 Silva-Palacios A Konigsberg M Zazueta C Nrf2 signaling and redox homeostasis in the aging heart: A potential target to prevent cardiovascular diseases? Ageing Res Rev 2016 26 81 95 26732035 102 Zhang C Shu L Kong AT MicroRNAs: New players in cancer prevention targeting Nrf2, oxidative stress and inflammatory pathways Curr Pharmacol Rep 2015 1 1 21 30 26618104 103 Schäfer M Dütsch S auf dem Keller U Navid F Schwarz A Johnson DA Johnson JA Werner S Nrf2 establishes a glutathione-mediated gradient of UVB cytoprotection in the epidermis Genes & Development 2010 24 10 1045 1058 20478997 104 Sethuraman M McComb ME Huang H Huang S Heibeck T Costello CE Cohen RA Isotope-coded affinity tag (ICAT) approach to redox proteomics: identification and quantitation of oxidant-sensitive cysteine thiols in complex protein mixtures J Proteome Res 2004 3 6 1228 33 15595732 105 Fisher EM Khan M Salisbury R Kuppusamy P Noninvasive Monitoring of Small Intestinal Oxygen in a Rat Model of Chronic Mesenteric Ischemia Cell Biochemistry and Biophysics 2013 67 2 451 459 23636684 106 Fan J Cai H Tan WS Role of the plasma membrane ROS-generating NADPH oxidase in CD34+ progenitor cells preservation by hypoxia J Biotechnol 2007 130 4 455 62 17610977

PMC005xxxxxx/PMC5131720.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 0404511 7473 Science Science Science (New York, N.Y.) 0036-8075 1095-9203 26941323 5131720 10.1126/science.aad3680 NIHMS828767 Article Activation of PKA leads to mesenchymal-to-epithelial transition and loss of tumor-initiating ability Pattabiraman Diwakar R. 1 Bierie Brian 1 Kober Katharina Isabelle 1 Thiru Prathapan 1 Krall Jordan 1 Zill Christina 1 Reinhardt Ferenc 1 Tam Wai Leong 145 Weinberg Robert A. 123* 1 Whitehead Institute for Biomedical Research, Cambridge, Massachusetts, USA 2 Department of Biology, Massachusetts Institute of Technology, Cambridge, Massachusetts, USA 3 Ludwig MIT Center for Molecular Oncology, Cambridge, Massachusetts, USA 4 Genome Institute of Singapore, 60 Biopolis Street, Singapore 5 Cancer Science Institute of Singapore, 14 Medical Drive, Singapore * Correspondence: weinberg@wi.mit.edu (R.A.W.) 10 11 2016 4 3 2016 01 12 2016 351 6277 aad3680aad3680 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. The epithelial-to-mesenchymal transition (EMT) enables carcinoma cells to acquire malignancy-associated traits and properties of tumor-initiating cells (TICs). TICs have emerged in recent years as important targets for cancer therapy owing to their ability to drive clinical relapse and enable metastasis. Here we propose a strategy to eliminate mesenchymal TICs by inducing their conversion to more epithelial counterparts that have lost tumor-initiating ability. We report that increases in intracellular levels of the second messenger, cAMP, and the subsequent activation of protein kinase A (PKA) induce a mesenchymal-to-epithelial transition (MET) in mesenchymal human mammary epithelial cells. Activation of PKA triggers PHF2-mediated epigenetic reprogramming of TICs, promoting their differentiation that leads to loss of tumor-initiating ability. This study provides proof-of-principle for inducing an MET as differentiation therapy for TICs and uncovers a novel role for PKA in enforcing and maintaining the epithelial state. One Sentence Summary We identify a novel role for the activation of PKA and downstream epigenetic reprogramming that results in the differentiation of tumor-initiating cells in aggressive breast cancers. Introduction Tumor-initiating cells (TICs), also known as cancer stem cells, are defined operationally by their ability to seed new tumors upon implantation in appropriate hosts. They have emerged in recent years as important targets for cancer therapy owing to their elevated resistance to conventional chemotherapy and their tumor-initiating ability; the latter allowing them to metastasize and drive clinical relapse (1, 2). While their mode of generation and biological properties have been explored in a diverse array of cancer types (3), our understanding of the biology of TICs remains superficial. Cytotoxic therapies designed specifically to eliminate TICs might be targeted, for example, to interdict the signaling pathways that are used preferentially or uniquely by these cells (4). At present, however, the nature of such TIC-specific signaling pathways remains to be fully elucidated. The epithelial-to-mesenchymal transition (EMT) is a cell-biological program that confers mesenchymal traits on both normal and neoplastic epithelial cells (5). In addition, activation of an EMT program enables both classes of cells to acquire stem-like properties (6, 7). Indeed, TICs from several carcinoma types possess distinct mesenchymal attributes, suggesting that they have passed, at least partially, through an EMT (7–9). This association between the EMT program and the TIC state has presented an attractive opportunity for drug development, using agents that preferentially target more mesenchymal carcinoma cells rather than their epithelial counterparts in an effort to eliminate TICs. At least two approaches might be taken to target mesenchymal TICs. One strategy would be to develop agents that show specific or preferential cytotoxicity toward TICs (1). In this study, we have embraced an alternative strategy that is designed to induce TICs to exit the more mesenchymal tumor-initiating state and enter into an epithelial non-stemlike state. Such induced differentiation should, we reasoned, place cells in a state where they would become more vulnerable to conventional cytotoxic treatments. Accordingly, we screened for agents that could induce a mesenchymal-to-epithelial transition (MET) and thereby uncovered the central role of 3’-5’-cyclic adenosine monophosphate (cAMP) and its downstream target, protein kinase A (PKA), in governing the transition of cells from the mesenchymal to the epithelial state. cAMP is a second messenger that transmits intracellular signals upon interaction of certain hormones and neurotransmitters with receptors on the plasma membrane (10). cAMP regulates multiple downstream effectors; the first of these to be identified and the most well-studied is protein kinase A (PKA) (11), which plays numerous roles in various cell types and operates in several subcellular locations (11). Being initially assembled as a heterotetrameric holoenzyme, the activity of PKA depends on binding of cAMP to its two regulatory subunits, which leads to the release of active catalytic subunits and the phosphorylation of a diverse array of substrates (12). In previous work, PKA has been shown, under some conditions, to promote an EMT; PKA was shown to regulate Snail in one study and another study demonstrated that HIF1α could regulate transcription of PRKACA under hypoxic conditions (13, 14). On the other hand, PKA signaling has been shown to favor the epithelial state, but the mechanistic understanding of this phenomenon is very limited. One report identified that schwannomas in Prkar1a (encoding the PKA regulatory subunit)-null mice exhibited loss of vimentin and gain of cytokeratins and E-cadherin (15), whereas another study revealed inhibition of formation of mesoderm-derived structures in Prkar1a null mice (16). A recent study reported that deletion of the Gαs subunit repressed the activity of PKA, limiting the proliferative potential of epithelial hair follicle stem cells (17). Nevertheless, the connection of PKA signaling to TICs and the stem-like state is poorly understood and the exploitation of this pathway as a differentiation-based cancer therapy has not been explored. Results Identification of agents that induce an MET in mammary epithelial cells Human breast cancers are characterized by cells exhibiting various degrees of epithelial and mesenchymal properties as revealed by the expression pattern of markers such as cytokeratins and vimentin (Fig S1). Almost 85% of the carcinomas we examined showed varied expression patterns of cytokeratins, indicating that the loss of epithelial properties is a commonly occurring event. Notably, ∼10% of the carcinomas we examined exhibited high degrees of intra-tumoral heterogeneity, created in part by the presence of subpopulations of neoplastic cells that exhibit both epithelial and mesenchymal properties. These are reminiscent of cells that have undergone an EMT, resembling TICs that possess a higher tumor-initiating propensity and an increased resistance to chemotherapy (18). To model the behavior of these subpopulations of carcinoma cells from human basal-like breast cancers, we used HMLE immortalized human mammary epithelial cells (19), which display an epithelial morphology, express E-cadherin at adherens junctions and low levels of mesenchymal markers such as vimentin and fibronectin. They also exhibit a CD44lo/CD24hi cell surface marker phenotype that is characteristic of previously reported cells that lack stemlike properties (non-CSCs) (20). We also used their spontaneously arising mesenchymal derivatives, termed NAMEC8 (N8) cells (21). Relative to their HMLE counterparts, N8 cells express mesenchymal markers such as vimentin and fibronectin as well as the EMT-inducing transcription factors Snail and Zeb1 at higher levels, lack expression of E-cadherin and prominent cell junctions, and display a CSC-like CD44hi/CD24lo cell surface marker profile (Fig 1A to C). They also possess a greater propensity to form mammospheres (Fig 1D, E), which is often used as an in vitro surrogate assay for the stemness of mammary epithelial cells. They are more efficient at migration through a transwell membrane and invasion through a Matrigel-coated Boyden chamber membrane (Fig 1F and G); both in vitro assays represent models of cancer cell invasiveness in vivo. N8 cells are also more resistant to treatment with chemotherapeutic drugs such as doxorubicin and paclitaxel (Fig 1H, I), as shown previously (21). Hence, two cell types represent epithelial and mesenchymal derivatives of mammary epithelial cells of common origin that were used to model the two cell states and how they impact tumor initiation and progression. To search for agents that can induce an MET, we performed a screen to identify compounds that could induce transcription of CDH1, which encodes E-cadherin, a key epithelial protein, in N8 cells. As a reporter for the activity of the CDH1 gene, we constructed a lentiviral vector that expresses a portion of the CDH1 promoter fused to luciferase (Fig S2A). We performed a screen using a 400-compound library for agents that were able to induce the CDH1-driven luciferase expression in N8 cells (Fig S2B). Most striking was the behavior of forskolin (Fsk), an adenylate cyclase activator that induced a 40-fold increase in luciferase activity (Fig S2C). Another adenylate cyclase activator, cholera toxin (CTx), was also able to induce an increase in luciferase activity (Fig S2D), suggesting that activation of adenylate cyclase could induce the acquisition of epithelial properties. Forskolin or Cholera Toxin and the induction of an MET in mammary epithelial cells We found that treatment of N8 cells in monolayer culture with either CTx or Fsk for a period of 14 days induced the formation of islands of cells with the characteristic cobblestone morphology of epithelial cells; such cells acquired the expression of E-cadherin at adherens junctions along with a loss of mesenchymal markers such as vimentin (Fig 1A–C). Also, the cell-surface marker expression profile of the N8 cells switched from a stem-like CD44hi/CD24lo to a non-stem CD44lo/CD24hi phenotype following this treatment (20) (Fig 1B). These shifts were accompanied by a 100-fold increase in CDH1 mRNA levels, as well as 4-, 5- and 7-fold decreases in the mRNAs levels of Snail, Twist1 and Zeb1 EMT-inducing transcription factors (EMT-TFs) to 25%, 20% and 14%, respectively, of the N8 cells before the transition (Fig S3A and B). Treatment of N8 cells with either CTx or Fsk resulted in a near-complete loss of mammosphere-forming ability (Fig 1D and E), as well as their ability to migrate and invade (Fig 1F and G). There were no significant differences in the rates of proliferation between the N8 cells and their CTx- and Fsk-treated derivatives (Fig S3C). Of additional interest, withdrawal of CTx after 14 days of treatment led to cell populations that continued to reside in an epithelial state for >2 months in culture. Reversion to an epithelial state, ostensibly similar to that of HMLE cells, rendered the N8 cells 8 times as sensitive to killing by doxorubicin (lowered the median inhibitory concentration (IC50) from 1.39µM to 0.159µM), and 13 times as sensitive to paclitaxel (lowered the IC50 from 4.79µM to 0.35µM) (Fig 1H and I). Additionally, the induced MET also resulted in increased sensitivity to a range of chemotherapeutic drugs and inhibitors including methotrexate, HSP90 inhibitors, proteasome inhibitors, and epidermal growth factor receptor-mitogen-activated protein kinase (EGFR/MAPK) pathway inhibitors, as observed when we screened against two small molecule libraries (Selleck Anti-cancer Compound Library and Enzo Kinase Inhibitor Library)(Fig S4). Hence, the induction of an MET rendered the N8 cells more sensitive to a range of drugs and inhibitors, pointing to its utility as a means of overcoming therapeutic resistance. Additionally, it also reinforces the notion that mesenchymal cells are more resistant to a range of cytotoxic agents. We then performed mRNA sequencing (mRNA-Seq) to compare the global gene expression profiles of the mesenchymal N8 and the reverted N8-CTx cells in order to view the transcriptional changes that occur following the induction of MET. As determined by differential gene expression (Fig 1J, Table S1 and S2) and principal component analyses (Fig S3D), the N8-CTx cells assume a gene expression profile that is almost completely converted to that of the epithelial HMLE cells and significantly differ from the mesenchymal N8 cells (Fig 1J). Gene set enrichment analyses showed that the changes in gene expression from N8 to the N8-CTx cells are highly similar to several previously published EMT/MET gene signatures (22–24) (Fig S3E). Taken together, these observations demonstrated a genuine transition of the N8 cells from a mesenchymal-like state to a bona fide epithelial state, rendering these cells more sensitive to a variety of drugs with potentially important therapeutic implications. Effects of Forskolin and Cholera Toxin on intracellular cAMP levels and PKA To confirm that both Fsk and CTx were working through alteration of cAMP levels, we measured the levels of this second messenger in both HMLE and N8 cells using liquid chromatography - mass spectrometry (LC-MS). Treatment with CTx resulted in a 6–8-fold increase in the intracellular levels of cAMP, which could be dampened by exposure to SQ22536, an inhibitor of adenylate cyclase, the enzyme responsible for the formation of cAMP (Fig 2A). The major downstream targets of cAMP are exchange proteins activated by cAMP (EPAC1/2) (25), cyclic nucleotide gated ion channels which are primarily found in cells of the kidney, heart, testis and central nervous system (26); and the most commonly studied downstream effector, protein kinase A (PKA) (11). To delineate the downstream pathways that are activated in response to increase in cAMP levels, we treated N8 cells with two cAMP analogs – 8-Bromoadenosine-3’,5’-cyclic monophosphate (8-Br-cAMP), which is known to preferentially activate PKA (27) or 8-(4-chlorophenylthio)-2’-O-methyladenosine-3’,5’-cyclic monophosphate (8-CPT-2Me-cAMP), which is a selective activator of the exchange proteins activated by cAMP (EPAC) (28). As was seen with Fsk/CTx, treatment with 8-Br-cAMP was also able to induce an MET in N8 cells, whereas 8-CPT-2Me-cAMP treatment had no effect on their mesenchymal properties (Fig 2B). This allowed us to conclude that PKA, rather than the cAMP-activated exchange proteins, was more likely to play a central role in the MET process. Knockdown of the catalytic subunit of PKA using two different shRNAs (Fig S5A) abrogated the CTx-induced MET process in N8 cells, as demonstrated by their inability to develop a clear epithelial morphology, acquire junctional E-cadherin, and to shed mesenchymal markers such as fibronectin (Fig 2C, D). Moreover, treatment of these PKA-knockdown cells with CTx failed to induce an effective transition from the CD44hi/CD24lo stem-like state to the CD44lo/CD24hi non-stem-like state, which was otherwise observable in the absence of PKA knockdown (Fig 2E). These results further reinforced the important role of PKA in the MET process. We proceeded to test whether PKA activity, independent of cAMP, was sufficient to induce an MET. Thus, we ectopically expressed a doxycycline-inducible constitutively active, cAMP-independent mutant form of PKA (caPKA) (29) in N8 cells and found that it was capable of inducing a reversion to the epithelial state in 7–10 days (Fig 2F). Hence, it appears as though PKA is both necessary and sufficient to induce an MET in the N8 cells. We tested the role of CTx/Fsk in inducing an epithelial state in other cell systems to assess the generality of our observations. Removing CTx from the standard culture medium of MCF10A immortalized human mammary epithelial cells (30), caused them to acquire mesenchymal properties, lose cell-cell adherens junctions, lose their characteristic cobblestone morphology, gain a CD44hi/CD24lo cell surface marker profile. They also lost E-cadherin expression and exhibited an increase in expression of Zeb1, Vimentin and Fibronectin (Fig S6A–E). Re-addition of CTx or forced expression of the constitutively active PKA mutant (caPKA) led to the re-acquisition of epithelial features (Fig S6A–E). Moreover, the MCF10A cells that lost epithelial properties upon CTx withdrawal were 4 times as resistant to treatment with doxorubicin, extending our observations made in N8 cells that the mesenchymal variants were more resistant to conventional chemotherapeutic agents (S6F). We then proceeded to test the role of CTx/Fsk in a series of other cell lines. MCF7-Ras human breast cancer cells (31) can be induced to undergo an EMT through the ectopic expression of EMT-inducing transcription factors, such as Slug. Co-treatment of the cells undergoing an EMT with CTx led to a 48-hour delay in the acquisition of mesenchymal morphology and CD44hi cell-surface marker expression (Fig S7A). Similarly, the ability of HMLE-Ras cells to undergo an EMT upon ectopic expression of Zeb1 was also hampered upon co-treatment with CTx (Fig S7B). PANC1 pancreatic adenocarcinoma cells undergo an EMT upon treatment with TGF-β1 for 48 hours (32). Co-treatment of PANC1 cells undergoing an EMT with either CTx or Fsk delayed the ability of TGF-β1 to induce an EMT by 48–72 hours, enabling the temporary retention of epithelial properties (Fig S7C, D). Treatment with CTx or Fsk induced the acquisition of epithelial properties in a range of cell lines that have mesenchymal traits including the Hs578T triple-negative breast cancer cell line (Fig S8A), the SUM149 breast cancer cell line (Fig S8B), the NCI-H596 lung adenosquamous carcinoma cell line (Fig S8C) and the mesenchymal EpCAMlo CD24lo fraction of the EF021 ovarian carcinoma cell line (Fig S8D). Induction of epithelial properties was also observed in PB3 cells (Fig S8E), which constitute an aggressive cell line isolated from mammary tumors of the genetically engineered MMTV-PyMT mouse model of breast cancer, in which the expression of the oncogene is driven by the mouse mammary tumor virus promoter (33). Finally, we note that others have recently reported that forskolin promotes the maintenance of an epithelial morphology in primary human mammary epithelial cells, the absence of which led spontaneously to acquisition of mesenchymal attributes, such as downregulation of E-cadherin expression and upregulation of mesenchymal markers (34). Taken together, these data signify the importance of PKA signaling in maintaining epithelial characteristics in a variety of normal and neoplastic epithelial cells. These data give an indication that these responses might be a general property of cAMP-induced activation of PKA in the reversal of phenotypes created by activation of an EMT program. Although CTx was able to induce entrance of the N8 cells and a range of other cell systems into a stably maintained epithelial state, there were a few models in which neither CTx nor Fsk was able to do so, namely the MDA-MB-231 and SUM159 human breast cancer cell lines, amongst others. These cell lines are maintained in the mesenchymal state through the deletion or stable silencing of several key epithelial genomic loci, including the repression of E-cadherin through strong DNA promoter hypermethylation (35). Hence, although the observed effects of PKA activation are applicable to some breast cancer lines and other carcinomas, they are not universal and depend instead on the specific genetic or epigenetic state of the cells. Essential role of PKA-induced activation of PHF2 in MET PKA is known to act on many substrates in both the cytoplasm and nucleus (36). Treatment of HMLE and N8 cells with CTx resulted in an immediate increase in the presence of both isoforms of the catalytic subunit in the nucleus (Fig S9A and B), which suggested that PKA might be regulating nuclear substrates following activation by cAMP. The most well-studied substrate of PKA, CREB1, translocates to the nucleus upon phosphorylation by PKA at Ser-133, thereafter altering the transcription of hundreds of target genes (37). In fact, about 300 distinct physiologic stimuli have been described in the literature that can induce CREB Ser-133 phosphorylation (38). It was, therefore, not surprising that CREB was already phosphorylated and present in the nucleus of the N8 cells even prior to their treatment with either CTx or Fsk (Fig 3A). Note that knockdown of CREB1 using at least two shRNAs (Fig S5C) did not affect the ability of CTx to induce an MET in the N8 cells (Fig 3B). Moreover, loss of CREB1 alone induced a partial MET in N8 cells (Fig 3B), consistent with previous reports of its role in the induction of an EMT (39, 40). From these observations, it was obvious that CREB1 did not play an important role in the PKA-induced MET. We then assessed the localization of Gli1, Gli2 and Gli3, which have been previously reported to be PKA substrates that are retained in the cytoplasm following phosphorylation (41), and found no retention of any of the Gli proteins in the cytoplasm following treatment with CTx or Fsk (Fig S9C). These observations indicate that the Gli proteins may not play a role in the observed PKA-induced MET. Having explored the two most commonly reported nuclear substrates of PKA, we then chose to focus on PHF2, an H3K9 histone demethylase, which is known to become activated upon phosphorylation by PKA (42). We found that knockdown of PHF2 expression in N8 cells using either of two shRNAs (Fig S5B) phenocopied the effects of PKA knockdown in that it prevented CTx-induced mesenchymal-epithelial transition (Fig 3C and D). In contrast, knockdown of PHF2 did not alter the ability of HMLE cells to undergo an EMT (Fig 3F and G), indicating that this enzyme, while necessary for induction of an MET, apparently plays no role in the reverse process -- the EMT, suggesting that it is specifically important for the derepression of silenced epithelial genes through its function as a H3K9 histone demethylase. PHF2 can be phosphorylated by PKA at four serine residues in its C-terminus (42) (Fig 3E). Accordingly, we engineered a phospho-mimetic form of PHF2 in which all four of these serines were replaced by aspartate residues. While expression of this mutant in N8 cells was not sufficient on its own to induce an MET, the phospho-mimetic PHF2 was able to accelerate the rate of CTx-induced transition from the mesenchymal to the epithelial state from 14 days to 6–7 days (Fig 3H, I). Hence, while PHF2 is essential for MET, it appears to be only one of the effectors of PKA operating during induction of epithelial transition. To test whether PHF2 can be directly phosphorylated by PKA in our system, we performed an immunoprecipitation of PHF2 followed by immunoblotting using an antibody that recognizes phospho-PKA substrate motif. As shown in Fig 3J, 24 hours after treatment of N8 cells with CTx, phosphorylation of PHF2 by PKA can be observed, providing evidence that PKA phosphorylates PHF2 in the N8 cells. Together these results suggest an important role for PHF2 as a PKA substrate in the induction of an MET. PKA-induced activation of PHF2 and the epigenetic reprogramming of mesenchymal cells The H3K9me2 and H3K9me3 marks have been associated with repression of gene transcription (43). Given the previously reported role of PHF2 as an H3K9me2/3 demethylase, we performed chromatin immunoprecipitation followed by deep sequencing (ChIP-Seq) using antibodies against the H3K9me3 and H3K9me2 marks to observe the presence of these marks in untreated N8 cells as well as CTx-treated counterparts in which PHF2 is active. In addition we also performed ChIP-Seq for PHF2, comparing genome-wide occupancy in N8 cells to the N8-CTx cells. We did so in order to monitor PHF2-associated alterations that might enable phenotypic shifts from the mesenchymal to epithelial states, including shifts that might relieve the H3K9-mediated silencing of epithelial genes. As seen in Fig 4A, there was a striking inverse correlation at specific loci of the presence of PHF2 with the repressive H3K9me2 or H3K9me3 marks. This suggests that presence of this demethylase may, on it own, suffice to relieve histone-mediated transcriptional silencing. As previously reported, PHF2 appears to occupy the promoter region of genes where it recognizes the H3K4me3 histone mark (Fig 4B) (9). Interestingly, the total H3K9me3 counts (greater than 4-fold enrichment above control) in N8-CTx cells was almost half of the total counts of the same mark in N8 cells (35,455 vs 18,675). Similarly, the total H3K9me2 counts in N8-CTx cells were also less than a half of that in the N8 cells (1295 vs 473). As shown in the representative circos plots, these data indicate a widespread loss of H3K9-mediated repression of genomic regions upon treatment of N8 cells with CTx and subsequent activation of PHF2 (Fig 4C). We then sorted for genomic regions present in the N8-CTx but not N8 cells that contained PHF2 binding and lacked repressive H3K9me2/3 marks (Table S3). This provided us with a list of genomic regions that were relieved of H3K9me2/3-mediated silencing in the N8-CTx cells, as compared to the N8 cells, owing to PHF2 occupancy. To ensure that these changes were specific for the loss of PHF2, we performed ChIP-Seq for H3K9me2/3 and PHF2 in CTx-treated N8 cells that had an shRNA against PHF2 preventing the MET (Table S4). These cells that remained morphologically mesenchymal also demonstrated a similar epigenetic profile to N8 cells with an overlap of 11,807 peaks than the reverted N8-CTx cells, which had an overlap of 6864 peaks. Hence the list of altered genomic regions outlined in Table S3 represents genes that were relieved of H3K9-mediated repression upon CTx-induced activation of PHF2. This suggests that PHF2 activity could be directly responsible for the derepression of these genes that are characteristic of the epithelial cell state. In addition, the expression values of genes that correspond to these genomic loci were also measured in reverted N8-CTx (epithelial) and parental N8 (mesenchymal) cells by RNA-seq which verified that gain of PHF2 occupancy and loss of H3K9 marks did indeed lead to increased expression (Table S3). Several genes that play a major role in the phenotype and profile of cells in the epithelial state were activated by CTx treatment. Amongst the list of genes that were relieved of silencing upon treatment with CTx include CDH1 and CDH3 (among other cadherin genes) that code for E-cadherin and P-cadherin (Fig S10A), respectively, which are essential components of adherens junctions and hallmark proteins of basal epithelial cells; KRT8 and KRT18 (Fig S10B), whose gene products are characteristic components of the cytoskeleton of epithelial cells; and AJAP1 and CLDN4 (Fig S10C, D), which specify genes coding for constituents of adherens and tight junctions that are formed by epithelial but not mesenchymal cells. Other regions include the TP63 gene (Fig S10E) whose product is a hallmark transcription factor of basal mammary epithelial cells and ITGB2, ITGB6 (Fig S10F) and ITGB8, which code for integrins that are typically expressed on epithelial cells. These observations reveal a mechanism by which activation of this demethylase enables the transcription of genes that induce an MET and ultimately define the state of the cells. Activation of PKA and the differentiation of TICs in vivo We tested the tumor-initiating ability of cells that have been induced to undergo an MET by activation of PKA in vitro. We transplanted at limiting dilutions the neoplastic, RAS-transformed derivatives of HMLE, N8 and the reverted N8-CTx cells, termed HMLE-Ras, N8-Ras, and N8-CTx-Ras, into the mammary fat pads of NOD/SCID mice. As anticipated, the frequency of tumor-initiating cells in the N8-Ras cells was far greater than in the HMLE-Ras cell population, in this case 100-fold higher. Significantly, the N8-CTx-Ras cells were as inefficient at tumor-initiation as the HMLE-Ras cells (Fig 5A). The primary tumors that arose upon orthotopic mammary stromal fat pad implantation of N8-Ras tumors spawned 20 to 30 micrometastases in the lungs by 12 weeks following implantation. This property was completely lost upon induction of an MET by CTx treatment prior to transplantation (Fig 5C and Fig S11A), which nevertheless formed primary tumors of comparable size (Fig 5B). Moreover, this confirmed previous observations that the phenotypic state of these cells prior to neoplastic transformation strongly influenced their behavior following transformation. To better mimic a clinical scenario, we next asked how the induction of an MET would impact pre-established tumors derived from mesenchymal N8 cells. While we wished to pharmacologically treat mice that already had established N8-Ras tumors, CTx is too toxic to be administered systemically, and the rapid clearance and poor pharmacodynamics of Fsk made it difficult to study its effects upon systemic administration. Such difficulties in treating mice with PKA agonists have also been reported previously (44, 45). For this reason, we focused our efforts on studying the proof-of-principle effects of PKA activation in vivo using the doxycycline-inducible version of constitutively active PKA (caPKA). Thus, we induced expression of the caPKA in already-formed N8-Ras tumors of 5mm diameter (Fig 5D). Upon visual inspection, the tumors that had been exposed for 14 days to doxycycline contained pasty, fluid-filled necrotic cores when compared to the tumors that had never been exposed to doxycycline, the latter being solid with a hard center of viable cells. Tumors from mice that received doxycycline weighed less than those that did not receive any (Fig 5F). Moreover, those tumors in which expression of caPKA had been induced developed a more differentiated histomorphology as revealed by H&E staining of tumor sections (Fig S11B, C). Importantly, when tumors were harvested and subjected to FACS analysis, the doxycycline-treated tumors showed a decrease in expression of the CD44 cell-surface marker associated with the stem-like population (20) in contrast to the untreated tumors (Fig S11D). Strikingly, secondary transplantation of cells isolated from the doxycycline-exposed tumors at limiting dilutions revealed a ∼20-fold loss of tumor-initiating ability (Fig 5E), showing that activation of PKA induces differentiation of TICs, diminishing their ability to subsequently seed new tumors. This result demonstrates that constitutive expression of PKA for a 14-day period in a growing tumor suffices on its own to induce the differentiation of TICs in the tumor, reducing the tumor-initiating properties of its associated cells as indicated by their subsequent inability to propagate tumors upon secondary transplantation. Discussion Cyclic AMP and its main effector, PKA, have been studied for four decades in a variety of cell-biological and physiologic settings, where its effects in activating a number of distinct, tissue-specific traits have been repeatedly documented (11). A role that it might play in governing the epithelial cell state and thus suppressing entrance into the alternative mesenchymal state in breast cancers has not been described. The present work makes its clear that this second messenger and its main effector, PKA, play a key role in determining this epithelial versus mesenchymal balance of mammary epithelial cells, as well as epithelial cells of other tissues. Indeed, in light of these results, it becomes plausible that maintenance of the residence of cells in an epithelial state depends on tonic elevated levels of intracellular cAMP. In retrospect, it now seems likely that the use of cholera toxin as an ingredient in the tissue culture medium of various epithelial cell types (including cells of the epidermis, mammary gland and bronchus (46, 47)) was motivated by the observation that loss of such cells in culture was accompanied by an overgrowth of fibroblast-like cells (46). These results collectively indicate a role for PKA in the differentiation of TICs by enforcing residence in the epithelial state and preventing or reversing the EMT program. Although PKA can act via a large number of substrates, we identified PHF2 as an important downstream effector of PKA that mediates the induction of epithelial characteristics through epigenetic reprogramming to a chromatin state that is more favorable for residence in the epithelial state. We find that activating this histone demethylase enables PKA to induce the transcription of genes that play a role in the entrance into and maintenance of residence in the epithelial state. The EMT program is known to represent one defined route for the generation of both normal and neoplastic epithelial stem cells (6, 7, 48). The observations that PKA-induced activation of PHF2 can either reverse or curtail this program present an opportunity to exploit such a mechanism for therapeutic gain. Indeed the differentiation of TICs through the induction of an MET is an attractive proposition - one that could be pursued through the induced increase of intracellular cAMP levels, activation of PKA, or activation of PHF2. Nonetheless, it is likely that many such approaches will result in widespread side-effect toxicities owing to the multitude of signaling pathways that are activated downstream of cAMP increase (11). Specific activation of the PHF2 histone-modifier enzyme may serve as a means of derepressing genes that are essential for the differentiated epithelial state without eliciting many of the toxicities of induced cAMP increases. Along the same lines, identification of a histone methyltransferase that counteracts PHF2 function may also provide an attractive target for therapeutic inhibition, a strategy that has proven successful in the case of DOT1L inhibition against MLL-driven leukemias (49). The role of the G9a histone methyltransferase in establishing the H3K9me2 mark for repression of the CDH1 promoter in breast cancer cells has been reported previously (50). This study provides mechanistic insight into the benefits of targeting such an enzyme in epithelial tumors, preventing the constituent cells from undergoing an EMT and thereby acquiring aggressive characteristics, including increases in the numbers of TICs. The pathways explored in this study provide novel insight into the functions of PKA in the induction of an MET and the differentiation of the more aggressive TICs within a tumor. This study reveals a new direction for targeting the TIC population through epigenetic rewiring that ultimately results in their differentiation and increased susceptibility to conventional chemotherapeutic drugs. Materials and Methods Cell Culture and treatments HMLE, NAMEC8 and all derived cell lines were grown in MEGM medium (Lonza, USA), MCF10A cells were grown in DMEM/F12 containing 5% horse serum (Sigma USA; H0146), EGF 20ng/mL, Hydrocortisone 0.5mg/mL, Cholera Toxin 100ng/mL and insulin 10µg/mL. EF021 and H596 cells were grown in RPMI containing 10% fetal bovine serum. MCF7Ras cells were grown in DME containing 10% fetal bovine serum. Hs578T were grown in DME containing 10% fetal bovine serum and 10µg/mL insulin. Cells were treated with either 100ng/mL of Cholera Toxin (Calbiochem USA; 227036), which was replenished every two days, or 1uM Forskolin (Tocris Biosciences USA; 1099), which was replenished daily over a period of 14 days. Cells were split to a ratio of 1:6 every 3 days during the treatments. Screening For the CDH1 reporter screen 500 N8 cells bearing wtCDH1 promoter luciferase were seeded into 384-well plates in a volume of 40ul. 24h later, 100nl of each compound (200uM stock) were added using a CyBio liquid handler, resulting in a final screen concentration of 0.5uM. Four days later, the plates were read for either firefly luciferase activity (Pierce, Cat# 16177) or CellTiter Glo (Promega USA, Cat# G7572). The Enzo compound library (Plate A and Plate B; 451 compounds, including repeats) was obtained from the Koch Institute Screening Facility at MIT. Firefly luciferase and CellTiter Glo assays were performed in triplicates. The vulnerabilities of the reverted cells were assessed by screening against the Selleck anti-cancer compound library (400 compounds) and the Enzo kinase library (80 compounds) at the Koch Institute Screening Facility at MIT. 1000 N8 or N8-CTx cells were seeded in 384-well plates in a volume of 50ul. 24h later, 50nl of each compound was added to assay a 5-point dose response. Three days later, the plates were read for CellTiter Glo, assays were performed in duplicate. Flow Cytometry Cells were prepared according to standard protocols and suspended in 2% IFS/PBS. DAPI (Life Technologies USA; D1306) was used to exclude dead cells. Cells were sorted on BD FACSAria SORP and analysed on BD LSRII, using BD FACSDiva Software (BD Biosciences, USA). Antibodies used were anti- CD44-PE-Cy7 (Biolegend USA #103029), anti- CD24-PE (BD Biosciences USA #555428), anti-CD45-Pacific Blue (Biolegend USA #103125), anti-CD31-Pacific Blue (Biolegend USA #102421). Mammosphere/Tumorsphere Culture Mammosphere/tumorsphere culture was performed as previously described (51). 1000 cells were seeded per well of a 96-well Corning Ultra-Low attachment plate (Corning USA; CLS3474) in replicates of 10, sphere numbers were counted between days 8 to 12. Migration and Invasion Assays Twenty-five thousand cells were seeded into 24-well cell culture inserts with 8 µm pores (BD Falcon, USA). After 12 to 24 hours, the cells on the upper surface of the filters were removed with a cotton swab. For visualization, cells on lower filter surfaces were fixed and stained with a Diff-Quick staining kit (Dade Behring/Siemens, Germany). Three to five fields per filter were counted. Data are presented as migrated cells per field. RNA preparation and qRT-PCR analysis Total RNA was isolated using the RNeasy Plus Mini kit (Qiagen USA; 74136) and reverse transcription was performed with the High Capacity RNA-to-cDNA Kit (Life Technologies USA; 4387406), both according to the manufacturer’s protocols. A cDNA sample prepared from 1 µg total RNA was used for quantitative RT-PCR. The PCR reactions were performed using the Fast SYBR Green Master Mix (Life Technologies; 4385612), data collection and data analysis were performed on the ABI7900 machine (Applied Biosystems, USA) using the SDS2.0 and RQ manager software. The thermal cycling parameters for the PCR were as follows: 95 °C for 5 min, followed by 45 cycles of 95 °C for 10 sec, 49 °C for 7 sec, and 72 °C for 25 sec. The relative mRNA quantity was normalized against the relative quantity of HPRT1 mRNA in the same sample. Name Primer sequence in 5’-3’ orientation E-CADHERIN F – TTGCACCGGTCGACAAAGGAC R - TGGATTCCAGAAACGGAGGCC FIBRONECTIN F - GAGAATGGACCTGCAAGCCCA R - AGTGCAAGTGATGCGTCCGC VIMENTIN F – ACCCGCACCAACGAGAAGGT R - ATTCTGCTGCTCCAGGAAGCG SNAI1 F - CTGGGTGCCCTCAAGATGCA R - CCGGACATGGCCTTGTAGCA SNAI2 F – TACCGCTGCTCCATTCCACG R - CATGGGGGTCTGAAAGCTTGG TWIST1 F - TGCGGAAGATCATCCCCACG R - GCTGCAGCTTGCCATCTTGGA ZEB1 F – TGCACTGAGTGTGGAAAAGC R - TGGTGATGCTGAAAGAGACG HPRTI F- CTCCGTTATGGCGACCC R- CACCCT TTCCAAATCCTCAG PHF2 F- CCCATGGGTTTTCTCACAGT R- GGCTCCCCTACGACGTTA PRKACA F- GTGTTCTGAGCGGGACTTTC R- GCCCTGAGAACAGGACTGAG PRKACB F- AAAGTCTTCTTTGGCTTTGGC R- CCTTCCCTGACCCCTTCTT Immunofluorescence (cultured cells) Cells were cultured on dishes containing coverslips for 2–3 days following which coverslips were washed in cold PBS, fixed in 4% paraformaldehyde for 10 minutes at 4°C and permeabilized in 0.2% tritonX in PBS for 2 minutes. Cells were then washed in PBS, blocked for 1 hour at room temperature in PBS containing 3% normal horse serum (Vector Labs USA; S-2000). Fixed cells were then incubated with the primary antibody in PBS containing 1%BSA solution overnight at 4°C. Cells were washed in PBS three times and secondary antibody was added in PBS containing 1%BSA solution for 1–2 hours at room temperature in the dark. Cells were washed three times in PBS and were incubated for 2 minutes in DAPI solution, following which they were washed in PBS and mounted with a drop of Prolong Gold antifade reagent (Life Technologies USA; P36961) and coverslipped. Slides were imaged on a PerkinElmer Ultraview Spinning Disk Confocal and analyzed using Volocity software. Immunofluorescence (Tissue microarrays) Slides were rehydrated by incubating in Histoclear solution twice for 5 minutes each, followed by incubation in 100% ethanol twice for 5 minutes each, in 95% ethanol twice for 5 minutes each, 70% ethanol twice for five minutes each, once in 35% ethanol for five minutes and in water for 5 minutes. Pressure cooker mediated heat induced epitope retrieval was carried out in 250mL of unmasking buffer containing sodium citrate at pH 6. Following retrieval, slides were blocked for 30 minutes in PBS containing 3% normal horse serum following which they were incubated with primary antibody in blocking solution overnight at 4°C. Slides were washed twice with PBS and incubated with secondary antibody at room temperature for 1 hour in the dark. Following two PBS washes, 20ul of mounting medium was added, coverslipped and stored in the dark for 24 hours before imaging. Target Company Catalog # host Dilution E-Cadherin BD Biosciences 610182 mouse 1:500 E-Cadherin Cell Signaling Technologies 3195 rabbit 1:200 Fibronectin BD Biosciences 610078 mouse 1:200 PKAC-α Cell Signaling Technologies 4782 rabbit 1:100 PKAC- α Abcam ab124390 mouse 10 µg/ml PKAC-β Santa Cruz Biotechnology sc-904 rabbit 1:100 Vimentin Cell Signaling Technologies 3932 rabbit 1:100 PHF2 Abcam ab154983 mouse 1:50 Proliferation Assays To measure rate of proliferation, 1000 cells were seeded onto 96-well plate in quadruplicate. Proliferation was measured using CyQuant (Life Technologies USA, C7026) according to manufacturer’s protocols. Protein Extraction and Western Blotting To obtain protein extracts, cells were washed with chilled PBS and scraped from culture dishes in aqueous lysis buffer (50mM Tris pH7.5, 150mM NaCl, 10mM EDTA pH8.0, 0.2% Sodium Azide, 50mM NaF, 0.5% NP40) containing cOmplete mini protease inhibitor cocktail (Roche, USA 04693159001) and stored at −80°C. Following thawing, they were centrifuged at top speed on a benchtop centrifuge at 4°C for 20 minutes and the supernatant was assayed for protein concentration with Bradford Reagent (Bio-Rad; 500-0006). 30ug of total protein were separated by SDS-PAGE on NuPage gels (Invitrogen, USA) and transferred to Hybond-P PVDF membrane (GE Healthcare, USA). Membranes were probed with specific primary antibodies and antibody-protein complex detected by HRP-conjugated secondary antibodies and SuperSignal West Dura Extended Duration Substrate (Life Technologies USA; 34075). Target Company Catalog # host Dilution E-Cadherin Cell Signaling Technology 3195 rabbit 1:1000 Fibronectin BD Biosciences 610078 mouse 1:1000 PKAC-α Cell Signaling Technology 4782 rabbit 1:1000 PKAC-β Santa Cruz Biotechnology sc-904 rabbit 1:1000 Zeb1 Cell Signaling Technology 3396 rabbit 1:1000 Snail1 Cell Signaling Technology 3879 rabbit 1:1000 Vimentin Cell Signaling Technology 3932 rabbit 1:1000 PHF2 Abcam ab154983 mouse 1:500 P-CREB1 Cell Signaling Technology 9198 rabbit 1:1000 CREB1 Cell Signaling Technology 9197 rabbit 1:1000 Gli1 Santa Cruz Biotechnology sc-20687 rabbit 1:1000 Gli2 Santa Cruz Biotechnology sc-271786 mouse 1:1000 Gli3 Santa Cruz Biotechnology sc-6154 goat 1:1000 p-PKA substrate Cell Signaling Technology 9621 rabbit 1:1000 PHF2 Cell Signaling Technology 3497 rabbit 1:1000 Cox IV Cell Signaling Technology 4850 rabbit 1:1000 Immunoprecipitation N8 cells in 15cm dishes were scraped and harvested in 0.5mL ice-cold IP buffer (Cell Signaling Technology, USA) containing protease and phosphatase inhibitors (Roche, Germany) following which they were sonicated with three pulses on ice. Lysates were spun down at 14,000g for 10 minutes and supernatants collected for IP. Anti-PHF2 antibody (Catalog# 3497; Cell Signaling Technology, USA) was added at 1:25 and incubated overnight at 4°C rotating. The following morning, 40ul of magnetic protein A/G beads (Catalog #26162; Thermo Scientific, USA) were added, incubated for 30 minutes on a rotator following which beads were washed 5 times using a magnetic separator with cold IP buffer. Beads were then boiled in LDS buffer and run on an SDS-PAGE gel followed by immunoblotting. RNAi-mediated knockdown To generate shRNA expressing plasmids, double-stranded oligonucleotides (oligos) encoding the desired shRNA were cloned into a Tet-pLKO-puro lentiviral vector (Addgene, plasmid 21915). In the absence of doxycycline, shRNA expression is repressed by constitutively expressed TetR protein. Upon the addition of doxycycline to the growth media, shRNA expression is triggered resulting in target gene knock-down. The cloning vector has a 1.9 kb stuffer that is released by digestion with AgeI and EcoRI. shRNA oligos are cloned into the AgeI and EcoRI sites in place of the stuffer. PKA hairpins Name Target Sequence Forward Oligo - full sequence 5’ flank sequence: CCGG Loop sequence: CTCGAG 3’ flank sequence: TTTTTG Reverse Oligo - full sequence 5’ flank sequence: AATTCAAAAA Loop sequence: CTCGAG 3’ flank sequence: (none) A1 CCCTTCATACCAA AGTTTAAA 5’- CCGGCCCTTCATACCAAAGTTTA AACTCGAGTTTAAACTTTGGTAT GAAGGGTTTTTG-3’ 5’- AATTCAAAAACCCTTCATACCAA AGTTTAAACTCGAGTTTAAACTT TGGTATGAAGGG-3’ B1 GACCAACCAATTC AGATTTAT 5’- CCGGGACCAACCAATTCAGATTT ATCTCGAGATAAATCTGAATTGG TTGGTCTTTTTG-3’ 5’- AATTCAAAAAGACCAACCAATTC AGATTTATCTCGAGATAAATCTG AATTGGTTGGTC-3’ B3 GTCTCAATAAGGC AATATATT 5’- CCGGGTCTCAATAAGGCAATATA TTCTCGAGAATATATTGCCTTATT GAGACTTTTTG-3’ 5’- AATTCAAAAAGTCTCAATAAGGC AATATATTCTCGAGAATATATTG CCTTATTGAGAC-3’ B4 AGACCAACCAATT CAGATTTA 5’- CCGGAGACCAACCAATTCAGATT TACTCGAGTAAATCTGAATTGGT TGGTCTTTTTTG-3’ 5’- AATTCAAAAAAGACCAACCAATT CAGATTTACTCGAGTAAATCTGA ATTGGTTGGTCT-3’ B10 GTCATGTAAATGC TGATATTG 5’- CCGGGTCATGTAAATGCTGATAT 5’- AATTCAAAAAGTCATGTAAATGC TGCTCGAGCAATATCAGCATTTA CATGACTTTTTG-3’ TGATATTGCTCGAGCAATATCAG CATTTACATGAC-3’ B12 GTTTAGAGGCTCT GGAGATAC 5’- CCGGGTTTAGAGGCTCTGGAGAT ACCTCGAGGTATCTCCAGAGCCT CTAAACTTTTTG-3’ 5’- AATTCAAAAAGTTTAGAGGCTCT GGAGATACCTCGAGGTATCTCCA GAGCCTCTAAAC-3’ PHF2 hairpins Name Target Sequence Forward Oligo Reverse Oligo P1 GGAGCCACCTGAC ATTGTAAA 5’- CCGGGGAGCCACCTGACATTGTA AACTCGAGTTTACAATGTCAGGT GGCTCCTTTTTG-3’ 5’-AATTCAAAAAGGAGCCACCTGAC ATTGTAAACTCGAGTTTACAATGT CAGGTGGCTCC-3’ P3 TTGCTGACCAGGT CGACAAAT 5’- CCGGTTGCTGACCAGGTCGACAA ATCTCGAGATTTGTCGACCTGGTC AGCAATTTTTG-3’ 5’-AATTCAAAAATTGCTGACCAGGT CGACAAATCTCGAGATTTGTCGA CCTGGTCAGCAA-3’ CREB1 hairpins Name Target Sequence CREB1 ACGGTGCCAACTCCAATTTAC CREB3 ACAGCACCCACTAGCACTATT Animal Studies Research involving animals complied with protocols approved by the MIT Committee on Animal Care. For tumor studies, cells suspended in 15µl 30% Matrigel(GFR)/PBS mix (BD Biosciences; 356230) were injected into the inguinal mammary gland fat pads of age-matched female NOD/SCID mice (Jackson Laboratory). Mice were sacrificed after 10 weeks or when tumors reached a diameter >1 cm. Lung surface metastases were counted with a fluorescent microscope. Chromatin Immunoprecipitation followed by Sequencing ChIP for PHF2, H3K9me2 and H3K9me3 was carried out using the SimpleChIP® Plus Enzymatic Chromatin IP Kit (Catalog# 9005; Cell Signaling Technology, USA) and the protocols within. The PHF2 rabbit monoclonal antibody (Catalog#3497; Cell Signaling Technology, USA) was used at 1:25 per IP, the H3K9me2 mouse monoclonal (Catalog# ab1220; Abcam, USA) and the H3K9me3 rabbit polyclonal antibodies (Catalog#ab8898) were used at 1:50 (10ug) per IP. The ChIP DNA was used to prepare libraries for sequencing, which was carried out in the Genome Technology Core at the Whitehead Institute. Library Preparation for Sequencing To prepare libraries for RNA-Seq, the TruSeq stranded mRNA protocol was followed to prep the libraries as described in the kit (Catalog # RS-122–2101, Illumina, USA) manual. To prepare libraries for the ChIP-Seq, the TruSeq ChIP protocol was followed as described in the kit (Catalog # IP-202–1012, Illumina, USA) manual. Deep Sequencing and data analysis Sequencing: Libraries were pooled together and sequenced on the HiSeq 2500 sequencer using the Standard sequencing protocols. Images analysis and base calling was done using the Standard Illumina pipeline, and then demultiplexed into fastq files. RNASeq paired-end reads from Illumina 1.5 encoding were aligned using TopHat (v 2.0.13) (52) to the human genome (GRCh37) with Ensembl annotation (GRCh37.75) in gtf format. Differential expression was assayed using HTSeq count (53) and DESeq (54). ChIPSeq data were aligned to the human genome (GRCh37) using Bowtie2 (v 2.2.5) (55), base encoding as above, and peaks were called using MACS2 (v 2.1.0.20150420) (56) with --nomodel option and fragment length was determined by strand cross-correlation (using phantompeakqualtools https://code.google.com/p/phantompeakqualtools/). Differential binding was determined using MACS’ bdgdiff tool. Peaks were annotated using Cis-regulatory Element Annotation System (CEAS) (57), and ChIPSeq data profiles were viewed in ngsplot (58). Overlap between peaks, and with expression data, were determined using bedtools (59). ChIPSeq data profiles were viewed in ngsplot (58) and IGV (60). RNA-seq and ChIP-seq data have been submitted to GEO and are awaiting a GSE ID. LC/MS-based metabolite profiling LC/MS analyses were conducted on a QExactive benchtop orbitrap mass spectrometer equipped with an Ion Max source and a HESI II probe, which was coupled to a Dionex UltiMate 3000 UPLC system (Thermo Fisher Scientific, San Jose, CA). External mass calibration was performed using the standard calibration mixture every 7 days. Polar metabolites were extracted using 1 ml of ice cold 80% methanol with 10 ng/ml phenylalanine-d8 or phenylalanine-13C9-15N as an internal standard. After 10 min vortex and centrifugation for 10 min at 10,000g, both at 4°C, samples were dried in a centrifugal evaporator. Dried samples were stored at −80°C and then resuspended in 100 µL water; 2.5 µl of each sample was injected onto a ZIC-pHILIC 2.1 × 150 mm (5 µm particle size) column (EMD Millipore). Buffer A was 20 mM ammonium carbonate, 0.1% ammonium hydroxide; buffer B was acetonitrile. The chromatographic gradient was run at a flow rate of 0.150 ml/min as follows: 0–20 min.: linear gradient from 80% to 20% B; 20–20.5 min.: linear gradient from 20% to 80% B; 20.5–28 min.: hold at 80% B. The column oven was held at 25°C. The mass spectrometer was operated with the spray voltage set to 3.0 kV, the heated capillary held at 275°C, and the HESI probe held at 350°C; the sheath gas flow was set to 40 units, the auxiliary gas flow was set to 15 units, and the sweep gas flow was set to 1 unit. To measure cAMP, a positive targeted SIM (tSIM) scan was performed at a resolution of 70,000, an AGC target of 1e5, and the maximum injection time at 250 msec. The tSIM window was set to a width of 1.0 m/z and centered at 330.05980 m/z, corresponding to the [M+H] ion of cAMP. To monitor other endogenous polar metabolites and the internal standard, the tSIM scans were interspersed with positive and negative mode scans in the range of 70–1000 m/z, with the resolution set to 70,000, the AGC target at 10 , and the maximum injection time at 80 msec. Relative quantitation of polar metabolites was performed with XCalibur QuanBrowser 2.2 (Thermo Fisher Scientific) using a 5 ppm mass tolerance and referencing an in-house library of chemical standards. Statistical Analysis Data are presented as mean ± SD. A Student’s t test (two-tailed) was used to compare two groups (p < 0.05 was considered significant) unless otherwise indicated. Supplementary Material Supplementary Info Table S1 Table S2 Table S3 Table S4 We thank the Keck Microscopy Facility, Metabolite Profiling Core Facility, Genome Technology Core, Bioinformatics and Research Computing Core and the Proteomics Core Facility at the Whitehead Institute and the Koch Institute Swanson Biotechnology Center (SBC), specifically the Histology Facility and the High Throughput Screening Facility. We are grateful for John Benson (SBC) for providing the compound library, Dr. Daniel Bachovchin (Broad Institute) for assistance with automation and Mr. Tom DiCesare for assistance with scientific illustration. We would like to thank Dr. Anushka Dongre, Dr. Sonia Iyer, and Ms. Julia Fröse for technical assistance, all members of the Weinberg Lab for helpful discussions and Dr. Arthur Lambert, Dr. Keerthana Krishnan and Dr. Satyaki Rajavasireddy for critical reading of the manuscript. D.R.P is supported by a C. J. Martin Overseas Biomedical Fellowship from the National Health and Medical Research Council of Australia (NHMRC APP1071853). W.L.T is supported by the National Research Foundation, Singapore (NRF-NRFF2015-04) and National Medical Research Council, Singapore (NMRC/TCR/007-NCC/2013). This research was supported by the Ludwig Center for Molecular Oncology at MIT (R.A.W), Breast Cancer Research Foundation (R.A.W), Samuel Waxman Cancer Research Foundation (R.A.W) and the National Institutes of Health R01-CA078461 (R.A.W). R.A.W. is an American Cancer Society and D. K. Ludwig Foundation Cancer Research Professor. R.A.W is a shareholder in and advisor to Verastem Inc. Whitehead Institute has filed a patent application on the subject matter of this manuscript. RNA-Seq and ChIP-Seq data from this study have been deposited at GEO under accession number GSE74883. Figure 1 Induction of a mesenchymal-to-epithelial transition (MET) upon treatment of N8 cells with cholera toxin (CTx) or forskolin (Fsk) Mesenchymal N8 cells acquire an epithelial morphology as adjudged by their morphology (A), loss of a stem-like CD44hi/CD24lo profile to assume a predominantly CD44lo/CD24hi profile (B) and expression of E-cadherin at cell junctions and loss of vimentin (C). Reverted N2-CTx and N3-Fsk cells lose their ability to form (D, E) mammospheres (mean ± SD, p<0.05, n=4), (F) migrate (mean ± SD, p<0.05, n=4) and (G) invade in transwell assays (mean ± SD, p<0.05, n=4) and acquire increased sensitivity to treatment with (H) doxorubicin and (I) paclitaxel (mean ± SD, p<0.05, n=4). (J) Heatmap of mRNA-Seq data demonstrating similarity in gene expression between HMLE, N8 and N8-CTx cells. Data in (E), (F) and (G) were analysed by student t-test, (H) and (I) were analysed by Bonferroni correction. All scale bars - 25µm. Figure 2 cAMP increases activate protein kinase A (PKA), which is both necessary and sufficient for the induction of an MET in N8 cells (A) Mass-spectrometry measurement of cAMP levels in N8 cells that have been treated with CTx or Fsk alone and in combination with adenylate cyclase inhibitor SQ (mean ± SD, p<0.05, n=3). (B) Treatment of N8 cells with either 8-CPT-2me-cAMP or 8-Br-cAMP to identify downstream pathways that are responsible for induction of an MET. Knockdown of either PRKACA or PRKACB prevents the ability of CTx to induce an MET in N8 cells as observed by changes in (C) morphology, (D) immunofluorescence for E-cadherin and vimentin and (E) CD44/CD24 status. (F) Morphological changes of N8 cells undergoing an MET upon ectopic expression of an active mutant of PKA (caPKA). Data in (A) were analysed using the Student t-test. All scale bars - 25µm. Figure 3 The PKA substrate PHF2, but not CREB1, is necessary for the MET-inducing properties of CTx Activation state of CREB1 as measured by levels of p-CREB1 across HMLE and N8 cells that have been treated with CTx or Fsk (A). Loss of CREB1 through shRNA-mediated knockdown induces a partial MET and permits CTx-mediated complete MET as shown by changes in morphology and immunofluorescence (B). shRNA-mediated knockdown of PHF2 abrogates the ability of CTx to induce an MET preventing changes in (C) morphology and immunofluorescence-based detection of E-cadherin and fibronectin expression as well as (D) blocking a shift from the CD44hi/CD24lo state to the CD44lo/CD24hi state. Expression of a PHF2 phosphomimetic where the C-terminal serines were modified to aspartate (E) accelerated the MET transition by 5 days as observed by changes in immunofluorescence (F) and quantitative EMT marker analysis by qPCR (G). Effects of shRNA-mediated knockdown of PHF2 on the ability of HMLE cells to undergo an EMT upon ectopic expression of Zeb1 (H, I) (qPCR data - mean ± SD, p<0.05, n=3). Immunoprecipitation of PHF2 followed by immunoblotting with a phospho-PKA substrate antibody showing direct phosphorylation of PHF2 by PKA 24hrs after treatment of N8 cells with CTx (J). (I) was analysed using the Student t-test. All scale bars - 25µm. Figure 4 Activation of PHF2 leads the epigenetic reprogramming of mesenchymal cells Genome-wide occupancy of H3K9me2, H3K9me3 and PHF2 marks shows the inverse correlation between the presence of the histone marks and the demethylase (A), which interacts mainly with the promoter and the first intronic region of genes (B). Circos plots of representative chromosomes 5 and 8 show widespread changes in the H3K9me2 and H3K9me3 profiles (C). Figure 5 PKA-induced MET is sufficient to deplete the tumor-initiating ability of N8-Ras cells in vivo (A) Table outlining differences in tumor-initiating ability of HMLE-Ras, N8-Ras and N8-CTx-Ras cells upon limiting dilution transplantation into NOD/SCID mice. Tumors that arose from transplantation of 2×106 cells were or similar size (B) with only the N8-Ras cells bring capable of forming micrometastases (C) (Each dot represents one mouse; data analysed using Student t-test; p<0.05, n=10). (D) Panel showing experimental outline to test the tumor-initiating ability of N8-Ras cells upon transient in vivo expression of PKA showing a (E) 20-fold decrease in tumor-initiating ability upon secondary transplantation with (F) no significant differences in the tumor volume. (Each dot represents one mouse; data analysed using Student t-test; p<0.05, n=10). Supplementary Figures 1–11, Supplementary Tables 1–4 REFERENCES AND NOTES 1 Gupta PB Identification of selective inhibitors of cancer stem cells by high-throughput screening Cell 2009 8 21 138 645 19682730 2 Chen J A restricted cell population propagates glioblastoma growth after chemotherapy Nature 2012 8 23 488 522 22854781 3 Kreso A Dick JE Evolution of the cancer stem cell model Cell stem cell 2014 3 6 14 275 4 Pattabiraman DR Weinberg RA Tackling the cancer stem cells - what challenges do they pose? Nature reviews. Drug discovery 2014 7 1 13 497 5 De Craene B Berx G Regulatory networks defining EMT during cancer initiation and progression Nature reviews. Cancer 2013 2 13 97 23344542 6 Mani SA The epithelial-mesenchymal transition generates cells with properties of stem cells Cell 2008 5 16 133 704 18485877 7 Morel AP Generation of breast cancer stem cells through epithelial-mesenchymal transition PloS one 2008 3 e2888 18682804 8 Chen C Evidence for epithelial-mesenchymal transition in cancer stem cells of head and neck squamous cell carcinoma PloS one 2011 6 e16466 21304586 9 Ahmed N Abubaker K Findlay J Quinn M Epithelial mesenchymal transition and cancer stem cell-like phenotypes facilitate chemoresistance in recurrent ovarian cancer Current cancer drug targets 2010 5 10 268 20370691 10 Beavo JA Brunton LL Cyclic nucleotide research -- still expanding after half a century Nat Rev Mol Cell Biol 2002 9 3 710 12209131 11 Tasken K Aandahl EM Localized effects of cAMP mediated by distinct routes of protein kinase A Physiological reviews 2004 1 84 137 14715913 12 Taylor SS Ilouz R Zhang P Kornev AP Assembly of allosteric macromolecular switches: lessons from PKA Nature reviews. Molecular cell biology 2012 10 13 646 22992589 13 MacPherson MR Phosphorylation of serine 11 and serine 92 as new positive regulators of human Snail1 function: potential involvement of casein kinase-2 and the cAMP-activated kinase protein kinase A Molecular biology of the cell 2010 1 15 21 244 19923321 14 Shaikh D cAMP-dependent protein kinase is essential for hypoxia-mediated epithelial-mesenchymal transition, migration, and invasion in lung cancer cells Cellular signalling 2012 12 24 2396 22954688 15 Nadella KS Targeted deletion of Prkar1a reveals a role for protein kinase A in mesenchymal-to-epithelial transition Cancer research 2008 4 15 68 2671 18413734 16 Amieux PS Increased basal cAMP-dependent protein kinase activity inhibits the formation of mesoderm-derived structures in the developing mouse embryo The Journal of biological chemistry 2002 7 26 277 27294 12004056 17 Iglesias-Bartolome R Inactivation of a Galpha(s)-PKA tumour suppressor pathway in skin stem cells initiates basal-cell carcinogenesis Nature cell biology 2015 6 17 793 25961504 18 Li X Intrinsic resistance of tumorigenic breast cancer cells to chemotherapy Journal of the National Cancer Institute 2008 5 7 100 672 18445819 19 Elenbaas B Human breast cancer cells generated by oncogenic transformation of primary mammary epithelial cells Genes & development 2001 1 1 15 50 11156605 20 Al-Hajj M Wicha MS Benito-Hernandez A Morrison SJ Clarke MF Prospective identification of tumorigenic breast cancer cells Proceedings of the National Academy of Sciences of the United States of America 2003 4 1 100 3983 12629218 21 Tam WL Protein kinase C alpha is a central signaling node and therapeutic target for breast cancer stem cells Cancer cell 2013 9 9 24 347 24029232 22 Onder TT Loss of E-cadherin promotes metastasis via multiple downstream transcriptional pathways Cancer research 2008 5 15 68 3645 18483246 23 Anastassiou D Human cancer cells express Slug-based epithelial-mesenchymal transition gene expression signature obtained in vivo BMC cancer 2011 11 529 22208948 24 Charafe-Jauffret E Gene expression profiling of breast cell lines identifies potential new basal markers Oncogene 2006 4 6 25 2273 16288205 25 de Rooij J Epac is a Rap1 guanine-nucleotide-exchange factor directly activated by cyclic AMP Nature 1998 12 3 396 474 9853756 26 Broillet MC Firestein S Cyclic nucleotide-gated channels. Molecular mechanisms of activation Annals of the New York Academy of Sciences 1999 4 30 868 730 10414360 27 Meyer RB Jr Miller JP Analogs of cyclic AMP and cyclic GMP: general methods of synthesis and the relationship of structure to enzymic activity Life sciences 1974 3 16 14 1019 4362776 28 Enserink JM A novel Epac-specific cAMP analogue demonstrates independent regulation of Rap1 and ERK Nature cell biology 2002 11 4 901 12402047 29 Orellana SA McKnight GS Mutations in the catalytic subunit of cAMP-dependent protein kinase result in unregulated biological activity Proc Natl Acad Sci U S A 1992 5 15 89 4726 1584809 30 Soule HD Isolation and characterization of a spontaneously immortalized human breast epithelial cell line, MCF-10 Cancer research 1990 9 15 50 6075 1975513 31 Sommers CL Papageorge A Wilding G Gelmann EP Growth properties and tumorigenesis of MCF-7 cells transfected with isogenic mutants of rasH Cancer research 1990 1 1 50 67 2403419 32 Ellenrieder V Transforming growth factor beta1 treatment leads to an epithelial-mesenchymal transdifferentiation of pancreatic cancer cells requiring extracellular signal-regulated kinase 2 activation Cancer research 2001 5 15 61 4222 11358848 33 Guy CT Cardiff RD Muller WJ Induction of mammary tumors by expression of polyomavirus middle T oncogene: a transgenic mouse model for metastatic disease Molecular and cellular biology 1992 3 12 954 1312220 34 Linnemann JR Quantification of regenerative potential in primary human mammary epithelial cells Development 2015 6 12 35 Hollestelle A Distinct gene mutation profiles among luminal-type and basal-type breast cancer cell lines Breast cancer research and treatment 2010 5 121 53 19593635 36 Shabb JB Physiological substrates of cAMP-dependent protein kinase Chemical reviews 2001 8 101 2381 11749379 37 Gonzalez GA Montminy MR Cyclic AMP stimulates somatostatin gene transcription by phosphorylation of CREB at serine 133 Cell 1989 11 17 59 675 2573431 38 Johannessen M Delghandi MP Moens U What turns CREB on? Cellular signalling 2004 11 16 1211 15337521 39 Wu D cAMP-responsive element-binding protein regulates vascular endothelial growth factor expression: implication in human prostate cancer bone metastasis Oncogene 2007 8 2 26 5070 17310988 40 Singh R Shankar BS Sainis KB TGF-beta1-ROS-ATM-CREB signaling axis in macrophage mediated migration of human breast cancer MCF7 cells Cellular signalling 2014 7 26 1604 24705025 41 Sheng T Chi S Zhang X Xie J Regulation of Gli1 localization by the cAMP/protein kinase A signaling axis through a site near the nuclear localization signal The Journal of biological chemistry 2006 1 6 281 9 16293631 42 Baba A PKA-dependent regulation of the histone lysine demethylase complex PHF2-ARID5B Nature cell biology 2011 6 13 668 21532585 43 Hublitz P Albert M Peters AH Mechanisms of transcriptional repression by histone lysine methylation The International journal of developmental biology 2009 53 335 19412890 44 Saunders MP A novel cyclic adenosine monophosphate analog induces hypercalcemia via production of 1,25-dihydroxyvitamin D in patients with solid tumors The Journal of clinical endocrinology and metabolism 1997 12 82 4044 9398710 45 Propper DJ Phase I study of the novel cyclic AMP (cAMP) analogue 8-chloro-cAMP in patients with cancer: toxicity, hormonal, and immunological effects Clinical cancer research : an official journal of the American Association for Cancer Research 1999 7 5 1682 10430069 46 Taylor-Papadimitriou J Purkis P Fentiman IS Cholera toxin and analogues of cyclic AMP stimulate the growth of cultured human mammary epithelial cells Journal of cellular physiology 1980 3 102 317 6248570 47 Green H Cyclic AMP in relation to proliferation of the epidermal cell: a new view Cell 1978 11 15 801 83196 48 Guo W Slug and Sox9 cooperatively determine the mammary stem cell state Cell 2012 3 2 148 1015 22385965 49 Daigle SR Selective killing of mixed lineage leukemia cells by a potent small-molecule DOT1L inhibitor Cancer cell 2011 7 12 20 53 21741596 50 Dong C G9a interacts with Snail and is critical for Snail-mediated E-cadherin repression in human breast cancer The Journal of clinical investigation 2012 4 122 1469 22406531 51 Dontu G In vitro propagation and transcriptional profiling of human mammary stem/progenitor cells Genes & development 2003 5 15 17 1253 12756227 52 Kim D TopHat2: accurate alignment of transcriptomes in the presence of insertions, deletions and gene fusions Genome biology 2013 14 R36 23618408 53 Anders S Pyl PT Huber W HTSeq--a Python framework to work with high-throughput sequencing data Bioinformatics 2015 1 15 31 166 25260700 54 Anders S Huber W Differential expression analysis for sequence count data Genome biology 2010 11 R106 20979621 55 Langmead B Salzberg SL Fast gapped-read alignment with Bowtie 2 Nature methods 2012 4 9 357 22388286 56 Zhang Y Model-based analysis of ChIP-Seq (MACS) Genome biology 2008 9 R137 18798982 57 Shin H Liu T Manrai AK Liu XS CEAS: cis-regulatory element annotation system Bioinformatics 2009 10 1 25 2605 19689956 58 Shen L Shao N Liu X Nestler E ngs.plot: Quick mining and visualization of next-generation sequencing data by integrating genomic databases BMC genomics 2014 15 284 24735413 59 Quinlan AR Hall IM BEDTools: a flexible suite of utilities for comparing genomic features Bioinformatics 2010 3 15 26 841 20110278 60 Robinson JT Integrative genomics viewer Nature biotechnology 2011 1 29 24

PMC005xxxxxx/PMC5131733.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 0110674 8015 Virology Virology Virology 0042-6822 1096-0341 27816895 5131733 10.1016/j.virol.2016.10.022 NIHMS827959 Article Dengue virus induces mitochondrial elongation through impairment of Drp1-triggered mitochondrial fission Barbier Vincent Lang Diane Valois Sierra Rothman Alan L. Medin Carey L. # Institute for Immunology and Informatics, Department of Cell and Molecular Biology, University of Rhode Island, Providence, RI 02903 # Address correspondence to Carey L. Medin, cmedin.uri@gmail.com 6 11 2016 4 11 2016 1 2017 01 1 2018 500 149160 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. Mitochondria are highly dynamic organelles that undergo continuous cycles of fission and fusion to maintain essential cellular functions. An imbalance between these two processes can result in many pathophysiological outcomes. Dengue virus (DENV) interacts with cellular organelles, including mitochondria, to successfully replicate in cells. This study used live-cell imaging and found an increase in mitochondrial length and respiration during DENV infection. The level of mitochondrial fission protein, Dynamin-related protein 1 (Drp1), was decreased on mitochondria during DENV infection, as well as Drp1 phosphorylated on serine 616, which is important for mitochondrial fission. Viral proteins NS4b and NS3 were also associated with subcellular fractions of mitochondria. Induction of fission through uncoupling of mitochondria or overexpression of Drp1 wild-type and Drp1 with a phosphomimetic mutation (S616D) significantly reduced viral replication. These results demonstrate that DENV infection causes an imbalance in mitochondrial dynamics by inhibiting Drp1-triggered mitochondrial fission, which promotes viral replication. Dengue virus Drp1 mitochondria fusion fission live-cell imaging dendritic cells Introduction DENV is a mosquito-borne human pathogen of global medical importance with an estimation of 390 million infections every year [1–3]. DENV causes an acute febrile illness that, in some patients, is associated with a life-threatening plasma leakage syndrome termed dengue hemorrhagic fever (DHF) [4]. The principal target cells for DENV infection are from the myeloid-derived lineage (monocytes, macrophages and dendritic cells), hepatocytes and endothelial cells [5–15]. Several studies indicate that higher viremia levels during DENV infection correlate with the severity of disease [16–19]. DENV infection is also often associated with liver dysfunction as a result of direct viral effect on hepatocytes and dysregulated immune response against the virus [20]. Both viral and host factors seem to play a role in disease pathogenesis, but little is known about DENV-induced mechanisms that trigger modulation of organelles during viral infection [17, 21, 22]. DENV is a member of the Flaviviridae family and its positive-sense single-stranded RNA genome of 10.7 kb encodes for 10 viral proteins. This limited coding capacity makes it fully reliant on the host machinery for its viral life cycle. Like many viruses, DENV needs to modulate host cell metabolism and organelles for its own replication, interfering with different cellular organelles [23–26]. Mitochondria are essential organelles involved in cellular energy production and sensing of metabolic homeostasis. They serve as an anti-viral signaling platform for MAVS signaling [27] and contribute to activation of apoptosis and cell death [28–30]. Recently, mitochondria were shown to release their DNA into the cytosol during stress to trigger a type I interferon response [31]. These processes indicate an important role of mitochondria in initiating responses against viruses. Mitochondrial morphology depends on highly dynamic processes, fusion and fission [32]. Mitochondrial fission involves the oligomerization of Dynamin-related protein 1 (Drp1) around the outer mitochondrial membrane, constricting the mitochondrion to trigger its division [33, 34]. Localization of Drp1 is regulated by differential phosphorylation of serine residues at positions 616 and 637 (S616-P and S637-P, respectively). Phosphorylation at S637 is thought to retain Drp1 in the cytoplasm whereas phosphorylation at S616 induces Drp1 localization to the fission site on the mitochondrion [35]. The endoplasmic reticulum and actin polymerization play a role in Drp1-triggered mitochondrial fission by initiating constriction of mitochondria [36, 37]. On the other hand, mitofusin 1 (Mfn1) and Mfn2, together with optic atrophy protein 1 (Opa1) are the core components of the mitochondrial fusion machinery. Mitochondrial fission and fusion proteins are regulated by post-translational modifications, including phosphorylation as described above for Drp1 [38]. The ability of mitochondria to undergo fission and fusion, and to move in the cells is critical for their function, as defects in mitochondrial dynamics are implicated in many neurodegenerative diseases [39–41]. DENV infection has previously been reported to induce structural and functional changes in organelles in vitro, including mitochondria [42–45]. DENV protease NS2b3 was shown to partially cleave Mfn1 and Mfn2 resulting in degradation of the proteins, which attenuated interferon responses and cell death [46]. A recent report was published while this manuscript was being prepared that showed that DENV NS4b induced mitochondrial elongation by inhibition of Drp1 activation and limited activation of the interferon response in a liver cell line [47]. Although these reports signify the importance of mitochondria and mitochondrial proteins during viral infection, the underlying mechanisms that link mitochondria and its proteins to DENV replication in target cell lines and primary cells has not been well studied. Using live-cell imaging of DENV-infected cell cultures, we observed that DENV infection induces significant mitochondrial elongation. Mitochondria from DENV-infected cells displayed increased respiration, which can lead to an increase in ATP production, when compared to uninfected cells, and viral proteins such as NS4b were preferentially localized in subcellular fractions of mitochondria. Characterization of fusion and fission proteins revealed a decrease in Drp1 on mitochondria in the Huh7 cell line and in monocyte-derived dendritic cells (DCs). Drp1 decrease was correlated with a reduction of Drp1 phosphorylation on S616 in the Huh7 cell line. Conversely, induction of mitochondrial fission or overexpression of wild-type (WT) Drp1 and Drp1 S616D inhibited DENV replication. These results demonstrate that DENV infection causes an imbalance in mitochondrial dynamics by inhibiting Drp1-triggered mitochondrial fission, which promotes viral replication. Results DENV infection induces mitochondrial elongation and reduces mitochondrial motility We recently described a reporter system for live-cell imaging of DENV-infected cells based on the pNS4b5-eGFP plasmid, which expresses a GFP fusion protein that translocates from the cytoplasm to the nucleus in infected cells [48]. To examine the structure of mitochondria in cells infected with DENV, we transfected cells with the pNS4b5-eGFP reporter and subsequently stained cells with MitoTracker Deep Red, a fluorescent stain that accumulates in active mitochondria. Uninfected liver cell lines (Huh7, Fig. 1A; HepG2, Fig. S1A and Huh7.5, Fig. S1B) and an epithelial cell line (Vero, Fig. 1B) showed both fused and fragmented mitochondria. In contrast, DENV-infected cells showed mostly fused mitochondria at 48h post-infection in all cell lines. The mitochondrial network in DENV-infected cells was filamentous and interconnected. To confirm that the mitochondrial elongation was not due to transfection of the reporter, which contains the DENV protein NS4b, cells were transfected with pTagRFP-mito (Huh7.5 and Huh7 cells, Fig. S1D and S1E, respectively) and infected 24h later with DENV. At 48h post-infection, cells were stained for DENV antigen and analyzed by confocal microscopy. Similarly, we stained DENV-infected and uninfected cells with COX-IV antibody, a mitochondrial marker, and DENV antigen (HepG2 cells, Fig. S1C). As observed previously, all cell lines infected with DENV showed elongated mitochondria at 48h post-infection compared to uninfected cells. Quantitative measurement of the change in mitochondrial lengths using a macro developed in ImageJ (NIH) showed that median lengths of mitochondria increased at 24h and became statistically significant by 48h post-infection when compared to uninfected cells (Fig. 1B). Analysis of DENV-infected Huh7 cells by transmission electron microscopy, characterized by the presence of convoluted membranes (DENV, Fig. 1C and 1D) and invaginated vesicles (DENV, Fig. 1D and 1E) specifically induced by DENV as previously described [24], also showed an increase of mitochondrial lengths when compared to uninfected cells (Fig. 1C and 1D). Defects in fusion and fission have been shown to reduce motility of mitochondria [49]. Therefore, we performed time-lapse imaging in Huh7 cells transfected with pTagRFP-mito and pNS4b5-eGFP and infected with DENV for 48h to evaluate the motility of mitochondria in cells during DENV infection, (Fig. S2, Movie S1 and S2). Prior to imaging mitochondria movements in cells, DENV infected and uninfected cells were identified by localization of GFP to the nucleus or cytoplasmic GFP only, respectively (Fig. S2). We observed not only an increase in mitochondrial lengths but also a reduction of mitochondrial movements in DENV-infected cells (Movie S2) compared to uninfected cells (Movie S1), indicating that mitochondrial motility is reduced in DENV-infected cells. DENV NS4b and NS3 proteins are enriched in the mitochondrial fraction Recently, HCV proteins C and NS4b were found to be associated with mitochondria [50–52]. Therefore, to assess whether DENV proteins interact with mitochondria, we analyzed the presence of DENV proteins in isolated mitochondria by western blot. We found that NS4b and NS3 were enriched in the mitochondrial fraction of DENV-infected Huh7 cells and DCs (Fig. 2A and 2B, respectively). Immunofluorescence analysis in DENV-infected Huh7 cells revealed that NS4b did not colocalize with mitochondria, but instead NS4b staining was mostly found juxtaposed to mitochondria, in dot-like structures colocalizing with the endoplasmic reticulum (ER) marker Sec61β (Fig. 2C). ER membranes have been reported to associate with mitochondria to form mitochondria-associated membranes (MAMs). To assess the presence of MAMs in the mitochondrial fractions, we analyzed the levels of FACL4, a MAM marker, by western blot. FACL4 was detected in the mitochondrial fractions of Huh7 cells and DCs (Fig. 2A and 2B), suggesting that DENV proteins interact with mitochondria through the MAMs. These results suggest that DENV may influence mitochondrial changes through viral protein interactions at the MAM-mitochondrial interface. We noticed the presence of additional bands of higher and lower molecular weight after FACL4 blotting. Whether they are a result of post-translational modifications or cross-reaction of FACL4 antibody with other isoforms of FACL is not clear. We also observed a decrease of the highest density band of FACL4 in the mitochondrial fraction of DENV-infected cells, particularly Huh7, whereas it is increased in the total cell lysate and cytosolic fraction, suggesting a decrease of ER-mitochondria contacts in DENV-infected cells. DENV-infected cells display increased mitochondrial respiration To assess if changes in mitochondrial morphology during DENV infection affect the function of mitochondria, the oxygen consumption rate (OCR), a measure of cellular respiration, was measured in uninfected and DENV-infected cells (Fig. 3A). Sequential addition of specific mitochondrial inhibitors to the cell culture medium was used to assess the function of each component of the respiratory chain (Fig. 3B). Basal levels of OCR in DENV-infected cells were increased compared to uninfected cells. In turn, mitochondrial respiration was also increased in DENV-infected cells. The addition of oligomycin, which inhibits F0/F1 ATPase (complex V), demonstrated that the potential for ATP production was significantly higher in DENV-infected cells, as compared with uninfected cells, but proton leak was unchanged. DENV-infected cells displayed increased respiration compared to uninfected cells. There was no significant difference in coupling efficiency between DENV-infected and uninfected cells. Maximal respiration was comparable between uninfected and DENV-infected cells, and similar to basal respiration in DENV-infected cells, suggesting that DENV increased mitochondrial respiration to their maximum. Accordingly, spare respiratory capacity was decreased in DENV-infected cells. These results indicate that elongated mitochondria in DENV-infected cells are functional and display increased respiration similar to maximal, resulting in more energy production when compared to uninfected cells. DENV infection decreases the level of Drp1 and Mfn2 on mitochondria Mitochondrial fusion requires fusion of the inner membrane of mitochondria by Opa1 and the outer membrane by Mfn1 and Mfn2 [53, 54]. Similarly, overexpression of Opa1 or mitofusins promote mitochondrial elongation. To analyze whether mitochondrial elongation seen during DENV infection was due to an increase in fusion events, mitochondrial and cytosolic fractions were isolated from DENV-infected and uninfected cells and fusion protein levels were assessed by western blot. Mfn1 and Opa1 protein levels in mitochondrial fractions from Huh7 cells and DCs were similar between uninfected and DENV-infected cells (Fig. 4A and 4B). By contrast, the level of Mfn2 protein on mitochondria was decreased in DENV-infected cells compared to uninfected cells. Immunofluorescence analysis also showed a decrease in Mfn2 fluorescence of DENV-infected cells, and no change in Mfn1 (Fig. S3A and S3B). These results indicate that elongated mitochondria are not due to an increase in fusion during DENV infection. Mitochondrial elongation due to a defect in mitochondrial fission and decreased expression of Drp1 has previously been described [55, 56]. To investigate whether DENV-induced mitochondrial elongation was a result of impaired mitochondrial fission, we analyzed Drp1 and Drp1 S616-P protein levels from uninfected and DENV-infected Huh7 cells and DCs by western blot. The level of Drp1 in total cell lysates from DENV-infected Huh7 cells was decreased when compared to uninfected cells by western blot and confirmed in cells by immunofluorescence analysis (Fig. 4A and S4A); there was no change in Drp1 mRNA levels (Fig. S5A). We observed at higher magnification that the remaining Drp1 in DENV-infected Huh7 cells was less associated with mitochondria compared to uninfected cells (Fig. S5B). In contrast, Drp1 was not decreased in total cell lysates in DENV-infected DCs when compared to uninfected cell lysates by western blot (Fig. 4B). Mitochondrial fractions from both DENV-infected Huh7 cells and DCs showed a significant decrease in Drp1 protein levels when compared to uninfected cells (Fig. 4A and 4B). Accordingly, the level of Drp1 S616-P was also markedly decreased in total cell lysate and in mitochondrial fraction of Huh7 cells (Fig. 4A). Immunofluorescence analysis of DENV-infected Huh7 cells showed a significant decrease in Drp1 S616-P (Fig. S4B). Drp1 S616-P also showed less colocalization with mitochondria in DENV-infected cells (Fig. S5C). In DCs, Drp1 S616-P was poorly detected by western blot (unpublished data). Overall, these results suggest that DENV infection inhibits Drp1-triggered mitochondrial fission by decreasing the levels of Drp1. Also, DENV infection reduced Drp1 S616-P protein on mitochondria in a cell type-specific manner. Induction of mitochondrial fission inhibits DENV replication Several approaches were used to investigate the effect of mitochondrial fission on DENV replication. DENV-infected cells treated with a reversible mitochondrial uncoupler carbonyl cyanide m-chlorophenylhydrazone (CCCP) [57] for 6h (from 24h to 30h post-infection) displayed short mitochondria and viral titers in culture supernatants were decreased compared to untreated cells (Fig. 5A and 5B). Removal of CCCP from the media resulted in an increase in mitochondrial lengths and viral titers back to the level in untreated DENV-infected cell cultures, demonstrating a reversible effect on mitochondrial function rather than a toxic effect on cells. The intensity of DENV antigen staining also was significantly decreased in DENV-infected cells treated with CCCP and increased back to levels comparable to DENV-infected untreated cells when CCCP was removed (Fig 5B and S6). These results suggest that mitochondrial fission reduces the production of DENV proteins and infectious viral particles. Since CCCP decouples mitochondria and impairs ATP production, the impairment in viral production may be due to lower availability of energy in the cells. Therefore, we wanted to induce fission of mitochondria by knocking down the fusion proteins, Mfn1 and Mfn2 (Fig. S7A). siRNA knockdown of Mfn2 protein, but not Mfn1 protein, decreased the expression of DENV proteins NS3 and NS5 (Fig. 5C and 5D). Viral titers were reduced by at least 40% after 24h of infection when compared to controls and were on the threshold of statistical significance (p = 0.06 and 0.17, by paired t-test) (Fig. 5E). To further examine the effect of mitochondrial fission on DENV infection, we knocked down endogenous Drp1 with siRNAs targeting the 3′ untranslated region (3′UTR) of Drp1 and transfected cells with plasmids expressing WT, phosphomimetic, or phosphodefective mutants of human Drp1 [55]. Endogenous Drp1 was observed as a double band and bands of higher molecular weight represent Drp1 WT or mutant constructs. In cells expressing Drp1 K38A there is an additional unidentified, intermediate-sized band (Fig. S7B). DENV-infected cells overexpressing WT Drp1 displayed a decrease in the levels of DENV NS3 and NS4b proteins by western blot analysis (Fig. 6A and 6B) and a reduction in viral titers (Fig. 6C) when compared to DENV-infected cells transfected with a control vector. Drp1 S616-P was reported to trigger mitochondrial fission [35] whereas S637-P results in a reduction in mitochondrial fission [58]. The overexpression of phosphomimetic Drp1 S616D caused a significant decrease in NS3 and NS4b levels (Fig 6A and 6B), as well as a decrease in virus production (Fig. 6C) when compared to the control, indicating that the ‘fission active’ form of Drp1 affects DENV protein expression and virus production. Overexpression of phosphodefective Drp1 S637A partially reduced virus titer, although the difference was not statistically significant. In contrast, the levels of DENV proteins and virus production were not affected by overexpression of phosphodefective Drp1 S616A or phosphomimetic Drp1 S637D, both of which are thought to be defective at inducing fission [55]. We further analyzed the effect of Drp1 GTPase activity on DENV infection by transfecting cells with a dominant negative form of Drp1 with defective GTPase activity (Drp1 K38A) [59]. Overexpression of Drp1 K38A increased levels of NS4b and NS3 proteins and viral titers when compared to cells expressing WT Drp1, indicating that GTPase activity of Drp1 is required for inhibition of DENV replication. Overall, our results indicate that inducing mitochondrial fission perturbs DENV production. Discussion Virus-induced mitochondrial changes The use of live-cell imaging of DENV-infected cells showed significant increase in mitochondrial elongation over time when compared to uninfected cells due to an inhibition of mitochondrial fission. Mitochondria are dynamic organelles that constantly undergo cycles of fusion and fission [32]. Many mitochondrial and cellular functions rely on mitochondria dynamics, suggesting that altering mitochondria dynamics could serve as a viral strategy to interfere with cellular signaling pathways. Several studies have reported that mitochondria dynamics can be altered during infection with pathogens such as viruses, bacteria and parasites, leading to changes in mitochondrial morphology [60–64]. Drp1 has also previously been shown to be targeted by viruses [60, 65]. Our results, together with those from Chatel-Chaix et al, show that inhibition of fission during DENV infection is due to a reduction in Drp1 protein levels on mitochondria. DENV NS4b is reported to trigger mitochondrial elongation, which was also observed with other DENV serotypes and Zika virus [47], and we also found a decrease of Drp S616-P after infection with two strains of Zika virus (unpublished data). In turn, overexpression of the ‘fission active’ form of Drp1 inhibits DENV titers and viral antigen suggesting that DENV-specific modulation of Drp1 promotes viral replication. Both liver cell line and DCs showed decrease of Drp1 (as well as Mfn2) on mitochondria during DENV infection. However, Drp1 S616-P was poorly detected in DCs compared to Huh7 cells by western blot. This result indicates that phosphorylation of S616 may be important in liver cells but not necessary in DCs. Phosphorylation state of other serines in Drp1 have been shown to regulate fission [35, 66] and these sites may be important for mitochondrial regulation in DCs. The advantage of using cellular fractionation in this study was the ability to define significant changes in the protein profile on mitochondrial fractions in both cell lines and DCs during DENV infection that, in some cases, would not have been identified using total cell lysates. Drp1 is the master regulator of mitochondrial fission [55]. The current model suggests that ER encircles mitochondria at sites of fission, and ER-associated inverted formin 2 (INF2) then stimulates actin polymerization, providing the force required for partial constriction of the mitochondria, thereby facilitating the translocation of Drp1 to these pre-constriction contact sites in the outer mitochondrial membrane [36]. The ER is also the site of DENV replication and many DENV proteins remain embedded in the ER through their transmembrane domains [67]. Whether DENV prevents Drp1 binding to receptors on mitochondria or DENV affects formation of ER-mitochondria contacts to prevent Drp1-triggered mitochondrial fission remains to be evaluated. We isolated mitochondrial fractions using magnetic beads coated with TOM22 antibody. Our results show that we not only pulled down mitochondria but also pulled down NS4b in DENV-infected mitochondrial fractions. Immunofluorescence did not reveal colocalization of mitochondria with NS4b but rather that they were in close proximity. Electron microscopy observations performed in our study indicate that DENV-induced convoluted membranes, which have previously been described as ER-derived membranes and contain non-structural proteins such as NS4b [68], are in close contact with elongated mitochondria. These results suggest that mitochondrial isolation using magnetic beads were also able to pull down mitochondrial-associated membranes. The decrease of FACL4 in mitochondrial fractions of DENV-infected cells suggests that interactions between the ER and mitochondria are reduced, which may also impair Drp1 recruitment to mitochondria to induce fission. On the other hand, we observed higher levels of FACL4 in DENV-infected cell lysates, which may be associated with changes in lipid metabolism during infection. There also remains the possibility that viral proteins in the ER membrane or MAMs may be located in regions where mitochondrial division is initiated and inhibit Drp1 binding to mitochondria-ER contact sites. Mitochondrial fusion proteins during DENV infection Our data revealed a decrease of Mfn2 protein on the mitochondria during DENV infection, whereas the level of Mfn1 protein was not significantly affected. Both proteins are anchored to the outer mitochondrial membrane and can form homotypic or heterotypic interactions that are both functional for mitochondrial fusion. Several studies have reported redundancy between the two proteins for mitochondrial fusion function [53, 69, 70]. Therefore, the decrease of Mfn2 on the mitochondria during DENV infection could be compensated by Mfn1 during fusion. One group recently reported that Mfn2, as well as Mfn1 protein, was cleaved by NS2b3 protease in A549 cells during DENV infection [46]. Cleavage could contribute to the decrease in Mfn2 levels observed in our experiments. However, the cleavage reported in A549 cells affected only a small fraction of the total Mfn2 protein [46], suggesting another mechanism is involved in the reduction of Mfn2. Mitochondrial fusion in A549 cells infected with DENV was reported to be Mfn1 mediated [46]. In contrast, our results in Huh7 cells and DCs show that mitochondrial elongation is due to inhibition of Drp1 and not due to Mfn1. The difference in findings from the previous study and our results may stem from the cell lines or cell types used for analysis. Uninfected A549 cells have been reported to display elongated mitochondria due to low expression levels of Drp1 [71, 72]. Although knockdown of Mfn2 indicates that it plays a role in DENV replication, Mfn2 did not play a role in mitochondrial elongation during DENV infection. Mfn2 has been reported to tether the ER to mitochondria, facilitating the transfer of calcium between the two organelles [73–75]. HIV-1 Vpr protein triggers the ubiquitinylation and turnover of Mfn2 by the CUL4 E3 ligase, leading to reduced ER-mitochondria contacts [76]. In contrast, other studies reported an increase of ER-mitochondria contacts in Mfn2 depleted cells compared to wild-type cells [77, 78]. It is possible that decreased levels of Mfn2 during DENV infection perturb the interaction between the ER and mitochondria, which could also contribute to a decrease in Drp1 recruitment to mitochondria during DENV infection. Accordingly, the level of MAM marker FACL4 was decreased in the mitochondria fraction of DENV-infected cells. Using live-cell imaging, we also noted a decrease in mitochondrial motility within cells infected with DENV. As described mainly in neurons, mitochondrial transport occurs along microtubules and relies on the activity of the Miro, Milton, and kinesin heavy chain motor complex [79, 80]. Miro depletion releases the mitochondria from the motor complex, preventing mitochondrial movement. Interestingly, Mfn2 also interacts with both Miro and Milton proteins and is essential for efficient mitochondrial mobility [81, 82]. Whether Mfn2 depletion is involved in reduction of mitochondrial motility during DENV infection remains to be determined. Mitochondrial elongation and function Previous reports have established that elongated mitochondria are associated with increased oxidative phosphorylation and ATP production [83–85]. An interconnected mitochondrial network induced during DENV infection may serve to optimize mitochondrial function, increasing bioenergetic capacity of the cell as a pro-survival response against stress. Our results indicate that DENV-infected cells increase mitochondrial respiration associated with ATP production with little to no change in proton leak or maximal respiration. These findings suggest that DENV-induced elongation of mitochondria enhances mitochondrial respiration and energy production, which is in contrast with previous observations indicating mitochondrial dysfunction and alterations in cellular ATP balance in DENV-infected HepG2 cells [42]. These differences may stem from cell-type specificity. Mitochondria hyperfusion has been described under conditions of stress such as starvation [84] or oxidative stress [86]. However, the induction of mitochondrial fragmentation with CCCP or overexpression of WT and S616-phosphorylated Drp1 reduce viral replication, suggesting that mitochondrial elongation is advantageous to DENV replication, and is not solely a consequence of metabolic stress. By contrast to DENV, HCV which is a strong inducer of oxidative stress [87] triggers mitochondrial fission via Drp1 overexpression [60], indicating that metabolic stress during virus infection is not always associated with mitochondrial elongation. Interestingly, DENV modulates the unfolded protein response in a time-dependent manner, preventing apoptosis and prolonging viral life cycle [88], as well as represses the formation of stress granules and processing (P) bodies [89], suggesting that DENV handles host cell stress responses. Whether DENV protein(s) specifically induce(s) mitochondrial elongation is under investigation. In summary, this report demonstrates DENV-induced mitochondrial elongation through inhibition of the fission protein, Drp1. The mechanism of Drp1 inhibition such as protein degradation and/or inhibition of kinases involved in Drp1 phosphorylation will need to be determined. Similarly, whether DENV proteins play a direct role in Drp1 regulation at the ER-mitochondrial interface remains to be investigated. Materials and Methods Cell culture and virus infection Vero, Huh7.5 (obtained from American Type Culture Collection (ATCC)) and Huh7 (a generous gift from Dr. Kate Fitzgerald) cells were maintained in Dulbecco’s modified minimal essential medium and HepG2 (ATCC) cells were maintained in Eagle’s Minimum Essential Medium, supplemented with 10% heat-inactivated fetal bovine serum (Sigma-Aldrich), 1% penicillin-streptomycin (Sigma-Aldrich), 1% non-essential amino acids (Lonza) and 1% L-Glutamine solution (Sigma-Aldrich). All cells were incubated in a humidified chamber at 37°C and 5% CO2. DENV-2 strain 16681 was originally obtained from Walter Reed Army Institute of Research and passaged in C6/36 cells (ATCC). Virus titers were determined by immunostained plaque assay on Vero cells [48, 90]. Antibodies and reagents The following primary antibodies were used for immunofluorescence analysis or western blot: rabbit anti-Drp1 (ab180769, Abcam), anti-Drp1-P Ser616 (D9A1, Cell Signaling Technology), anti-Mfn1 (D6E2S, Cell Signaling Technology), anti-Mfn2 (D1E9, Cell Signaling Technology), anti-OPA1 (GTX48589, GeneTex), anti-TOMM20 (ab186734, Abcam), anti-beta-actin-HRP (ab20272, Abcam), anti-NS3 (GTX124252, GeneTex), anti-NS4b (GTX103349, GeneTex), anti-COX-IV (3E11, Cell Signaling Technology) and mouse anti-Dengue Virus Type II (MAB8702, EMD Millipore). Secondary antibodies were purchased from Thermo Fisher Scientific: Pierce anti-rabbit poly-HRP, anti-rabbit conjugated with Alexa Fluor® 647 or Alexa Fluor® 488 and anti-mouse conjugated with Alexa Fluor® 647. Carbonyl cyanide 3-chlorophenylhydrazone (CCCP, Sigma-Aldrich) was resuspended in medium and incubated at the concentration and time indicated in the figure legend. Generation of DCs Peripheral blood mononuclear cells (PBMCs) were isolated from buffy coat of healthy donors obtained from Oklahoma Blood Institute (Oklahoma City, OK) by density gradient centrifugation on Ficoll-Paque Premium (GE Healthcare). Monocytes were isolated from PBMCs by positive selection using CD14+ microbeads (Miltenyi Biotec). Monocytes were cultured in RPMI 1640 medium supplemented with 10% heat-inactivated FBS (Sigma-Aldrich), 1% penicillin-streptomycin (Sigma-Aldrich), 50 ng/mL human interleukin-4 (IL-4) and 160 ng/mL human granulocyte-macrophage colony-stimulating factor (GM-CSF) for 5 days to generate monocyte-derived dendritic cells (DCs). Fresh medium with cytokines was added at day 3, and cells were infected with DENV at day 5. DCs were screened by flow cytometric analysis (MACSQuant, Miltenyi Biotec) for expression of CD11c (VioBlue), CD14 (VioGreen), HLA-DR (FITC), CD3 (PE), CD19 (PE), CD86 (PE-Vio770), CD83 (APC), CD1a (APC-Vio770) (Miltenyi Biotec). Cell fractionation Total cell lysates were prepared by solubilizing cells in lysis buffer (Miltenyi Biotec) supplemented with a protease and phosphatase inhibitor cocktail that consists of 1% Halt protease and phosphatase inhibitor cocktail (Thermo Fisher Scientific), 5 mM EDTA (Thermo Fisher Scientific), 1% phosphatase inhibitor cocktail 3 (Sigma-Aldrich), and 0.1% pepstatin A (Research Products International Corp.). The cytosolic fraction was obtained from supernatant of total cell lysate centrifuged at 13,000 g for 10 min, 4°C. Mitochondria were isolated from total cell lysates using the Mitochondria Isolation Kit (Miltenyi Biotec), according to manufacturer’s instructions. Briefly, cells were lysed using the lysis buffer mentioned earlier and homogenized using a glass dounce tissue grinder (Sigma-Aldrich), with 250–300 strokes per sample, on ice. Anti-TOM22 magnetic beads were added to the cell lysates to label the mitochondria, followed by mitochondria isolation in the magnetic field of a MidiMACS separation unit. Isolated mitochondria were resuspended in RIPA buffer (150 mM NaCl, 1% Triton X-100, 0.5% sodium deoxycholate, 0.1% SDS, 50 mM Tris-HCl, pH=8.0) supplemented with a protease and phosphatase inhibitor cocktail. Western Blot analysis Cells were lysed in RIPA buffer freshly supplemented with a protease and phosphatase cocktail. Lysates were incubated on ice for 30 min and centrifuged at 10,000 rpm for 10 min at 4°C. Supernatants were used as total cell lysates. Protein concentrations were determined using Pierce BCA Protein Assay kit (Thermo Fisher Scientific) following manufacturer’s instructions and measured using an Envision plate reader (PerkinElmer). BOLT LDS Sample Buffer and Reducing Agent (Thermo Fisher Scientific) were added to cell lysates at a final concentration of 1x and samples were denatured at 70°C for 10 min. Proteins were separated on 4–12% or 8% BOLT Bis-Tris Plus gel (Thermo Fisher Scientific) and transferred onto nitrocellulose membranes using the Trans-Blot® Turbo™ RTA Mini Nitrocellulose Transfer kit (Bio-Rad). Membranes were blocked in 5% skim milk for 1h at room temperature (rt) followed by incubation with the appropriate primary antibody overnight at 4°C. Membranes were washed and incubated with the appropriate secondary antibody diluted for 1h at rt. For β-actin analysis, membranes were stained for 1 h with anti-β-actin-HRP. Detection was performed using Amersham ECL Select Western Blotting Detection Reagent (GE Healthcare Life Sciences) following manufacturer’s guidelines and images were captured with ChemiDoc™ XRS+ System (Bio-Rad). Detection and quantification of band intensities was performed using Image Lab 5.1 (Bio-Rad). Band intensities were normalized to loading control TOMM20 for mitochondrial fractions. Plasmids, dsiRNA and transfections The following plasmids were used in our study: pNS4b5-eGFP [48], pTagRFP-mito (Evrogen), pAcGFP-Sec61β (gift from Tom Rapoport, Addgene plasmid #15108), pEYFP-Drp1, pEYFP-Drp1 S616D, pEYFP-Drp1 S616A, pEYFP-Drp1 S637D, pEYFP-Drp1 S637A (Drp1 phosphomimetic and phosphodefective mutant plasmids were a gift from Dr. Luca Scorrano, University of Padua), pDrp1 K38A-CFP (gift from Dr. Gyorgy Hajnoczky, Thomas Jefferson University)). Transfection of cells with plasmids was performed using GeneJuice® Transfection Reagent (EMD Millipore), according to the manufacturer’s instructions. Dicer substrate siRNAs (dsiRNAs) were purchased from Integrated DNA Technologies (IDT). Drp1, Mfn1 and Mfn2 dsiRNAs were designed with Custom RNAi Design Tool (IDT). Drp1 dsiRNA targets the 3′UTR region of human Drp1. DsiRNA sequences were as follows: Drp1 sense strand 5′-GCAGGAAUGCCUACAUUAAUUCCTA-3′ Drp1 anti-sense strand 5′-UAGGAAUUAAUGUAGGCAUUCCUGCUU-3′; Set (A) Mfn1 sense strand 5′-AGCGGAGACUUAGCAUAAUGGCAGA-3′ Mfn1 antisense strand 5′-UCUGCCAUUAUGCUAAGUCUCCGCUCC-3′; Set (B) Mfn1 sense strand 5′-AGAAUCCUAACAAUAGAGAUUGCTT-3′ Mfn1 antisense strand 5′-AAGCAAUCUCUAUUGUUAGGAUUCUUC-3′; Set (A) Mfn2 sense strand 5′-GCAUGGUACCAAGGAGUUAAGUUGA-3′ Mfn2 antisense strand 5′-UCAACUUAACUCCUUGGUACCAUGCUG-3′; Set (B) Mfn2 sense strand 5′-GGCUCAAGACUAUAAGCUGCGAATT-3′ Mfn2 antisense strand 5′-AAUUCGCAGCUUAUAGUCUUGAGCCAA-3′. A dsiRNA with no homology to any human gene was used as a negative control (DS NC1, IDT). Transfection of cells with dsiRNAs was performed at 5 nM final concentration using RiboJuice™ siRNA Transfection Reagent (EMD Millipore), according to the manufacturer’s instructions. Immunofluorescence analysis Cells were plated in 8-wells Nunc® Lab-Tek® II glass chamber slides (Sigma-Aldrich) and treated as indicated. Cells were fixed in 4% paraformaldehyde (Affymetrix) for 15 min at rt. Fixed cells were blocked and permeabilized in PBS with 0.1% Triton-X-100 and 10% normal goat serum (Cell Signaling Technology) for 30 min at rt. After blocking, cells were incubated with the appropriate primary antibody overnight at 4°C. Cells were washed and incubated with the appropriate secondary antibody for 1h 30min at rt. After washing, chambers were removed and slides mounted on coverslips with VectaShield (Vector Laboratories) mounting medium containing 6-diamino-2-phenylindole (DAPI) for detection of nuclei. Slides were analyzed on a LSM 800 confocal laser-scanning microscope (Zeiss) or Nikon C1si confocal microscope. Orthogonal projections were created from z-stacks, processed using median filter to reduce noise and improve image quality, and adjusted for brightness and contrast identically between each condition using ImageJ (NIH). The average mitochondrial lengths for cells within each condition were calculated using a macro developed in ImageJ (NIH). Briefly, the red channel image containing the MitoTracker Deep Red fluorescence was converted to 32 bit to keep image integrity and then to 16 bit and inverted. The image was further converted to black and white for analysis. The background was subtracted (rolling= 40 light) and Gaussian blur (sigma=0.90) was applied. The threshold was set (0, 235) and the image was converted to mask and then made binary. Skeletonization was applied to reduce pixel widths of captured mitochondria to a single pixel. Analyze particles was applied (size=0-infinity, circularity=0.00–1.00). The pixel lengths were determined for each particle captured from the image that corresponded to mitochondria. Statistical analysis was done on the average mitochondrial lengths of each cell using hierarchical linear regression models to adjust for the lack of independence of mitochondria within cells. Statistical analyses were performed utilizing JMP version 8 software (SAS institute). Live cell imaging Cells were plated in 8-wells Nunc™ Lab-Tek™ chambered coverglass (Thermo Fisher Scientific) and treated as indicated. Mitochondria were visualized in cells transfected with pTagRFP-mito (Evrogen) or after incubation with 200 nM MitoTracker® Deep Red FM (Thermo Fisher Scientific) for 15 min at 37°C. Nuclei were visualized using NucBlue® Live ReadyProbes® Reagent (Thermo Fisher Scientific) following manufacturer’s instructions. Images and time-lapse videos were captured on EVOS fluorescence microscope (Thermo Fisher Scientific) with 100× oil immersion objective. For time-lapse imaging, images were collected every 10 sec for 4 min and then processed identically between each condition to adjust brightness and contrast using ImageJ software (NIH). Transmission Electron Microscopy Huh7 cells were cultured on Permanox Lab-Tek chamber slides, fixed with 1.25% glutaraldehyde in 0.15M sodium cacodylate buffer at 4°C. Following fixation, cells were buffer rinsed and post-fixed in 1% osmium tetroxide. After distilled water rinses, samples were stained en bloc with 1% aqueous uranyl acetate for 30 min in the dark. Slides were rinsed and dehydrated through a graded ethanol series. Media chambers and gaskets were removed. Slides were covered with Epox 812 resin and placed over resin filled slide-duplicating molds and polymerized overnight at 60°C. The resin was separated from the slide, and cultures were examined by light microscopy to select regions of interest. Selected areas were cut out and mounted on blocks for sectioning. Ultra-thin sections (50nm) were prepared using a Reichert Ultracut S microtome (Leica Biosystems), retrieved onto 300 mesh copper grids, and contrasted with uranyl acetate and lead citrate stains. Sections were examined at 80 kV using a CM-10 electron microscope (FEI). Images were collected with a model 785 Erlangshen ES1000W CCD camera (Gatan). qRT-PCR analysis Total RNA extracted from cell pellets using the RNeasy Mini Kit (Qiagen) and the QIAcube. cDNA was synthetized from 250 ng of RNA with the RT2 First Strand Kit (Qiagen). Standard quantitative RT-PCR was performed with the RT2 SYBR® Green qPCR Mastermix (Qiagen) on a Bio-Rad CFX machine, and gene expression was normalized to the beta-actin expression level. Primer sets were designed using Primer-BLAST (NCBI) and purchased from Integrated DNA Technologies (IDT). Primer sequences were as follows: forward primer Drp1: 5′-CACCCGGAGACCTCTCATTC-3′, reverse primer Drp1: 5′-TTTACCCCATTCTTCTGCTTCCA-3′; forward primer beta-actin 5′-AGAGCTACGAGCTGCCTGAC-3′, reverse primer beta-actin 5′-AGTTTCGTGGATGCCACAGG-3′. Mitochondrial function profile Oxygen consumption rate (OCR), an indicator of mitochondrial respiration, was measured using a Seahorse Bioscience XF24 extracellular flux analyzer (Seahorse Bioscience). Huh7 cells were seeded in a XF24 microplate using complete growth medium and infected 1 day later with DENV (MOI 1) for 48h. The day before the assay, the cartridge sensor was hydrated overnight with 1mL Seahorse Bioscience XF24 Calibration Buffer at 37°C without CO2. On the day of the assay, the growth medium was replaced with XF24 Base Assay medium and cells were incubated at 37°C in a non-CO2 incubator for 1h. OCR values were monitored under basal condition and measured after sequential injection of oligomycin (1 μM), FCCP (1 μM), and antimycin A/rotenone (0.5 μM) using XF Cell Mito Stress kit (Seahorse Bioscience). Mitochondrial function parameters were analyzed according to Seahorse Bioscience instructions. Supplementary Material 1 Fig. S1. DENV infection induces mitochondrial elongation A. and B. Immunofluorescence analysis of mitochondrial morphology in uninfected (cytoplasmic GFP) or DENV-infected (cytoplasmic and nuclear GFP) HepG2 (A.) and Huh7.5 (B.) cells. Cells were transfected with pNS4b5-eGFP reporter and infected with DENV (MOI of 0.1) for 48h. Cells were subsequently stained with MitoTracker Deep Red FM prior to imaging. Images were converted to 8-bit and inverted for ImageJ analysis of mitochondrial lengths. Images were taken at 100× magnification and are representative of at least three independent experiments. C. Immunofluorescence analysis of mitochondrial morphology in HepG2 cells uninfected or infected with DENV (MOI of 1) for 48h. Cells were fixed and stained with antibodies against DENV antigen and COX-IV. Images were taken with Nikon Leica C1si confocal microscope ×120 magnification. D. and E. Immunofluorescence analysis of mitochondrial morphology in Huh7.5 (D.) and Huh7 (E.) cells transfected with pTagRFP-mito and uninfected or infected with DENV (MOI of 1) for 48h. Cells were fixed and stained with antibody against DENV antigen. Images were taken with confocal LSM800 at ×63×1.5 magnification and are representative of at least three independent experiments. 2 Fig. S2. Infection state of cells for mitochondrial motility assay Immunofluorescence analysis of Huh7 cells transfected with pTagRFP-mito and pNS4b5-eGFP and uninfected or infected with DENV (MOI of 1) for 48h. Live cells were imaged with EVOS fluorescence microscope at ×100 magnification. Infected cells (cytoplasmic and nuclear GFP) were discriminated from uninfected cells (cytoplasmic GFP) and processed for time-lapse imaging (Movie S1 and S2). Time-lapse live cell imaging was performed on RFP channel to visualize mitochondria in uninfected cell (Movie S1) and infected cell (Movie S2). Time-lapse videos show images collected every 10 sec for 4 min and then processed using ImageJ software (NIH). 3 Fig. S3. DENV infection decreases the level of Mfn2 protein, but not Mfn1 A. and B. Mfn1 and Mfn2 protein levels in uninfected or DENV-infected (MOI of 1) Huh7 cells were analyzed by immunofluorescence at 48h pi. Cells were fixed and stained with antibodies against Mfn1 (A.), Mfn2 (B.) and DENV antigen. Images were taken with confocal LSM800 at ×20 magnification. 4 Fig. S4. DENV infection decreases the level of Drp1 and Drp1 S616-P A. and B. Drp1 and Drp1 S616-P levels in uninfected or DENV-infected (MOI of 0.1 or 1) Huh7.5 cells were analyzed by immunofluorescence at 48h pi. Cells were fixed and stained with antibodies against Drp1 (A.) or Drp1 S616-P (B.) protein and DENV antigen. Images were taken with confocal LSM800 at ×20 magnification. Quantification of Mean Fluorescence Intensity (MFI) in individual cells from image corresponding to Drp1 or Drp1 S616-P staining was performed using ImageJ and compared for statistical analysis using unpaired Student-t test. 5 Fig. S5. DENV infection decreases Drp1 and Drp1 S616-P protein levels, but does not affect Drp1 gene expression A. qRT-PCR analysis of Drp1 gene expression was performed in uninfected and DENV-infected DCs after 8h, 12h and 24h of infection. Ct values were normalized to the housekeeping gene β-actin. Data are represented as mean +/− SD of three independent experiments (except 12h, n=2). Statistical analysis was done with unpaired Student-t test. B and C. Drp1 and Drp1 S616-P levels in uninfected or DENV-infected Huh7.5 cells were analyzed by immunofluorescence at 48h pi. Cells were transfected with pTagRFP-mito, challenged with DENV (MOI of 1) and after 48h of infection, cells were fixed and stained with antibodies against Drp1 (B.) or Drp1 S616-P (C.) and DENV antigen. Images were taken with confocal LSM800 at ×63×1.5 magnification. 6 Fig. S6. CCCP-induced mitochondrial fragmentation decreases DENV antigen staining DENV-2 antigen levels in uninfected or DENV-infected (MOI of 1) Vero cells, treated as indicated, were analyzed by immunofluorescence with confocal LSM800 at ×20 magnification. Cells were fixed and stained with antibody against DENV antigen. 7 Fig. S7. Control for knockdown of mitochondrial fusion proteins Mfn1 and Mfn2, and mitochondrial fission protein Drp1 A. Huh7.5 cells were transfected with dsiRNAs against Mfn1 or Mfn2 and infected with DENV (MOI of 0.1) for 24h. Western blots were probed with Mfn1 and Mfn2 antibodies and equal protein loading was assessed using an antibody against β-actin. Blot is representative of at least four independent experiments. B. Huh7.5 cells were transfected with a negative control dsiRNA or dsiRNA targeting the untranslated region of Drp1. Cells were transfected with plasmids expressing Drp1 wild-type or mutants and infected with DENV (MOI of 0.1) for 24h. Western blots were probed with antibodies directed against Drp1 protein and equal protein loading was assessed using an antibody against β-actin. 8 Movie S1. Time-lapse live cell imaging of mitochondria in uninfected Huh7 cell. 9 Movie S2. Time-lapse live cell imaging of mitochondria in DENV-infected Huh7 cell. We would like to thank Dr. Robbert Creton (Brown University) for his guidance in development of the ImageJ macro to capture mitochondrial lengths, Jennifer Friedman (Rhode Island Hospital) for help in the statistical analysis of mitochondria, Dr. Meenakshi Khare and Siraj Janoudi (University of Rhode Island) for technical assistance, Dr. Luca Scorrano (University of Padua) for generously providing the Drp1 phosphorylation mutant plasmids, Dr. David Guertin lab (UMass Medical School) for access to the Seahorse Bioanalyzer and Dr. Gyorgy Hajnoczky (Thomas Jefferson University) for providing the pDrp1 K38A-CFP plasmid. We also thank Dr. Kate Fitzgerald (UMass Medical School) and lab members for helpful discussions. Funding information This work was supported by National Institute of General Medical Sciences of the National Institutes of Health under grant number P20 GM104317 (COBRE) and P20 GM103430 (INBRE). Fig. 1 DENV infection induces mitochondrial elongation A. Immunofluorescence analysis of mitochondrial morphology in uninfected (cytoplasmic GFP) or DENV-infected (cytoplasmic and nuclear GFP) Huh7 cells. Cells were transfected with pNS4b5-eGFP reporter and infected with DENV (MOI of 0.1) for 48h. Cells were subsequently stained with MitoTracker Deep Red FM prior to imaging. Images were taken at 100× magnification and are representative of at least three independent experiments. B. Quantitative analysis of mitochondrial lengths in uninfected and infected Vero cells after 24h and 48h of infection. Vero cells transfected with pNS4b5-eGFP reporter were infected with DENV (MOI of 1) for 24h and 48h and mitochondria were stained with MitoTracker Deep Red FM. The average mitochondrial lengths for cells within each condition were calculated using a macro developed in ImageJ (NIH). Statistical analysis was done on the average mitochondrial lengths of each cell as described. The results reflect three independent experiments. C. and D. Ultra-thin sections TEM images of uninfected and DENV-infected (MOI of 1) resin-embedded Huh7 cells at different magnifications (C., ×7900, scale bar: 500 nm; D., ×19000, scale bar: 200 nm). Uninfected Huh7 cells exhibit typical ultrastructural morphology of non-infected cells. DENV-infected cells display elongated mitochondria (Mt), along with previously described convoluted membranes (CM) and invaginated vesicles (Ve) in proximity to the endoplasmic reticulum (ER) membranes. Nucleus (Nu) appears on lower magnification images (C.). E. Magnified image of box area illustrates invaginated vesicles (Ve) present in DENV-infected cells. Fig. 2 DENV NS4b and NS3 proteins are enriched in the mitochondrial fraction A. and B. Western blot analysis of NS4b and NS3 protein levels was performed in subcellular fractions from uninfected and DENV-infected Huh7 cells (A.) and DCs (B.). After 48h of infection, cells were fractionated to separate cytosolic and mitochondrial fractions, and compared to total cell lysates. TOMM20 was used as a loading and purity control for mitochondrial fraction. FACL4 antibody served as a marker for MAMs. β-actin antibody was used as a loading control. Western blots are representative of three independent experiments. C. Immunofluorescence analysis of NS4b protein localization in uninfected and DENV-infected Huh7 cells after 48h (MOI of 1). Cells were transfected with pTagRFP-mito and pAcGFP-Sec61β. Images were taken with confocal LSM800 at ×63×1.5 magnification. Fig. 3 DENV-infected cells display increased mitochondrial respiration A. OCR in uninfected and DENV-infected Huh7 cells at 48h post-infection (MOI of 1) were measured in real time under basal conditions and in response to sequential injections of oligomycin (1 μM), FCCP (1 μM) and antimycin A/rotenone (0.5 μM) using XF24 Seahorse analyzer. Data are shown as percent of baseline OCR relative to first measurement from uninfected cells and represent mean +/− SEM of three independent experiments with at least five biological replicates per condition and measurement. B. Mitochondrial function parameters were analyzed according to Seahorse Bioscience instructions and represented as fold change in OCR relative to uninfected cells. Data are shown as mean +/− SD, analyzed with unpaired Student-t test. Fig. 4 DENV infection decreases the level of Drp1 and Mfn2 on mitochondria A. and B. DENV-infected and uninfected Huh7 cells (A.) and DCs (B.) were harvested at 48h post-infection (pi) and total, cytosolic and mitochondrial fractions were analyzed by western blot for Mfn1, Mfn2, Opa1, Drp1 and Drp1 S616-P protein levels. NS4b antibody served as a control for DENV infection. TOMM20 antibody was used as a loading and purity control for mitochondrial fraction. β-actin antibody was used as a loading control for all fractions. For quantification of protein levels in mitochondrial fraction, blots were evaluated after normalization to loading controls and expressed as mean +/− SD of at least three independent experiments. Statistical analysis was performed using unpaired Student-t test. Fig. 5 CCCP-induced mitochondrial fragmentation or depletion of mitochondrial fusion protein Mfn2 affects DENV infection A. Supernatants from DENV infected Vero cells were collected after 24h of infection. Cells were treated with medium containing DMSO or CCCP 10 μM and supernatants were again collected after 6h post treatment (30h pi). To reverse CCCP treatment, DMSO and CCCP-treated cells were washed once with PBS and supernatant was collected (48h pi). DENV titers from collected supernatants were analyzed using an ELISpot assay. Data are representative of one out of two independent experiments, analyzed with unpaired Student-t test. Error bars represent SD of three biological replicates per condition. B. Immunofluorescence analysis of mitochondrial morphology of uninfected and DENV-infected Vero cells after indicated treatments. Cells were transfected with pTagRFP-mito, challenged with DENV (MOI of 1) and treated as indicated before being fixed and stained with antibody against DENV antigen. Images were taken with confocal LSM800 at ×63×1.5 magnification. Magnified images of boxed area show mitochondria stained with pTagRFP-mito for each condition. Images were converted to black and white using ImageJ. C. Huh7.5 cells were transfected with two independent sets of dsiRNAs against Mfn1 or Mfn2 (set (A), left and set (B), right) and infected with DENV (MOI of 0.1) for 24h. Western blots were probed with NS3 and NS5 antibodies and equal protein loading was assessed using an antibody against β-actin. Blot is representative of at least four independent experiments. D. For quantification of NS3 and NS5 protein levels, blots of four independent experiments were evaluated after normalization to β-actin and expressed as mean +/− SEM. Statistical analysis was performed using unpaired Student-t test. E. DENV titers from collected supernatants of Huh7.5 cells treated with dsiRNAs against Mfn1 or Mfn2 were analyzed using an ELISpot assay. DENV titers were transformed to log (FFU/ml) and displayed as mean +/− SEM. Statistical analysis was performed using paired t-test. Titers are representative of independent experiments (n=3). Fig. 6 Mitochondrial fission inhibits DENV replication A. Huh7.5 cells were transfected with dsiRNA against the untranslated region of Drp1. Cells were transfected with plasmids expressing Drp1 wildtype, phosphomimetic mutants of human Drp1 (S616D and S637D), phosphodeficient mutants (S616A and S637A) and dominant negative protein Drp1 K38A and infected with DENV (MOI of 0.1) for 24h. Western blots were probed with antibodies directed against NS3 and NS4b proteins and equal protein loading was assessed using an antibody against β-actin. Blot is representative of at least three independent experiments. B. For quantification of protein levels, blots of three independent experiments were evaluated after normalization to β-actin and expressed as mean +/− SEM. Statistical analysis was performed using unpaired Student-t test. C. DENV titers from collected supernatants of Huh7.5 cells expressing indicated Drp1 proteins were analyzed using an ELISpot assay. Titers are representative of independent experiments (n=3). Highlights Mitochondrial length and respiration are increased during DENV infection. DENV inhibits Drp1-triggered mitochondrial fission. DENV titers are reduced by mitochondrial fragmentation, Drp1 WT and S616D expression. Viral proteins NS4b and NS3 are associated with subcellular fractions of mitochondria. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. 1 Bhatt S Gething PW Brady OJ Messina JP Farlow AW Moyes CL Drake JM Brownstein JS Hoen AG Sankoh O The global distribution and burden of dengue Nature 2013 496 7446 504 507 23563266 2 Brady OJ Gething PW Bhatt S Messina JP Brownstein JS Hoen AG Moyes CL Farlow AW Scott TW Hay SI Refining the global spatial limits of dengue virus transmission by evidence-based consensus PLoS Negl Trop Dis 2012 6 8 e1760 22880140 3 Simmons CP Farrar JJ Nguyen v V Wills B Dengue N Engl J Med 2012 366 15 1423 1432 22494122 4 Dengue: Guidelines for Diagnosis, Treatment, Prevention and Control: New Edition Geneva 2009 5 Aye KS Charngkaew K Win N Wai KZ Moe K Punyadee N Thiemmeca S Suttitheptumrong A Sukpanichnant S Prida M Pathologic highlights of dengue hemorrhagic fever in 13 autopsy cases from Myanmar Hum Pathol 2014 45 6 1221 1233 24767772 6 Balsitis SJ Coloma J Castro G Alava A Flores D McKerrow JH Beatty PR Harris E Tropism of dengue virus in mice and humans defined by viral nonstructural protein 3-specific immunostaining Am J Trop Med Hyg 2009 80 3 416 424 19270292 7 Couvelard A Marianneau P Bedel C Drouet MT Vachon F Henin D Deubel V Report of a fatal case of dengue infection with hepatitis: demonstration of dengue antigens in hepatocytes and liver apoptosis Hum Pathol 1999 30 9 1106 1110 10492047 8 Durbin AP Vargas MJ Wanionek K Hammond SN Gordon A Rocha C Balmaseda A Harris E Phenotyping of peripheral blood mononuclear cells during acute dengue illness demonstrates infection and increased activation of monocytes in severe cases compared to classic dengue fever Virology 2008 376 2 429 435 18452966 9 Hall WC Crowell TP Watts DM Barros VL Kruger H Pinheiro F Peters CJ Demonstration of yellow fever and dengue antigens in formalin-fixed paraffin-embedded human liver by immunohistochemical analysis Am J Trop Med Hyg 1991 45 4 408 417 1951849 10 Huerre MR Lan NT Marianneau P Hue NB Khun H Hung NT Khen NT Drouet MT Huong VT Ha DQ Liver histopathology and biological correlates in five cases of fatal dengue fever in Vietnamese children Virchows Arch 2001 438 2 107 115 11253111 11 Jessie K Fong MY Devi S Lam SK Wong KT Localization of dengue virus in naturally infected human tissues, by immunohistochemistry and in situ hybridization J Infect Dis 2004 189 8 1411 1418 15073678 12 Kou Z Quinn M Chen H Rodrigo WW Rose RC Schlesinger JJ Jin X Monocytes, but not T or B cells, are the principal target cells for dengue virus (DV) infection among human peripheral blood mononuclear cells J Med Virol 2008 80 1 134 146 18041019 13 Povoa TF Alves AM Oliveira CA Nuovo GJ Chagas VL Paes MV The pathology of severe dengue in multiple organs of human fatal cases: histopathology, ultrastructure and virus replication PLoS One 2014 9 4 e83386 24736395 14 Schmid MA Diamond MS Harris E Dendritic cells in dengue virus infection: targets of virus replication and mediators of immunity Front Immunol 2014 5 647 25566258 15 Suksanpaisan L Cabrera-Hernandez A Smith DR Infection of human primary hepatocytes with dengue virus serotype 2 J Med Virol 2007 79 3 300 307 17245728 16 Brandt WE McCown JM Gentry MK Russell PK Infection enhancement of dengue type 2 virus in the U-937 human monocyte cell line by antibodies to flavivirus cross-reactive determinants Infect Immun 1982 36 3 1036 1041 6284641 17 Halstead SB Mahalingam S Marovich MA Ubol S Mosser DM Intrinsic antibody-dependent enhancement of microbial infection in macrophages: disease regulation by immune complexes Lancet Infect Dis 2010 10 10 712 722 20883967 18 Kliks SC Nisalak A Brandt WE Wahl L Burke DS Antibody-dependent enhancement of dengue virus growth in human monocytes as a risk factor for dengue hemorrhagic fever Am J Trop Med Hyg 1989 40 4 444 451 2712199 19 Laoprasopwattana K Libraty DH Endy TP Nisalak A Chunsuttiwat S Vaughn DW Reed G Ennis FA Rothman AL Green S Dengue Virus (DV) enhancing antibody activity in preillness plasma does not predict subsequent disease severity or viremia in secondary DV infection J Infect Dis 2005 192 3 510 519 15995967 20 Samanta J Sharma V Dengue and its effects on liver World J Clin Cases 2015 3 2 125 131 25685758 21 Rothman AL Cellular immunology of sequential dengue virus infection and its role in disease pathogenesis Curr Top Microbiol Immunol 2010 338 83 98 19802580 22 Whitehorn J Farrar J Dengue Br Med Bull 2010 95 161 173 20616106 23 de Armas-Rillo L Valera MS Marrero-Hernandez S Valenzuela-Fernandez A Membrane dynamics associated with viral infection Rev Med Virol 2016 26 3 146 160 26817660 24 Welsch S Miller S Romero-Brey I Merz A Bleck CK Walther P Fuller SD Antony C Krijnse-Locker J Bartenschlager R Composition and three-dimensional architecture of the dengue virus replication and assembly sites Cell Host Microbe 2009 5 4 365 375 19380115 25 Allonso D Andrade IS Conde JN Coelho DR Rocha DC da Silva ML Ventura GT Silva EM Mohana-Borges R Dengue Virus NS1 Protein Modulates Cellular Energy Metabolism by Increasing Glyceraldehyde-3-Phosphate Dehydrogenase Activity J Virol 2015 89 23 11871 11883 26378175 26 Heaton NS Randall G Dengue virus-induced autophagy regulates lipid metabolism Cell Host Microbe 2010 8 5 422 432 21075353 27 Tait SW Green DR Mitochondria and cell signalling J Cell Sci 2012 125 Pt 4 807 815 22448037 28 Green DR Reed JC Mitochondria and apoptosis Science 1998 281 5381 1309 1312 9721092 29 Li XD Sun L Seth RB Pineda G Chen ZJ Hepatitis C virus protease NS3/4A cleaves mitochondrial antiviral signaling protein off the mitochondria to evade innate immunity Proc Natl Acad Sci U S A 2005 102 49 17717 17722 16301520 30 Sun Q Sun L Liu HH Chen X Seth RB Forman J Chen ZJ The specific and essential role of MAVS in antiviral innate immune responses Immunity 2006 24 5 633 642 16713980 31 West AP Khoury-Hanold W Staron M Tal MC Pineda CM Lang SM Bestwick M Duguay BA Raimundo N MacDuff DA Mitochondrial DNA stress primes the antiviral innate immune response Nature 2015 520 7548 553 557 25642965 32 Twig G Elorza A Molina AJ Mohamed H Wikstrom JD Walzer G Stiles L Haigh SE Katz S Las G Fission and selective fusion govern mitochondrial segregation and elimination by autophagy EMBO J 2008 27 2 433 446 18200046 33 Ingerman E Perkins EM Marino M Mears JA McCaffery JM Hinshaw JE Nunnari J Dnm1 forms spirals that are structurally tailored to fit mitochondria J Cell Biol 2005 170 7 1021 1027 16186251 34 Bui HT Shaw JM Dynamin assembly strategies and adaptor proteins in mitochondrial fission Curr Biol 2013 23 19 R891 899 24112988 35 Taguchi N Ishihara N Jofuku A Oka T Mihara K Mitotic phosphorylation of dynamin-related GTPase Drp1 participates in mitochondrial fission J Biol Chem 2007 282 15 11521 11529 17301055 36 Friedman JR Lackner LL West M DiBenedetto JR Nunnari J Voeltz GK ER tubules mark sites of mitochondrial division Science 2011 334 6054 358 362 21885730 37 Korobova F Ramabhadran V Higgs HN An actin-dependent step in mitochondrial fission mediated by the ER-associated formin INF2 Science 2013 339 6118 464 467 23349293 38 Elgass K Pakay J Ryan MT Palmer CS Recent advances into the understanding of mitochondrial fission Biochim Biophys Acta 2013 1833 1 150 161 22580041 39 Knott AB Bossy-Wetzel E Impairing the mitochondrial fission and fusion balance: a new mechanism of neurodegeneration Ann N Y Acad Sci 2008 1147 283 292 19076450 40 Burte F Carelli V Chinnery PF Yu-Wai-Man P Disturbed mitochondrial dynamics and neurodegenerative disorders Nat Rev Neurol 2015 11 1 11 24 25486875 41 Chan DC Mitochondria: dynamic organelles in disease, aging, and development Cell 2006 125 7 1241 1252 16814712 42 El-Bacha T Midlej V Pereira da Silva AP Silva da Costa L Benchimol M Galina A Da Poian AT Mitochondrial and bioenergetic dysfunction in human hepatic cells infected with dengue 2 virus Biochim Biophys Acta 2007 1772 10 1158 1166 17964123 43 Ho LJ Wang JJ Shaio MF Kao CL Chang DM Han SW Lai JH Infection of human dendritic cells by dengue virus causes cell maturation and cytokine production J Immunol 2001 166 3 1499 1506 11160189 44 Mosquera JA Hernandez JP Valero N Espina LM Anez GJ Ultrastructural studies on dengue virus type 2 infection of cultured human monocytes Virol J 2005 2 26 15801983 45 Zargar S Wani TA Jain SK Morphological changes in vero cells postinfection with dengue virus type-2 Microsc Res Tech 2011 74 4 314 319 20687131 46 Yu CY Liang JJ Li JK Lee YL Chang BL Su CI Huang WJ Lai MM Lin YL Dengue virus impairs mitochondrial fusion by cleaving mitofusins PLoS Pathog 2015 11 12 e1005350 26717518 47 Chatel-Chaix L Cortese M Romero-Brey I Bender S Neufeldt CJ Fischl W Scaturro P Schieber N Schwab Y Fischer B Dengue Virus Perturbs Mitochondrial Morphodynamics to Dampen Innate Immune Responses Cell Host Microbe 2016 48 Medin CL Valois S Patkar CG Rothman AL A plasmid-based reporter system for live cell imaging of dengue virus infected cells J Virol Methods 2015 211 55 62 25445884 49 Chen H Chan DC Mitochondrial dynamics–fusion, fission, movement, and mitophagy–in neurodegenerative diseases Hum Mol Genet 2009 18 R2 R169 176 19808793 50 Okuda M Li K Beard MR Showalter LA Scholle F Lemon SM Weinman SA Mitochondrial injury, oxidative stress, and antioxidant gene expression are induced by hepatitis C virus core protein Gastroenterology 2002 122 2 366 375 11832451 51 Schwer B Ren S Pietschmann T Kartenbeck J Kaehlcke K Bartenschlager R Yen TS Ott M Targeting of hepatitis C virus core protein to mitochondria through a novel C-terminal localization motif J Virol 2004 78 15 7958 7968 15254168 52 Ramage HR Kumar GR Verschueren E Johnson JR Von Dollen J Johnson T Newton B Shah P Horner J Krogan NJ A combined proteomics/genomics approach links hepatitis C virus infection with nonsense-mediated mRNA decay Mol Cell 2015 57 2 329 340 25616068 53 Chen H Detmer SA Ewald AJ Griffin EE Fraser SE Chan DC Mitofusins Mfn1 and Mfn2 coordinately regulate mitochondrial fusion and are essential for embryonic development J Cell Biol 2003 160 2 189 200 12527753 54 Cipolat S Martins de Brito O Dal Zilio B Scorrano L OPA1 requires mitofusin 1 to promote mitochondrial fusion Proc Natl Acad Sci U S A 2004 101 45 15927 15932 15509649 55 Cereghetti GM Stangherlin A Martins de Brito O Chang CR Blackstone C Bernardi P Scorrano L Dephosphorylation by calcineurin regulates translocation of Drp1 to mitochondria Proc Natl Acad Sci U S A 2008 105 41 15803 15808 18838687 56 Mai S Klinkenberg M Auburger G Bereiter-Hahn J Jendrach M Decreased expression of Drp1 and Fis1 mediates mitochondrial elongation in senescent cells and enhances resistance to oxidative stress through PINK1 J Cell Sci 2010 123 Pt 6 917 926 20179104 57 Ishihara N Jofuku A Eura Y Mihara K Regulation of mitochondrial morphology by membrane potential, and DRP1-dependent division and FZO1-dependent fusion reaction in mammalian cells Biochem Biophys Res Commun 2003 301 4 891 898 12589796 58 Chang CR Blackstone C Cyclic AMP-dependent protein kinase phosphorylation of Drp1 regulates its GTPase activity and mitochondrial morphology J Biol Chem 2007 282 30 21583 21587 17553808 59 Frank S Gaume B Bergmann-Leitner ES Leitner WW Robert EG Catez F Smith CL Youle RJ The role of dynamin-related protein 1, a mediator of mitochondrial fission, in apoptosis Dev Cell 2001 1 4 515 525 11703942 60 Kim SJ Syed GH Khan M Chiu WW Sohail MA Gish RG Siddiqui A Hepatitis C virus triggers mitochondrial fission and attenuates apoptosis to promote viral persistence Proc Natl Acad Sci U S A 2014 111 17 6413 6418 24733894 61 Stavru F Palmer AE Wang C Youle RJ Cossart P Atypical mitochondrial fission upon bacterial infection Proc Natl Acad Sci U S A 2013 110 40 16003 16008 24043775 62 Chen TT Wu LS Hsu PW Pang CY Lee KM Cheng PC Peng SY Mitochondrial dynamics in the mouse liver infected by Schistosoma mansoni Acta Trop 2015 148 13 23 25910628 63 Gorbunov NV McDaniel DP Zhai M Liao PJ Garrison BR Kiang JG Autophagy and mitochondrial remodelling in mouse mesenchymal stromal cells challenged with Staphylococcus epidermidis J Cell Mol Med 2015 19 5 1133 1150 25721260 64 Lum M Morona R Dynamin-related protein Drp1 and mitochondria are important for Shigella flexneri infection Int J Med Microbiol 2014 304 5–6 530 541 24755420 65 Kim SJ Khan M Quan J Till A Subramani S Siddiqui A Hepatitis B virus disrupts mitochondrial dynamics: induces fission and mitophagy to attenuate apoptosis PLoS Pathog 2013 9 12 e1003722 24339771 66 Yan J Liu XH Han MZ Wang YM Sun XL Yu N Li T Su B Chen ZY Blockage of GSK3beta-mediated Drp1 phosphorylation provides neuroprotection in neuronal and mouse models of Alzheimer’s disease Neurobiol Aging 2015 36 1 211 227 25192600 67 Murray CL Jones CT Rice CM Architects of assembly: roles of Flaviviridae nonstructural proteins in virion morphogenesis Nat Rev Microbiol 2008 6 9 699 708 18587411 68 Paul D Bartenschlager R Flaviviridae Replication Organelles: Oh, What a Tangled Web We Weave Annu Rev Virol 2015 2 1 289 310 26958917 69 Chen H Chomyn A Chan DC Disruption of fusion results in mitochondrial heterogeneity and dysfunction J Biol Chem 2005 280 28 26185 26192 15899901 70 Detmer SA Chan DC Complementation between mouse Mfn1 and Mfn2 protects mitochondrial fusion defects caused by CMT2A disease mutations J Cell Biol 2007 176 4 405 414 17296794 71 Rehman J Zhang HJ Toth PT Zhang Y Marsboom G Hong Z Salgia R Husain AN Wietholt C Archer SL Inhibition of mitochondrial fission prevents cell cycle progression in lung cancer FASEB J 2012 26 5 2175 2186 22321727 72 Thomas KJ Jacobson MR Defects in mitochondrial fission protein dynamin-related protein 1 are linked to apoptotic resistance and autophagy in a lung cancer model PloS one 2012 7 9 e45319 23028930 73 Rowland AA Voeltz GK Endoplasmic reticulum-mitochondria contacts: function of the junction Nat Rev Mol Cell Biol 2012 13 10 607 625 22992592 74 Raturi A Simmen T Where the endoplasmic reticulum and the mitochondrion tie the knot: the mitochondria-associated membrane (MAM) Biochim Biophys Acta 2013 1833 1 213 224 22575682 75 de Brito OM Scorrano L Mitofusin 2 tethers endoplasmic reticulum to mitochondria Nature 2008 456 7222 605 610 19052620 76 Huang CY Chiang SF Lin TY Chiou SH Chow KC HIV-1 Vpr triggers mitochondrial destruction by impairing Mfn2-mediated ER-mitochondria interaction PLoS One 2012 7 3 e33657 22438978 77 Cosson P Marchetti A Ravazzola M Orci L Mitofusin-2 independent juxtaposition of endoplasmic reticulum and mitochondria: an ultrastructural study PLoS One 2012 7 9 e46293 23029466 78 Filadi R Greotti E Turacchio G Luini A Pozzan T Pizzo P Mitofusin 2 ablation increases endoplasmic reticulum-mitochondria coupling Proc Natl Acad Sci U S A 2015 112 17 E2174 2181 25870285 79 Saxton WM Hollenbeck PJ The axonal transport of mitochondria J Cell Sci 2012 125 Pt 9 2095 2104 22619228 80 Schwarz TL Mitochondrial trafficking in neurons Cold Spring Harb Perspect Biol 2013 5 6 81 Misko A Jiang S Wegorzewska I Milbrandt J Baloh RH Mitofusin 2 is necessary for transport of axonal mitochondria and interacts with the Miro/Milton complex J Neurosci 2010 30 12 4232 4240 20335458 82 Baloh RH Schmidt RE Pestronk A Milbrandt J Altered axonal mitochondrial transport in the pathogenesis of Charcot-Marie-Tooth disease from mitofusin 2 mutations J Neurosci 2007 27 2 422 430 17215403 83 Mitra K Wunder C Roysam B Lin G Lippincott-Schwartz J A hyperfused mitochondrial state achieved at G1-S regulates cyclin E buildup and entry into S phase Proc Natl Acad Sci U S A 2009 106 29 11960 11965 19617534 84 Tondera D Grandemange S Jourdain A Karbowski M Mattenberger Y Herzig S Da Cruz S Clerc P Raschke I Merkwirth C SLP-2 is required for stress-induced mitochondrial hyperfusion EMBO J 2009 28 11 1589 1600 19360003 85 Rossignol R Gilkerson R Aggeler R Yamagata K Remington SJ Capaldi RA Energy substrate modulates mitochondrial structure and oxidative capacity in cancer cells Cancer Res 2004 64 3 985 993 14871829 86 Redpath CJ Bou Khalil M Drozdzal G Radisic M McBride HM Mitochondrial hyperfusion during oxidative stress is coupled to a dysregulation in calcium handling within a C2C12 cell model PLoS One 2013 8 7 e69165 23861961 87 Paracha UZ Fatima K Alqahtani M Chaudhary A Abuzenadah A Damanhouri G Qadri I Oxidative stress and hepatitis C virus Virol J 2013 10 251 23923986 88 Pena J Harris E Dengue virus modulates the unfolded protein response in a time-dependent manner J Biol Chem 2011 286 16 14226 14236 21385877 89 Emara MM Brinton MA Interaction of TIA-1/TIAR with West Nile and dengue virus products in infected cells interferes with stress granule formation and processing body assembly Proc Natl Acad Sci U S A 2007 104 21 9041 9046 17502609 90 Liu L Wen K Li J Hu D Huang Y Qiu L Cai J Che X Comparison of plaque- and enzyme-linked immunospot-based assays to measure the neutralizing activities of monoclonal antibodies specific to domain III of dengue virus envelope protein Clin Vaccine Immunol 2012 19 1 73 78 22116689

PMC005xxxxxx/PMC5131786.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 0402313 3651 Exp Hematol Exp. Hematol. Experimental hematology 0301-472X 1873-2399 26860989 5131786 10.1016/j.exphem.2016.01.009 NIHMS758334 Article Leukemia and chromosomal instability in aged Fancc−/− mice Cerabona Donna 12 Sun Zejin 1 Nalepa Grzegorz 1234 1 Department of Pediatrics, Indiana University School of Medicine, Indianapolis, IN, USA 2 Department of Biochemistry and Molecular Biology, Indiana University School of Medicine, Indianapolis, IN, USA 3 Department of Medical and Molecular Genetics, Indiana University School of Medicine, Indianapolis, IN, USA 4 Division of Pediatric Hematology–Oncology, Riley Hospital for Children, Indianapolis, IN, USA. Corresponding Author: Grzegorz Nalepa, M.D., Ph.D., Indiana University School of Medicine, Wells Center for Pediatric Research; 1044 W Walnut St., R4-421; Indianapolis, IN 46202, gnalepa@iu.edu 19 11 2016 6 2 2016 5 2016 01 5 2017 44 5 352357 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. Fanconi anemia (FA) is an inherited disorder of genomic instability associated with high risk of myelodysplasia and AML. Young mice deficient in FA core complex genes do not naturally develop cancer, hampering preclinical studies on malignant hematopoiesis in FA. Here we show that aging Fancc−/− mice are prone to genomically unstable AML and other hematologic neoplasms. We demonstrate that aneuploidy precedes malignant transformation during Fancc−/− hematopoiesis. Our observations reveal that Fancc−/− mice develop hematopoietic chromosomal instability followed by leukemia in age-dependent manner, recapitulating the clinical phenotype of human Fanconi anemia and providing a proof of concept for future development of preclinical models of FA-associated leukemogenesis. INTRODUCTION The Fanconi anemia (FA) signaling network protects genomic integrity and prevents cancer by facilitating interphase DNA repair and orchestrating cell division1-3. Germline biallelic mutations of any FA genes cause Fanconi anemia (FA), an inherited bone marrow failure syndrome associated with myelodysplasia (MDS) and acute myeloid leukemia (AML). The overall risk of leukemia in FA is increased 600-fold4. Young mice deficient in core FA genes do not spontaneously recapitulate clinical hematopoietic manifestations of Fanconi anemia5. Fancc−/− mice demonstrate hypersensitivity to cross-linking agents6, decreased hematopoietic stem cell repopulating ability7,8, and hypersensitivity to interferon-gamma8 reflecting disruption of the FA signaling network during hematopoiesis. However, young Fancc−/− mice do not develop spontaneous leukemia or bone marrow failure6,8. One observation study of a small Fancc−/− mouse cohort (n=8) did not detect decreased survival9. However, FA patients very rarely develop AML in their first year of life10 and two soft-tissue tumors (adenocarcinoma and histiocytic sarcoma) had been reported in >13-month old Fancc−/− mice11. Thus, we hypothesized that aging Fancc−/− mice may be predisposed to hematopoietic malignancies. If the absolute time to the onset of leukemia is similar in FA humans and mice, a long-term observation of FA−/− mice may be crucial to detect cancer predisposition. To address this translationally relevant question, we asked whether Fancc−/− mice develop malignancies as they age. RESULTS & DISCUSSION We observed cohorts of wt (n=20) and Fancc−/− (n=18) mice for 24 months, and noticed decreased survival of Fancc−/− mice (p=0.01, Fig. 1A). Five Fancc−/− mice (27.8%) died between 8 and 24 months of age due to leukemia or lymphoma (Fig. 1B-L). Specifically, we diagnosed acute myeloid leukemia (AML) in two moribund Fancc−/− mice with peripheral blasts, predominance of Gr-1+ (Ly-6G) peripheral blood low-density mononuclear cells (LDMNCs), and myeloid infiltrates around the liver vessels (Fig. 1C-E). One Fancc−/− mouse developed lethal B-cell acute lymphoblastic leukemia (ALL), as evidenced by expansion of B220+ blasts that replaced >90% of bone marrow and infiltrated the liver (Fig. 1F-H). Additionally, two Fancc−/− mice died due to metastatic abdominal T-cell lymphoma manifested by massive mesenteric lymph node conglomerates (Fig. 1I-J) accompanied by Cd3+ liver infiltrates (Fig. 1K-L) in the absence of bone marrow or peripheral blood abnormalities. After 24 months of observation, all surviving Fancc−/− and wt mice were sacrificed and examined by necropsy. 4/13 (30.8%) 2-year old Fancc−/− animals had hematopoietic solid tumors and/or peripheral blasts, consistent with leukemia/lymphoma. Serial blood counts did not reveal progressive pancytopenia in aging Fancc−/− mice, suggesting that the development of leukemia may not be preceded by bone marrow failure in this animal model of FA. Together, 9/18 Fancc−/− mice developed hematopoietic malignancies by two years of age (including 5 animals that died prematurely due to disease), compared to 0/20 control wt mice (Fig. 1B, p=0.0003). Thus, aging Fancc−/− mice are prone to hematopoietic neoplasms, reflecting the age-dependent risk of leukemia in FA patients4,10,12. We next asked whether Fancc−/− AML can be propagated in wt mice via competitive stem cell transplantation. We mixed donor Fancc−/− Cd45.2+ LDMNCs isolated from a moribund AML Fancc−/− mouse (Fig. 1C-E) with wt Cd45.1+ competitor LDMNCs at a 1:1 ratio and transplanted the mixed cells into 3 lethally irradiated wt recipients. Three wt recipients of age-matched wt Cd45.2+ LDMNCs mixed with wt Cd45.1+ LDMNCs served as controls (Fig. 2A). By 50 days post-transplantation, all recipients of Fancc−/− LDMNCs had died of AML, while control recipients of wt LDMNCs remained healthy (Fig. 2B). The diagnosis of AML was confirmed in all recipients by flow cytometry, peripheral blood smears (Fig. 2C) and splenomegaly (p=0.0216, Fig. 2D). Peripheral blood flow cytometry demonstrated increased Cd45.2+ chimerism (p=0.0436) in recipients of leukemic Fancc−/− LDMNCs compared to controls at one month post-transplant (Fig. 2E-F), highlighting the malignant potential of leukemic Fancc−/− LDMNCs to outcompete wt hematopoietic cells in the host marrow. The FA signaling network maintains genomic integrity during Fancc−/− hematopoiesis in vivo13 and genomic instability promotes cancer14. Thus, we asked whether leukemic Fancc−/− mice displayed increased chromosomal instability and whether chromosomal instability precedes the onset of leukemia during Fancc−/− hematopoiesis. We compared karyotypes of LDMNCs isolated from leukemic Fancc−/− to age-matched wt and healthy Fancc−/− marrows. Bone marrow cells isolated from healthy Fancc−/− mice had higher incidence of aneuploidy and increased frequency of abnormal mitotic figures than wt LDMNCs (Fig. 3A-D), showing that Fancc−/− hematopoietic cells become chromosomally unstable before overt leukemogenesis occurs. Similarly, FA patients develop hematopoietic chromosomal and nuclear abnormalities prior to the onset of leukemia13,15,16. Leukemic Fancc−/− bone marrows were more aneuploid (Fig. 3B) with higher mitotic index compared to both age-matched wt (p=0.001) and Fancc−/− non-leukemic marrows (p<0.0001, Fig. 3D). This observation is consistent with further exacerbation of genomic instability and acquisition of bizarre karyotypic abnormalities reported in human FA-associated AML17,18. In summary, Fancc−/− mice develop chromosomally unstable hematopoietic malignancies as they age, recapitulating clinical and genomic abnormalities seen in Fanconi anemia patients (Figs. 1 and 3). Interestingly, similar incidence of tumors had been reported in old mice deficient in another FA core gene, Fanca, although that observation did not reach statistical significance due to small sample sizes19. Thus, late-onset carcinogenesis may be a common phenotype of murine FA core gene knockouts. AML arising in Fancc−/− mice can be propagated via hematopoietic stem cell transplant and produce rapid onset of lethal leukemia in wt transplant recipients (Fig. 2). As large-scale cohorts of leukemic mice are essential for preclinical drug testing, our observations may facilitate the development of future preclinical models of FA−/− AML. METHODS Mice C57Bl/6J Fancc−/− mice were a gift of David W. Clapp (Indiana University). Mice were PCR-genotyped as described9. B6.SJL-PtprcaPepcb/BoyJ mice were purchased from the IU In Vivo Therapeutics Core. All studies were approved by the Institutional Animal Care and Use Committee at IU. Marrow harvest and transplantation Bone marrow cells were flushed from mouse femurs using a 23-gauge needle/syringe (Becton Dickinson). LDMNCs were isolated by density gradient using Histopaque-1119 (Sigma) centrifuging for 30 minutes at 1800rpm with no brake. After centrifugation, LDMNCs were removed from the interface and utilized for experiments. Cytospins were made by resuspending LDMNCs in PBS and centrifuging onto slides at 450rpm for 5 minutes on a Shandon Cytospin 3 Cytocentrifuge (Thermo Scientific). 1.5×106 donor test LDMNCs (C57Bl/6J background) and 1.5×106 donor competitor BoyJ LDMNCs were transplanted into recipients via tail vein injection. Recipients were 8 week-old female B6.SJL-PtprcaPepcb/BoyJ mice that underwent whole-body split-dose 1100 rads irradiation (700 rads/400 rads, 4 hours apart). For chimerism analysis, peripheral blood was collected from lateral tail veins into EDTA-coated tubes, incubated with RBC lysis solution (Qiagen) for 10 minutes at room temperature, washed, stained with anti-Cd45.2-FITC (BD Biosciences) and anti-Cd45.1-PE (BD Biosciences) as described9, and analyzed on a FacsCalibur machine (Becton-Dickinson). At least 10,000 events/sample were acquired and analyzed using FlowJo Software. Metaphase spreads Bone marrow cells flushed from tibias were cultured in IMDM plus 20% FBS, murine SCF (100ng/ml), and IL-6 (200ng/ml) for 2 days. Cells were then exposed to 0.2 μg/ml colcemid (Life Tech) for 4 hours and pelleted at 800rpm for 5 minutes. Cells were resuspended dropwise in pre-warmed (37°C) 75mM KCl, and incubated at 37°C for 15 minutes. After pelleting, cells were resuspended in a 3:1 methanol:glacial acetic acid fixative. Cells were pelleted and resuspended in fixative two additional times before being dropped onto slides and dried overnight. Spreads were stained with Vectashield mounting medium with DAPI (Vector Laboratories). Histology and flow cytometry Murine tissues obtained post-mortem were fixed in 10% formalin, paraffin-embedded, sectioned (5 μM sections), and stained with hematoxylin and eosin. Peripheral blood smears and bone marrow cytospins were stained with Giemsa using the automated Siemens Hematek 3000 (Fisher) system. For flow cytometry, peripheral blood cells were incubated in RBC lysis solution and bone marrow LDMNCS were isolated as described above. Cells were stained with either Gr-1-APC (Ly6G, clone: RB6-8C5) or B220-FITC (clone: RA3-6B2), analyzed on a FacsCalibur machine using live gating followed by data quantification with FlowJo software. Leukemia diagnoses were made using criteria established in the Bethesda proposal for classification of nonlymphoid neoplasms in mice20, and were independently validated by a veterinary pathologist at IU School of Medicine. Microscopy Images of smears, cytospins, and histological sections were obtained using a Zeiss Axiolab microscope with a color camera. Metaphase spreads were imaged on a Deltavision personalDx deconvolution microscope (Applied Precision). Image stacks (distance between z-sections: 0.2 μm) were deconvolved using Softworx and analyzed using Imaris software suite (Bitplane). Statistics Statistical analysis was performed with the GraphPad Prism 6 software. ACKNOWLEDGEMENTS The authors thank the following funding sources: NIH K12 Pediatric Scientist Award, the Barth Syndrome/Bone Marrow Failure Research Fund at Riley Children's Foundation, the Heroes Foundation, and NIH T32 HL007910 “Basic Science Studies on Gene Therapy of Blood Diseases” grant. We wish to thank Dr. George Sandusky for his pathology expertise, the IU In Vivo Therapeutics Core for use of their irradiation facility, and Li Jiang for transplantation assistance. We wish to acknowledge the work of our colleagues that we were unable to cite in this report due to space limitations. Fig 1 Aging Fancc−/− mice develop hematologic malignancies (A) Kaplan-Meier survival curve of wt (n=20) and Fancc−/− (n=18) cohorts. (B) Table demonstrating incidence of leukemias and lymphomas in wt and Fancc−/− mice by 24 months of age. Statistical significance for (A, B) was determined using a log-rank (Mantel-Cox) test. Peripheral blood smear of a moribund Fancc−/− mouse (C) shows leukemic blasts (arrows). Diagnosis of acute myeloid leukemia was confirmed with flow cytometry demonstrating increased expression of the Gr1 myeloid marker in the peripheral blood compared to wt and healthy Fancc−/− controls (D) and the presence of leukemic infiltrates in the liver (E). Bone marrow cytospin (F) of another moribund Fancc−/− mouse demonstrated multiple blasts (arrows). Flow cytometry demonstrated increased expression of the B220 B-cell marker on bone marrow blasts (G) and necropsy revealed leukemic infiltrates in the liver (H), consistent with B-cell ALL. Necropsy of another Fancc−/− mouse demonstrated conglomerates of mesenteric lymph nodes (I, J). Liver infiltrates in this mouse were Cd3+ (K, L), consistent with T-cell lymphoma. Fig 2 Transplanted Fancc−/− AML LDMNCs induce aggressive leukemia in wt recipients (A) Competitive repopulation transplant schematic. Cd45.2+ LDMNCs obtained from a leukemic Fancc−/− mouse or wt control were mixed 1:1 with Cd45.1+ wt competitor LDMNCs and transplanted into lethally irradiated wt recipients (n=3 recipients per genotype). (B) Recipients of Fancc−/− LDMNCs died within 2 months after transplantation. Log-rank (Mantel-Cox) test was used to assess significance. (C) AML in wt recipients of Fancc−/− LDMNCs. Gr1 was used as a myeloid marker for flow cytometry (right panel). (D) Splenomegaly in moribund wt mice transplanted with Fancc−/− LDMNCs. (E, F) Leukemic Cd45.2+ Fancc−/− LDMNCs overpopulate recipient bone marrow compared to wt Cd45.1+ competitor LDMNCs. Statistical significance was determined using an unpaired student's t-test; error bars represent SEM. Fig 3 Genomic instability and abnormal mitosis in leukemic and pre-leukemic Fancc−/− mice (A) Representative images of LDMNC metaphase spreads from wt and leukemic Fancc−/− mice. (B) Increased aneuploidy in leukemic Fancc−/− LDMNCs. At least 74 spreads were counted from wt, Fancc−/− non-leukemic, and Fancc−/− leukemic mice. Note increased chromosomal instability in non-leukemic Fancc−/− LDMNCs compared to age-matched wt controls. Fisher's exact test was used to determine statistical significance. Leukemic Fancc−/− LDMNCs undergo abnormal mitosis (C) and have a higher mitotic index (D) compared to LDMNCs from wt and Fancc−/− non-leukemic mice (n=3 mice/genotype; at least 500 cells were counted per genotype). Statistical analyses were performed using chi-square tests with Yates correction. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. The authors declare no conflicts of interest. REFERENCES 1 Kottemann MC Smogorzewska A Fanconi anaemia and the repair of Watson and Crick DNA crosslinks. Nature 2013 493 7432 356 363 23325218 2 Nalepa G Clapp DW Fanconi anemia and the cell cycle: new perspectives on aneuploidy. F1000Prime Rep 2014 6 23 24765528 3 D'Andrea AD Susceptibility pathways in Fanconi's anemia and breast cancer. N Engl J Med 2010 362 20 1909 1919 20484397 4 Alter BP Fanconi anemia and the development of leukemia. Best Pract Res Clin Haematol 2014 27 3-4 214 221 25455269 5 Parmar K D'Andrea A Niedernhofer LJ Mouse models of Fanconi anemia. Mutat Res 2009 668 1-2 133 140 19427003 6 Chen M Tomkins DJ Auerbach W Inactivation of Fac in mice produces inducible chromosomal instability and reduced fertility reminiscent of Fanconi anaemia. Nat Genet 1996 12 4 448 451 8630504 7 Haneline LS Gobbett TA Ramani R Loss of FancC function results in decreased hematopoietic stem cell repopulating ability. Blood 1999 94 1 1 8 10381491 8 Whitney MA Royle G Low MJ Germ cell defects and hematopoietic hypersensitivity to gamma-interferon in mice with a targeted disruption of the Fanconi anemia C gene. Blood 1996 88 1 49 58 8704201 9 Pulliam-Leath AC Ciccone SL Nalepa G Genetic disruption of both Fancc and Fancg in mice recapitulates the hematopoietic manifestations of Fanconi anemia. Blood 2010 116 16 2915 2920 20606166 10 Alter BP Giri N Savage SA Malignancies and survival patterns in the National Cancer Institute inherited bone marrow failure syndromes cohort study. Br J Haematol 2010 150 2 179 188 20507306 11 Carreau M Not-so-novel phenotypes in the Fanconi anemia group D2 mouse model. Blood 2004 103 6 2430 14998919 12 Kutler DI Singh B Satagopan J A 20-year perspective on the International Fanconi Anemia Registry (IFAR). Blood 2003 101 4 1249 1256 12393516 13 Abdul-Sater Z Cerabona D Sierra Potchanant E FANCA safeguards interphase and mitosis during hematopoiesis in vivo. Exp Hematol 2015 14 Gordon DJ Resio B Pellman D Causes and consequences of aneuploidy in cancer. Nat Rev Genet 2012 13 3 189 203 22269907 15 Mehta PA Harris RE Davies SM Numerical chromosomal changes and risk of development of myelodysplastic syndrome--acute myeloid leukemia in patients with Fanconi anemia. Cancer Genet Cytogenet 2010 203 2 180 186 21156231 16 Barton JC Parmley RT Carroll AJ Preleukemia in Fanconi's anemia: hematopoietic cell multinuclearity, membrane duplication, and dysgranulogenesis. J Submicrosc Cytol 1987 19 2 355 364 3599130 17 Woo HI Kim HJ Lee SH Yoo KH Koo HH Kim SH Acute myeloid leukemia with complex hypodiploidy and loss of heterozygosity of 17p in a boy with Fanconi anemia. Ann Clin Lab Sci 2011 41 1 66 70 21325258 18 Quentin S Cuccuini W Ceccaldi R Myelodysplasia and leukemia of Fanconi anemia are associated with a specific pattern of genomic abnormalities that includes cryptic RUNX1/AML1 lesions. Blood 2011 117 15 e161 170 21325596 19 Wong JC Alon N McKerlie C Huang JR Meyn MS Buchwald M Targeted disruption of exons 1 to 6 of the Fanconi Anemia group A gene leads to growth retardation, strain-specific microphthalmia, meiotic defects and primordial germ cell hypoplasia. Hum Mol Genet 2003 12 16 2063 2076 12913077 20 Kogan SC Ward JM Anver MR Bethesda proposals for classification of nonlymphoid hematopoietic neoplasms in mice. Blood 2002 100 1 238 245 12070033

PMC005xxxxxx/PMC5131797.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 8300886 3025 Dermatol Clin Dermatol Clin Dermatologic clinics 0733-8635 1558-0520 27890234 5131797 10.1016/j.det.2016.07.002 NIHMS827950 Article Current Status of Dedicator of Cytokinesis-Associated Immunodeficiency: DOCK8 and DOCK2 Dimitrova Dimana M.D 1 Freeman Alexandra F M.D. 2 1 Experimental Transplantation and Immunology Branch, National Cancer Institute, National Institutes of Health, 10 Center Drive, Bethesda, MD 20892 2 Laboratory of Clinical Infectious Diseases, National Institute of Allergy and Infectious Diseases, National Institutes of Health, 10 Center Drive, Bethesda, MD 20892 Corresponding author: Alexandra Freeman MD, NIH Building 10 Room 12C103, Bethesda, MD 20892, P (301)594-9045, F (301)496-0773, freemaal@mail.nih.gov 5 11 2016 1 2017 01 1 2018 35 1 1119 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. Synopsis DOCK8 deficiency is an autosomal recessive combined immunodeficiency disease associated with elevated IgE, atopy, recurrent sinopulmonary and cutaneous viral infections, and malignancy. The DOCK8 protein is critical for cytoskeletal organization, and deficiency impairs dendritic cell transmigration, T cell survival, and NK cell cytotoxicity. Early hematopoietic stem cell transplant is gaining prominence as a definitive treatment given the potential for severe complications and mortality in this disease. Recently, DOCK2 deficiency has been identified in several patients with early-onset invasive bacterial and viral infections. Immunodeficiency dedicator of cytokinesis 8 dedicator of cytokinesis 2 atopic dermatitis cutaneous viral infection malignancy INTRODUCTION Dedicator of cytokinesis 8 (DOCK8) deficiency is an autosomal recessive combined immunodeficiency syndrome characterized by recurrent sinopulmonary and cutaneous viral infections, as well as an increased IgE level and atopy. While patients with an autosomal recessive variant of hyper-IgE syndrome had been described as early as 2004, a genetic basis involving bi-allelic mutations often with large deletions was not established until 2009.1–3 In the intervening years, definitive treatment with early hematopoietic stem cell transplant (HSCT) has gained prominence, and advances have been made in understanding the functions of DOCK8 in dendritic cell and lymphocyte activity. Recently, another syndrome with differing phenotype but similar immunopathogenic basis, dedicator of cytokinesis 2 (DOCK2) deficiency, has been described.4 CLINICAL PRESENTATION OF DOCK8 DEFICIENCY Atopy Patients with DOCK8 deficiency demonstrate atopy early on. Nearly all patients exhibit atopic dermatitis (AD) which ranges from mild to very severe and difficult to treat (see Fig. 1a, 1b). Unlike patients with autosomal dominant hyper-IgE syndrome from STAT3 mutations, many have food allergies with anaphylaxis, as well as asthma. Eosinophilic esophagitis has also been seen with increased frequency.3,5–7 Infections Cellulitis and skin abscesses are common, as is mucocutaneous candidiasis. There is a striking susceptibility to cutaneous infections by viruses, such as human papillomavirus (HPV) leading to widespread and recalcitrant warts, extensive and disfiguring molluscum contagiosum, herpes simplex virus (HSV) with recurrent or persistent lesions or herpes keratitis, and varicella zoster virus (VZV) with severe primary infection or recurrent zoster (see Fig. 2a, 2b, 2c). Chronic EBV viremia is frequent, and may be associated with transformation to malignancy. Interestingly, severe systemic viral infections are less common, although several patients have suffered from CMV disease, encephalitis and progressive multifocal leukoencephalopathy.3,5–7 Most patients have a history of recurrent sinusitis and otitis media requiring tympanostomy tubes. Most also have had multiple pneumonias, with development of bronchiectasis in over a third but infrequent pneumatocele formation (see Fig. 3).3,5–7 Malignancy Increased risk of neoplasms, especially hematological and epithelial, is an important feature of DOCK8 deficiency, and malignancy is often particularly aggressive and has early onset. In one large cohort of 136 patients, 17% of patients were diagnosed with malignancy at a median age of 12 years.6 Malignancy most frequently arises from poor control of viruses including squamous cell carcinomas from HPV infection and EBV related lymphomas. Microcystic adnexal carcinoma, aggressive cutaneous T-cell lymphoma, and diffuse large B cell lymphoma have been described.3,8 Of note, not all tumors noted have been associated with viral infection. Other Clinical Manifestations Vascular abnormalities have been recognized in over 10% of patients in two recently described large cohorts (see Fig. 4). Cerebral aneurysms and stenosis is seen, and has been associated with stroke. Vaccine strain varicella was identified as the etiologic agent in one case, but in others an infectious has not been identified. Aortic aneurysm and abdominal arterial vasculitis has also been described, without known etiology. Autoimmunity rarely has manifest in other forms such as hemolytic anemia.6,7 Liver disease, both associated with and without cryptosporidia has been described as well, and can be quite significant leading to liver transplant.9,10 Laboratory Features DOCK8 deficiency is a combined T and B cell immunodeficiency. In the initial cohort of patients described by Zhang et al, all were noted to have normal neutrophils and monocytes. Ninety percent had low total T cells and CD8+ T cells, and all had low CD4+ T cells; 36% had low B and 60% low NK cells. In addition, there was poor CD8+ but not CD4+ T cell proliferation in response to stimulation.3 Engelhardt et al observed lower rates of T cell lymphopenia in their 2009 and 2015 cohorts (38% and 27% respectively), while Aydin et al observed low total lymphocyte counts in 20% of patients but low total T cells and CD4+ T cells in nearly half of patients.2,6,7 Elevated IgE and eosinophilia were nearly ubiquitous. Consistently, IgG levels were usually normal or elevated, IgA levels were variable, and IgM levels tended to be low and to decline with age.6 Vaccine responses to polysaccharide and protein antigens were variable, but the patients followed by Zhang et al showed protective titers to rubella and VZV.3 Half of patients had low or absent specific antibody responses to pneumococcus, diphtheria, tetanus, or Candida.7 Memory B cells in DOCK8-deficient patients were near absent, as were switched memory B cells.7,11 Memory T cell numbers were variable, but in one study most CD8+ cells had an exhausted CD45RA+/CCR7− phenotype.7,12 Caracciolo et al noted low numbers of naive and recent thymic emigrant T lymphocytes, along with Th2 skewing.11 This is consistent with low T cell receptor excision circles in 3 children with DOCK8 deficiency, a finding with ramifications for potential early detection of this disease.13 Making the Diagnosis Given the potential for severe infection and malignancy, it is important to recognize DOCK8 deficiency before development of serious complications whenever possible. Diagnosis may be difficult, and, particularly in infants and young children, early presentation may significantly overlap with severe AD in both laboratory and clinical features. Genetic sequencing is key to making the diagnosis, but the expense makes this prohibitive for screening. Thus, several groups have sought to identify markers that can clue in the clinician to an underlying monogenetic disorder. Furthermore, distinguishing features of different monogenetic hyper-IgE syndromes on presentation is important for targeting subsequent evaluations. When compared to severe AD patients, DOCK8-deficient patients were more likely to have low total T cells, low CD4+ T cells, and decreased naive CD8+ T cells in one small study. Total B lymphocyte numbers did not differ significantly between the two groups, but subsets revealed decreased memory and increased naive and transitional B cells in the DOCK8-deficient patients.14 When examining IgE-sensitization patterns in AD, STAT3 deficiency, and DOCK8 deficiency, Boos et al found that AD patients had the highest ratios of aeroallergen-specific IgE to total IgE, while patients with DOCK8 deficiency showed the highest serum specific IgE against food antigens, followed by AD patients.15 Using the NIH-developed Hyper-IgE Syndrome scoring system, Engelhardt et al compared clinical and laboratory scoring for DOCK8 and STAT3 patients and identified several objective features that were helpful in distinguishing the two syndromes: parenchymal lung abnormalities, retained primary teeth, and minor trauma fractures were deemed most consistent with STAT3 deficiency.7,16 Characteristic facies was also significantly associated with STAT3 deficiency but was considered a subjective assessment. By assigning negative points to the three features above and adding points for based on absolute eosinophil count and frequency of sinus and ear infections, the group developed a DOCK8 score that appears promising but has yet to be validated.7 Another study featured long-term follow-up of biomarker trends in individual DOCK8- and STAT3-deficient patients. DOCK8 patients demonstrated consistently lower total, CD4+, and CD8+ T cell numbers but normal Th17 cells as opposed to low Th17 cell but otherwise normal numbers of T cells in STAT3-deficient patients. In terms of clinical characteristics, the authors suggested that a history of recurrent viral infections, bronchial hyperreactivity, food allergies, and consanguinity should prompt greater concern for DOCK8 deficiency.17 In addition to severe atopic dermatitis and STAT3 deficiency, the differential diagnosis for a patient with atopic dermatitis, elevated IgE, and recurrent infections includes several other genetic disorders. Wiskott-Aldrich syndrome (WAS) is characterized by T cell lymphopenia, poor lymphocyte proliferation, impaired NK cytotoxicity, autoimmunity, and malignancy. The WAS protein (WASp) coordinates cytoskeletal reorganization downstream from DOCK8, which accounts for some overlap in phenotype including recurrent bacterial and viral infections, eczema and vascular abnormalities. Distinguishing features of WAS include X-linked inheritance and microthrombocytopenia.18 Phosphoglucomutase 3 (PGM3) deficiency, a congenital disorder of glycosylation, has been recently identified in patients who, in addition to severe atopy and hypergammaglobulinemia, usually have lymphopenia and neutropenia. Developmental delay or neurologic impairment is common.19 Omenn syndrome is a form of severe combined immunodeficiency associated with several different genetic defects. Severe erythroderma and exfoliative dermatitis are evident in early infancy, along with elevated IgE, infections, lymphadenopathy and hepatosplenomegaly.20 STK4 or Macrophage Stimulating 1 (MST1) deficiency, discovered within the last few years, has a phenotype similar to DOCK8 deficiency, with cutaneous viral, bacterial, and fungal infections, recurrent respiratory infections, and CD4 lymphopenia. Atopic dermatitis seems to be milder, and IgG and IgA are elevated as well as IgE. Cardiac anomalies have been noted in multiple patients.21 Tyk2 deficiency was described in 2006 in a patient with elevated IgE, atopic dermatitis, recurrent skin staphylococcal abscesses, and mycobacterial infection. However, recently, seven new Tyk2-deficient patients were identified, all with normal IgE, calling into question the classification of Tyk2 deficiency as a hyper-IgE syndrome.22 FUNCTIONS OF THE DOCK8 PROTEIN The DOCK8 protein is a member of the DOCK180-related family of atypical guanine nucleotide exchange factors that activate small Rho GTPases such as Rac and Cdc42.23,24 These GTPases interact with DOCK8 at actin projections called lamellipodia, which are found at the leading edge of motile cells such as endothelial cells, neurons, immune cells, and epithelial cells. Early studies suggested DOCK8 as an important player in dynamic actin reorganization because of its accumulation at lamellipodia and its ability to induce formation of vesicles containing filamentous actin.23 As DOCK8 is highly expressed only in the immune system (and at low levels in non-immune tissues such as the placenta, kidney, lung, and pancreas), more recent investigation has shed light on the protein’s specific role in the survival and function of dendritic cells (DCs) and lymphocytes.23,25 Dendritic Cells DCs adapt their shape to facilitate amoeboid migration through the interstitium. In a DOCK8 knockout mouse model, DCs were unable to crawl through 3-dimensional (3D) fibrillar networks and transmigrate into the lymph node for T cell priming, due to impaired Cdc42 activation at the leading edge membrane.26 The peripheral blood of DOCK8-deficient patients shows severe deficiency of plasmacytoid dendritic cells and correspondingly low interferon alpha (IFN-α) levels.27,28 B Lymphocytes The B cells of mice lacking DOCK8 cannot develop into marginal zone B cells, survive in germinal centers, and undergo affinity maturation, leading to normal initiation but poor persistence of antibody response post-immunization. DOCK8 appears to be required for organization of a B cell immunological synapse by recruiting the integrin ligand ICAM-1.29 In a study of DOCK8-deficient patients who showed either poor initial antibody responses to vaccination or poor persistence of protective titers, DOCK8 was shown to function as an adaptor linking TLR9 to MyD88 and downstream signaling pathways to effect B cell activation.30 T, NK, and NKT Lymphocytes Recent studies of DOCK8-deficient mice show a lack of CD4+ T cell infiltration into HSV-infected skin, associated with poor control of primary cutaneous herpes simplex lesions and increased virus loads.31 T cell lymphopenia is a prominent feature in DOCK8-deficient mice, due to both decreased survival of CD4+ and CD8+ T cells and decreased egress of mature CD4+ T cells from the thymus.32 Primary T cell response to infection or immunization is near-normal, but there is poor secondary expansion given reduced survival of memory CD8+ T cells.12,32 Diminished recruitment of LFA-1 to the CD8+ T cell/DC synapse and delay in the first cell division likely result in impaired generation of long-lived CD8+ T cells.12 As mentioned previously, DOCK8 activates Cdc42, which via its effector WASp is necessary for reorganizing the F-actin cytoskeleton in NK and dendritic cells. WASp itself also interacts directly with DOCK8, and WASp function may thus also be reduced in DOCK8 patients.33,34 DOCK8-deficient NK cells show decreased natural and receptor-mediated cytotoxicity, with decreased polarization of LFA-1, F-actin, and cytolytic granules toward the cytotoxic synapse.33,35 Zhang et al examined lymphocyte migration through the dermis and found that DOCK8-deficient T and NK cells develop an abnormal shape when moving in confined spaces.36 In a 3D collagen gel matrix simulating the dermis, the cells were able to move, unlike the nonmotile dendritic cells described by Harada et al, and chemotaxis was not impaired.26 However, hours later, the cells underwent fragmentation (cytothripsis) and died, suggesting that DOCK8 is essential for coordinating lymphocyte cytoskeletal integrity in this milieu. The early cell death prevents the generation of long-lived skin-resident memory CD8+ T cells, which may explain the preponderance of severe cutaneous viral infections in these patients. Low NKT cell numbers have been noted in DOCK8-deficient humans, while in mice ongoing NKT proliferative and cytokine responses are impaired.37 A recent analysis of the cytokine profile of DOCK8 deficiency shows that unstimulated DOCK8-deficient PBMCs secrete higher levels of inflammatory cytokines such as IFNgamma, IL1beta, IL4, and IL6 as compared to healthy control cells. Interestingly, stimulated PBMCs secreted less IFNgamma, suggesting impaired Th1 cell function. As Cdc42 is required for IFNgamma secretion at the immunological synapse, this finding is consistent with an intrinsic defect secondary to DOCK8 deficiency.17 DOCK8 Expression in Tumors Many but not all of the malignancies described in DOCK8 deficiency are virus-associated, which has prompted the question of whether the DOCK8 protein may have some intrinsic tumor suppressive function. Loss of chromosome arm 9p, where the DOCK8 gene is also located, is common in lung cancer.38 Deletions and other chromosomal alterations encompassing the DOCK8 gene have been associated with lung, gastric, pancreatic, and head and neck squamous cell carcinomas, while decreased expression of DOCK8 has been noted in certain lung and liver tumors, as well as in high-grade gliomas.39–44 However, increased expression of DOCK8 has been noted in a radiosensitive esophageal cancer line and in hepatocellular carcinoma cells, so a consistent role for DOCK8 in tumorigenesis has not emerged.45,46 THERAPEUTIC APPROACHES TO DOCK8 DEFICIENCY Nearly two thirds of patients receive immunoglobulin replacement therapy, as well as prophylactic antibiotics. Some patients also receive antiviral and antifungal prophylaxis.6 Even in patients who are otherwise well-controlled, HSV lesions, molluscum, and warts may be particularly recalcitrant, disfiguring, and problematic from a quality of life standpoint. Systemic IFN-α 2b therapy, which may act by inhibiting viral replication and activating effector lymphocytes, has yielded dramatic improvement of viral infection in 3 published cases of DOCK8-deficient patients.27,28 However, significant side effects may be associated with IFN-α 2b therapy, and careful monitoring is essential. Nevertheless, DOCK8 deficiency is associated with significant mortality, mainly due to infection or malignancy. In the cohort described by Engelhardt et al, mean age of death was 9 years and 3 months, while Aydin et al report probability of survival of 37% at age 30 years if not transplanted.6,7 Hematopoietic stem cell transplant (HSCT) has been repeatedly shown to be curative and more recently is being offered at an early stage. In their initial post-transplant course, patients may have a transient worsening of warts and chronic bacterial pretransplant infections. However, within several months, marked improvement or, more commonly, complete resolution of all skin manifestations has consistently been noted even in patients who previously experienced particularly severe or disfiguring skin disease. Complete immunological correction has generally been reported, even in several cases of mixed donor chimerism. Two deaths have been described in the literature, one considered transplant-related and one due to Klebsiella sepsis in the context of congenital asplenia, and unpublished experience has shown transplant-related mortality in the setting of pre-transplant significant end-organ disease.47–57 Somatic reversions were identified in 17 DOCK8 patients followed at the National Institutes of Health, and these patients demonstrated longer survival and a milder disease course; however, experience with these patients has shown that they may nevertheless have life-threatening complications and require HSCT.58 DOCK2 DEFICIENCY DOCK2, like DOCK8, is a member of the DOCK180 superfamily of proteins. DOCK2-deficient mice were known to have immunological defects even before DOCK8 deficiency was described in humans. Notable features included T cell lymphopenia and decreased T cell proliferation, loss of marginal zone B cells, and decreased myeloid and lymphocyte migration, with the potential for developing hyper-IgE.25 Recently, biallelic DOCK 2 mutations were identified in 5 patients with invasive bacterial and viral infections, lymphopenia and impaired antibody responses.4 Three of the children were born to consanguineous parents. Infections included recurrent pneumonia, disseminated varicella, Mycobacterium avium, mumps meningoencephalitis, and Klebsiella pneumoniae sepsis. Unfortunately, only 3 of the patients survived to receive HSCT and attain clinical improvement. Further investigation of the patients’ T, B, and NK cells revealed defective chemotaxis, actin polymerization, and NK cell degranulation. Interestingly, viral replication and virus-induced cell death were increased in DOCK2-deficient fibroblasts, and inducing lentiviral-mediated DOCK2 expression in the presence of interferon alfa-2b protected the cells. Thus, DOCK2 may impair non-hematopoietic immunity as well. While some similarities exist between the DOCK2 and DOCK8 deficiency phenotypes, only one of the DOCK2 patients had elevated IgE, and severe eczema and allergies were not prominent features in these patients. SUMMARY In the years since DOCK8 deficiency was described, much progress has been made in delineating the phenotype, establishing the role of DOCK8 in leukocyte function, and defining HSCT as a necessary treatment. The phenotype of DOCK2 deficiency in humans remains to be further characterized beyond the index patients. The growing experience with DOCK8-deficient patients highlights the concept that severe skin disease can be an indicator of underlying immunodeficiency, and thus dermatologists may be key to the early diagnosis of this disorder. Abbreviations and Acronyms Cdc42 G protein activated by DOCK guanine exchange factors DOCK2 Dedicator of cytokinesis 2, one of a class of guanine nucleotide exchange factors whose role is to activate the G protein Rac. DOCK8 Dedicator of cytokinesis 8, one of a class of guanine nucleotide exchange factors whose role is to activate G proteins such as Rac and Cdc42. Deficiency leads to impaired cytoskeletal organization and a phenotype of combined immunodeficiency with eczema, elevated IgE, and malignancy. HSCT Hematopoietic stem cell transplant, a treatment for primary immunodeficiencies that result from genetic defects in hematopoietic cells. ICAM-1 Intercellular adhesion molecule 1, a ligand for LFA-1 necessary for leukocyte endothelial transmigration. LFA-1 Lymphocyte function-associated antigen 1, binds to ICAM-1 and functions as an adhesion molecule. MST1 Macrophage stimulating 1 (also known as STK4). Deficiency results in a rare form of immunodeficiency wuth a phenotype similar to DOCK8 deficiency. MyD88 Myeloid differentiation primary response 88, a protein used by toll-like receptors to activate the transcription factor NF-κB. PGM3 Phosphoglucomutase 3, which mediates glycosylation. Deficiency results in a phenotype of severe atopy, hypergammaglobulinemia, leukopenia, and developmental delay. Rac G protein activated by DOCK guanine exchange factors. Rho GTPase G protein such as Cdc42 and Rac, activated by DOCK guanine exchange factors. These proteins regulate various aspects of cytoskeletal dynamics. STAT3 Signal transducer and activator of transcription 3, a transcription factor. Deficiency leads to autosomal dominant Hyper-IgE syndrome. TLR9 Toll-like receptor 9, important for activation of innate immunity via MyD88 Tyk2 Tyrosine kinase 2, a protein involved in IL-10 and IFN-α signaling. Deficiency has been associated with a variable phenotype that includes susceptibility to mycobacterial infection. WAS Wiskott-Aldrich syndrome is a rare X-linked recessive disease classically characterized by a triad of recurrent sinopulmonary infections, eczema, and thrombocytopenia with small platelets. Many patients do not exhibit the classic triad and may have autoimmune disease among other manifestations. WASp Wiskott-Aldrich syndrome protein, a protein that coordinates cytoskeletal reorganization. Deficiency leads to WAS. Fig. 1 Fig. 1a, 1b: Chronic severe eczematous dermatitis in a 6 year old male with DOCK8 deficiency. Fig. 2 Fig. 2a, 2b: Large, widespread warts as a manifestation of severe human papilloma virus infection in a 15 year old female with DOCK8 deficiency. Fig. 2c: Extensive molluscum contagiosum in a patient with DOCK8 deficiency. Fig. 3 Computed tomography of the chest of a 15 year old female with DOCK8 deficiency shows right middle lobe bronchiectasis with thickened bronchial walls. Fig. 4 Magnetic resonance angiography demonstrates vasculopathy in a 19 year old patient with DOCK8 deficiency. Shown below are a dilated ascending and transverse thoracic aorta with diffuse irregularity and foci of narrowing in the descending aorta. Other findings (not shown) included narrowing at the bifurcation of the right brachial cephalic artery, dilatation of the right subclavian artery at its origin, narrowing of the left common carotid, narrowing of the common iliacs, and narrowing of the right external iliac artery. Key Points DOCK8 deficiency is an autosomal recessive hyper-IgE syndrome associated with atopy, recurrent sinopulmonary and cutaneous viral infections, and malignancies. The DOCK8 protein plays an important role in cytoskeletal organization, impacting dendritic cell transmigration. DOCK8 deficiency leads to persistence of B cells in germinal centers, early T cell death, and lowered NK cell cytotoxicity. DOCK2 deficiency has been recently described in several patients with early-onset invasive bacterial and viral infections. Neither author has any commercial or financial conflicts of interest. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. 1 Renner ED Puck JM Holland SM Autosomal recessive hyperimmunoglobulin E syndrome: a distinct disease entity The Journal of pediatrics 2004 144 93 9 14722525 2 Engelhardt KR McGhee S Winkler S Large deletions and point mutations involving the dedicator of cytokinesis 8 (DOCK8) in the autosomal-recessive form of hyper-IgE syndrome The Journal of allergy and clinical immunology 2009 124 1289 302 e4 20004785 3 Zhang Q Davis JC Lamborn IT Combined immunodeficiency associated with DOCK8 mutations N Engl J Med 2009 361 2046 55 19776401 4 Dobbs K Dominguez Conde C Zhang SY Inherited DOCK2 Deficiency in Patients with Early-Onset Invasive Infections The New England journal of medicine 2015 372 2409 22 26083206 5 Zhang Q Davis JC Dove CG Su HC Genetic, clinical, and laboratory markers for DOCK8 immunodeficiency syndrome Disease markers 2010 29 131 9 21178272 6 Aydin SE Kilic SS Aytekin C DOCK8 deficiency: clinical and immunological phenotype and treatment options - a review of 136 patients Journal of clinical immunology 2015 35 189 98 25627830 7 Engelhardt KR Gertz ME Keles S The extended clinical phenotype of 64 patients with dedicator of cytokinesis 8 deficiency J Allergy Clin Immunol 2015 136 402 12 25724123 8 Chu EY Freeman AF Jing H Cutaneous manifestations of DOCK8 deficiency syndrome Arch Dermatol 2012 148 79 84 21931011 9 Al-Herz W Ragupathy R Massaad MJ Clinical, immunologic and genetic profiles of DOCK8-deficient patients in Kuwait Clinical immunology (Orlando, Fla) 2012 143 266 72 10 Shah NN Freeman AF Parta M Haploidentical transplantation for DOCK8 deficiency 57th Annual Meeting and Exposition of the American Society of Hematology Orlando, FL 2015 11 Caracciolo S Moratto D Giacomelli M Expansion of CCR4+ activated T cells is associated with memory B cell reduction in DOCK8-deficient patients Clin Immunol 2014 152 164 70 24674883 12 Randall KL Chan SS Ma CS DOCK8 deficiency impairs CD8 T cell survival and function in humans and mice The Journal of experimental medicine 2011 208 2305 20 22006977 13 Dasouki M Okonkwo KC Ray A Deficient T Cell Receptor Excision Circles (TRECs) in autosomal recessive hyper IgE syndrome caused by DOCK8 mutation: implications for pathogenesis and potential detection by newborn screening Clinical immunology (Orlando, Fla) 2011 141 128 32 14 Janssen E Tsitsikov E Al-Herz W Flow cytometry biomarkers distinguish DOCK8 deficiency from severe atopic dermatitis Clin Immunol 2014 150 220 4 24440647 15 Boos AC Hagl B Schlesinger A Atopic dermatitis, STAT3- and DOCK8- hyper-IgE syndromes differ in IgE-based sensitization pattern Allergy 2014 69 943 53 24898675 16 Grimbacher B Schaffer AA Holland SM Genetic linkage of hyper-IgE syndrome to chromosome 4 American journal of human genetics 1999 65 735 44 10441580 17 Hagl B Heinz V Schlesinger A Key findings to expedite the diagnosis of hyper-IgE syndromes in infants and young children Pediatric allergy and immunology : official publication of the European Society of Pediatric Allergy and Immunology 2015 18 Moulding DA Record J Malinova D Thrasher AJ Actin cytoskeletal defects in immunodeficiency Immunological reviews 2013 256 282 99 24117828 19 Zhang Y Yu X Ichikawa M Autosomal recessive phosphoglucomutase 3 (PGM3) mutations link glycosylation defects to atopy, immune deficiency, autoimmunity, and neurocognitive impairment The Journal of allergy and clinical immunology 2014 133 1400 9 9e1 5 24589341 20 Chinn IK Shearer WT Severe Combined Immunodeficiency Disorders Immunology and allergy clinics of North America 2015 35 671 94 26454313 21 Halacli SO Ayvaz DC Sun-Tan C STK4 (MST1) deficiency in two siblings with autoimmune cytopenias: A novel mutation Clinical immunology (Orlando, Fla) 2015 161 316 23 22 Kreins AY Ciancanelli MJ Okada S Human TYK2 deficiency: Mycobacterial and viral infections without hyper-IgE syndrome The Journal of experimental medicine 2015 212 1641 62 26304966 23 Ruusala A Aspenstrom P Isolation and characterisation of DOCK8, a member of the DOCK180-related regulators of cell morphology FEBS letters 2004 572 159 66 15304341 24 Cote JF Vuori K Identification of an evolutionarily conserved superfamily of DOCK180-related proteins with guanine nucleotide exchange activity Journal of cell science 2002 115 4901 13 12432077 25 Su HC Dedicator of cytokinesis 8 (DOCK8) deficiency Current opinion in allergy and clinical immunology 2010 10 515 20 20864884 26 Harada Y Tanaka Y Terasawa M DOCK8 is a Cdc42 activator critical for interstitial dendritic cell migration during immune responses Blood 2012 119 4451 61 22461490 27 Al-Zahrani D Raddadi A Massaad M Successful interferon-alpha 2b therapy for unremitting warts in a patient with DOCK8 deficiency Clin Immunol 2014 153 104 8 24743019 28 Keles S Jabara HH Reisli I Plasmacytoid dendritic cell depletion in DOCK8 deficiency: rescue of severe herpetic infections with IFN-alpha 2b therapy J Allergy Clin Immunol 2014 133 1753 5 e3 24767873 29 Randall KL Lambe T Johnson AL Dock8 mutations cripple B cell immunological synapses, germinal centers and long-lived antibody production Nat Immunol 2009 10 1283 91 19898472 30 Jabara HH McDonald DR Janssen E DOCK8 functions as an adaptor that links TLR-MyD88 signaling to B cell activation Nature immunology 2012 13 612 20 22581261 31 Flesch IE Randall KL Hollett NA Delayed control of herpes simplex virus infection and impaired CD4(+) T-cell migration to the skin in mouse models of DOCK8 deficiency Immunol Cell Biol 2015 93 517 21 25776845 32 Lambe T Crawford G Johnson AL DOCK8 is essential for T-cell survival and the maintenance of CD8+ T-cell memory Eur J Immunol 2011 41 3423 35 21969276 33 Ham H Guerrier S Kim J Dedicator of cytokinesis 8 interacts with talin and Wiskott-Aldrich syndrome protein to regulate NK cell cytotoxicity J Immunol 2013 190 3661 9 23455509 34 McGhee SA Chatila TA DOCK8 immune deficiency as a model for primary cytoskeletal dysfunction Disease markers 2010 29 151 6 21178274 35 Mizesko MC Banerjee PP Monaco-Shawver L Defective actin accumulation impairs human natural killer cell function in patients with dedicator of cytokinesis 8 deficiency J Allergy Clin Immunol 2013 131 840 8 23380217 36 Zhang Q Dove CG Hor JL DOCK8 regulates lymphocyte shape integrity for skin antiviral immunity J Exp Med 2014 211 2549 66 25422492 37 Crawford G Enders A Gileadi U DOCK8 is critical for the survival and function of NKT cells Blood 2013 122 2052 61 23929855 38 Sato M Takahashi K Nagayama K Identification of chromosome arm 9p as the most frequent target of homozygous deletions in lung cancer Genes, chromosomes & cancer 2005 44 405 14 16114034 39 Takahashi K Kohno T Ajima R Homozygous deletion and reduced expression of the DOCK8 gene in human lung cancer International journal of oncology 2006 28 321 8 16391785 40 Takada H Imoto I Tsuda H Genomic loss and epigenetic silencing of very-low-density lipoprotein receptor involved in gastric carcinogenesis Oncogene 2006 25 6554 62 16715143 41 Heidenblad M Schoenmakers EF Jonson T Genome-wide array-based comparative genomic hybridization reveals multiple amplification targets and novel homozygous deletions in pancreatic carcinoma cell lines Cancer research 2004 64 3052 9 15126341 42 Saelee P Wongkham S Puapairoj A Novel PNLIPRP3 and DOCK8 gene expression and prognostic implications of DNA loss on chromosome 10q25. 3 in hepatocellular carcinoma Asian Pacific journal of cancer prevention : APJCP 2009 10 501 6 19640199 43 Idbaih A Carvalho Silva R Criniere E Genomic changes in progression of low-grade gliomas Journal of neuro-oncology 2008 90 133 40 18618226 44 Marescalco MS Capizzi C Condorelli DF Barresi V Genome-wide analysis of recurrent copy-number alterations and copy-neutral loss of heterozygosity in head and neck squamous cell carcinoma Journal of oral pathology & medicine : official publication of the International Association of Oral Pathologists and the American Academy of Oral Pathology 2014 43 20 7 45 Ogawa R Ishiguro H Kuwabara Y Identification of candidate genes involved in the radiosensitivity of esophageal cancer cells by microarray analysis Diseases of the esophagus : official journal of the International Society for Diseases of the Esophagus/ISDE 2008 21 288 97 46 Wang SJ Cui HY Liu YM CD147 promotes Src-dependent activation of Rac1 signaling through STAT3/DOCK8 during the motility of hepatocellular carcinoma cells Oncotarget 2015 6 243 57 25428919 47 McDonald DR Massaad MJ Johnston A Successful engraftment of donor marrow after allogeneic hematopoietic cell transplantation in autosomal-recessive hyper-IgE syndrome caused by dedicator of cytokinesis 8 deficiency The Journal of allergy and clinical immunology 2010 126 1304 5 e3 20810158 48 Pai SY de Boer H Massaad MJ Flow cytometry diagnosis of dedicator of cytokinesis 8 (DOCK8) deficiency The Journal of allergy and clinical immunology 2014 134 221 3 24698323 49 Bittner TC Pannicke U Renner ED Successful long-term correction of autosomal recessive hyper-IgE syndrome due to DOCK8 deficiency by hematopoietic stem cell transplantation Klin Padiatr 2010 222 351 5 21058221 50 Gatz SA Benninghoff U Schutz C Curative treatment of autosomal-recessive hyper-IgE syndrome by hematopoietic cell transplantation Bone Marrow Transplant 2011 46 552 6 20622910 51 Metin A Tavil B Azik F Successful bone marrow transplantation for DOCK8 deficient hyper IgE syndrome Pediatr Transplant 2012 16 398 9 22248482 52 Boztug H Karitnig-Weiss C Ausserer B Clinical and immunological correction of DOCK8 deficiency by allogeneic hematopoietic stem cell transplantation following a reduced toxicity conditioning regimen Pediatr Hematol Oncol 2012 29 585 94 22897717 53 Ghosh S Schuster FR Fuchs I Laws HJ Borkhardt A Meisel R Treosulfan-based conditioning in DOCK8 deficiency: complete lympho-hematopoietic reconstitution with minimal toxicity Clinical immunology (Orlando, Fla) 2012 145 259 61 54 Cuellar-Rodriguez J Freeman AF Grossman J Matched related and unrelated donor hematopoietic stem cell transplantation for DOCK8 deficiency Biol Blood Marrow Transplant 2015 21 1037 45 25636378 55 Purcell C Cant A Irvine AD DOCK8 primary immunodeficiency syndrome Lancet 2015 386 982 25798541 56 Ghosh S Schuster FR Adams O Haploidentical stem cell transplantation in DOCK8 deficiency - Successful control of pre-existing severe viremia with a TCRass/CD19-depleted graft and antiviral treatment Clinical immunology (Orlando, Fla) 2014 152 111 4 57 Barlogis V Galambrun C Chambost H Successful allogeneic hematopoietic stem cell transplantation for DOCK8 deficiency The Journal of allergy and clinical immunology 2011 128 420 22 e2 21546070 58 Jing H Zhang Q Zhang Y Somatic reversion in dedicator of cytokinesis 8 immunodeficiency modulates disease phenotype The Journal of allergy and clinical immunology 2014 133 1667 75 24797421

PMC005xxxxxx/PMC5131799.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 0370534 520 Anal Chim Acta Anal. Chim. Acta Analytica chimica acta 0003-2670 1873-4324 27871605 5131799 10.1016/j.aca.2016.10.023 NIHMS824228 Article Neutral hydrophilic coatings for capillary electrophoresis prepared by controlled radical polymerization Navarro Fabian H. Gomez Jorge E. 1 Espinal Jose H. Sandoval Junior E. Department of Chemistry, Universidad del Valle, Cali, Colombia Correspondence: Professor Junior E. Sandoval, Department of Chemistry, Universidad del Valle, Calle 13 # 100-00, Cali, Colombia, junior.sandoval@correounivalle.edu.co, Phone: +572-321 2128, Fax: +572-339 3248 1 Present address: Department of Chemistry, Universidad Nacional de Colombia, Av. Carrera 30 # 45-03, Bogota, Colombia 26 10 2016 19 10 2016 15 12 2016 15 12 2017 948 104112 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. In the present study, porous silica particles as well as impervious fused-silica wafers and capillary tubes were modified with hydrophilic polymers (hydroxylated polyacrylamides and polyacrylates), using a surface-confined grafting procedure based on atom transfer radical polymerization (ATRP) which was also surface-initiated from α-bromoisobutyryl groups. Initiator immobilization was achieved by hydrosilylation of allyl alcohol on hydride silica followed by esterification of the resulting propanol-bonded surface with α-bromoisobutyryl bromide. Elemental analysis, IR and NMR spectroscopies on silica micro-particles, atomic force microscopy, ellipsometry and profilometry on fused-silica wafers, as well as CE on fused-silica tubes were used to characterize the chemically modified silica substrate at different stages. We studied the effect of monomer concentration as well as cross-linker on the ability of the polymer film to reduce electroosmosis and to prevent protein adsorption (i. e., its non-fouling capabilities) and found that the former was rather insensitive to both parameters. Surface deactivation towards adsorption was somewhat more susceptible to monomer concentration and appeared also to be favored by a low concentration of the cross-linker. The results show that hydrophilic polyacrylamide and polyacrylate coatings of controlled thickness can be prepared by ATRP under very mild polymerization conditions (aqueous solvent, room temperature and short reaction times) and that the coated capillary tubes exhibit high efficiencies for protein separations (0.3–0.6 million theoretical plates per meter) as well as long-term hydrolytic stability under the inherently harsh conditions of capillary isoelectric focusing. Additionally, there was no adsorption of lysozyme on the coated surface as indicated by a complete recovery of the basic enzyme. Furthermore, since polymerization is confined to the inner capillary surface, simple precautions (e.g., solution filtration) during the surface modification process are sufficient to prevent capillary clogging. Graphical Abstract ATRP grafting from hydride silica hydrosilylation hydroxylated monomers hydrolytic stability 1. Introduction Fused-silica capillary tubes used in GC and micro-bore LC have also found extensive use in the modern version of electrophoresis, CE(1, 2). In contrast with chromatography, chemical modification of the fused silica capillary has been aimed primarily at eliminating unwanted interactions between the inner wall of the CE tube and the sample undergoing separation. Many chemical modifications schemes used to modify silica-based chromatographic substrates have been readily extended to electrophoretic capillaries (3–5). Strictly speaking, the type and number of active surface functionalities on silica surfaces (silanols, siloxanes, etc.) depend on factors such as synthesis method, thermal history and the presence of humidity. Additionally, the irregular porous structure of silica micro-particles governs the accessibility of reagents to the active silica sites during surface modification. This kinetic factor aside, the chemistries of porous and flat silicas are essentially equivalent. It is also accepted that wafer surfaces are geometrically and chemically equivalent to the inner surface of fused-silica capillaries; they are both flat at the molecular level. Surface coverage resulting from organic groups grafted on a flat silica surface should be significantly denser than that of the same groups on a curved surface (particulate porous silica) since the later is sterically more constrained resulting in dissimilar group conformations (6). One of the greatest challenges in the current practice of CE is the requirement of surface deactivation of capillaries toward protein adsorption. Such unsolved bottleneck arises from the nonspecific interactions (ion exchange, hydrogen bonding, dispersion, etc.) between the inner capillary wall and the solute. Such unwanted interactions are responsible for excessive peak tailing, incomplete solute recovery and unreliable quantification. As a result, the high speed, separation efficiency, selectivity and versatility, minimal sample size requirement and easy automation of CE cannot be fully exploited in the case of many proteins. Researchers have suggested that a polymeric film that furnishes a hydrophilic, stable and adsorption-resistant (non-fouling) barrier between the capillary wall and the solute is required in protein analysis (7, 8). Such materials are also of great interest in many other fields such as medical sciences, biosensors, contact lenses technology, enzyme-based immuno assays, to name just a few (9). The attachment of polymers to solid surfaces via chemical bonding is an important strategy to modify the substrate properties in such a manner that the nonspecific interactions between the protein molecules and silica surface are minimized (7, 10). Anchoring of a polymer onto a silica surface has been achieved by two main strategies, the “grafting-to” and the “graftingfrom” methods. The grafting-to approach involves the attachment of a prefabricated polymer from solution, via the formation of a covalent bond between polymer active groups and matching groups on the substrate surface. Water-soluble polymers such as poly(ethylene glycol) and poly(vinyl alcohol) are typical examples of preformed polymers used in the grafting-to technique. Hydroxyl groups of the polymer react with surface-immobilized active species, such as glycidyl (11). Despite its experimental simplicity, the grafting-to strategy is limited by strong steric hindrance effects that worsen with increasing polymer size and eventually hinder contact of incoming polymer with surface reactive sites (10). In the graftingfrom approach, propagating polymer chains grow from surface-immobilized initiator groups. Among the various procedures for the grafting-from method, atom transfer radical polymerization (ATRP) is especially attractive for its remarkable control over the molar mass of the grafted polymer, great versatility (works well with a variety of functionalized monomers), compatibility with water and the possibility of mild polymerization temperatures (12–14). In the ATRP-based grafting-from approach, chain transfer and thermal self-initiation processes are essentially negligible and polymeric chains grow exclusively from the surface; i.e., the process is both surface-initiated and surface-confined. The initiator moieties are most usually anchored to the silica surface via silane coupling chemistry. Halogenated ATRP initiator groups, such as benzyl chloride or, to a much greater extent, 2-bromo-isobutanoyl have been used in the past to grow neutral hydrophilic films on fused silica capillaries. Wirth and her research group reported the first covalent bonding of a polymer film for CE by ATRP in 1998 (15). Surprisingly, very few papers have been published about ATRP applied to CE since then (16–18). Although acrylamide has been the most commonly used monomer to make polymeric coatings for CE (4, 7), the limited stability at moderately high pH of polyacrylamide has been a well-known fact from traditional slab gel electrophoresis (19). The slow deamidation of the polymer under this condition leads to a considerable deterioration of the coating, evidenced by the formation of fully-dissociated carboxylic groups that cause strong electroosmosis, polymeric layer swelling and analyte band distortion. The use of N-substituted acrylamide derivatives such as N-acryloylamino-ethoxyethanol, has resulted in coated capillaries with superior resistance to hydrolysis and hence improved long-term CE separations at alkaline pH (20). N-acryloyl-aminoethanol (AAE) (21–23) and its relative N-acryloyl-aminopropanol (AAP) (24, 25), also N-substituted acrylamide derivatives, should provide durable coatings as well. It appears that the N-substitution with hydroxyl-terminated chains produces polymeric coatings that are not only hydrolytically more stable, but have also higher hydrophilicity compared to polyacrylamide (19, 24, 25). When it comes to electrophoretic performance, hydrophilicity turns out to be of paramount importance since a high hydrophilicity of the polymer effectively precludes proteins to compete with water for its potentially adsorptive sites (26). It has been suggested that any monomer less hydrophilic than acrylamide should be considered unsuitable to produce a good quality gel “since acrylamide is itself already at the border-line between hydrophilicity and hydrophobicity” (25). There is a definite need for hydrophilic coatings of improved hydrolytic stability that enable the use of CE at its full potential, and such improved materials are frequently targeted at protein separations. In the present work we explore the combination of several promising synthetic schemes to modify the fused-silica surface of capillaries for a lasting resistance to protein adsorption. More specifically, our work attempts to bring together the best of three worlds: (i) a stable anchorage of the polymerization initiating group to the inner wall of the capillary tube by means of Si–C linkages formed by hydrosilylation; (ii) a stable polymeric film whose strength arises from the N-substitution on acrylamide; and (iii) a surface-confined in-situ polymerization method (ATRP) that prevents clogging of the capillary tube while permitting an easy control of the coating thickness. 2. Materials and Methods 2.1 Instrumentation IR spectroscopy was performed on a Shimadzu, Model FTIR-8400 spectrometer (Columbia, MD, U.S.) equipped with a diffuse reflectance infrared Fourier transform (DRIFT) accessory. Solid state NMR characterization was carried out using a Bruker Advance II-400 MHz NMR spectrometer (Rheinstetten, Germany) equipped with a Bruker MAS II probe. Electrophoretic separations were performed on Agilent model 7100 or 1600 CE Systems (Palo Alto, CA, U.S.). Atomic force microscopy (AFM) images were obtained with an Asylum Research model MFP-3D-SA (Santa Barbara, CA, U.S.) instrument. Carbon percent measurements were carried out on a FlashEA 1112 elemental analyzer Thermo Inc. (Waltham, MA, U.S.). 2.2 Materials Fused-silica capillaries (50-μm id, 360-μm od and 20-μm polyimide coating) were purchased from Biotaq Inc. (Silver Springs, MD, U.S.) and Polymicro Technologies (Phoenix, AZ, U.S.). UV-grade fused-silica wafers (25.4 × 9.0 × 1.0 mm) were purchased from Laser Optex Inc. (Beijin, China). Silica micro-particles (Nucleosil, 7-μm diameter and 91.0-m2 g−1 surface area, and YMC Gel, 10-μm diameter and 288-m2 g−1 surface area) were obtained from Macherey-Nagel (Duren, Germany) and YMC Co. (Kioto, Japan) respectively. Piperazine-N, N′-bis(3-propanesulfonic acid) (PIPPS) was purchased from GFS Chemicals (Columbus, OH, U.S.). Nacryloyl-aminoethanol, N-acryloyl-aminopropanol, 2-hydroxyethyl methacrylate, α-bromoisobutyryl bromide, tris[2-(dimethylamino)ethyl]amine (Me6TREN), bipyridine (bpy), allyl alcohol, platinum(0)-1,3-divinyl-1,1,3,3-tetramethyldisiloxane (Karstedt’s catalyst, ~2 % Pt in xylene), hexachloroplatinic acid, cuprous chloride, 2,5-di-tert-butylhydroquinone, potassium cyanide, EDTA, 3-(N-morpholino)propanesulfonic acid (MOPS), methylcellulose, Ampholytes pH 3-10 and pH 6–8 (40% aqueous solutions) and DMSO were purchased from Sigma-Aldrich (St. Louis, MO, U.S.). Analytical grade solvents were obtained from various vendors; they were freshly distilled from sodium shavings before use. 2.3 Surface modification Prior to modification, fused-silica capillaries (typically, a 6.0-m length) and wafers were conditioned by etching with NaOH followed by leaching with HCl, according to procedures described elsewhere (27). Porous silica particles did not undergo any conditioning. Hydride silica substrates were prepared as described previously (28). 3-hydroxypropyl (propanol) bonded silica phase was prepared by hydrosilylation of allyl alcohol on hydride silica, as described in a recent report (29). Esterification of the bound propanol groups was carried out by treatment for 6 h at room temperature (20 ± 2 °C) with a solution containing 0.25 M of α-bromoisobutyryl bromide and 0.25 M of pyridine in dry DMF under a nitrogen blanket (30–32). This solution was prepared in a nitrogen-purged glove bag. The product was sequentially rinsed with 10%v/v water in DMF and THF, and then dried under nitrogen. Polymerization of 2-hydroxyethyl acrylate (HEA), AAE and AAP was catalyzed with CuI(M6TREN)+Cl− in 4:1 v/v ethanol/water solvent (23). 2-Hydroxyethyl methacrylate (HEMA) was polymerized with CuI(bpy)2+Cl− in 1:1 v/v methanol/water solvent (33). For a typical polymerization in a capillary, a solution containing 1.0 M monomer and 20 mM Cu(I) catalyst in alcohol/water solvent was passed through the column for 3 h at room temperature (18, 30). This solution was prepared with helium-degassed solvent in a nitrogen-purged glove bag. At the reaction end, the column was sequentially rinsed with solvent and water for 30 min each. The column was emptied with nitrogen gas for a few minutes. Fused-silica wafers were modified by immersion in the reagent solution with gentle shaking, while 65 mg of porous silica were used per mL of reagent solution with magnetic stirring. 2.4 Spectroscopy and microscopy DRIFT, solid state NMR and AFM measurements were obtained accordingly to previously described procedures (28). An Uvisel 2 Spectroscopic Phase Modulated ellipsometer was used for film thickness measurements at Horiba Scientific laboratories (France). Perfilometry measurements were carried out with a KLA Tencor D-120 stylus profiler (U.S.) at the Materials Engineering School of Universidad del Valle. 2.5 Surface coverage The concentration, ΓR, of surface-bonded R-groups on porous silica was obtained from elemental analysis of the bonded material along with the BET specific surface area of the hydride intermediate substrate (34). 2.6 CE experiments Prior to use, capillaries were conditioned with the electrolyte (typically, 25 mM PIPPS, adjusted to pH 4 with NaOH) under high pressure (7 bar) for 20 min, after which a stable electroosmotic mobility (EOM) value was obtained (27). EOM determinations followed methodology previously described in the literature (35, 36). CIEF separation conditions were adapted from a method devised by Hempe et al. (37, 38). 3. Results and discussion 3.1 Silica substrate modification The synthetic scheme used to chemically modify the inner capillary surface with a polymeric film is outlined in Figure 1. Initiator immobilization is accomplished by hydrosilylation of allyl alcohol on hydride silica followed by esterification of the resulting propanol-bonded surface with α-bromoisobutyryl bromide. The first step is based on the formation of a stable Si–C bond by the catalytic addition of silicon hydride to an olefin (hydrosilylation). We have recently demonstrated that allyl alcohol can also be anchored to the hydride silica surface and that it does so with a high coverage (3–5 μmol m−2) (29). In the esterification of the anchored alcohol groups with α-bromoisobutyryl bromide, we found that using DMF as solvent and pyridine as HBr-binding base maximize the solubilization of the bromide salts produced during the esterification process. This is important to minimize clogging of the tube as the reaction is performed on fused-silica capillaries. When this two-step procedure is applied on porous silica micro-particles, ligand density information from the primary (hydrosilylation) as well as the secondary (esterification) modifications can be readily estimated from percent carbon contents along with specific surface area and structural information (molar mass and number of carbon atoms) of the anchored groups (34). Our results indicate that a typical propanol-modified substrate with a surface coverage of 3.5 μmol m−2 (2.1 chain nm−2) produced an α-bromoisobutyryl coverage of 1.1 μmol m−2 (0.75 chain nm−2), which means that about 30% of the surface propanol groups were esterified. One can think of the resultant functionalized silica surface as consisting of active α-bromoisobutyryl initiator moieties surrounded by a dense layer of inactive propanol groups. Naturally, denser surface coverages of the α-bromoisobutyryl groups should be expected on a flat silica surface in comparison to the same groups on a curved surface such as that of porous silica. Since the average cross sectional area of a polymer chain (~1.8–2.0 nm2) is much larger than that of a typical initiator group (~0.20 nm2) (39, 40), our relatively low α-bromoisobutyryl coverage (1.5 nm2 chain−1) should provide the footprint surface area required for a dense polymer grafting quite well, as described below. Previous studies show that such modest α-bromoisobutyryl group density favors an efficient polymer growth (41). When we applied our ATRP-based grafting-from approach to several hydroxylated monomers–namely, HEA, HEMA, AAE, and AAP– efficient polymerization took place. Monomer structures are depicted in Figure S-1 of the Supplementary Material. It is important to point out that we used mild conditions for the polymerization reaction: aqueous solvent, room temperature and short reaction times, as suggested by Armes (31, 33) and Kakuchi (23). These conditions represent a significant departure from those proposed by Wirth in her original report (100°C, dry DMF solvent, and 40-h reaction time) (15). Undoubtedly, the milder conditions are more attractive while still favoring good grafting density of the polymer. Furthermore, we used a much more common and (about 500 times) more active α-bromoisobutyryl group as initiator. Figure 2 shows IR spectroscopic evidence that confirms polymerization from the immobilized α-bromoisobutyryl initiating groups. A substantial increase of stretching bands in the 2950–2850 cm−1 region is observed, which is indicative of symmetric and asymmetric aliphatic C-H stretching, along with the appearance of characteristic C=O stretching bands at 1655 and 1730 cm−1 for amides (curves B and C) and esters (curves D and E) respectively. Most likely, the bands at 3085 and 1550 cm−1 (curves B and C) can be assigned to the amido N–H stretching and bending modes respectively. Furthermore, 13C CP-MAS NMR spectra clearly verifies chemical bonding of the polymeric film to the silica surface, as shown in Figures S-2 to S-4 of the Supplementary Material. To further characterize the polymeric film on the silica surface, AFM was used on coated silica plates by probing the surface after polymerization. As shown in Figure 3, the surface is uniformly coated with the polymeric film and exhibits a more rugose surface topography (rugosity = 3.16 nm), in comparison to an unmodified plate (~ 0.3 nm)(28). Thickness measurements with ellipsometry revealed a uniform film with a layer thickness of 53 ± 2 nm. Additionally, profilometry measurements provide a thicknesses of 59 ± 8 nm, which is quite consistent with that from ellipsometry. One intrinsic feature of ATRP-based grafting-from process is that the polymerization is confined to the substrate surface. Since free initiator is absent in the monomer/catalyst mix, no polymer is formed in the bulk of the solution, avoiding blockage of the capillary column. Additionally, a simple rinsing procedure provides adequate film cleanup. These details are of critical importance in the preparation of chemically modified CE tubes. 3.2 Electrophoretic performance Chemical modification of CE tubes is accompanied by a drastic drop of electroosmosis, the electrokinetic bulk flow of electrolyte originated at the wall/solution boundary. We evaluated the effect of monomer concentration on the ability of the polymer film to reduce electroosmosis and it was found to be very small, as expected, and essentially insensitive to variations of monomer concentration. This is shown in the Supplementary Material for polyAAE, polyHEA and polyHEMA coatings (Figure S-5). We next examined the CE performance of polymer-coated capillary tubes with respect to their ability to prevent protein adsorption (i. e., its non-fouling capabilities). Peak asymmetry and separation efficiency of two strong silanol-sensitive compounds, Ru(bpy)32+ and lysozyme, were used as indicators of the extent of adsorption on the coated capillaries upon varying monomer concentration. The Ru-complex is very sensitive to electrostatic interactions with silica surfaces, while lysozyme is also sensitive to other nonspecific interactions such as hydrogen bonding, hydrophobic effects, etc. The effect of increasing monomer concentration is evident in the case of polyAAE (1–2 M-levels appear to favor low asymmetry values and high plate count) and essentially absent in the case of polyHEMA (This is shown in Figure S-6 of the Supplementary Material). In the case of polyHEA, a concentration ≥ 2 M appears to favor high plate counts. Similar observations, albeit to a lesser extent, apply to the Ru probe (not shown). We also examined the effect of cross-linker concentration on polymer film behavior (chemical structures of the cross-linkers used are also depicted in Figure S-1 of the Supplementary Material) and found that, similarly to monomer concentration, there is no evident effect on EOM (see Figure S-7 of the Supplementary Material). On the other hand, low concentration of cross-linker (0.1% molar) appear to furnish lower asymmetry and higher plate count in the case of polyAAE, while there was very little or slightly detrimental effect for polyHEA and polyHEMA (see Figure S-8 of the Supplementary Material). Results are summarized in Figure 4, where it appears that, generally speaking, the presence of cross-linker improves the symmetry, separation efficiency and time-corrected area of lysozyme when the amide monomers, AAE and AAP, are used. In contrast, when the two ester monomers, HEA and HEMA, are used, the presence of cross-linker appears to be somewhat detrimental to the electrophoretic performance of the basic enzyme. It should be pointed out that there does not seem to be a most favorable monomer among the ones studied. PolyAAE exhibits the largest corrected peak area, which suggests maximum solute recovery. Cross-linked polyAAP shows the greatest separation efficiency, while uncross-linked HEMA has a remarkably low peak asymmetry. Several proteins with high pI were separated to test the capability of the polymeric coating to resist adsorption. Figure 5 shows a typical electropherogram for a mixture of four of these proteins (cytochrome c, lysozyme, trypsin and ribonuclease A) at pH 4.0, with separation efficiencies comparable to those of former reports (15). Whereas separation efficiencies are very high (typically 0.25 to 0.65 million plates per meter; see Table S-1 of the Supplementary Material), CE performance at its full potential –a few million plates per meter– seems to be elusive, and some residual fouling appears to be an inevitable limitation of our hydrophilic coatings at this stage of development. The graphical abstract shows another example of the great selectivity obtained for two cytochromes C on a polyAAE-Bis modified capillary under the same CE conditions used in Figure 5. To further complement our observations, such residual adsorption can be quantitatively assessed by analyte recovery measurements. This is accomplished by linear fitting of logarithmic peak area vs. effective length data (42). For a given tube id the slope, k, is proportional to the surface density of adsorption sites and, hence, is a useful figure to comparatively assess protein adsorption on different capillary coatings. Preliminary experiments, summarized in Table 1, show that the k-values for both polyHEMA and polyAAP coatings are statistically undistinguishable from zero which, in turn, indicates a complete (100%) recovery of the basic enzyme; i.e., there is no adsorption on the coated surface. In contrast, the oligoamine (tetraethylenepentamine) used to dynamically coat a native capillary, while quite effective to reduce lysozyme adsorption under the same CE conditions, exhibits some measurable analyte lost during its passage through the capillary tube as indicated by a k-value which is statistically different from zero. With regards to precision, while the repeatability of migration times was 0.2-0.7 %RSD, the typical peak area precision was 5-13 %RSD (n = 10 successive injections). The later is likely due to the difficulties associated with the various isoforms that accompany the enzyme samples, as illustrated in Figure 5. 3.3 Long-term hydrolytic stability We have chosen the very demanding electrolyte conditions of capillary isoelectric focusing (CIEF) to assess long-term hydrolytic stability of our polymeric coatings. While we used a hydrolytically durable N-substituted amide as the starting monomer along with a stable Si–C link to the silica surface, when a sodium hydroxide solution is used as cathode electrolyte –as it is done in CIEF, the underlying siloxane bridges that hold the silica backbone will eventually collapse. Although to a lesser extent, the phosphoric acid solution used as anode electrolyte will also be very aggressive towards the polymeric coating. Figure 6 shows typical CIEF separation profiles of blood samples on a polyAAP-Bis coated capillary tube. The two upper panels of Figure 7 illustrate measured migration time and area for two hemoglobin (Hb) peaks of a hemolyzed blood sample mix as a function of injection number during a reiterating CIEF experiment. Within the experimental error of this several-day-long experiment, peak migration time and, to a lesser extent, corrected peak area do not seem to deteriorate significantly. It should be pointed out that there is a certain trend associated with each 10-injection cycle (we renewed the sample after 10 consecutive injections) which contributes to the overall variability and may mask the long-term profile of the parameter under study. This effect of sample renewal is more pronounced for the peak area (mid panel of Figure 7) and for polyAAE coatings (not shown). Long-term precision for Hb S and A0 was about 5 and 23 %RSD for migration times and time-corrected peak areas respectively (n = 260). Regression analysis on the data (see Table S-2 of the Supplementary Material) reveals nearly zero slopes along with low correlation coefficients. The very small slope of, for instance, migration times vs. run number indicates that it would take an average of 15 and 45 runs to decrease by one second the migration times of Hb S and A0 respectively. Figure S-11 of the Supplementary Material visually corroborates the fact that the Hb separation profile does not deteriorate significantly after many sample injections. Perhaps more important from a stability point of view, the hydrolytic deterioration of a CE coating ultimately leads to an increased exposure of ionizable silanol groups and, hence, greater EOM. The impact of a prolonged exposure of a coated capillary to the harsh conditions prevailing in CIEF should provide an “ageing” curve that assesses its hydrolytic stability. Indeed, the EOM appears to be more revealing with regard to the effect of coating exposure to the harsh conditions prevailing in CIEF (see lower panel of Figure 7): precision of EOM stars declining after 200 runs and its value definitely increases above 260 runs. Yet, acceptable band profiles are obtained even after 270 runs, as depicted in Figure S-11 of the Supplementary Material. Unfortunately, comparable ageing experiments are not easy to find in the CIEF literature probably because these tests are expensive in terms of effort and time. 4. Conclusions and future perspectives In this report, we describe surface-initiated and -confined ATRP methodology for chemically modifying silica surfaces with low-fouling polymer coatings. The results indicate that hydroxylated polyacrylamides and polyacrylates coatings of controlled thickness can be prepared by ATRP under very mild polymerization conditions (aqueous solvent, room temperature and short reaction times) and that the coated capillary tubes exhibit high performance for protein separations along with remarkable hydrolytic stability which affords favorable long-term reproducibility. Since polymerization is confined to the inner capillary surface, clogging of the narrow tube conduit becomes very unlikely if simple precautions are taken during the surface modification process. The controlled polymerization method can be easily extended not only to other hydrophilic monomers such as N-acryloylamino-ethoxyethanol (20, 43), but also to betain-related (e.g., sulfobetaineacrylamide) monomers that are known to form zwitterionic polymer surfaces that exhibit high resistance to nonspecific protein adsorption (44). These families of N-substituted acrylamide monomers should provide a step ahead towards the solution to the need for non-adsorbing coatings that enable the use of CE at its full potential, particularly for protein separations. Although capable of efficiently initiating a variety of polymer chains, immobilization of the α-bromoisobutyryl group did not occur with the high yields expected from hydrosilylation and esterification (see Figure 1). Clearly, further work is needed to achieve a better control of silica surface coverage. There is also a concern regarding the relatively poor stability of the ester functionality at the opposite end of the chain containing the α-bromoisobutyryl initiator group (see the final structure of the sequence depicted in Figure 1). Such labile ester group is in clear contrast with the highly stable Si–C bond (formed by hydrosilylation) which fastens the propyl chain to the inner capillary wall, and with the polymeric film itself –a stable N-substituted polyacrylamide. It is also very important to simplify the synthetic scheme where immobilization of the initiator α-bromoisobutyryl group takes two steps. Our early attempts to anchor initiator groups to the hydride silica via the direct hydrosilylation of 4-vinylbenzyl chloride as well as allyl α-bromoisobutyrate (both commercially available) were unsuccessful, as described in the last section of the Supplementary Material. We are currently working on a new procedure to anchor the α-bromoisobutyryl initiator by means of a single-step reaction, whose results will be the subject of another report. Supplementary Material supplement This work was supported by the National Institute of General Medical Sciences of the National Institutes of Health [grant number R01GM089759], and Universidad del Valle [project number CI-7832]. The content is solely the responsibility of the authors and does not necessarily represent the official views of the supporters. Abbreviations EOM electroosmotic mobility ATRP atom transfer radical polymerization AFM atomic force microscopy AAE N-acryloyl-aminoethanol AAP N-acryloyl-aminopropanol HEMA 2-hydroxyethyl methacrylate HEA 2-hydroxyethyl acrylate Me6TREN tris[2-(dimethylamino)ethyl]amine bpy bipyridine PIPPS piperazine-N, N′-bis(3-propanesulfonic acid) CIEF capillary isoelectric focusing MOPS 3-(N-morpholino)propanesulfonic acid Hb hemoglobin Figure 1 Schematic diagram illustrating the hydrosilylation of allyl alcohol on the Si–H surface, followed by esterification of the propanol-functionalized silica surface to immobilize the α-bromoisobutyryl initiator, and then by surface-initiated and -confined polymerization of N-acryloyl-aminopropanol via ATRP. Figure 2 DRIFT spectra of porous silica particles (Nucleosil) functionalized with (A) α-bromoisobutyryl initiator, and the polymerization products of (B) AAP, (C) AAE, (D) HEA and (E) HEMA monomers. Monomer concentrations, 1.0 M; catalyst concentrations, 20.0 mM; temperature, 25°C; reaction time, 3 h. Catalyst and solvent (B) through (D), [CuI(Me6TREN)]+Cl− and ethanol/water 4:1 v/v, (E), [CuI(bpy)2]+Cl− and methanol/water 1:1 v/v. Figure 3 Typical AFM image of a silica plate modified with a polyAAE film. The plate as probed in the intermittent noncontact mode with a force constant of 42 N/m and resonance frequency of 300 kHz. A rectangular silicon cantilever with tetrahedral tip and nominal spring constant of 0.06 N/m was used. The scan size was 3 μm × 3 μm. Preparation conditions as described in Figure 2. Figure 4 Comparison of lysozyme CE peak features on several polymeric coatings with and without their corresponding cross-linkers. Error bars represent ± 1 SD (n = 3). ATRP conditions equivalent to those in Figure 2. Cross-linker concentration, 0.1% molar with respect to the monomer. CE conditions: capillary, 35.0 cm (effective length, 26.5 cm); electric field, 400 V cm−1 (9.8 μA); electrolyte, 25.0 mM PIPPS buffer with pH 4.0; protein concentration, 0.5 mg mL−1 each in buffer; injection, 3 s at 50 mbar; detection, 210 nm. Figure 5 CE separation of a mixture of basic proteins in a polyAAE-coated capillary. Conditions: capillary, 35.0cm (effective length, 26.5 cm); electric field, 400 V cm−1 (8.1 μA); electrolyte, 25.0 mM PIPPS buffer with pH 4.0; protein concentration, 0.1 mg mL−1 each in 10-fold diluted buffer; injection, 15 s at 20 mbar; detection at 200 nm. Solutes: (a) cytochrome C, equine, (b) lysozyme, chicken, (c) trypsin, bovine, (d) α-chymotrypsinogen A, porcine. Figure 6 Hemoglobin constituents in normal blood (left), and blood from a patient with sickle cell disease by CIEF. Whole blood (10 μL) was treated with 200 μL of hemolyzing solution (EDTA, 5 mM; KCN, 10 mM), and then introduced into the capillary by low-pressure injection (30 mbar for 17 sec). Cathode and anode solutions were 20 mM NaOH and 100 mM H3PO4 in 0.375% methylcellulose, respectively. Prior to each assay, the capillary was flushed (5 min under 0.95 bar) with electrolyte containing 0.375% methylcellulose and 2.5% v/v Ampholyte mix solution (pH 6–8 / pH 3–10, 10/1 v/v). Sample constituents were focused for 4 min at 25 kV, then eluted past the detector (UV absorbance at 415 nm) under simultaneous low pressure (20 mbar) and same voltage. Capillary: Bis-crosslinked polyAAP coating, 50-μm id and 35.0-cm length (26.5-cm effective length). Figure 7 Long-term stability test under typical CIEF conditions for a Bis-crosslinked polyAAP coating immobilized on a 50-μm id capillary with 35.0-cm length (26.5-cm effective length). Experimental CIEF conditions were the same described in Figure 6, except that the sample was a 1:1 v/v mixture of hemolysates from same normal and abnormal blood samples. 10 consecutive injections per sample aliquot were carried out from the same hemolysis lot. Electroosmotic mobility (EOM) was measured after every 10 CIEF runs. CE conditions for EOM measurements: hydrodynamic injections (3 s at 50 mbar) of 10 mM DMSO (neutral marker) in the electrolyte solution (25-mM MOPS buffer, pH 7.0), 28.0 kV voltage applied (18 μA) for 5.0 min, hydrodynamic mobilization at 50 mbar, detection at 215 nm, 1 min electrolyte rinse between runs. Error bars represent ± 1 SD (n = 3). Table 1 CE recovery of lysozyme as a function of the nature of the coating grafted on the inner wall of the capillary (42). coating k, cm−1 a sk %Recovery b (l=26.5 cm) s%recovery polyHEMA c 0.0015 0.0092 96 23 polyAAP d 0.0029 0.0044 93 11 none (native) e 0.0055 0.0018 86 4 a Estimated from linear fitting of the logarithm of lysozyme peak area (referred to the internal standard) vs. effective length, l. b Estimated from the equation R = exp(-kl) with R as the analytical recovery of lysozyme and 26.5 cm being the minimum possible l in our instrument. c Uncross-linked. Other preparation conditions as described in Figure 2. d Cross-linked with 0.10% molar Bis with respect to the monomer. e Native tube etched with 1.0 M NaOH, leached with 1:1 v/v HCl and conditioned with electrolyte. CE conditions: electrolyte, 25.0 mM PIPPS buffer adjusted to pH 4.00 ± 0.05 with 12.5 mM γ-aminobutyric acid; also containing 1.5 mM tetraethylenepentamine in the case of the native capillary tube. Sample: 1.0 mg mL−1 lysozyme and 3.0 mM histidine as internal standard in electrolyte (10.0 mM benzylamine as internal standard with the native capillary tube). Applied field, 400 V cm−1 (current, 5.8 μA). Five replicate injections (5 s at 0.67 mbar cm−1) were performed on a 50-μm capillary tube with varying effective length. Highlights A controlled radical polymerization scheme is devised for stable, non-fouling CE coatings. Polymers are surface-initiated from α-bromoisobutyryl groups under mild conditions. High separation efficiency and complete solute recovery are obtained for lysozyme. Clog-free capillaries inherently result from the surface-confined grafting method. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. 1 Jorgenson JW Lukacs KD High-resolution separations based on electrophoresis and electroosmosis Journal of Chromatography A 218 209 216 1981 2 Hjerten S High-performance electrophoresis: the electrophoretic counterpart of highperformance liquid chromatography Journal of Chromatography A 270 1 6 1983 3 Buchmeiser MR New synthetic ways for the preparation of high-performance liquid chromatography supports Journal of Chromatography A 918 233 266 2001 11407572 4 Doherty EAS Meagher RJ Albarghouthi MN Barron AE Microchannel wall coatings for protein separations by capillary and chip electrophoresis ELECTROPHORESIS 24 34 54 2003 12652571 5 Pallandre A de Lambert B Attia R Jonas AM Viovy J-L Surface treatment and characterization: Perspectives to electrophoresis and lab-on-chips ELECTROPHORESIS 27 584 610 2006 16400705 6 Unger KK Porous silica: Its properties and use as support in column liquid chromatography Unger KK Journal of Chromatography Library Elsevier 1979 16 7 Horvath J Dolnik V Polymer wall coatings for capillary electrophoresis ELECTROPHORESIS 22 644 655 2001 11296918 8 El Rassi Z Electrophoretic and electrochromatographic separation of proteins in capillaries: An update covering 2007–2009 ELECTROPHORESIS 31 174 191 2010 20039288 9 Banerjee I Pangule RC Kane RS Antifouling coatings: Recent developments in the design of surfaces that prevent fouling by proteins, bacteria, and marine organisms Advanced Materials 23 690 718 2011 20886559 10 Zhao B Brittain WJ Polymer brushes: surface-immobilized macromolecules Progress in Polymer Science 25 677 710 2000 11 Zhao Z Malik A Lee ML Adsorption on polymer-coated fused-silica capillary electrophoresis columns using selected protein and peptide standards Analytical Chemistry 65 2747 2752 1993 8250261 12 Matyjaszewski K Xia J Atom Transfer Radical Polymerization Chemical Reviews 101 2921 2990 2001 11749397 13 Barbey R Polymer Brushes via Surface-Initiated Controlled Radical Polymerization: Synthesis, Characterization, Properties, and Applications Chemical Reviews 109 5437 5527 2009 19845393 14 Matyjaszewski K Atom Transfer Radical Polymerization (ATRP): Current Status and Future Perspectives Macromolecules 45 4015 4039 2012 15 Huang X Doneski LJ Wirth MJ Surface-Confined Living Radical Polymerization for Coatings in Capillary Electrophoresis Analytical Chemistry 70 4023 4029 1998 21651239 16 Huang X Wirth MJ Surface Initiation of Living Radical Polymerization for Growth of Tethered Chains of Low Polydispersity Macromolecules 32 1694 1696 1999 17 Feldmann A Clausnitzer U Otto M Optimization of capillary coating by hydroxyethyl methacrylate for capillary zone electrophoresis of proteins Journal of Chromatography B 803 149 157 2004 18 Miller MD Baker GL Bruening ML Polymer-brush stationary phases for opentubular capillary electrochromatography Journal of Chromatography A 1044 323 330 2004 15354455 19 Righetti PG Gelfi C Electrophoresis gel media: the state of the art Journal of Chromatography B: Biomedical Sciences and Applications 699 63 75 1997 9392368 20 Chiari M Nesi M Sandoval JE Pesek JJ Capillary electrophoretic separation of proteins using stable, hydrophilic poly(acryloylaminoethoxyethanol)-coated columns Journal of Chromatography A 717 1 13 1995 21 Albarghouthi MN Stein TM Barron AE Poly-N-hydroxyethylacrylamide as a novel, adsorbed coating for protein separation by capillary electrophoresis ELECTROPHORESIS 24 1166 1175 2003 12707908 22 Zhao C Zheng J Synthesis and Characterization of Poly(N-hydroxyethylacrylamide) for Long-Term Antifouling Ability Biomacromolecules 12 4071 4079 2011 21972885 23 Narumi A Poly(N-hydroxyethylacrylamide) Prepared by Atom Transfer Radical Polymerization as a Nonionic, Water-Soluble, and Hydrolysis-Resistant Polymer and/or Segment of Block Copolymer with a Well-Defined Molecular Weight Macromolecular Chemistry and Physics 210 349 358 2009 24 Simo-Alfonso E Gelfi C Sebastiano R Citterio A Righetti PG Novel acrylamido monomers with higher hydrophilicity and improved hydrolytic stability: I. Synthetic route and product characterization ELECTROPHORESIS 17 723 731 1996 8738335 25 Simo-Alfonso E Gelfi C Sebastiano R Citterio A Righetti PG Novel acrylamido monomers with higher hydrophilicity and improved hydrolytic stability: II. Properties of N-acryloylaminopropanol ELECTROPHORESIS 17 732 737 1996 8738336 26 Chen S Li L Zhao C Zheng J Surface hydration: Principles and applications toward low-fouling/nonfouling biomaterials Polymer 51 5283 5293 2010 27 Gomez JE Sandoval JE The effect of conditioning of fused-silica capillaries on their electrophoretic performance ELECTROPHORESIS 29 381 392 2008 18080253 28 Gomez JE Sandoval JE New Approaches to Prepare Hydride Silica Analytical Chemistry 82 7444 7451 2010 20712322 29 Gomez JE Navarro FH Sandoval JE Novel 3-hydroxypropyl-bonded phase by direct hydrosilylation of allyl alcohol on amorphous hydride silica ELECTROPHORESIS 35 2579 2586 2014 24934906 30 Matyjaszewski K Polymers at Interfaces: Using Atom Transfer Radical Polymerization in the Controlled Growth of Homopolymers and Block Copolymers from Silicon Surfaces in the Absence of Untethered Sacrificial Initiator Macromolecules 32 8716 8724 1999 31 Wang XS Armes SP Facile Atom Transfer Radical Polymerization of Methoxy-Capped Oligo(ethylene glycol) Methacrylate in Aqueous Media at Ambient Temperature Macromolecules 33 6640 6647 2000 32 Ohno K Morinaga T Koh K Tsujii Y Fukuda T Synthesis of Monodisperse Silica Particles Coated with Well-Defined, High-Density Polymer Brushes by Surface-Initiated Atom Transfer Radical Polymerization Macromolecules 38 2137 2142 2005 33 Robinson KL Khan MA de Paz Banez MV Wang XS Armes SP Controlled Polymerization of 2-Hydroxyethyl Methacrylate by ATRP at Ambient Temperature Macromolecules 34 3155 3158 2001 34 Sandoval JE Equation for calculating surface coverage from end-capping of chromatographic bonded phases Journal of Chromatography A 852 375 381 1999 10481976 35 Williams BA Vigh G Fast, Accurate Mobility Determination Method for Capillary Electrophoresis Analytical Chemistry 68 1174 1180 1996 21619150 36 Sandoval JE Chen S-M Method for the Accelerated Measurement of Electroosmosis in Chemically Modified Tubes for Capillary Electrophoresis Analytical Chemistry 68 2771 2775 1996 21619350 37 Hempe JM Granger JN Craver RD Capillary isoelectric focusing of hemoglobin variants in the pediatric clinical laboratory ELECTROPHORESIS 18 1785 1795 1997 9372271 38 Hempe JM Craver RD Separation of hemoglobin variants with similar charge by capillary isoelectric focusing: Value of isoelectric point for identification of common and uncommon hemoglobin variants ELECTROPHORESIS 21 743 748 2000 10733215 39 Kim J-B Bruening ML Baker GL Surface-Initiated Atom Transfer Radical Polymerization on Gold at Ambient Temperature Journal of the American Chemical Society 122 7616 7617 2000 40 Jones DM Brown AA Huck WTS Surface-Initiated Polymerizations in Aqueous Media: Effect of Initiator Density Langmuir 18 1265 1269 2002 41 Ma H Surface Initiated Polymerization from Substrates of Low Initiator Density and Its Applications in Biosensors ACS Applied Materials & Interfaces 2 3223 3230 2010 20942448 42 Espinal JH Gomez JE Sandoval JE Closer look at the operating definition of protein recovery in CE ELECTROPHORESIS 34 1141 1147 2013 23400851 43 Chen H Synthesis and Characterization of Antifouling Poly(Nacryloylaminoethoxyethanol) with Ultralow Protein Adsorption and Cell Attachment Langmuir 30 10398 10409 2014 25127733 44 Schlenoff JB Zwitteration: Coating Surfaces with Zwitterionic Functionality to Reduce Nonspecific Adsorption Langmuir 30 9625 9636 2014 24754399

PMC005xxxxxx/PMC5131882.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 101130617 29778 Cancer Cell Cancer Cell Cancer cell 1535-6108 1878-3686 27505672 5131882 10.1016/j.ccell.2016.07.004 NIHMS829253 Article MITOCHONDRIAL Akt REGULATION OF HYPOXIC TUMOR REPROGRAMMING Chae Young Chan 1 Vaira Valentina 234 Caino M. Cecilia 1 Tang Hsin-Yao 4 Seo Jae Ho 1 Kossenkov Andrew V. 5 Ottobrini Luisa 46 Martelli Cristina 4 Lucignani Giovanni 78 Bertolini Irene 34 Locatelli Marco 9 Bryant Kelly G. 1 Ghosh Jagadish C. 1 Lisanti Sofia 1 Ku Bonsu 10 Bosari Silvano 34 Languino Lucia R. 11 Speicher David W. 512 Altieri Dario C. 1 1 Prostate Cancer Discovery and Development Program, Tumor Microenvironment and Metastasis Program, The Wistar Institute, Philadelphia, PA 19104, USA 2 Istituto Nazionale Genetica Molecolare “Romeo and Enrica Invernizzi”, Milan 20122, Italy 3 Division of Pathology, Fondazione IRCCS Ca’ Granda Ospedale Maggiore Policlinico, Milan 20122, Italy 4 Department of Pathophysiology and Transplantation, University of Milan, Milan 20122, Italy 5 Center for Systems and Computational Biology, The Wistar Institute, Philadelphia, PA 19104, USA 6 Institute for Molecular Bioimaging and Physiology (IBFM), National Research Council (CNR), Milan 20090, Italy 7 Department of Health Sciences, University of Milan, Milan 20142, Italy 8 Department of Diagnostic Services, Unit of Nuclear Medicine, San Paolo Hospital, Milan 20142, Italy 9 Division of Neurosurgery, Fondazione IRCCS Ca’ Granda-Ospedale Maggiore Policlinico, Milan 20122, Italy 10 Functional Genomics Research Center, Korea Research Institute of Bioscience and Biotechnology. 125 Gwahak-ro, Yuseong-gu, Daejeon 305-806, Republic of Korea 11 Department of Cancer Biology, Kimmel Cancer Center, Thomas Jefferson University, Philadelphia, PA 19107, USA 12 Molecular and Cellular Oncogenesis Program, The Wistar Institute, Philadelphia, PA 19104, USA Correspondence to: Dario C. Altieri, M.D., The Wistar Institute, 3601 Spruce Street, Philadelphia, PA 19104, Tel. (215) 495-6970; (215) 495-2638; daltieri@wistar.org 11 11 2016 8 8 2016 08 8 2017 30 2 257272 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. SUMMARY Hypoxia is a universal driver of aggressive tumor behavior, but the underlying mechanisms are not completely understood. Using a phosphoproteomics screen, we now show that active Akt accumulates in the mitochondria during hypoxia and phosphorylates pyruvate dehydrogenase kinase 1 (PDK1) on Thr346 to inactivate the pyruvate dehydrogenase complex. In turn, this pathway switches tumor metabolism towards glycolysis, antagonizes apoptosis and autophagy, dampens oxidative stress, and maintains tumor cell proliferation in the face of severe hypoxia. Mitochondrial Akt-PDK1 signaling correlates with unfavorable prognostic markers and shorter survival in glioma patients and may provide an “actionable” therapeutic target in cancer. Mitochondria Akt PDK1 hypoxia metabolism tumor cell proliferation INTRODUCTION Hypoxia is a nearly universal feature of tumor growth (Hockel and Vaupel, 2001), conferring worse disease outcome via protection from apoptosis (Graeber et al., 1996), resistance to therapy (Tredan et al., 2007), and enhanced metastatic competence (Cox et al., 2015). This pathway requires the transcriptional activity of hypoxia-inducible factor 1 (HIF1), a master regulator of oxygen homeostasis (Keith et al., 2012) that becomes stabilized upon drops in oxygen pressure by escaping prolyl hydroxylation and proteasome-dependent destruction by the von Hippel-Lindau tumor suppressor (Semenza, 2013). In turn, nuclear localized HIF1 contributes to oncogene signaling (Mazumdar et al., 2010), angiogenesis (Ravi et al., 2000), cell invasion (Gilkes et al., 2014), and tumor metabolic reprogramming. In this context, mitochondria are the primary site of hypoxia-induced metabolic reprogramming in tumors (Denko, 2008). This response involves HIF1-dependent transcription of mitochondrial pyruvate dehydrogenase kinase (PDK) (Kim et al., 2006; Papandreou et al., 2006), which in turn phosphorylates the pyruvate dehydrogenase complex (PDC) on three separate sites (Patel et al., 2014). By suppressing the oxidative decarboxylation of pyruvate into acetyl-CoA (Patel et al., 2014), an active PDK shuts off oxidative phosphorylation, lowers the production of toxic ROS, and switches tumor bioenergetics towards glycolysis (Denko, 2008), a driver of more aggressive disease traits (Gatenby and Gillies, 2004). What has remained unclear, however, is whether HIF1-dependent transcription is the sole mechanism for PDK activation in hypoxia (Kim et al., 2006; Papandreou et al., 2006), and the existence of other potential regulators of this response has not been widely investigated. In this study, we examined mechanisms of the tumor response to hypoxia. RESULTS A mitochondrial Akt phosphoproteome in hypoxia We began this study by profiling the mitochondrial phosphoproteome of prostate adenocarcinoma PC3 cells exposed to severe hypoxia (<0.5% oxygen for 48 hr) versus normoxia. A total of 4,236 phosphosites were identified in the phosphopeptide-enriched samples, with a large number of changes in phosphorylation level in hypoxia/normoxia samples (Figures 1A and S1A). In total, 1,329 phosphosites showed a significant change (minimum fold-change of 1.6) in at least one sample analyzed (Figure 1A). By bioinformatics analysis, the mitochondrial phosphoproteome in hypoxia contained regulators of organelle integrity, bioenergetics, gene regulation and proteostasis (Figure 1B), which are functionally implicated in tumor cell proliferation, motility, invasion and apoptosis (Figure S1B). To complement these data, we also examined changes in the global mitochondrial proteome in hypoxia versus normoxia. A total of 5,583 proteins were identified in this analysis, and 267 of these proteins showed a significant change in hypoxia/normoxia samples (Figure S1C). Many of the phosphosites were not modulated at the protein level, suggesting that these phosphorylation events were independent of protein expression. In addition, the hypoxia-regulated mitochondrial phosphoproteome contained a discrete “Akt signature” (Figure 1A), characterized by increased phosphorylation of six Akt target proteins in hypoxia versus normoxia (Figure 1C). Based on these results, we next looked at a role of Akt in the tumor response to hypoxia. Exposure of PC3 cells to hypoxia resulted in increased recruitment of Akt to mitochondria, whereas the cytosolic levels of Akt were unchanged between hypoxia and normoxia (Figures 1D and S1D). The hypoxia-regulated pool of mitochondrial Akt was “active” as it was phosphorylated on Ser473 (Figure 1D) and persisted for up to 24 hr after re-oxygenation (Figures 1E and S1E). Consistent with these results, hypoxia was accompanied by increased phosphorylation of a set of mitochondrial proteins recognized by an antibody to the Akt consensus phosphorylation sequence, RxRxxS/T (Akt cons Ab) (Figure 1F). Preincubation of mitochondrial extracts with Akt cons Ab (Figure 1F), or silencing Akt2 by small interfering RNA (siRNA) (Figure S1F), removed the mitochondrial proteins recognized by Akt cons Ab in hypoxia, confirming the specificity of this response and Akt-directed phosphorylation activity in mitochondria in hypoxia. Silencing Akt1 had minimal effect (Figure S1F). siRNA silencing of HIF1α did not affect Akt recruitment to mitochondria in hypoxia (Figure S1G), suggesting that this pathway did not require HIF1-dependent transcription. In addition, depletion of HIF1α did not affect Akt levels in the cytosol or mitochondria under normoxic conditions, whereas phosphorylated Akt2 levels were increased in the cytosol in response to hypoxia (Figure S1G). As detected by Akt cons Ab, the expression levels of downstream Akt-phosphorylated target molecules were unchanged in normoxic or hypoxic conditions (Figure 1F). In response to hypoxia, active Akt was found predominantly in the mitochondrial inner membrane, and, to a lesser extent, the matrix (Figure S1H). The mechanism(s) of how Akt is recruited to mitochondria in hypoxia was further investigated. We found that blocking the chaperone activity of heat shock protein-90 (Hsp90) with 17-allylaminogeldanamycin (17-AAG) prevented the accumulation of mitochondrial Akt in hypoxia (Figure 1G). Also, scavenging mitochondrial ROS with MitoTempo (MT) inhibited Akt recruitment to mitochondria (Figure 1H). The antioxidant N-acetyl cysteine (NAC) had no effect (Figure 1H), identifying mitochondria-derived ROS as a critical stimulus for mitochondrial accumulation of Akt in hypoxia. PDK1 is a phosphorylation target of mitochondrial Akt in hypoxia We next set up a 1D proteomics screen to identify mitochondrial proteins phosphorylated by Akt in hypoxia (Figure 2A). Immune complexes precipitated with Akt cons Ab from normoxic or hypoxic PC3 cells contained bands with ~35 to ~120 kDa molecular weight that were more abundant in hypoxia (Figure S2A). Preclearing mitochondrial extracts with Akt cons Ab removed most of these proteins, validating the specificity of the immunoprecipitation step. From these experiments, we identified by mass spectrometry 84 high-confidence Akt substrates differentially expressed in hypoxia (Table S1). Sixteen of these molecules were known mitochondrial proteins (Figure 2B), including hypoxia- and HIF1-regulated effectors of bioenergetics (UGP2, SLC2A1, PDK1, HK2), extracellular matrix remodeling (P4HA1), Ca2+ homeostasis at the ER-mitochondria interface (Ero1L), oxidative phosphorylation (LonP1, IBA57), and metabolism (Acot9). Due to previously published work suggesting the importance of PDK1 in the tumor hypoxic response, we next focused on PDK1 as a potential substrate of mitochondrial Akt in hypoxia. In kinase assays, active Akt1 or Akt2 readily phosphorylated PDK1, as well as control GSK3β, as determined by Western blotting with Akt cons Ab (Figure 2C). This phosphorylation event was selective for PDK1, as related PDK2, PDK3 or PDK4 isoforms were unreactive (Figure 2D). In addition, PDK1 immune complexes reacted with Akt cons Ab preferentially in hypoxia (Figure 2E), and reciprocally, immune complexes precipitated with Akt cons Ab in hypoxia contained PDK1 (Figure S2B), consistent with the model of Akt phosphorylation of PDK1 in hypoxia. We next looked for potential Akt phosphorylation sites in PDK1 by LC-MS/MS analysis of chymotrypsin digests of Akt-phosphorylated PDK1 in a kinase assay separated by SDS-PAGE (Figure S2C). We identified Thr346 (T346) in a number of PDK1 chymotryptic peptides, including the sequence STAPRPRVEpTSRAVPL (m/z=908.9751) as the sole phospho-amino acid modified by Akt1 or Akt2, compared to control (Figure 2F). The PDK1 sequence surrounding T346 matched an Akt consensus phosphorylation site, RxRxxS/T (Figure S2D), which was not present in PDK2, PDK3 or PDK4 (Figure S2E). Consistent with these data, active Akt2 phosphorylated wild type (WT) PDK1 but not a phosphorylation-defective Thr346→Ala (T346A) PDK1 mutant in transfected PC3 cells (Figure 2G). In the PDK1 crystal structure, T346 is predicted to localize to a flexible, “ATP lid” hinge region (Figure 2H), positioned to affect ATP loading and kinase activation. To independently validate these findings, we next generated a phospho-specific antibody to phosphorylated T346 (pT346 Ab) in PDK1. The pT346 Ab dose-dependently reacted with the phosphorylated PDK1 peptide CAPRPRVE(pT)SRAVPLA, but not the non-phosphorylated sequence (Figure S2F). A second antibody raised against the non-phosphorylated sequence recognized the non-phosphorylated PDK1 peptide (Figure S2G). Under these conditions, pT346 Ab reacted with Akt2-phosphorylated WT PDK1, but not T346A PDK1 mutant (Figure 2I). Consistent with the model that T346 phosphorylation is hypoxia-sensitive, WT PDK1, but not T346A PDK1, precipitated from hypoxic PC3 cells reacted with pT346 Ab (Figure 2J). pT346 Ab only weakly reacted with WT or T346A PDK1 precipitated from normoxic cells (Figure 2J). Finally, we generated clones of PC3 cells stably silenced for endogenous PDK1 by short hairpin RNA (shRNA). pT346 Ab did not react with these cells in normoxia (Figure S2H). In contrast, pLKO transfectants reacted with pT346 Ab in hypoxia, and this response was abolished by shRNA silencing of PDK1 (Figure S2H). Akt-PDK1 phosphorylation axis in hypoxia Expression of WT PDK1 in hypoxic PC3 cells increased the phosphorylation of the E1α catalytic subunit (PDHE1) of the PDC (Figure 3A) on site 1 (Ser264 in the mature protein), one of three regulatory phosphorylation sites (Patel et al., 2014). Conversely, expression of T346A PDK1 mutant reduced PDHE1 phosphorylation in hypoxia (Figure 3A), and no PDHE1 phosphorylation was detected in normoxia (Figure 3A). Immune complexes of WT or T346A PDK1 mutant contained comparable amounts of the PDC component, PDHE1α, suggesting that T346 does not contribute to a PDK1-PDC complex (Figure S3A). In a kinase assay, active Akt2 increased PDK1 phosphorylation of PDHE1 (Figure 3B). While WT PDK1 phosphorylated PDHE1 in the presence of Akt2 (Figure 3C), T346A PDK1 mutant was ineffective (Figure 3C). Consistent with these data, increased PDHE1 phosphorylation was detected only in the presence of Akt2 and PDK1, but not PDK2, PDK3 or PDK4 (Figure S3B). Silencing of Akt2 inhibited PDHE1 phosphorylation in hypoxia, whereas Akt1 knockdown only had a partial effect (Figure 3D). As a complementary approach, we used a pan-Akt small molecule inhibitor, MK2206, which indistinguishably suppressed Akt phosphorylation in hypoxia and normoxia (Figure S3C). Incubation of PC3 cells with MK2206 suppressed PDHE1 phosphorylation in hypoxia (Figure 3E). This response was specific because PDK1 immunoprecipitated from MK2206-treated cells also failed to phosphorylate PDHE1 in a kinase assay in hypoxia (Figure S3D). In normoxia, MK2206 had no effect on PDHE1 phosphorylation in cell extracts (Figure 3E) or in a kinase assay with immunoprecipitated PDK1 (Figure S3D), validating the specificity of Akt-directed phosphorylation in hypoxic conditions. As an independent approach, we next generated WT or kinase-dead (KD) Akt2 constructs targeted to the mitochondria by the cytochrome c oxidase subunit 8 mitochondrial import sequence. Similar to the endogenous protein, mitochondrial-targeted Akt2 accumulated in the various submitochondrial fractions (Figure S3E). Functionally, mitochondrial-targeted Akt2-KD inhibited PDHE1 phosphorylation in hypoxic PC3 cells (Figure 3F), whereas non-mitochondrial targeted Akt2-KD had no effect. There was no PDHE1 phosphorylation in the cytosol of hypoxic or normoxic tumor cells, and Akt2-KD or mitochondrial-targeted Akt2-KD had no effect in these settings (Figure S3F). Reciprocally, forced expression of mitochondrial-targeted WT Akt2 was sufficient to increase PDHE1 phosphorylation even in the absence of hypoxia (Figure S3G). Finally, we reconstituted PDK1-depleted cells with various PDK1 cDNAs. Expression of WT PDK1 in these settings restored PDHE1 phosphorylation in hypoxia, whereas T346A PDK1 mutant had no effect (Figure 3G). With respect to its enzymatic function, PDK1 knockdown increased PDH activity in normoxic PC3 cells (Figure 3H). Hypoxic cells showed reduced PDH activity, and this response was partially rescued by shRNA silencing of PDK1 (Figure 3H). Reconstitution of these cells with WT PDK1, but not T346A PDK1 mutant, suppressed PDH activity in hypoxia (Figure 3I). In addition, expression of Akt-KD or mitochondrial-targeted Akt-KD in PC3 cells had no effect on PDH activity in normoxia, but modestly elevated PDH function in hypoxia (Figure S3H), consistent with loss of an Akt-regulated inhibitory function of PDK1 in these settings. Vector or non-mitochondrial targeted Akt2-KD had no effect (Figure S3H). Mitochondrial Akt-PDK1 phosphorylation controls tumor metabolic reprogramming To understand how mitochondrial Akt-PDK1 signaling affects tumor behavior, we first looked at potential changes in cancer metabolism. Consistent with previous studies, hypoxia stimulated glycolytic metabolism in tumor cells, characterized by increased glucose consumption (Figure 3J) and lactate production (Figure 3K). Mitochondrial Akt-PDK1 signaling was important for this response, as PDK1 knockdown reduced glucose consumption in hypoxia, whereas reconstitution of targeted cells with WT PDK1, but not T346A PDK1 mutant, restored glycolysis (Figure 3J). Similarly, Akt inhibition with MK2206 (Figure 3K) or silencing of Akt2 (Figure 3L) impaired metabolic reprogramming, reducing lactate production in hypoxia. Normoxic cultures were not affected (Figures 3K and 3L), and Akt1 knockdown had only partial effect (Figure 3L). Consistent with a metabolic switch towards glycolysis (Kim et al., 2006; Papandreou et al., 2006), PC3 cells reconstituted with WT PDK1 exhibited reduced oxygen consumption, a marker of oxidative phosphorylation, whereas expression of T346A PDK1 mutant restored oxygen consumption (Figure 3M), further supporting a role of mitochondrial Akt-PDK1 signaling in hypoxic metabolic reprogramming. Mitochondrial Akt-PDK1 phosphorylation in vivo When analyzed in time-course experiments, hypoxia increased phosphorylation of Akt1 and Akt2, as well as PDHE1 starting at 3 and 6 hr, respectively (Figure S4A). The overall hypoxic response under these conditions was cell type-specific. Akt inhibition strongly reduced PDHE1 phosphorylation in prostate adenocarcinoma (DU145), lung adenocarcinoma (A549) and glioblastoma (LN229), but had no effect on PDHE1 phosphorylation in breast adenocarcinoma cells MCF-7 (ER+) or MDA-231 (ER−) (Figure S4B). Knockdown of PTEN in MCF-7 cells increased PDHE1 phosphorylation in normoxia and, to a lesser extent, hypoxia, whereas LN229 cells were unaffected (Figure S4C), suggesting that PTEN status may differentially affect hypoxia-stimulated mitochondrial Akt-PDK1 signaling depending on the tumor cell type. To examine a more “physiologic” model of tumor hypoxia, we next looked at 3D cultures of patient-derived, stem cell-enriched GBM neurospheres (Di Cristofori et al., 2015). These cultures become hypoxic in their “core”, as determined by expression of a hypoxia probe (Figures 4A and S4D and S4E). Under these conditions, GBM neurospheres exhibited strong phosphorylation of PDK1, as determined by immunofluorescence with pT346 Ab (Figure 4A). Conversely, differentiated GBM cells depleted of stem cells and growing as monolayers were normoxic, contained cytosolic HIF1α, and did not react with pT346 Ab (Figure 4A). Pre13 absorption of pT346 Ab with the immunizing peptide abolished reactivity with GBM (Figure S4D). Next, we looked at Akt phosphorylation of PDK1 in primary, patient-derived GBM tissue samples (Table S2). GBMs with a high score (≥2) for nuclear HIF1α showed increased phosphorylation of PDK1 by Akt (pT346 Ab), as well as phosphorylation of PDHE1 and Src, a major determinant of glioma invasiveness (Du et al., 2009), in hypoxic areas (Figures 4B and S4F). In contrast, GBMs with undetectable nuclear HIF1α (score = 0) showed low to undetectable levels of PDK1-PDHE1 phosphorylation (Figures 4C and S4F). In these patients, phosphorylation of PDK1 (pT346 Ab) (Figure 4D) or PDHE1 phosphorylation (Figure 4E) correlated with expression of nuclear HIF1α. Reciprocally, PDHE1 phosphorylation correlated with the expression of Akt-phosphorylated PDK1 (pT346 Ab) (Figure 4F), reinforcing a link between hypoxia and mitochondrial Akt-PDK1 phosphorylation in primary patient samples. Mitochondrial Akt-PDK1 regulation of tumor cell proliferation in hypoxia To test a role of a mitochondrial Akt-PDK1 signaling in tumor growth in vivo, we first utilized human U251 GBM cells expressing a luciferase reporter under the control of a HIF1-responsive element (HRE) and a mCherry reporter under a constitutive PGK promoter to quantify cell viability. Stereotactic intracranial injection of these cells in immunocompromised mice gave rise to GBMs characterized by HIF1-directed luciferase activity and reactivity with a hypoxia-sensitive marker (Figures 5A and 5B). Despite low oxygenation, these orthotopic GBMs remained viable, as determined by high mCherry expression (Figures 5A and 5B) and exhibited a time-dependent increase in the number of mitotic tumor cells (Figure S5A). These proliferating cells stained intensely positive for Akt-phosphorylated PDK1 (Figures 5C and S5B and S5C), correlating with increased HIF1 activity (Figure S5D). PDHE1 was also highly phosphorylated in intracranial GBMs (Figure 5C). We next tested the requirement of mitochondrial Akt-PDK1 signaling in regulating proliferation under hypoxic conditions. siRNA knockdown of Akt1 or Akt2 (Figure 5D) or stable shRNA knockdown of PDK1 (Figure 5E) suppressed tumor cell proliferation in hypoxia. Normoxic cultures were partially affected (Figures 5D and 5E). When cells were analyzed for cell cycle transitions, MK2206 or the PDK1 inhibitor dichloroacetate (DCA) suppressed S-phase progression in hypoxia and increased the population of tumor cells in G1/sub-G1 phase (Figure S5E). Finally, stable silencing of PDK1 abolished PC3 colony formation in hypoxia, a marker of tumorigenicity (Figures 5F and 5G), whereas normoxic growth was not significantly affected. Together, these data point to an important role of mitochondrial Akt-PDK1 signaling in maintaining tumor cell proliferation in hypoxia. Mitochondrial Akt regulation of stress signaling in hypoxia The downstream implications of defective mitochondrial Akt-PDK1 signaling were next investigated. First, inhibition of Akt with MK2206 (Figure 6A) or stable shRNA silencing of PDK1 (Figure 6B) increased aberrant ROS production in tumor cells, especially in hypoxia. This was associated with decreased tumor cell viability (Figure 6C), and appearance of cleaved caspase 3 (Figure 6D), a marker of apoptosis. In normoxia, cleaved caspase 3 was undetectable. Confirming the specificity of this response, exposure of tumor cells to a small molecule inhibitor of PI3K, PX-866 did not result in caspase activation (Figure 6D). Reconstitution of these cells with WT PDK1, but not T346A PDK1 mutant, partially rescued tumor cell viability in hypoxia (Figure 6E). Normoxic cultures were not affected, further supporting a role of PDK1 signaling selectively in hypoxia. As a second downstream pathway of tumor maintenance modulated by bioenergetics, we next observed that stable knockdown of PDK1 (Figure 6F) or siRNA silencing of Akt1 or Akt2 (Figure 6G) in hypoxia increased the phosphorylation of the energy stress sensor, AMP-regulated kinase (AMPK). This response was associated with concomitant activation of autophagy, as determined by LC3 conversion to a lipidated form (Figures 6F and 6G), and punctate LC3 fluorescence staining (Figures 6H and 6I). Normoxic cultures showed a minimal level of autophagy induction after PDK1 silencing (Figures 6F and 6G and S6A). In PDK1-depleted cells, re-expression of WT PDK1, but not T346A PDK1 mutant, attenuated AMPK phosphorylation and reduced autophagy in hypoxia (Figures 6H and 6I and S6B). Requirement of hypoxic mitochondrial reprogramming for tumor growth in vivo Next, we asked if mitochondrial Akt-PDK1 signaling was important for tumor growth in vivo. shRNA silencing of PDK1 significantly impaired the growth of PC3 xenograft tumors implanted in immunocompromised mice (Figure 7A). Re-expression of WT PDK1 in these cells restored tumor growth in vivo (Figures 7B and 7C), whereas T346A PDK1 mutant further impaired tumor growth (Figure 7C). By immunohistochemistry, PC3 tumors harboring WT PDK1 showed increased cell proliferation, reduced apoptosis and lower levels of autophagy, compared to tumors reconstituted with T346A PDK1 mutant (Figures 7D and 7E). In addition, tumors with loss of endogenous PDK1 showed a trend towards increased apoptosis and heightened autophagy in vivo, whereas tumor cell proliferation by Ki-67 staining was unchanged (Figures 7F and G). Taken together, these results suggest that mitochondrial Akt-PDK1 signaling promotes tumor adaptation to hypoxia, and specifically enables continued tumor cell proliferation despite an unfavorable microenvironment (Figure 7H). To test the relevance of this model in human cancer, we next looked at the prognostic impact of Akt phosphorylation of PDK1 in a clinically-annotated cohort of 116 glioma patients (Table S3). Undetectable in normal brain parenchyma, the expression of Akt phosphorylated PDK1 on T346 progressively increased in gliomas, with the highest reactivity observed in glioblastoma (Figures 8A and 8B). PDK1 phosphorylation on T346 segregated with other markers of disease progression, including HIF1α expression (Figure 8C), wild type status of isocitrate dehydrogenase-1 (IDH1) (Figure 8D), and unmethylated MGMT promoter (Figure 8E). Consistent with this prognostic profile, elevated expression of Akt-phosphorylated PDK1, as determined by ROC curves analysis (Figures S7A and S7B), was significantly associated with reduced overall survival in patients with gliomas (p=0.006; HR=2.2; 95%CI: 1.17–4.12; Figure 8F) as well as patients with GBM (p=0.032; HR=2.03; 95%CI: 0.95–4.32; Figure 8G). DISCUSSION In this study, we have shown that hypoxia, a universal driver of malignancy, promotes the recruitment of active Akt to mitochondria of tumor cells. In turn, mitochondrial Akt, in particular Akt2, phosphorylates the bioenergetics regulator PDK1 on a Thr346 target site, resulting in increased kinase activity and phosphorylation of its downstream substrate in the PDC, PDHE1α. This pathway improves tumor fitness, stimulating glycolysis, countering autophagy and apoptosis, dampening oxidative stress, and enabling continued cell proliferation in face of severe hypoxia in vivo. Accordingly, mitochondrial Akt-PDK1 signaling emerged as a powerful negative prognostic factor in glioma patients, correlating with markers of unfavorable outcome and shortened survival. Akt is an essential signaling node exploited in most cancers, integrating growth factor responses with mechanisms of cell proliferation, survival and bioenergetics (Manning and Cantley, 2007). The role of this pathway as a regulator of tumor adaptation to hypoxia has not been previously described, and its spatial arrangement in subcellular compartments, in particular mitochondria, has only recently begun to emerge (Ghosh et al., 2015). Data presented here suggest that the recruitment of predominantly active Akt to mitochondria during hypoxia (Santi and Lee, 2010) may be part of a broader stress response in tumors, enabled by the chaperone function of cytosolic Hsp90 in mitochondrial pre-protein import (Young et al., 2003) and mitochondrial ROS production, which may participate in subcellular trafficking of signaling molecules (Nakahira et al., 2006), including mitochondrial import (Fischer and Riemer, 2013). In our phosphoproteomics screen, Akt recruitment to mitochondria was associated with a discrete Akt phosphorylation signature that included regulators of organelle homeostasis and glycolytic reprogramming in hypoxia, such as 6-phosphofructose-2-kinase/fructose-2,6-bisphosphatase 3 (PFKB3) and PDK1 (De Bock et al., 2013). In the case of PDK1, Akt phosphorylation took place exclusively in hypoxia, did not involve other PDK isoforms, and targeted a unique Thr346 site in the “ATP lid” (Zhang et al., 2015), ideally positioned to affect ATP loading, and kinase activation (Patel et al., 2014). Thr346 did not affect PDK1 binding to the PDC, thus differently from another proposed post-translational modification of PDK1 involving Tyr phosphorylation of the “ATP lid” (Hitosugi et al., 2011). Extensively studied as part of HIF1 signaling (Semenza, 2013), the tumor response to hypoxia has been linked to metabolic reprogramming (Kim et al., 2006; Papandreou et al., 2006), with suppression of mitochondrial respiration in favor of glycolysis (Denko, 2008). However, there is evidence that this pathway may extend well beyond bioenergetics, as PDK1 signaling has been implicated in senescence (Kaplon et al., 2013), metastatic tropism (Dupuy et al., 2015), and multiple mechanisms of tumor maintenance (McFate et al., 2008). The fact that this response is “druggable” and that small molecule PDK1 inhibitors have entered clinical testing in cancer patients (Michelakis et al., 2010) adds to the relevance of PDK1 signaling as a potential driver of tumor progression. Against this backdrop, the pathway of mitochondrial Akt-PDK1 signaling described here appears ideally poised to affect a plethora of downstream signaling mechanisms important for tumor adaptation and improved fitness. This involves suppression of apoptosis via mitochondrial Akt phosphorylation of hexokinase-II at the outer membrane (Roberts et al., 2013) and cyclophilin D in the permeability transition pore (Ghosh et al., 2015), as well as inhibition of oxidative phosphorylation through Akt phosphorylation of PDK1 and subsequent inhibition of the PDC (Patel et al., 2014). In turn, this lowers the production of toxic ROS, prevents the phosphorylation of stress energy sensor, AMPK (Liang and Mills, 2013), and inhibits downstream activation of autophagy (White, 2012). Although these pathways have been implicated in both tumor suppression and oncogenesis (Liang and Mills, 2013; White, 2012), there is evidence that AMPK inhibition and suppression of autophagy may promote malignant expansion (White, 2012), and heightened metastatic competence (Caino et al., 2013), including in hypoxia (Liu et al., 2015), further compounding the more aggressive traits of glycolytic tumors (Gatenby and Gillies, 2004). Here, a pivotal feature of mitochondrial Akt-PDK1 signaling was its activation in mitotic cells and requirement to support tumor cell proliferation in the face of hypoxia, in vivo. Whether this response can be entirely attributed to the suppression of ROS or involves other mechanisms of mitochondria-to-nuclei retrograde signaling (Jazwinski, 2013) remains to be elucidated. On the other hand, hypoxic reprogramming may participate in cell cycle transitions via regulation of p27 expression (Gardner et al., 2001) or Myc transcriptional activity (Gordan et al., 2007), and effector(s) of glycolysis have been linked to chromosomal segregation and mitotic progression, selectively in hypoxia (Jiang et al., 2014). In line with the broad impact of mitochondrial Akt-PDK1 signaling on multiple mechanisms of tumor adaptation, this pathway was found to have significant implications for disease progression in humans. Accordingly, expression of PDK1 phosphorylated on T346 was undetectable in normal brain, but increased steadily in the hypoxic environment of gliomas, including glioblastomas, segregated with unfavorable prognostic markers, and correlated with shortened overall survival in these patients. Although these results await further confirmation in larger patient cohorts, detection of Akt-phosphorylated PDK1 may provide an easily accessible biomarker for clinical decision-making in patients with gliomas, including glioblastoma (Wick et al., 2014). Despite expectations for personalized, or precision medicine, small molecule antagonists of Akt and its associated signaling nodes, PI3K and mTOR (Manning and Cantley, 2007), have produced only limited responses in the clinic, hampered by drug resistance and significant toxicity (Fruman and Rommel, 2014). While these results may reflect mechanisms of tumor adaptation (Ghosh et al., 2015), including mitochondrial reprogramming (Caino et al., 2015), the identification of mitochondrial Akt (Ghosh et al., 2015) as a post-translational regulator of PDK1 activity and tumor progression may rationally repurpose Akt-directed molecular therapies as an approach to impair tumor adaptation to hypoxia. In this context, targeted inhibition of the mitochondrial pool of Akt may selectively disable a host of metabolic, survival and proliferative requirements for tumor fitness (McFate et al., 2008) and reawaken endogenous tumor suppressor mechanisms important for anticancer activity in patients. EXPERIMENTAL PROCEDURES Patients All patient-related studies were reviewed and approved by an Institutional Review Board at Fondazione IRCCS, Ca’ Granda Ospedale Maggiore Policlinico Milan, Italy. A first cohort of 26 patients diagnosed with de novo glioma were enrolled at Fondazione IRCCS Ca’ Granda Ospedale Maggiore Policlinico (Milan, Italy) between 2010 and 2011, and described previously (Di Cristofori et al., 2015). All patients were treated with surgical resection with curative intent. Gliomas were staged according to the WHO classification (Louis et al., 2007), and the clinicopathological and molecular characteristics of the patients’ series used in this study are described in Table S2. This cohort was used to evaluate the expression of phosphoT346-PDK1 (pPDK1), phosphoPDHE1α (pPDHE1), phosphoT416-Src (pSrc) and nuclear HIF1α reactivity, by immunohistochemistry. Tissue microarrays (TMAs) of glioma or normal brain tissues were as described (Di Cristofori et al., 2015). Immunohistochemistry slides were digitalized using an Aperio scanner at 20× magnification, and HIF1α nuclear staining was quantified using a nuclear-specific algorithm implemented in Genie Histology Pattern Recognition software (Aperio, Leica Microsystems). To specifically quantify nuclear HIF1α expression in primary GBM culture, the Volocity algorithm that counts and displays red signals (HIF1α) within the Hoechst signal (nuclei) in each z-stack was used. A second series of 116 patients with de novo glioma who underwent surgery with curative intent between 2008 and 2009 and for which complete clinical and follow-up records were available (Table S3), was retrieved from the archives of the Pathology Division. This cohort was used in the present study to correlate the expression of Akt-phosphorylated PDK1 on T346 with prognostic markers of glioma progression, including nuclear HIF1α, MGMT promoter methylation and IDH1 mutational status (wild type (WT)/R132H), and patients’ overall survival. Xenograft tumor growth studies All experiments involving animals were approved by an Institutional Animal Care and Use Committee (IACUC) at The Wistar Institute in accordance with the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health (NIH), or, alternatively, at the University of Milan, in compliance with the Italian Ministry of Health. In a first set of experiments, PC3 cells stably transduced with control pLKO or PDK1-directed shRNA were reconstituted with vector, WT PDK1 or T346A PDK1 mutant cDNA at 80% confluency, suspended in PBS, pH 7.4, and injected (0.2 ml containing 2×106 cells) s.c. into the flank of 6–8 week old male NOD SCID γ (NSG, NOD.Cg-Prkdcscid Il2rgtm1Wjl/SzJ) immunocompromised mice (Jackson Laboratory, 3 mice per condition/2 tumors per mouse). The width and length of superficial tumors were measured with a caliper at the indicated time intervals, and tumor volume was calculated according to the formula Vol= width2 × length/2. After 21 days Xenograft tumors were harvested, fixed and processed for immunohistochemistry. An orthotopic murine model of glioblastoma (GBM) was obtained by stereotactic injection (coordinates: 1.5 mm lateral to the bregma, 0 mm behind and 3.0 mm ventral to the dura) (Maes et al., 2009) of 1×105 U251-HRE-mCherry GBM cells in 2 µl of PBS into 7–8-week-old female nude mice (Harlan Laboratories) at day 0 (Lo Dico et al., 2014). Following surgery, mice were monitored for recovery until complete awakening. Six animals per time point were used and mice were euthanized after 20 or 34 days. Intracranial GBM samples were harvested from the various groups and processed for differential expression of phosphorylated PDK1 or PDHE1, HIF1α or Ki-67 by immunohistochemistry on serial sections. Statistical analysis Data were analyzed using the two-sided unpaired t or chi-square tests using a GraphPad software package (Prism 6.0) for Windows. Correlation parameters between immunohistochemical (IHC) scores in glioma patients and clinicopathological variables were derived using Mann-Whitney U test or chi-square test for continuous or discrete variable, respectively, using GraphPad Prism or MedCalc (Mariakerke, Belgium) statistical package. Receiver operating characteristics (ROC) curves method was used to test the accuracy of T346 phosphorylated PDK1 to correctly discriminate between glioma patients according to their survival status (alive or dead for the disease) and to generate cut-offs for phosphorylated PDK1 IHC score using the non-arbitrary criterion derived from the Youden’s statistic (J, MedCalc Software) as described (Di Cristofori et al., 2015). The pPDK1 IHC score value that more accurately discriminated between alive or dead patients was >25 and >40 for gliomas or GBM patients, respectively (Youden criterion). Glioma patients were then sorted into low or highexpressor categories and Kaplan-Meier survival curves were compared using the Log-Rank test (MedCalc Software). Data are expressed as mean±SD or mean±SEM of replicates from a representative experiment out of at least two or three independent determinations. A p value of <0.05 was considered as statistically significant. Supplementary Material S1 S2 This work was supported by the National Institutes of Health (NIH) grants P01 CA140043 (D.C.A. and L.R.L.), R01 CA78810 and CA190027 (D.C.A.), R01 CA089720 (L.R.L.), F32CA177018 (M.C.C.), the Office of the Assistant Secretary of Defense for Health Affairs through the Prostate Cancer Research Program under Award No. W81XWH-13-1-0193 (D.C.A.), and a Challenge Award from the Prostate Cancer Foundation (PCF) to M.C.C., L.R.L. and D.C.A. V.V. is supported by an award from Fondazione Cariplo (2014-1148), and L.O. is supported by FP7-INSERT project (HEALTH-2012-INNOVATION-1, GA305311). I.B. is supported by a fellowship from the Doctorate School of Molecular and Translational Medicine at the University of Milan, Italy. Support for Core Facilities utilized in this study was provided by Cancer Center Support Grant (CCSG) CA010815 to The Wistar Institute. Figure 1 Mitochondrial phosphoproteome in hypoxia (A) Phosphoproteome of prostate adenocarcinoma PC3 cells in hypoxia versus normoxia. Identified phosphosites met a minimum MaxQuant localization probability of 0.75 and a score difference of 5. Fold changes were calculated from the normalized Heavy/Light SILAC ratio. Six Akt target proteins showing increased phosphorylation in hypoxia are indicated. Grey, not significant; red, upregulated; blue, downregulated; yellow squares, Akt targets. (B) Ingenuity pathway analysis of mitochondrial phospho- and global proteome in hypoxia. (C) Kinases for which at least 5 known targets showed significant changes in phosphorylation in hypoxic versus normoxic conditions as in (A). Up, upregulation; Dn, downregulation. *, The modulated genes are: ARID1A; HIST1H1E; HMGA1; LARP1; LIG1; LIG3; LMNB2; LRCH3; LRWD1; MARCKS; MED1; MKI67; NCL; NPM1; NUCKS1; PDS5B; PTPN2; RB1; RBL1; RBL2; SAMHD1; SETDB1; TERF2; VIM. **, The modulated genes are: DUT; EEF1D; HIST1H1E; HMGA1; IRS2; LIG1; LIG3; LMNA; LMNB1; MAP4; NOLC1; NPM1; NUCKS1; PDS5B; PTPN2; RB1; SAMHD1; TCOF1; TOP2A; TPX2; VIM. (D) PC3 cells in normoxia (N) or hypoxia (H) were fractionated in cytosol (Cyto) or mitochondrial (Mito) extracts and analyzed by Western blotting. pAkt, phosphorylated Akt (Ser473). TCE, total cell extracts. (E) PC3 cells in hypoxia (H) were exposed to reoxygenation (O2) for the indicated time intervals and analyzed by Western blotting. N, normoxia. (F) The indicated subcellular fractions isolated from normoxic (N) or hypoxic (H) PC3 cells were analyzed with an antibody to the Akt consensus phosphorylation site RxRxxS/T (Akt cons Ab) by Western blotting. Mito Sup, supernatant of mitochondrial extracts after preclearing with Akt cons Ab. (G) PC3 cells in normoxia (N) or hypoxia (H) were treated with vehicle (Veh) or Hsp90 small molecule inhibitor 17-AAG (5 µM for 6 hr), and cytosolic (Cyto) or mitochondrial (Mito) extracts were analyzed by Western blotting. (H) PC3 cells in normoxia (N) or hypoxia (H) were treated with vehicle (Veh), the antioxidant N-acetyl cysteine (NAC, 1 mM) or mitochondria-specific ROS scavenger, MitoTempo (MT, 25 µM), and subcellular fractions were analyzed by Western blotting. See also Figure S1. Figure 2 Mitochondrial Akt phosphorylation of PDK1 (A) Schematic diagram for the identification of a mitochondrial Akt phosphoproteome in hypoxic versus normoxic PC3 cells. (B) Mitochondrial proteins reacting with Akt cons Ab showing differential expression in hypoxic versus normoxic PC3 cells. (C) Recombinant PDK1 or GSK3β was mixed in a kinase assay with active Akt1 or Akt2, and phosphorylated bands were detected with Akt cons Ab by Western blotting. (D) The indicated PDK isoforms were mixed in the presence or absence of active Akt2 in a kinase assay and phosphorylated bands were detected with Akt cons Ab, by Western blotting. (E) PC3 cells in normoxia (N) or hypoxia (H) were immunoprecipitated (IP) with an antibody to PDK1 followed by Western blotting. HIF1α reactivity (bottom) was used as a marker of hypoxia. TCE, total cell extracts. Bottom, densitometric quantification of phosphorylated (p) PDK1 bands. U, arbitrary units. (F) Extracted ion chromatogram of the PDK1 phosphorylated T346 chymotryptic peptide (STAPRPRVEpTSRAVPL, m/z 908.9751) resulting from incubation with or without active Akt1 or Akt2 in a kinase assay. (G) PC3 cells were transfected with vector or Flag-tagged wild type (WT) PDK1 or T346A PDK1 mutant, immunoprecipitated with an antibody to Flag and immune complexes were mixed with active Akt2 in a kinase assay followed by Western blotting with Akt cons Ab. Bottom, densitometric quantification of phosphorylated (p) PDK1 bands. U, arbitrary units. (H) Molecular dynamics simulation of the structure of PDK1 (ribbon) with stick representation of residues 336–356 comprising the “ATP lid”. The ATP molecule is derived from the structure of PDK3-L2-ATP (PDB code 1Y8P) superimposed onto the structure of PDK1. The predicted location of Thr346 as well as Arg343 and Arg348 is shown. (I) The experimental conditions are as in (G) except that Flag-PDK1 immune complexes mixed with active Akt2 in a kinase assay were analyzed with phospho-specific pT346 Ab by Western blotting. Exp., exposure. Bottom, densitometric quantification of phosphorylated (p) PDK1 bands. U, arbitrary units. (J) Flag-PDK1 immune complexes as in (G) were precipitated from PC3 cells in normoxia (N) or hypoxia (H) and analyzed with pT346 Ab by Western blotting. p, phosphorylated. Bottom, densitometric quantification of pPDK1 bands. U, arbitrary units. See also Figure S2 and Table S1. Figure 3 A mitochondrial Akt-PDK1-PDHE1 phosphorylation axis in hypoxia (A) PC3 cells in normoxia (N) or hypoxia (H) were transfected with vector, WT PDK1 or T346A PDK1 mutant and analyzed by Western blotting. Bottom, densitometric quantification of phosphorylated (p) PDHE1 bands. U, arbitrary units. (B) The indicated recombinant proteins were mixed in a kinase assay and analyzed by Western blotting. (C) PC3 cells transfected with vector or the indicated Flag-tagged WT PDK1 or T346A PDK1 mutant were immunoprecipitated (IP) with an antibody to Flag, and immune complexes were mixed in a kinase assay with recombinant Akt2 and PDHE1 followed by Western blotting. (D) PC3 cells in normoxia (N) or hypoxia (H) were transfected with control siRNA (Ctrl) or siRNA to Akt1 or Akt2, and analyzed by Western blotting. (E) PC3 cells in normoxia (N) or hypoxia (H) were treated with vehicle control (Veh) or a small molecule Akt inhibitor, MK2206 (1 µM), and analyzed by Western blotting. (F) PC3 cells in normoxia (N) or hypoxia (H) were transfected with vector, Akt-kinase dead (Akt-KD) or mitochondrial-targeted Akt-KD (mtAkt-KD) mutant, and mitochondrial extracts (Mito) were analyzed by Western blotting. (G) PC3 cells in normoxia (N) or hypoxia (H) were transduced with pLKO or PDK1-directed shRNA, reconstituted with vector, WT PDK1 or T346A PDK1 mutant cDNA and analyzed by Western blotting. Bottom, densitometric quantification of phosphorylated (p) PDHE1 bands. U, arbitrary units. (H) PC3 cells transduced with pLKO or PDK1-directed shRNA were analyzed for PDH activity in normoxia (N) or hypoxia (H) conditions. Left, representative tracings (n=4). Right, quantification of PDH activity. ns, not significant. Mean±SEM. *, p=0.03. (I) PC3 cells in hypoxia were transduced with PDK1-directed shRNA, reconstituted with vector, WT PDK1 or T346A PDK1 mutant cDNA and analyzed for PDH activity. Left, representative tracings (n=3). Right, quantification of PDH activity. Mean±SEM. **, p=0.009. (J) PC3 cells transduced with pLKO or PDK1-directed shRNA were reconstituted with vector, WT PDK1 or T346A PDK1 cDNA and analyzed for glucose consumption (n=4). Mean±SEM. ***, p<0.0002. (K) PC3 cells in normoxia (N) or hypoxia (H) were treated with vehicle control (Veh) or Akt inhibitor, MK2206 (1 µM), and analyzed for lactate production (n=3). Mean±SEM. **, p=0.001–0.004; ***, p=0.0005–0.0009. (L) PC3 cells in normoxia (N) or hypoxia (H) were transfected with control siRNA (Ctrl) or siRNA to Akt1 or Akt2 and analyzed for lactate production (n=2). Mean±SD. **, p=0.004; ***, p=0.0005. (M) PC3 cells stably silenced for PDK1 were transfected with vector (Vec), WT PDK1 or T346A PDK1 mutant, and analyzed for oxygen (O2) consumption (n=3). Mean±SEM. For all panels, data were analyzed using the two-sided unpaired Student’s t tests. See also Figure S3. Figure 4 Mitochondrial Akt-PDK1 phosphorylation, in vivo (A) GBM neurospheres (top) or differentiated GBM cultures (bottom) were stained for DNA (DAPI), HIF1α, pT346-phosphorylated PDK1, or hypoxia (hypoxia-sensitive probe). Merged images of nuclear-localized HIF1α in hypoxic neurospheres (by velocity mask) are indicated (Merge). Yellow box, Volocity analysis to identify cells with nuclear HIF1α in each single z-stack. Scale bar, 20 µm. (B and C) Immunohistochemical staining of primary, patient-derived GBM samples with high (≥2) (B) or low (0) (C) score for HIF1α and phosphorylated protein (pProt) expression. Scale bar, 100 µm. p, phosphorylated. (D–F) Quantitative immunohistochemical correlation of patient-derived GBM samples (n=24) or grade II gliomas (n=2) for HIF1α expression and pPDK1 (D), or pPDHE1 (E), or between pPDK1 and pPDHE1 (F). Four tissue microarray (TMA) cores/patient. The scoring is as follows: 0, no staining; 1, staining in at least one TMA core; 2, staining in ≥2 TMA cores. The individual p values per each analysis are indicated (Chi-Square test). See also Figure S4 and Table S2. Figure 5 Requirement of mitochondrial Akt for tumor cell proliferation in hypoxia (A and B) Bioluminescence imaging of immunocompromised mice carrying U251 intracranial GBMs (3 animals/group) expressing luciferase under the control of HIF1-responsive elements (Luc) and mCherry (cell viability) and exposed to a hypoxia-sensitive probe (Hypox). Scans were obtained at days 20 and 34 (A) and fluorescence signals were quantified (B). *, p=0.016–0.057 by Mann-Whitney test. (C) Tissue samples from intracranial GBMs as in (A) were harvested at day 34 and analyzed for expression of HIF1α, phosphorylated (p) PDK1 (pT346 Ab) or pPDHE1, by immunohistochemistry. Yellow lines were used to delineate the tumor mass within mice’ brain. Scale bar, 100 µm. Asterisks, mitotic cells; Insets (H&E and pPDK1 panels), high-power magnification of mitotic cells. Scale bar, 25 µm. (D and E) PC3 cells transfected with control siRNA (Ctrl) or Akt1- or Akt2-directed siRNA (D) or stably transduced with pLKO or PDK1-directed shRNA (E) were analyzed for cell proliferation in normoxia or hypoxia by direct cell counting (n=5). Mean±SEM. ***, p<0.001; **, p=0.002 (F and G) PC3 cells stably transduced with pLKO or PDK1-directed shRNA were analyzed in normoxia or hypoxia for colony formation by crystal violet staining after 10 days (F) and quantified (n=3) (G). Mean±SEM. ns, not significant. **, p=0.003. For all panels, data were analyzed using the two-sided unpaired Student's t test. See also Figure S5. Figure 6 Mitochondrial Akt regulation of stress signaling in hypoxia (A and B) PC3 cells in normoxia (N) or hypoxia (H) were treated with vehicle (Veh) or MK2206 (1 µM) (A) or transduced with pLKO or PDK1-directed shRNA (B) and analyzed for ROS production by CELLROX Green staining and flow cytometry. Upper panels, representative tracings. Bottom panels, quantification of ROS production under the various conditions tested (n=2). Mean±SD for both datasets. *, p=0.01–0.02; **, p=0.004; ns, not significant. (C) The experimental conditions are as in (A) and treated cells were analyzed for cell viability by direct cell counting relative to control (n=3). Mean±SEM. ***, p<0.0001. (D) PC3 cells in normoxia (N) or hypoxia (H) were incubated with vehicle (Veh) or small molecule inhibitors of Akt (MK2206, 1 µM) or PI3K (PX-866, 10 µM) and analyzed by Western blotting. (E) PC3 cells stably silenced for PDK1 were reconstituted with vector, WT PDK1 or T346A PDK1 mutant and analyzed for cell viability by direct cell counting relative to control (n=3). Mean±SEM. ***, p=0.0002. (F and G) PC3 cells in normoxia (N) or hypoxia (H) were transduced with pLKO or PDK1- directed shRNA (F) or control siRNA (Ctrl) or Akt1- or Akt2-directed siRNA (G), and analyzed by Western blotting. (H and I) PC3 cells as in (E) were analyzed for LC3 reactivity by fluorescence microscopy, Scale bars, 10 µm (H), and cells with LC3 puncta (>3) were quantified (n=250–860 cells) (I). Mean±SEM.*, p=0.014; ***, p=0.0005. ns, not significant. For all panels, data were analyzed using the two-sided unpaired Student's t test. See also Figure S6. Figure 7 Mitochondrial Akt-directed hypoxic reprogramming supports tumor growth in vivo (A) PC3 cells transduced with pLKO or PDK1-directed shRNA were injected s.c. in the flanks of male NSG immunocompromised mice (3 animals/group; 2 tumors/mouse) and superficial tumor growth was quantified with a caliper at the indicated time intervals for 20 days. Data were analyzed using the two-sided unpaired Student's t test. Mean±SEM. ***, p<0.0001. (B) PC3 cells stably transduced with pLKO or PDK1-directed shRNA were reconstituted with WT PDK1 or T346A PDK1 mutant and injected s.c. in the flanks of immunocompromised mice (5 mice/group; 2 tumors/mouse). Tumor growth in the various groups was quantified at the indicated time intervals for 20 days. Data were analyzed using the two-sided unpaired Student's t test. Mean±SEM. *, p=0.01–0.02; ***, p<0.0001. (C) PC3 cells stably transduced with pLKO or PDK1-directed shRNA were reconstituted with vector, WT PDK1 or T346A PDK1 mutant and injected s.c. in immunocompromised mice with determination of tumor growth after 18 days. Each point corresponds to an individual tumor. (D and E) Tumors harvested from the animals in (C) were analyzed for histology (D) and cell proliferation (top, Ki-67), autophagy (middle, LC3-II) or apoptosis (bottom, TUNEL) was quantified (E). The statistical analysis of the various groups by ANOVA is as follows: Ki-67, p<0.0001; LC3, p=0.024; TUNEL, p=0.039. Scale bars, 100 µm. (F and G) Superficial flank tumors of PC3 cells transduced with control pLKO or PDK1-directed shRNA were harvested after 18 day and processed for immunohistochemistry (F) with quantification of reactivity for Ki-67 (top), LC3 (middle) or TUNEL (bottom) (H). Representative images per each condition are shown. (n=3, 10 images per mouse), Mean±SD. Scale bars, 100 µm. (H) Schematic model of a mitochondrial Akt-PDK1-PDHE1 phosphorylation axis in hypoxic tumor reprogramming. Figure 8 Mitochondrial Akt phosphorylation of PDK1 is a negative prognostic marker in human gliomas (A) Representative micrographs of immunohistochemical staining of non-neoplastic human brain parenchyma (normal) or grade II-IV gliomas (WHO classification) with PDK1 pT346 Ab. OD, oligodendroglioma; AOD, anaplastic OD; GBM, glioblastoma. Scale bar, 100 µm. (B) Quantification of pT346 staining in a series of human brain tumors (n=116) and 85 nonneoplastic brain parenchyma using a two-factor scoring system that considers the percentage of positive cells and the intensity of the staining (pPDK1 score). ***, p<0.0001; **, p=0.002 by Mann Whitney U-test. Each symbol represents an individual patient. (C–E) Differences in pPDK1 score in human brain tumors as in (B) (n=116) according to nuclear HIF1α expression (C, **, p=0.008 by Mann Whitney U-test), IDH1 mutation status (D; *, p=0.02 by Mann Whitney U-test), or MGMT promoter methylation (D; *, p=0.01 by Mann Whitney U-test). Data are presented as Tukey box-and-whisker plots. The bottom and top of the box represent the first and third quartiles, and the band inside the box represents the median (i.e. the 2nd quartile). The bottom end of the whisker represents the lowest datum within the 1.5 interquartile range (IQR) of the lower quartile, and the top end of the whisker represents the highest datum within 1.5 IQR of the upper quartile. Outlier data, if any, are represented by single points. (F and G) Kaplan-Meier curves were generated with either the complete series of glioma patients (n=116; F) or with GBM cases only (n=61; G) sorted into “Low” or “High” groups according to pPDK1 score. Cutoffs to rank patients in these two categories were generated using ROC curves and the Youden’s J statistic. Overall survival curves were compared using the Log-Rank test. HR, Hazard Ratio; CI, Confidence Interval. See also Figure S7 and Table S3. SIGNIFICANCE The ability to flexibly adapt to an unfavorable microenvironment is a distinctive feature of tumor cells, engendering treatment resistance and unfavorable disease outcome. Low oxygen pressure, or hypoxia, is a powerful driver of tumor adaptation, but “druggable” therapeutic target(s) in this response have remained elusive. Here, we show that hypoxic tumors recruit a pool of active Akt to mitochondria, culminating with Akt phosphorylation of the metabolic gatekeeper, PDK1. This phosphorylation step improves tumor fitness, preserves tumor cell proliferation in the face of severe hypoxia, and is a negative prognostic factor in glioma patients. Repurposing small molecule Akt inhibitors currently in the clinic may provide an approach to prevent hypoxic reprogramming and improve anticancer therapy. ACCESSION NUMBERS The mass spectrometry proteomics data have been deposited to the MassIVE data repository (http://massive.ucsd.edu) with the MassIVE accession MSV000079671 and ProteomeXchange accession PXD004024. AUTHOR CONTRIBUTIONS Y.C.C. and D.C.A. conceived the project, Y.C.C., K.G.B., M.C.C., J.C.G., J.H.S. and S.L. performed experiments of Akt-PDK-PDHE1 phosphorylation, metabolic reprogramming, modulation of autophagy, mitochondrial localization, cell proliferation and survival, I.B. and V.V. performed experiments with GBM neurospheres and patient-derived GBM samples, M.L. provided primary, patient-derived GBM samples, L.O. performed experiments with an orthotopic mouse GBM model, H.Y.T and D.W.S. performed proteomics experiments, B.K. performed molecular modeling of the PDK1 phosphorylation site, A.V.K. analyzed bioinformatics data of global phosphoproteomics and proteomics identification of PDK1 phosphorylation, S.B., L.R.L., D.W.S. and D.C.A. analyzed data and Y.C.C., V.V. and D.C.A. wrote the paper. COMPETING FINANCIAL INTEREST The authors declare that no competing financial interest exists. REFERENCES Caino MC Chae YC Vaira V Ferrero S Nosotti M Martin NM Weeraratna A O'Connell M Jernigan D Fatatis A Metabolic stress regulates cytoskeletal dynamics and metastasis of cancer cells J Clin Invest 2013 123 2907 2920 23921130 Caino MC Ghosh JC Chae YC Vaira V Rivadeneira DB Faversani A Rampini P Kossenkov AV Aird KM Zhang R PI3K therapy reprograms mitochondrial trafficking to fuel tumor cell invasion Proc Natl Acad Sci U S A 2015 112 8638 8643 26124089 Cox TR Rumney RM Schoof EM Perryman L Hoye AM Agrawal A Bird D Latif NA Forrest H Evans HR The hypoxic cancer secretome induces pre-metastatic bone lesions through lysyl oxidase Nature 2015 522 106 110 26017313 De Bock K Georgiadou M Schoors S Kuchnio A Wong BW Cantelmo AR Quaegebeur A Ghesquiere B Cauwenberghs S Eelen G Role of PFKFB3-driven glycolysis in vessel sprouting Cell 2013 154 651 663 23911327 Denko NC Hypoxia, HIF1 and glucose metabolism in the solid tumour Nat Rev Cancer 2008 8 705 713 19143055 Di Cristofori A Ferrero S Bertolini I Gaudioso G Russo MV Berno V Vanini M Locatelli M Zavanone M Rampini P The vacuolar H+ ATPase is a novel therapeutic target for glioblastoma Oncotarget 2015 6 17514 17531 26020805 Du J Bernasconi P Clauser KR Mani DR Finn SP Beroukhim R Burns M Julian B Peng XP Hieronymus H Bead-based profiling of tyrosine kinase phosphorylation identifies SRC as a potential target for glioblastoma therapy Nat Biotechnol 2009 27 77 83 19098899 Dupuy F Tabaries S Andrzejewski S Dong Z Blagih J Annis MG Omeroglu A Gao D Leung S Amir E PDK1-Dependent Metabolic Reprogramming Dictates Metastatic Potential in Breast Cancer Cell Metab 2015 22 577 589 26365179 Fischer M Riemer J The mitochondrial disulfide relay system: roles in oxidative protein folding and beyond Int J Cell Biol 2013 2013 742923 24348563 Fruman DA Rommel C PI3K and cancer: lessons, challenges and opportunities Nat Rev Drug Discov 2014 13 140 156 24481312 Gardner LB Li Q Park MS Flanagan WM Semenza GL Dang CV Hypoxia inhibits G1/S transition through regulation of p27 expression J Biol Chem 2001 276 7919 7926 11112789 Gatenby RA Gillies RJ Why do cancers have high aerobic glycolysis? Nat Rev Cancer 2004 4 891 899 15516961 Ghosh JC Siegelin MD Vaira V Faversani A Tavecchio M Chae YC Lisanti S Rampini P Giroda M Caino MC Adaptive mitochondrial reprogramming and resistance to PI3K therapy J Natl Cancer Inst 2015 107 Gilkes DM Semenza GL Wirtz D Hypoxia and the extracellular matrix: drivers of tumour metastasis Nat Rev Cancer 2014 14 430 439 24827502 Gordan JD Bertout JA Hu CJ Diehl JA Simon MC HIF-2alpha promotes hypoxic cell proliferation by enhancing c-myc transcriptional activity Cancer Cell 2007 11 335 347 17418410 Graeber TG Osmanian C Jacks T Housman DE Koch CJ Lowe SW Giaccia AJ Hypoxia-mediated selection of cells with diminished apoptotic potential in solid tumours Nature 1996 379 88 91 8538748 Hitosugi T Fan J Chung TW Lythgoe K Wang X Xie J Ge Q Gu TL Polakiewicz RD Roesel JL Tyrosine phosphorylation of mitochondrial pyruvate dehydrogenase kinase 1 is important for cancer metabolism Mol Cell 2011 44 864 877 22195962 Hockel M Vaupel P Tumor hypoxia: definitions and current clinical, biologic, and molecular aspects J Natl Cancer Inst 2001 93 266 276 11181773 Jazwinski SM The retrograde response: When mitochondrial quality control is not enough Biochim Biophys Acta 2013 1833 400 409 22374136 Jiang Y Li X Yang W Hawke DH Zheng Y Xia Y Aldape K Wei C Guo F Chen Y PKM2 regulates chromosome segregation and mitosis progression of tumor cells Mol Cell 2014 53 75 87 24316223 Kaplon J Zheng L Meissl K Chaneton B Selivanov VA Mackay G van der Burg SH Verdegaal EM Cascante M Shlomi T A key role for mitochondrial gatekeeper pyruvate dehydrogenase in oncogene-induced senescence Nature 2013 498 109 112 23685455 Keith B Johnson RS Simon MC HIF1alpha and HIF2alpha: sibling rivalry in hypoxic tumour growth and progression Nat Rev Cancer 2012 12 9 22 Kim JW Tchernyshyov I Semenza GL Dang CV HIF-1-mediated expression of pyruvate dehydrogenase kinase: a metabolic switch required for cellular adaptation to hypoxia Cell Metab 2006 3 177 185 16517405 Liang J Mills GB AMPK: a contextual oncogene or tumor suppressor? Cancer Res 2013 73 2929 2935 23644529 Liu XD Yao J Tripathi DN Ding Z Xu Y Sun M Zhang J Bai S German P Hoang A Autophagy mediates HIF2alpha degradation and suppresses renal tumorigenesis Oncogene 2015 34 2450 2460 24998849 Lo Dico A Valtorta S Martelli C Belloli S Gianelli U Tosi D Bosari S Degrassi A Russo M Raccagni I Validation of an engineered cell model for in vitro and in vivo HIF-1alpha evaluation by different imaging modalities Mol Imaging Biol 2014 16 210 223 24002614 Louis DN Ohgaki H Wiestler OD Cavenee WK Burger PC Jouvet A Scheithauer BW Kleihues P The 2007 WHO classification of tumours of the central nervous system Acta Neuropathol 2007 114 97 109 17618441 Maes W Deroose C Reumers V Krylyshkina O Gijsbers R Baekelandt V Ceuppens J Debyser Z Van Gool SW In vivo bioluminescence imaging in an experimental mouse model for dendritic cell based immunotherapy against malignant glioma J Neurooncol 2009 91 127 139 18787761 Manning BD Cantley LC AKT/PKB signaling: navigating downstream Cell 2007 129 1261 1274 17604717 Mazumdar J O'Brien WT Johnson RS LaManna JC Chavez JC Klein PS Simon MC O2 regulates stem cells through Wnt/beta-catenin signalling Nat Cell Biol 2010 12 1007 1013 20852629 McFate T Mohyeldin A Lu H Thakar J Henriques J Halim ND Wu H Schell MJ Tsang TM Teahan O Pyruvate dehydrogenase complex activity controls metabolic and malignant phenotype in cancer cells J Biol Chem 2008 283 22700 22708 18541534 Michelakis ED Sutendra G Dromparis P Webster L Haromy A Niven E Maguire C Gammer TL Mackey JR Fulton D Metabolic modulation of glioblastoma with dichloroacetate Sci Transl Med 2010 2 31ra34 Nakahira K Kim HP Geng XH Nakao A Wang X Murase N Drain PF Wang X Sasidhar M Nabel EG Carbon monoxide differentially inhibits TLR signaling pathways by regulating ROS-induced trafficking of TLRs to lipid rafts J Exp Med 2006 203 2377 2389 17000866 Papandreou I Cairns RA Fontana L Lim AL Denko NC HIF-1 mediates adaptation to hypoxia by actively downregulating mitochondrial oxygen consumption Cell Metab 2006 3 187 197 16517406 Patel MS Nemeria NS Furey W Jordan F The pyruvate dehydrogenase complexes: structure-based function and regulation J Biol Chem 2014 289 16615 16623 24798336 Ravi R Mookerjee B Bhujwalla ZM Sutter CH Artemov D Zeng Q Dillehay LE Madan A Semenza GL Bedi A Regulation of tumor angiogenesis by p53- induced degradation of hypoxia-inducible factor 1alpha Genes Dev 2000 14 34 44 10640274 Roberts DJ Tan-Sah VP Smith JM Miyamoto S Akt phosphorylates HK-II at Thr-473 and increases mitochondrial HK-II association to protect cardiomyocytes J Biol Chem 2013 288 23798 23806 23836898 Santi SA Lee H The Akt isoforms are present at distinct subcellular locations Am J Physiol Cell Physiol 2010 298 C580 C591 20018949 Semenza GL HIF-1 mediates metabolic responses to intratumoral hypoxia and oncogenic mutations J Clin Invest 2013 123 3664 3671 23999440 Tredan O Galmarini CM Patel K Tannock IF Drug resistance and the solid tumor microenvironment J Natl Cancer Inst 2007 99 1441 1454 17895480 White E Deconvoluting the context-dependent role for autophagy in cancer Nat Rev Cancer 2012 12 401 410 22534666 Wick W Weller M van den Bent M Sanson M Weiler M von Deimling A Plass C Hegi M Platten M Reifenberger G MGMT testing--the challenges for biomarker-based glioma treatment Nat Rev Neurol 2014 10 372 385 24912512 Young JC Hoogenraad NJ Hartl FU Molecular chaperones Hsp90 and Hsp70 deliver preproteins to the mitochondrial import receptor Tom70 Cell 2003 112 41 50 12526792 Zhang SL Hu X Zhang W Yao H Tam KY Development of pyruvate dehydrogenase kinase inhibitors in medicinal chemistry with particular emphasis as anticancer agents Drug Discov Today 2015 20 1112 1119 25842042

PMC005xxxxxx/PMC5131925.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 101690683 45625 Lancet Gastroenterol Hepatol Lancet Gastroenterol Hepatol The lancet. Gastroenterology & hepatology 2468-1253 27917405 5131925 10.1016/S2468-1253(16)30015-2 NIHMS807722 Article Efficacy and safety of 3-week response-guided triple direct-acting antiviral therapy for chronic hepatitis C infection: a phase 2, open-label, proof-of-concept study Lau George * Benhamou Yves Chen Guofeng Li Jin Shao Qing Ji Dong Li Fan Li Bing Liu Jialiang Hou Jinlin Sun Jian Wang Cheng Chen Jing Wu Vanessa Wong April Wong Chris L P Tsang Stella T Y Wang Yudong Bassit Leda Tao Sijia Jiang Yong Hsiao Hui-Mien Ke Ruian Perelson Alan S Schinazi Raymond F * Division of Gastroenterology and Hepatology, Humanity and Health Medical Centre, Hong Kong, Hong Kong SAR, China (Prof G Lau MD, C Wang MD, J Chen PhD, V Wu BSc, A Wong BSc, Y Wang PhD); Second Liver Cirrhosis Diagnosis and Treatment Center (Prof G Lau, Prof G Chen MD, Prof Q Shao MD, D Ji MD, F Li MD, B Li MD, J Liu MD) and Institute of Infectious Disease (Prof J Li MD), 302 Hospital, Beijing, China; Service d’Hépatologie, Hôpital Pitié-Salpêtrière, Paris, France (Y Benhamou MD); State Key Laboratory of Organ Failure Research, Guangdong Provincial Key Laboratory of Viral Hepatitis Research, Department of Infectious Diseases, Nanfang Hospital, Southern Medical University, Guangzhou, China (Prof J Hou MD, Prof J Sun MD, C Wang); Hong Kong Molecular Pathology Diagnostic Centre, Hong Kong SAR, China (C L P Wong PhD, S T Y Tsang PhD); Center for AIDS Research, Department of Pediatrics, Emory University School of Medicine, Atlanta, GA, USA (L Bassit PhD, S Tao PhD, Y Jiang PhD, H-M Hsiao MS, Prof R F Schinazi PhD); Theoretical Biology and Biophysics, MS-K710, Los Alamos National Laboratory, Los Alamos, NM, USA (R Ke PhD, A S Perelson PhD); and Department of Mathematics, North Carolina State University, Raleigh, NC, USA (R Ke) Correspondence to: Prof Raymond F Schinazi, Emory University School of Medicine, Department of Pediatrics, 1760 Haygood Drive NE, Room E420, Atlanta, GA 30322, USA, rschina@emory.edu or Prof George Lau, Beijing 302 Hospital of PLA-Hong Kong-Humanity and Health Hepatitis C Diagnosis and Treatment Center, Beijing 302 Hospital, Beijing, 100039, China, gkklau@netvigator.com * Contributed equally to the study design and manuscript 21 8 2016 25 7 2016 10 2016 01 10 2017 1 2 97104 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. Summary Background To shorten the course of direct-acting antiviral agents for chronic hepatitis C virus (HCV) infection, we examined the antiviral efficacy and safety of 3 weeks of response-guided therapy with an NS3 protease inhibitor and dual NS5A inhibitor–NS5B nucleotide analogue. Methods In this open-label, phase 2a, single centre study, Chinese patients with chronic HCV genotype 1b infection without cirrhosis were randomly allocated by a computer program to one of three treatment groups (sofosbuvir, ledipasvir, and asunaprevir; sofosbuvir, daclatasvir, and simeprevir; or sofosbuvir, daclatasvir, and asunaprevir) until six patients in each group (1:1:1) achieved an ultrarapid virological response (plasma HCV RNA <500 IU/mL by day 2, measured by COBAS TaqMan HCV test, version 2.0). Patients with an ultrarapid virological response received 3 weeks of therapy. Patients who did not achieve an ultrarapid response were switched to sofosbuvir and ledipasvir for either 8 weeks or 12 weeks. The primary endpoint was the proportion of patients with a sustained virological response at 12 weeks (SVR12) after treatment completion, analysed in the intention-to-treat population. All patients who achieved an ultrarapid virological response were included in the safety analysis. This trial is registered with ClinicalTrials.gov, number NCT02470858. Findings Between April 5, 2015, and April 15, 2015, 26 eligible patients were recruited. 12 patients were assigned to sofosbuvir, ledipasvir, and asunaprevir; six to sofosbuvir, daclatasvir, and simeprevir; and eight to sofosbuvir, daclatasvir, and asunaprevir. Six patients in each group achieved an ultrarapid virological response (18 [69%]). All patients with an ultrarapid virological response who were given 3 weeks of triple therapy achieved SVR12. The most common adverse events were fatigue (one [17%] of six patients receiving sofosbuvir, ledipasvir, and asunaprevir; one [17%] of six patients receiving sofosbuvir, daclatasvir, and simeprevir; and two [33%] of six patients receiving sofosbuvir, daclatasvir, and asunaprevir) and headache (one [17%] patient in each group). No patients experienced any serious adverse events. Interpretation In this proof-of-concept study, all patients with chronic HCV without cirrhosis who achieved an ultrarapid virological response on triple direct-acting antiviral regimens by day 2 and received 3 weeks of treatment were cured, with excellent tolerability. By shortening the duration of therapy from the currently recommended 12 weeks to 3 weeks, we could drastically reduce the cost of therapy and the rate of adverse events. Further large-scale studies should be done to confirm our findings. Funding Center for AIDS Research, National Institutes of Health, US Department of Energy, National Center for Research Resources and the Office of Research Infrastructure Programs, Cheng Si-Yuan (China-International) Hepatitis Research Foundation, and Humanity and Health Medical Group. Introduction Treatment of hepatitis C virus (HCV) infection has entered a new era with the emergence of direct-acting antiviral agents. By 2015, the US Food and Drug Administration (FDA) and EU had approved three new direct-acting antiviral drugs—sofosbuvir, simeprevir, and daclatasvir—for the treatment of HCV infection as part of combination regimens.1,2 91–100%3–5 of individuals infected with HCV genotype 1 treated with 8–12 weeks of sofosbuvir and ledipasvir once daily, and 98–100%6 of patients who received 12–24 weeks of sofosbuvir and daclatasvir once daily, achieved sustained virological response at 12 weeks (SVR12). The cost of such regimens is onerous7 and this has adversely affected treatment access and drug compliance, and has encouraged drug counterfeiting.8,9 A major challenge is to reduce treatment cost without affecting efficacy by shortening the duration of treatment.10,11 Attempts to reduce the duration of therapy to 6 weeks through the addition of ribavirin to sofosbuvir and ledipasvir resulted in many patients relapsing after treatment,12 but the addition of an experimental NS3/4A protease inhibitor (GS-9451) or an experimental non-nucleoside polymerase inhibitor (GS-9669) yielded an SVR12 in 95% of patients.13 Mathematical modelling of HCV RNA changes during therapy suggests that a more rapid second-phase viral decline should allow for a shorter treatment duration.14 Nucleoside analogue inhibitors do not generate fast second-phase declines such as those with HCV protease inhibitors.14–17 Thus, we postulated that addition of an approved protease inhibitor, such as simeprevir or asunaprevir to sofosbuvir and ledipasvir or daclatasvir might induce a more rapid second-phase HCV RNA decline, allowing for a shorter treatment duration. Therefore, we did a proof-of-concept, response-guided therapy clinical study to investigate the efficacy and the safety of 3 weeks of triple direct-acting antiviral therapy, containing NS5B, NS3, and NS5A inhibitors, in Chinese patients with chronic HCV genotype 1b infection without cirrhosis who achieved an ultrarapid initial viral response (uRVR), defined by a serum HCV RNA lower than 500 IU/mL within the first 2 days of dosing. We focused on patients infected with genotype 1b because this is the predominant strain in Asian populations (estimated at about 50 million), although this genotype can also be found in other populations.18,19 Methods Study design and participants This was an open-label, proof-of-concept, phase 2a study done at a single centre (Humanity and Health Medical Centre, Hong Kong SAR, China). Participants were identified using the Beijing 302 Hospital of PLA–Hong Kong Humanity and Health Medical Group, Hepatitis C Diagnosis and Treatment Centre database, which had records for 503 patients infected with HCV by the time of the study. People who satisfied all the inclusion and exclusion criteria and who consented were consecutively enrolled. Key inclusion criteria were: older than 18 years; documented chronic HCV genotype 1b infection for more than 6 months; a baseline plasma HCV RNA concentration of 104–107 IU/mL; and absence of cirrhosis as assessed by liver biopsy or by liver stiffness measurement less than 12·5 kPa. Key exclusion criteria were: hepatitis B virus (HBV) or HIV infection; chronic liver disease of a non-HCV aetiology; hepatocellular carcinoma or other malignancy; drug or alcohol misuse; pregnant or nursing woman; known hypersensitivity to pharmaceutical products used in this study; and any other medical disorders or clinical conditions (eg, substantial cardiopulmonary, neurological, renal, haematological, autoimmune disorders, and any malignancy) that could interfere with the study. Treatment-experienced patients had been exposed to interferon-based therapy previously. Written informed consent was obtained from all patients. The study was approved by the independent ethics committee at the study centre and was done in compliance with the Declaration of Helsinki, Good Clinical Practice guidelines, and local regulatory requirements. An independent data and safety monitoring committee reviewed the progress of the study. The full protocol is included in the appendix. Randomisation and masking All eligible patients were randomly allocated by a computerised system to one of three treatment groups (sofosbuvir, ledipasvir, and asunaprevir; sofosbuvir, daclatasvir, and simeprevir; or sofosbuvir, daclatasvir, and asunaprevir) until six patients achieved a uRVR in each group. The computer sequence was generated by a biostatistician (JC) who assigned them to trial groups but was not involved in the rest of the trial. The trial was open label; patients and investigators were aware of group assignment. Procedures Patients received sofosbuvir, ledipasvir, and asunaprevir; sofosbuvir, daclatasvir, and simeprevir; or sofosbuvir, daclatasvir, and asunaprevir. Doses were as follows: sofosbuvir 400 mg once daily; ledipasvir 90 mg once daily; daclatasvir 60 mg once daily; simeprevir 150 mg once daily, and asunaprevir 100 mg twice daily. Patients with a uRVR were treated for 3 weeks. Patients who did not achieve a uRVR were switched to sofosbuvir and ledipasvir for either 8 weeks or 12 weeks and followed up (although not included in subsequent studies). A plasma HCV RNA threshold of less than 500 IU/mL by day 2 was chosen because in China this is the most commonly used initial screening criteria, although more sensitive assays are available. Additionally, our preliminary data (not shown) suggested that using this threshold at 48 h was a surrogate for SVR12 with 3 weeks of pan-oral direct-acting antiviral therapy, irrespective of previous interferon-based therapy response. Screening assessments included measurement of the plasma HCV RNA concentration, IL28B genotyping with a TaqMan genotyping assay (Applied Biosystems, Foster City, CA, USA) for the rs12979860 single-nucleo tide polymorphism, IFNL4 (rs368234815) genotyping, and standard laboratory tests. Plasma HCV RNA concentration was measured at baseline (0 h), 1 h, 2 h, 4 h, 8 h, and 24 h after initial dosing, and at days 2, 4, 7, 14, and 21 or at the end of treatment. After the end of treatment all patients had plasma HCV RNA measured at 4-week intervals until week 12. Plasma HCV RNA concentrations were measured by a COBAS TaqMan 48 analyser, version 2.0 (Roche Molecular Systems, Branchburg, NJ, USA), with a lower limit of quantification of 25 IU/mL and a lower limit of detection of 6 IU/mL. HCV genotype and subtype were determined at screening using the Versant HCV Genotype INNOLiPA 2.0 assay (Siemens Healthcare Diagnostics, Tarrytown, NY, USA). Population sequencing was used to determine resistance-associated variants in all patients at baseline, and again at the time of failure in those who had virological failure.20 Virological failure resulting in study drug discontinuation was defined as: failure to achieve plasma HCV RNA concentrations less than 25 IU/mL during the 3 weeks of therapy; confirmed HCV RNA concentrations of 25 IU/mL or more at two consecutive measurements at any point in patients with on-treatment HCV RNA concentrations less than 25 IU/mL; or a confirmed increase in HCV RNA at two consecutive measurements of greater than 1 log10 IU/mL above the nadir at any timepoint during treatment. Liver stiffness was measured with FibroScan (Echosens, Paris, France) according to the manufacturer’s instructions. Vital signs and symptom-directed physical examinations and adverse events graded according to the NIAID Division of AIDS toxicity table (version 1.0) were recorded at baseline, day 2, day 4, day 7, week 2, and week 3 during treatment, and then at weeks 4 and 12 after the end of treatment. Blood was taken for haematology and chemistry at screening, baseline, day 2, day 4, day 7, week 2, and week 3 during treatment, and then at weeks 4 and 12 after the end of treatment. 12-lead electrocardiograms were done at screening, baseline, week 2, and week 3 during treatment, and then at weeks 4 and 12 after the end of treatment. Plasma samples from patients infected with HCV were stored in aliquots at −80°C until further processing. Viral RNA was extracted and the genome sequenced using an Applied Biosystems 3730×l DNA analyser as previously described.21 Data were analysed with the ABL-DeepChek-HCV 1.4 software (ABL SA, Luxembourg, Luxembourg). To confirm compliance, a liquid chromatographytandem mass spectrometry (LC-MS/MS) method was developed to detect and quantify the anti-HCV drug concentrations in plasma collected from each patient. Briefly, blood samples were collected from 25 patients at 8 h, 24 h, and 48 h after initiating treatment. 50 µL of plasma was extracted with 3 mL of ethyl acetate containing abacavir (as an internal standard). The supernatant was collected, air dried, then reconstituted in 2 mmol/L ammonium acetate solution with 75% methanol. Five target compounds were simultaneously monitored and quantified by LC-MS/MS in the multiple reaction monitoring mode, including sofosbuvir (m/z 530·2/243·2), daclatasvir (m/z 739·5/339·3), ledipasvir (m/z 889·4/637·3), asunaprevir (m/z 748·4/648·4), and simeprevir (m/z 750·4/315·2). A Dionex Ultimate 3000 HPLC system (Thermo Scientific, Waltham, MA, USA) coupled with an AB SCIEX API5000 triple quadrupole mass spectrometer (AB SCIEX, Framingham, MA, USA) was used for analysis. Outcomes The primary efficacy endpoint was the proportion of individuals achieving SVR12 as defined by HCV RNA below the lower limit of quantification (25 IU/mL). The main safety endpoint was the frequency and severity of adverse events. Secondary endpoints were: the proportion of patients with undetectable HCV viral load at specified timepoints during treatment (day 2, day 4, day 7, week 2, and week 3) and after treatment (week 4 and week 12); the kinetics of circulating HCV RNA from baseline; the proportion of patients with adverse events; discontinuation rates related to adverse events; safety laboratory changes; and the occurrence of HCV resistance mutations. Statistical analysis A minimum of 14 patients were needed to power this proof-of-concept clinical study and therefore guide the decision for comparative late-stage clinical studies. Considering the feasibility and cost of the study, and accounting for dropouts, a sample size of 18 patients with six in each treatment group was selected to meet power calculation requirements. Assuming at least a 1% increase in SVR12 rate in each group from the literature,13 with a sample size of six in each group, the power to detect such an increase ranged from 78% to 99% at the α level of 0·05, depending on the magnitude of the increase. Baseline characteristics between groups were compared using the Kruskal-Wallis test for continuous outcomes and Fisher’s exact tests for binary outcomes. The difference in HCV viral load decline between groups was compared using the Kruskal-Wallis test and corrected by Dunn’s test for multiple comparison between groups. HCV RNA concentrations and liver stiffness before treatment, at the end of treatment, and 12 weeks after treatment were compared using Wilcoxon signed-rank test for each group. Statistical analysis was based on the intention-to-treat principle. Since this was a phase 2a proof-of-concept study, the small sample size did not allow the analysis of predictive factors for uRVR using a multivariate model. Statistical analyses were done with Stata (release 13; StataCorp). Viral kinetic modelling was done with a multiscale model22,23 for all 26 participants (appendix p 12). Role of the funding source The funder of the study had no role in study design, data collection, data analysis, data interpretation, or writing of the report. All authors had full access to all the data in the study and had final responsibility for the decision to submit for publication. Results Between April 5, 2015, and April 15, 2015, 27 patients with chronic hepatitis C were assessed for eligibility. One patient with cirrhosis (liver stiffness >12·5 kPa) was excluded and 26 were randomly assigned to the three treatment groups (figure 1). Six of 12 (50%, 95% CI 25–75) patients receiving sofosbuvir, ledipasvir, and asunaprevir, six of six (100%, 61–100) receiving sofosbuvir, daclatasvir, and simeprevir, and six of eight (75%, 41–95) receiving sofosbuvir, daclatasvir, and asunaprevir achieved plasma HCV RNA concentrations less than 500 IU/mL by day 2, and subsequently received triple direct-acting antiviral therapy for 3 weeks (p=0·10). Baseline characteristics of the patients who achieved a uRVR are shown in table 1. All participants were Chinese, six (33%) were male, seven (39%) had the IL28 non-CC genotype, and eight (44%) had baseline HCV RNA less than 800 000 IU/mL. Four (22%) patients had stage 2 liver fibrosis. Patients in each of the 3-week treatment groups were well matched except for body-mass index. The eight patients who did not achieve uRVR were switched to sofosbuvir and ledipasvir by day 3 and followed up, but were not included in subsequent studies, and hence had a choice of duration of therapy of 8 weeks (n=4) or 12 weeks (n=4). Patients with uRVR had a significantly lower mean baseline HCV RNA than those without uRVR (7·0 log10 IU/mL [95% CI 6·8–7·2] vs 6·0 log10 IU/mL [5·6–6·3], p<0·0001; appendix p 2). The eight patients without uRVR were not considered in the efficacy and safety analyses. However, none of them had grade 3–4 adverse events and all of these patients achieved SVR12. At week 3, all 18 patients who achieved uRVR had undetectable plasma HCV RNA irrespective of the allocated direct-acting antiviral regimen. The median time to achieve plasma HCV RNA less than 25 IU/mL was shorter in patients receiving sofosbuvir, ledipasvir, and asunaprevir than in patients receiving sofosbuvir, daclatasvir, and asunaprevir (4 days [IQR 0] vs 14 days [7]; p=0·01). Plasma HCV RNA remained below the limit of quantification up to week 12 after completion of treatment (table 2). The median time to achieve plasma HCV RNA less than 25 IU/mL was significantly shorter in patients who achieved uRVR (7 days [IQR 10]) than in those who did not achieve this threshold (17·5 days [10·5]; p=0·003; appendix p 3). A post-hoc analysis of liver stiffness showed a significant decrease from baseline to 12 weeks post-treatment in patients receiving sofosbuvir, ledipasvir, and asunaprevir (6·1 kPa vs 4·9 kPa, p=0·028) and in patients receiving sofosbuvir, daclatasvir, and simeprevir (5·8 kPa vs 5·1 kPa, p=0·046), but not in patients receiving sofosbuvir, daclatasvir, and asunaprevir (5·5 kPa vs 5·4 kPa, p=0·53). The effect of overall treatment on HCV clearance was estimated by taking into account the effectiveness of the regimen in inhibiting HCV RNA synthesis (εα) and virion secretion or export (εS), with ε=1 being 100% effective and enhancing intracellular HCV RNA decay (by a factor κ). The model fit individual patient data well as shown in the appendix (p 9). HCV RNA kinetics exhibit a three-phase viral decline, with a rapid first phase, moderate decay during an intermediate phase, and then a slower third phase. No significant difference was recorded in the decline rate of the mean (figure 2) and median HCV load between the three treatment groups (appendix p 8). Based on 18 patients with uRVR, the estimated mean effectiveness of the treatments was εα=0·9962 and εS=0·9987 (appendix p 7). The mean estimated rate of viral clearance was 29·7 per day and the mean effectiveness of the direct-acting antiviral therapy in enhancing intracellular HCV RNA degradation, κ, was 1·49, consistent with previous estimates.22 Using data from all 26 treated patients, parameter values were not distributed differently among patients with uRVR and those without uRVR (data not shown). However, the estimated baseline viral load (V0) was significantly lower in those with uRVR than in those without uRVR (log10 V0 5·81 vs 7·23, p<0·0001; appendix p 11). Resistance-associated variant analysis was done at baseline to elucidate the naturally occurring resistance profile. Since virus clearance was achieved rapidly in all patients during the treatment, it was difficult to do resistance-associated variant analysis at other timepoints. Sequencing of NS3/4A, NS5A, and NS5B was successful in all 26 baseline samples. Five naturally occurring protease inhibitor resistance-associated variants (V36L, T54S, S122R, I132V, or D168H/N/Y) were identified in NS3/4A in 23 (88%) of 26 baseline samples. The I132V variant was found in all 23 samples and among them, five had one additional mutation (two D168H/N/Y, one S122R, one T54S, and one V36L). Four NS5A variants associated with resistance to daclatasvir or ledipasvir (Q30R, L31F/V, H54N, or Y93H) were identified in 23 (88%) of 26 samples. The Q30R variant was found in all 23 samples and five had one or two additional mutations (appendix p 13). Other minor mutations are summarised in the appendix (p 13). None had the 282T mutation associated with sofosbuvir resistance. Plasma drug concentrations provided a confirmation of drug delivery and patient compliance during the study (appendix pp 14, 15). Drugs taken by patients were consistent with their regimen, with only a few exceptions noted in which plasma drug concentrations were lower than the limit that could be quantified by HPLC-MS. Of the 18 patients who completed treatment, the most common adverse events were headache and fatigue (table 3), and all adverse events were mild to moderate. Two patients had grade 3 laboratory abnormalities: one had anaemia due to menorrhagia related to uterine fibroids and another had transient hyperbilirubinaemia. The cause of the transient hyperbilirubinaemia is unknown. Previously, this adverse event was also reported in a clinical study of daclatasvir plus asunaprevir.24 In our study, it was only recorded in one of six patients receiving sofosbuvir, ledipasvir, and asunaprevir. No grade 4 laboratory abnormalities were reported. Discussion In this exploratory cohort of Chinese patients with chronic genotype 1b HCV who had no cirrhosis and had a uRVR to triple direct-acting antiviral regimens, we show that these regimens are well tolerated and a 100% cure rate is achievable. This study reports the shortest treatment duration of pan-oral direct-acting antivirals used for patients with chronic HCV infection. The incidence of adverse events was low compared with other reported studies on pan-oral direct-acting antiviral agents,3–6 which might be related to the shorter duration of therapy. The patients are still being followed up and no relapse or adverse events have been recorded as of March, 2016. Current recommendations for genotype 1b patients without cirrhosis include: 8–12-week therapy with ledipasvir and sofosbuvir; or 12-week therapy with paritaprevir with ritonavir plus ombitasvir and dasabuvir, or sofosbuvir plus simeprevir, or sofosbuvir plus daclatasvir, or grazoprevir (MK-5172) and elbasvir (MK-8742), yielding an SVR12 in more than 95% of recipients.1,2,25 A major factor restricting the wide availability of direct-acting antivirals to patients with HCV is the high cost of treatment, about US$200–1200 per day, even when discounted.10 Several clinical trials were designed to shorten treatment duration to 6–8 weeks, with more than an 80–90% success rate.5,13,21 However, attempts to further shorten therapy to 4 weeks have so far failed. In the C-SWIFT genotype 1 study,26 only three of the five patients with genotype 1b chronic hepatitis C without cirrhosis achieved an SVR at 4 weeks or 8 weeks after the end of therapy with a 4-week fixed-dose combination of grazoprevir (MK-5172), elbasvir (MK-8742), and sofosbuvir 400 mg once daily. Similarly, in the NIH/UMD Synergy trial27 of patients with early F0–F2 fibrosis treated with ledipasvir and sofosbuvir plus GS-9451 and GS-9669, only five (20%) of 25 achieved SVR12, whereas ten (40%) of 25 in another cohort treated with ledipasvir and sofosbuvir plus GS-9451 achieved an SVR12. Modelling of HCV RNA kinetics during anti-HCV drug therapy has predicted faster first phases in protocols using an NS5A inhibitor, such as daclatasvir22 or ledipasvir,28 and faster second phases using a protease inhibitor.17 This finding together with the use of sofosbuvir,29 should allow a shorter treatment duration. Our pilot study strongly suggests that combining the three different approved oral anti-HCV drug classes leads to a complete cure (SVR12) in 3 weeks in all patients with genotype 1b HCV without cirrhosis who had plasma HCV RNA reduced to less than 500 IU/mL within the first 2 days of therapy. No differences were observed between the three combinations of the drugs used in both day 2 response and SVR12. Nevertheless, patients receiving sofosbuvir, ledipasvir, and asunaprevir had shorter time to antiviral efficacy than did patients receiving sofosbuvir, daclatasvir, and asunaprevir; the reason is unclear. The most apparent distinction between these two groups is the inclusion of ledipasvir (sofosbuvir, ledipasvir, and asunaprevir) versus daclatasvir (sofosbuvir, daclatasvir, and asunaprevir). Review of the scientific literature only found one related comparison, daclatasvir plus asunaprevir versus sofosbuvir and ledipasvir for hepatitis C genotype 1 in Japanese patients, but this was an indirect comparison.30 No direct comparison of ledipasvir with daclatasvir has been done. A substantial proportion of the enrolled individuals (18 [69%] of 26) achieved a very rapid and profound drop in plasma HCV RNA. We found that this rapid viral decline could be affected by the presence of baseline NS5A resistance-associated variants (appendix p 13). The reasons for the naturally occurring resistance-associated variants are still unknown. However, in our study, baseline resistance-associated variants did not affect SVR12 in patients, irrespective of whether they had a uRVR. Even for patients who did not have a uRVR, baseline resistance-associated variants did not affect SVR12 in patients treated for longer than 3 weeks (8 weeks or 12 weeks). To our knowledge, this study is the first to use the principle of response-guided therapy with an all direct-acting antiviral regimen to shorten treatment duration, with a high proportion of patients achieving an SVR12. This concept, if validated in larger studies, could greatly reduce duration of therapy and subsequently compliance and cost, and improve the accessibility and affordability of direct-acting antiviral drugs, especially in middle-income and low-income countries. Future larger studies are currently being planned in Mongolia to shorten the duration of treatment with sofosbuvir, daclatasvir, and simeprevir to 3 weeks. Additionally, this should also reduce the emergence of resistance, development of side-effects, and help to curtail the use of counterfeit direct-acting antiviral drugs in some countries, which could be potentially harmful. Our findings are not consistent with the concept that cure corresponds to an end of treatment viral load below a cure boundary of less than one viral particle in the extracellular body fluid (ie, 15 L), which corresponds to a concentration of about 10−4·22 IU/mL.13,14,31 The multiscale model predicted that none of the 18 treated patients would reach the cure boundary (appendix p 9). This finding prompts a reconsideration of whether reducing the amount of virus to less than one virion is really necessary to achieve a cure. A weakness of our study is the small number of patients, and the results should be validated in larger studies. The small number of individuals in each group limits a comparison of different direct-acting antiviral regimens, and a direct comparison might be warranted in future studies. Furthermore, the patients were Chinese with genotype 1b HCV infection and without cirrhosis, of whom up to 85% had the IL28B CC genotype. IL28B CC genotype and genotype 1b HCV infection are predictive factors for an SVR, and some might argue that dual direct-acting antiviral therapies could achieve a good response in this population. Treatment with direct-acting antiviral drugs has been reported to increase the risk of recurrence of hepatocellular carcinoma in patients with previous HCV-related hepatocellular carcinoma.32 However, to our knowledge, this study is the first to report that SVR12 can be achieved after 3 weeks of direct-acting antiviral therapies. Normally, interferon-based treatment can achieve SVR at 24 weeks in 44–79% of patients with HCV genotype 1 infection,33 but still needs more than 48 weeks of treatment. Limited data support the use of dual direct-acting antiviral therapies to shorten treatment duration. Since nucleoside analogue inhibitors do not generate second-phase declines as fast as HCV protease inhibitors, the addition of a protease inhibitor to sofosbuvir and ledipasvir or daclatasvir could allow for a shorter treatment duration. This hypothesis is supported by our data. All patients are being followed up and long-term follow-up data will be reported later. In our study, a high proportion of patients were young women with a low baseline HCV viral load. These characteristics could be associated with a favourable response to short duration pan-oral direct-acting antiviral therapy to less than 6 weeks.27 Nevertheless, it has been estimated that there are at least 4 million people with such characteristics in China,34 and we believe these findings could benefit more patients with chronic hepatitis C worldwide in the same setting. Application of this response-guided treatment approach to patients with different ethnic backgrounds, with different genotypes, and with cirrhosis should be addressed by larger clinical studies. Supplementary Material 1 We thank Ruy M Ribeiro, Los Alamos National Laboratory, and Becky Kinkead and James Kohler at Emory University for helpful editorial comments. We thank ABL TherapyEdge’s team (ABL SA, Luxembourg), including Dimitri Gonzalez, Virginie Duc, Ronan Boulme, and Chalom Sayada for their outstanding support with sequence analysis processing. The salary of RFS is supported in part by Center for AIDS Research NIH grant 5P30-AI-50409. Portions of this work were done under the auspices of the US Department of Energy under contract DE-AC52-06NA25396 and supported by NIH grants R01-AI028433, R01-AI078881, and the National Center for Research Resources and the Office of Research Infrastructure Programs through grant R01-OD011095 (to ASP). Figure 1 Trial profile uRVR=ultrarapid virological response. Figure 2 Decline in mean hepatitis C viral load Solid lines are mean model trajectories calculated from predicted viral load (see appendix p 9). Dashed horizontal line indicates the assay lower limit of quantification. Table 1 Baseline demographic and clinical characteristics of the patients who achieved an ultrarapid virological response Sofosbuvir, ledipasvir, and asunaprevir 3 weeks (n=6) Sofosbuvir, daclatasvir, and simprevir 3 weeks, (n=6) Sofosbuvir, daclatasvir, and asunaprevir 3 weeks (n=6) Age (years [mean, range]) 41 (25–66) 40 (23–59) 31 (21–47) Sex   Male 2 (33%) 2 (33%) 2 (33%)   Female 4 (67%) 4 (67%) 4 (67%) BMI (kg/m2 [mean, range]) 23·7 (4·5) 21·1 (2·6) 19·2 (2·0) HCV RNA (log10 IU/mL)   Mean (SD) 6·3 (0·3) 5·7 (1·0) 5·9 (0·8)   Median (IQR) 6·2 (6·1–6·6) 5·6 (5·3–6·7) 5·9 (5·2–6·5) HCV RNA ≥800 000 IU/mL 5 (83%) 2 (33%) 3 (50%) Previous treatment   Treatment experienced 3 (50%) 2 (33%) 1 (17%)   Treatment naive 3 (50%) 4 (67%) 5 (83%) Previous response*   Relapser 1 (33%) 2 (100%) 1 (100%)   Partial responder 1 (33%) 0 0   Null responder 1 (33%) 0 0 IL28B genotype   CC 4 (67%) 4 (67%) 3 (50%)   CT 2 (33%) 2 (33%) 3 (50%) IFNL4 genotype   TT/TT 4 (67%) 4 (67%) 5 (83%)   ΔG/TT 2 (33%) 2 (33%) 1 (17%) Liver stiffness measure (kPa) 6·1 (1·7) 5·8 (2·0) 5·5 (0·9) Fibrosis stage (METAVIR)†   F0–F1 (LSM≤7·0 kPa) 4 (67%) 4 (67%) 6 (100%)   F2 (7·0 kPa<LSM≤9·5 kPa) 2 (33%) 2 (33%) 0   F3 (9·5 kPa<LSM≤12·5 kPa) 0 0 0 Data are mean (SD) or n (%), unless otherwise stated. BMI=body-mass index. HCV=hepatitis C virus. * Data indicate the response to the most recent previous regimen: a relapser is a patient who received at least 36 weeks of pegylated interferon and ribavirin with HCV DNA undetectable at the end of treatment, but detectable within 52 weeks of follow-up; a partial responder is a patient who received at least 20 weeks of pegylated interferon and ribavirin for the treatment of HCV and achieved ≥2 log10 IU/mL reduction in HCV RNA at week 12, but failed to achieve HCV RNA undetectable at the end of treatment; a null responder is a patient who received at least 12 weeks of pegylated interferon and ribavirin and failed to achieve a 2 log10 IU/mL reduction in HCV RNA at week 12. † Liver stiffness was determined by Fibroscan; no patient in this study had cirrhosis (fibrosis score >12·5 kPa). Table 2 Virological response of individuals who achieved an ultrarapid virological response at day 2 and received 3 weeks of treatment Sofosbuvir, ledipasvir, and asunaprevir (3 weeks) Sofosbuvir, daclatasvir, and simprevir (3 weeks) Sofosbuvir, daclatasvir, and asunaprevir (3 weeks) During treatment period Day 2 1 (16·7% [3·0–56·3]) 2 (33·3% [9·7–70·0]) 0 Day 4 5 (83·3% [43·6–96·9]) 2 (33·3% [9·7–70·0]) 0 Day 7 6 (100·0% [61·0–100·0]) 4 (66·7% [30·0–90·3]) 2 (33·3% [9·7–70·0]) Week 2 6 (100·0% [61·0–100·0]) 6 (100·0% [61·0–100·0]) 5 (83·3% [43·6–96·6]) Week 3 6 (100·0% [61·0–100·0]) 6 (100·0% [61·0–100·0]) 6 (100·0% [61·0–100·0]) During post-treatment period Week 4 6 (100·0% [61·0–100·0]) 6 (100·0% [61·0–100·0]) 6 (100·0% [61·0–100·0]) Week 12 6 (100·0% [61·0–100·0]) 6 (100·0% [61·0–100·0]) 6 (100·0% [61·0–100·0]) The lower limit of quantification for plasma HCV RNA by Roche COBAS TaqMan test was 25 IU/mL. Table 3 Adverse events and laboratory abnormalities during the treatment period Sofosbuvir, ledipasvir, and asunaprevir (n=6) Sofosbuvir, daclatasvir, and simeprevir (n=6) Sofosbuvir, daclatasvir, and asunaprevir (n=6) Common adverse events Fatigue 1 (17%) 1 (17%) 2 (33%) Nausea 1 (17%) 0 0 Headache 1 (17%) 1 (17%) 1 (17%) Dizziness 0 1 (17%) 0 Insomnia 1 (17%) 0 1 (17%) Abdominal pain 0 1 (17%) 1 (17%) Constipation 1 (17%) 0 0 Diarrhoea 0 1 (17%) 1 (17%) Dermatitis 0 1 (17%) 1 (17%) Common cold 1 (17%) 0 1 (17%) Laboratory abnormalities Any grade 3 laboratory abnormality during treatment 1 (17%) 1 (17%) 0 Decreased haemoglobin 0 1 (17%) 0 Raised total bilirubin 1 (17%) 0 0 There were no deaths or discontinuations due to adverse events. All adverse events and serious adverse events were recorded from the time the consent form was signed until 14 days after cessation of treatment. The relatedness (probable or possible) of the adverse event to the regimen was determined by the investigator; patients could have had more than one adverse event. Research in context Evidence before this study We searched PubMed from Jan 1, 2010, to April 14, 2016, for articles published in English using a combination of the medical subject headings “HCV treatment” and “antiviral agent” as search terms. Two related clinical trials have been reported for hepatitis C virus (HCV) combination therapy studies (phase 2a). These trials have shown promising safety and efficacy using combination direct-acting antiviral drugs. Added value of this study Although our study is small, we showed high rates of sustained viral response at 12 weeks with use of the 3-week triple direct-acting antiviral therapy, which supports the possibility that this ultrashort regimen might be effective for some patients infected with HCV. Implications of all the available evidence Our study demonstrated that the duration of therapy of pan-oral direct-acting antiviral agents could be drastically shortened from the current recommended 12 weeks to only 3 weeks with triple direct-acting antiviral therapy, containing NS5B, NS3, and NS5A inhibitors, in non-cirrhotic Chinese patients infected with chronic HCV genotype 1b and who had an ultrarapid virological response. This could have great therapeutic, public health, and economic implications because genotype 1b rather than 1a is the predominant strain in Asian populations (estimated at about 50 million). Further large-scale studies with the use of a response-guided approach are needed in other populations with different genotypes and in cirrhotic patients See Online for appendix Contributors GL, YB, and RFS designed the study. JC randomly assigned the patients. GL, YB, RFS, GC, JLi, QS, DJ, FL, BL, JLiu, JH, JS, CW, VW, AW, and YW recruited the patients or did the study. CLPW, STYT, YW, LB, ST, YJ, and H-MH did the laboratory work, RK and ASP did the biostatistics analysis and viral kinetic mathematical modelling. GL and RFS wrote the manuscript and all authors contributed to the final version of the manuscript. Declaration of interests RFS was involved in the discovery of sofosbuvir. All other authors declare no competing interests. References 1 AASLD–IDSA HCV Guidance Panel Hepatitis C guidance: AASLD–IDSA recommendations for testing, managing, and treating adults infected with hepatitis C virus Hepatology 2015 62 932 954 26111063 2 European Association for the Study of the Liver EASL recommendations on treatment of hepatitis C 2015 J Hepatol 2015 63 199 236 25911336 3 Afdhal N Zeuzem S Kwo P Ledipasvir and sofosbuvir for untreated HCV genotype 1 infection N Engl J Med 2014 370 1889 1898 24725239 4 Afdhal N Reddy KR Nelson DR Ledipasvir and sofosbuvir for previously treated HCV genotype 1 infection N Engl J Med 2014 370 1483 1493 24725238 5 Kowdley KV Gordon SC Reddy KR Ledipasvir and sofosbuvir for 8 or 12 weeks for chronic HCV without cirrhosis N Engl J Med 2014 370 1879 1888 24720702 6 Sulkowski MS Gardiner DF Rodriguez-Torres M Daclatasvir plus sofosbuvir for previously treated or untreated chronic HCV infection N Engl J Med 2014 370 211 221 24428467 7 Jayasekera CR Barry M Roberts LR Nguyen MH Treating hepatitis C in lower-income countries N Engl J Med 2014 370 1869 1871 24720680 8 McCarthy M Fake medicines are undermining global efforts to combat infectious disease, says US journal BMJ 2015 350 h2137 25899508 9 Attaran A Stopping murder by medicine: introducing the Model Law on Medicine Crime Am J Trop Med Hyg 2015 92 6 suppl 127 132 25897071 10 Hagan LM Sulkowski MS Schinazi RF Cost analysis of sofosbuvir/ribavirin versus sofosbuvir/simeprevir for genotype 1 hepatitis C virus in interferon-ineligible/intolerant individuals Hepatology 2014 60 37 45 24677184 11 Hill A Khoo S Fortunak J Simmons B Ford N Minimum costs for producing hepatitis C direct-acting antivirals for use in large-scale treatment access programs in developing countries Clin Infect Dis 2014 58 928 936 24399087 12 Gane EJ Stedman CA Hyland RH Efficacy of nucleotide polymerase inhibitor sofosbuvir plus the NS5A inhibitor ledipasvir or the NS5B non-nucleoside inhibitor GS-9669 against HCV genotype 1 infection Gastroenterology 2014 146 736 743 24262278 13 Kohli A Osinusi A Sims Z Virological response after 6 week triple-drug regimens for hepatitis C: a proof-of-concept phase 2A cohort study Lancet 2015 385 1107 1113 25591505 14 Perelson AS Guedj J Modelling hepatitis C therapy—predicting effects of treatment Nat Rev Gastroenterol Hepatol 2015 12 437 445 26122475 15 Guedj J Dahari H Shudo E Smith P Perelson AS Hepatitis C viral kinetics with the nucleoside polymerase inhibitor mericitabine (RG7128) Hepatolgy 2012 55 1030 1037 16 Guedj J Pang PS Denning J Analysis of hepatitis C viral kinetics during administration of two nucleotide analogues: sofosbuvir (GS-7977) and GS-0938 Antivir Ther 2014 19 211 220 24464551 17 Guedj J Perelson AS Second-phase hepatitis C virus RNA decline during telaprevir-based therapy increases with drug effectiveness: implications for treatment duration Hepatology 2011 53 1801 1808 21384401 18 Messina JP Humphreys I Flaxman A Global distribution and prevalence of hepatitis C virus genotypes Hepatology 2015 61 77 87 25069599 19 Liu CH Kao JH Nanomedicines in the treatment of hepatitis C virus infection in Asian patients: optimizing use of peginterferon alfa Int J Nanomed 2014 9 2051 2067 20 Kai Y Hikita H Tatsumi T Emergence of hepatitis C virus NS5A L31V plus Y93H variant upon treatment failure of daclatasvir and asunaprevir is relatively resistant to ledipasvir and NS5B polymerase nucleotide inhibitor GS-558093 in human hepatocyte chimeric mice J Gastroenterol 2015 50 1145 1151 26208695 21 Sulkowski M Hezode C Gerstoft J Efficacy and safety of 8 weeks versus 12 weeks of treatment with grazoprevir (MK-5172) and elbasvir (MK-8742) with or without ribavirin in patients with hepatitis C virus genotype 1 mono-infection and HIV/hepatitis C virus co-infection (C-WORTHY): a randomised, open-label phase 2 trial Lancet 2015 385 1087 1097 25467560 22 Guedj J Dahari H Rong L Modeling shows that the NS5A inhibitor daclatasvir has two modes of action and yields a shorter estimate of the hepatitis C virus half-life Proc Natl Acad Sci U S A 2013 110 3991 3996 23431163 23 Rong L Guedj J Dahari H Analysis of hepatitis C virus decline during treatment with the protease inhibitor danoprevir using a multiscale model PLoS Comput Biol 2013 9 e1002959 23516348 24 Kumada H Suzuki Y Ikeda K Daclatasvir plus asunaprevir for chronic HCV genotype 1b infection Hepatology 2014 59 2083 2091 24604476 25 Omata M Kanda T Wei L APASL consensus statements and recommendation on treatment of hepatitis C Hepatol Intl 2016 published online April 29 26 Poordad F Lawitz E Gutierrez JA C-swift: grazoprevir/elbasvir + sofosbuvir in cirrhotic and noncirrhotic, treatment-naive patients with hepatitis C virus genotype 1 infection, for durations of 4, 6 or 8 weeks and genotype 3 infection for durations of 8 or 12 weeks J Hepatol 2015 62 S192 S193 27 Kattakuzhy S Sidharthan S Wilson E Predictors of sustained viral response to 4–6 week duration therapy with ledipasvir + sofosbuvir + GS-9451 +/− GS-9669 in early and advanced fibrosis (NIH/UMD synergy trial) J Hepatol 2015 62 S669 28 Lawitz EJ Gruener D Hill JM A phase 1, randomized, placebo-controlled, 3-day, dose-ranging study of GS-5885, an NS5A inhibitor, in patients with genotype 1 hepatitis C J Hepatol 2012 57 24 31 22314425 29 Svarovskaia ES Dvory-Sobol H Parkin N Infrequent development of resistance in genotype 1–6 hepatitis C virus-infected subjects treated with sofosbuvir in phase 2 and 3 clinical trials Clin Infect Dis 2014 59 1666 1674 25266287 30 Swallow E Kelley C Signorovitch J Wygant G McPhee F Daclatasvir + asunaprevir versus sofosbuvir/ledipasvir for hepatitis C genotype 1 in Japanese patients: an indirect comparison J Comp Eff Res 2016 5 273 279 26793987 31 Snoeck E Chanu P Lavielle M A comprehensive hepatitis C viral kinetic model explaining cure Clin Pharmacol Ther 2010 87 706 713 20463660 32 Reig M Marino Z Perello C Unexpected early tumor recurrence in patients with hepatitis C virus-related hepatocellular carcinoma undergoing interferon-free therapy: a note of caution J Hepatol 2016 published online April 12 33 Nguyen LH Nguyen MH Systematic review: Asian patients with chronic hepatitis C infection Aliment Pharmacol Ther 2013 37 921 936 23557103 34 Wei L Lopez-Talavera J Roa H Prevalence of HCV viral and host IL28B genotypes in China Hepatology 2011 54 563A 564A

PMC005xxxxxx/PMC5131926.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 101176023 32461 Stat Appl Genet Mol Biol Stat Appl Genet Mol Biol Statistical applications in genetics and molecular biology 2194-6302 1544-6115 27248122 5131926 10.1515/sagmb-2015-0043 NIHMS831610 Article The use of vector bootstrapping to improve variable selection precision in Lasso models Laurin Charles Integrative Epidemiology Unit, School of Social and Community Medicine, University of Bristol, Bristol, BS8 2BN, UK. Department of Psychology, University of Notre Dame, Notre Dame, IN 46556, USA Boomsma Dorret Department of Biological Psychology, VU University Amsterdam, Amsterdam, 1081 HV, Netherlands Lubke Gitta Department of Psychology, University of Notre Dame, Notre Dame, IN 46556, USA. Department of Biological Psychology, VU University Amsterdam, Amsterdam, 1081 HV, Netherlands 26 11 2016 1 8 2016 01 8 2017 15 4 305320 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. The Lasso is a shrinkage regression method that is widely used for variable selection in statistical genetics. Commonly, K-fold cross-validation is used to fit a Lasso model. This is sometimes followed by using bootstrap confidence intervals to improve precision in the resulting variable selections. Nesting cross-validation within bootstrapping could provide further improvements in precision, but this has not been investigated systematically. We performed simulation studies of Lasso variable selection precision with and without nesting cross-validation within bootstrapping. Data were simulated to represent genomic data under a polygenic model as well as under a model with effect sizes representative of typical GWAS results. We compared these approaches to each other as well as to software defaults for the Lasso. Nested cross-validation had the most precise variable selection at small effect sizes. At larger effect sizes, there was no advantage to nesting. We illustrated the nested approach with empirical data comprising SNPs and SNP-SNP interactions from the most significant SNPs in a GWAS of borderline personality symptoms. In the empirical example, we found that the default Lasso selected low-reliability SNPs and interactions which were excluded by bootstrapping. Variable selection Lasso Bootstrap Additive-by-Additive Epistasis Association Polygenic model 1 Introduction Multiple linear regression is most useful when it is applied to samples in which the number of predictors is small relative to the number of observations. However, if the number of predictors is large relative to the number of observations, there will be considerable sampling variability in the estimated coefficients. Under those conditions, the increased variability of coefficients can be managed by applying shrinkage methods. Shrinkage means estimating coefficients under a constraint that leads them to have reduced absolute values, drawing them toward 0, thus reducing sampling variability. There are many such constraints, hence many shrinkage methods. One of the most important shrinkage methods is the Lasso (Tibshirani, 2011). When the Lasso is applied to a multiple regression problem, small-valued coefficient estimates are reduced to 0 and the remaining coefficient estimates are shrunk by a fixed amount (Tibshirani, 2013). Because of this property, the Lasso is often used for variable selection (Tibshirani, 1996). In Lasso variable selection, the predictors having nonzero coefficient estimates after shrinkage are selected into the regression model, and those that are shrunk to 0 are excluded from the model. Variable selection with the Lasso is the task of deciding whether the predictor variables that could be included in a regression model are “unimportant” or as “important.” The unimportant, or “noise,” predictors are not associated with the outcome in the population, but could be strongly associated in a given sample because of sampling fluctuation. Important predictors are consistently associated with the outcome over independent samples. Unimportant predictors should be excluded from the regression model (i.e., their coefficient estimates shrunken to 0), while important predictors should be included in the model. In practice, using the Lasso for variable selection translates to classifying the excluded predictors as unimportant and the included predictors as important. The accuracy of this claim has been studied in simulation studies, and the Lasso has often been found to have low variable selection precision (VSP), meaning that among the included predictors, a relatively large proportion (> 50%) were false positives (Devlin et al., 2003, Ayers and Cordell, 2010, He and Lin, 2011). In this paper, we propose a method for controlling the Lasso’s VSP. We build on a variety of research that has investigated how to control the Lasso’s VSP. A main current in this research has been to estimate standard errors (SEs) and confidence intervals (CIs) for Lasso coefficient estimates with the bootstrap (e.g. Minnier et al., 2011, Chatterjee, 2011). Lasso SEs and CIs are used in a secondary variable selection step based on analogy with hypothesis testing (Freedman and Lane, 1983). The SEs are used to generate t-statistics; predictors having t-statistics below a user-set cutoff (e.g. |t| ≤ 2) are excluded. Similarly, predictors with CIs that contain 0 are excluded. Successfully applying a second variable selection step requires knowledge of why the Lasso includes too many variables in the model during the first step. Lasso variable selection depends on its degree of shrinkage. The degree of shrinkage in a Lasso model is controlled by a user-set metaparameter called λ. The shrinkage metaparameter λ is a scalar; its value determines the number of variables that the Lasso selects. The larger the value of λ, the more conservative the model. The choice of λ is thus related to the Lasso’s VSP. In general, each value of λ is associated with a single Lasso model. λ is commonly chosen through model-comparison methods: the user proposes a set of candidate λs and chooses one based on, e.g., BIC or cross-validation indices. Of these common methods, using cross-validation to choose λ has been associated with overfitting, leading to an excess of false positives and low VSP (Meinshausen and Bühlmann, 2010, James and Radchenko, 2009). Fan and colleagues (2012) attribute this to the large sample correlations that can arise between noise predictors and the outcome. They demonstrate empirically that noise correlations increase in magnitude with increasing numbers of noise predictors. Further, when sample size N is less than number of predictors p, the largest noise correlation can easily exceed true correlations in magnitude. In cross-validation, the sample size used for model-fitting is always smaller than that in the entire sample, exacerbating the problems identified by Fan et al, 2012. There has been little investigation into using bootstrap SEs and CIs to mitigate low VSP associated with cross-validated selection of λ. Asymptotic analyses have found that under bootstrap resampling, Lasso coefficients for noise predictors are expected to fluctuate between positive and negative values (Chatterjee, 2011, Camponovo, 2014). This should lead noise predictors to have larger bootstrap SEs and CIs than important predictors, which suggests that these statistics are useful for identifying false positive associations. To investigate this expectation, we introduced and evaluated a bootstrap-based method with the goal of improving Lasso VSP by excluding false positive predictor selections under cross-validation. We investigated bootstrap SEs and CIs for the Lasso when cross-validated selection of λ is done before bootstrapping, which is the standard approach (e.g. Sartori 2009, D’Angelo et al. 2009). An alternative is to nest cross-validated selection of λ within each bootstrap replication, which could lead to larger SEs and wider CIs than in the standard approach (Buckland et al., 1997, Bühlmann et al., 2011). Such larger SEs and wider CIs lead to more conservative variable selection, and possibly to improved VSP. We compared the VSP resulting from the standard approach to bootstrapping to the VSP resulting from nested selection of λ. We did this comparison in simulated and empirical data. The simulated data were generated to resemble data observed in genetic association studies. The simulated data were high-dimensional, with a small number of weak predictors and a large number of noise predictors. The empirical data were drawn from a genome-wide association study (GWAS). We made this choice because of the prominence of genetics, in particular GWAS and the use of polygenic (risk) scores, as a context for the application and development of the Lasso (Waldron et al., 2011, Lange et al., 2014). 2 Approach Estimation of SEs and CIs for nonzero Lasso coefficient estimates may provide a way to control the Lasso’s VSP as well as to assess the relative importance of predictors. In general, Lasso SEs or CIs cannot be estimated using a closed form (Osborne et al., 2000). A large variety of approaches has been tried to estimate Lasso SEs and CIs: see review paragraphs in e.g. Bühlmann et al. (2014), Chatterjee (2011), Kyung et al. (2010). Bootstrap estimation has received substantial interest, but many issues remain less explored, most importantly, the effects of choosing λ through cross-validation when using the bootstrap. Next, we briefly review the two most common approaches to bootstrapping Lasso SEs and CIs: vector bootstrapping and residual bootstrapping. We follow this with a selective review of applied and methodological research using these approaches. We review both approaches to show that the behavior of the residual bootstrap has been studied in detail, but that comparatively little methodological research has been done on the vector bootstrap, despite its popularity in applied research. Hence, this paper focuses on the vector bootstrap. 2.1 Vector and Residual Bootstrapping In nonparametric bootstrapping, samples are repeatedly drawn from observed sample data. The statistic of interest is calculated in each bootstrap sample. In this case, Lasso coefficient estimates are calculated, leading to an approximate sampling distribution. The approximate sampling distribution of Lasso coefficient estimates then permits calculation of SEs and CIs (Efron and Tibshirani, 1994). We denote an estimate of the sampling distribution for the Lasso coefficient for predictor j as F̂(β̂j). F̂(β̂j) represents the marginal distribution of β̂j values, as estimated using the nonparametric bootstrap. Two ways to use non-parametric bootstrapping to find F̂(β̂j) are vector bootstrapping and residual bootstrapping (Sartori, 2009). 2.1.1 The vector bootstrap Vector bootstrapping begins with an observed sample of N observations measured on p predictors x1, …, xp and an outcome y. Each observation in the sample is considered as a row vector zi = (xi1, …, xip, yi),which consists of p predictor values xij and a single outcome value yi. A Lasso model can be fit to every bootstrap sample of N observations zi, yielding a set of Lasso coefficient estimates (Camponovo, 2014). F̂(β̂j) is defined as the distribution of β̂j values from each of all possible bootstrap samples of size N. The number of unique bootstrap samples increases faster than exponentially in N; to save computation time in practice, F̂(β̂j) is estimated using Monte Carlo simulation. The Monte Carlo estimate of F̂(β̂j) is written F̂*(β̂j). 2.1.2 The residual bootstrap Residual bootstrapping begins by fitting a linear regression model to a sample of N observations measured on p predictors x1, …, xp with outcome y, and generating the N residuals e. Residual bootstrapping uses the sampling distribution of residuals to simulate the distribution of y values about their conditional means Xβ. Importantly, this requires treating the observed predictor values x1, …, xp as fixed and assuming that only the correct predictors are in the model (Efron and Gong, 1983), likely an inappropriate assumption in the context of variable selection. The residuals are then resampled. Each bootstrap sample of N residuals, stored in the vector e*, can be used to generate N outcomes y*, defined as y* = Xβ̂+e*. The y* values are regressed on X using the Lasso, yielding, as in vector bootstrapping, a set of Lasso coefficient estimates for each resample. This set of coefficient estimates is used to define F̂(β̂j), which is typically approximated through Monte Carlo simulation, as F̂*(β̂j). 2.2 Previous research in bootstrapping the Lasso 2.2.1 Research with residual bootstrap Residual bootstrapping of Lasso SEs and CIs has received more methodological research interest than has vector bootstrapping. Detailed investigations of the residual bootstrapped Lasso have been undertaken by Chatterjee (2011), Minnier et al. (2011), Kyung et al. (2010), and Knight and Fu (2000), among others, with the theoretical results in Chatterjee (2011) synthesizing much of the previous work. By comparison, the behavior of Lasso SEs and CIs under vector bootstrapping has been under-studied, particularly with λ selected through cross-validation. 2.2.2 Research with vector bootstrap The vector-bootstrapped Lasso has often been applied in statistical genetics, sometimes with λ selected through cross-validation. Further, the vector-bootstrapped Lasso is closely related to several other prominent variable-selection methods proposed in statistical genetics (Valdar et al., 2012, Cho et al., 2010, Motyer et al., 2011). D’Angelo et al. (2009) proposed using vector resampling to estimate SEs of Lasso coefficients of SNP-SNP and gene-gene interaction terms. Sartori (2009) compared residual and vector bootstrapping of Lasso CIs and SEs in the context of statistical genetics, including the selection of λ through cross-validation before resampling. She found that: 1) residual and vector bootstrap SEs of Lasso coefficient estimates had similar degrees of bias in linear models; and 2) vector bootstrap CIs had superior coverage in linear and in logistic models. Camponovo (2014) used vector bootstrapping to generate simultaneous confidence regions in linear models with random predictors. He found poor coverage rates, and used an asymptotic argument to propose two modified vector bootstrapping procedures. The modified procedures generated confidence regions with adequate coverage. The present study builds on the above research: in it, we compare the vector bootstrap with and without cross-validation nested within bootstrap replications. An important difference in the present study is that VSP, rather than coverage rate, is the criterion of comparison: if the CI for an important predictor excludes the true coefficient value but also excludes 0, the variable is correctly selected. Both versions of the bootstrapped Lasso are straightforward to implement in existing statistical software; e.g. R package glmnet, PLINK 1.9. (Friedman et al., 2010, Chang et al., 2014). This makes them attractive and approachable to applied researchers, who would benefit from understanding the trade-off in VSP involved in choosing one over the other. 2.2.3 Research with cross-validated selection of λ Many methods have been proposed to increase the VSP of Lasso regression and related methods (Chatterjee, 2011, Lockhart et al., 2013, Fan et al., 2012). However, little methodological research has addressed the combination of vector bootstrapping and cross-validated selection of λ that is used in practice (Sartori, 2009, Cho et al., 2010). Despite the suggestion, in a recent textbook, that cross-validated selection of λ should be nested within bootstrap samples when applying the Lasso (Bühlmann et al., 2011), little published work has evaluated this procedure (Okser et al., 2014). The goal of the present paper is to address this deficiency and to investigate the conditions in which nested selection of λ leads to improved VSP. In addition, we address the question whether bootstrapped t-statistics are useful for the identification of false positives. 2.3 Role of λ when fitting Lasso models The metaparameter λ controls the bias and parsimony of a fitted Lasso model. Equation 1 gives the definition of a Lasso model for predictors X and outcome y when λ is known (Tibshirani, 1996). (1) β^=argβmin12∑i(yi-∑jxijβj)2+λ∑j∣βj∣ The value of λ determines the degree of shrinkage toward 0 and serves as a threshold for variable selection. A predictor is selected if the absolute value of its covariance with the outcome is larger than λ, and excluded otherwise (Efron et al., 2004). This thresholding property limits the useful range of values that λ can take. The minimum value that λ can take is 0, where the Lasso fit is the same as that of OLS regression. Such solutions are unbiased, but, because of the improbability of any OLS coefficients equaling 0 exactly, they are also unparsimonious. The maximum value that λ can take depends on the largest sample covariance of any predictor with the outcome. More specifically, when λ is equal to or greater than that covariance, all coefficients are shrunk to 0, and the fitted model is intercept-only, thus parsimonious but biased (Friedman et al., 2007). 2.3.1 Selection of λ through K-fold cross-validation In general, each value of λ is associated with a single Lasso model (Tibshirani, 2013, Efron et al., 2004). K-fold cross-validation (Zhang, 1993) is often used to select the best-performing model. Lasso model fitting, λ selection, and K-fold cross-validation has been described in detail for its implementation in the R package glmnet (Friedman et al., 2010). Following this procedure, the λ value that is associated with the best-performing model is selected. The best-performing model is the one having the minimum cross-validation index, which is computed as the sum of squared residuals averaged over the K cross-validations. The selected λ value is then used to fit a finalized model by solving Equation 1 in the entire sample. This produces a set of selected predictors that are then indexed in set s. Different λ values might be selected in different samples from the same population due to the influence of noise correlations (Fan et al., 2012). In the next section, we interpret selection of λ as a source of variation in Lasso coefficient estimates. 2.3.2 Lasso variance estimates: Contribution of λ selection The variance of Lasso coefficient estimates depends on the joint distribution of the p predictor variables X and the outcome y, as well as on the value of λ (Pötscher and Leeb, 2009). The conditional distribution of estimates for a single predictor, denoted gj(β̂|λ), is the distribution of β̂j coefficients at a fixed λ value. The marginal distribution, hj(β̂), is the distribution of β̂j averaged over λ values. The variance of β̂j can be found using either gj or hj. Heuristically, hj treats the selected λ value as a realization of a random variable (Zhang, 1993, Bühlmann et al., 2014). We argue that using hj might improve VSP because using gj treats λ as fixed, which can underestimate the variance of coefficients. To support this claim, consider the inequality: (2) Var(β^j)=Eλ{Varβ(β^j|λ)}+Varλ{Eβ(β^j|λ)}≥Varβ(β^j|λ) (Chatfield, 1995). If β̂j and λ were independent, then Var(β̂j) = Varβ(β̂j|λ) and there would be little difference between the fixed and random λ approaches in practice. However, β̂j and λ are not necessarily independent: the range of possible λ values is bounded by (0, rmax). Thus, although using gj (treating β̂j and λ as independent) has the practical advantage of using fewer computational resources, it will only be acceptable if the resulting underestimate of the standard error of β̂j is small. 2.3.3 λ selection in bootstrapping In practice, both gj, the conditional distribution of β̂j given λ, and hj, the distribution of β̂j averaged over all λs, are unknown. Both distributions can be estimated using the vector bootstrap. Finding the bootstrap estimate of the conditional distribution, g^j∗ is done by fitting Lasso models to resampled X and y values, given the λ value chosen through K-fold cross-validation in the original sample. Finding the bootstrap estimate of the marginal distribution h^j∗ requires treating the selected λ value as random. Nesting λ-selection within each bootstrap replication approximates the effect of sampling error on the value of λ selected. Our simulations compared the fixed- and random-λ approaches with respect to VSP, and suggest effect sizes at which the increased computational burden of the random approach is worthwhile. 3 Methods The purpose of the current paper is to propose and to evaluate the use of the vector bootstrap, with λ selected through K-fold cross-validation, as a method for estimating Lasso SEs and CIs. In particular, we compared three variants of this approach: a software default approach to variable selection using the Lasso (Method 1, see Figure 1); an approach involving selection of λ before resampling (Method 2, see Figure 2); and a third approach where λ-selection is nested within bootstrap samples (Method 3, see Figure 3). Our evaluation was in terms of VSP and of accuracy of ranking predictors by relative importance. Relative importance was calculated using coefficients of variation, (|t−1|, where t is a bootstrapped t-statistic). In the first step of each variable selection method, a Lasso model is fit to the entire sample. This requires selection of λ, which is done through K-fold cross-validation. The initial model fit produces a set of selected predictors, which are indexed in the set s. This step is the default application of the Lasso in the R packages glmnet and grpreg. We denote it Method 1. Methods 2 and 3 differ from Method 1 by having a second variable selection step. In this step, further reduction of the set of selected predictors is done using bootstrap SEs or CIs. All predictors are used in vector bootstrap resampling, but, to save computational resources, SEs and CIs are not calculated for predictors that were excluded in the initial variable selection step. Method 2 uses the same value of λ in every bootstrap sample. Method 3 differs from Method 2 by re-selecting λ in each bootstrap sample. In both methods, after SEs or CIs are calculated, variables that: 1) include 0 in their confidence intervals; or that: 2) have a coefficient of variation greater than a certain cutoff; are excluded. 3.1 Lasso CIs and SEs: Secondary selection or ranking Bootstrap CI or SE estimates improve Lasso models through a second step of selecting or ranking predictors. CI and SE estimates are both directly related to the sampling variance of a Lasso coefficient estimate, discussed above. A 1 − α CI for the coefficient estimate β̂j is generated either using the α2,1-α2 quantiles of the bootstrap distribution g^j∗, or using an approximate inverted z-test, which gives the interval β̂j ± zα/2SE*(β̂j), where zα/2 is the α2 quantile of a standard normal distribution and SE*(β̂j) is the bootstrap estimate of the standard error of β̂j. Using either CI method, predictors that have CIs that contain 0 are excluded since this can be regarded as evidence that predictor xj is a false positive selection. SEs can also be used to give information about the relative importance of predictors in addition to improved VSP. We propose the use of the coefficient of variation for each nonzero Lasso regression weight as an index of relative importance and to apply cutoffs to this statistic to exclude false positives. The coefficient of variation of a random variable X, denoted Cvar (X), is the ratio of its standard error to the absolute value of its mean. This index is sensitive to small differences in mean values and thus may be better able to distinguish small true positives from false positives. We used the vector bootstrap to estimate Cvar for individual Lasso coefficients, denoted Cvar∗(β^j). 4 Simulation Studies We first compared Methods 1–3 using a factorial simulation study. The simulation had two goals: first, evaluating the Methods’ ability to distinguish signal from noise; second, evaluating their ability to correctly order signals of differing strengths. The data generation models were as simple as possible while still representing two empirically interesting scenarios based on statistical genetics: 1) a low probability of selecting important predictors at random; and 2) a spectrum of small true effect sizes. To this end, data were generated under two different linear models: first, a few-important-predictors model with under 5% of predictors having true effects, and with effect sizes (given in r2) representing 2.5% or less of outcome variance attributable to any important predictor; and second, a polygenic model, in which there were thousands of predictors, each of which had an effect drawn from a normal distribution, all of which together accounted for 60% of the variance in the outcome. 2500 Monte Carlo (MC) replications were used in each cell of the few-important-predictors design. This number of replications was chosen based on pilot studies, in which at least 2500 replications were required in order to generate relatively smooth empirical distributions of coefficient estimates (see also Sartori 2009). However, only 250 MC replications were used when data were simulated under the polygenic model due to the substantial use of computational resources needed to bootstrap such data. B = 1000 bootstrap replications were used within each MC replication. The average performances of the three methods across samples were compared; within each MC replication, each method was employed on an independent sample drawn from the population distribution. This was done in order to avoid creating dependence among results that might arise from fitting the methods to the same data. 4.1 Simulation Design Three factors were manipulated in the simulations: method, data-generating model, and effect size. As described above, the methods compared were the fixed- (Method 2) and random-λ (Method 3) variants of the vector bootstrapped Lasso, with the default application, Method 1. The second factor manipulated in the simulation study was the data-generating model. Three data-generating models were used: a single important predictor and 99 noise predictors; a 5 important predictors and 99 noise predictors, with the important predictors having different r2 values, enabling us to rank them; and a polygenic model with 3000 predictors with effect sizes drawn from a 𝒩 (0, 0.60) distribution–the three predictors with the largest (absolute) effects were treated as the important predictors. Each data generating model was a linear regression model having a standard normal outcome and binomial(2, 0.5) distributed predictors; N = 2500 was used as the sample size. This was chosen as a rough approximation of the sample size and predictor structure of smaller genome-wide association studies of quantitative phenotypes (Balding, 2006). The third factor manipulated in the simulation study was effect size. Effect sizes of r2 = 0.01, 0.0033, 0.001 were used in the single-important-predictor analyses. In the five-important-predictors analyses, each important predictor had a different effect size: the set r2 = 0.01, 0.0067, 0.0033, 0.0022, 0.001 was used. In the polygenic model, the three strongest predictors were expected to account for 5.1% of phenotypic variance together. This estimate is based on treating the largest simulated-SNP effects as being drawn from a truncated normal distribution representing the upper 0.15% tail area of the 𝒩(0, 0.60), and then treating the strongest negative effects as being drawn independently from the corresponding part of the lower tail(Barr and Sherrill, 1999). We manipulated the effect size of important predictors for two reasons: we used effect size as a measure of the “difficulty” of correctly selecting important predictors, giving us a way to use the data to influence the methods’ VSP; and because previous simulation studies (e.g. Tibshirani, 2011, Leng et al., 2006, Meinshausen and Bühlmann, 2010), used effect sizes that are now considered to be unrealistically large in the context of statistical genetics (Stefansson et al., 2009, Park et al., 2011). In the next section, we describe the evaluation criteria. 4.2 Evaluation Criteria The main question asked in this paper is: when does nested selection of λ lead to improved VSP over other approaches? A subsidiary question is: can Cvar∗-statistics give information about the relative importance of predictors that might be used to identify false positives? Addressing these questions requires quantifying the performance of the different methods. We used variable selection precision (VSP) to quantify the methods’ performance and additionally used the False Negative Rate (FNR) to identify the risk that each method might be over-conservative. VSPis#ImportantPredictorsSelected#SelectedPredictors=#TruePositives#Positives. VSP is set to 0 if there are no positives. Thus, in each replication of the few-important-predictors simulation, VSP ranged from 0, 1104,1103,…,12, 1, while in the polygenic model, the denominator was 3000. FNR is the proportion of truly important predictors that have been classified as unimportant by a variable selection method (Fawcett, 2006). A positive in the vector bootstrap Lasso (Methods 2 and 3) was defined as a predictor for which the (1 − α) × 100% bootstrap percentile confidence interval (Efron and Tibshirani, 1994) excluded 0. Confidence level (α) was set to 0.05, and intervals were symmetric. In sensitivity analyses, confidence levels of 0.02, 0.10 and 0.20 were also used. A positive in the default Lasso (Method 1) was defined as a predictor having a coefficient in the finalized Lasso model (i.e. a predictor indexed in s). To quantify the performance of Cvar∗ as an importance measure, we used the median rank and median absolute deviation of ranks of each predictor’s Cvar∗ over MC replications. We chose this measure to obtain both a typical rank and the variability of rankings for important and for noise predictors. This importance measure was not used with data that had been simulated under a polygenic model because the randomness of the effects meant that there was no consistent mapping between a predictor’s index and its effect size. 4.3 Simulation Results 4.3.1 Improved Variable Selection Precision Use of the vector bootstrap (Methods 2 and 3) was associated with increased VSP at all effect sizes, but also with increased FNR at all effect sizes. This is shown in Table 1, for nominal α = 0.05 and percentile bootstrap confidence intervals; normal-theory bootstrap confidence intervals are not shown because their performance was similar to, but slightly worse than that of the percentile intervals. At the smaller effect sizes, the increased precision was only apparent with random λ, and, at the smallest effect size, the magnitude of this advantage was small. The increased FNR suggests that this pattern was due to the bootstrapping procedures being more conservative than Method 1. In sensitivity analyses, we found that the nominal coverage rate α chosen for confidence intervals interacted with the bootstrapping method used. For example, at r2 = 0.0033, the random approach (Method 3) is more precise than the fixed approach (Method 2) at α = 0.01, 0.05 (Table 1), but less precise at α = 0.10 (not shown). The bootstrapped coefficient of variation ( Cvar∗) was in general much larger for false positives than for true positives. Bootstrap means for noise predictors were never 0, preventing the occurrence of division-by-0 errors in computing Cvar∗ values. These observations support the use of a Cvar∗ cutoff as a way of increasing Lasso VSP. However, Cvar∗ cutoffs did not increase VSP to the extent that CIs did. They offered no improvement over the default Method 1 at effect sizes r2 = 0.0033, 0.01 or in multiple-predictor models. 4.3.2 Ranking Predictors The usefulness of Cvar∗ to rank predictors by relative importance depended on the effect size as well as the number of true predictors. At r2 = 0.0033, 0.01 or with multiple true predictors, using Cvar∗ to rank predictors by relative importance yielded no improvement over ranking predictors by the absolute value of their non-bootstrapped coefficients. This is shown in Table 2. Additionally, true predictor ranks were identical regardless of whether λ was treated as random or as fixed. For a single true predictor with r2 = 0.001, ranking by Cvar∗ led to better discrimination of the true predictor than did ranking by non-bootstrapped coefficients. The true predictor did not always have the smallest Cvar∗, but that it was within the top 6 predictors at least half the time. Without bootstrapping, the true predictor was often selected out of the model. Overall, it was in the top 50% of predictors close to half the time but that it was not frequently among the highest ranks. 4.3.3 Variability of Coefficients under Different Methods Treating the metapararameter λ as random was associated with greater variability of coefficient estimates, as illustrated in Table 3 by larger standard deviations, and wider confidence intervals for Method 3 when compared to Method 2. Table 3 presents summaries for important predictors only; results for noise predictors were very similar (results in the table are averages over MC replications and are not conditional on the important predictors being selected into the model). Confidence interval lengths were similar between quantile-based bootstrap CIs and normal-theory bootstrap CIs. These results are consistent with inequality (2), which suggests that the increased variability is attributable to the variance in expected coefficient values with respect to the distribution of λ. Table 4 compares five-number summaries of the distribution of λ from Method 1 to those from Method 3. The results for Method 1 are five-number summaries of the λ values that were selected across replications, while those for Method 3 are the averages across replications of the five-number-summary of λ values. Selection of λ via Method 3 (cross-validation nested within bootstrap replications) tends to produce lower values of λ (distribution shifted left) which are also less-variable (smaller IQR). The smaller median λ means that Method 3 performed less regularization than did Method 1, hence had coefficients with larger values and may have included more predictors in each bootstrap replication. Thus, bootstrap CIs from Method 3 would have been relatively wide, leading to increased VSP because of liberal variable selection within each bootstrap replication. The results of the simulation studies quantified the methods’ relative performance in idealized data. For a more critical evaluation of their practical utility, we applied them in a GWAS data set, using them to select pairwise interactions as well as main effects. 5 Empirical Illustration We used data gathered for a genome-wide association study (GWAS) of Borderline Personality Disorder features to compare the vector bootstrap with λ-selection nested within bootstrap samples to the standard Lasso. This comparison serves as a representative analysis for possible applications of Method 3. The original GWAS was based on a sample of N = 7124 individuals who participated in a twin-family study of mental and somatic health (Boomsma et al., 2006, Willemsen et al., 2013); see Willemsen et al. for detailed methods including IRB approval, genotyping, and quality control procedures. Responses to a psychiatric inventory measuring Borderline Personality features were used as outcomes because they have shown a promising signal that was replicated in an independent sample (Lubke et al., 2014). Borderline features were measured using total scores on the PAI-BOR inventory, a 24-item test (Morey, 1991). More specifically, the outcome we used in this study was the residual of PAI-BOR score after OLS regression on age, gender, and their interaction, as well as a principal component score representing ancestry (Abdellaoui et al., 2013, Price et al., 2010). Following up on D’Angelo et al. (2009)’s proposal, we fit Lasso multiple regressions of Borderline features on SNP main effects and SNP-SNP interactions. To control the computational resources required, we limited the analysis to pairwise interactions and main effects of the 125 SNPs having the strongest univariate association with the Borderline features phenotype, as listed in Lubke et al. 2014. R’s memory limitations limit the application of bootstrapping to interactions between 1500 or fewer variables (15002 × 1000 bootstrap samples ≈ 231 objects) (R Core Team, 2013). To avoid multicollinearity, these 125 were then pruned to the set of 77 SNPs that had pairwise correlations of less than r = 0.6 among each other. From these, 2926 pairwise interaction terms were calculated, resulting in a total p = 3003, and hence up to 3003 Lasso partial regression weights needing CIs and Cvar∗. The primary purpose of this analysis was to compare the different methods for lasso variable selection in a data set with SNPs having different allele frequencies and unknown effects on the outcome. The secondary purpose was an exploratory analysis of epistatic effects between SNPs as predictors of borderline personality symptoms, generating hypotheses that can be tested in independent samples. We did not emphasize the need for the selected models to be biologically plausible; that is, we did not group SNPs for selection by gene or pathway. We did not rule out models of pure interaction in the absence of main effects (Cordell, 2009). Accordingly, we did not force selection of hierarchical models: that is, it was not the case that any interaction terms considered for selection must have had main effects in the model. Because of this and because of the bias caused by Lasso estimation, the values of the coefficients for (pure) interaction terms cannot be interpreted straightforwardly as magnitudes of effect modifications. Results from applying the vector bootstrap with random λ (Method 3) were compared to those from the default Lasso (Method 1). Method 3 was used to generate percentile CIs and Cvar∗. The resulting variable selections and rankings were compared to Method 1’s selections and to ranking coefficient estimates from the finalized model by absolute value. 5.1 Empirical illustration: Results The empirical illustration concerned the application of the bootstrapped Lasso to pairwise interaction effects. The bootstrapped Lasso produced different variable selections and importance rankings than did the default Lasso. The bootstrapped Lasso tended to select better quality predictors than did the default Lasso. The default Lasso selected predictors with low minor allele frequency (MAF), hence had very large standard error estimates. This suggests that the default Lasso can ignore important aspects of the data. Figure 4 and Table 3 present comparisons of the default Lasso and the approach using vector bootstrapping. Figure 4 plots bootstrap mean estimates on the horizontal axis and bootstrap standard error estimates on the vertical axis. Points falling outside of the dark gray V have Cvar∗ values less than 0.5 (bootstrapped t-statistics greater than 2). Predictors that were selected by the default Lasso are plotted as light gray diamonds. There is no obvious relationship between default Lasso selection of a predictor and its bootstrap moments. dbSNP lookup of the predictors in Table 3 showed that the bootstrapped Lasso was less prone to selecting interactions between SNPs having low MAF than was the default Lasso. The default Lasso, in selecting these interactions, was in effect including interactions between binomial predictors that have low success probabilities. These interactions tended to have large bootstrap standard errors, hence are excluded when a Cvar∗ cutoff is used. Interestingly, the low-MAF SNPs involved these interactions tended to have moderately strong main effects in the conventional GWAS analyses. Using vector bootstrapping of Lasso coefficients (CI or Cvar∗ cutoff of 0.5) suggested a single interaction for followup. 6 Conclusion Using vector bootstrap CIs on Lasso regression coefficients offers a valid way to distinguish false positive selections from true positives. Percentile CIs were associated with increased precision of variable selection at all effect sizes. At the smallest effect sizes, including those in a polygenic model, gains were only achieved when using Method 3, which treated the metaparameter λ as random. Additionally, the bootstrapped methods were more conservative in variable selection than was Method 1, the default Lasso. This suggests that, if several small effects are expected and if avoiding false positives is more important than avoiding false negatives, treating λ as random justifies increased computational cost. The (bootstrapped) coefficient of variation, Cvar∗ does measure the relative importance of Lasso predictors, but offers little to recommend it over using the absolute value of Lasso coefficients. We observed a “rising tide lifts all boats” effect for all methods, where a predictor having a given small effect size was more likely to be selected when the data were generated to have other important predictors. This was despite the predictors and their effects being independent of one another. The more complex models tended to have lower (more lenient) thresholds λ selected by cross-validation, regardless of the method used. A possible explanation is that a λ causing inclusion of a solitary small effect might not be able to consistently decrease the residual sum of squares in different cross-validation subsamples, but that a λ that admits multiple small effects could. The low degree of overlap in the distributions of λ estimated by Method 1 and Method 3 suggests a need for explanation. The two distributions should have the same mean value, because the bootstrap distribution of a statistic should approximate the distribution of a statistic across repeated independent samples. In the empirical analysis, the vector-bootstrapped Lasso excluded unreliable predictors that had been selected by the default Lasso. However, it is possible that residual bootstrapping could have led to better performance with low MAF predictors. Under vector bootstrapping, “monoallelic” SNPs are possible within the bootstrap samples. The result would be inflation of the intercept term in the regression model, which would affecting model-fitting through cross-validation. Further, VSP is dependent on the base rates of positives and negatives, and would have been skewed if important SNPs tended to have low MAFs and noise SNPs high MAFs, or vice-versa, which limits the generalizability of the simulation results to empirical data. Three follow-up studies are suggested by this result: a simulation study comparing the two Lasso methods after manipulating the reliability of predictors; an attempt to replicate the promising interaction between rs118160379 and rs59194015; and a comparison of VSP from Method 3 to that from other Lasso confidence intervals, e.g. those in Camponovo (2014). In addition, both our empirical and our simulation results are also relevant to research that uses polygenic scores to predict complex traits and to investigate the polygenic architecture of closely related traits. Lasso regression has recently been implemented in PLINK2, a software package that is very widely used in the analysis of complex traits. In consequence, Lasso variable selections are being used to construct polygenic scores, but have not yielded significant improvements over conventional methods (Warren et al., 2014). Our simulation results suggest that vector bootstrapping (with nested selection of λ) may be able to yield polygenic scores with greater predictive accuracy, but will require relatively large sample sizes to avoid excluding important variants from the score. Large samples are relatively common in the GWAS context, however using this approach efficiently will require careful application and data management. There were several limitations to this study. The data generating model had predictors that were independently and identically distributed as well as a normally distributed outcome. These attributes are unlikely to hold in empirical data. We plan to extend the current simulations with correlated and differently scaled predictors as well as skewed outcomes. Second, the simulated effect sizes, while small, were still somewhat larger than those that are typically observed in the statistical genetics of complex traits (Stefansson et al., 2009). The phenotypic variance explained by the polygenic model was consistent with a highly heritable trait, e.g. height, and is larger than would that expected for most complex traits of interest. The number of important predictors used was also much smaller than the number of genetic loci expected to influence complex phenotypes (Sivakumaran et al., 2011). Similarly, our results suggest that bootstrapping is most useful when there are many predictors to consider, to be reduced to a relatively small number of important ones, and that the single-important predictor case (tested here) is perhaps suboptimal for evaluating the precision and conservatism of the bootstrap methods. Finally, the argument used to justify nesting λ-selection within bootstrapping was intuitive. A more rigorous argument might be able to identify specific conditions on X or y that would lead to Method 3 consistently outperforming Method 2, or vice-versa. On the other hand, simulations could be used to estimate the components of bootstrapped variance due to the individual terms in Equation (2). Vector bootstrapping CIs of Lasso coefficients led to increased variable selection precision, especially at small effect sizes. Our illustration with empirical data showed that this is also an effective approach to select important interactions between predictors. In consequence, vector bootstrapping CIs is a very promising approach for identifying sets of SNP-SNP and SNP-environment interactions. Fig. 1 Method 1–Default approach to Lasso variable selection using K-fold cross-validation Fig. 2 Method 2–Vector bootstrap for improved Lasso variable selection precision with λ treated as fixed Fig. 3 Method 3–Vector bootstrapping with metaparameter λ treated as random Figure 4 Bootstrap means and SEs of 77 SNPs, 2926 pairwise interactions; light gray diamonds represent predictors selected without bootstrapping. The lines represent mean=±2×SE^∗: points and diamonds ouside the V-shape (i.e. in lower corners) are promising signals. A cube-root transformation was used on both axes. Table 1 Variable selection precisions and false negative rates of default lasso vs. bootstrapped percentile CI The random-λ bootstrap CI (Method 3) is more precise than the default at each effect size, but improvement is marginal for the very smallest effects. In all cases, the default approach showed low false-negative rates. R2\Method Default Fixed λ Random λ VSP Ranks 0.3646 0.9529 0.8956 0.01 0.3720 0.8304 0.8528 0.0033 0.2049 0.1536 0.3372 0.001 0.0503 0.0096 0.0507 Polygenic 0.0096 0.1154 0.0990 FNR Ranks 0.2008 0.5784 0.4594 0.01 0.0084 0.1500 0.0408 0.0033 0.3776 0.8420 0.6088 0.001 0.7928 0.9888 0.9404 Polygenic 0.1343 0.8804 0.8372 Bold text indicates the largest VSP/smallest FNR in each row. Nominal α = 0.05. Table 2 Median and MAD of ranks of 5 important predictors out of 104 total Bootstrapping resulted in no improvement of predictor ranks. Both bootstrapping methods (2 and 3) showed identical performance. R2\Method Default (1) Bootstrapped ( Cvar∗) (2 and 3) 0.01 1 [0] 1 [0] 0.0067 2 [0] 1 [0] 0.0033 3 [0] 3 [0] 0.0022 4 [1.48] 4 [1.48] 0.001 5 [2.97] 6 [4.48] Table 3 Standard deviations and confidence interval lengths of Lasso estimates of important predictors Measures of the variability of coefficient estimates for important predictors, averaged over MC replications, showed that Method 3 had increased variability compared to Method 2. Model Method SD qCI Length ntCI Length Ranks 2 (Fixed λ) 0.0099 0.0351 0.0405 3 (Nested λ) 0.0177 0.0693 0.0692 Single/R2=0.0033 2 (Fixed λ) 0.00637 0.02060 0.02500 3 (Nested λ) 0.0167 0.0650 0.0656 polygenic 2 (Fixed λ) 0.0172 0.0604 0.0674 3 (Nested λ) 0.0239 0.0860 0.0857 SD = standard deviation; qCI = bootstrap quantile confidence interval; ntCI = bootstrapped normal approximation confidence interval; Nominal α = 0.05 Table 4 Five-number summaries of the distribution of λ values in different simulations Summaries of the distribution of λ values suggest that Method 3 (selection nested within bootstrap samples) produces a less-variable, left-shifted distribution. Model Method Min 1Q Med 3Q Max Ranks 1 (CV only) 0.0154 0.0282 0.0324 0.0367 0.0582 3 (Nested λ) 0.00753 0.01410 0.01640 0.01910 0.04030 Single/R2=0.0033 1 (CV only) 0.0196 0.0365 0.0446 0.0505 0.0685 3 (Nested λ) 0.00812 0.01540 0.01820 0.02190 0.06020 Polygenic 1 (CV only) 0.00796 0.01320 0.01490 0.01680 0.03180 3 (Nested λ) 0.00182 0.00463 0.00543 0.00625 0.00968 Table 5 Promising SNP-SNP interactions Chrs rsIDs MAFs Cvar∗ CI 16, 1 rs118160379 × rs59194015 .05, .27 .49 [−.078, −.002] Selected by default but rejected by bootstrap: Chrs rsIDs MAFs Cvar∗ CI 9, 4 rs112188788 × rs139344595 .02, .01 .73 [0, .501] 6, 1 rs117666484 × rs73008417 .01, .01 .94 [0, .562] 9, 1 rs112188788 × rs73008417 .02, .01 1.17 [−2.904, 0] 12, 9 rs117256451 × rs112188788 .02, .02 1.57 [−.056, .451] Abdellaoui A Hottenga JJ de Knijff P Nivard MG Xiao X Scheet P Brooks A Ehli EA Hu Y Davies GE 2013 Population structure, migration, and diversifying selection in the netherlands European Journal of Human Genetics 21 11 1277 1285 23531865 Ayers KL Cordell HJ 2010 SNP Selection in genome-wide and candidate gene studies via penalized logistic regression Genetic epidemiology 34 8 879 891 21104890 Balding DJ 2006 A tutorial on statistical methods for population association studies Nature Reviews Genetics 7 10 781 791 Barr DR Sherrill ET 1999 Mean and variance of truncated normal distributions The American Statistician 53 4 357 361 Boomsma DI de Geus EJC Vink JM Stubbe JH Distel MA Hottenga JJ Posthuma D Van Beijsterveldt TCEM Hudziak JJ Bartels M Willemsen G 2006 Netherlands twin register: From twins to twin families Twin Research and Human Genetics 9 6 849 857 17254420 Buckland ST Burnham KP Augustin NH 1997 Model selection: an integral part of inference Biometrics 603 618 Bühlmann PL van de Geer SA Van de Geer S 2011 Statistics for high-dimensional data Springer Bühlmann P Meier L van de Geer S 2014 Discussion: “A significance test for the lasso” Ann Statist 42 2 469 477 Camponovo L 2014 On the Validity of the Pairs Bootstrap for Lasso Estimators Social Science Research Network Working Paper Series Chang CC Chow CC Tellier LC Vattikuti S Purcell SM Lee JJ 2014 Second-generation plink: rising to the challenge of larger and richer datasets arXiv preprint arXiv:1410.4803 Chatfield C 1995 Model uncertainty, data mining and statistical inference Journal of the Royal Statistical Society. Series A (Statistics in Society) 158 3 419 466 Chatterjee A 2011 Bootstrapping lasso estimators Journal of the American Statistical Association 106 494 608 625 Cho S Kim K Kim YJ Lee JK Cho YS Lee JY Han BG Kim H Ott J Park T 2010 Joint identification of multiple genetic variants via elastic-net variable selection in a genome-wide association analysis Annals of human genetics 74 5 416 428 20642809 Cordell HJ 2009 Detecting gene–gene interactions that underlie human diseases Nature Reviews Genetics 10 6 392 404 D’Angelo G Rao DC Gu CC 2009 Combining least absolute shrinkage and selection operator (LASSO) and principal-components analysis for detection of gene-gene interactions in genome-wide association studies BMC Proceedings 3 Suppl 7 S62 20018056 Devlin B Roeder K Wasserman L 2003 Analysis of multilocus models of association Genetic Epidemiology 25 1 36 47 12813725 Efron B Gong G 1983 A leisurely look at the bootstrap, the jackknife, and Cross-Validation The American Statistician 37 1 36 48 Efron B Hastie T Johnstone L Tibshirani R 2004 Least angle regression Annals of Statistics 32 407 499 Efron B Tibshirani RJ 1994 An Introduction to the Bootstrap (Chapman & Hall/CRC Monographs on Statistics & Applied Probability) Chapman and Hall/CRC Fan J Guo S Hao N 2012 Variance estimation using refitted cross-validation in ultrahigh dimensional regression Journal of the Royal Statistical Society: Series B (Statistical Methodology) 74 1 37 65 22312234 Fawcett T 2006 An introduction to roc analysis Pattern recognition letters 27 8 861 874 Freedman D Lane D 1983 A nonstochastic interpretation of reported significance levels Journal of Business & Economic Statistics 1 4 292 298 Friedman J Hastie T Höfling H Tibshirani R 2007 Pathwise coordinate optimization Annals of Applied Statistics 1 2 302 332 Friedman J Hastie T Tibshirani R 2010 Regularization paths for generalized linear models via coordinate descent Journal of statistical software 33 1 1 20808728 He Q Lin DYY 2011 A variable selection method for genome-wide association studies Bioinformatics (Oxford, England) 27 1 1 8 James GM Radchenko P 2009 A generalized dantzig selector with shrinkage tuning Biometrika 96 2 323 337 Knight K Fu W 2000 Asymptotics for Lasso-Type estimators The Annals of Statistics 28 5 1356 1378 Kyung M Gill J Ghosh M Casella G 2010 Penalized regression, standard errors, and bayesian lassos Bayesian Analysis 5 369 412 Lange K Papp JC Sinsheimer JS Sobel EM 2014 Next-generation statistical genetics: Modeling, penalization, and optimization in high-dimensional data Annual review of statistics and its applications 1 279 300 Leng C Lin Y Wahba G 2006 A note on the lasso and related procedures in model selection Statistica Sinica 16 4 1273 Lockhart R Taylor J Tibshirani RJ Tibshirani R 2013 A significance test for the lasso arXiv preprint arXiv:1301.7161 Lubke G Laurin C Amin N Hottenga J Willemsen G van Grootheest G Abdel-laoui A Karssen L Oostra B van Duijn C 2014 Genome-wide analyses of borderline personality features Molecular psychiatry 19 8 923 929 23979607 Meinshausen N Bühlmann P 2010 Stability selection Journal of the Royal Statistical Society: Series B (Statistical Methodology) 72 4 417 473 Minnier J Tian L Cai T 2011 A perturbation method for inference on regularized regression estimates Journal of the American Statistical Association 106 496 1371 1382 22844171 Morey LC 1991 Personality Assessment Inventory: Professional manual Psychological Assessment Resources Odessa, FL Motyer A McKendry C Galbraith S Wilson S 2011 LASSO model selection with post-processing for a genome-wide association study data set BMC Proceedings 5 Suppl 9 S24 Okser S Pahikkala T Airola A Salakoski T Ripatti S Aittokallio T 2014 Regularized machine learning in the genetic prediction of complex traits PLoS Genet 10 11 e1004754 25393026 Osborne MR Presnell B Turlach BA 2000 On the LASSO and its dual Journal of Computational and Graphical Statistics 9 2 319 337 Park JHH Gail MH Weinberg CR Carroll RJ Chung CC Wang Z Chanock SJ Fraumeni JF Chatterjee N 2011 Distribution of allele frequencies and effect sizes and their interrelationships for common genetic susceptibility variants Proceedings of the National Academy of Sciences of the United States of America 108 44 18026 18031 22003128 Pötscher BM Leeb H 2009 On the distribution of penalized maximum likelihood estimators: The LASSO, SCAD, and thresholding Journal of Multivariate Analysis 100 9 2065 2082 Price AL Zaitlen NA Reich D Patterson N 2010 New approaches to population stratification in genome-wide association studies Nature Reviews Genetics 11 7 459 463 R Core Team 2013 R: A Language and Environment for Statistical Computing R Foundation for Statistical Computing Vienna, Austria Sartori S 2009 Penalized Regression: bootstrap confidence intervals and variable selection for high dimensional data sets PhD thesis Universitá Degli Studi di Milano Sivakumaran S Agakov F Theodoratou E Prendergast JG Zgaga L Manolio T Rudan I McKeigue P Wilson JF Campbell H 2011 Abundant pleiotropy in human complex diseases and traits The American Journal of Human Genetics 89 5 607 618 22077970 Stefansson H Ophoff RA Steinberg S Andreassen OA Cichon S Rujescu D Werge T Pietiläinen OP Mors O Mortensen PB 2009 Common variants conferring risk of schizophrenia Nature 460 7256 744 747 19571808 Tibshirani R 1996 Regression Shrinkage and Selection via the Lasso Journal of the Royal Statistical Society. Series B (Methodological) 58 1 267 288 Tibshirani R 2011 Regression shrinkage and selection via the lasso: a retrospective Journal of the Royal Statistical Society: Series B (Statistical Methodology) 73 3 273 282 Tibshirani RJ 2013 The lasso problem and uniqueness Electronic Journal of Statistics 7 0 1456 1490 Valdar W Sabourin J Nobel A Holmes CC 2012 Reprioritizing genetic associations in hit regions using LASSO-based resample model averaging Genetic Epidemiology 36 5 451 462 22549815 Waldron L Pintilie M Tsao MS Shepherd FA Huttenhower C Jurisica I 2011 Optimized application of penalized regression methods to diverse genomic data Bioinformatics 27 24 3399 3406 22156367 Warren H Casas JP Hingorani A Dudbridge F Whittaker J 2014 Genetic prediction of quantitative lipid traits: comparing shrinkage models to gene scores Genetic epidemiology 38 1 72 83 24272946 Willemsen G Vink JM Abdellaoui A den Braber A van Beek JH Draisma HH van Dongen J van’t Ent D Geels LM van Lien R 2013 The adult netherlands twin register: Twenty-five years of survey and biological data collection Twin Research and Human Genetics 16 01 271 281 23298648 Zhang P 1993 Model selection via multifold cross validation The Annals of Statistics 21 1 299 313

PMC005xxxxxx/PMC5132183.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 101694381 45813 GMS Infect Dis GMS Infect Dis GMS infectious diseases 2195-8831 27917362 5132183 10.3205/id000018 NIHMS822785 Article Correlation of (1→3)-β-D-glucan with other inflammation markers in chronically HIV infected persons on suppressive antiretroviral therapy Hoenigl Martin 123 de Oliveira Michelli Faria 1 Pérez-Santiago Josué 1 Zhang Yonglong 4 Woods Steven Paul 5 Finkelman Malcolm 4 Gianella Sara 1 1 Division of Infectious Diseases, University of California San Diego, San Diego, California, USA 2 Section of Infectious Diseases and Tropical Medicine, Department of Internal Medicine, Medical University of Graz, Graz, Austria 3 Division of Pulmonology, Department of Internal Medicine, Medical University of Graz, Graz, Austria 4 Research Laboratory, Associates of Cape Cod, Inc., Falmouth, Massachusetts, USA 5 Department of Psychology, University of Houston, Texas, USA Corresponding author: Martin Hoenigl, MD, Division of Infectious Diseases, Department of Medicine, University of California, San Diego, 200 West Arbor Drive #8208, San Diego, CA 92103, USA, Phone: +1 (619) 543-5605, mhoenigl@ucsd.edu Martin Hoenigl and Michelli Faria de Oliveira contributed equally to this work. 5 11 2016 22 12 2015 2015 01 12 2016 3 Doc3This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. We evaluated associations between levels of BDG and other biomarkers of inflammation in blood from 41 virologically suppressed persons with chronic HIV-infection. We found a significant correlation between BDG and neopterin levels (r=0.68), and trends to significance for correlations with other inflammation markers (tumor-necrosis-factor-α: r=0.30; interleukin-8: r=0.30; interleukin-6: r=0.28). In conclusion, BDG levels correlated with inflammation markers in a cohort of virologically suppressed individuals with chronic HIV infection. Future studies are needed to evaluate whether BDG may be a marker for morbidity in chronic HIV infection. beta-D-glucan HIV virologically suppressed neopterin plasma interleukin Introduction Antiretroviral therapy (ART) can suppress HIV replication for long-term in most HIV-infected individuals who have good adherence to therapy [1]. Although the advent of ART has increased the life expectancy and decreased morbidity of HIV-infected individuals, they still present higher rates of non-AIDS defining disorders, such as cardiovascular disease and neurocognitive impairment, than HIV-uninfected individuals [2]. These events have been associated with residual immune dysfunction which persists in some individuals despite long term suppressive ART [3]. The exact mechanism of chronic immune dysfunction is incompletely understood and most likely multifactorial in origin. Previous reports have indicated that microbial translocation of bacterial and fungal commensals of the gastrointestinal tract into systemic circulation is one important factor, which is associated with immune dysfunction, persistent inflammation and likely also plays a role in the disease progression, and non-AIDS associated comorbidities [4], [5]. The enteropathy associated with HIV infection is characterized by microbial over-growth in the intestinal lumen and disrupted intestinal permeability resulting in increased levels of lipopolysaccharides (LPS) and 16S rRNA in blood plasma. Such microbial translocation most likely leads to local and systemic immune activation, characterized by increased levels of pro-inflammatory cytokines such as tumour-necrosis-factor α (TNF-α), neopterin, cluster of differentiation 14 (CD14), interleukin-8 (IL-8), and interleukin-6 (IL-6) [6], [7], [8], [9]. In particular, plasma neopterin is an established marker of monocyte activation and was repeatedly associated with HIV disease progression, greater peripheral monocyte HIV DNA reservoirs and negative neurocognitive and cardiovascular outcomes [10], [11], [12]. (1→3)-β-D-glucan (BDG) is a polysaccharide component of the cell wall of Pneumocystis jirovecii (PJ), Candida spp. and several other fungal species excluding Mucorales and Cryptococcus spp. A number of previous studies have shown that blood BDG levels, determined by the Fungitell® assay (Associates of Cape Cod, USA) in serum are useful for early diagnosis of invasive fungal infections (IFI) or PJ-pneumonia (PJP) [13], [14], [15], [16], [17]. In the absence of an active IFI or PJP, serum BDG may be a reasonable indicator of gut mucosal barrier impairment [18], [19] and microbial translocation [20]. The latter was recently reported also for a cohort of HIV-infected subjects [5]. Methods In this study we measured plasma BDG and compared levels with those of established biomarkers of immune activation and microbial translocation in a cohort of virologically suppressed individuals with chronic HIV infection. Study samples were collected as part of a prospective study between May 2008 and February 2013 at the University of California, San Diego. Plasma samples were stored at −80°C at the day of collection and 41 samples from 41 subjects with suppressed levels of HIV RNA were randomly selected for retrospective evaluation of BDG levels and other biomarkers. sCD14 (Trillium Diagnostics, Brewer, ME, USA) and neopterin (Thermo Scientific, Waltham, MA, USA) were measured by enzyme-linked immunosorbent assays (ELISAs), while IL-8, IL-6 and TNF-α were measured by electrochemiluminescence multiplex assay (Meso Scale Diagnostics, Rockville, MD, USA), all according to the manufacturer’s procedures. BDG testing of plasma samples was performed in March 2015 at Associates of Cape Cod, Inc., research laboratories using the Fungitell assay (Cape Cod, Inc., East Falmouth, USA). For statistical analysis SPSS 21 (SPSS Inc., Chicago, IL, USA) was used. BDG levels were squareroot transformed to achieve a distribution close to normal. Correlation between levels of BDG and levels of other biomarkers was calculated using Pearson correlation analysis. The UCSD Human Research Protections Program approved the study protocol, consent and all study related procedures. All study participants provided voluntary, written informed consent before any study procedures were undertaken. Results Median age of the study population was 51 years (range 22–71), 32 participants were males, 9 females. Twenty-six were Caucasian, 9 African-American and 6 reported other race. Median estimated duration of infection was 14.4 years (range 0.4–26.3 years), median CD4 cell count was 643 (range 196–1,740). All participants were virologically suppressed at the time of sampling, with a minority (25%) still being on their first ART regimen. None of the participants had an active fungal infection and none was treated with systemic antifungal agents during the 6 months before participating in the study. Median BDG level was 15 pg/mL (range: 5–238 pg/mL). BDG levels, levels of other biomarkers and correlations are displayed in Table 1. Higher levels of BDG were associated with higher levels of neopterin (r=0.68; p<0.001). We also found some non-significant trends for positive correlations between BDG and other inflammation markers, while no correlation was found between BDG and sCD14. Results are shown in Table 1. In addition, higher levels of BDG were correlated with higher percentage of neutrophils among white blood cell count (r=0.35, p=0.024). No correlations were found between BDG and age, sex, and estimated duration of infection. BDG was significantly higher in those with a CD4 count below 300 cells/mL (n=4), when compared to those above that threshold (n=37; p<0.001, two-tailed t-test). Discussion We found that BDG levels were low in a cohort of virologically suppressed individuals with chronic HIV infection, but nevertheless correlated strongly with plasma neopterin levels, and also slightly with TNF-α, IL-6, and IL-8 levels. Our results confirm findings of a previous study, that evaluated BDG levels in a cohort of in HIV-infected outpatients (majority not virologically suppressed) and found that plasma IL-8 (p=0.03), and TNF-α (p=0.03) were increased in those with high BDG levels (i.e. >40 pg/mL), while IL-6 was not significantly different [5]. Another study that evaluated BDG as a marker for cryptococcal meningitis among HIV-infected individuals in cerebrospinal fluid (CSF) (21) found positive correlations between BDG and IL-8 (p<0.01), and also TNF-α (p=0.02), while again no correlation was found with IL-6 [21]. The observed correlation of BDG with IL-8 may be explained by findings of another study reporting that (1→3)-β-D-glucans powerfully co-stimulate cytokine production (IL-6/IL-8) [22]. However it is unclear why a similar association was not observed with IL-6. Our results further indicate that BDG levels are higher in asymptomatic individuals with CD4 counts below 300 cells/mL. Explanations include potential colonization or subclinical infection with Candida spp. or Pneumocystis that may occur more frequently in individuals with lower CD4 counts [13]. It has been shown previously that BDG levels were markedly higher (mean 142 pg/mL) in a HIV infected cohort with lower median CD4 counts (26, IQR 10–53, all without opportunistic infections), when compared to the cohort studied here (with a median CD4 count >600 pg/mL) [13]. In another study, high serum BDG levels (>40 pg/mL) were more likely to occur in individuals with CD4 counts less than 200 cells/mL (31.8% vs. 8.4%, p<0.01), higher HIV viral levels (2.85 vs. 2.13 log10 copies/mL, p<0.01), and those without ART (68.2% vs. 90.0%, p<0.01) [5]. Major limitations of our pilot study include the small sample size. To further examine the role of BDG as a potential biomarker for microbial translocation and its correlation with immune dysfunction and non-AIDS clinical events during HIV infection more comprehensive studies will be necessary. Also BDG levels were determined in plasma samples. While the test is officially licensed for testing of serum samples, previous studies have indicated that detection of BDG in plasma may be associated with similar performance characteristics [14]. Conclusion In conclusion, BDG levels correlated with inflammation markers in a cohort of virologically suppressed individuals with chronic HIV infection. Future studies are needed to evaluate whether BDG may be a marker for morbidity in chronic HIV infection. Acknowledgements/Funding The study cohort was derived from National Institutes of Health grant number MH073419. This work was also supported by funds from the following: the Max Kade Foundation, New York (Max Kade Postdoctoral Research grant), Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq-Brazil), Interdisciplicinary Research Fellowship in NeuroAIDS (R25-MH081482), HNRP developmental grant PST-HN39, and grants from the National Institutes of Health: MH101012, AI100665, MH097520, DA034978, AI036214, AI007384, AI027763, AI106039, AI43638, AI074621. Martin Hoenigl served on the speakers’ bureau of Merck. Yonglong Zhang and Malcolm Finkelman are employees of Associates of Cape Cod. Table 1 Results for all investigated biomarkers [median and (IQR) or mean ± standard deviation (SD) are displayed] and correlation of β-D-glucan (BDG; squareroot transformed to achieve distribution close to normal) with other biomarkers Biomarkers Results [median (IQR) or mean ± SD] Correlation between squareroot BDG and other biomarkers p-value BDG (pg/mL) 15 (8–21) – – Square root BDG 4.96 ± 2.33 – – TNF-α (pg/mL) 1.82 ± 0.88 r=0.29 0.06 Neopterin (ng/mL) 2.94 ± 1.86 r=0.68 <0.001 IL-8 (pg/mL) 6.31 ± 3.30 r=0.30 0.06 IL-6 (pg/mL) 0.6319 ± 0.4880 r=0.28 0.07 sCD14 (ng/mL) 1,425 ± 650 r=0.19 0.23 Competing interests All other authors declare that they have no competing interests. References 1 Palella FJ Jr Baker RK Moorman AC Chmiel JS Wood KC Brooks JT Holmberg SD HIV Outpatient Study Investigators Mortality in the highly active antiretroviral therapy era: changing causes of death and disease in the HIV outpatient study J Acquir Immune Defic Syndr 2006 9 43 1 27 34 16878047 2 Hunt PW HIV and inflammation: mechanisms and consequences Curr HIV/AIDS Rep 2012 6 9 2 139 147 22528766 3 Deeks SG Verdin E McCune JM Immunosenescence and HIV Curr Opin Immunol 2012 8 24 4 501 506 22658763 4 Marchetti G Tincati C Silvestri G Microbial translocation in the pathogenesis of HIV infection and AIDS Clin Microbiol Rev 2013 1 26 1 2 18 23297256 5 Morris A Hillenbrand M Finkelman M George MP Singh V Kessinger C Lucht L Busch M McMahon D Weinman R Steele C Norris KA Gingo MR Serum (1→3)-β-D-glucan levels in HIV-infected individuals are associated with immunosuppression, inflammation, and cardiopulmonary function J Acquir Immune Defic Syndr 2012 12 61 4 462 468 22972021 6 Karris MY Kao YT Patel D Dawson M Woods SP Vaida F Spina C Richman D Little S Smith DM Predictors of virologic response in persons who start antiretroviral therapy during recent HIV infection AIDS 2014 3 28 6 841 849 24401640 7 Klatt NR Chomont N Douek DC Deeks SG Immune activation and HIV persistence: implications for curative approaches to HIV infection Immunol Rev 2013 7 254 1 326 342 23772629 8 Canipe A Chidumayo T Blevins M Bestawros M Bala J Kelly P Filteau S Shepherd BE Heimburger DC Koethe JR A 12 week longitudinal study of microbial translocation and systemic inflammation in undernourished HIV-infected Zambians initiating antiretroviral therapy BMC Infect Dis 2014 9 29 14 521 25266928 9 Chevalier MF Petitjean G Dunyach-Rémy C Didier C Girard PM Manea ME Campa P Meyer L Rouzioux C Lavigne JP Barré-Sinoussi F Scott-Algara D Weiss L The Th17/Treg ratio, IL-1RA and sCD14 levels in primary HIV infection predict the T-cell activation set point in the absence of systemic microbial translocation PLoS Pathog 2013 9 6 e1003453 23818854 10 Wang SX Ho EL Grill M Lee E Peterson J Robertson K Fuchs D Sinclair E Price RW Spudich S Peripheral neuropathy in primary HIV infection associates with systemic and central nervous system immune activation J Acquir Immune Defic Syndr 2014 7 66 3 303 310 24732871 11 Mildvan D Spritzler J Grossberg SE Fahey JL Johnston DM Schock BR Kagan J Serum neopterin, an immune activation marker, independently predicts disease progression in advanced HIV-1 infection Clin Infect Dis 2005 3 40 6 853 858 15736019 12 Nyamweya S Townend J Zaman A Steele SJ Jeffries D Rowland-Jones S Whittle H Flanagan KL Jaye A Are plasma biomarkers of immune activation predictive of HIV progression: a longitudinal comparison and analyses in HIV-1 and HIV-2 infections? PLoS ONE 2012 7 9 e44411 22970212 13 Sax PE Komarow L Finkelman MA Grant PM Andersen J Scully E Powderly WG Zolopa AR AIDS Clinical Trials Group Study A5164 Team Blood (1→3)-beta-D-glucan as a diagnostic test for HIV-related Pneumocystis jirovecii pneumonia Clin Infect Dis 2011 7 53 2 197 202 21690628 14 Karageorgopoulos DE Vouloumanou EK Ntziora F Michalopoulos A Rafailidis PI Falagas ME β-D-glucan assay for the diagnosis of invasive fungal infections: a meta-analysis Clin Infect Dis 2011 3 52 6 750 770 21367728 15 Hoenigl M Prattes J Spiess B Wagner J Prueller F Raggam RB Posch V Duettmann W Hoenigl K Wölfler A Koidl C Buzina W Reinwald M Thornton CR Krause R Buchheidt D Performance of galactomannan, beta-d-glucan, Aspergillus lateral-flow device, conventional culture, and PCR tests with bronchoalveolar lavage fluid for diagnosis of invasive pulmonary aspergillosis J Clin Microbiol 2014 6 52 6 2039 2045 24671798 16 Prattes J Hoenigl M Rabensteiner J Raggam RB Prueller F Zollner-Schwetz I Valentin T Hönigl K Fruhwald S Krause R Serum 1,3-beta-d-glucan for antifungal treatment stratification at the intensive care unit and the influence of surgery Mycoses 2014 11 57 11 679 686 25040144 17 Prüller F Wagner J Raggam RB Hoenigl M Kessler HH Truschnig-Wilders M Krause R Automation of serum (1→3)-beta-D-glucan testing allows reliable and rapid discrimination of patients with and without candidemia Med Mycol 2014 7 52 5 455 461 24906361 18 Shahid Z Sanathkumar N Restrepo A Haider S Muzaffar J Grazziutti M Nucci M Anaissie E Elevated serum beta-D-glucan (BDG) as a marker for chemotherapy-induced mucosal barrier injury (MBI) in adults with hematologic malignancies: a retrospective analysis 49th Annual Meeting of the Infectious Diseases Society of America (IDSA) 2011 10 20–23 Boston, USA 2011 Abstract 1334. Available from: https://idsa.confex.com/idsa/2011/webprogram/Paper31834.html 19 Ellis M Al-Ramadi B Finkelman M Hedstrom U Kristensen J Ali-Zadeh H Klingspor L Assessment of the clinical utility of serial beta-D-glucan concentrations in patients with persistent neutropenic fever J Med Microbiol 2008 3 57 Pt 3 287 295 18287290 20 Held J Kohlberger I Rappold E Busse Grawitz A Häcker G Comparison of (1→3)-β-D-glucan, mannan/anti-mannan antibodies, and Cand-Tec Candida antigen as serum biomarkers for candidemia J Clin Microbiol 2013 4 51 4 1158 1164 23363830 21 Rhein J Bahr NC Morawski BM Schutz C Zhang Y Finkelman M Meya DB Meintjes G Boulware DR Detection of High Cerebrospinal Fluid Levels of (1→3)-β-d-Glucan in Cryptococcal Meningitis Open Forum Infect Dis 2014 11 26 1 3 ofu105 25734173 22 Kikkert R Bulder I de Groot ER Aarden LA Finkelman MA Potentiation of Toll-like receptor-induced cytokine production by (1→3)-beta-D-glucans: implications for the monocyte activation test J Endotoxin Res 2007 13 3 140 149 17621556

PMC005xxxxxx/PMC5133176.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 8218602 3203 Crisis Crisis Crisis 0227-5910 2151-2396 27245814 5133176 10.1027/0227-5910/a000383 NIHMS812845 Article Suicide Attempts and Deaths in Sofala, Mozambique, From 2011 to 2014 Who, Where, and From What Wagenaar Bradley H. 12 Raunig-Berhó Manuela 1 Cumbe Vasco 34 Rao Deepa 15 Napúa Manuel 6 Sherr Kenneth 12 1 Department of Global Health, University of Washington, Seattle, WA, USA 2 Health Alliance International, Seattle, WA, USA 3 Sofala Provincial Health Directorate, Department of Mental Health, Ministry of Health, Beira, Mozambique 4 Beira Central Hospital, Department of Medicine, Psychiatric Services, Beira, Mozambique 5 Department of Psychiatry and Behavioral Sciences, University of Washington, Seattle, WA, USA 6 Beira Operations Research Center, Ministry of Health, Beira, Mozambique Bradley H. Wagenaar Department of Global Health, University of Washington, Harris Hydraulics Laboratory, Box 357965, Seattle, WA 98195-7965, USA, Tel. +1 (206) 221-4970, Fax +1 (206) 685-8519, wagenaarb@gmail.com 27 8 2016 01 6 2016 11 2016 01 11 2017 37 6 445453 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. Background Mozambique was recently estimated to have the highest suicide rate in Africa. Aims To fill a knowledge gap on suicide attempts and deaths in Mozambique. Method We reviewed a census of 898 emergency psychiatric consultations from March 2013 to July 2014 and 1,173 violent death autopsy records from June 2011 to August 2014 at Beira Central Hospital in Sofala, Mozambique. Results In all, 18.0% of emergency psychiatric consultations were suicide attempts. Females were disproportionately represented (68.3%, p < .001), and the mean age was 26.8 years. Rat poison was used in 66% of attempts, followed by unspecified methods (19.8%), and unspecified poisoning (6.8%). Of the violent death autopsies, 10% were suicides. Suicide deaths were more likely to be male (67.3%, p < .001), and the mean age was 30.8 years. Common methods were hanging (43.2%), unspecified substance (28.0%), or rat poison (26.3%). Common places of death were the hospital or hospital transit (46.4%) and the household (35.7%). Female suicide deaths more often involved toxic substances and males more often employed hanging. Conclusion Females more often present with suicide attempts, but deaths due to suicide are more frequent among males. Females more often use toxic substances, whereas males more often use lethal methods, such as hanging. Policies to reduce the availability or toxicity of rat poison should be considered. suicide attempts methods of suicide mental health systems epidemiology Mozambique Over 800,000 individuals globally are estimated to have died from suicide in 2012 (World Health Organization [WHO], 2014), although this number is very likely an under-estimate due to misclassification of suicide deaths (Rockett & Thomas, 1999) as well as limited coverage of vital registration (Mathers, Fat, Inoue, Rao, & Lopez, 2005) and autopsy systems (Kapusta et al., 2011) in many countries. Trends in suicide rates have shown disparate patterns globally over the past decade. While estimated suicide rates have dropped over 45% in low- and middle-income countries (LMICs) in the Western Pacific Region, rates have increased over 35% in LMICs in the African region (WHO, 2014). Across all countries, there are an estimated 1.9 male suicides for every female suicide and 20 suicide attempts for each completed suicide (WHO, 2014). Unfortunately, WHO member states only report suicide mortality statistics; globally there are no official or routinely collected data on suicide attempts. Sex ratios for completed suicide differ markedly across the world, most often attributed to differences in methods of suicidal behavior and specific sociocultural factors (Biddle et al., 2010; Booth, 1999; Gunnell & Eddleston, 2003; Park, Ahn, Lee, & Hong, 2014; Phillips, Li, & Zhang, 2002; Phillips, Yang, et al., 2002). Globally, women are more likely to attempt suicide, but often employ a less lethal method than males do, leading to the so-called suicide paradox whereby females are more likely to attempt suicide but males have higher rates of death due to suicide (Bertolote et al., 2005; Nock et al., 2008; Schrijvers, Bollen, & Sabbe, 2012). This is not the case in China, where females have historically had higher rates of completed suicide than males, although this pattern has been changing rapidly. This distinctive pattern of high rates of female suicide in China has been attributed to elevated rates of completed suicide among rural women using highly toxic dichlorvos and parathion organophosphate pesticides, along with unique sociocultural factors in China (Eyer et al., 2003; Hendin et al., 2008; Meng, 2002; Phillips, Li, et al., 2002; WHO, 2014; Zhang & Ma, 2012). A number of risk factors for death due to suicide have been consistently observed across countries and cultures, including: the presence of a mental disorder, family history of psychopathology, stressful life events in the past month, young or old age, low socioeconomic status, and previous suicide attempts (Phillips, Yang, et al., 2002; Vijayakumar, John, Pirkis, & Whiteford, 2005; Vijayakumar & Rajkumar, 1999; Yoshimasu, Kiyohara, & Miyashita, 2008). Across eight diverse emergency-care settings in LMICs, self-poisoning has been identified as the primary method of attempted suicide across all locations, accounting for 69–98% of attempts. In neighboring South Africa, 71% of suicide attempts identified in the ER were female and the mean age was 21 years (Fleischmann et al., 2005). Of 300 attempted suicides presenting at the Department of Emergency Medicine at a large referral hospital in Dar es Salaam, Tanzania, the mean age of attempted suicide was 23.7 years, 68.7% of attempts were females, and 91% of attempts employed self-poisoning in the form of medication or poison (Ndosi & Waziri, 1997). In a South African study of patients seen for a suicide attempt at a referral hospital in Bloemfontein, the median age of patients was 22 years, 68.9% were female, and drug overdose was the most common method of attempt (du Toit et al., 2008). Previous suicide prevention efforts in diverse settings have focused on: (a) reducing the availability or toxicity/danger of commonly used suicide methods such as pesticides, domestic gasoline, or handguns; (b) media interventions to ensure responsible reporting practices around suicide to limit imitation or glamorization of suicide; (c) school-based interventions around crisis management and coping skills; and (d) ensuring positive attitudes toward suicidal patients and understanding of local idioms of distress among medical professionals (Bhana, Petersen, Baillie, Flisher, & The Mhapp Research Programme Consortium, 2010; Etzersdorfer, Vijayakumar, Schony, Grausgruber, & Sonneck, 1998; Hagaman et al., 2013; Hendin et al., 2008; Khan, 2005; WHO, 2001, 2008a, 2008b). Up to 45% of individuals who die by suicide visit their primary care physician within 1 month of suicidal death and over 75% of suicide deaths have contact with their provider within 1 year of death, highlighting the missed opportunities for prevention without adequate integration of mental health and targeted suicidal screening into primary care settings (Luoma, Martin, & Pearson, 2002). According to the 2014 WHO world suicide report, Mozambique has the seventh highest suicide rate in the world (27.4/100,000), with a rate over double the global average of 11.4 per 100,000 (WHO, 2014). However, Mozambique has no published national suicide statistics, no comprehensive vital registration system, and, to our knowledge, there are only two peer-reviewed publications that mention suicide. The first details that suicides made up 4.2% of deaths due to injury in Maputo City in the year 2000 (Nizamo, Meyrowitsch, Zacarias, & Konradsen, 2006), and the second indicates that, of injury-related maternal deaths from 1991 to 1995 in Maputo, 33% were due to suicide (Granja, Carla, Zacarias, & Bergstrom, 2002). In terms of emergency room (ER) psychiatric services for self-harm, at psycho-trauma centers in northern Uganda, 5.2% of consultations were for suicide attempts (Nakimuli-Mpungu et al., 2013). In South Africa, 17.7% of all consultations of youth under age 19 referred in a psychiatric department were for attempted suicide (Schlebusch, 1985). In Ethiopia, 19.2% of patients attending an adult outpatient psychiatry clinic had previously attempted suicide, with the most common attempt method being hanging (Mekonnen & Kebede, 2011). In Malawi, deaths due to suicide accounted for 17% of autopsies conducted at the Queen Elizabeth Central Hospital, with a mean age of 33.4 and 77% of deaths being male; in this setting the most common method of suicide death was self-poisoning with carbamate rat poison or organophosphate pesticides (79%), followed by hanging (19%; Dzamalala, Milner, & Liomba, 2006). In Uganda, a retrospective review of suicide death records found a mean age of 30.6, a majority of suicide deaths among males (77%), and hanging as the most common method of death (63%), followed by self-poisoning (26%; Kinyanda, Wamala, Musisi, & Hjelmeland, 2011). To date, there are no peer-reviewed studies in Mozambique on the epidemiologic profile (age, gender profile, method used) of suicide deaths. In addition to a lack of detailed understanding around suicide deaths in Mozambique, there are no peer-reviewed studies in Mozambique detailing the demographic and epidemiologic profile of utilization of ER psychiatric services, including suicide attempts. The purpose of the present study was to address these gaps in the mental health literature, representing the first assessment of attempted suicide identified in the ER by age, sex, and diagnosis, as well as analyzing deaths due to suicide and suicide methods used. We aim for these data to inform future policies and programs to improve the prevention, care, and treatment for mental disorders and suicidal behavior across Mozambique and other similar LMICs. Method Study Setting Sofala Province, Mozambique, has approximately two million inhabitants (United States Central Intelligence Agency, 2014a) and 14 psychiatric technicians, two adult psychiatrists, one child psychiatrist, and 11 clinical psychologists providing mental health services operating out of 18 health facilities. Twelve of 13 districts have at least one clinic providing outpatient mental health-care services, primarily located at large central or district-level referral hospitals. In the Mozambican system, psychiatric technicians can diagnose and prescribe psychotropic medications following a 2-year training program in place since the first cadre graduated in 1996 (dos Santos, 2011). The Beira Central Hospital is one of three quaternary-level specialist facilities nationwide and provides the largest number of outpatient psychiatric consultations of any facility in Sofala Province, in addition to inpatient and emergency room psychiatric services. Adult outpatient, inpatient, and emergency psychiatric services at the Beira Hospital are staffed by two adult psychiatrists (one Mozambican and one Cuban), two psychiatric technicians, and three clinical psychologists. All Mozambican Ministry of Health clinics use the International Classification of Diseases, Tenth Edition (ICD-10) code (WHO, 1992) system to categorize diagnoses of mental disorders. Emergency Room Record Review We reviewed 898 ER psychiatric consultations, representing a census of those conducted at Beira Central Hospital in Sofala, Mozambique, from March 2013 to July 2014. At intake to the ER, if the attending provider recognizes that a given patient has an issue that is psychiatric in nature, they refer the patient to the on-call psychiatric specialist (psychiatrist or psychiatric technician) to conduct a specialized psychiatric consultation. The ER psychiatric consultation registries are hand-written, bound books filled out by the psychiatric specialist at the time of consultation and include the variables of: date of consultation, age, gender, visit number (first visit or second+ visit), and diagnosis. Two abstractors entered data in Excel (Version 2013). Inconsistencies between data abstractors and illegible hand-writing were resolved by revisiting the registry and cross-checking with the psychiatric specialist responsible for a given entry. Legal Medicine Record Review We reviewed 1,173 autopsies for violent deaths, representing a census of those conducted at Beira Central Hospital legal medicine department from June 2011 to August 2014. For suicide data, two data abstractors entered data in Excel 2013, and any inconsistencies between data abstractors or illegible handwriting were resolved by revisiting the registry and/or crosschecking with the legal medicine expert responsible for the registry entry. All available variables were abstracted and included: age, sex, race, date of autopsy, location of death, and cause of death. Statistical Analyses and Variable Classification ER diagnoses were tabulated and two-sample t tests were used to compare the continuous age distributions of subjects with each diagnosis with the mean age of the rest of the sample. Chi-squared tests were used to test for gender differences among diagnoses. Fisher’s exact test was used if any cell was less than 5. Owing to reviewer concerns about missing data and multiple testing, we used a Bonferroni adjustment for the 11 statistical tests of diagnostic categories, resulting in an alpha level of 0.05/11, or 0.0045. Little’s test (Little, 1988) was implemented to test for the missing completely at random (MCAR) assumption for all ER covariates (age, gender, % first visit); this resulted in a p value of .21, failing to reject the null hypothesis of MCAR. Suicide death records were tabulated by suicide method, and procedures similar to the aforementioned ones were used to compare the continuous age distributions of each method (two-sample t tests), as well as gender differences (χ2 tests or Fisher’s exact). For each suicide method, the three most common places of death were tabulated. A one-sample test for proportions using the null hypothesis of 50% female and 50% male was used to test gender differences among all suicide deaths. Little’s test applied to age and gender resulted in a p value of .44, failing to reject the null hypothesis of MCAR. A Bonferroni adjustment was applied for the five statistical tests of suicide methods, resulting in a p-value cut-off of .05/5, or .01. We used Stata 13 for all statistical analyses. Results Emergency Room Psychiatric Consultations Of those without missing data (83.4% for age, 89.8 for gender, and 92.2 for visit number), the mean age of ER consultations was 30.0 (SD = 11.5), 59.1% of consultations were males, and 64.5% were first-visit consultations (see Table 1). A total of 149 (16.6%) records were missing data for age, 92 (10.2%) for gender, 70 (7.8) for visit number, and 30 (3.3%) were missing diagnostic information. Those missing gender information had a mean age of 30.1 years (SD = 13.1), and 29.4% (n = 27) were diagnosed with delirium, 21.7% (n = 20) with a suicide attempt, and 17.4% (n = 16) with psychomotor agitation. Of those missing age information, 46.0% were female (n = 63), and 27.5% (n = 41) were diagnosed with delirium, 17.5% (n = 26) with psychomotor agitation, and 14.8% (n = 22) with a suicide attempt. Of the 898 ER psychiatric consultations, 18% (n = 162) were due to suicide attempt, with other common consultations for delirium (n = 259, 28.8% of consultations), psychomotor agitation (n = 132, 14.7%), psychosis (n = 66, 7.3%), and behavioral disorder due to psychoactive substance use (n = 52, 5.8%). Intoxication by rat poison (known locally as Ratex) was the most common method of suicide attempt (n = 107, 66.0% of attempts), followed by unspecified method (n = 32, 19.8%), unspecified medication intoxication (n = 11, 6.8%), and unspecified chemical intoxication (n = 5, 3.1%; see Table 2). Those presenting for a suicide attempt were significantly younger (mean age = 26.8 years, p < .001) than the average psychiatric ER patient, and were significantly more likely to be female (68.3% female, p < .001; Table 2). Suicide Deaths From Legal Medicine Violent Death Autopsy Records Of the 1,173 autopsies for violent death conducted from June 2011 to August 2014, 118 (10.1%) were suicides, with other common records being accidents (n = 777, 66.2%), homicides (n = 185, 15.8%), and natural deaths (n = 93, 66.2%). The mean age of suicide deaths was 30.8 (SD = 15.8, range = 11–81), and there were significantly more male deaths (n = 76, 64.4% male, p < .001; Table 1). Seven individuals (5.9%) were missing age information and five (4.2%) were missing gender information. Missing age and gender data were co-occurring: of the seven individuals missing age information, four were missing gender information, and of the five individuals missing gender information, four were missing age information. Of those with missing age data, 57% (n = 4) died by hanging and 43% (n = 3) died by rat poison. Of those with missing gender data, 40% (n = 2) died by hanging, 40% (n = 2) died by rat poison, and 20% (n = 1) died by unspecified toxic substance. The most common method of completed suicide was hanging (n = 51, 43.2%), followed by unspecified toxic substance (n = 33, 28.0%), intoxication by rat poison (n = 31, 26.3%), jumping from a high place (n = 2, 1.7%), and asphyxia (n = 1, 0.8%; Table 3). Men more often employed hanging (22.5% female) and women more often employed an unspecified toxic substance (46.9% female), although this difference was not statistically significant after controlling for multiple comparisons (p = .04). The majority of deaths for toxic substance ingestion occurred at the hospital (54.8%–75.0%), whereas the majority who used hanging died in their household (n = 31, 67.4%). Discussion This study sought to elucidate the epidemiologic profile of suicide attempts within the emergency care setting, as well as suicide deaths from violent death autopsies conducted at a large referral hospital in Sofala, Mozambique. We found that suicide attempts were the second most common cause of ER psychiatric visits in Central Mozambique, following delirium. Individuals presenting for suicide attempts were predominantly young females (2.2:1 female-to-male ratio) who had ingested rat poison. By contrast, suicide deaths were most often young males (2.1:1 male-to-female ratio) who had ingested a toxic substance or employed hanging. Our data indicate that the Mozambican suicide profile is similar to well-established Western suicide patterns whereby women attempt suicide at a higher rate than males, but die from suicide at lower rates, potentially due to the use of less-deadly suicide methods. Our findings are in line with a recent systematic review of suicidal behavior in African countries, finding that the most frequent methods of suicide across diverse settings are hanging and pesticide poisoning, and that men are, on average, around three times as likely to die from suicide as women (Mars, Burrows, Hjelmeland, & Gunnell, 2014). Our findings that females make up 68% of attempted suicide patients parallels similar studies conducted in Tanzania and South Africa, finding that females made up between 69 and 71% of attempted suicides in hospital settings (du Toit et al., 2008; Fleischmann et al., 2005; Ndosi & Waziri, 1997). Furthermore, our finding of 64% of suicide deaths among males is similar to other retrospective autopsy reviews in Malawi and Uganda, which both found that 77% of suicide deaths occurred among males (Dzamalala et al., 2006; Kinyanda et al., 2011). That both suicide attempts and deaths were mostly among young individuals (48% of deaths under age 26) is not necessarily surprising given the young age structure of the population in Mozambique (67% of the population is under the age of 25 (United States Central Intelligence Agency, 2014b), yet highlights the importance of targeting youth and teenagers in future suicide prevention interventions. In our sample, suicide deaths were an average of 4 years older than suicide attempts, a trend that is similar across other studies in sub-Saharan Africa (du Toit et al., 2008; Dzamalala et al., 2006; Fleischmann et al., 2005; Kinyanda et al., 2011; Ndosi & Waziri, 1997). The preponderance of suicide attempts and deaths attributable to the ingestion of rat poison is similar to high rates of carbamate rodenticide self-poisoning in Malawi, suicide attempt case reports from Nigeria, and unintentional and intentional poisonings in South Africa (Dzamalala et al., 2006; Eze, 2014; Veale, Wium, & Muller, 2013). Given that the majority of suicide deaths in our sample follow the ingestion of a toxic substance, and that a combined 64% of these patients die only after reaching the hospital, the improvement of hospital treatment protocols and training of emergency providers in poisoning interventions could have a significant effect on survival rates. Additional mixed-methods implementation science could target increased access and effectiveness of emergency transport or improved linkages and referral networks for patients at rural or peripheral facilities. Assessing and implementing strategies to manage stigma toward suicide and, more generally, to individuals suffering from mental health issues among the general population and health-care workers may be essential to obtain more accurate reporting of suicide and suicide attempts (Hagaman et al., 2013). Furthermore, the potentially stigmatizing attitudes of health workers, as well as the community as a whole, toward suicide victims likely decreases help-seeking and may negatively influence health worker behavior toward the care and treatment for suicide victims (Keusch, Wilentz, & Kleinman, 2006; Osafo, Knizek, Akotia, & Hjelmeland, 2012). The next steps for these analyses could be to conduct specific mixed-method studies on: (a) community and health worker stigmatizing attitudes toward suicide victims and associated family or friends; and (b) death rates for suicide attempts with various methods to determine how to better improve treatment and suicide case-fatality rates. Future studies could focus on rural areas, as they may have higher suicide case-fatality rates, either due to less access to effective treatments or the use of different suicide methods, such as fertilizer, which may have higher death rates. Mixed-method studies should additionally be conducted to understand why rat poison is so frequently used and to determine other common methods used for suicide attempts in rural and urban areas. Policies shown to be effective in other LMIC settings should then be adapted and tested around potential restriction of access to these common toxic substances, such as the use of safe storage boxes, or the reformulation of pesticides to be less toxic (Bhana et al., 2010; Etzersdorfer et al., 1998; Hagaman et al., 2013; Hendin et al., 2008; Khan, 2005; Konradsen et al., 2007; WHO, 2001, 2008a, 2008b). Historically, many deaths in the United States were linked to unintentional and intentional poisoning by highly toxic carbamate rodenticide (Waseem, Perry, & Bomann, 2010), leading to policies restricting public access to these substances and the substitution of Coumadin and other long-acting derivatives that are considered relatively safe (Centers for Disease Control and Prevention, 1997). More generally, residential access to organophosphate pesticides in the United States has been limited since the 1996 Food Quality Protection Act, which rapidly led to a 62% decrease in hospitalizations linked to organophosphate exposures from 1995 to 2004 (Sudakin & Power, 2007). Similar local, national, and regional regulatory policies and systems-level interventions are necessary to avoid preventable deaths from the ingestion of highly toxic rodenticides or other organophosphate poisons in LMICs; however, enforcement of bans on toxic substances can prove difficult, as South Africa has experienced following its 2012 ban on the toxic carbamate rodenticide Aldicarb (London & Rother, 2013). While it is easier to affect pesticide restriction/toxicity than restricting access to means of hanging, prevention interventions for hanging could center on changing perceptions of hanging as an easy, painless, effective, or rapid method of suicide (Biddle et al., 2010). With the well-established connection between media reporting practices and the potential for suicide contagion, efforts could be directed toward policies and norms around suicide reporting in Mozambique. Case-control and other population-based epidemiologic studies should be prioritized to understand the population burden of suicidal behavior and Mozambican-specific risk factors for suicide attempts or deaths. Currently, less than 1% of the population is estimated to have access to basic mental health services in Mozambique (WHO, 2011); thus, the improved integration of mental health care, depression/mood disorder screening, and suicide prevention into primary care could have a positive impact on population rates of suicide attempts and deaths. One major limitation of the present analyses is that our data are exclusively from one large referral hospital serving the northern region of Mozambique. While our data are a census of available records from this central hospital, we have no clear understanding of the population coverage of these suicide attempt or death data. Previous efforts to improve vital registration systems in Mozambique have often excluded violent deaths from natural deaths. Going forward we urge these systems to be financed and built in parallel, rather than in a fragmented fashion – all deaths should be treated equally, whether violent or natural. In the shorter term, efforts should be made to triangulate all available suicide death information to estimate population and subpopulation burden. While the WHO has recently published data indicating that Mozambique has the highest suicide rate in Africa, the current low levels of vital registration, lack of an effective national reporting system on violent cause of death, lack of transparency in data sources used for the WHO estimates for Mozambique, and the lack of other rigorous peer-reviewed population-level suicide analyses preclude any strong statements regarding the validity of the published WHO statistics. Another limitation centers on missing information regarding age, gender, and visit number. This issue was most pronounced in the ER psychiatric review, with 17% of age information, 10% of gender information, and 8% of visit number information missing. We cannot rule out systematic causes of missing data, although we hypothesize that missing data are a random subset of the dataset as the pattern of emergency diagnoses for those missing age or gender information mirrors that seen in the full sample. In addition, the age patterns of those for whom gender information was missing and the gender distribution of those for whom age data were missing also do not deviate significantly from those seen in the full sample. Last, given the highly stigmatized nature of suicide attempts and deaths in Mozambique, it is likely that those who seek care from an ER setting or those who have their death registered in local records may differ systematically from those missed by these routine systems. However, violent death registration and autopsy coverage is supposed to cover the entire population catchment area of the Beira Central Hospital. Conclusion The majority of suicide attempts and deaths in Sofala, Mozambique, are among young individuals under the age of 30 who ingest toxic substances, with the single most common substance being rat poison. Females more often attempt suicide using toxic substances, while males make up the majority of suicide deaths and more often use hanging. Given recent WHO publications highlighting Mozambique as having the highest suicide rate in Africa, and seventh globally, this initial systems-level study should urgently be followed up with larger population-based studies to determine the risk factors, burden, and epidemiologic profile of suicide attempts across Mozambique. Policies and interventions to decrease access, toxicity, and/or allure around utilizing rat poison as a suicide method should be concurrently examined. Regional approaches may be necessary to regulate the sale and current widespread availability of highly toxic organophosphate pesticides and rodenticides in southern Africa. Systems-level assessments should be carried out to ensure those who attempt suicide receive optimal care given the constraints on the public-sector health system in Mozambique. Acknowledgments This work was supported by the African Health Initiative of the Doris Duke Charitable Foundation and the University of Washington’s Royalty Research Fund. K. Sherr was supported by Grant Number K02TW009207 from the Fogarty International Center. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. This study was approved by the institutional review boards at the University of Washington and the Mozambican National Institute of Health. Table 1 Demographic characteristics of 898 emergency room (ER) psychiatric consultations conducted March 2013 to July 2014 and 118 suicide deaths from autopsies conducted June 2011 to August 2014 at Beira Central Hospital, Sofala Province, Mozambique Characteristic 898 ER psychiatric consultations 118 suicide deaths N (%) unless noted N (%) unless noted Age, M (SD) 30.0 (11.5) 30.8 (15.8)  <18 36 (4.0) 19 (16.1)  18-25 274 (30.5) 38 (32.2)  26-35 273 (30.4) 19 (16.1)  36-45 100 (11.1) 14 (11.9)  46-55 28 (3.1) 10 (8.5)  56+ 38 (4.2) 11 (9.3)  Missing 149 (16.6) 7 (5.9) Gender  Female 330 (36.8) 37 (31.4)*  Male 476 (53.0) 76 (64.4)*  Missing 92 (10.2) 5 (4.2) Visit number  1 534 (59.5) Not recorded  2+ 294 (32.7)  Missing 70 (7.8) Note. * p < .001 using one-sample test for proportion with null hypothesis of 50% female, 50% male. Table 2 Proportion of consultations for suicide attempt, along with: visit number, age, and gender breakdown of 898 patients seeking care from emergency room psychiatric services from March 2013 to July 2014 at Beira Central Hospital, Sofala Province, Mozambique Consultation type (n, %)a Mean age (SD) Percent female % First visit All other consultations combined (706, 78.6) 30.7 (11.7)* 35.2* 58.2* Suicide attempt (162, 18.0) 26.8 (10.1)* 68.3* 89.5*  Chemical intoxication – rat poison (107, 66.0) 26.9 (9.9)* 65.5* 88.0*  Suicide attempt – unspecifed (32, 19.8) 27.9 (12.3) 68.8* 90.3*  Medication intoxication – unspecifed (11, 6.8) 24.9 (7.5) 81.8 90.0  Chemical intoxication – unspecifed (5, 3.1) 21.8 (5.5) 80.0 100.0  Chemical intoxication – battery acid (3, 1.9) All missing 100.0 100.0  Chemical intoxication – batteries (2, 1.2) 24 (0) 50.0 100.0  Chemical intoxication – gasoline (1, 0.62) Missing 100.0 100.0  Suicide attempt – hanging (1, 0.62) 36 (0) 100.0 100.0 Missing consultation data (30, 3.3) 30.8 (10.9) 32.0 74.1 Note. a Percentages for subgroupings are out of the subgroup total, not out of the total N = 898. * Signifcantly different from the average age, gender distribution, or % first visit using Bonferroni adjustment for 11 statistical tests, resulting in p-value cut-off of .05/11 = .0045. Table 3 Most common suicide methods and place of death by age and gender for 118 suicide deaths from violent death autopsies conducted June 2011 to August 2014 at Beira Central Hospital’s legal medicine department, Sofala Province, Mozambique Suicide method (n, %)a Mean age (SD) Proportion female Three most common places of death (n, %) Hanging (51, 43.2) 33.7 (18.2) 22.5 Household (31, 67.4) Transit to hospital (3, 6.5) In public (3, 6.5) Missing (5, 9.8) Toxic substance – unspecifed (33, 28.0) 27.5 (13.0) 46.9 Beira Central Hospital (24, 75.0) Household (3, 9.4) In public (3, 9.4) Missing (1, 3.0) Chemical intoxication – rat poison (31, 26.3) 29.0 (14.6) 34.5 Beira Central Hospital (17, 54.8) Household (5, 16.1) Transit to hospital (3, 9.7) Missing (0, 0) Jumped from high place (2, 1.7) 35 (2.8) 50.0 Beira Central Hospital (2, 100.0) Missing (0, 0) Asphyxia (1, 0.8) 51 (0.0) 0.0 Household (1, 100.0) Missing (0, 0) Note. a No missing data for suicide method. * Signifcantly different from the average age, gender distribution, or % frst visit using Bonferroni adjustment for fve statistical tests of suicide methods, resulting in a p-value cut-off of .05/5 = .01. The authors have no conflicts of interest to report. About the authers Dr. Wagenaar is Acting Instructor in the Department of Global Health at the University of Washington. He is also a technical advisor for Health Alliance International, a center of the Department of Global Health at the University of Washington, WA, USA. Ms. Raunig-Berhó is an MPH student in the Department of Global Health at the University of Washington, WA, USA. Dr. Cumbe is a licensed psychiatrist, Director of Psychiatry and the Mental Health Service at the Beira Central Hospital, and Provincial Head of the Mental Health Service in Sofala Province, Mozambique. He is also a PhD student in the Department of Psychiatry at the Federal University of São Paulo, Brazil. Dr. Rao is a licensed clinical psychologist and Associate Professor in the Departments of Global Health and Psychiatry and Behavioral Sciences at the University of Washington (UW), WA, USA. She is Director of the UW Program on Global Mental Health and Associate Director of the UW Global Health MPH Program. Dr. Napúa is a medical doctor and Scientific Director at the Beira Operations Research Center in Beira, Mozambique. Dr. Sherr is Associate Professor in the Department of Global Health (DGH) at the University of Washington, WA, with adjunct appointments in epidemiology and industrial and systems engineering. He is also Director of Implementation Science for Health Alliance International and Co-Director of the DGH’s PhD program in implementation science. References Bertolote JM Fleischmann A De Leo D Bolhari J Botega N De Silva D Wasserman D Suicide attempts, plans, and ideation in culturally diverse sites: The WHO SUPRE-MISS community survey Psychological Medicine 2005 35 10 1457 1465 doi:10.1017/S0033291705005404 16164769 Bhana A Petersen I Baillie KL Flisher AJ The Mhapp Research Programme Consortium Implementing the World Health Report 2001 recommendations for integrating mental health into primary health care: A situation analysis of three African countries: Ghana, South Africa and Uganda International Review of Psychiatry 2010 22 6 599 610 doi:10.3109/09540261.2010.536152 21226648 Biddle L Donovan J Owen-Smith A Potokar J Longson D Hawton K Gunnell D Factors influencing the decision to use hanging as a method of suicide: Qualitative study The British Journal of Psychiatry: The Journal of Mental Science 2010 197 4 320 325 doi:10.1192/bjp.bp.109.076349 20884956 Booth H Pacific island suicide in comparative perspective Journal of Biosocial Science 1999 31 4 433 448 10581875 Centers for Disease Control and Prevention Poisonings associated with illegal use of Aldicarb as a rodenticide – New York City, 1994–1997 Morbidity and Mortality Weekly Report (MMWR) 1997 46 41 961 963 9347907 dos Santos PF Availability of mental health services in Mozambique 2011 University of Lisbon Portugal (Unpublished master's thesis) Retrieved from http://run.unl.pt/handle/1036217708 du Toit E Kruger J Swiegers S van der Merwe M Calitz F Philane L Joubert G The profile analysis of attempted-suicide patients referred to Pelonomi Hospital for psychological evaluation and treatment from 1 May 2005 to 30 April 2006 South African Journal of Psychiatry 2008 14 1 20 25 Dzamalala CP Milner D. a Liomba NG Suicide in Blantyre, Malawi (2000-2003) Journal of Clinical Forensic Medicine 2006 13 2 65 69 doi:10.1016/j.jcfm.2005.08.006 16271492 Etzersdorfer E Vijayakumar L Schony W Grausgruber A Sonneck G Attitudes towards suicide among medical students: Comparison between Madras (India) and Vienna (Austria) Social Psychiatry and Psychiatric Epidemiology 1998 33 104 110 9540384 Eyer F Meischner V Kiderlen D Thiermann H Worek F Haberkorn M Eyer P Human parathion poisoning. A toxicokinetic analysis Toxicological Reviews 2003 22 3 143 163 15181664 Eze CU Attributes of patients with suicide attempts seen at the Niger Delta University Teaching Hospital within a year: A case series Tropical Medicine & Surgery 2014 2 3 2 6 doi:10.4172/2329-9088.1000170 Fleischmann A Bertolote JM De Leo D Botega N Phillips M Sisask M Wasserman D Characteristics of attempted suicides seen in emergency-care settings of general hospitals in eight low- and middle-income countries Psychological Medicine 2005 35 10 1467 1474 doi:10.1017/S0033291705005416 16164770 Granja Carla A Zacarias E Bergstrom S Violent deaths: The hidden face of maternal mortality BJOG – An International Journal of Obstetrics and Gynaecology 2002 109 5 8 11843374 Gunnell D Eddleston M Suicide by intentional ingestion of pesticides: A continuing tragedy in developing countries International Journal of Epidemiology 2003 32 6 902 909 doi:10.1093/ije/dyg307 14681240 Hagaman AK Wagenaar BH McLean KE Kaiser BN Winskell K Kohrt BA Suicide in rural Haiti: Clinical and community perceptions of prevalence, etiology, and prevention Social Science & Medicine (1982) 2013 83 61 69 doi:10.1016/j.socscimed.2013.01.032 23465205 Hendin H Phillips MR Vijayakumar L Pirkis J Wang H Yip P Fleischmann A Suicide and suicide prevention in Asia 2008 WHO Geneva, Switzerland Kapusta ND Tran US Rockett IRH De Leo D Naylor CPE Niederkrotenthaler T Sonneck G Declining autopsy rates and suicide misclassification: A cross-national analysis of 35 countries Archives of General Psychiatry 2011 68 10 1050 1057 doi:10.1001/archgenpsychiatry.2011.66 21646567 Keusch G Wilentz J Kleinman A Stigma and global health: Developing a research agenda The Lancet 2006 367 525 527 Khan MM Suicide prevention and developing countries Journal of the Royal Society of Medicine 2005 98 10 459 463 doi:10.1258/jrsm.98.10.459 16199814 Kinyanda E Wamala D Musisi S Hjelmeland H Suicide in urban Kampala , Uganda: A preliminary exploration African Health Sciences 2011 11 2 219 227 21857853 Konradsen F Pieris R Weerasinghe M van der Hoek W Eddleson M Dawson A Community uptake of safe storage boxes to reduce self-poisoning from pesticides in rural Sri Lanka BMC Public Health 2007 7 13 Little RJ A test of missing completely at random for multivariate data with missing data Journal of the American Statistical Association 1988 83 404 1198 1202 London L Rother H-A Poisoning and pesticides The South African Medical Journal 2013 103 9 595 596 24344418 Luoma J Martin C Pearson J Contact with mental health and primary care providers before suicide: A review of the evidence American Journal of Psychiatry 2002 159 6 909 916 12042175 Mars B Burrows S Hjelmeland H Gunnell D Suicidal behaviour across the African continent: A review of the literature BMC Public Health 2014 14 1 606 doi:10.1186/1471-2458-14-606 24927746 Mathers CD Fat DM Inoue M Rao C Lopez AD Counting the dead and what they died from: An assessment of the global status of cause of death data Bulletin of the World Health Organization 2005 83 171 177 15798840 Mekonnen D Kebede Y The prevalence of suicidal ideation and attempts among individuals attending an adult psychiatry out-patient clinic in Gondar, Ethiopia African Health Sciences 2011 11 1 103 107 21572865 Meng L Rebellion and revenge: The meaning of suicide of women in rural China International Journal of Social Welfare 2002 11 4 300 309 Nakimuli-Mpungu E Alderman S Kinyanda E Allden K Betancourt TS Alderman JS Musisi S Implementation and scale-up of psycho-trauma centers in a post-conflict area: A case study of a private-public partnership in northern Uganda PLoS Medicine 2013 10 4 doi:10.1371/journal.pmed.1001427 Ndosi NK Waziri MC The nature of parasuicide in Dar es Salaam, Tanzania Social Science and Medicine 1997 44 1 55 61 Nizamo H Meyrowitsch D Zacarias E Konradsen F Mortality due to injuries in Maputo City, Mozambique International Journal of Injury Control and Safety Promotion 2006 13 1 1 6 16537218 Nock MK Borges G Bromet EJ Alonso J Angermeyer M Beautrais A Williams D Cross-national prevalence and risk factors for suicidal ideation, plans and attempts The British Journal of Psychiatry: The Journal of Mental Science 2008 192 2 98 105 doi:10.1192/bjp.bp.107.040113 18245022 Osafo J Knizek B Akotia C Hjelmeland H Attitudes of psychologists and nurses toward suicide and suicide prevention in Ghana: A qualitative study International Journal of Nursing Studies 2012 49 691 700 22177569 Park S Ahn MH Lee A Hong JP Associations between changes in the pattern of suicide methods and rates in Korea, the US, and Finland International Journal of Mental Health Systems 2014 8 1 22 doi:10.1186/1752-4458-8-22 24949083 Phillips MR Li X Zhang Y Suicide rates in China, 1995-99 Lancet 2002 359 9309 835 840 doi:10.1016/S0140-6736(02)07954-0 11897283 Phillips MR Yang G Zhang Y Wang L Ji H Zhou M Risk factors for suicide in China: A national case-control psychological autopsy study Lancet 2002 360 9347 1728 1736 doi:10.1016/S0140-6736(02)11681-3 12480425 Rockett IRH Thomas BM Reliability and sensitivity of suicide certification in higher-income countries Suicide and Life-Threatening Behavior 1999 29 2 141 149 10407967 Schlebusch L Self-destructive behavior in adolescents South African Medical Journal 1985 68 11 792 795 4071325 Schrijvers DL Bollen J Sabbe BGC The gender paradox in suicidal behavior and its impact on the suicidal process Journal of Affective Disorders 2012 138 1-2 19 26 doi:10.1016/j.jad.2011.03.050 21529962 Sudakin DL Power LE Organophosphate exposures in the United States:A longitudinal analysis of incidents reported to poison centers Journal of Toxicology and Environmental Health. Part A 2007 70 2 141 147 doi:10.1080/15287390600755224 17365575 United States Central Intelligence Agency Central Intelligence Agency World Factbook: Mozambique 2014a Retrieved from https://www.cia.gov/library/publications/the-world-factbook/geos/mz.html United States Central Intelligence Agency Central Intelligence Agency World Factbook: Mozambique 2014b Retrieved from https://www.cia.gov/library/publications/the-world-factbook/geos/mz.html Veale D Wium C Muller G Toxicovigilance I: A survey of acute poisonings in South Africa based on Tygerberg Poison Information Centre data The South African Medical Journal 2013 103 5 293 297 23971117 Vijayakumar L John S Pirkis J Whiteford H Suicide in developing countries (2): Risk factors Crisis: The Journal of Crisis Intervention and Suicide Prevention 2005 26 3 112 119 Vijayakumar L Rajkumar S Are risk factors for suicide universal? A case-control study in India Acta Psychiatrica Scandinavica 1999 99 6 407 411 10408261 Waseem M Perry C Bomann S Cholinergic crisis after rodenticide poisoning Western Journal of Emergency Medicine 2010 11 5 524 527 21293782 World Health Organization The ICD-10 classification of mental and behavioral disorders: Clinical descriptions and diagnostic guidelines 1992 Author Geneva, Switzerland World Health Organization The World Health Report 2001: Mental health – new understanding, new hope 2001 Author Geneva, Switzerland World Health Organization Clinical management of acute pesticide intoxication: Prevention of suicidal behaviours 2008a Author Geneva, Switzerland World Health Organization Preventing suicide: A Resource for media professionals 2008b Author Geneva, Switzerland World Health Organization Mental Health Atlas 2011: Mozambique 2011 Author Geneva, Switzerland World Health Organization Preventing suicide – A global imperative 2014 Author Geneva, Switzerland Yoshimasu K Kiyohara C Miyashita K Suicidal risk factors and completed suicide: Meta-analyses based on psychological autopsy studies Environmental Health and Preventive Medicine 2008 13 5 243 256 doi:10.1007/s12199-008-0037-x 19568911 Zhang J Ma Z Patterns of life events preceding the suicide in rural young Chinese: A case control study Journal of Affective Disorders 2012 140 2 161 167 doi:10.1016/j.jad.2012.01.010.Patterns 22595373

PMC005xxxxxx/PMC5133186.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 101318567 35738 J Biophotonics J Biophotonics Journal of biophotonics 1864-063X 1864-0648 27243910 5133186 10.1002/jbio.201600035 NIHMS820475 Article Low-level laser therapy stimulates the oxidative burst in human neutrophils and increases their fungicidal capacity Cerdeira Cláudio Daniel 1 Brigagão Maísa Ribeiro Pereira Lima 1 de Carli Marina Lara 2 de Souza Ferreira Cláudia 1 de Oliveira Isac Moraes Gabriel 1 Hadad Henrique 2 Hanemann João Adolfo Costa 2 Hamblin Michael R. 345 Sperandio Felipe Fornias *346 1 Department of Biochemistry, Institute of Biomedical Sciences, Federal University of Alfenas (UNIFAL-MG), Alfenas, MG 37130-000, Brazil 2 Department of Clinics and Surgery, School of Dentistry, Federal University of Alfenas (UNIFAL-MG), Alfenas, MG 37130-000, Brazil 3 Wellman Center for Photomedicine, Massachusetts General Hospital, Boston, MA 02114, USA 4 Department of Dermatology, Harvard Medical School, Boston, MA 02115, USA 5 Harvard-MIT Division of Health Sciences and Technology, Cambridge, MA 02139, USA 6 Department of Pathology and Parasitology, Institute of Biomedical Sciences, Federal University of Alfenas (UNIFAL-MG), Alfenas, MG 37130-000, Brazil * Corresponding author: ff.sperandio@gmail.com, felipe.fornias@unifal-mg.edu.br, Phone: 55 35 3299 1301, Fax: 55 35 3299 1063 2 10 2016 31 5 2016 12 2016 01 12 2017 9 11-12 11801188 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. Low-level laser therapy (LLLT) is known to enhance mitochondrial electron transfer and ATP production; thus, this study asked whether LLLT could stimulate the oxidative burst in human neutrophils (PMN) and improve their ability to kill microorganisms. Blood from healthy human subjects was collected and PMN were isolated from the samples. PMN were treated in vitro with 660 nm or 780 nm CW laser light at 40 mW power and increasing energies up to 19.2 J and were subsequently incubated with Candida albicans cells. Generation of hydroxyl radicals, hypochlorite anions and superoxide anions by PMN were checked using fluorescent probes and chemiluminescence assays; a microbicidal activity assay against C. albicans was also performed. LLLT excited PMN to a higher functional profile, which was translated as superior production of reactive oxygen species (ROS) and increased fungicidal capacity. The most efficacious energy was 19.2 J and, interestingly, the 660 nm light was even more efficacious than 780 nm at increasing the respiratory burst of PMN and the fungicidal capacity. Human neutrophils (PMN) were stimulated in vitro with 660 nm or 780 nm CW laser light at 40 mW of power and a total energy of 19.2 J. Low-level laser therapy (LLLT) excited PMN to a higher functional profile, which was translated as a superior production of reactive oxygen species (ROS) such as hydroxyl radicals (HO•) and hypochlorite anions (ClO−) (Figure) and increased fungicidal capacity against Candida albicans cells. Graphical Abstract low-level laser therapy photobiomodulation neutrophils reactive oxygen species oxidative burst Candida albicans 1. Introduction Low-level laser therapy (LLLT) or photobiomodulation uses light to stimulate cells and tissues via non-ionizing and non-thermal red or near infrared (NIR) wavelengths [1–3] and has been used to treat several medical conditions, including those typified by pain or inflammation [1, 4, 5]. However, there is uncertainty in the literature about whether LLLT has mainly an anti-inflammatory effect [6], or whether it can be regarded as a pro-inflammatory stimulus [7, 8]. Authors have used different models to study the pro-inflammatory and anti-inflammatory responses after LLLT. In order to recruit immune cells to a given anatomical site, some studies have employed chemical irritants such as formaldehyde to induce lung inflammation that mimics inflammation caused by pollutants [5], LPS to simulate mastitis in rats [9], or to cause acute respiratory distress syndrome (ARDS) in mice [10], or have even used live fungal cells to test the fungicidal response of neutrophils [11]. Many reports focus on polymorphonuclear (PMN) cells or neutrophils, as they possess specific receptors that allow them to recognize fungi and bacteria [12, 13] and are one of the principal mediators of the innate immune response against pathogens [14–16]. The degranulation process and the respiratory burst that are typical of PMN involve endogenous peroxidases and NADPH-oxidase, and these processes are essential for PMN to fight fungi [17]. We have previously studied the effects of LLLT upon the behavior of PMN acting against an important fungal infection (Paracoccidioides brasiliensis), and observed that LLLT induced lesser PMN migration to the infection site; however, the PMN that did reach the site of infection were significantly more activated towards the fungal cells and could efficiently kill the microorganism due to higher PMN viability, and greater production of ROS as well as pro-inflammatory proteins [11]. Therefore, in order to better investigate the activation of the respiratory burst of PMN after light irradiation we extracted and isolated these cells from the blood of healthy human subjects and subjected them to LLLT with different in vitro protocols, then assessed the production of intracellular and extracellular ROS and their fungicidal capacity against C. albicans cells. 2. Materials and methods All experiments were performed in compliance with the relevant laws and institutional guidelines in accordance with the ethical standards of the Declaration of Helsinki (http://www.wma.net/en/30publications/10policies/b3/index.html). The present study was performed after approval of Human Research Ethics Committee of our institution (Protocol #292274), and experimentation with human subjects was only possible after obtaining an informed consent. 2.1 Human neutrophil isolation Human PMN were obtained from eight (8) healthy male donors (aged 20–40 years) who had not received any medication for at least 7 days prior to blood collection and who signed an informed consent. The cells were purified from heparin blood samples diluted in phosphate-buffered saline solution (PBS) by sequential Ficoll–Hypaque differential density centrifugation (5810R Centrifuge, Eppendorf, NY, USA) according to Boyum [18]. The degree of purity was greater than 95% and was verified by light microscopy, after Leishman staining, in order to confirm the neutrophil morphology [19]. PMN in the pellet were resuspended in 2 mL of PBS. The cell viability was evaluated with 0.2% Trypan blue (Sigma-Aldrich, St. Louis, MO, USA) and found to be >92%. PMN were quantified using a hemocytometer, and the final concentration was adjusted to 106 PMN/mL [19, 20]. The extracted and isolated PMN from individuals were gathered to compose pools of cells that were further distributed into distinct experimental groups. 2.2 Low-level laser irradiation of PMN Cells were irradiated using a CW semiconductor diode laser (Twin Flex; MMO, São Carlos, SP, Brazil) with a spot size of 0.04 cm2, at 660 nm or 780 nm wavelengths with 40 mW total energy, irradiance of 1 W/cm2 per irradiation, fluence of 6 J/cm2 per irradiation, and giving 0.24 J of energy per irradiation period (6 seconds per irradiation). Different numbers of irradiations were used up to a maximum of 80 pulses (480 sec total) equivalent to 19.2 J total energy. The laser tip always touched the bottom of the well-plate and the laser beam remained perpendicular to the surface of the plate at all times. PMN viability was assessed with an MTT assay (Sigma-Aldrich) and the minimum viability was 98 ± 2% for non-irradiated and 97 ± 3% for irradiated groups. There were no statistically significant differences among the groups. Irradiated and non-irradiated PMN (1 × 106 cells/mL) were added into a 96-well plate. After 3 hours of incubation (5% CO2 and 37 °C) 10 μL of MTT (5 mg/mL) was added to the wells and the plate was further incubated for 4 hours. Subsequently, 10 μL of DMSO (Sigma-Aldrich) was added and the plate was read in a microplate reader at 570 nm (Anthos Zenyth 200, Biochrom, Cambridge, UK). Triton-X-100 (10%) was used as positive control. 2.3 Measurement of hydroxyl radical and hypochlorite anion intra- and extracelullar Intracellular and extracellular hydroxyl radical (HO•) and hypochlorite anion (ClO−) were measured with the fluorescent probes, hydroxyphenyl fluorescein (HPF; H36004) and aminophenyl fluorescein (APF; A36003) (Molecular Probes, Life Technologies, New York, NY, USA). PMN suspension (1 × 106 cells/mL) was incubated with 10 mM glucose, 1 mM CaCl2, 1.5 mM MgCl2 in a 96-well plate. Laser groups were irradiated with 660 nm or 780 nm wavelengths until fluorescence emission reached a plateau (maximum of 80 irradiation periods – total energy of 19.2 J delivered for 480 seconds). APF or HPF were dissolved in 25 μM DMSO and were added to the wells. The same concentration of DMSO was also added to the remaining experimental groups. The fluorescence emission was measured (Varian Cary Eclipse Fluorescence Spectrophotometer, Varian, Mulgrave, Australia) using excitation/emission wavelengths of 500/520 nm [21]. 2.4 Measurement of superoxide anion production by cytochrome c reduction Superoxide dismutase (SOD)-inhibitable cytochrome c reduction assay was used to measure the superoxide anion ( O2•-) production by PMN. The experiment was performed in triplicate. PMN suspension (5 × 105 cells/mL) in PBS was added into a 96-well plate with 50 U/ml SOD for 10 minutes, followed by addition of 10 mM glucose, 1 mM CaCl2, 1.5 mM MgCl2, and 25 μM cytochrome c. In the control group, PMN were stimulated with non-opsonized C. albicans at a m.o.i of 10 : 1 [22], or 40 μM N-formyl-methionyl-leucyl-phenylalanine (fMLP). In the laser groups, PMN were previously irradiated with laser 660 nm or 780 nm wavelengths for 80 irradiations (total energy of 19.2 J delivered over 480 seconds), following stimulation with non-opsonized C. albicans at a m.o.i of 10 : 1, or 40 μM fMLP. The 96-well plate was incubated at 37 °C for 3 hours. The reaction was monitored by spectrophotometry at a wavelength of 550 nm. The results were expressed in nmol of O2•- per 5 × 105 cells per minute, as follows [23, 24]: ΔOD550=OD550(sample)-OD550(reference[SOD])ΔOD550X12.64=nmolO2•-/5×105cells/min 2.5 Chemiluminescence The quantification of total and extracellular ROS produced by the human PMN oxidative “burst” was carried out using the luminol or isoluminol chemiluminescence assays, respectively. While luminol is thought to cross cell membranes, hence measuring the general chemiluminescence [25, 26], isoluminol possess the amino group located distant from the oxygen group [27], thus making this molecule not trespass membranes [28, 29]; in that way, isoluminol only detects the extracellular chemiluminescence [26]. PMN suspension (1 × 106 cells/mL) was incubated with 10 mM glucose, 1 mM CaCl2, 1.5 mM MgCl2, 50 μM luminol or 50 μM isoluminol, and 8 U/mL horseradish peroxidase (HRP); then, in the control group, PMN were stimulated with non-opsonized C. albicans at a m.o.i of 10 : 1, or 40 μM fMLP. In the laser groups, PMN were previously irradiated with 660 nm or 780 nm wavelengths at 80 irradiation periods (total energy of 19.2 J delivered for 480 seconds), following stimulation with non-opsonized C. albicans at an m.o.i of 10 : 1. A luminometer (Glomax 20/20 Luminometer, Promega, USA) was used to measure the chemiluminescence signal over 30 minutes. The results were expressed as relative light units (RLU) [23]. Negative control (32 μM DPI, Sigma-Aldrich) was performed. The experiment was performed in triplicate. 2.6 Yeast preparation C. albicans yeast (ATCC 10231 – American Type Culture Collection, VA, USA) was cultived in brain heart infusion (BHI) medium overnight at 37 °C. Fungal cells (10 mL) were centrifuged (5810R Centrifuge, Eppendorf, NY, USA) at 2200 g for 8 minutes, and the pellet was re-suspended in PBS. The fungal suspension was adjusted to 3 × 107 colony-forming units (CFU)/mL using a spectrophotometer (1 OD600 = 3 × 107 CFU/mL). C. albicans was inactivated by heat for 30 minutes at 100 °C for the oxidative burst experiments, whereas for the microbicidal activity assay we used fully viable microorganisms. C. albicans was opsonized by 20% human serum incubated at 37 °C for 1 hour followed by centrifugation at 1000 g for 10 minutes. The pellet was re-suspended in PBS. 2.7 Microbicidal activity assay The microbicidal activity assay was performed according to Green et al. [22]. Opsonized C. albicans suspension (1 × 107 CFU/mL) was incubated with PMN (106 cells/mL), 10 mM glucose, 1 mM CaCl2, and 1.5 mM MgCl2 at 37 °C over 90 minutes. In the laser groups, PMN were previously irradiated with laser 660 nm or 780 nm wavelengths for 80 irradiation periods (total energy of 19.2 J delivered for 480 seconds). After different incubation times (5, 30 and 90 minutes) all experimental groups were treated with cold PBS and were diluted in H2O (pH = 11) at 3 different concentrations: 103, 106 and 109 cells/mL. The suspensions were dispensed on Tryptone Soy Agar (TSA) into Petri dishes, and they were incubated at 37 °C for 24 hours. Viable C. albicans were measured from the number of colonies grown on Petri dishes and was expressed in CFU/mL. Negative control (32 μM DPI, Sigma-Aldrich) was performed. 2.8 Statistical analysis Experimental groups were compared using one-way analysis of variance (ANOVA) followed by Tukey test for multiple comparisons. The analyses adopted a significance level of 5% (p < 0.05). 3. Results 3.1 HO• and ClO− generation by PMN HO• and ClO− were produced in significantly higher quantities by irradiated PMN than they were by control PMN (p < 0.01). Both red and infrared wavelengths stimulated ROS production and the most efficacious energy dose for 660 nm light was 10.08 J (p < 0.01), while for 780 nm it was 19.2 J (p < 0.01) (Figure 1A depicts red LLLT and 1B depicts infrared LLLT). 3.2 O2•- generation by PMN stimulated with LLLT, fMLP or C. albicans O2•- was highly generated by both 660 nm or 780 nm irradiated PMN (p < 0.01) compared to control PMN; no significant differences were found between the irradiated cells and the cells stimulated with fMLP (Figure 2). In addition, the increased levels of O2•- by irradiated PMN were similar to the O2•- levels of cells that were stimulated with C. albicans (Figure 2). Finally, SOD was able to significantly quench the O2•- levels generated by irradiated PMN. This was done as a control to confirm extracellular generated superoxide (Figure 2). 3.3 Intracellular and extracellular production of total oxidants by PMN The total amount of ROS produced by the 660 nm-irradiated PMN (detected with luminol) was clearly higher than that of control or 780 nm-irradiated PMN, and the oxidants produced after LLLT were efficaciously inhibited by the use of a specific inhibitor of NADPH oxidase called diphenyleneiodonium (DPI) (Figure 3). Similarly, we only found significantly higher levels of extracellular ROS (detected by isoluminol) with red LLLT, and DPI neutralized the greater part of these oxidants (Figure 4). It is worth mentioning that the 660 nm irradiation was more effective than fMLP or C. albicans stimulation for both intra and extracellular production of oxidants (Figures 3 and 4). 3.4 Fungicidal capacity of irradiated PMN We observed that the ability to kill C. albicans was enhanced after red light stimulation. This was demonstrated by a diminished number of fungal colonies after 80 minutes of co-culture; however, 40 minutes of fungal co-culture did not lead to any difference in the results between the control and irradiated groups (Figure 5). 4. Discussion In this study we detected elevated HO• and ClO− formation upon the illumination of human PMN; such ROS generation is considered to be an important mechanism for PMN to kill microorganisms [30–32]. ROS generation by PMN has even been implicated in the killing of neoplastic cells [32–34] and could be involved in the immune response against tumors. Immune cells play a crucial role in virtually all types of infections, and need to be effective against a wide range of bacteria, viruses and fungi. Cellular and humoral immune responses essentially work jointly to eradicate microbes such as C. albicans, which is a common opportunistic pathogen that can lead to a form of invasive fungal disease in immunocompromised individuals [35]. Candida ranks as one of the most important nosocomial fungal infections, with the potential to be lethal if it reaches the bloodstream [36, 37]. PMN are well known to be the principal participants in the innate immune response that develops after Candida infections [14–16]. These leukocytes possess receptors that allow them to recognize fungi and other microbes, namely the beta-2 subfamily of integrins also known as CD18 [12, 13]. Thus, we asked whether LLLT could cause PMN to fight C. albicans more effectively by stimulating their inner production and discharge of ROS such as hydroxyl radicals. In fact, hydroxyl radicals are effective against Gram-negative bacteria in third-degree burn infections [31] and can also kill neoplastic cells [33]. We found that not only were HO• and ClO− produced in higher quantities after LLLT, but superoxide ( O2•-) was also increased to a level equivalent to the fMLP treated positive-control. The chemotactic peptide fMLP induces the respiratory burst in PMN by increasing the phosphorylation of various tyrosine domains of proteins [38] and increasing NADPH oxidase activity [39]. Thus, increasing the production of extracellular superoxide with LLLT to the same extent as fMLP is remarkable, and indicates that 19.2 J (80 irradiations of 6 seconds each) represented an effective stimulatory energy dose. Moreover, stimulating PMN with C. albicans led to the same superoxide generation as LLLT. This again means that PMN were effectively stimulated to produce ROS to very high levels. The assay of the amount of oxidants using luminol and isoluminol showed that light stimulation increased generation of both intracellular and extracellular ROS; the concentrations of oxidants were raised by light by as much or even more than they were with other stimuli (fMLP and Candida). In that way, LLLT probably elevated the oxidative burst to the highest extent possible, since fMLP did not increase ROS production higher than light irradiation. Additionally, we have reasons to believe that the mechanism of stimulation of ROS generation in PMN by light is NADPH oxidase dependant, since the NADPH oxidase inhibitor DPI efficaciously reduced the amount of oxidants to their basal levels. DPI prevents flavin adenine dinucleotide from binding to NADPH oxidase [40], and therefore impairs the ROS production of PMN even when treated with fMLP [39]. There is a dichotomy that underlies the role of LLLT on inflammation: sometimes, LLLT has pro-inflammatory effects [7, 8], while at other times LLLT has mainly anti-inflammatory effects [6]. The anti-inflammatory effects are usually manifested as a lesser influx of PMN into a given site of inflammation; for instance, 650 nm LLLT diminished the number of LPS-induced PMN entering the breast alveoli in a rat model of mastitis [9], while a 830 nm LLLT protocol was able to significantly reduce the total number of cells and the number of PMN in a model in BALB/c mice in which LPS was used to induce a disease that mimicked acute respiratory distress syndrome (ARDS) [10]. In addition, three out of four doses of 660 nm LLLT significantly reduced myeloperoxidase (MPO) activity in PMN in an ARDS model in mice [41]. Nevertheless, we believe that although LLLT can reduce the influx of PMN into a given site of inflammation, somewhat paradoxically, LLLT may also be able to increase the individual activity levels of those PMN. In a recent study we were able to demonstrate that even though irradiated PMN were present in lower numbers, they were far more active against virulent forms of Paracoccidioides brasiliensis (Pb), which is the fungus responsible for a neglected tropical disease known as paracoccidioidomycosis (PCM) [11]. The highest fungicidal capacity was activated in the PMN by red light compared to NIR light. We do not understand exactly why the shorter wavelength was more effective, but the literature suggests that red light is indeed efficacious at accelerating the electron transport in the mitochondria leading to higher ATP production since red light is absorbed by the recognized photoacceptor, cytochrome C oxidase [42–45]. It is known that the respiratory burst in PMN is driven by ATP as an energy source. Repeated 632.8 nm laser treatment to the thymus projection areas and to the hind limbs of mice led to a rise in production of cytokines, NO and Hsp70 by immune cells [46]. Moreover, even blue light from a light emitting diode may increase ROS generation in a dose-dependent manner and without altering the viability of fibroblasts [47]; blue laser light may also result in greater production of ROS in fibroblasts, human cervix adenocarcinoma cells, macrophages, and several other types of cells [48]. Inducing the production of ROS by PMN makes them better at killing C. albicans, and considering that some Candida strains may present resistance to commonly used antimycotics [49, 50], while the ROS burst is not target-specific and so far no fungal resistance has been found against PMN killing. 5. Conclusion In summary, the paradox of LLLT being a potent anti-inflammatory or pro-inflammatory tool, as well as the pros and cons of shining light on an infected tissue have to be considered before recommending laser-therapy to any clinical application; here we show how specific and safe low-level light parameters can boost the innate immune response towards an opportunistic fungus, what may possibly help treating clinical fungal infections in the near future. The authors are grateful to Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES AUX PE PNPD 2386/2011) and Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq –INCT of Redox Processes in Biomedicine – Redoxoma) for financially supporting this study. MR Hamblin was supported by US NIH grant R01AI050875. Figure 1 ROS generation is significantly increased with LLLT. Fluorescent probes APF and HPF helped detecting the ROS production after increasing doses of (A) red and (B) infrared LLLT. Results represented by means ± standard deviations. Asterisks indicate a p < 0.01 from that point on and in comparison to any other group. All groups contained the same amount of diluted DMSO. Figure 2 LLLT shifts up the O2•- generation in PMN. O2•- generation determined by SOD-inhibitable reduction of cytochrome c for Control and Laser groups. Human PMN were irradiated with 660 nm or 780 nm laser light; p < 0.01 (*) between PMN irradiated with laser 660 nm and non-irradiated PMN; p < 0.01 (*) between PMN irradiated with laser 780 nm and non-irradiated PMN; p < 0.01 (*) between PMN irradiated with 660 nm and PMN irradiated with 660 nm along with SOD; p < 0.01 (*) between PMN irradiated with 780 nm and PMN irradiated with 780 nm and SOD. Irradiated and non-irradiated PMN stimulated by fMLP showed no statistically differences between them (p > 0.05). Irradiated and non-irradiated PMN stimulated by C. albicans showed no statistically differences between them (p > 0.05). Results represent means ± standard deviations. p value obtained with ANOVA followed by Tukey test (level of significance of 5%). fMLP, N-Formyl-methionyl-leucyl-phenylalanine; SOD, superoxide dismutase. Figure 3 LLLT enhances PMN luminol chemiluminescence. p < 0.01 (*) between non-irradiated and irradiated (660 nm) PMN; p < 0.01 (*) between non-irradiated and irradiated (660 nm) PMN stimulated with C. albicans; p < 0.01 (*) between irradiated (660 nm) PMN and irradiated (660 nm) PMN with DPI; p < 0.01 (*) between irradiated (780 nm) PMN and irradiated (780 nm) PMN with DPI. Results represented by means ± standard deviations. RLU, relative light units; fMLP, N-Formyl-methionyl-leucyl-phenylalanine; DPI, diphenyleneiodonium. Figure 4 LLLT enhances PMN isoluminol chemiluminescence. p < 0.01 (*) between non-irradiated and irradiated (660 nm) PMN; p < 0.01 (*) between non-irradiated and irradiated (660 nm) PMN stimulated with C. albicans; p < 0.05 (**) between irradiated (660 nm) PMN and irradiated (660 nm) PMN with DPI; p < 0.01 (*) between irradiated (780 nm) PMN and irradiated (780 nm) PMN with DPI. Results represented by means ± standard deviations. RLU, relative light units; fMLP, N-Formyl-methionyl-leucyl-phenylalanine; DPI, diphenyleneiodonium. Figure 5 Fungicidal capacity of PMN is improved with LLLT. Colony-forming units (CFU) of C. albicans at 40 (A) Control; (B) PMN + C. albicans; (C) PMN + C. albicans + laser 660 nm; (D) PMN + C. albicans + laser 780 nm) and 80 minutes (E) Control; F – PMN + C. albicans; (G) PMN + C. albicans + laser 660 nm; (H) PMN + C. albicans + laser 780 nm) evaluation for non-irradiated and irradiated PMN, being 0: no dilution; 1: 103 cells/mL; 2: 106 cells/mL and 3: 109 cells/mL; and (I) Quantification of CFU. p < 0.05 (*) between non-irradiated and irradiated (laser 660 nm) PMN after 80 minutes of fungal growth; p < 0.05 (*) between C. albicans and irradiated (laser 660 nm) PMN after 80 minutes of fungal growth; p < 0.05 (*) between C. albicans and irradiated (laser 780 nm) PMN after 80 minutes of fungal growth; p < 0.05 (*) between non-irradiated PMN and non-irradiated PMN with DPI after 40 minutes and 80 minutes of fungal growth; p < 0.05 (*) between irradiated (laser 660 nm) PMN and irradiated (laser 660 nm) PMN with DPI after 40 minutes; p < 0.01 (**) between irradiated (laser 660 nm) PMN and irradiated (laser 660 nm) PMN with DPI after 80 minutes; p < 0.01 (**) between irradiated (laser 780 nm) PMN and irradiated (laser 780 nm) PMN with DPI after 40 minutes and 80 minutes of fungal growth. Results represented by means ± standard deviations. DPI, diphenyleneiodonium. Conflict of interest All authors of the present manuscript declare no conflict of interest. Author biographies Please see Supporting Information online. 1 Chung H Dai T Sharma SK Huang YY Carroll JD Hamblin MR Ann Biomed Eng 40 516 533 2012 22045511 2 Sperandio FF Simoes A Correa L Aranha AC Giudice FS Hamblin MR Sousa SC J Biophotonics 8 795 803 2015 25411997 3 Sperandio FF Giudice FS Correa L Pinto DS Jr Hamblin MR de Sousa SC J Biophotonics 6 839 847 2013 23554211 4 Sperandio FF Simoes A Aranha AC Correa L Orsini Machado de Sousa SC Photomed Laser Surg 28 581 587 2010 20961226 5 Miranda da Silva C Peres Leal M Brochetti RA Braga T Vitoretti LB Saraiva Camara NO Damazo AS Ligeiro-de-Oliveira AP Chavantes MC Lino-Dos-Santos-Franco A PLoS One 10 e0142816 2015 26569396 6 Lim W Choi H Kim J Kim S Jeon S Zheng H Kim D Ko Y Kim D Sohn H Kim O J Oral Pathol Med 44 94 102 2015 25066944 7 Viegas VN Abreu ME Viezzer C Machado DC Filho MS Silva DN Pagnoncelli RM Photomed Laser Surg 25 467 473 2007 18158747 8 Woodruff LD Bounkeo JM Brannon WM Dawes KS Barham CD Waddell DL Enwemeka CS Photomed Laser Surg 22 241 247 2004 15315732 9 Wang Y He X Hao D Yu D Liang J Qu Y Sun D Yang B Yang K Wu R Wang J J Vet Med Sci 76 1443 1450 2014 25452258 10 Oliveira MC Jr Greiffo FR Rigonato-Oliveira NC Custodio RW Silva VR Damaceno-Rodrigues NR Almeida FM Albertini R Lopes-Martins RA de Oliveira LV de Carvalho PT Ligeiro de Oliveira AP Leal EC Jr Vieira RP J Photochem Photobiol B 134 57 63 2014 24792475 11 Burger E Mendes AC Bani GM Brigagao MR Santos GB Malaquias LC Chavasco JK Verinaud LM de Camargo ZP Hamblin MR Sperandio FF PLoS Negl Trop Dis 9 e0003541 2015 25675431 12 Mayadas TN Cullere X Trends Immunol 26 388 395 2005 15922663 13 McFarland HI Nahill SR Maciaszek JW Welsh RM J Immunol 149 1326 1333 1992 1500720 14 Netea MG Gijzen K Coolen N Verschueren I Figdor C Van der Meer JW Torensma R Kullberg BJ Microbes Infect 6 985 989 2004 15345229 15 Mahanty S Greenfield RA Joyce WA Kincade PW Infect Immun 56 3162 3166 1988 3182076 16 Diamond RD Arch Med Res 24 361 369 1993 8118160 17 Dias MF Mesquita J Filgueira AL De Souza W Med Mycol 46 241 249 2008 18404552 18 Boyum A Scand J Clin Lab Invest Suppl 97 7 1968 5707208 19 Maqbool M Vidyadaran S George E Ramasamy R Med J Malaysia 66 296 299 2011 22299545 20 Nauseef WM Methods Mol Biol 412 15 20 2007 18453102 21 Setsukinai K Urano Y Kakinuma K Majima HJ Nagano T J Biol Chem 278 3170 3175 2003 12419811 22 Green JN Winterbourn CC Hampton MB Methods Mol Biol 412 319 332 2007 18453121 23 Dahlgren C Karlsson A Bylund J Methods Mol Biol 412 349 363 2007 18453123 24 Cohen HJ Chovaniec ME J Clin Invest 61 1081 1087 1978 26695 25 Dahlgren C Stendahl O Infect Immun 39 736 741 1983 6299947 26 Rajecky M Lojek A Ciz M Int J Lab Hematol 34 136 142 2012 21834798 27 Jancinova V Drabikova K Nosal R Rackova L Majekova M Holomanova D Redox Rep 11 110 116 2006 16805965 28 Lundqvist H Dahlgren C Free Radic Biol Med 20 785 792 1996 8728025 29 Lundqvist H Kricka LJ Stott RA Thorpe GH Dahlgren C J Biolumin Chemilumin 10 353 359 1995 8588512 30 Sperandio FF Huang YY Hamblin MR Recent Pat Antiinfect Drug Discov 8 108 120 2013 23550545 31 Huang L Wang M Dai T Sperandio FF Huang YY Xuan Y Chiang LY Hamblin MR Nanomedicine (Lond) 9 253 266 2014 23738632 32 Wang M Huang L Sharma SK Jeon S Thota S Sperandio FF Nayka S Chang J Hamblin MR Chiang LY J Med Chem 55 4274 4285 2012 22512669 33 Sperandio FF Sharma SK Wang M Jeon S Huang YY Dai T Nayka S de Sousa SC Chiang LY Hamblin MR Nanomedicine 9 570 579 2013 23117043 34 Huang L Wang M Sharma SK Sperandio FF Maragani S Nayka S Chang J Hamblin MR Chiang LY ECS Trans 45 1 11 10.1149/04520.0065ecst 2013 35 Soloviev DA Jawhara S Fonzi WA Infect Immun 79 1546 1558 2011 21245270 36 Abelson JA Moore T Bruckner D Deville J Nielsen K Pediatrics 116 61 67 2005 15995032 37 Snydman DR Chest 123 500S 503S 2003 12740235 38 Torres M Hall FL O’Neill K J Immunol 150 1563 1577 1993 7679431 39 Hidalgo MA Carretta MD Teuber SE Zarate C Carcamo L Concha II Burgos RA J Immunol Res 2015 120348 2015 26634216 40 O’Donnell BV Tew DG Jones OT England PJ Biochem J 290 1 41 49 1993 8439298 41 de Lima FM Aimbire F Miranda H Vieira RP de Oliveira AP Albertini R J Lasers Med Sci 5 63 70 2014 25653801 42 Passarella S Ostuni A Atlante A Quagliariello E Biochem Biophys Res Commun 156 978 986 1988 3190685 43 Pastore D Di Martino C Bosco G Passarella S Biochem Mol Biol Int 39 149 157 1996 8799337 44 Pastore D Greco M Passarella S Int J Radiat Biol 76 863 870 2000 10902741 45 Eells JT Henry MM Summerfelt P Wong-Riley MT Buchmann EV Kane M Whelan NT Whelan HT Proc Natl Acad Sci USA 100 3439 3444 2003 12626762 46 Novoselova EG Glushkova OV Cherenkov DA Chudnovsky VM Fesenko EE Photodermatol Photoimmunol Photomed 22 33 38 2006 16436179 47 Mamalis A Garcha M Jagdeo J Lasers Surg Med 47 210 215 2015 25655579 48 Kushibiki T Hirasawa T Okawa S Ishihara M Photomed Laser Surg 31 95 104 2013 23390956 49 Pfaller MA Boyken L Hollis RJ Messer SA Tendolkar S Diekema DJ J Clin Microbiol 43 5425 5427 2005 16272464 50 Tavanti A Davidson AD Fordyce MJ Gow NA Maiden MC Odds FC J Clin Microbiol 43 5601 5613 2005 16272493

PMC005xxxxxx/PMC5133190.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 0401650 3402 Differentiation Differentiation Differentiation; research in biological diversity 0301-4681 1432-0436 27262401 5133190 10.1016/j.diff.2016.05.006 NIHMS792166 Article Sox9 overexpression in uterine epithelia induces endometrial gland hyperplasia Gonzalez Gabriel a Mehra Shyamin a Wang Ying a Akiyama Haruhiko b Behringer Richard R. a a Department of Genetics, University of Texas M.D. Anderson Cancer Center, Houston, TX 77030, USA b Department of Orthopedic Surgery, Gifu University, Gifu City 501-1194, Japan 1 7 2016 01 6 2016 Oct-Nov 2016 01 10 2017 92 4 204215 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. SOX9 is a high mobility group transcription factor that is required in many biological processes, including cartilage differentiation, endoderm progenitor maintenance, hair differentiation, and testis determination. SOX9 has also been linked to colorectal, prostate, and lung cancer. We found that SOX9 is expressed in the epithelium of the adult mouse and human uterus, predominantly marking the uterine glands. To determine if SOX9 plays a role in the development of endometrial cancer we overexpressed Sox9 in the uterine epithelium using a progesterone receptor-Cre mouse model. Sox9 overexpression in the uterine epithelium led to the formation of simple and complex cystic glandular structures in the endometrium of aged-females. Histological analysis revealed that these structures appeared morphologically similar to structures present in patients with endometrial hyperplastic lesions and endometrial polyps that are thought to be precursors of endometrial cancer. The molecular mechanisms that cause the glandular epithelium to become hyperplastic, leading to endometrial cancer are still poorly understood. These findings indicate that chronic overexpression of Sox9 in the uterine epithelium can induce the development of endometrial hyperplastic lesions. Thus, SOX9 expression may be a factor in the formation of endometrial cancer. Sox9 transgenic mouse uterus endometrium cancer Introduction In the United States, it is estimated that there were 54,870 women newly diagnosed with endometrial cancer last year, and this malignancy will prove to be fatal for 10,170 (National Cancer Institute). Uterine adenocarcinoma is the most prevalent type of endometrial cancer arising from uncontrolled growth of the glandular epithelium (GE). Adenocarcinoma development occurs in stages (Silverberg, 2000). The first stage is widely thought to be the manifestation of uterine lesions known as endometrial hyperplasia, sub-classified into simple or complex. In simple hyperplasia, the uterine glands become dilated cysts, showing pseudostratified epithelium, while the stroma appears normal and contains small blood vessels uniformly spaced. In complex hyperplasia the uterine glands show a crowded architecture and are irregular in shape and size often containing numerous side buds, while the stroma appears normal. Various studies have indicated that the presence of cytological atypia, accompanying either simple or complex hyperplasia indicates progression to adenocarcinoma (Silverberg, 2000). In atypical hyperplasia, the uterine gland epithelium loses its normal columnar morphology and the nuclei become rounder. Upon transition to adenocarcinoma, the uterine glands grow in close proximity with minimal stroma between them. One study showed that the progression rate to adenocarcinoma is 4.3% from simple hyperplasia, 16.1% from complex hyperplasia, 7.4% from atypical simple hyperplasia, and 47% from atypical complex hyperplasia (Silverberg, 2000). A study of 7,835 women with endometrial hyperplasia revealed that the cumulative risk of developing adenocarcinoma within 20 years was 4.6% in patients with non-atypical endometrial hyperplasia or disordered proliferative endometrium (Lacey et al., 2010). In contrast, the presence of atypical hyperplasia (simple or complex) raised the cumulative risk of developing adenocarcinoma within 20 years to 27.5%. Endometrial polyposis is another type of endometrial lesion frequently diagnosed in women (Anastasiadis et al., 2000). Endometrial polyps are described as benign growths that develop in the endometrium. The size of the polyps ranges from a few millimeters up to several centimeters. 20-25% of menopausal and postmenopausal women will develop endometrial polyps (Humphrey et al., 2008). Although mostly benign, endometrial polyps are thought to transform into endometrial cancer (Savelli, 2003). Histological analysis of endometrial polyps usually shows hyperplastic lesions, however, unlike in simple hyperplasia, polyps also display fibrotic stroma and thick-walled blood vessels (Silverberg, 2000). Early discovery and removal is the key factor for treating endometrial polyps before they transform into endometrial cancer. The molecular events that occur during each stage of endometrial cancer progression are not fully understood. It is thought that the primary driving force that gives rise to the initial hyperplastic lesions is altered estrogen signaling (Bokhman, 1983; Sherman, 2000). Unopposed estrogen potentially exerts a mitogenic signal that can lead to hyperplasia. Later, mutations in KRAS and PTEN are thought to drive the clonal expansion of malignant cells that give rise to adenocarcinomas (Sherman, 2000). Subsequently, it is thought that the malignant cells inactivate the p53 pathway, which regulates cell cycle checkpoints. An understanding of the molecular mechanisms that initiate hyperplasia and subsequently the transformation from benign lesions to cancer will assist with diagnosis and more effective treatment of endometrial cancers. The main focus of this study was to examine the role of the transcription factor SOX9 during development, tissue homeostasis, and disease. SOX9 is a member of the SRY (sex determination region Y)-related HMG box (SOX) family. The SOX genes are critical in multiple developmental and physiological processes (Kiefer, 2007). Members of the SOX family contain a conserved high mobility group DNA binding domain. SOX9, SOX8, and SOX10 are transcription factors in the SoxE subgroup that contain an activation domain. Thus, SOX9 primarily acts as a transcriptional activator of target genes. SoxE subgroup genes have been shown to play roles in diverse biological processes, including sex determination, chondrogenesis, and differentiation of Müller glial cells (Foster et al., 1994; Lefebvre and de Crombrugghe, 1998; Poché et al., 2008). SOX9 is expressed in the epithelium of many organs and has also been linked to colorectal cancer, prostate cancer and lung adenocarcinoma (Darido et al., 2008; Jiang et al., 2010; Wang et al., 2008). SOX9 is activated by fibroblast growth factor (FGF) signaling (Ling et al., 2011) and mutations that result in a constitutively active form of the FGFR2-IIIb have been discovered in 12% of endometrial cancers (Dutt et al., 2008). Moreover, SOX9 expression has been detected in human endometrial cell lines and endometrial tumors (Saegusa et al., 2012). Therefore, we hypothesized that overexpressing SOX9 in the epithelial compartment of the mouse uterus would lead to the development of cancer. To test this idea, we conditionally overexpressed Sox9 in the uterine epithelium and examined these mice for uterine pathologies. Material and methods Mice The Pgr-Cre (Pgrtm2(cre)Lyd) mouse strain was obtained from Dr. Franco DeMayo (Baylor College of Medicine, Houston) (Soyal et al., 2005). Sox9-Cre (Sox9tm3(cre)Crm) and Sox9 conditional overexpression transgenic mice (CAG-Sox9) were described previously (Akiyama et al., 2005; Kim et al., 2011). R26R-RG (Shioi et al., 2011) mice were obtained from Dr. Go Shioi (RIKEN Center for Developmental Biology, Kobe, Japan). All mice were maintained on a C57BL/6J × 129/SvEv mixed genetic background. Mice were genotyped using tail snips and PCR according to previous reports (Soyal et al., 2005; Akiyama et al., 2005; Kim et al., 2011; Shioi et al., 2011). Estrous cycle status was determined by microscopic examination of vaginal washes (Behringer et al., 2014). All animals were maintained in compliance with the Public Health Service Policy on Humane Care and Use of Laboratory Animals, the U.S. Department of Health and Humane Services Guide for the Care and Use of Laboratory Animals, and the United States Department of Agriculture Animal Welfare Act. All protocols were approved by the University of Texas M.D. Anderson Cancer Center Institutional Animal Care and Use Committee. Histology Female reproductive organs were dissected into ice-cold phosphate buffered saline (PBS). The specimens were then fixed in 4% paraformaldehyde (PFA) (Electron Microscopy Sciences, cat# 19208, Hatfield, PA USA) overnight on a nutator at 4°C, dehydrated and embedded in paraffin using standard procedures. The tissues were placed in metal base molds either perpendicular to the bottom for cross-sections or on the side for longitudinal sections. Cross-sections of the uterus were made starting at the most anterior end (oviduct) to the most posterior end (cervix). Hematoxylin and eosin (H&E) staining was performed using standard procedures. Masson's Trichrome staining was performed using the Accustain Trichrome Stain Kit (Sigma-Aldrich, cat# HT15-1KT, St. Louis, MO USA), following the manufacturer's protocol. Alcian blue staining was performed by placing the tissue sections in a 3% acetic acid solution for 3 minutes. The slides were transferred into 1% Alcian Blue 8GX Stain (Sigma-Aldrich), 3% acetic acid for 30 minutes at room temperature. The slides were then washed in tap water for 10 minutes and rinsed in distilled water. The tissue sections were counterstained with a 0.1% nuclear fast red solution (Sigma-Aldrich) for 5 minutes. The slides were washed in tap water for 1 minute and dehydrated in 95% ethanol. The slides were then transferred into Histo-Clear (National Diagnostics) for 2 minutes three times. Glass cover slips were mounted using Permount (Fisher Scientific). Immunostaining Female reproductive organs were dissected and fixed as described above. The next day, the tissues were washed twice for 30 minutes in ice-cold PBS. For frozen sections, the tissues were then transferred into a 15% sucrose solution until they sank to the bottom of the tube. This was repeated using a 30% sucrose solution. The sucrose was discarded and the tissues were embedded and frozen in OCT (Fisher Scientific, cat# 23-730-571, Waltham, MA, USA). Frozen blocks were sectioned using a cryostat. The thickness of the sections varied from 20 μm to 80 μm. The tissue sections were air dried for 5 minutes and then placed in ice-cold acetone for 5 minutes. For paraffin sections, the tissues were fixed and washed as described above, then processed into paraffin using standard methods and 5 μm sections were cut. Prior to performing immunostaining, the tissue sections were placed in a 55°C oven for 30 min and then deparaffinized and rehydrated. The sections were placed in a 10 mM sodium citrate solution, heated in a microwave oven for 20 minutes, and then cooled to room temperature. The sections were washed with PBS for 5 minutes, and then placed in Hydrogen Peroxide Blocking Reagent (Abcam, ab94666, Cambridge, UK) for 10 minutes. The slides were transferred into 0.1% Triton X-100 in PBS and for 10 minutes. Blocking solution consisting of 5% sheep serum was added directly on top of each tissue section and incubated for 30 minutes at room temperature. Several primary antibodies used in this study were produced in mice. Therefore, tissue sections were incubated in mouse IgG (M.O.M. ™ Kit, cat# BMK-2202, Vector Laboratories, Burlingame, CA) for 1 hour at room temperature. The primary antibodies, SOX9 (Millipore, AB5535, Billerica, MA, USA, 1:200), E-Cadherin (BD Biosciences, cat# 610181, San Jose, CA, USA, 1:200), p63 (Santa Cruz Biotechnology, sc-25268, Dallas, Texas, USA, 1:100), FOXA2 (Abcam, ab40874, Cambridge, UK, 1:200), Ki67 (Abcam, ab8191, Cambridge, UK, 1:100), Phosphohistone H3 (Millipore, 06-570, Billerica, MA, USA, 1:200), p21(Origene, TA307018, Rockville, MD, USA, 1:200) were diluted in 10% sheep serum and added onto the tissue sections and incubated overnight at 4°C. The next day, the tissue sections were washed 3 times with PBS/0.1% Tween-20 for 10 minutes. For immunofluorescent staining, the sections were then incubated with a 1:400 dilution of the secondary antibody (AlexaFluor, BD Biosciences, San Jose, CA, USA) for 2 hours at room temperature or overnight at 4°C. The tissue sections were washed 3 times with PBS/0.1% Tween-20 for 5 minutes. One drop of Vectashield mounting medium with DAPI (4′, 6-diamidino-2-phenylindole) (Vector Laboratories, cat# H-1200, Burlingame, CA) was added onto the tissue sections and mounted with a glass coverslip. For immunohistochemical staining, following overnight incubation with primary antibody solution and washing steps, the sections were incubated with 1:500 dilution of the secondary antibody conjugated with horseradish peroxidase for 2 hours at room temperature. The tissue sections were washed 3 times with PBS/0.1% Tween-20 for 5 minutes. The 3, 3′ diaminobenzidine (DAB) chromogen and DAB substrate were mixed together following the manufacturer's protocol (Abcam, ab94665, Cambridge, UK). One drop was added on top of each section and the color was allowed to develop. The slides were washed in tap water for 1 minute and dehydrated in 95% ethanol. The slides were then transferred into Histo-Clear (National Diagnostics, cat# HS-200, Atlanta, GA, USA) for 2 minutes three times. Glass cover slips were mounted using Permount (Fisher Scientific, cat# SP15-100, Waltham, MA, USA). Image capture and post-acquisition processing Confocal fluorescent images were obtained using a Nikon C2 Confocal System (Nikon Corporation, Tokyo, Japan) with 405/488/561/640 nm solid-state lasers and inverted eclipse Ti-E microscope with CFI Plan Apochromat Lambda 10x, 20x, and 60x objectives. Z-stacks were captured and a maximum intensity Z-projection was applied to the stacks. Z-stacks were processed using Imaris (Bitplane.com), which allowed us to adjust the color histogram of each color channel and background. In some cases, processed images show a pseudo-colored signal with a light-colored background. TUNEL analysis Paraffin sections were deparaffinized and rehydrated for TUNEL analysis, using the In Situ Cell Death Detection Kit, Fluorescein (Sigma-Aldrich, cat# 11684795910, St. Louis, MO USA) according to the manufacturer's instructions. Vectashield mounting medium with DAPI was added to the processed slides that were mounted with a glass coverslip. H-score analysis of Ki-67 immunostaining Semi-quantitative analysis was performed using an H-score analysis by two independent observers. The proportion (0-100) and intensity of Ki-67 nuclear immunostaining (0: no staining; 1: weak staining; 2: moderate staining, 3: strong staining) were used to calculate an H-score. A t-test was used to compare potential differences in H-scores between control and experimental samples. Results SOX9 is expressed in mouse and human uterine epithelial tissues SOX9 is expressed in diverse tissues and organs, including the human endometrium and endometrial cancer (Furuyama et al., 2011; Lefebvre and de Crombrugghe, 1998; Poché et al., 2008; Saegusa et al., 2012). To determine if SOX9 was expressed in the adult mouse uterus, we performed immunofluorescent staining for SOX9. Nuclear-localized SOX9 was detected primarily in the glandular epithelium (GE) (Fig. 1A). A few SOX9-positive cells were also found in the luminal epithelium (LE). SOX9 expression was also examined in the adult human endometrium (Fig. 1B). Similar to the mouse, nuclear-localized SOX9 was detected in the GE of the human endometrium. Nuclear-localized SOX9 was also found in the human LE. Thus, SOX9 is expressed in the GE and a subset of LE in both mice and woman, suggesting that SOX9 may have a conserved role in uterine gland biology. Uterine cells expressing Sox9 were also visualized using a transgenic fluorescent protein reporter mouse. We took advantage of a R26R-RG mouse line that carries a Cre-dependent conditional allele that can express a histone 2B fusion with monomeric Cherry fluorescent protein (H2B-mCherry or R) and an enhanced green fluorescent protein localized to cell membranes by fusion to a protein domain that confers a GPI linkage (EGFP-GPI or G) (Shioi et al., 2011; Stewart et al., 2009). Thus, cells that expressed Cre activated RG expression, resulting in a red fluorescent nucleus and green plasma membrane. The fluorescent nuclear marker provided single cell resolution. To express these fluorescent proteins in Sox9-expressing cells, we used Sox9-Cre knock-in mice that carry an allele that preserves Sox9 expression but also expresses Cre in a Sox9-specific pattern (Akiyama et al., 2005). R26R-RG mice were bred to Sox9-Cre knock-in mice and their progeny, Sox9-Cre; R26R-RG double heterozygotes, were used to generate fixed frozen sections of adult uteri that were imaged by confocal microscopy for H2B-mCherry fluorescence. The fixation protocol was compatible with H2B-mCherry activity but inactivated GFP, therefore GFP was not used in the current study. Similar to the SOX9 immunofluorescence studies described above, H2B-mCherry was detected in the uterine glands (Fig. 1C, D). In addition, some luminal epithelial cells also expressed H2B-mCherry. SOX9 has been shown to regulate cell cycle progression in a rat chondrocytic cell line (Panda et al., 2001). The uterine epithelium undergoes rapid changes in cell cycle as it undergoes waves of proliferation and apoptosis during the estrous cycle (Garry et al., 2010; Huang et al., 2012). Thus, we were curious if SOX9 expression in the uterine epithelium would vary during the estrous cycle. Uteri from nulliparous females at four stages of the estrous cycle, proestrus, estrus, metestrus, and diestrus were dissected. Longitudinal uterine sections were examined for SOX9 by immunofluorescence (Fig. 2). Similar to our initial observations, we detected nuclear-localized SOX9 predominantly in the uterine glands but also in cells and subregions of the LE. No overt differences in SOX9 expression were detected in the uterine glands throughout the estrous cycle. In contrast to the uterine glands, the number of SOX9 positive cells appeared to change in the LE during the estrous cycle. There were more SOX9 positive cells found in proestrus, estrus, and metestrus, compared to diestrus. Thus, SOX9 expression in the LE varies during the estrous cycle. In contrast, SOX9 expression in the uterine glands appears to be constant during the different stages of the estrous cycle. Generation of a mouse model to overexpress Sox9 in the uterine epithelium Higher SOX9 labeling indices were observed in human atypical hyperplasia and endometrial carcinoma compared to normal uterine tissue (Saegusa et al., 2012), correlating SOX9 expression with malignancy. To experimentally test the idea that Sox9 can act as a cancer-initiating gene, we created a new mouse model to overexpress Sox9 in the uterine epithelium. CAG-Sox9 transgenic mice carry a construct with a CAG promoter for ubiquitous expression and a loxP-flanked monomeric red fluorescent protein (RFP) gene followed by multiple polyadenylation signals to terminate transcription. This is followed by a Sox9 cDNA and an intraribosomal entry site (IRES) EGFP expression cassette. In the absence of Cre expression the CAG-Sox9 transgene ubiquitously expresses RFP. However, Cre expression will delete the mRFP polyA cassette, leading to Sox9 and EGFP expression. Pgr-cre knock-in mice were used to express Cre in the postnatal uterine epithelium (Soyal et al., 2005). Pgr is expressed in the mouse uterine epithelium starting at approximately postnatal day (P) 14. Pgr is also active in the uterine stroma and myometrium after female mice reach sexual maturity. To activate Sox9 expression throughout the uterine epithelium, CAG-Sox9 mice were bred with Pgr-Cre mice to generate Pgr-Cre/+; CAG-Sox9/+ males. Pgr-Cre/+; CAG-Sox9/+ males were then crossed with homozygous CAG-Sox9/CAG-Sox9 females to generate Pgr-Cre/+; CAG-Sox9/CAG-Sox9 (Sox9-cOE) females for analysis. +/+; CAG-Sox9/CAG-Sox9 females served as controls. Sox9-cOE females were fertile up to 6 months. We did not breed Sox9-cOE females older than 6 months old. Overexpression of SOX9 in the uterus of prepubescent females We compared the gross morphological features of uteri dissected from Sox9-cOE and control females at P21, 28 and 35 prior to full sexual maturity. In these prepubescent females, gross examination of the reproductive tracts revealed no obvious morphological differences between Sox9-cOE and controls (data not shown). In addition, no overt histological differences were observed in H&E stained tissue sections from Sox9-OE and control uteri (data not shown). We analyzed SOX9 expression in the uterine epithelium of Sox9-cOE and control females at P21, 28, and 35 by immunofluorescence staining (Fig. 3). At P21, SOX9 immunofluorescence in the LE and GE appeared comparable between Sox9-cOE controls, with SOX9 positive cells present throughout the uterine glands with some SOX9 positive in the luminal epithelium. However, at P28, a few stromal cells appear to ectopically express SOX9 (Fig. 3I, J). This observation was expected because Pgr-Cre becomes active in the stroma as female mice reach sexual maturity. Moreover, at P35, in addition to SOX9 positive cells in the uterine glands, SOX9 positive cells were found throughout the LE (Fig. 3K, L). Thus, in this model, Sox9 overexpression initiates between P28 and P35 that subsequently expands throughout the LE by P35. Abnormal uterine structures develop in Sox9-cOE mice Although SOX9 is expressed throughout the LE as early as P35 in Sox9-cOE mice, no gross morphological or histological differences were observed compared to controls. Therefore, we aged these mice to determine if chronic Sox9 expression could alter the uterus. Sox9-cOE and control uteri were examined at 1-, 2-, 4-, 8-, and 12-month-old time points (Fig. 4, Fig. S1). H&E staining of histological sections showed dilated, cystic uterine glands starting at 2-months of age in Sox9-OE mice that persisted and became more severe at later time points (Fig. 4D, Fig. S1). There were numerous severely dilated uterine gland cysts observed along the entire length of the uterus (Fig. 4D, E, F, Fig. S1). The cysts appeared to encroach upon the myometrium. In 12-month-old Sox9-cOE females, uterine glands were cystically dilated and showed crowding and cribriform appearance (Fig. S1K-N). In contrast, in 12-month-old control females, the uterine glands showed normal histological morphology (Fig. S1I, J). The cysts contained large amounts of eosin-positive material, presumably glandular secretions (Fig. S1K, M, N). In some cases, the Sox9-cOE uterine glands were microscopically dilated (Fig. 4D, F, S1G, H, K-N), comparable to what is seen in human simple endometrial hyperplasia (Silverberg, 2000). In other cases, the uterine glands were more cribriform and crowded (Fig 4J), similar to what is seen in human complex endometrial hyperplasia (Silverberg, 2000). In a few cases, we observed loss of normal columnar morphology and nuclei that appeared rounder (Fig. S1L, N) similar to cases of atypical hyperplasia. These findings suggest that overexpressing Sox9 in the mouse uterine epithelium leads to the appearance of histopathological findings similar to cases of human endometrial hyperplasia. The stroma of Sox9-OE female uteri developed fibrosis as early as 2 months of age and this persisted and appeared to become more severe at 4- and 8-months of age (Fig. 4D, J, Fig. S1). At 4 months of age, Mason's Trichrome staining in Sox9-cOE uteri revealed strong staining in the stromal region, indicating increased connective tissue/collagen fibers compared to controls (Fig. 4H, I, K, L). Thus, chronic overexpression of Sox9 appears to influence epithelial organization and stromal differentiation. In this study, Sox9-cOE and control were aged up to 1 year. Gross examination of the uteri from 1-year-old females revealed gross morphological differences between Sox9-cOE and control uteri (Fig. 5). Under the dissecting microscope, numerous cystoid structures inside the uteri of 1-year-old Sox9-cOE females were observed (Fig. 5D-F). In control females, only 1 or 2 cystic structures were found along the entire length of both uterine horns (Fig. 5A-C). In comparison with controls, the uteri of Sox9-cOE females were filled with cystic structures of variable sizes. A longitudinal incision along the length of the uterus, providing a luminal view, showed that the cystic structures were filled with a white substance (Fig. 5F). Cellular behaviors in endometrial hyperplastic lesions in Sox9-cOE mice To determine if the endometrial hyperplastic lesions of Sox9-cOE females are formed by increased cell proliferation of the GE, immunofluorescent staining for markers of cell proliferation was performed. The mouse uterine epithelium undergoes rapid changes between cell proliferation and apoptosis, depending on the stage of the estrous cycle (Garry et al., 2010; Huang et al., 2012). To minimize potential differences in cell proliferation that are typically dependent on estrous cycle stage, estrous cycle status for each female was determined prior to sacrifice. Uteri were examined from 2- to 4-month-old females. The number of cells undergoing mitosis was assessed by staining for phosphohistone H3 (PH3). No differences in the numbers of PH3 positive cells between Sox9-cOE and controls were detected (Fig. 6A, D). Tissue sections were also stained for Ki-67, a marker of cell proliferation. No major differences in Ki-67 immunostaining were observed in uterine tissue sections between estrous stage-matched Sox9-cOE and controls (Fig. 6B, E). Both groups displayed Ki-67 positive staining in the LE during estrus, whereas during diestrus, the number of Ki-67 positive cells was negligible (data not shown). In the uterine glands, low numbers of Ki-67 positive cells were detected during estrus and diestrus. Therefore, there were no overt changes in cell proliferation detected between Sox9-cOE and controls females at 2 and 4-months of age. To determine if there were changes in the number of cells undergoing apoptosis during metestrus, we performed TUNEL staining. No differences in the numbers of TUNEL positive cells between Sox9-cOE and controls were detected (Fig. 6C, F). Aged Sox9-cOE females showed greater number of uterine glands that appeared cystically dilated and showed crowding and cribriform appearance. Therefore, Ki-67 immunostaining was performed in uterine sections of control and Sox9-cOE mice 8- and 12-months of age (Fig. 7A, B). Control tissues contained few Ki-67 positive GE cells (Fig. 7C, D). Compared to controls, Sox9-cOE uteri contained areas with high numbers of Ki-67 positive GE cells (Fig. 7E) and adjacent areas with few Ki-67 positive GE cells (Fig. 7F). An H-score was used to evaluate the proportion and intensity of Ki-67 immunostaining and a t-test was used to compare potential differences between the groups (Fig. 7G). A slight difference, although not statistically different, was noticed between Sox9-cOE and controls females examined at 8- and 12-months of age (Fig. 7, data not shown). We did not observe a correlation between Ki-67 staining and morphology. Differentiation status of endometrial hyperplastic lesions in Sox9-cOE mice The differentiation status of the Sox9-cOE uterine tissues was examined by molecular marker analysis. FOXA2 is a winged helix transcription factor that is expressed in mouse uterine glands but not the LE (Jeong et al., 2010). FOXA2 was expressed in the uterine glands of 8-month-old controls and some cells of the cystically-dilated hyperplastic uterine glands of Sox9-cOE females (Fig. 8A, B). This suggests that even though the Sox9-cOE uterine glands show structural abnormalities, they retain uterine gland identity, at least for FOXA2. TP63 (tumor protein p63) is a transcription factor expressed in stratified epithelium but not in columnar epithelium (Kurita et al., 2005). Conditionally inactivating Wnt4 in uterine glands led to abnormal glandular differentiation and appearance of a p63-expressing basal epithelial layer in mouse uterine glands (Franco et al., 2011). SOX9 and WNT4 play opposing roles during gonadal differentiation (Uhlenhaut et al., 2009). Therefore, we examined the differentiation state of the uterine epithelium in Sox9-cOE mice by performing immunofluorescent staining for p63 and SOX9 (Fig. 8C-F). As a control, we included tissue sections with cervix, which possesses a stratified epithelium. The cervical epithelium stained positively for p63 (Fig. 8E, F). However, p63 immunostaining was not detected in the uterine glands or luminal epithelium of Sox9-cOE and controls at 2-, 4-, 8-months of age, and also in 1-year-old tissue sections that display epithelial hyperplasia (Fig. 8C, D). This indicates that SOX9 does not appear to influence the expression of Tp63 in the mouse GE. Discussion SOX9 marks uterine glands The uterus is an organ that undergoes coordinated cyclic waves of cell death followed by rapid tissue regeneration by proliferation and differentiation (Gray et al., 2001; Medh and Thompson, 2000). During menstruation the stratum functionalis is shed and then regenerated. It has been proposed that glandular epithelial cells within the stratum basalis form the new stratum functionalis (Spencer et al., 2005). Specifically, adult stem cells within the stratum basalis, are thought to support the rapid regeneration of the stratum functionalis (Gargett, 2007). Cre-based lineage-tracing experiments in the mouse suggest that Sox9-expressing cells can differentiate into various cell types in the intestine, liver and pancreas (Furuyama et al., 2011). This suggests that Sox9-expressing cells may function as epithelial progenitor cells in these organs. In human endometrial samples, it was shown that SOX9 was expressed in the deeper stratum basalis region, the area retained after the stratum functionalis is shed during menstruation (Valentijn et al., 2013). Cells expressing SOX9 also showed a pattern of expression similar to that of SSEA-1, a marker of pluripotent cells, and were found to express high telomerase activity and had longer telomeres (Valentijn et al., 2013). Here, we show that nuclear-localized SOX9 is expressed in the uterine epithelium in both mouse and human. SOX9 expression was predominantly found throughout the uterine glands. In sections of the human endometrium, a similar expression pattern was observed by the Human Protein Atlas (Uhlen et al., 2010). In addition, other studies have shown that SOX9 is expressed in human endometrial glands (Saegusa et al., 2012; Valentijn et al., 2013). Therefore, there is ample evidence demonstrating that SOX9 is expressed in the glandular epithelium of the uterus, in both human and mice. In our Sox9-RG female mice, the uterine glands expressed mCherry fluorescent protein, suggesting that the GE was expressing Sox9 or was derived from cells that had expressed Sox9 at one point. It was previously shown that the GE retains BrdU for longer periods than the LE, suggesting that the GE harbors uterine epithelial progenitor cells (Kaitu'u-Lino et al., 2010). We observed that SOX9 was expressed in the GE throughout the estrous cycle of mice but SOX9 expression varied in the LE, depending on estrous cycle stage. In humans, SOX9 expression varies during the menstrual cycle, becoming elevated during the proliferative phase compared to the secretory phase (Saegusa et al., 2012). It is possible that SOX9-expressing progenitor epithelial cells divide during the proliferative stage and migrate towards the luminal surface. There, they either undergo apoptosis or perhaps incorporate into the luminal epithelium. A minority of the LE cells in the uterus of Sox9-RG mice were mCherry positive, suggesting that there may also be a resident pool of progenitor cells for the luminal epithelium. Overexpression of Sox9 in the uterine epithelium leads to endometrial hyperplasia We have shown that SOX9 is expressed in uterine glands in mouse and human, suggesting that it regulates processes in the uterine GE. We generated a new mouse model to overexpress Sox9 in the female reproductive tract. We used the Pgr promoter, driving Cre expression to activate Sox9 expression in the uterine epithelium, stroma and myometrium (Soyal et al., 2005). Previously, Kim et al., (2011) showed that the CAG-Sox9 transgene, when activated by Sox9-Cre, was expressed at ∼1.5X the levels of endogenous Sox9 in neonatal rib cartilage and that the downstream chondrocyte markers Aggregan and Col2a1 were up-regulated to similar levels. We speculate that similar levels of Sox9 were achieved in the Sox9-cOE uterine glands, whereas the fold increase in the luminal epithelium may be much higher because SOX9 is predominantly expressed in uterine glands. We found that before puberty, the uteri of Sox9-cOE females appeared morphologically and histologically similar to control littermates, even though SOX9 was detected in the entire uterine epithelium by P35. Therefore, overexpressing Sox9 in the uterine epithelium does not result in immediate morphological abnormalities. However, the endometrium showed a progression to hyperplasia starting at 2-months of age, indicating that the cystically dilated uterine gland lesions developed after the females became sexually mature. Perhaps the continuous remodeling of the uterine tissues during repeated estrous cycles facilitates Sox9-induced abnormalities. The Pgr-Cre allele also expresses in postnatal stroma (Soyal et al., 2005). We did detect some SOX9-expressing cells in the uterine stroma from at least P21 to P35. It is possible that ectopic expression of SOX9 in the uterine stroma also contributed to the subsequent glandular hyperplasia. These abnormal uterine glands in our mouse model resembled human endometrial polyps (Anastasiadis et al., 2000; Deligdisch et al., 2000). Histologically, these structures appeared as cystically dilated uterine glands present in the endometrium surrounded by fibrotic stroma that could invade the myometrium (adenomyosis). In contrast to human endometrial polyps, the abnormal uterine glands in our mouse model did not grow into the uterine cavity (Deligdisch et al., 2000). 20-25% of women over the age of 40 will eventually develop endometrial polyps (Humphrey et al., 2008). Endometrial polyps are usually benign, although it is thought that they can transform into endometrial cancer (Savelli, 2003). Treating mice that contain only one functional p53 allele (p53 +/-) with N-ethyl-N-nitrosourea (ENU), a potent mutagen, led to the development of endometrial polyps with a similar histology to those observed in our mouse model (Mitsumori et al., 2000). These ENU-treated mice also developed endometrial stromal sarcomas that contained point mutations in the remaining p53 allele. Our data suggest that the Sox9-induced uterine gland abnormalities in our mouse model may be preneoplastic lesions for endometrial cancer. The oldest Sox9-cOE uteri also showed complex hyperplasia; glands that varied in size and contained numerous buds. The uterine glands showed a crowded cytoarchitecture, although stroma was observed between the epithelium, indicating that although hyperplastic, the lesions had not progressed to endometrial adenocarcinoma. Therefore, Sox9-cOE females develop cellular changes that are considered hallmarks of disease progression. The cytological abnormalities resembled what is seen in human simple endometrial hyperplasia (Deligdisch et al., 2000; Silverberg, 2000). Moreover, Sox9-cOE females developed uterine glands that grew in a more cribriform, crowded fashion, similar to what is seen in human complex endometrial hyperplasia (Silverberg, 2000). Endometrial polyps and endometrial hyperplastic lesions are thought to develop into endometrioid adenocarcinoma (Savelli, 2003). Thus, the polypoid-like lesions observed in Sox9-cOE uteri may be progressing to a disease state comparable to that found in humans. Cellular mechanisms of uterine gland hypertrophy in Sox9-cOE mice The hyperplastic lesions that develop in Sox9-cOE females may be the result of abnormal cell proliferation. SOX9 is expressed primarily in the GE, which has a lower proliferation rate compared to the LE (Kaitu'u-Lino et al., 2010). In adult females of reproductive age, there was no statistical difference in GE cell proliferation between Sox9-cOE and controls. Similarly, there was no statistical difference for Ki-67 immunostaining assessed by H-score analysis in females (> 8-months of age) past their reproductive age. However, Sox9-cOE uteri contained focal areas with high number of Ki-67 positive GE cells. Therefore, focal hypertrophy in females past reproductive age may explain the appearance of the large number of cysts observed by gross morphology in uteri of mice 8-months-old and older. Our studies are consistent with the idea that Sox9 must be tightly regulated to maintain uterine gland homeostasis. In summary, our findings demonstrate that overexpression of Sox9 is sufficient to induce changes to the tissue architecture of the female mouse reproductive tract and plays a role in the development of histological lesions that resemble human endometrial polyps and hyperplasia. This implicates SOX9 in the pathogenesis of these endometrial diseases and may contribute to the formation of endometrial cancer. Supplementary Material 1 Supplemental Fig. S1. (A) Crosses used to generate mice conditionally overexpressing Sox9. (B) Diagram of CAG-Sox9 allele. Monomeric red fluorescent protein (mRFP) is constitutively expressed driven by the CAG promoter. Sox9 is not expressed because of a transcriptional stop sequence associated with mRFP. Ires-EGFP, intraribosomal entry site-enhanced green fluorescent protein. (C) Females that inherit the CAG-Sox9 allele express mRFP in the entire reproductive tract. WT, wild type. (D) Pgr-Cre; CAG-Sox9 (Sox9-cOE) females express Cre under the control of the Pgr locus, leading to Cre-mediated excision of the mRFP sequence and expression of Sox9 and EGFP. 2 Supplemental Fig. S2. Histological analysis of endometrial lesions in Sox9-cOE uteri. H&E sections of uteri from Sox9-cOE females and controls. (A, B) Uterus from 1-month-old female control. (B) Higher magnification image of A. (C, D) Uterus from 1-month-old Sox9-cOE female shows similar morphology as A. (D) Higher magnification image of C. (E, F) Uterus from 2-month-old female control. (F) Higher magnification image of E. (G, H) Uterus from 2-month-old Sox9-cOE female shows multiple dilated cystic uterine glands and denser stroma compared to E. (H) Higher magnification image of uterus from 2-month-old Sox9-cOE female to highlight abnormal glandular morphology. (I, J) Uterus from 12-month-old female control. (J) Higher magnification image of I. (K-N) Uteri from 12-month-old Sox9-cOE females show abnormal morphology. (K) Uterus shows crowded uterine glands with cribriform appearance, multiple dilated cystic uterine glands and denser stroma. (M) Uterus shows multiple dilated cystic uterine glands and denser stroma. (L, N) Higher magnification of uteri from 12-month-old Sox9-cOE female showing GE with rounder nuclei and loss of normal columnar morphology (arrows). Scale bar, 50 μm. We thank Dr. Russell Broaddus (M.D. Anderson Cancer Center) for providing human uterine tissues, Dr. Franco DeMayo (Baylor College of Medicine) for providing the Pgr-Cre mouse strain, and David Stewart for strain rederivation. Supported by National Institutes of Health (NIH) grant HD030284 and the Ben F. Love Endowment to R.R.B. G.G. was supported by National Cancer Institute T32 grant CA09299. Veterinary resources were supported by NIH grant CA16672. Fig. 1 SOX9 expression in the adult mouse and human uterus. (A, B) Immunofluorescent staining using an α-SOX9 antibody (green) counterstained with DAPI (blue) of proestrus stage adult mouse (A) and postmenopausal human (B) uterine sections. (C, D) Fluorescent images of 8-week-old Sox9-Cre; R26R-RG double heterozygous female, showing H2BmCherry-positive cells (red) present mostly in the GE. (C) Nuclei were counterstained with DAPI (blue). (D) The blue channel (DAPI) was removed to show that most H2BmCherry-positive nuclei are present in the GE. Dotted line outlines the luminal epithelium. GE, glandular epithelium; LE, luminal epithelium. Scale bars, 50 μm. Fig. 2 SOX9 expression during the estrous cycle. (A, C, E, G) Immunofluorescent staining using an α-SOX9 antibody (green) counterstained with DAPI (black) of adult uterine sections. (B, D, F, H) higher magnifications of A, C, E, G, respectively. Note that most mitotic figures in B and D (black arrowheads and insets) are devoid of SOX9. Scale bar, 50 μm Fig. 3 Wild-type and SOX9 overexpression in the uteri of prepubertal female mice. (A-L) Immunofluorescent staining for SOX9 (green) and E-cadherin (red) at P21, 28, and 35. Uteri from control (A-F) and Sox9-cOE (G-L) females. (B, D, F, H, J, L) Higher magnification views of A, C, E, G, I, K, respectively, with the background subtracted to emphasize SOX9. (A-F) SOX9 overexpression is observed in the LE and GE of uteri from Sox9-cOE at P35 (K, L) compared to controls (E, F). Some SOX9 was also detected in the stroma and myometrium of uteri from Sox9-cOE at P28 (I, J) and P35 (K, L) compared to controls (C-F). DAPI, blue. Scale bar, 50 μm. Fig. 4 Histological analysis of Sox9-OE uteri. (A-F) H&E stained uterine sections from control (A-C) and Sox9-cOE (D-F) females at 2-, 4-, and 8-months of age. (D-F) The stroma of Sox9-cOE uteri appeared denser (asterisk) and contained dilated cystic uterine glands (arrowheads). Scale bar, 500 μm. (G, J) High magnification image of H&E stained uterine sections from 8-month-old females showing fibrotic stroma (asterisk) in Sox9-cOE females (J, asterisk). (H, I, K, L) Masson's Trichrome stained uterine sections from control (H, I) and Sox9-cOE (K, L) females at 4-months of age. More intense blue staining in the Sox9-cOE uterine sections compared to controls indicates an increase in connective tissue/collagen fibers. (I, L) Scale bar, 50 μm. GE, glandular epithelium. Fig. 5 Gross morphology of 1-year old Sox9-OE uteri. (A-F) Brightfield images of 1-year-old uteri. (A-C) controls, (D-F) Sox9-cOE. (C, F) Uteri cut longitudinally (mesometrial side at the top). (A-C) 1-2 cystic structures are observed in the control uteri (arrowheads). (D-F) Numerous cystic structures (arrowheads) are present inside the uteri of Sox9-cOE females, some appeared filled with white debris (asterisk). Scale bar, 2 mm. Fig. 6 Cell proliferation and cell death analysis in Sox9-OE uteri. (A-C) Control and (D-F) Sox9-cOE 4-month-old uteri. (A, D) Immunofluorescent staining for phosphohistone H3 (red) on uterine sections taken during estrus. (B, E) Immunofluorescent staining for SOX9 (green) and Ki-67 (red) on uterine sections taken during estrus. DAPI, blue. (C, F) TUNEL assay on uterine sections taken during metestrus. TUNEL positive cells (green), DAPI (blue). Scale bar, 50 μm. Fig. 7 Ki-67 analysis of cell proliferation in aged females. (A-F) Immunohistochemical staining for Ki-67. (A, C, D) Control and (B, E, F) Sox9-cOE 8-month-old uteri. (C, D) Representative areas of A. (E, F) Representative areas of B to show local difference in Ki67 staining in one uterus. (G) Boxplot of H-Score analysis with standard error of the mean of 4 control and 6 Sox9-cOE uteri obtained from age-matched females. Scale bar, 100 μm. Fig. 8 Differentiation of Sox9-OE uterine glands. Immunofluorescent staining of the uterus for FOXA2 (red), in 8-month old control (A) and Sox9-OE (B) females. (B) Cribriform uterine glands in the Sox9-OE females showed less FOXA2 immunostaining. DAPI, blue. (C-F) Immunofluorescent staining of the uterus containing dilated cystic uterine glands for p63 (red) and SOX9 (green) in 1-year-old Sox9-OE female. (E, F) Sections from 1-year-old Sox9-cOE females containing cervical epithelium. DAPI, black. Scale bar, 50 μm. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. Akiyama H Kim JE Nakashima K Balmes G Iwai N Deng JM Zhang Z Martin JF Behringer RR Nakamura T de Crombrugghe B 2005 Osteo-chondroprogenitor cells are derived from Sox9 expressing precursors Proc Natl Acad Sci U S A 102 14665 14670 10.1073/pnas.0504750102 16203988 Anastasiadis PG Koutlaki NG Skaphida PG Galazios GC Tsikouras PN Liberis VA 2000 Endometrial polyps: prevalence, detection, and malignant potential in women with abnormal uterine bleeding Eur J Gynaecol Oncol 21 180 183 10843481 Behringer R Gertsenstein M Vintersten Nagy K Nagy A 2014 Manipulating the Mouse Embryo: A Laboratory Manual Fourth Edi Cold Spring Harbor Laboratory Press Bokhman JV 1983 Two pathogenetic types of endometrial carcinoma Gynecol Oncol 15 10 17 6822361 Darido C Buchert M Pannequin J Bastide P Zalzali H Mantamadiotis T Bourgaux JF Garambois V Jay P Blache P Joubert D Hollande F 2008 Defective claudin-7 regulation by Tcf-4 and Sox-9 disrupts the polarity and increases the tumorigenicity of colorectal cancer cells Cancer Res 68 4258 4268 10.1158/0008-5472.CAN-07-5805 18519685 Deligdisch L Kalir T Cohen CJ de Latour M Le Bouedec G Penault-Llorca F 2000 Endometrial histopathology in 700 patients treated with tamoxifen for breast cancer Gynecol Oncol 78 181 6 10.1006/gyno.2000.5859 10926800 Dutt A Salvesen HB Chen TH Ramos AH Onofrio RC Hatton C Nicoletti R Winckler W Grewal R Hanna M Wyhs N Ziaugra L Richter DJ Trovik J Engelsen IB Stefansson IM Fennell T Cibulskis K Zody MC Akslen La Gabriel S Wong KK Sellers WR Meyerson M Greulich H 2008 Drug-sensitive FGFR2 mutations in endometrial carcinoma Proc Natl Acad Sci U S A 105 8713 8717 10.1073/pnas.0803379105 18552176 Foster JW Dominguez-Steglich MA Guioli S Kwok C Weller PA Stevanovic M Weissenbach J Mansour S Young ID Goodfellow PN Brook JD Schafer AJ 1994 Campomelic dysplasia and autosomal sex reversal caused by mutations in an SRY-related gene Nature 372 525 530 7990924 Franco HL Dai D Lee KY Rubel Ca Roop D Boerboom D Jeong JW Lydon JP Bagchi IC Bagchi MK DeMayo FJ 2011 WNT4 is a key regulator of normal postnatal uterine development and progesterone signaling during embryo implantation and decidualization in the mouse FASEB J 25 1176 1187 10.1096/fj.10-175349 21163860 Furuyama K Kawaguchi Y Akiyama H Horiguchi M Kodama S Kuhara T Hosokawa S Elbahrawy A Soeda T Koizumi M Masui T Kawaguchi M Takaori K Doi R Nishi E Kakinoki R Deng JM Behringer RR Nakamura T Uemoto S 2011 Continuous cell supply from a Sox9-expressing progenitor zone in adult liver, exocrine pancreas and intestine Nat Genet 43 34 41 10.1038/ng.722 21113154 Gargett CE 2007 Uterine stem cells: What is the evidence? Hum Reprod Updat 13 87 101 10.1093/humupd/dml045 Garry R Hart R Karthigasu Ka Burke C 2010 Structural changes in endometrial basal glands during menstruation BJOG 117 1175 85 10.1111/j.1471-0528.2010.02630.x 20560946 Gray CA Bartol FF Tarleton BJ Wiley AA Johnson GA Bazer FW Spencer TE Al GET 2001 Developmental Biology of Uterine Glands Biol Reprod 65 1311 1323 10.1095/biolreprod65.5.1311 11673245 Huang CC Orvis GD Wang Y Behringer RR 2012 Stromal-to-epithelial transition during postpartum endometrial regeneration PLoS One 7 e44285 10.1371/journal.pone.0044285 22970108 Humphrey PA Dehner LP Pfeifer JD 2008 The Washington Manual of Surgical Pathology Lippincott Williams & Wilkins 1 Jeong JW Kwak I Lee KY Kim TH Large MJ Stewart CL Kaestner KH Lydon JP DeMayo FJ 2010 Foxa2 is essential for mouse endometrial gland development and fertility Biol Reprod 83 396 403 10.1095/biolreprod.109.083154 20484741 Jiang SS Fang WT Hou YH Huang SF Yen BL Chang JL Li SM Liu HP Liu YL Huang CT Li YW Jang TH Chan SH Yang SJ Hsiung Ca Wu CW Wang LH Chang IS 2010 Upregulation of SOX9 in lung adenocarcinoma and its involvement in the regulation of cell growth and tumorigenicity Clin Cancer Res 16 4363 4373 10.1158/1078-0432.CCR-10-0138 20651055 Kaitu'u-Lino TJ Ye L Gargett CE 2010 Reepithelialization of the uterine surface arises from endometrial glands: evidence from a functional mouse model of breakdown and repair Endocrinology 151 3386 3395 10.1210/en.2009-1334 20444944 Kiefer JC 2007 Back to basics: Sox genes Dev Dyn 236 2356 2366 10.1002/dvdy.21218 17584862 Kim Y Murao H Yamamoto K Deng JM Behringer RR Nakamura T Akiyama H 2011 Generation of transgenic mice for conditional overexpression of Sox9 J Bone Miner Metab 29 123 129 10.1007/s00774-010-0206-z 20676705 Kurita T Cunha GR Robboy SJ Mills Aa Medina RT 2005 Differential expression of p63 isoforms in female reproductive organs Mech Dev 122 1043 55 10.1016/j.mod.2005.04.008 15922574 Lacey JV Sherman ME Rush BB Ronnett BM Ioffe OB Duggan Ma Glass AG Richesson Da Chatterjee N Langholz B 2010 Absolute risk of endometrial carcinoma during 20-year follow-up among women with endometrial hyperplasia J Clin Oncol 28 788 792 10.1200/JCO.2009.24.1315 20065186 Lefebvre V de Crombrugghe B 1998 Toward understanding S0X9 function in chondrocyte differentiation Matrix Biol 16 529 540 9569122 Ling S Chang X Schultz L Lee TK Chaux A Marchionni L Netto GJ Sidransky D Berman DM 2011 An EGFR-ERK-SOX9 signaling cascade links urothelial development and regeneration to cancer Cancer Res 71 3812 3821 10.1158/0008-5472.CAN-10-3072 21512138 Medh R Thompson B 2000 Hormonal regulation of physiological cell turnover and apoptosis Cell Tissue Res 301 101 124 10928284 Mitsumori K Onodera H Shimo T Yasuhara K Takagi H Koujitani T Hirose M Maruyama C Wakana S 2000 Rapid induction of uterine tumors with p53 point mutations in heterozygous p53-deficient CBA mice given a single intraperitoneal administration of N-ethyl-N-nitrosourea Carcinogenesis 21 1039 1042 10783330 Panda DK Miao D Lefebvre V Hendy GN Goltzman D 2001 The Transcription Factor SOX9 Regulates Cell Cycle and Differentiation Genes in Chondrocytic CFK2 Cells J Biol Chem 276 41229 41236 10.1074/jbc.M104231200 11514554 Poché RA Furuta Y Chaboissier MC Schedl A Behringer RR 2008 Sox9 is expressed in mouse multipotent retinal progenitor cells and functions in Müller Glial cell development J Comp Neurol 510 237 250 18626943 Saegusa M Hashimura M Suzuki E Yoshida T Kuwata T 2012 Transcriptional upregulation of Sox9 by NF-κB in endometrial carcinoma cells, modulating cell proliferation through alteration in the p14(ARF)/p53/p21(WAF1) pathway Am J Pathol 181 684 92 10.1016/j.ajpath.2012.05.008 22698986 Savelli L 2003 Histopathologic features and risk factors for benignity, hyperplasia, and cancer in endometrial polyps Am J Obstet Gynecol 188 927 931 10.1067/mob.2003.247 12712087 Sherman ME 2000 Theories of endometrial carcinogenesis: a multidisciplinary approach Mod Pathol 13 295 308 10.1038/modpathol.3880051 10757340 Shioi G Kiyonari H Abe T Nakao K Fujimori T Jang CW Huang CC Akiyama H Behringer RR Aizawa S 2011 A mouse reporter line to conditionally mark nuclei and cell membranes for in vivo live-imaging Genesis 49 570 578 10.1002/dvg.20758 21504045 Silverberg SG 2000 Problems in the differential diagnosis of endometrial hyperplasia and carcinoma Mod Pathol 13 309 327 10.1038/modpathol.3880053 10757341 Soyal SM Mukherjee A Lee KYS Li J Li H DeMayo FJ Lydon JP 2005 Cre-mediated recombination in cell lineages that express the progesterone receptor genesis 41 58 66 15682389 Spencer TE Hayashi K Hu J Carpenter KD 2005 Comparative developmental biology of the mammalian uterus Curr Top Dev Biol 68 85 122 10.1016/S0070-2153(05)68004-0 16124997 Stewart MD Jang C Hong NW Austin AP Behringer RR 2009 Dual fluorescent protein reporters for studying cell behaviors in vivo Genesis 47 708 717 10.1002/dvg.20565.Dual 19813259 Uhlen M Oksvold P Fagerberg L Lundberg E Jonasson K Forsberg M Zwahlen M Kampf C Wester K Hober S Wernerus H Bjorling L Ponten F 2010 Towards a knowledge-based Human Protein Atlas Nat Biotech 28 1248 1250 Uhlenhaut NH Jakob S Anlag K Eisenberger T Sekido R Kress J Treier AC Klugmann C Klasen C Holter NI Riethmacher D Schütz G Cooney AJ Lovell-Badge R Treier M 2009 Somatic sex reprogramming of adult ovaries to testes by FOXL2 ablation Cell 139 1130 42 10.1016/j.cell.2009.11.021 20005806 Valentijn AJ Palial K Al-Lamee H Tempest N Drury J Von Zglinicki T Saretzki G Murray P Gargett CE Hapangama DK 2013 SSEA-1 isolates human endometrial basal glandular epithelial cells: phenotypic and functional characterization and implications in the pathogenesis of endometriosis Hum Reprod 28 2695 2708 10.1093/humrep/det285 23847113 Wang H Leav I Ibaragi S Wegner M Hu G Lu ML Balk SP Yuan X 2008 SOX9 is expressed in human fetal prostate epithelium and enhances prostate cancer invasion Cancer Res 68 1625 1630 10.1158/0008-5472.CAN-07-5915 18339840

PMC005xxxxxx/PMC5133193.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 8710523 21484 Adv Drug Deliv Rev Adv. Drug Deliv. Rev. Advanced drug delivery reviews 0169-409X 1872-8294 27262925 5133193 10.1016/j.addr.2016.05.020 NIHMS798594 Article PLA Micro- and Nano-Particles Lee Byung Kook Yun Yeonhee Park Kinam * Purdue University, Departments of Biomedical Engineering & Pharmaceutics, West Lafayette, Indiana, U.S.A. Correspondence to: Kinam Park, Ph.D., Purdue University, Departments of Biomedical Engineering & Pharmaceutics, 206 S. Martin Jischke Drive, West Lafayette, IN 47907, USA, Tel: 765-494-7759, kpark@purdue.edu 29 6 2016 01 6 2016 15 12 2016 15 12 2017 107 176191 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. Poly(D,L-lactic acid) (PLA) has been widely used for various biomedical applications for its biodegradable, biocompatible, and nontoxic properties. Various methods, such as emulsion, salting out, and precipitation, have been used to make better PLA micro and nano-particle formulations. They are widely used as controlled drug delivery systems of therapeutic molecules, including proteins, genes, vaccines, and anti-cancer drugs. Even though PLA-based particles have challenges to overcome, such as low drug loading capacity, low encapsulation efficiency, and terminal sterilization, continuous innovations in particulate formulations will lead to development of clinically useful formulations. Graphical abstract PLA Micro-particles Nano-particles Fabrication methods Drug delivery system 1. Introduction Shortly after poly(D,L-lactic acid) (PLA), poly(glycolic acid) (PGA), and poly(lactic-co-glycolic acid) (PLGA) were developed for use in surgical implants and tissue repair in the 1960s, they have been used widely for various biomedical applications, including sutures, bone plates, abdominal mesh, and controlled release drug delivery [1-5]. These polymers, which are biocompatible, biodegradable, and nontoxic, have been used for various biomedical applications for decades [5, 6]. The first PLGA-based drug delivery system approved by the Food and Drug Administration (FDA) was the Lupron Depot drug delivery system. It is made of PLGA (L:G ratio of 75:25) and leuprolide acetate for treatment of advanced prostate cancer. This system delivers the drug over a period of 4 months after a single injection [7]. The list of FDA-approved controlled release products of PLA and PLGA formulations are given in Table 1. The FDA requires cGMP (current good manufacturing practice) protocols to ensure efficacy, safety, and stability for pharmaceuticals. PLGA and PLA for clinical applications are manufactured under cGMP regulation [9, 10]. The list of cGMP grade PLAs from major suppliers is shown in Table 2. Researchers can get easy access to the non-GMP grade polymers through commercial suppliers, such as Sigma-Aldrich, Vornia, Akina, and other suppliers. Polyesters such as PLA and PLGA have been used extensively in drug delivery because of their biodegradable and mechanical properties that can be adjustable [11, 12]. They can be designed and synthesized with different molecular weights and L:G ratios for individual applications with high reproducibility at low cost. These advantages allow researchers to make micro and nano-particles using PLA and PLGA. They have been used to make various delivery systems for low molecular weight drugs as well as peptide and protein drugs [1-4, 13, 14]. Most nano-particle formulations based on PLA and PLGA have been focused on drug delivery to target tumors [13]. One of the advantages of using PLA to make micro and nano-particles is the flexibility. Physical properties, such as size and shape, and chemical properties, including molecular weight and L:G ratio, can be easily controlled to obtain desirable pharmacokinetic and biodegradable properties. Typically, the particles include spheres, capsules, cubes, and other shapes [14, 15]. An active pharmaceutical ingredient (API) is usually dispersed homogeneously within the PLA matrix [2]. PLA-based micro and nano-particles are useful in drug delivery and biomedical applications, but there are also a variety of limitations, such as high initial burst release, as listed in Table 3. Nano-particles are internalized in cells partly through fluid phase pinocytosis and also through clathrin-mediated endocytosis. They rapidly escape the endo-lysosomes and enter the cytoplasm within 10 min of incubation [12]. The micro and nano-particles can effectively incorporate the drug into their structure [8]. Micro and nano-particles with large surface-to-volume ratios provide a greater number of reaction sites than macro size particles with smaller surface areas [16]. Various methods to modify properties of PLA micro and nano-particles are available [1, 3, 5, 6, 11, 16-19]. Bulk and surface properties can be modified. The bulk modification methods include blending with different polymers, plasticization, copolymerization, and cross-linking. The surface modification methods include surface coating, entrapment, and plasma treatment. Supplemental biomaterials, such as poly(ethylene glycol) (PEG), polysaccharides, and extracellular matrix (ECM) proteins, have been used for coating PLA micro and nano-particles [2, 20]. All modifications are designed to achieve high drug encapsulation efficiency and loading, and controlled drug release rate. The following sections cover different preparation techniques for making PLA micro and nano-particles, and their applications, handling challenges, and strategies. 2. Preparation techniques of micro and nano-particles There are several techniques useful for the preparation of PLA-based micro and nano-particles [21-28]. The techniques are classified into four categories. Category 1 is traditional emulsion based methods which are single emulsion, double emulsion, and multiple emulsions. Category 2 is precipitation-based methods which include nano precipitation, rapid expansion of supercritical fluid into liquid, salting out, and dialysis. Category 3 is direct compositing methods, such as melting technique, spray drying, supercritical fluid and in situ forming micro-particles. Category 4 includes new approaches including microfluidic technique and template/mold based technique. Other criteria depend on the mode of drug encapsulation. The drug is either entrapped inside of the particles of "capsules" or dispersed in polymer matrices [6, 17, 29, 30]. Fig. 1 shows a representative micro and nano-particle structure with the four categories of preparation techniques. 2.1. Single emulsion One of the simplest methods to make micro and nano-particle is the single emulsion/solvent extraction technique. Many hydrophobic drugs are dissolved with PLA in various water-immiscible organic solvents which are then emulsified in a water phase containing a stabilizer. Emulsion, such as o/w (oil in water), o/o (oil in oil) or w/o (water in oil), can be formed to accommodate different types of dispersed phase and the dispersion medium [6, 17]. The emulsification is exposed to a high energy source, such as ultrasound, homogenizer, or milling. The oil phase is removed by evaporation under low pressure or vacuum or by solvent extraction using a large volume of water, leading to the formation of particles dispersed in the water phase. The particles are collected by centrifugation or filtration and washed with pure water or buffer solution to remove residual stabilizers and any free drug. The harvested particles is lyophilized for storage [13, 17, 31]. The emulsion/solvent extraction technique can produce various size particles ranging from nanometers to micrometers by controlling the agitation rate and other experimental parameters. The parameters for loading a water soluble drug into particles include the phase volumes of oil and water, concentration of polymer and drug, presence of oil soluble surfactant in oil, stabilizer/surfactant in oil/water, saturation solubility of drug in water, and stirring rate [1, 17, 32-34]. 2.2. Double emulsion The single emulsion/extraction methods have challenges of poor encapsulation of hydrophilic drugs due to their diffusion and dispersion from the emulsified oil phase into the aqueous continuous phase. Thus, double-emulsion/extraction methods have been frequently used for improving the encapsulation efficiency of water soluble drugs, such as peptides and proteins [17, 35, 36]. In some cases, solid/oil/water (s/o/w) emulsion has been used for a high drug loading of water soluble peptides, such as insulin [17, 37, 38]. This involves addition of a water-soluble drug solution to an organic polymer solution under high energy stirring to form a w/o (or reversely o/w) emulsion. This w/o emulsion is added into a second water phase containing a stabilizer with stirring, resulting in the formation of a w/o/w (or o/w/o) emulsion. The organic solvent is removed under reduced pressure or vacuum to produce polymer particles. The harvested particles are thoroughly washed using pure water or buffer to remove residual raw materials before lyophilization [2, 17]. There are several parameters that have to be adjusted to optimize the characteristics of particles prepared by the double emulsion method. These include the amount of hydrophilic drug to be added, polymer concentration, type of solvent, stabilizer concentration, volume of the second aqueous phase, stirring rate and other variables [1, 17, 29, 32-34]. Fig. 2 shows the single and double emulsion processes with oil, emulsifier, and water. In one example, PLA three different molecular weights were used to prepare micro-particles containing prilocaine (PRL), which is an amino-amide type local anesthetic. PRL-loaded PLA micro-particles were prepared by w/o/w double emulsion methods. The particle sizes were 32 μm, 40 μm, and 68 μm. The SEM images and release profiles of three different types of PLA micro-particles are shown in Fig. 3 and Fig. 4, respectively [39]. 2.3. Salting out An alternative to the widely applied emulsion-based technique is the salting-out method. This method involves the addition of polymer and drug solution in a water-miscible solvent such as acetone, acetonitrile, or tetrahydrofuran to an aqueous solution containing the salting-out agent (e.g., magnesium chloride and calcium chloride) and a colloidal stabilizer, such as polyvinylpyrrolidone, under high speed stirring (Fig. 5) [17]. When this o/w emulsion is diluted with a large amount of water, it induces the formation of particles by enhancing the diffusion of the miscible solvent into the water phase. The particles can be purified and harvested by centrifugation or cross-flow filtration [29, 40]. One of the important advantages of this method is minimizing tension to the loaded protein [17, 40]. Salting out does not need a heating process, and thus may be useful when heat sensitive drugs have to be encapsulated [41-43]. The salting out process requires optimization of the process conditions, e.g., the salt type and concentration, the type of polymer and solvent, and the ratios of these compounds in order to obtain micro-particles [17]. 2.4. Nanoprecipitation The nanoprecipitation method is a relatively easy and reproducible technique for the preparation of PLA based nano-particles. The nanoprecipitation is a one step process, also known as the solvent displacement method (Fig. 6) [29, 30]. The advantages of this method are: narrow size distribution; less toxic and eco-friendly solvents; and low energy source of a stirring device. The nanoprecipitation method can be applied in various ways: (i) direct pouring of an anti-solvent (e.g., water phase) into an organic solution; (ii) slow drop-wise addition of water to an organic solution; (iii) direct pouring of an organic solution into the water phase; and (iv) dilution of a polymeric dispersed phase using an anti-solvent. The solvent is then removed from the suspension under reduced pressure or vacuum [44, 45]. There are variable key parameters for forming nano-particles. The injection rate of the organic phase into the anti-solvent phase affects the particle size. The mixing rate affects both particle size and drug encapsulation yield. The type of the organic solvent also affects the size and encapsulation efficiency of particles. Typical solvents used for nanoprecipitation are acetone, acetonitrile, dimethylacetamide, dimethylformamide, dimethylsulfoxide (DMSO), 2-pyrrolidone, N-methyl-2-pyrrolidone (NMP), PEG, and tetrahydrofuran. Acetone is the most preferred solvent. Usually, a binary mixture of solvents is used, e.g., acetone-ethanol [28]. Other factors are the drug: polymer ratio, surfactant, and anti-solvent phase volume ratio. The difficulty faced in this method is the choice of combination of the drug/polymer/solvent/non-solvent system (usually called Ouzo region [46-48]) in which the nano-particles would be formed with a high drug encapsulation efficiency [30]. Producing the successful nano-particles, however, is restricted to a narrow condition of the Ouzo region. Beyond the Ouzo region, micro-particles rather than nano-particles are produced [29, 30]. Improved loading of procaine hydrochloride, a water soluble drug, into polymer nano-particles was achieved by increasing the aqueous phase pH and replacing procaine hydrochloride with a procaine dihydrate base [44]. Docetaxel (DTX)-loaded nano-particles were prepared by the nanoprecipitation method using polydopamine-modified tocopherol polyethylene glycol succinate (TPGS)-PLA. The distribution of the nano-particle size was around 126~209 nm. The DTX loaded PLA nano-particles reduced the tumor size most significantly on hepatoma-bearing nude mice (Fig. 7) [49]. 2.5. Dialysis Dialysis is an effective, simple method for forming small size and narrowly distributed nano-particles similar to nanoprecipitation. A polymer is dissolved in an organic solvent and placed in a dialysis tube. The dialysis is performed in a non-solvent miscible with an organic solvent. The organic solvent is displaced by non-solvent resulting in a loss of polymer solubility, and subsequent formation of polymer aggregates. Nano-particles of homogeneous suspension are collected after fully displacing the organic solvent with non-solvent (Fig. 8) [6, 50-54]. 2.6. Spray drying Spray drying is a useful, continuous particle production method, where the drug is dissolved or dispersed in an organic phase with a polymer that is then sprayed as ultra fine droplets in dry air flow [17, 55-58]. The organic phase is instantly evaporated. The dried particles are collected under low pressure with dry air flow (Fig. 9). This technique is easy to set up, but at the same time it is hard to control the drug distribution in the particles [4, 28]. Spray drying has been studied for protein encapsulation to improve the stability of biomacromolecules [17]. It is also useful for large scale hydrophobic drug particle production [56, 59-63]. The size distribution and morphology of the spray dried PLA micro-particles were analyzed by SEM. The largest diameter was 3 μm for making 5% of the PLA solution. There was no significant difference between 0.5% and 1% of PLA solutions (Fig. 10) [64]. 2.7. In situ forming micro-particle In situ forming depots have been used for site forming micro-particles or micro depots. This approach overcomes some drawbacks of conventional techniques, including manufacturing costs and complexities of other methods, e.g., drying and resuspension [17]. A drug-polymer solution is administered via injection at the target site where it is precipitated into an implant or forms micro-particles (Fig. 11), a concept that has been employed in FDA-approved long-acting release products [17]. The drug/polymer solutions are dissolved in water-miscible solvents, such as NMP and DMSO. The toxicity of solvents must be examined before selection. Some solvents show lower myotoxicities of in situ implants and in situ micro-particles [27]. Water-miscible solvents result in hardening of emulsion droplets in vivo, and the solvent removal process may be responsible for a high burst release [65, 66]. The safety issues may limit types of oils, e.g., paraffin/mineral oils, that can be used [17]. The delivery of leuprolide acetate for months was based on in situ forming micro-particle formulation. A conventional formulation is a two syringe/connector system. A solution of leuprolide and PLGA or PLA in NMP was emulsified into an external oil phase. In situ forming PLGA micro-particles showed a high initial release (~40%) because of their high porosity (Fig. 12). In situ forming PLA micro-particles exhibited a much lower initial release (~9%), which had a slow and continuous drug release (Fig. 13) [67]. 2.8. Melting technique The melting technique provides another option for encapsulating drugs into polymers. The melting process avoids the use of organic solvents, but the drug is dispersed in a polymer melt. The resulting drug/polymer melt is solidified by a cooling water phase or a cooling chamber with dry air flow (Fig. 14) [17, 28]. The drug/polymer melt is cooled down and then ground or milled to form particles [17]. If spherical particles and a smaller distribution are desired, the ground melt can be emulsified in a hot solution containing emulsifier or a hot gel [68]. Limitations of this approach are the thermal treatment of the drug and the multitude of steps to obtained smooth micro-particles [17]. 2.9. Supercritical fluids technique (SCF) Supercritical fluid and dense gas technology offer an interesting and effective technique for particle production, avoiding most of the drawbacks of the traditional methods. The supercritical fluid method uses more environmentally friendly solvents and has the potential to produce nano-particles with high purity and no residual solvents [69-71]. Two principal processes have been developed for the production of nano-particles using supercritical fluids: rapid expansion of supercritical solution (RESS) and rapid expansion of supercritical solution into a liquid solvent (RESOLV) (Fig. 15). A limitation of the RESS is the use of low concentrations and low molecular weight PLAs [6, 17]. It was hard to control particle quality, such as size, and morphology [72]. The effects of operating conditions, such as pressure, flow rate and concentration of drug and polymer, were evaluated using different sizes and morphologies of particles. The PLA particles presented mean diameters between 5.4 ~ 20.5 μm (Table 4 and Fig. 16) [73]. 2.10. Microfluidic technique The microfluidic technique can fabricate uniform, biodegradable PLA-based particles and implants. The uniform particles may allow precisely controlled release systems, because the size of the particles is a primary determinant of drug release kinetics [1]. Droplet microfluidics deals with discrete droplets having precisely controlled volume and composition, restricted dispersion, which are ideal templates for fabricating complex particles. A number of microfluidic approaches have been developed and are widely used for fabricating single emulsion, double or multiple emulsions (Fig. 17). The polymeric micro-particles are generally fabricated by making o/w emulsions in microfluidic devices, where polymers are dissolved in an organic solvent (oil phase), by droplet solidification through solvent evaporation, diffusion or extraction. The polymer solution in the organic solvent is filled in a T- or Y-junction microfluidic device and then ejected into a large amount of stabilizing solution. The solvent is diffused from droplets into the water phase, and then the droplets are solidified to microspheres due to the relatively different solubility of the solvent in water [2, 15, 74, 75]. When the polymer solution droplets are ejected into the stabilizing solution, the size of the particles are determined by properties of the solutions (density and viscosity), the flow rate of the polymer solution, the diameter of the nozzle, and the interfacial tension between the polymer solution and the nozzle tip. There are similar limitations as other methods using physical devices such as a nozzle, a channel, a template, and a mold. The fabrication of nano-particles would apparently require a smaller nozzle or channels [18]. Applications of microfluidic based micro-particles are rapidly growing for development of controlled drug delivery systems. They have been useful for complex and multifunctional drug delivery systems as multi core-shell micro-particles [15]. With further development of microfluidic techniques and manufacturing processes, micro-particles with desired drug loading and release kinetics can be prepared. Furthermore, the low cost and high reproducibility make this technology promising for mass production of specific drug delivery systems [15, 76]. PLA particles over a wide range of size were made from w/o/w emulsion produced in a three-dimensional (3D) flow focusing glass capillary device. The droplet size is usually controlled by the fluid rate and orifice size. Also, a numerical model of drop generation in a 3D flow focusing device was developed to understand the mechanism of drop generation in the dripping regime [77]. The microfluidics technique was used to make polymeric micro-particles of two different sizes (11 μm and 41 μm) to study in vitro release profile of bupivacaine (Fig. 18). Mono dispersed particles prepared using microfluidics released drug more slowly than similar size particles prepared using a conventional method such as the emulsion method [78]. While the micro-particle manufacturing method is one parameter affecting the drug release kinetics, it is important to realize that other factors also contribute significantly to the drug release profile. 2.11. Template/mold based technique The hydrogel template method is based on the unique properties of physical gels that can undergo sol-gel phase transition upon changes in environmental conditions such as temperature. The hydrogel template is useful in producing particles in homogeneous sizes. The first step is to form a certain pattern on a hard master template. A warm aqueous hydrogel solution (e.g., gelatin solution) is poured on top of the master template, and then the template is placed under low temperature conditions for imprinting the template by a formed hydrogel mold. The solidified mold is peeled off and a polymer/drug solution, in a suitable organic solvent, is poured on the hydrogel mold and evenly spread for filling empty cavities. The filled hydrogel mold is dried to evaporate the solvent. The drug-loaded particles are collected by dissolving the hydrogel mold in water. The particles are washed in water, and collected by centrifugation or filtration [79]. The gelatin hydrogel mold, however, is easy to be damaged while spreading a drug-polymer solution on the template. Thus, an alternative polymer, poly(vinyl alcohol) (PVA), was used to make a water-soluble polymer mold. The PVA mold has many advantages over the gelatin gel mold, including stronger mechanical strength, ease of handling for making a mold, and storage in a dry chamber before use. After drying of the PVA mold having drug-containing microparticles, it is placed in distilled water and stirred at room temperature. Only the PVA mold is dissolved completely and then the drug-containing microparticles are floating freely in the water. The microparticles are washed by filtering through dual-layer meshes made of stainless steel. After collection using centrifugation, the microparticles are washed again in double distilled water to remove residual PVA, centrifuged, and the supernatant is removed. The microparticles are vacuum dried overnight. Fig. 19 shows the procedure and fabricated PLA (average MW 25,000~35,000) micro-particles produced by using this system. The shape of the microparticles is cylinderical because PVA mold having a cylinder pattern with flat bottom is used. The shape of microparaticles depends on the PVA mold pattern. Geometry of microparticles is an important parameter for the drug release rate and interaction with cells and tissues. For cylinder and spherical shapes, the surface areas may be different but the overall release mechanism, whether diffusion-controlled reservoir or matrix systems, is not expected to change due to the shape. Thus, the shape of microparticles may not affect the drug release kinetics significantly. Since the microparticle size is around 50 μm, it is not expected to make any difference in administration. The homogeneous microparticles, in fact, make it easier for administration. The PVA mold method was used to make micro-particles of three poorly water-soluble drugs: risperidone (RIS), methylprednisolone acetate (MPA), and paclitaxel (PTX). The fabricated micro-particles showed great conformity to the original template design for a wide range of formulation conditions. In addition, the micro-particles produced showed narrow size distribution, which provided advantages compared with the conventional emulsion-based method (Fig. 20) [80]. The PVA mold method was extended to develop the Vacuum SpinSwiper machine to make micro-particles more efficiently in large quantities [81]. There are several advantages of the template based techniques. The preferred advantages are mono-dispersed and predetermined micro-particles dimension, easy scale-up, and a reproducible process. The drug loading by the PVA mold approach can be high for water-soluble drugs, which is not easy to produce by conventional methods. There are also limitations, however. The size of the particle is currently in the micron scale, i.e., larger than 1 μm because of the use of UV lithographic technique in fabricating silicon water master templates. Fabrication of nano-particles requires preparation of the master templates with nano size patterns. 3. Applications of PLA micro and nano-particles PLA micro and nano-particles have been proposed for improving oral bioavailability of poorly water soluble drugs. Nano-particles are thought to be absorbed from the gastrointestinal tract after oral administration [2, 82]. Poorly water soluble drugs are difficult to make into suitable dosage forms with adequate oral bioavailability [16]. Particles loaded with a poorly soluble drug can significantly increase the drug dissolution rate. The intestine has a special mechanism to absorb particles of certain sizes. The 100 nm particles showed a significantly higher uptake than larger particles [83, 84]. Although gene therapy has been extensively studied for treating genetic diseases and acquired diseases [85], the safety and efficiency of gene delivery have not been examined in depth. PLA-based micro and nano-particles have shown particular promise in improving protection from plasma enzymes, alternative routes of administration (e.g., nasal, oral, pulmonary, and mucosal), and prolonged gene delivery efficacy [86-89]. Cationic PEG-PLA nano-particles are one of the major delivery systems for the small interference RNA (siRNA) system. Systemic delivery of small interfering polo-like kinase 1 (siPlk1) by PEG-PLA nano-particles significantly suppressed tumor growth in an MDA-MB-435s (cancer cells) murine xenograft model [90]. siRNA encapsulated in PEG-PLA nano-particles were shown to have successfully entered the cells and resulted in remarkable gene-specific knockdown in the adult zebrafish heart [91]. PLA based nano-particles containing polyethyleneimine (PEI) on their surfaces were used for incorporating genes. PEO-PLA-PEI was also used for co-delivery of supercoiled minicircle (mc) DNA vectors and Dox. These nano carrier systems have the advantage of non-fouling oxazolines to confer biological stability, of PLA to provide hydrophobicity for Dox encapsulation and of bioreductive PEI to provide gene complexation. The dual delivery of mcDNA-Dox to B16F10 (Musmusculus skinmelanoma cell line; ATCC® CRL-6475™)-Luciferase tumor bearing mice resulted in significantly reduced tumor size and cancer cells' viability [92]. Recently, dual or multi drug delivery systems have shown great potential in the drug delivery field for cancer and gene therapy. Vaccinations have been highly successful for preventing many infectious diseases using micro-particles [93, 94]. New vaccines are focused on AIDS, hepatitis B, anthrax, SARS, and MERS. Many research groups are focused on developing micro-particle based single shot vaccines using PLA-based materials [95-98]. The T cell activation in response to antigen-encapsulated micro-particles has increased up to 100~1,000 fold more than antigens alone [99]. HIV Gag antigens (p24)-coated PLA nano-particles captured by monocyte-derived dendritic cells (MDDCs) from HIV-1 individuals stimulated MDDC maturation and increased HIV-specific CD8+ T-cell proliferation as compared with p24 alone [100]. PEG-PLA-PEG block copolymer nano-particles were evaluated for encapsulating the hepatitis B surface antigen (HBsAg) as an oral vaccine delivery system. HBsAg encapsulated copolymer and PLA nano-particles were used for adjuvanticity in generating immune stimulations after oral administration. PEG-PLA-PEG copolymer nano-particles exhibited effective levels of humoral immunity along with the mucosal (sIgA) and cellular immune response (TH1) [101]. In the study of the relationship between PLA-PEG particle size and efficacy of transport across the nasal mucosal, tetanus toxoid was encapsulated into PLA-PEG particles of different sizes (200 nm, 1.5 μm, 5 μm, and 10 μm) prepared by the w/o/w double emulsion solvent evaporation technique. The nasal bioavailability of tetanus toxoid encapsulated into 200 nm nano-particles was higher than into larger particles. PLA-PEG nano-particles and aluminum phosphate have been used as a potential adjuvant system using tetanus toxoid. The encapsulation efficiency was increased to nearly 90% in PLA-PEG nano-particles as compared to 55% in a conventional vaccine. PLA-PEG-aluminum (Al) and PLA-Al showed 80% and 50% survival rates, respectively, even at 180 days as compared to a 30% survival rate in the conventional tetanus vaccine [102]. Cyclosporine A (CyA) entrapped in PLA micro and nano-particles showed enhanced bioavailability and sustained release kinetics for extended periods of time [103]. The effects of the concentration of PLA, surfactant, and aqueous phase volume on the PLA microsphere size and nimesulide (a non-steroidal anti-inflammatory drug) encapsulation efficiency were studied using particles made by emulsion methods. A specific aqueous phase volume was selected for small size particles, because an increased volume resulted in microdroplet's coalescence [104]. The in vitro release of albumin from PLA micro-particles was sustained for one month after the particles were blended with PEG [105]. Insulin loaded PEG-PLA nano-particles provided a sustained release for more than two months. The burst release amount increased as PEG molecular weight or PEG content increased [106, 107]. Insulin-loaded PLA based particles had more specific and selective release at high pH conditions, which might increase the effect of the insulin in the blood stream (pH 7.4) [108]. One of the extensive drug delivery fields is cancer chemotherapy. Paclitaxel loaded particles have significantly enhanced anti-tumoral efficacy as compared with free drugs. Paclitaxel loaded PLA-PEG-PLA micro-particles showed 49.6% sustained release of paclitaxel within 1 month [109]. In vitro cytotoxicity testing in cancer cell lines revealed that the PLA-PEG nanoparticles compared with free paclitaxel exhibited similar cytotoxicity [110]. Gemcitabine hydrochloride (GEM) loaded PEG-PLA nanoparticles had zero-order release profiles. The particles increased antitumor effect compared with the free drug on different cancer cell lines and showed a significant improvement of cell interaction. Two xenograft murine models of human solid tumors were used for in vivo anticancer activity of the particles. GEM-PEG-PLA nano-particles significantly inhibited the tumor growth and the mice survival rate increased compared with the free drug [111]. Docetaxel and tamoxifen are potent drugs against breast cancer. There is an antagonistic problem when both drugs are used in combination because they have different metabolisms. Docetaxel and tamoxifen loaded TPGS-PLA showed a significant reduction of the drug antagonism in the MCF7 cell line [112]. Some representative published results on PLA/PLGA based particles as drug delivery systems are summarized in Table 5 [5, 16]. There are variable drugs and suitable methods applied for their encapsulation. The suitability of the method was evaluated by loading efficiency and pharmacokinetic release results. 4. Challenges and strategies Current micro and nano-particle production methods have been constrained by limitations of processes. The conventional techniques have several disadvantages, including the relatively high cost of particle production, the potential toxicity of solvents and reagents like stabilizers, emulsifiers, and other additives for forming particles, the use of a high energy mechanical mixer and homogenizer, which may be damaging to biological drugs such as proteins, peptides and macromolecules, the difficulty of reproducing biologically stable particles, and the low drug encapsulation efficiency [2, 7, 8, 17, 29, 146]. The high energy mechanical mixer and homogenizer generate high shearing forces. These high energy shearing forces can cause disorder which changes the natural structure of the macromolecules. Biopharmaceuticals, or protein drugs, may be denatured by exposure to the water/solvent interface or organic solvents. The reported processes are small laboratory scale or small test production scale under 1 g. Some of the large production scales have resulted in different particle size distributions increasing the process volume because the solvent evaporation rate may be different [30]. The most significant challenge is to understand the particle forming mechanism and encapsulation process [2, 8, 147]. Physico-chemical characterization of nano- and micro-particles have not been complete [2, 8, 148]. The nano-particles have several unique physico-chemical properties that can present difficulty in characterization. The nano-particles need higher cost characterization methods related to size, shape, surface charge, surface area and other physico-chemical properties [7, 149]. Incomplete characterization of the nano-particles may lead to an incomplete understanding of the correlation between nano-particle properties and various biological effects. The properties of the nano-particles may be easily changed by the surrounding environments, such as the blood stream, cell types, and physico-chemical environments (temperature, pH, pressure, volume, etc.). PLA-based nano-particles need to be analyzed both in dry or lyophilized form and in the test media, such as with or without serum based culture media for complete characterization [2, 7, 8, 17, 149]. One important consideration for making drug delivery systems is sterilization of PLA-based particles [7, 149]. Most useful sterilization methods cannot be applied to PLA based particles. Steam sterilization cannot be used with PLA based particles because high temperature and pressure can affect the particle that softens, melts, deforms, and undergoes hydrolysis. Heat sterilization exposes the particle to high temperature for long periods of time, which can destroy the PLA matrix structure and drug. Ethylene oxide (EO) is known as a polymer softener and plasticizer. The residual EO gas causes mutagenic, carcinogenic, and allergenic effects. Gamma radiation can breakdown polymer chains, resulting in decreased molecular weight and increased biodegradation rates, significantly altering drug release profiles. Therefore, GMP grade production of particles has to be done by aseptic processing. It is very effective for preventing contamination of particles but an expensive technique for manufacturing PLA based particles. It requires clean room control and the use of GMP protocols. Various particle production techniques have been investigated for improving the encapsulation efficiency of drugs, optimizing the scale up process for mass production, and enhancing the reproducibility of the methods [2, 147, 149]. Table 6 lists a summary of challenges associated with developing PLA based particles as drug delivery applications [7-9, 150]. One of the most useful strategies to overcoming certain challenges is surface modification of PLA based micro and nano-particles for improving the stability of the particles. Surface modification is important for escaping the immune system when administrating particles to the bloodstream [29, 151]. Similarly, other strategies have been used to make a hydrophilic cloud around the particles to reduce their uptake by RES systems. These strategies comprise surface modifications of particles with Tween 80, PEG or PEO, poloxamers and poloxamines, polysorbate 80, TPGS, functional amino acids and polysaccharides [29, 124]. The most preferred surface modification is the adsorption or grafting of PEG (known as PEGylation) to the surface of particles. Grafting of PEG and PEG-containing copolymers onto the surface of particles augmented the blood circulation half-life. Increasing the molecular weight of the PEG chains has been shown to reduce opsonization of particles and improve retention in the circulation [29, 110, 151-153]. In addition, PEG may have good interactions with blood components. The other option is a copolymerization with PGA/PCL, PEG, and other polymers. PLA-copolymers have worked well as biocompatible polymer particles for drug delivery systems [2, 9, 29]. There is a huge amount of knowledge scattered around the world. The data on each PLA-based micro and nano-particles are unique in that the fabrication method, the drug used, and the efficacy testing methods are all different. This makes it difficult to compare properties of PLA particles, and PLGA particles for that matter, to find the right formulation for specific applications. It is time to assemble a data bank that presents detailed information correlating PLA particle properties and their in vivo functions. Such a data bank will propel more systematic development of future PLA micro and nano-particles that can be developed for specific in vivo applications. Acknowledgments This work was supported by the Showalter Research Trust Fund and the National Institute of Health through CA129287 and GM095879. Figure 1 Schematic description of a PLA micro and nano-particle and its preparation techniques. Figure 2 Schematic diagram of emulsion-based methods for preparation of polymer particles. Figure 3 SEM photos of prilocaine-loaded micro-particles with (a) Resomer R202 (PLA, average MW 10,000-18,000), (b) Resomer R203S (PLA, average MW 18,000-28,000), (c) Resomer R207 (PLA, average MW 209,000) polymers. Images of the same batches recuperated after 96 hr. of drug release studies are also reported: (d) Resomer R202, (e) Resomer R203S, (f) Resomer R207 [39]. Figure 4 Drug release profiles of prilocaine-loaded micro-particles in pH 7.4 phosphate buffer. The total amount was calculated from the average value of encapsulation efficiency percentage obtained by the direct and the indirect method [39]. Figure 5 Schematic description of the salting out method for preparation of polymer particles. Figure 6 Schematic description of nanoprecipitation for preparation of polymer particles. Figure 7 (A) Field emission SEM images of DTX-loaded TPGS-PLA/NPs, polydopamine (pD)-TPGS-PLA/NPs and galactosamine (Gal)-pD-TPGS-PLA/NPs; local images in the box are shown in the lower panel. (B) TEM images of DTX-loaded TPGS-PLA/NPs, pD-TPGS-PLA/NPs and Gal-pD-TPGS-PLA/NPs [49]. Figure 8 Schematic description of dialysis for preparation of polymer particles. Figure 9 Schematic description of spray drying for preparation of polymer particles. Figure 10 PLA micro-spheres made of PEG-distearate obtained by the spray drying technique. (A) PLA 1% (w/v) and PEG-distearate 10% (w/v), (B) PLA 1% (w/v) and PEG-distearate 1% (w/v), (C) PLA 3% (w/v) and PEG-distearate 1% (w/v), (D) PLA 5% (w/v) and PEG-distearate 1% (w/v) [64]. Figure 11 Schematic description of in situ forming micro-particle for preparation of polymer particles. Figure 12 Leuprolide release and SEM images of in situ forming micro-particles (PLGA (50:50) Resomer® (RG 503H, average MW 24,000~38,000), standard formulation) [67]. Figure 13 Leuprolide release and SEM images of in situ forming micro-particles (PLA Resomer® (R 202H, average MW 10,000~18,000), 10% or 15%, w/w, drug loading and 30% or 40%, w/w, polymer concentration) [67]. Figure 14 Schematic description of melting technique for preparation of polymer particles. Figure 15 Schematic description of supercritical fluids technique. Figure 16 Typical morphological features by scanning electron microscopy of the particles produced in the assays A1–A11 [73]. Figure 17 Schematic description of microfluidic technique for preparation of polymer particles. Figure 18 SEM images of micro-particles prepared via microfluidics with 41 μm (A) and 11 μm (B). Drug-release profiles from monodispersed micro-particles prepared with microfluidic devices and polydisperse micro-particles prepared using the conventional single emulsion technique (right) [78]. Figure 19 Schematic of template/mold method process and fabricated PLA micro-particles. Figure 20 Comparison of release profiles of risperidone (RIS), methylprednisolone acetate (MPA), and paclitaxel (PTX) from 85:15 PLGA (inherent viscosity 0.55~0.75 dL/g) micro-particles (left). Comparison of release profiles of RIS-loaded micro-particles prepared using hydrogel template and emulsion methods (right) [80]. Table 1 Sustained release depot formulations based on PLGA/PLA currently available for clinical use. Product Name Active Ingredient Company Application Formulation Administration Lupron®Depot (1989) Leuprolide acetate TAP Prostate cancer, endometriosis Microparticle I.M. Zoladex® (1989) Goserelin acetate AstraZeneca Pharmaceuticals Prostate cancer, Endometriosis Implant S.C. Sandostatin LAR® Depot (1998) Octreotide acetate Novartis Acromegaly Microparticle S.C. Atridox® (1998) Doxycycline hyclate Zila, Inc. Chronic adult periodontitis In situ forming Local delivery Nutropin®Depot (1999) Growth hormone Genetech Pediatric growth hormone deficiency Microparticle I.M. Surodex® (1999) (Approved in China, Singapore and several other countries) Dexamethasone Allergan Post surgical inflammation after cataract surgery Implant Local delivery Trelstar TM Depot (2000 and 2010) Triptorelin pamoate Pfizer Prostate cancer Microparticle I.M. Somatuline® Depot (2000) Lanreotide Ipsen Acromegaly Microparticle S.C. Suprefact® Depot (2000) (Approved in Sweden, Netherlands and several other countries) Buserelin acetate Sanofi-Aventis Prostate cancer Implant S.C. Arestin® (2001) Minocycline Orapharma Periodontal disease Microparticle Local delivery Suprecur® MP (2002) (Approved in Japan) Buserelin acetate Sanofi-Aventis Prostate cancer Microparticle S.C. Risperidal® ConstaTM (2003) Risperidone Johnson &Johnson Antipsychotic Microparticle I.M. Eligard® (2004) Leuprolide acetate injectable TOLMAR Pharmaceuticals Inc. Prostate cancer In situ forming implant S.C. Somatuline LA® (2004) (Approved in UK) Lanreotide acetate Ipsen Acromegaly Microparticle I.M. Decapeptyl® (2006) (Approved in EU) Triptorelin pamoate Ipsen Prostate cancer Microparticle I.M. Vivitrol® (2006) Naltrexone Alkermes Alcohol abuse Microparticles I.M. Decapeptyl® (2006) (Approved in EU) Triptorelin pamoate Ipsen Prostate cancer Microparticle I.M. Ozurdex® (2009) Dexamethasone Allergan Diabetic macular edema Implant Local delivery Lutrate Depot® (2010) Leuprolide acetate G P Pharm Type 2 diabetes Microparticle I.M. Bydureon® (2012) Exenatide Amylin Pharmaceuticals Inc. Type 2 diabetes Microparticle S.C. PropelTM (2012) Mometasone furoate Intersect ENT Post surgical Inflammation Implant Local delivery Lupaneta Pack (2012) Leuprolide acetate Norethindrone acetate AbbVie endocrine Prostate cancer Microparticle I.M. Pamorelin LA® (2012) (Approved in EU) Triptorelin pamoate Galenica Prostate cancer Microparticle I.M. (Information from Reference [7] was included in the table). Table 2 The list of cGMP grade PLA and PLGA manufacturers. Company Trade name Corbion Purasorb® Lactel Lactel® Alkermes Medisorb® Evonik Resomer® PCAS Expansorb® Table 3 Advantages and limitations of PLA micro and nano-particles [8]. Advantages Limitations Micro-particles Subcutaneous injections Intramuscular injections Controlled release Reproducible processes Unintended toxic side effect due to high initial burst Wasteful use of expensive drugs due to initial burst release Nano-particles Direct injection to the blood Potentially improved vaccine responses Non-specific uptake by reticuloendothelial system (RES) systems Potential immunotoxicity Table 4 Results of the factorial design experiment used to study the effects of pressure, concentration and flow rate of polymer solution on PLA particle mean diameters and DCM residual levels [73]. Assay Pressure (MPa) Concentration of polymer solution (%) Flow rate of polymer solution (mL/min) Mean diameter (μm) Residual DCM (ppm) A1 8 0.5 0.5 6 ± 1 790 A2 8 0.5 2.5 21 ± 1 1,130 A3 16 0.5 0.5 5 ± 0 1,280 A4 16 0.5 2.5 18 ± 0 6,490 A5 8 1.5 0.5 7 ± 0 <600 A6 8 1.5 2.5 12 ± 0 5,980 A7 16 1.5 0.5 7 ± 0 2,040 A8 16 1.5 2.5 11 ± 0 2,630 A9 12 1.0 1.5 10 ± 0 3,400 A10 12 1.0 1.5 6 ± 0 4,160 A11 12 1.0 1.5 9 ± 0 4,670 Table 5 Investigations on PLA/PLGA particles as drug delivery systems. Material Drug Method Result Ref. PLA Irinotecan hydrochloride Single emulsion Smooth surface and less initial burst release [113] PLA Nimesulide Single emulsion Initial burst followed by sustained release [114] PLA-PEG Tetanus toxoid Single emulsion Enhanced transport across the rat nasal mucosa [115] PLA Vanillin Single emulsion Slow and sustained release, stable particles over 3 months Inferior free radical scavenging activity than free vanillin [116] PLA BSA Double emulsion Encapsulation efficiency up to 71.6% [23] PLA-PEG Hemoglobin Double emulsion Less macrophage uptake [117] PLA Protein-C Double emulsion Release of protein C seems to increase with the hydrophilic character of PLA [118] PLA- TPGS BSA Double emulsion longer blood circulation time than free drug [119] PLA Neurotoxin-1 Double emulsion Brain delivery of NT-1 enhanced [120] PLA Triclosan Double emulsion High encapsulation efficiency [121] PLA Oridonin Modified spontaneous emulsion solvent diffusion Slow drug release up to 72hrs. [22] PLA-PEG Lactoferrin Modified double emulsion/solvent evaporation Increased uptake by bEnd.3 cells [122] PLA-PEG Zidovudine Solvent evaporation Less phagocytosis [21] PLA- TPGS Paclitaxel Modified solvent extraction/evaporation Initial burst followed by sustained release [24] PLA- mPEG Salting out Less interaction with leukocytes [123] PLA- PEG-PLA Savoxepine Salting out Controlled drug release up to 1 week [124] PLA BSA Salting out/coacervation High encapsulation efficiency and acceptable burst release [125] PLA Cloricromene Nanoprecipitation Faster dissolution than free drug [26] PLA Tamoxifen Nanoprecipitation Significant therapeutic efficacy with reduced side effects [126] PLA- Pluronic Stevioside Nanoprecipitation High potential safe and effective [127] Chitosan- PLA Anthraquinone Nanoprecipitation Continuous and sustained release, pH dependent release profiles [128] PLA- pluronic Insulin Dialysis/ nanoprecipitation Good control over blood glucose concentration [129] PEG Gene delivery Dialysis Improved transfection activity [25] PLA HIV p24 protein Dialysis Induced mucosal antibody production [130] PLA Progesterone Theophylline Vitamin D3 Spray drying Alternative method [56, 59] PLA Piroxicam Spray drying Small initial burst release [60] PLA Ketotifen Spray drying Released in plasma between 336 and 384 hr, and the mean residence time increased between 30 and 70 times compared to solution treatment [131] PLA/ PLGA Implant In situ forming micro- particles Lower myotoxicity [27, 132] PLA Ivermectin In situ forming gel Slow in vitro release and 80% cumulative release in 80 days 110-120 days maintained effective gel in vivo pharmacokinetic results [133] PLA/ PLGA Diltiazem hydrochloride, Buserelin acetate In situ forming micro- particles Significantly reduced burst effect [134] PLA Bupivacaine hydrochloride In situ forming micro- particles Significantly slower release compared to conventional particle formulation [135] HA-PLA Methylprednisolone In situ forming micro gel Entrap a hydrophobic drug and prolong release profile [136] PLA Steroid norethisterone Melting Zero order release [137, 138] PLA Naltrexone Melting Long effective blocking action to morphine [139] PLA Prednisolone Melting Sustained release over 30 days [140] PLA/F68 Dexamethasone Hot melt extrusion No negative influence in the body [141] PLA Hyoscine Butylbromide Indomethacin Piroxicam Thymopentin Supercritical fluids technique High encapsulation efficiency [142] PLA Enzyme Insulin Calcitonin Supercritical fluids technique Low loss of enzyme activity, retention of protein activity [143] PLA Rifampin Gentamycin Naltrexone Supercritical fluids technique Lower initial burst [144] PLA Paclitaxel Microfluidic technique Monodisperse paclitaxel particles [145] PLA/ PLGA Risperidone Paclitaxel Template method Low initial burst and controlled drug release for 1 month [80] Table 6 Various challenges for making PLA particles for drug delivery [7]. General issues ○ Low drug loading capacity and low encapsulation efficiency ○ Low reproducibility between batches ○ Poor scalability of manufacturing process from laboratory to large scale production ○ Lack of compendial method for measurement of drug release rates ○ High initial burst release with incomplete release of the drug ○ Presence of residual organic solvents Physico-chemical characterization ○ Particle heterogeneity in shape and size ○ Unknown correlation between particle properties and in vivo behavior ○ Lack of reference particles for validation of tools and techniques ○ Lack of standardized test protocols. ○ Difficulty in characterizing residual solvents affecting particle properties Other issues ○ Difficulty in terminal sterilization of particles ○ No accelerated testing methods for long term stability studies ○ Little information on particle formulations for production of generic products This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. References [1] Freiberg S Zhu XX Polymer microspheres for controlled drug release International Journal of Pharmaceutics 2004 282 1 18 15336378 [2] Kamaly N Xiao Z Valencia PM Radovic-Moreno AF Farokhzad OC Targeted polymeric therapeutic nanoparticles: design, development and clinical translation Chemical Society reviews 2012 41 2971 3010 22388185 [3] Stloukal P Kucharczyk P Sedlarik V Bazant P Koutny M Low molecular weight poly(lactic acid) microparticles for controlled release of the herbicide metazachlor: preparation, morphology, and release kinetics Journal of agricultural and food chemistry 2012 60 4111 4119 22480233 [4] Thomasin C Merkle HP Gander B Drug microencapsulation by PLA/PLGA coacervation in the light of thermodynamics. 2. Parameters determining microsphere formation J Pharm Sci 1998 87 269 275 9523977 [5] Lassalle V Ferreira ML PLA nano- and microparticles for drug delivery: an overview of the methods of preparation Macromol Biosci 2007 7 767 783 17541922 [6] Allouche J Brayner R Fiévet F Coradin T Synthesis of Organic and Bioorganic Nanoparticles: An Overview of the Preparation Methods Nanomaterials: A Danger or a Promise? 2013 27 74 Springer London [7] Kumar G Shafiq N Malhotra S Drug-loaded PLGA nanoparticles for oral administration: fundamental issues and challenges ahead Critical reviews in therapeutic drug carrier systems 2012 29 149 182 22475089 [8] Mitragotri S Burke PA Langer R Overcoming the challenges in administering biopharmaceuticals: formulation and delivery strategies Nat Rev Drug Discov 2014 13 655 672 25103255 [9] Madhavan Nampoothiri K Nair NR John RP An overview of the recent developments in polylactide (PLA) research Bioresource Technology 2010 101 8493 8501 20630747 [10] Bala I Hariharan S Kumar MN PLGA nanoparticles in drug delivery: the state of the art Critical reviews in therapeutic drug carrier systems 2004 21 387 422 15719481 [11] Soppimath KS Aminabhavi TM Kulkarni AR Rudzinski WE Biodegradable polymeric nanoparticles as drug delivery devices Journal of Controlled Release 2001 70 1 20 11166403 [12] Anderson JM Shive MS Biodegradation and biocompatibility of PLA and PLGA microspheres Advanced Drug Delivery Reviews 2012 64 Supplement 72 82 [13] Danhier F Ansorena E Silva JM Coco R Le Breton A Preat V PLGA-based nanoparticles: an overview of biomedical applications J Control Release 2012 161 505 522 22353619 [14] Pagels RF Prud'homme RK Polymeric nanoparticles and microparticles for the delivery of peptides, biologics, and soluble therapeutics Journal of Controlled Release 2015 219 519 535 26359125 [15] Zhao C-X Multiphase flow microfluidics for the production of single or multiple emulsions for drug delivery Advanced Drug Delivery Reviews 2013 65 1420 1446 23770061 [16] Coelho JF Ferreira PC Alves P Cordeiro R Fonseca AC Góis JR Gil MH Drug delivery systems: Advanced technologies potentially applicable in personalized treatments The EPMA Journal 2010 1 164 209 23199049 [17] Wischke C Schwendeman SP Principles of encapsulating hydrophobic drugs in PLA/PLGA microparticles Int J Pharm 2008 364 298 327 18621492 [18] Freitas S Merkle HP Gander B Microencapsulation by solvent extraction/evaporation: reviewing the state of the art of microsphere preparation process technology Journal of Controlled Release 2005 102 313 332 15653154 [19] Rao JP Geckeler KE Polymer nanoparticles: Preparation techniques and size-control parameters Progress in Polymer Science 2011 36 887 913 [20] Lin YM Zhang A Rippon HJ Bismarck A Bishop AE Tissue engineering of lung: the effect of extracellular matrix on the differentiation of embryonic stem cells to pneumocytes Tissue engineering. Part A 2010 16 1515 1526 20001250 [21] Mainardes RM Gremiao MP Brunetti IL da Fonseca LM Khalil NM Zidovudine-loaded PLA and PLA-PEG blend nanoparticles: influence of polymer type on phagocytic uptake by polymorphonuclear cells J Pharm Sci 2009 98 257 267 18425813 [22] Xing J Zhang D Tan T Studies on the oridonin-loaded poly(d,l-lactic acid) nanoparticles in vitro and in vivo International Journal of Biological Macromolecules 2007 40 153 158 16901536 [23] Gao H Wang YN Fan YG Ma JB Synthesis of a biodegradable tadpole-shaped polymer via the coupling reaction of polylactide onto mono(6-(2-aminoethyl)amino-6-deoxy)-beta-cyclodextrin and its properties as the new carrier of protein delivery system J Control Release 2005 107 158 173 16095747 [24] Zhang Z Feng S-S Nanoparticles of poly(lactide)/vitamin E TPGS copolymer for cancer chemotherapy: Synthesis, formulation, characterization and in vitro drug release Biomaterials 2006 27 262 270 16024075 [25] Ataman-Onal Y Munier S Ganee A Terrat C Durand PY Battail N Martinon F Le Grand R Charles MH Delair T Verrier B Surfactant-free anionic PLA nanoparticles coated with HIV-1 p24 protein induced enhanced cellular and humoral immune responses in various animal models J Control Release 2006 112 175 185 16563545 [26] Leo E Brina B Forni F Vandelli MA In vitro evaluation of PLA nanoparticles containing a lipophilic drug in water-soluble or insoluble form International Journal of Pharmaceutics 2004 278 133 141 15158956 [27] Kranz H Brazeau GA Napaporn J Martin RL Millard W Bodmeier R Myotoxicity studies of injectable biodegradable in-situ forming drug delivery systems International Journal of Pharmaceutics 2001 212 11 18 11165816 [28] Makadia HK Siegel SJ Poly Lactic-co-Glycolic Acid (PLGA) as Biodegradable Controlled Drug Delivery Carrier Polymers 2011 3 1377 1397 22577513 [29] Sadat Tabatabaei Mirakabad F Nejati-Koshki K Akbarzadeh A Yamchi MR Milani M Zarghami N Zeighamian V Rahimzadeh A Alimohammadi S Hanifehpour Y Joo SW PLGA-based nanoparticles as cancer drug delivery systems Asian Pacific journal of cancer prevention : APJCP 2014 15 517 535 24568455 [30] Sah E Sah H Recent Trends in Preparation of Poly(lactide-co-glycolide) Nanoparticles by Mixing Polymeric Organic Solution with Antisolvent Journal of Nanomaterials 2015 2015 22 [31] Quintanar-Guerrero D Allemann E Fessi H Doelker E Preparation techniques and mechanisms of formation of biodegradable nanoparticles from preformed polymers Drug development and industrial pharmacy 1998 24 1113 1128 9876569 [32] Martín-Sabroso C Fraguas-Sánchez AI Aparicio-Blanco J Cano-Abad MF Torres-Suárez AI Critical attributes of formulation and of elaboration process of PLGA-protein microparticles International Journal of Pharmaceutics 2015 480 27 36 25578370 [33] Li M Rouaud O Poncelet D Microencapsulation by solvent evaporation: State of the art for process engineering approaches International Journal of Pharmaceutics 2008 363 26 39 18706988 [34] Tan MXL Danquah MK Drug and Protein Encapsulation by Emulsification: Technology Enhancement Using Foam Formulations Chemical Engineering & Technology 2012 35 618 626 [35] Jain RA The manufacturing techniques of various drug loaded biodegradable poly(lactide-co-glycolide) (PLGA) devices Biomaterials 2000 21 2475 2490 11055295 [36] Lamprecht A Ubrich N Hombreiro Pérez M Lehr CM Hoffman M Maincent P Biodegradable monodispersed nanoparticles prepared by pressure homogenization-emulsification International Journal of Pharmaceutics 1999 184 97 105 10425355 [37] Zambaux MF Bonneaux F Gref R Maincent P Dellacherie E Alonso MJ Labrude P Vigneron C Influence of experimental parameters on the characteristics of poly(lactic acid) nanoparticles prepared by a double emulsion method Journal of Controlled Release 1998 50 31 40 9685870 [38] Lu JM Wang X Marin-Muller C Wang H Lin PH Yao Q Chen C Current advances in research and clinical applications of PLGA-based nanotechnology Expert review of molecular diagnostics 2009 9 325 341 19435455 [39] Bragagni M Beneitez C Martín C de la Ossa DHP Mura PA Gil-Alegre ME Selection of PLA polymers for the development of injectable prilocaine controlled release microparticles: Usefulness of thermal analysis International Journal of Pharmaceutics 2013 441 468 475 23164705 [40] Jung T Kamm W Breitenbach A Kaiserling E Xiao JX Kissel T Biodegradable nanoparticles for oral delivery of peptides: is there a role for polymers to affect mucosal uptake? European Journal of Pharmaceutics and Biopharmaceutics 2000 50 147 160 10840198 [41] Lambert G Fattal E Couvreur P Nanoparticulate systems for the delivery of antisense oligonucleotides Advanced Drug Delivery Reviews 2001 47 99 112 11251248 [42] Ghaderi R Artursson P Carlfors J A new method for preparing biodegradable microparticles and entrapment of hydrocortisone in dl-PLG microparticles using supercritical fluids European Journal of Pharmaceutical Sciences 2000 10 1 9 10699378 [43] Allémann E Gurny R Doelker E Preparation of aqueous polymeric nanodispersions by a reversible salting-out process: influence of process parameters on particle size International Journal of Pharmaceutics 1992 87 247 253 [44] Govender T Stolnik S Garnett MC Illum L Davis SS PLGA nanoparticles prepared by nanoprecipitation: drug loading and release studies of a water soluble drug Journal of Controlled Release 1999 57 171 185 9971898 [45] Barichello JM Morishita M Takayama K Nagai T Absorption of insulin from Pluronic F-127 gels following subcutaneous administration in rats International Journal of Pharmaceutics 1999 184 189 198 10387948 [46] Francois G Katz JL Nanoparticles and nanocapsules created using the Ouzo effect: spontaneous emulisification as an alternative to ultrasonic and high-shear devices Chemphyschem : a European journal of chemical physics and physical chemistry 2005 6 209 216 15751338 [47] Vitale SA Katz JL Liquid Droplet Dispersions Formed by Homogeneous Liquid−Liquid Nucleation:  “The Ouzo Effect” Langmuir 2003 19 4105 4110 [48] Anton N Benoit JP Saulnier P Design and production of nanoparticles formulated from nano-emulsion templates-a review J Control Release 2008 128 185 199 18374443 [49] Zhu D Tao W Zhang H Liu G Wang T Zhang L Zeng X Mei L Docetaxel (DTX)-loaded polydopamine-modified TPGS-PLA nanoparticles as a targeted drug delivery system for the treatment of liver cancer Acta Biomaterialia 2016 30 144 154 26602819 [50] Kostag M Köhler S Liebert T Heinze T Pure Cellulose Nanoparticles from Trimethylsilyl Cellulose Macromolecular Symposia 2010 294 96 106 [51] Chronopoulou L Fratoddi I Palocci C Venditti I Russo MV Osmosis based method drives the self-assembly of polymeric chains into micro- and nanostructures Langmuir 2009 25 11940 11946 19572495 [52] Na K Lee KH Lee DH Bae YH Biodegradable thermo-sensitive nanoparticles from poly(l-lactic acid)/poly(ethylene glycol) alternating multi-block copolymer for potential anti-cancer drug carrier European Journal of Pharmaceutical Sciences 2006 27 115 122 16253487 [53] Hornig S Heinze T Efficient approach to design stable water-dispersible nanoparticles of hydrophobic cellulose esters Biomacromolecules 2008 9 1487 1492 18393524 [54] Jeong Y-I Cho C-S Kim S-H Ko K-S Kim S-I Shim Y-H Nah J-W Preparation of poly(DL-lactide-co-glycolide) nanoparticles without surfactant Journal of Applied Polymer Science 2001 80 2228 2236 [55] Liu W Chen XD Selomulya C On the spray drying of uniform functional microparticles Particuology 2015 22 1 12 [56] Bodmeier R Chen HG Preparation of biodegradable poly(+/−)lactide microparticles using a spray-drying technique J Pharm Pharmacol 1988 40 754 757 2907552 [57] Sosnik A Seremeta KP Advantages and challenges of the spray-drying technology for the production of pure drug particles and drug-loaded polymeric carriers Advances in Colloid and Interface Science 2015 223 40 54 26043877 [58] Chavalitkul J Likittanaprasong N Seansala P Suttiruengwong S Seadan M Application of Ultrasonic Atomization for Biopolymer Particles and Tube Fabrication Energy Procedia 2014 56 458 465 [59] Pavanetto F Genta I Giunchedi P Conti B Evaluation of spray drying as a method for polylactide and polylactide-co-glycolide microsphere preparation J Microencapsul 1993 10 487 497 8263677 [60] Wagenaar BW Müller BW Piroxicam release from spray-dried biodegradable microspheres Biomaterials 1994 15 49 54 8161657 [61] Benelli P Conti B Genta I Costantini M Montanari L Clonazepam microencapsulation in poly-D,L-lactide-co-glycolide microspheres J Microencapsul 1998 15 431 443 9651865 [62] Mu L Feng SS Fabrication, characterization and in vitro release of paclitaxel (Taxol) loaded poly (lactic-co-glycolic acid) microspheres prepared by spray drying technique with lipid/cholesterol emulsifiers J Control Release 2001 76 239 254 11578739 [63] Mu L Teo M-M Ning H-Z Tan C-S Feng S-S Novel powder formulations for controlled delivery of poorly soluble anticancer drug: Application and investigation of TPGS and PEG in spray-dried particulate system Journal of Controlled Release 2005 103 565 575 15820404 [64] Lacasse FX Hildgen P McMullen JN Surface and morphology of spray-dried pegylated PLA microspheres International Journal of Pharmaceutics 1998 174 101 109 [65] Royals MA Fujita SM Yewey GL Rodriguez J Schultheiss PC Dunn RL Biocompatibility of a biodegradable in situ forming implant system in rhesus monkeys Journal of biomedical materials research 1999 45 231 239 10397981 [66] Jain RA Rhodes CT Railkar AM Malick AW Shah NH Comparison of various injectable protein-loaded biodegradable poly(lactide-co-glycolide) (PLGA) devices: in-situ-formed implant versus in-situ-formed microspheres versus isolated microspheres Pharmaceutical development and technology 2000 5 201 207 10810750 [67] Luan X Bodmeier R In situ forming microparticle system for controlled delivery of leuprolide acetate: Influence of the formulation and processing parameters European Journal of Pharmaceutical Sciences 2006 27 143 149 16243496 [68] Wichert B Rohdewald P A new method for the preparation of drug containing polylactic acid microparticles without using organic solvents Journal of Controlled Release 1990 14 269 283 [69] Pinto Reis C Neufeld RJ Ribeiro AJ Veiga F Nanoencapsulation I. Methods for preparation of drug-loaded polymeric nanoparticles, Nanomedicine: Nanotechnology Biology and Medicine 2006 2 8 21 [70] Davies OR Lewis AL Whitaker MJ Tai H Shakesheff KM Howdle SM Applications of supercritical CO2 in the fabrication of polymer systems for drug delivery and tissue engineering Advanced Drug Delivery Reviews 2008 60 373 387 18069079 [71] Kim J-H Paxton TE Tomasko DL Microencapsulation of Naproxen Using Rapid Expansion of Supercritical Solutions Biotechnology Progress 1996 12 650 661 [72] Yeo Y Baek N Park K Microencapsulation methods for delivery of protein drugs Biotechnology and Bioprocess Engineering 6 213 230 [73] Sacchetin PSC Morales AR Moraes ÂM Rosa P.d.T.V.e. Formation of PLA particles incorporating 17α-methyltestosterone by supercritical fluid technology The Journal of Supercritical Fluids 2013 77 52 62 [74] Watanabe T Ono T Kimura Y Continuous fabrication of monodisperse polylactide microspheres by droplet-to-particle technology using microfluidic emulsification and emulsion-solvent diffusion Soft Matter 2011 7 9894 9897 [75] Liao CY Su YC Formation of biodegradable microcapsules utilizing 3D, selectively surface-modified PDMS microfluidic devices Biomedical microdevices 2010 12 125 133 19851872 [76] Jyothi NV Prasanna PM Sakarkar SN Prabha KS Ramaiah PS Srawan GY Microencapsulation techniques, factors influencing encapsulation efficiency J Microencapsul 2010 27 187 197 20406093 [77] Vladisavljević GT Shahmohamadi H Das DB Ekanem EE Tauanov Z Sharma L Glass capillary microfluidics for production of monodispersed poly (dl-lactic acid) and polycaprolactone microparticles: Experiments and numerical simulations Journal of Colloid and Interface Science 2014 418 163 170 24461831 [78] Xu Q Hashimoto M Dang TT Hoare T Kohane DS Whitesides GM Langer R Anderson DG Preparation of monodisperse biodegradable polymer microparticles using a microfluidic flow-focusing device for controlled drug delivery, Small (Weinheim an der Bergstrasse Germany) 2009 5 1575 1581 [79] Acharya G Shin CS McDermott M Mishra H Park H Kwon IC Park K The hydrogel template method for fabrication of homogeneous nano/microparticles Journal of Controlled Release 2010 141 314 319 19822178 [80] Lu Y Sturek M Park K Microparticles produced by the hydrogel template method for sustained drug delivery Int J Pharm 2014 461 258 269 24333903 [81] Park K Skidmore SM Fultz L Yun Y Lee B.k. Device for large-scale microparticle production and method of using the same patent pending [82] Takeuchi H Yamamoto H Kawashima Y Mucoadhesive nanoparticulate systems for peptide drug delivery Advanced Drug Delivery Reviews 2001 47 39 54 11251244 [83] Hsu CH Cui Z Mumper RJ Jay M Preparation and characterization of novel coenzyme Q10 nanoparticles engineered from microemulsion precursors AAPS PharmSciTech 2003 4 E32 14621964 [84] Murakami H Kobayashi M Takeuchi H Kawashima Y Utilization of poly(dl-lactide-co-glycolide) nanoparticles for preparation of mini-depot tablets by direct compression Journal of Controlled Release 2000 67 29 36 10773326 [85] Dwarki VJ Belloni P Nijjar T Smith J Couto L Rabier M Clift S Berns A Cohen LK Gene therapy for hemophilia A: production of therapeutic levels of human factor VIII in vivo in mice Proceedings of the National Academy of Sciences of the United States of America 1995 92 1023 1027 7862626 [86] Chen J Tian B Yin X Zhang Y Hu D Hu Z Liu M Pan Y Zhao J Li H Hou C Wang J Zhang Y Preparation, characterization and transfection efficiency of cationic PEGylated PLA nanoparticles as gene delivery systems Journal of Biotechnology 2007 130 107 113 17467097 [87] Munier S Messai I Delair T Verrier B Ataman-Önal Y Cationic PLA nanoparticles for DNA delivery: Comparison of three surface polycations for DNA binding, protection and transfection properties Colloids and Surfaces B: Biointerfaces 2005 43 163 173 15941652 [88] Chen CK Jones CH Mistriotis P Yu Y Ma X Ravikrishnan A Jiang M Andreadis ST Pfeifer BA Cheng C Poly(ethylene glycol)-block-cationic polylactide nanocomplexes of differing charge density for gene delivery Biomaterials 2013 34 9688 9699 24034497 [89] Basarkar A Devineni D Palaniappan R Singh J Preparation, characterization, cytotoxicity and transfection efficiency of poly(dl-lactide-co-glycolide) and poly(dl-lactic acid) cationic nanoparticles for controlled delivery of plasmid DNA International Journal of Pharmaceutics 2007 343 247 254 17611054 [90] Yang X-Z Dou S Sun T-M Mao C-Q Wang H-X Wang J Systemic delivery of siRNA with cationic lipid assisted PEG-PLA nanoparticles for cancer therapy Journal of Controlled Release 2011 156 203 211 21839126 [91] Gu G Hu Q Feng X Gao X Menglin J Kang T Jiang D Song Q Chen H Chen J PEG-PLA nanoparticles modified with APTEDB peptide for enhanced anti-angiogenic and anti-glioma therapy Biomaterials 2014 35 8215 8226 24974009 [92] Gaspar VM Baril P Costa EC de Melo-Diogo D Foucher F Queiroz JA Sousa F Pichon C Correia IJ Bioreducible poly(2-ethyl-2-oxazoline)-PLA-PEI-SS triblock copolymer micelles for co-delivery of DNA minicircles and Doxorubicin J Control Release 2015 213 175 191 26184050 [93] Yu XJ Luo C Lin JC Hao P He YY Guo ZM Qin L Su J Liu BS Huang Y Nan P Li CS Xiong B Luo XM Zhao GP Pei G Chen KX Shen X Shen JH Zou JP He WZ Shi TL Zhong Y Jiang HL Li YX Putative hAPN receptor binding sites in SARS_CoV spike protein Acta pharmacologica Sinica 2003 24 481 488 12791172 [94] Thermet A Rollier C Zoulim F Trepo C Cova L Progress in DNA vaccine for prophylaxis and therapy of hepatitis B Vaccine 2003 21 659 662 12531335 [95] Liu Y Chen X Wang L Yang T Yuan Q Ma G Surface charge of PLA microparticles in regulation of antigen loading, macrophage phagocytosis and activation, and immune effects in vitro Particuology 2014 17 74 80 [96] Gupta RK Goswami DG Singh RR Surolia A Panda AK Soybean agglutinin coated PLA particles entrapping candidate vaccines induces enhanced primary and sustained secondary antibody response from single point immunization European Journal of Pharmaceutical Sciences 2012 45 282 295 22155747 [97] Johansen P Moon L Tamber H Merkle HP Gander B Sesardic D Immunogenicity of single-dose diphtheria vaccines based on PLA/PLGA microspheres in guinea pigs Vaccine 1999 18 209 215 10506644 [98] Guillon C Mayol K Terrat C Compagnon C Primard C Charles M-H Delair T Munier S Verrier B Formulation of HIV-1 Tat and p24 antigens by PLA nanoparticles or MF59 impacts the breadth, but not the magnitude, of serum and faecal antibody responses in rabbits Vaccine 2007 25 7491 7501 17904700 [99] Walsh MC Banas JA Mudzinski SP Preissler MT Graziano RF Gosselin EJ A two-component modular approach for enhancing T-cell activation utilizing a unique anti-FcgammaRI-streptavidin construct and microspheres coated with biotinylated-antigen Biomolecular engineering 2003 20 21 33 12485681 [100] Climent N Munier S Piqué N García F Pavot V Primard C Casanova V Gatell JM Verrier B Gallart T Loading dendritic cells with PLA-p24 nanoparticles or MVA expressing HIV genes induces HIV-1-specific T cell responses Vaccine 2014 32 6266 6276 25240755 [101] Jain AK Goyal AK Mishra N Vaidya B Mangal S Vyas SP PEG–PLA–PEG block copolymeric nanoparticles for oral immunization against hepatitis B International Journal of Pharmaceutics 2010 387 253 262 20005936 [102] Bansal V Kumar M Dalela M Brahmne HG Singh H Evaluation of synergistic effect of biodegradable polymeric nanoparticles and aluminum based adjuvant for improving vaccine efficacy International Journal of Pharmaceutics 2014 471 377 384 24939616 [103] Lee W.-k. Park J.-y. Yang EH Suh H Kim SH Chung DS Choi K Yang CW Park J.-s. Investigation of the factors influencing the release rates of cyclosporin A-loaded micro- and nanoparticles prepared by high-pressure homogenizer Journal of Controlled Release 2002 84 115 123 12468215 [104] Tan Q Jiang R Xu M Liu G Li S Zhang J Nanosized sustained-release pyridostigmine bromide microcapsules: process optimization and evaluation of characteristics International journal of nanomedicine 2013 8 737 745 23459707 [105] Jiang W Schwendeman SP Stabilization and controlled release of bovine serum albumin encapsulated in poly(D, L-lactide) and poly(ethylene glycol) microsphere blends Pharmaceutical research 2001 18 878 885 11474795 [106] Elvassore N Bertucco A Caliceti P Production of insulin-loaded poly(ethylene glycol)/poly(l-lactide) (PEG/PLA) nanoparticles by gas antisolvent techniques J Pharm Sci 2001 90 1628 1636 11745721 [107] Caliceti P Salmaso S Elvassore N Bertucco A Effective protein release from PEG/PLA nano-particles produced by compressed gas anti-solvent precipitation techniques Journal of Controlled Release 2004 94 195 205 14684283 [108] Mimi N Belkacemi H Sadoun T Sapin A Maincent P How the composition and manufacturing parameters affect insulin release from polymeric nanoparticles Journal of Drug Delivery Science and Technology 2015 30 458 466 Part B [109] Ruan G Feng S-S Preparation and characterization of poly(lactic acid)–poly(ethylene glycol)–poly(lactic acid) (PLA–PEG–PLA) microspheres for controlled release of paclitaxel Biomaterials 2003 24 5037 5044 14559017 [110] Adesina SK Holly A Kramer-Marek G Capala J Akala EO Polylactide-based paclitaxel-loaded nanoparticles fabricated by dispersion polymerization: characterization, evaluation in cancer cell lines, and preliminary biodistribution studies J Pharm Sci 2014 103 2546 2555 24961596 [111] Cosco D Paolino D De Angelis F Cilurzo F Celia C Di Marzio L Russo D Tsapis N Fattal E Fresta M Aqueous-core PEG-coated PLA nanocapsules for an efficient entrapment of water soluble anticancer drugs and a smart therapeutic response European Journal of Pharmaceutics and Biopharmaceutics 2015 89 30 39 25460850 [112] Tan G-R Feng S-S Leong DT The reduction of anti-cancer drug antagonism by the spatial protection of drugs with PLA–TPGS nanoparticles Biomaterials 2014 35 3044 3051 24439415 [113] Yoshizawa H Nishino S Shiomori K Natsugoe S Aiko T Kitamura Y Surface morphology control of polylactide microspheres enclosing irinotecan hydrochloride Int J Pharm 2005 296 112 116 15885462 [114] Freitas MN Marchetti JM Nimesulide PLA microspheres as a potential sustained release system for the treatment of inflammatory diseases International Journal of Pharmaceutics 2005 295 201 211 15848005 [115] Vila A Sánchez A Évora C Soriano I McCallion O Alonso MJ PLA-PEG particles as nasal protein carriers: the influence of the particle size International Journal of Pharmaceutics 2005 292 43 52 15725552 [116] Dalmolin LF Khalil NM Mainardes RM Delivery of vanillin by poly(lactic-acid) nanoparticles: Development, characterization and in vitro evaluation of antioxidant activity Materials Science and Engineering: C 2016 62 1 8 26952391 [117] Sheng Y Yuan Y Liu C Tao X Shan X Xu F In vitro macrophage uptake and in vivo biodistribution of PLA-PEG nanoparticles loaded with hemoglobin as blood substitutes: effect of PEG content Journal of materials science. Materials in medicine 2009 20 1881 1891 19365612 [118] Zambaux MF Bonneaux F Gref R Dellacherie E Vigneron C Preparation and characterization of protein C-loaded PLA nanoparticles Journal of Controlled Release 1999 60 179 188 10425324 [119] Lee SH Zhang Z Feng S-S Nanoparticles of poly(lactide)—tocopheryl polyethylene glycol succinate (PLA-TPGS) copolymers for protein drug delivery Biomaterials 2007 28 2041 2050 17250886 [120] Cheng Q Feng J Chen J Zhu X Li F Brain transport of neurotoxin-I with PLA nanoparticles through intranasal administration in rats: a microdialysis study Biopharm Drug Dispos 2008 29 431 439 18837064 [121] Pinon-Segundo E Ganem-Quintanar A Alonso-Perez V Quintanar-Guerrero D Preparation and characterization of triclosan nanoparticles for periodontal treatment Int J Pharm 2005 294 217 232 15814246 [122] Hu K Li J Shen Y Lu W Gao X Zhang Q Jiang X Lactoferrin-conjugated PEG– PLA nanoparticles with improved brain delivery: In vitro and in vivo evaluations Journal of Controlled Release 2009 134 55 61 19038299 [123] Nguyen CA Allémann E Schwach G Doelker E Gurny R Cell interaction studies of PLA-MePEG nanoparticles International Journal of Pharmaceutics 2003 254 69 72 12615412 [124] Leroux J-C Allémann E De Jaeghere F Doelker E Gurny R Biodegradable nanoparticles — From sustained release formulations to improved site specific drug delivery Journal of Controlled Release 1996 39 339 350 [125] Thomasin C Merkle HP Gander BA Physico-chemical parameters governing protein microencapsulation into biodegradable polyesters by coacervation International Journal of Pharmaceutics 1997 147 173 186 [126] Pandey SK Ghosh S Maiti P Haldar C Therapeutic efficacy and toxicity of tamoxifen loaded PLA nanoparticles for breast cancer International Journal of Biological Macromolecules 2015 72 309 319 25151954 [127] Barwal I Sood A Sharma M Singh B Yadav SC Development of stevioside Pluronic-F-68 copolymer based PLA-nanoparticles as an antidiabetic nanomedicine Colloids and Surfaces B: Biointerfaces 2013 101 510 516 23022553 [128] Jeevitha D Amarnath K Chitosan/PLA nanoparticles as a novel carrier for the delivery of anthraquinone: Synthesis, characterization and in vitro cytotoxicity evaluation Colloids and Surfaces B: Biointerfaces 2013 101 126 134 22796782 [129] Xiong XY Li YP Li ZL Zhou CL Tam KC Liu ZY Xie GX Vesicles from Pluronic/poly(lactic acid) block copolymers as new carriers for oral insulin delivery Journal of Controlled Release 2007 120 11 17 17509718 [130] Aline F Brand D Pierre J Roingeard P Séverine M Verrier B Dimier-Poisson I Dendritic cells loaded with HIV-1 p24 proteins adsorbed on surfactant-free anionic PLA nanoparticles induce enhanced cellular immune responses against HIV-1 after vaccination Vaccine 2009 27 5284 5291 19450633 [131] Guerrero S Muniz E Teijon C Olmo R Teijon JM Blanco MD Ketotifen-loaded microspheres prepared by spray-drying poly(D,L-lactide) and poly(D,L-lactide-co-glycolide) polymers: characterization and in vivo evaluation J Pharm Sci 2008 97 3153 3169 18023014 [132] Rungseevijitprapa W Brazeau GA Simkins JW Bodmeier R Myotoxicity studies of O/W-in situ forming microparticle systems European Journal of Pharmaceutics and Biopharmaceutics 2008 69 126 133 18036794 [133] Geng Z Luo X Zhang Z Li H Tian J Yu Z Study of an injectable in situ forming gel for sustained-release of Ivermectin in vitro and in vivo Int J Biol Macromol 2016 85 271 276 26708436 [134] Kranz H Bodmeier R A novel in situ forming drug delivery system for controlled parenteral drug delivery International Journal of Pharmaceutics 2007 332 107 114 17084049 [135] Kranz H Bodmeier R Structure formation and characterization of injectable drug loaded biodegradable devices: In situ implants versus in situ microparticles European Journal of Pharmaceutical Sciences 2008 34 164 172 18501569 [136] Pitarresi G Palumbo FS Fiorica C Calascibetta F Di Stefano M Giammona G Injectable in situ forming microgels of hyaluronic acid-g-polylactic acid for methylprednisolone release European Polymer Journal 2013 49 718 725 [137] Anderson LC Wise DL Howes JF An injectable sustained release fertility control system Contraception 1976 13 375 384 1248260 [138] Nykamp G Carstensen U Müller BW Jet milling—a new technique for microparticle preparation International Journal of Pharmaceutics 2002 242 79 86 12176228 [139] Yolles S Leafe TD Woodland JH Meyer FJ Long acting delivery systems for narcotic antagonists II: release rates of naltrexone from poly(lactic acid) composites J Pharm Sci 1975 64 348 349 1127596 [140] Smith A Hunneyball l.M. Evaluation of poly(lactic acid) as a biodegradable drug delivery system for parenteral administration International Journal of Pharmaceutics 1986 30 215 220 [141] Li D Guo G Fan R Liang J Deng X Luo F Qian Z PLA/F68/dexamethasone implants prepared by hot-melt extrusion for controlled release of anti-inflammatory drug to implantable medical devices: I. Preparation, characterization and hydrolytic degradation study Int J Pharm 2013 441 365 372 23178216 [142] Bleich J Muller BW Production of drug loaded microparticles by the use of supercritical gases with the aerosol solvent extraction system (ASES) process J Microencapsul 1996 13 131 139 8999119 [143] Whitaker MJ Hao J Davies OR Serhatkulu G Stolnik-Trenkic S Howdle SM Shakesheff KM The production of protein-loaded microparticles by supercritical fluid enhanced mixing and spraying Journal of Controlled Release 2005 101 85 92 15588896 [144] Falk R Randolph TW Meyer JD Kelly RM Manning MC Controlled release of ionic compounds from poly (l-lactide) microspheres produced by precipitation with a compressed antisolvent Journal of Controlled Release 1997 44 77 85 [145] He T Liang Q Zhang K Mu X Luo T Wang Y Luo G A modified microfluidic chip for fabrication of paclitaxel-loaded poly(l-lactic acid) microspheres Microfluidics and Nanofluidics 2011 10 1289 1298 [146] van de Weert M Hennink WE Jiskoot W Protein instability in poly(lactic-co-glycolic acid) microparticles Pharmaceutical research 2000 17 1159 1167 11145219 [147] Katou H Wandrey AJ Gander B Kinetics of solvent extraction/evaporation process for PLGA microparticle fabrication International Journal of Pharmaceutics 2008 364 45 53 18782610 [148] Thomasin C Ho NT Merkle HP Gander B Drug microencapsulation by PLA/PLGA coacervation in the light of thermodynamics. 1. Overview and theoretical considerations J Pharm Sci 1998 87 259 268 9523976 [149] Astete CE Sabliov CM Synthesis and characterization of PLGA nanoparticles Journal of biomaterials science. Polymer edition 2006 17 247 289 16689015 [150] Yun YH Lee BK Park K Controlled Drug Delivery: Historical perspective for the next generation J Control Release 2015 219 2 7 26456749 [151] Sah H Thoma LA Desu HR Sah E Wood GC Concepts and practices used to develop functional PLGA-based nanoparticulate systems International journal of nanomedicine 2013 8 747 765 23459088 [152] Ma G Microencapsulation of protein drugs for drug delivery: strategy, preparation, and applications J Control Release 2014 193 324 340 25218676 [153] Palacio J Agudelo NA Lopez BL PEGylation of PLA nanoparticles to improve mucus-penetration and colloidal stability for oral delivery systems Current Opinion in Chemical Engineering 2016 11 14 19

PMC005xxxxxx/PMC5133388.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 0410462 6011 Nature Nature Nature 0028-0836 1476-4687 27533040 5133388 10.1038/nature19081 EMS69189 Article Tumor hypoxia causes DNA hypermethylation by reducing TET activity Thienpont Bernard 12 Steinbacher Jessica 3 Zhao Hui 12 D’Anna Flora 12 Kuchnio Anna 14 Ploumakis Athanasios 5 Ghesquière Bart 1 Van Dyck Laurien 12 Boeckx Bram 12 Schoonjans Luc 14 Hermans Els 6 Amant Frederic 6 Kristensen Vessela N. 7 Peng Koh Kian 8 Mazzone Massimiliano 19 Coleman Mathew 5 Carell Thomas 3 Carmeliet Peter 14 Lambrechts Diether 12 1 Vesalius Research Center, VIB, Leuven, Belgium 2 Laboratory of Translational Genetics, Department of Oncology, KU Leuven, Leuven, Belgium 3 Department für Chemie und Pharmazie, Ludwig-Maximilians-Universität, München, Germany 4 Laboratory of Angiogenesis and Vascular Metabolism, Department of Oncology, KU Leuven, Leuven, Belgium 5 Institute of Cancer and Genomic Sciences, University of Birmingham, Birmingham, UK 6 Gynecologic Oncology, University Hospitals Leuven, Department of Oncology, KU Leuven, Leuven, Belgium 7 Department of Genetics, Institute for Cancer Research, Oslo University Hospital Radiumhospitalet, Oslo, Norway 8 Department of Development and Regeneration, and Stem Cell Institute Leuven, KU Leuven, Leuven, Belgium 9 Laboratory of Molecular Oncology and Angiogenesis, Department of Oncology, KU Leuven, Leuven, Belgium Correspondence and requests for materials should be addressed to BT (bernard.thienpont@vib-kuleuven.be) or DL (diether.lambrechts@vib-kuleuven.be). 17 8 2016 17 8 2016 01 9 2016 17 2 2017 537 7618 6368 This file is available to download for the purposes of text mining, consistent with the principles of UK copyright law. Summary Hypermethylation of tumor suppressor gene (TSG) promoters confers growth advantages to cancer cells, but how these changes arise is poorly understood. Here, we report that tumor hypoxia reduces the activity of oxygen-dependent TET enzymes, which catalyze DNA de-methylation through 5-methylcytosine oxidation. This occurs independently of hypoxia-associated alterations in TET expression, proliferation, metabolism, HIF activity or reactive oxygen, but directly depends on oxygen shortage. Hypoxia-induced loss of TET activity increases hypermethylation at gene promoters in vitro. Also in patients, TSG promoters are markedly more methylated in hypoxic tumors, independently of proliferation, stromal cell infiltration and tumor characteristics. Our data suggest cellular selection of hypermethylation events, with almost half of them being ascribable to hypoxia across tumor types. Accordingly, increased hypoxia after vessel pruning in murine breast tumors increases hypermethylation, while restored tumor oxygenation by vessel normalization abrogates this effect. Tumor hypoxia thus acts as a novel regulator underlying DNA methylation. Mutational processes underlying oncogenesis are well studied. Apart from genetic changes, tumors are also epigenetically distinct from their tissue of origin. Most established are DNA methylation changes, but the mechanisms underlying these are poorly understood1. In tumors, DNA methylation changes involve global hypomethylation, and local hypermethylation (HM) of CpG-rich gene promoters1. HM frequently affects tumor suppressor genes (TSGs), down-regulating their expression and thus contributing to oncogenesis. How methylation changes arise remains debated. Following an instructive model, genetic changes are a prerequisite for methylation changes2. For instance, BRAF mutations lead to HM in colorectal tumors3. A limitation of this model is that, while pervasive, HM of TSGs can be explained by somatic mutations in only a fraction of tumors. As a striking example, extensive HM was found in ependymomas devoid of somatic mutations4. In contrast to methylation, DNA de-methylation mechanisms have remained elusive, until recently, when ten-eleven translocation methylcytosine dioxygenases (TET1, TET2 and TET3) were shown to oxidize 5-methylcytosine (5mC) to 5-hydroxymethylcytosine (5hmC)5. 5hmC and its further oxidized derivatives are subsequently replaced with an unmodified C by base-excision repair to achieve de-methylation6. Reduced 5mC oxidation due to decreased TET activity thus increases DNA methylation. Mutations suppressing TET activity and thus reducing 5hmC are often found in myeloid leukemia and glioblastoma6–9, but less frequently in other tumor types. In contrast, 5hmC loss is pervasive in tumors and even proposed as a cancer hallmark10. Thus, similar to HM, somatic mutations explain the loss of 5hmC in only a fraction of tumors, and it remains unclear which other factors trigger this loss2. Interestingly, TET enzymes are Fe2+ and α-ketoglutarate-(αKG)-dependent dioxygenases, similar to HIF-prolyl-hydroxylase domain proteins (PHDs)11. The latter are sensitive in their activity to oxygen and act as oxygen sensors: under normoxic conditions PHDs hydroxylate the HIF transcription factors, targeting them for proteasomal degradation, whereas under hypoxia they fail to hydroxylate, leading to HIF stabilization and hypoxia response activation12. Expanding tumors continuously become disconnected from their vascular supply, resulting in vicious cycles of hypoxia followed by HIF activation and tumor vessel formation13. Consequently, hypoxia pervades in solid tumors, with oxygen levels ranging from 5% to anoxia, and about a third of tumor areas containing <0.5% oxygen14. Although DNA HM and hypoxia are well-recognized cancer hallmarks, the impact of hypoxia on TET hydroxylase activity and subsequent DNA (de)methylation has not been assessed. We here hypothesize that a hypoxic micro-environment decreases TET hydroxylase activity in tumors, leading to an accumulation of 5mC and acquisition of HM. Impact of hypoxia on DNA hydroxymethylation activity To assess whether hypoxia affects TET activity, we exposed 10 human and 5 murine cell lines with detectable 5hmC levels for 24 hours to 21% or 0.5% O2, a level commonly observed in tumors14. Hypoxia induction was verified and DNA was extracted and profiled for nucleotide composition using LC/MS. 11 cell lines, including eight cancer cell lines, displayed 5hmC loss (Figure 1a). However, this did not translate into global 5mC increases (Extended data figure 1), presumably because 5mC is more abundant and at many sites not targeted by TETs15. The effect of hypoxia was concentration- and time-dependent: a dose-response revealed gradual reductions from 1-2% O2 onwards and a time course respectively, a 20% and 40% reduction after 15 and >24 hours (Figure 1b-c). Loss of 5hmC was not secondary to increased 5hmC oxidation to 5fC16, as hypoxia also decreased 5fC levels in ES cells (Extended data figure 1). In some cell lines, 5hmC failed to decrease under hypoxia. Particularly, 5hmC was unaffected in H1299 and 4T1, and even increased in SHSY5Y and SK-N-Be2c neuroblastoma cells, as reported previously17 (Figure 1a). When profiling TET expression, neuroblastoma cells displayed potent hypoxia-induction of TET1 and TET2, H1299 and 4T1 exhibited intermediate increases, and all other cell lines no or only modest increases of some TET paralogues (Figure 1a). Tet expression changes were confirmed at the protein level in murine cell lines, and HIF1β-ChIP-seq further confirmed that HIF binds near the promoters of TETs that are upregulated, but not near those that are unaltered (Extended data figure 2a-b), in keeping with the cell-type specificity of the hypoxia response12. Importantly, no cell line showed decreased TET expression, indicating that 5hmC loss is not due to reduced TET expression. Since hypoxia differentially affects TET expression, we correlated hypoxia-associated changes in overall TET expression (the combined abundances of TET1, TET2 and TET3) with changes in 5hmC levels. Hypoxia reduced 5hmC on average by 44% (P=0.0097) in each cell line (Figure 1d), independently of TET expression changes. Nevertheless, changes in TET expression also determined 5hmC levels. This was confirmed by siRNA knockdown of TET2, which constitutes ~60% of all TET expression in MCF7 cells: this reduced 5hmC levels also by ~60% (Extended data figure 2c). Likewise, Tet1-KO ES cells displayed lower 5hmC levels than wild-type ES cells, in which Tet1 is the predominantly expressed Tet paralogue, both under 21% or 0.5% O2 (Figure 1a, Extended data figure 2d). Hence, 5hmC levels after hypoxia appear to be determined by altered oxygen availability and by changes in TET abundance. This explains why cell lines without hypoxia-induced upregulation of TETs display 5hmC loss, whereas cell lines strongly upregulating TETs compensate this, resulting in equal or increased 5hmC levels. Changes secondary to hypoxia do not affect DNA hydroxymethylation Apart from gene expression, TET activity is affected by a variety of cellular processes, including changes in reactive oxygen species (ROS), Krebs cycle metabolites and proliferation7,11,17,18. Since such changes might also occur secondary to hypoxia, we investigated whether they underlie 5hmC reductions in hypoxia. Firstly, ROS could affect TETs in the nucleus through inactivation of Fe2+ in their catalytic domain. Although ROS was overall increased upon hypoxia, no increase in nuclear ROS was detected by a nucleus-specific ROS probe or 8-oxo-guanine quantification (Extended data figure 3a-f). Ascorbate supplementation to counteract ROS increases19, moreover failed to rescue 5hmC loss (Figure 1e). Secondly, changes in metabolites such as succinate and fumarate affect TET function by competing with its cofactor αKG7. The concentration of these metabolites was however not increased in hypoxic MCF10A or ES cells, and only 3-4-fold in MCF7 cells (Extended data figure 3g-i). The onco-metabolite 2-hydroxyglutarate was also increased in hypoxic MCF7 and MCF10A cells, but levels were only ~5-10% of αKG (Extended data figure 3h,j), and therefore unlikely to affect TET activity, as affinity of these competing metabolites for hydroxylases is lower or similar to αKG7,20. Indeed, culturing MCF7 cells in glutamine-free medium to decrease these metabolite concentrations did not alter 5hmC levels (Extended data figure 3k). Exogenously adding cell-permeable αKG under hypoxia to counteract putative competing metabolites likewise did not rescue the 5hmC loss (Figure 1f). This excludes that metabolite competition underlies hypoxia-associated 5hmC loss. Thirdly, increases in cell proliferation have been linked to 5hmC loss21. However, cell growth was unaffected or decreased upon exposure to hypoxia in all cell lines tested, indicating that increased proliferation does not underlie 5hmC reduction (Extended data figure 3l). Fourthly, to exclude cellular changes secondary to HIF activation, we pharmacologically activated the hypoxia response program by exposing 5 cell lines grown in atmospheric conditions to IOX2, a small molecule inhibitor displaying high specificity for PHDs22 (Extended data figure 3m). Cell lines not characterized by hypoxia-induced TET expression changes (i.e., MCF10A, A549 and MCF7) showed no change in 5hmC under IOX2, while SK-N-Be2c and SHSY5Y, characterized by TET upregulation, did show an increase in 5hmC (Figure 1g). Thus, upon IOX2 exposure, 5hmC changes mirrored changes in TET transcription. We also prepared nuclear protein extracts from MCF7 cells grown under hypoxic and atmospheric conditions, and then compared their 5mC oxidative capacities at the same oxygen tension in vitro; these were however identical (Extended data figure 3n). Loss of 5hmC was therefore not secondary to activation of the hypoxia response program. In a final experiment, we assessed the effect of varying oxygen concentrations on the activity of recombinant purified Tet1 or Tet2, by measuring conversion of 5mC to 5hmC on double-stranded genomic DNA. We observed a dose-dependent loss of 5hmC production with decreasing oxygen concentration. Importantly, under the hypoxic conditions applied in this study (0.5% O2), Tet1 and Tet2 activity were reduced by 45±7% and 52±8% (P=0.01; Figure 1h-i). Together, these data demonstrate that decreased oxygen availability directly diminishes the oxidative activity of TETs, independently of changes in HIF activity, competing metabolites, proliferation, nuclear ROS or TET expression. Genomic loci displaying differential DNA hydroxymethylation To analyze where in the genome hypoxia reduces 5hmC, DNA from hypoxic and control MCF7 cells was immunoprecipitated using antibodies targeting 5mC or 5hmC, and subjected to high-throughput sequencing (DIP-seq). We detected 290,382 sites enriched for 5hmC. Upon hypoxia, 10,001 of these peaks exhibited a decrease in 5hmC (5% FDR), versus only 18 exhibiting an increase, thereby confirming the global 5hmC loss (Figure 2a; Supplementary table 1). Genomic annotation of these peaks using chromHMM23 revealed they were predominantly found at gene promoters, but also at enhancers and actively transcribed regions, in line with known TET binding (Figure 2b)15. For example, 5hmC was decreased near transcription start sites of NSD1, FOXA1 and CDKN2A (Extended data figure 4). Analysis of 5mC-DIP signals at these 10,001 regions highlighted that, in 724 out of 875 altered regions at P<0.05, the 5mC content was increased, although only 1 of these sites survived 5% FDR correction (Figure 2c; Supplementary table 2). Increases in 5mC were thus more subtle than decreases observed for 5hmC. Several days may be required for 5hmC changes to cause 5mC changes19. We therefore cultured cells for 48 (instead of 24) hours under hypoxia, and used targeted bisulfite-sequencing (BS-seq) to obtain base-resolution quantitation of 5mC at ~85Mb of promoters and enhancers. Using this approach, we could assess increases in 5mC for 1,894 of the 10,001 regions displaying 5hmC loss. As observed upon 5mC-DIP-seq, out of 402 altered sites (P<0.05), 301 displayed increased methylation. Likewise, 60 out of 99 altered sites at 5% FDR were increased (P=2.8×10-3; Figure 2d; Supplementary table 3). ChromHMM annotation revealed that these 60 sites were predominantly in gene promoters and enhancers. To assess the impact of HM on gene expression, we performed RNA-seq on hypoxic MCF7 cells. Genes depleted in 5hmC and at the same time increased in 5mC, were characterized by decreased expression upon hypoxia (Figure 2e; P=2.5×10-42 and 7.4×10-4, respectively for 3,660 genes with 5hmC loss and 55 genes with both 5hmC loss and 5mC gain; Supplementary table 4). Reduced TET activity thus leads to an accumulation of 5mC, decreasing expression of associated genes. Selection of HM events in hypoxic tumors We next analyzed whether 5hmC loss and concomitant 5mC gain also occur in vivo. We focused on gene promoters as they are more frequently affected upon hypoxia, and directly linked to gene expression. Moreover, as cancer cells go through multiple rounds of sustained hypoxia14, we hypothesized that changes in 5mC might be enriched for, as they provide a substrate for cellular selection of cancer cells, similar to somatic mutations. First, we assessed 5hmC levels in three patient-derived tumor xenografts, wherein we marked hypoxic areas with pimonidazole (Extended data figure 5a). Immunofluorescence analysis revealed decreased 5hmC in hypoxic areas, linking tumor hypoxia to 5hmC loss in vivo. To model whether hypoxia-associated HM contributes to the oncogenic process, we analyzed tumors profiled in the pan-cancer study of The Cancer Genome Atlas (TCGA)24. We selected 8 solid tumor types (3,141 tumors) for which both DNA methylation (450K array) and gene expression (RNA-seq) data were available for >100 samples, and classified each as hypoxic, normoxic or intermediate using an established gene signature (Extended data figure 5b)25. Next, we analyzed tumor-associated DNA HM in each tumor type by performing unsupervised clustering of 1,000 CpGs that displayed the strongest HM in tumor versus normal tissue (Extended data figure 5c). In the 3 first clusters, displaying low, intermediate and high average HM, we analysed the enrichment of hypoxic tumors. For all 8 tumor types, hypoxic tumors predominated in the hypermethylated cluster and normoxic tumors in the hypomethylated cluster (Figure 3a; P=2×10-4), suggesting that hypoxia leads to increased methylation in tumors. Whereas the above analysis identifies uniform increases in methylation based on average changes, it poorly captures exceptional increases in HM known to occur in a subset of tumors1,26. We therefore also modeled tumor HM by annotating increases in CpG methylation at gene promoters using a stringent threshold (Bonferroni-corrected P<0.05) as HM events. In each tumor type the promoters of 187±38 out of 29,649 genes frequently displayed HM events (Supplementary table 5). Importantly, hypoxic tumors had on average 4.8-fold more HM events in these genes than normoxic tumors (Figure 3b; P=4.1×10-13). These events were functional, reducing gene expression in tumors carrying these HM events (Extended data figure 5d). They primarily affected promoters with a high or intermediate CpG content, in line with TET target preference (Extended data figure 5e)15. Furthermore, they were not restricted to a small subset: 77±6.5%, 49±9.3% or 39±9.1% of hypoxic tumors was affected by ≥1, ≥10 or ≥20 HM events. When considering HM frequencies in normoxic tumors as baseline, up to 48% of HM events were hypoxia-related. As HM can also be genetically-encoded, mutations in some genes correlated positively with HM (e.g. IDH1, TET1, TET3 and BRAF; Supplementary table 6). Importantly, hypoxia predicted HM independently of the mutation status (P=6.1×10-12). Mutations inhibiting TET activity were moreover infrequent (~1.8% of tumors), indicating that HM is not genetically-encoded in most tumors. TET-mutant tumors were also not more hypoxic, suggesting that hypoxia induces HM, and not vice versa (Extended data figure 5f). Hypoxia-associated HM events occurred independently of other tumor characteristics, such as tumor percentage, immune cell infiltration, tumor size, proliferation or metastasis (P=4×10-13), and were significant in 7 of 8 tumor types (Supplementary tables 7-8). In line with an earlier report21, high proliferation was the only other variable significantly predicting HM (P=5.3×10-10), although only in 4 of 8 tumor types (Extended data figure 5g-h). Using multiple regression, we estimated contributions of tumor characteristics to HM variance. Based on partial correlation coefficients, proliferation predicted 12.1±4.1% and hypoxia 33.3±5.7% of HM events explained by the model (Extended data figure 5i). Given the enrichment of HM events in hypoxic tumors, we next selected genes enriched for HM events in hypoxic versus normoxic tumors (5% FDR). This revealed 263±94 genes per tumor type, with 9.0±1.6% being shared between any 2 types (Supplementary table 9). Ontology analysis of hypermethylated genes revealed common biological processes, such as cell cycle arrest, DNA repair and apoptosis. In line with tumor hypoxia inducing glycolysis, angiogenesis and metastasis, HM was also observed in genes suppressing these processes (Extended data figure 6a-c). Reduced TET activity underlies HM Three strategies were used to confirm the role of TET activity in hypoxia-associated HM. First, we correlated TET expression with HM events, while correcting for hypoxia and proliferation. TET2 and TET3 expression correlated inversely with HM (P=0.046 and 0.0028, Extended data figure 7a), as did hypoxia and proliferation (P<1.2×10-13 for both). Similar to our in vitro observations, this implicates reduced TET activity in HM. Secondly, we assessed the overlap of HM events induced by hypoxia and IDH1R132 mutations8 in 63 glioblastomas. Among IDH1-wildtype glioblastomas, the HM frequency was 3.4-fold higher in hypoxic tumors (Figure 4a, Extended data figure 7b). As expected, IDH1R132 tumors showed HM, albeit 3.9-fold more than hypoxic tumors (Figure 4a), indicating that TET enzymes, being fully inactivated in IDH-mutant tumors9, were only partially inactivated in hypoxia, similar to our in vitro observations. Of 228 genes frequently hypermethylated in glioblastomas, hypermethylated genes in the hypoxic and IDH-mutant subgroups displayed a 58% overlap (P<10-16; Figure 4b) and a reduced expression (Extended data figure 7c), indicating that loss of TET activity affects the same genes, regardless of the underlying trigger. Finally, to link hypoxia-associated HM to 5hmC loss, we profiled 24 non-small cell lung tumors for 5mC and 5hmC using 450K arrays (Extended data figure 7d). This revealed a generalized loss of 5hmC in hypoxic tumors (-7.1±1.1%; P=3.7×10-3; Figure 4c). Also individual probes mostly displayed 5hmC loss and 5mC gain in hypoxic tumors (respectively, 96.7% and 65.4% of probes altered at P<0.01; Supplementary table 10). Of all probes displaying 5mC gain, most (87%) also displayed 5hmC loss, and of probes altered both in 5hmC and 5mC (P<0.01), 92% showed 5hmC loss and 5mC gain (Figure 4d; P<10-16). This directly implicates hypoxia-induced loss of 5hmC in HM of hypoxic tumors. Rescue and exacerbation of hypoxia-induced HM in murine breast tumors To manipulate tumor oxygenation and confirm its impact on HM, we used mice expressing the polyoma middle T-antigen under the mouse mammary tumor virus promoter (MMTV-PyMT). These mice spontaneously develop breast tumors, with hypoxic areas emerging from 7 weeks onwards, encompassing ∼20% of tumor at 16 weeks27. Hypoxic areas in these tumors were also depleted in 5hmC (Figure 5a-b). We monitored HM changes by targeted BS-seq of TSG promoters commonly inactivated in cancer28. Hypoxic human breast tumors indeed display a specific increase in HM at these TSG promoters, whereas no effect was observed for oncogenes (Extended data figure 8a). In line with the age-associated increase in tumor hypoxia27, HM events increased dramatically with age or tumor size, but not in normal mammary glands (Extended data figure 8b-d). Importantly, >95% of cells in these tumors were PyMT-positive, whereas cell proliferation and immune cell infiltration were comparable between hypoxic and normoxic areas (Extended data figure 8e-g). HM changes are therefore unlikely secondary to changes in proliferation or cellular heterogeneity. To test whether reduced tumor oxygenation increases HM, 9-week-old MMTV-PyMT mice were hydrodynamically injected with a soluble-Flk1 (sFlk1)-expressing plasmid. After 3 weeks, this caused tumor vessel pruning and hypoxia (Extended data figure 9a-d). Shallow whole-genome sequencing for 5hmC (TAB-seq) revealed a global loss of 5hmC upon sFlk1 overexpression (-12.4±3.5%, P=0.040), predominantly at gene-dense regions and affecting the entire gene (Figure 5c, Extended data figure 9e), consistent with previously described 5hmC distributions15. Moreover, targeted BS-seq revealed an exacerbated HM phenotype after sFlk1 overexpression at 12 weeks, and this in TSGs but not oncogenes (10 out of 15 TSGs contained ≥1 HM event; P=0.010, Figure 5d, Extended data figure 9f). Tumor growth and expression of proliferation markers, Tet paralogues and the immune cell marker CD45 were unaffected by sFlk1 overexpression, indicating that HM occurs independently (Extended data figure 9g-j). To rescue this effect, we normalized the tumor vasculature by intercrossing a heterozygous Phd2 loss-of-function allele with the PyMT transgene. This significantly reduced tumor hypoxia at 16 weeks27 (Extended data figure 9k). TAB-seq revealed a 5hmC gain (+25.3±4.7%, P=0.0098), primarily at gene-dense regions and affecting the entire gene (Figure 5c, Extended data figure 9l). Interestingly, BS-seq revealed that, whereas 8 out of 15 TSGs displayed ≥1 HM event in Phd2+/+ tumors, no HM was observed in Phd2+/- tumors (P=2.6×10-7, Figure 5e). Again, oncogenes were unaffected (Extended data figure 9m). Importantly, effects were independent of Phd2 haplodeficiency in tumor cells, as similar effects were observed in PyMT mice having endothelial-cell-specific Phd2 haplodeficiency (Extended data figure 9n-o)27. Like the sFlk1 model, also increasing tumor oxygenation by Phd2 haplodeficiency did not affect tumor growth, expression of proliferation markers, Tets or CD45 (Extended data figure 9p-u). Discussion We here show that tumor hypoxia directly reduces TET activity, causing a 5hmC decrease predominantly at gene promoters and enhancers. Concomitantly, 5mC increases at these sites, and, similar to genetic mutations, becomes a substrate for oncogenic selection in vivo26. Since hypoxia prevails in tumors, 5mC changes in TSG promoters are enriched for, rendering hypoxic tumors hypermethylated at these sites. HM events in tumors have long been suspected to occur through selection of random DNA methylation variants29. However, the identification of genetically-encoded HM challenged this stochastic model2. By demonstrating that hypoxia drives HM, we show that genetically-encoded and tumor microenvironment-driven models of epimutagenesis co-exist. However, since hypoxia is pervasive, the mechanism described here is relevant for most solid tumors: up to 48% of HM events was hypoxia-related, and effects were replicated in all tumor types investigated, independently of mutation- and proliferation-induced HM. Importantly, modest hypoxia (2-5% O2) did not affect TET activity, indicating that TET enzymes are not physiological oxygen sensors like the PHDs, as reported30. TET activity only becomes limiting under pathophysiological oxygen concentrations found in tumors14, and analogous to somatic TET haploinsufficiency, this partial reduction in TET activity contributes to oncogenesis. Our findings also suggest intriguing avenues of investigation into other ischemia-related pathologies. Our model provides an elegant mechanism for the association between hypoxia and (mal)adaptive oncogenic processes: genes affected by HM were involved in cell-cycle arrest, DNA repair and apoptosis, but also glycolysis, metastasis and angiogenesis. Interestingly, high levels of angiogenesis inhibitors stimulate metastatic spreading in murine cancer models, at least in specific settings31, and tumor hypoxia is considered a driver of this behavior. The mechanism described here, by which HM accumulates under hypoxia, may underlie these escape mechanisms. Contrastingly, low levels of VEGF inhibition can induce tumor vessel normalization and improve oxygenation32. Our observations in normalized PyMT tumors suggest that therapeutic benefits of vessel normalization, such as decreased metastatic burden27, might occur by inhibiting hypoxia-associated HM. Countering this HM, for instance through drugs inhibiting DNA methylation and/or by normalizing tumor blood supply, may thus prove therapeutically beneficial. Methods Materials All materials were molecular biology grade. Unless noted otherwise, all were from Sigma (Diegem, Belgium). Analysis of global 5mC and 5hmC levels in cultured cells Cell lines MCF7, MCF10A, A549, H1299, SHSY5Y, Hep G2, Hep 3B2, HT-1080, NCI-H358, LLC, Neuro-2a, 4T1 and SK-N-BE2c cells lines were obtained from the American Type Culture Collection and their identity was not further authenticated. These are not listed in the database of commonly misidentified cell lines maintained by ICLAC. LLC, Neuro-2a, 4T1, Hep G2, HT-1080, Hep 3B2, MCF7 and A549 cells were cultured at 37°C in Dulbecco’s modified Eagle medium (DMEM) with 10% fetal bovine serum (FBS), 5ml of 100 U/ml Penicillin-Streptomycin (Pen Strep, Life Technologies) and 5ml of L-Glutamine 200mM. NCI-H358, H1299 and SK-N-BE2c cell lines were cultured at 37°C in Roswell Park Memorial Institute (RPMI) 1640 Medium (RPMI) 10% FBS 1% Pen Strep and 1% L-Glutamine. MCF10A cells were cultured at 37°C in DMEM/F-12 (Dulbecco's Modified Eagle Medium/Nutrient Mixture F-12) supplemented with 5% horse serum (Life Technologies), 20 ng/ml human Epidermal Growth Factor (Prepotec), 0.5 μg/ml hydrocortisone, 100ng/ml cholera toxin, 10 μg/ml insulin, and 100 U/ml Pen Strep. The SHSY5Y cell line was cultured at 37°C in DMEM/F-12 supplemented with 10% FBS, 2% (PenStrep) and 1% Non Essential Amino Acids (MEM). Mouse J1 ES cells were cultured feeder-free in fibroblast-conditioned medium. Cell cultures were confirmed to be mycoplasma-free every month. Cell line treatment conditions Control cell cultures were grown at atmospheric oxygen concentrations (21%) with 5% CO2. To render cultures hypoxic, they were incubated in an atmosphere of 0.5% oxygen, 5% CO2 and 94.5% N2. Where indicated, IOX2 (50 µM), ascorbate (0.5 mM, a dose known to support TET activity19) or dimethyl α-ketoglutarate (0.5 mM) were added to fresh culture medium, using an equal volume of the carrier (DMSO) as a control for IOX2. Cells were plated at a density tailored to reach 80-95% confluence at the end of the treatment. Fresh medium was added to the cells just before hypoxia exposure. For glutamine-free culture experiments, dialysed FBS was added to glutamine-free DMEM, and supplemented with glutamine (4 mM) for the control. Mouse J1 ES cells and Tet1-gene-trap ES cells were cultured feeder-free in fibroblast-conditioned medium. DNA extraction After exposure to the aforementioned stimuli, cultured cells were washed on ice with ice-cold phosphate-buffer saline (PBS) with deferoxamin (PBS-DFO, 200 µM), detached using cell scrapers and collected by centrifugation (400 ×G, 4°C). Nucleic acids were subsequently extracted using the Wizard Genomic DNA Purification (Promega, Leiden, The Netherlands) kit according to instructions, with all buffers supplemented with DFO (200 µM), dissolved in 80 µL PBS-DFO with RNAse A (200 units, NEB, Ipswich, MA, USA), incubated for 10 minutes at 37°C. After proteinase K addition (200 units) and incubation for 30 minutes at 56°C, DNA was purified using the QIAQuick blood and tissue kit (all buffers supplemented with DFO), eluted in 100 µL of a 10 mM Tris, 1mM EDTA solution (pH 8) and stored at -80°C until further processing. LC-ESI-MS/MS of DNA to measure 5mC, 5hmC and 8-oxoG levels To measure the cytosine, 5-methylcytosine (5mC), 5-hydroxymethylcytosine (5hmC) and 8-oxo-7,8-dihydroguanine (8-oxo-G) content of DNA samples, three technical replicates were run for each sample. More specifically, 0.5 to 2 µg DNA in 25 µL H2O were digested as follows: an aqueous solution (7.5 µL) of 480 μM ZnSO4, containing 42 units Nuclease S1, 5 units antarctic phosphatase, and specific amounts of labeled internal standards were added and the mixture was incubated at 37 °C for 3 h in a Thermomixer comfort (Eppendorf). After addition of 7.5 μL of a 520 µM [Na]2-EDTA solution containing 0.2 units snake venom phosphodiesterase I, the sample was incubated for another 3 h at 37 °C. The total volume was 40 µL. The sample was then kept at -20 °C until the day of analysis. Samples were then filtered by using an AcroPrep Advance 96 filter plate 0.2 μm Supor (Pall Life Sciences) and then analyzed by LC-ESI-MS/MS, which are performed using an Agilent 1290 UHPLC system and an Agilent 6490 triple quadrupole mass spectrometer coupled with the stable isotope dilution technique. DNA samples were digested to give a nucleoside mixture and spiked with specific amounts of the corresponding isotopically labeled standards before LC-MS/MS analysis. The nucleosides were analyzed in the positive ion selected reaction monitoring mode (SRM). In the positive ion mode, [M+H]+ species were measured. Determination and comparison of nucleoside concentrations The resulting cytosine, 5mC, 5hmC and 8-oxo-G peak areas were normalized using the isotopically labeled standards, and expressed relative to the total cytosine content (i.e. C + 5mC + 5hmC). Concentrations were depicted as averages of independent replicates grown on different days, and compared between hypoxia and normoxia (21% O2), or between control and treated conditions, using a paired Student’s t-test. No statistical methods were used to predetermine sample size. TET mRNA concentrations and hypoxia marker gene induction RNA extraction, cDNA synthesis and qPCR For RNA extractions, cell culture medium was removed, TRIzol (Life Technologies) added and processed according to manufacturers guidelines. Reverse transcription and qPCR were performed using 2× TaqMan® Fast Universal PCR Master Mix (Life Technologies), TaqMan probes and primers (IDT, Leuven, Belgium), whose sequence is available under Supplementary table 12. Thermal cycling and fluorescence detection were done using a LightCycler 480 Real-Time PCR System (Roche). Taqman assay amplification efficiencies were verified using serial cDNA dilutions, and estimated to be >95%. mRNA concentration analysis and statistics Ct values were determined for each sample and gene of interest in technical duplicates, and normalized according to the corresponding amplification efficiency. Per sample, TET expression was expressed relative to β-2-microglobulin (human) or Hypoxanthine Phosphoribosyltransferase 1 (mouse) levels by subtraction of their average Cts. Concentrations were expressed as averages of at least 5 replicates extracted on different days. For Figure 1a, copy number estimates for TET1, TET2 and TET3 were expressed for each cell line, relative to the summed copy number estimates of TET1, TET2 and TET3 under control conditions (21% O2). Concentrations were compared between hypoxia and normoxia, or between control and treatment conditions using a Student’s t-test. No statistical methods were used to predetermine sample size. Hypoxia marker gene induction To further verify induction of the hypoxia response program, hypoxia marker gene expression was verified. We analyzed mRNA levels of genes encoding the E1B 19K/Bcl-2-binding protein Nip3 (BNIP3) and fructose-bisphosphate aldolase (ALDOA), 2 established hypoxia marker genes33. RT-qPCR was performed as described for the TET mRNA concentration assays, and differential expression was calculated using the ΔΔCt method34. We moreover excluded that the increase in HIF1α protein concentrations was secondary to a transcriptional upregulation, by assessing HIF1A mRNA expression in parallel. mRNA concentrations were expressed relative to normoxic controls (21% O2). Differences in mRNA concentration were assessed using a Student’s t-test on 5 or more independent replicates grown on different days. Validation of hypoxia induction and Tet protein expression Western blotting for Hif1α, Tet1, Tet2 and Tet3 To assess HIF1α protein stabilization, proteins were extracted from cultured cells as follows: cells were placed on ice, and washed twice with ice-cold PBS. Proteins were extracted with extraction buffer (50 mM Tris HCl, 150 mM NaCl, 1% Triton X-100, 0.5% Na-deoxycholate and 0.1% SDS) with 1× protease inhibitor cocktail. Protein concentrations were determined using a bicinchoninic acid protein assay (BCA, Thermo Scientific) following the manufacture’s protocol, and an estimated 60 µg protein was loaded per well on a NuPAGE Novex 3-8% Tris-Acetate Protein gel (Life Technologies), separated by electrophoresis and blotted on polyvinylidene fluoride membranes. Membranes were activated with methanol and washed, and incubated with antibodies targeting β-actin (4967, Cell Signaling), Tet1 (09-872, Millipore) and Tet3 (61395, Active Motif), at 1:1000 dilution, targeting Tet2 (124297, Abcam) at 1:250 dilution, and targeting HIF-1α (C-Term) (Cayman Chemical Item 10006421) at 1:3000 dilution. Secondary antibodies and detection were according to routine laboratory practices. Western blotting was done on 6 independent replicates grown on different days. Analysis of HIF1β target genes using ChIP-seq To confirm that hypoxia-associated differential expression of TET genes is induced by the HIF pathway, we performed HIF1β ChIP-seq. Because HIF1β is the obligate binding partner of all 3 HIFα proteins stabilized and activated upon hypoxia35, HIF1β ChIP-seq reveals all direct HIF target genes. Chromatin immunoprecipitation 25-30×106 cells were incubated in hypoxic conditions for 16 hours. Cultured cells were subsequently immediately fixed by adding 1% Formaldehyde (16% Formaldehyde (w/v), Methanol-free, Thermo Scientific) directly in the medium and incubating for 8 minutes. Fixed cells were incubated with 150 µM of glycine for 5 min to revert the cross-links, washed twice with ice-cold PBS 0.5% Triton-X100, scraped and collected by centrifugation (1000 ×G 5min at 4°C). The pellet was resuspended in 1400 µL of RIPA buffer (50 mM Tris-HCl pH 8, 150 mM NaCl, 2 mM EDTA pH 8, 1% Triton-X100, 0.5% Sodium deoxycholate, 1% SDS, 1% protease inhibitors) and transferred in a new eppendorf tube. The lysate was homogenized by passing through an insulin syringe, and incubated on ice for 10 min. The chromatin was sonicated for 3 min by using a Branson 250 Digital Sonifier with 0.7 s ‘On’ and 1.3 s ‘Off’ pulses at 40% power amplitude, yielding a size of 100 to 500 bp. The sample was kept ice-cold at all times during the sonication. The samples were centrifuged (10 min at 16000 ×G at 4°C) and the supernatant were transferred in a new eppendorf tube. The protein concentration was assessed using a BCA assay. Fifty µL of shared chromatin was used as “input” and 1.4 µg of primary ARNT/HIF-1β monoclonal antibody (NB100-124, Novus) per 1 mg of protein was added to the remainder of the chromatin, and incubated overnight at 4°C in a rotator. Pierce Protein A/G Magnetic Beads (Life Technologies) were added to the samples in a volume that is 4X the volume of the primary Ab and incubated at 4°C for at least 5 hours. A/G Magnetic Beads were collected and the samples were washed 5 times with the washing buffer (50 mM Tris-HCl, 200 mM LiCl, 2 mM EDTA, pH 8, 1% Triton, 0.5% Sodium deoxycholate, 0.1% SDS, 1% protease inhibitors), and twice with TE buffer. The A/G magnetic beads were resuspended in 50 µL of TE buffer, and 1.5 µL of RNAse A (200 units, NEB, Ipswich, MA, USA) were added to the A/G beads samples and to the input, incubated for 10 minutes at 37°C. After addition of 1.5 µL of proteinase K (200 units) and overnight incubation at 65°C, the DNA was purified using 1.8× volume of Agencourt AMPure XP (Beckman Coulter) according to the manufactory instructions, and then eluted in 15 µL of TE buffer. The input DNA was quantified on NanoDrop. ChIP-seq, mapping and analysis Five µg of input and all of the immunoprecipitated DNA were converted into sequencing libraries using the NEBNext DNA library prep master mix set. A single end of these libraries was sequenced for 50 bases on a HiSeq 2000, mapped using Bowtie and extended for the average insert size (250 bases). ChIP peaks were called by Model-based Analysis for ChIP-Seq36, with standard settings and using a sequenced input sample as baseline. Patient-derived xenografted tumors Patient-derived xenografts To assess whether tumor-associated hypoxia reduces 5hmC levels in vivo, redundant material from 2 endometrial tumors and a breast tumor, removed during surgery, was grafted in the interscapular region of nude mice. Informed consent was obtained from the patient, following the ethical approval of the local ethical committee. All animal experiments were approved by the local ethical committee (P098/2014). Each tumor was allowed to grow until 1 cm3, after which it was harvested. 10% of this tumor was reimplanted in a nude mouse, and the tumor was thus propagated for 3 generations until it was used for this experiment. To mark hypoxic areas, mice were injected with pimonidazole (60 mg/kg, Hypoxyprobe, Massachusetts, USA) i.p. 1 hour before sacrifice. Immunofluorescence staining and analysis Tumors were harvested, fixed in formaldehyde and embedded in paraffin using standard procedures. Slides were deparafinated and rehydrated 2 xylene baths (5 minutes), followed by 5 times 3 minutes in EtOH baths at decreasing concentrations (100%, 96%, 70%, 50% and water) and a 3 minute Tris-buffered saline (TBS; 50 mM Tris, 150 mM NaCl, pH 7.6) bath. The following antibodies were used for immunofluorescence staining: primary antibodies were FITC-conjugated mouse anti-pimonidazole (HP2-100, Hydroxyprobe), rabbit anti-5hmC (39791, Active Motif), rat anti-polyoma middle T (AB15085, Abcam), rat anti-CD31 (557355, BD Biosciences), rat anti-CD45 (553076, BD Biosciences), rabbit anti-Ki67 (AB15580, Abcam) and mouse anti-pan cytokeratin (C2562, Sigma). Secondary antibodies were Alexa fluor 405-conjugated goat anti-rabbit (A31556, Thermo Fisher), Alexa Fluor 647 conjugated goat anti-rat (A-21247, Life technologies), peroxidase-conjugated goat anti-FITC (PA1-26804, Pierce), biotinylated goat anti-rat (A10517, Thermo Fisher) and biotinylated goat anti-rabbit (E043201, Dako). Signal amplification was done using the TSA Fluorescein System (NEL701A001KT, Perkin Elmer) or the TSA Cyanine 5 System (NEL705A001KT, Perkin Elmer). Different protocols were implemented depending on the epitopes of interest. Staining for the following epitopes was combined: CD45, 5hmC, pimonidazole and DNA; PyMT, 5hmC, pimonidazole and DNA; Ki67, pimonidazole and DNA; CD31 and pimonidazole; and pan-cytokeratin, 5hmC, pimonidazole and DNA. Antigen retrieval for CD31, CD45 and pan-cytokeratin was done by a 7 min trypsin digestion, for pimonidazole and Ki67 using AgR at 100°C for 20 min, followed by cooling for 20 min. Slides were washes in TBS for 5 min, permeabilized in 0.5% Triton-X100 in PBS for 20 min. For 5hmC antigen retrieval, slides were next denatured in 2 N HCl for 10 min, with the HCl being neutralized for 2 min in borax, 1% in PBS pH 8.5, and washed twice for 5 min in PBS. For all slides, endogenous peroxidase activity was quenched using H2O2 (0.3% in MeOH), followed by three 5 min washes in TBS. Slides were blocked using pre-immune goat serum (X0907, Dako; 20% in TNB; TSA Biotin System kit, Perkin Elmer, Waltham, MA). Binding of primary antibodies (anti-5hmC, anti-CD45, anti-CD31 and anti-pan cytokeratin or FITC-conjugated anti-pimonidazole; all 1/100 in TNB) was allowed to proceed overnight. Slides were washed 3× in TNT (0.5% Triton-X100 in TBS) for 5 min, after which secondary antibodies (all 1/100 in TNB with 10% pre-immune sheep serum) were allowed to bind for 45 min: sheep-anti-FITC-PO (for pimonidazole), goat anti-rabbit-Alexa Fluor 405 (for 5hmC), goat anti-rat-Alexa Fluor 647 (for CD45), and biotinylated goat anti-mouse (for pan-cytokeratin). Slides were washed 3× 5 min in TNT, after which signal amplification was done for 8 min using Fluorescein Tyramide (1/50 in amplification diluent). Slides stained for pimonidazole that required co-staining slides for Ki67 or PyMT, or slides stained for pan-cytokeratin that required co-staining for pimonidazole were subjected to a second indirect staining for the latter epitopes: after 5 min of TNT and 5 min of TBS, slides were quenched again for peroxidase activity using H2O2 and blocked using pre-immune goat serum, prior a second overnight round of primary antibody binding (anti-Ki67, FITC-anti-pimonidazole or anti-PyMT, all 1/100). The next day, 3× 5 min washes with TNT were followed by a 1 h incubation with a biotinylated goat anti-rabbit antibody (for Ki67) or goat anti-rat (for PyMT), again 3× 5 min washes with TNT, a 30 min incubation with peroxidase conjugated to streptavidine (for Ki67 and PyMT) or to anti-FITC (for pimonidazole), again 3× 5 min washes with TNT and signal amplification for 8 min using, for pimonidazole, Fluorescein Tyramide and for others Cyanine 5 Tyramide (1/50 in amplification diluent). Finally, slides were stained with propidium iodide + RNAse (550825; BD biosciences) for 15 min, washed for 5 min in PBS and mounted with Prolong Gold (Life Technologies). Slides were imaged on a Nikon A1R Eclipse Ti confocal microscope. 3-5 sections per slide were imaged, and processed using Image J. More specifically, nuclei were identified using the propidium iodide signal, and nuclear signal intensities for Fluorescein and Cy3 (pimonidazole and 5hmC) measured. Analyses were exclusively performed on slide regions showing a regular density and shape of nuclei, in order to avoid inclusion of acellular or necrotic areas. The pimonidazole signal will also not stain necrotic/acellular areas 37, and was used to stratify viable cell nuclei into normoxic (pimonidazole negative) and hypoxic (pimonidazole positive) regions; and the 5hmC signal in both populations was compared using ANOVA. PyMT-negative and CD45-positive cells were counted directly. The fraction of pimonidazole and CD31-positive areas was directly quantified using ImageJ across 10 images per slide. Metabolite levels Metabolite and protein extraction For metabolite extractions, 12-well cell culture dishes were placed on ice and washed twice with ice-cold 0.9% NaCl, after which 500 µL of ice-cold 80% methanol was added to each well. Cells were scraped and 500 µL was transferred to a vial on ice. Wells were washed with 500 µL 80% methanol, which was combined with the initial cell extracts. The insoluble fraction was pelleted at 4°C by a 10 minute 21,000×G centrifugation. The pellet (containing the proteins) was dried, dissolved in 0.2 N NaOH at 96°C for 10 minutes and quantified using a bicinchoninic acid protein assay (BCA, Pierce, Erenbodegem, Belgium), whereas the supernatant fraction was processed for metabolite profiling. Derivation and measurement of metabolites The supernatant fraction containing metabolites was transferred to a new vial and dried in a Speedvac. The dried supernatant fraction was dissolved in 45 µL of 2% methoxyamine hydrochloride in pyridine and held for 90 minutes at 37°C in a horizontal shaker, followed by derivatization through the addition of 60 µL of N-(tert-butyldimethylsilyl)-n-methyl-trifluoroacetamide with 1% tert-butyldimethylchlorosilane and a 60 minute incubation at 60°C. Samples were subsequently centrifuged for 5 minutes at 21,000 ×G, and 85 µL was transferred to a new vial and analysed using a gas-chromatography based mass spectrometer (triple quadrupole, Agilent) operated in Multiple Reaction Monitoring (MRM) mode. Analysis of metabolite concentrations For each sample, metabolite measurements were normalized per sample to the corresponding protein concentration estimates, and expressed relative to control-treated samples. Four technical replicates were run for each sample, and the experiment was repeated 4 times using independent samples (n=16). Differences in metabolite concentration were assessed using a two-tailed paired Student’s t-test or using analysis of variance with post-hoc Tukey HSD when repeated measures were compared. ROS measurement using 2',7'-dichlorodihydrofluorescein diacetate MCF7 cells were cultured in 24 well plates and exposed to 21% (control) or 0.5% O2 (hypoxia) for 24 hours. DMEM used for staining was pre-equilibrated to the required O2 tension, and all steps performed at 21% (control) or 0.5% O2 (hypoxia) using a glove box. The cells were washed 2× with 500 µL DMEM, and incubated for 30 min in 2',7'-dichlorodihydrofluorescein diacetate (DCF-DA; 10 μM) in 500µL DMEM, keeping 2 wells unstained by DMEM without DCF-DA. Cells were treated with the indicated concentrations of H2O2 in DMEM for 30 min at 37 °C, and fixed by adding 33.3 µL of 16% methanol-free paraformaldehyde (Thermo Fisher) for 8 min at RT. The fixative was quenched using glycine (150 µM), cells were washed 2× in ice-cold PBS, scraped to detach them and transfer them to pre-cooled FACS tubes over cell strainers. Cells were kept on ice until they were analysed by flow cytometry using a FACSVerse (BD Biosciences). Nuclear ROS measurement using Nuclear Peroxy Emerald 1 MCF7 cells were seeded on 12 well glass bottom plates and after 24 h exposed to 21% (control) or 0.5% O2 (hypoxia) for 24 h. PBS used for subsequent staining was pre-equilibrated to the required O2 tension, and all washing, treatment and staining steps were performed at the appropriate O2 tension (21% or 0.5%) using a glove box. Cells were loaded with Nuclear Peroxy Emerald 1 (NucPE1; 5 μM)38,39 and Hoechst 33342 (10 μg/mL) in PBS for 15 min at 37 °C. After washing 3× in PBS, control cells were incubated with H2O2 (0.5 mM in PBS) as a positive control, or with water (control and hypoxia cells) in PBS at 37 °C for 20 min. Cells were washed 3× in PBS, placed on ice and immediately imaged by confocal microscopy. The nuclear NucPE1 signal was measured, and averaged across >100 nuclei per replicate using ImageJ. This experiment was repeated 5 times on different days, and signals compared using a t-test. Cell growth measurement using Sulforhodamine B 5,000 cells/well were seeded in three 96-well plates. After 48 h, one plate was fixed using trichloroacetic acid (3.3% wt/vol) for 1 h at 4 °C, one plate incubated for 24 h at 37 °C under hypoxic and one under control conditions (resp. 0.5% and 21% O2). The latter 2 plates were subsequently also fixed using trichloroacetic acid (3.3% wt/vol) for 1 h at 4 °C, and all 3 plates were next analyzed using the In Vitro Toxicology Assay Kit, Sulforhodamine B-based (Sigma) as per the manufacturers instructions. Growth inhibition was calculated as described40. siRNA transfection siRNA ON-TARGETplus SMART pools (Thermo) were diluted in Optimem I reduced serum medium using Lipofectamine RNAiMAX (Life technologies) to reverse transfect MCF7 cells in 10 cm dishes (for DNA) or 6 well plates (for RNA). Cells were transfected 72 h before RNA and DNA extraction as described. Hydroxylation assay using nuclear extracts MCF7 cells were cultured for 24 h under control or hypoxic conditions (resp. 21 and 0.5% O2), chilled on ice and processed for extraction of nuclear proteins using the NE-PER Nuclear and Cytoplasmic Extraction Kit (Thermo Scientific). The activity of control and hypoxic extracts was assessed in parallel using the Colorimetric Epigenase 5mC-Hydroxylase TET Activity/Inhibition Assay Kit (Epigentek, Farmingdale, USA) according to manufacturers instructions. Reactions were allowed to proceed for one hour, after which washing and detection of 5hmC were done according to manufacturers instructions. Differences between hypoxia and control were analyzed using ANOVA, for 5 independent experiments. DNA hydroxymethylation assay using purified Tet enzyme The genomic DNA used in this assay was extracted from Tet-triple-knockout ES cells (a gift from Prof. Guo-Liang Xu, State Key Laboratory of Molecular Biology, CAS, Shanghai, China), and it therefore was devoid of 5hmC41. To enable efficient denaturation, it was digested using MseI prior to the assay and purified using solid phase reversible immobilisation paramagnetic beads (Agencourt AMPure XP, Beckman Coulter, USA). The assays were performed in Whitley H35 Hypoxystations (don Whitley Scientific, UK) at 37° C, 5% CO2, N2, plus the following oxygen tensions: 0.1%, 0.25%, 0.5%, 1%, 2.5%, 5%, 10% and 21%. Hypoxystations were calibrated less than 1 month prior to all experiments. Optimized assay components were as follows: 1.0 µg/µL bovine serum albumin (New England Biolabs), 50 mM Tris (pH 7.8), 100 µM dithiothreitol (Life Technologies), 2ng/µL digested gDNA, 250 µM α-ketoglutarate, 830 µM ascorbate, 200 µM FeSO4 and 45 ng/µL Tet1 enzyme (Wisegene, USA). The major assay components (H20, BSA and Tris) used for all samples were allowed to pre-equilibrate at 0.1% O2 for 1 hour. These and the remaining assay buffer components (<100 µL) were then pre-equilibrated at the desired oxygen tension for 15 min, and mixed prior to addition of Tet1 enzyme in a total reaction volume of 25 µL. Reactions were allowed to proceed for 3 min, longer incubations showed a decrease in activity. Reactions were stopped with 80 mM EDTA and stored at -80° C. To measure the resulting 5hmC content of the DNA, reactions were diluted to 100 µL, denatured for 10 min at 98° C and analysed in duplicate using the Global 5-hmC Quantification Kit (Active Motif) following manufacturers instructions. Michaelis-Menten and Lineweaver-Burk plots and the resulting KM values were estimated using R. Hypoxia-induced changes in genomic distribution of 5(h)mC in MCF7 cells DIP-seq To assess where in the genome the levels of 5mC and 5hmC were altered, we performed DNA immunoprecipitations coupled to high-throughput sequencing (DIP-seq). MCF7 cells were selected for these experiments as they were a cancer cell line with high levels of 5hmC and expression of TETs under control conditions, and a cell growth that is unaffected by hypoxia, thus enabling us to study effects of hypoxia on TET activity in a cell line that shows high endogenous activity, but that is isolated from hypoxia-induced changes in cell proliferation. MCF7 cell culture and DNA extractions were as described for LC/MS analyses. Library preparations and DNA immunoprecipitations were as described42, using established antibodies targeting 5mC (clone 33D3, Eurogentec, Liege, Belgium) and 5hmC (Active Motif catalogue number.39791, La Hulpe, Belgium). For 5hmC-DIP-seq, paired barcoded libraries prepared from DNA of hypoxic and control samples were mixed prior to capture, to enable a direct comparison of 5hmC-DIP-seq signal to the input. A single end of these libraries was sequenced for 50 bases on a HiSeq 2000, mapped using Bowtie and extended for the average insert size (150 bases). Mapping statistics are summarized in Supplementary table 11. For analysis of sequencing data, MACS peak calling, read depth quantification and annotation with genomic features as annotated in EnsEMBL build 77 was done using using SeqMonk. Differential (hydroxy-)methylation was quantified by EdgeR43, using either 3 or 5 independent pairs of control and hypoxic samples (resp. for 5hmC-DIP-seq and 5mC-DIP-seq). These cells were cultured and exposed to hypoxia (0.5% O2) or control conditions (21% O2) on different days. Results were reported for 5hmC peak areas that exhibited a change significant at a P<0.05 and 5% FDR. Target enrichment BS-seq using SeqCapEpi To confirm enrichment of 5mC at gene promoters using an independent method, DNA libraries were prepared using methylated adapters and the NEBNext DNA library prep master mix set following manufacturer recommendations. Libraries were bisulfite-converted using the Imprint DNA modification kit (Sigma) as recommended, and PCR amplified for 12 cycles using barcoded primers (NEB) and the KAPA HiFi HS Uracil+ ready mix (Sopachem, Eke, Belgium) according to manufacturers instructions. Fragments were selected from these libraries using the SeqCapEpi CpGiant Enrichment Kit (Roche) following the manufacturers instructions, sequenced from both ends for 100 bases on a HiSeq 2000. For analyzing these sequences, sequencing reads were trimmed for adapters using TrimGalore and mapped on a bisulfite-converted human genome (GRCh37) using BisMark. The number of methylated and unmethylated cytosines in captured regions were quantified using Seqmonk for each experiment. Differential methylation of regions of interest was assessed by Fisher’s exact test and for 5 independent replicates grown on different days. t-scores were averaged following Fisher’s method. Mapping statistics are summarized in Supplementary table 11. RNA-seq To assess the impact of the increased 5mC occupancy at gene promoters on their expression, RNA-seq was performed. Briefly, total RNA was extracted using TRIzol (Invitrogen), and remaining DNA contaminants in 17-20ug of RNA was removed using Turbo DNase (Ambion) according to the manufacturers instruction. RNA was repurified using RNeasy Mini Kit (Qiagen). Ribosomal RNA present was depleted from 5ug of total RNA using the RiboMinus Eukaryote System (Life technologies). cDNA synthesis was performed using SuperScript® III Reverse Transcriptase kit (Invitrogen). 3 µg of Random Primers (Invitrogen), 8 µL of 5× First-Strand Buffer and 10 µL of RNA mix was incubated at 94°C for 3 min and then at 4°C for 1 min. Then, 2 µL of 10 mM dNTP Mix (Invitrogen), 4 µL of 0.1 M DTT, 2 µL of SUPERase• In™ RNase Inhibitor 20U/ µL (Ambion), 2 µL of SuperScript™ III RT (200 units/µL) and 8 µL of Actinomycin D (1µg/µL) were added and the mix were incubated 5 min at 25°C, 60 min at 50°C and 15 min at 70°C to heat inactivating the reaction. The cDNA was purified by using 80 µL (2× volume) of Agencourt AMPure XP and eluted in 50 µL of the following mix: 5 µL of 10X NEBuffer 2, 1.5 µL of 10 mM dNTP mix (10mM dATP, dCTP, dGTP, dUTP, Sigma), 0.1µL of RNaseH (10 U/µL, Ambion), 2.5 µL of DNA Polymerase I Klenov (10U/µL, NEB) and water until 50 µL. The eluted cDNA was incubated for 30 min at 16°C, purified by Agencourt AMPure XP and eluted in 30 µL of dA-Tailing mix (2 µL of Klenow Fragment, 3 µL of 10X NEBNext dA-Tailing Reaction Buffer and 25 µL of water). After 30 min incubation at 37°C, the DNA was purified by Agencourt AMPure XP, eluted in TE buffer and quantified on NanoDrop. Subsequent library preparation was done using the DNA library prep master mix set and sequencing was performed as described for ChIP-seq. Expression levels (reads per million) of genes displaying significant increases in methylation at their gene promoter, as determined using SeqCapEpi, was compared between control and hypoxic samples using t-test. Mapping statistics are summarized in Supplementary table 11. TCGA samples and data analysis Sample description From the TCGA pan-cancer analysis, we selected all solid tumor types for which >100 tumors were available with both gene expression data (RNA-seq) and DNA methylation data (Illumina Infinium HumanMethylation450 BeadChip). These were 408 bladder carcinomas, 691 breast carcinomas, 243 colorectal adenocarcinomas, 520 head and neck squamous cell carcinomas, 290 kidney renal cell carcinomas, 430 lung adenocarcinomas, 371 lung squamous cell carcinomas, and 188 uterine carcinomas, representing in total 3,141 unique patients. Corresponding RNA-seq read counts as well as DNA methylation data from Infinium HumanMethylation450 BeadChip arrays were downloaded from the TCGA server. Breast tumor subtype was annotated for 208 tumors, and for the remaining tumors imputed by unsupervised hierarchical clustering of genes in the PAM50 gene expression signature44. Other clinical and histological variables were available for >95% of tumors, and missing values were encoded as not available. Gene mutation data was available for 129 bladder carcinomas, 646 breast carcinomas, 200 colorectal adenocarcinomas, 306 head and neck squamous cell carcinomas, 241 kidney renal cell carcinomas, 182 lung adenocarcinomas, 74 lung squamous cell carcinomas, and 3 uterine carcinomas. Stratification of tumors for hypoxia and proliferation To identify which of these tumor samples were hypoxic or normoxic, we performed unsupervised hierarchical clustering based a modification (Ward.D of the clusth function in R`s stats package) of the Ward error sum of squares hierarchical clustering method45, on normalized log-transformed RNA-seq read counts for 14 genes that make up the hypoxia metagene signature (ALDOA, MIF, TUBB6, P4HA1, SLC2A1, PGAM1, ENO1, LDHA, CDKN3, TPI1, NDRG1, VEGFA, ACOT7 and ADM)25. In each case the top 3 subclusters identified were annotated as normoxic, intermediate and hypoxic. To identify which of these tumor samples were high- or low-proliferative, we performed unsupervised hierarchical clustering based a modification (Ward.D of the clusth function in R`s stats package) of the Ward error sum of squares hierarchical clustering method45, and this for all genes annotated to an established tumor proliferation signature (MKI67, NDC80, NUF2, PTTG1, RRM2, BIRC5, CCNB1, CEP55, UBE2C, CDC20 and TYMS)46. Tumors in the top 2 subclusters identified were labeled as high- or low-proliferative. Analysis of the top 1000 CpGs most hypermethylated versus normal tissue To identify tumor-associated HM events, we compared 450K methylation data from tumors and normal tissues. All available DNA methylation data from normal tissue (matched or unmatched to tumor samples, on average 59 per tumor type, representing 472 in total, range = 21-160) were downloaded. For each of the 8 tumor types investigated, we selected the top 1,000 CpGs that showed the highest average tumor-associated increases in DNA methylation. Per tumor type, unsupervised hierarchical clustering based on a modification of the Ward error sum of squares hierarchical clustering method (Ward.D of the clusth function in R`s stats package)45 annotated the first 3 clusters identified as having low, intermediate and high hypermethylation. Cluster co-membership for methylation and hypoxia metagene expression were analysed using the Cochran-Armitage test for trend. Analyses using the top 100, 500, 5,000 or 10,000 CpGs yielded near identical results (not shown). Analysis of HM events We next applied a method to identify those CpGs that exhibit exceptional increases in HM but that are hypermethylated only in a subset of all tumors. Such more rare events are typically found in cancer, where HM inactivates a gene in only a subset of tumors. HM of individual CpGs at gene promoters (i.e. on average 3.7 CpGs per promoter are represented on the 450K array) in individual tumors was assessed as follows: To achieve a normal distribution, all β-values were transformed to M-values47 using M = log2(β/(1-β)). For each tumor type, the mean μ and standard deviation σ of the M value across all control (normoxic) tumors was next calculated, irrespective of mutational status, for each CpG, and used to assign Z-values to each CpG in each tumor using Z = (M - μ)/σ. These Z-values describe the deviation in normal methylation variation for that probe. To identify CpGs that display an extreme deviation, we selected those for which the Z-value exceeded 5.6 (i.e. the mean plus 5.6 times the standard deviation, corresponding to a Bonferroni-adjusted P-value of 0.01); they were considered as hypermethylation events in that particular tumor. This analysis was preferred over Wilcoxon-based models that assess differences in the average methylation level between subgroups, as the latter do not enable the identification or quantification of the more rare HM events in individual promoters or CpGs. To identify genes with frequently hypermethylated CpGs in their promoter, the number of HM events in that promoter was counted in all tumors, and contrasted to the expected number of HM events in that promoter (i.e. the general HM frequency × the number of CpGs assessed in that promoter × the number of tumors) using Fisher’s exact test. Genes with an associated Bonferroni-adjusted P-value below 0.01 were retained and considered as frequently hypermethylated in that tumor type. To assess what fraction of these HM events are hypoxia-related, we assumed that the fraction of events detected under normoxia was hypoxia–unrelated, and that all excess events detected in intermediate and hypoxic tumors were hypoxia-related. For example, in 691 breast carcinomas, 0.25% of CpGs were hypermethylated in 251 normoxic tumors, 0.81% in 350 intermediate and 1.40% in 90 hypoxic tumors. So, 0.56% and 1.15% of HM events in respectively intermediate and hypoxic tumors were hypoxia-related. Taking into account the number of tumors, 0.25% of HM events (i.e. (0.25% × 251 + 0.25% × 350 + 0.25% × 90) ÷ 691) are not hypoxia-related, and 0.43% are hypoxia related (i.e. (0% × 251 + 0.56% × 350 + 1.15% × 90) ÷ 691). Hence, 63% of all HM events (i.e. 0.43÷(0.43+0.25)). To assess the contribution of hypoxia to HM relative to other covariates, partial R2 values were calculated for the contribution of each covariate in a linear model, and compared to the total R2 achieved by the model. To identify genes with CpGs in their promoter that are more frequently hypermethylated in hypoxic than normoxic tumors, the number of HM events in that promoter was counted in all hypoxic tumors, and contrasted to the number found in normoxic tumors. Differences in frequencies were assessed using Fisher’s exact test, and genes with a Bonferroni-adjusted P<0.01 were retained and considered hypermethylated upon hypoxia. Enrichment of ontologies associated with these genes was assessed using Fisher’s exact test as implemented in R`s topGO package. Analysis of the impact of HM events on the expression of associated genes To enable a direct comparison between the expression of different genes, we transformed gene expression values (reads per million) to their respective z-scores. To assess the impact of HM events on the expression of associated genes, we compared the expression z-scores of all frequently HM genes that contain one or more HM events in their promoter (i.e., on average each promoter contains 3.7 CpGs; if one of these is hypermethylated the associated gene is considered hypermethylated in that particular tumor), to the expression of all frequently HM genes that do not contain a HM event. The effect of HM on gene expression was plotted for the 8 main tumor types stratified into normoxic, intermediately hypoxic and hypoxic tumors, and for glioblastomas stratified into normoxic, intermediately hypoxic, hypoxic and IDH-mutant tumors (n=4). The difference in expression z-scores between genes not carrying and carrying a HM event in their promoter was assessed using a t-test. Analysis of the impact of frequent mutations on the occurrence of HM events To assess the impact of somatic mutations on hypoxia-associated HM frequencies, we analyzed the top 20 genes described to be most frequently mutated in the pan-cancer analysis,24 and supplemented this list with genes known to cause HM upon mutation (i.e. IDH1, IDH2, SDHA, FH, TET1, TET2 and TET3). Mutations in IDH1 and IDH2 were retained if they respectively affected amino acid R132, and amino acids R140 or R172. Mutations in other genes were scored using Polyphen, and only mutations classified as probably damaging were retained. 7 mutations were found in lung tumors, 3 mutations in colorectal tumors, 8 mutations in breast tumors and 6 mutations (all IDH1R132) in glioblastomas. None of these mutations were enriched in hypoxic subsets. In multivariate analyses of variance, in each of the tumor types analyzed, mutations in these genes were significantly associated with increased HM frequencies, but also hypoxia was independently and significantly associated with the HM frequency. Correlation between HM and expression of TET or DNMT enzymes Gene expression values (reads per million) of DNMT and TET enzymes were determined for each tumor using available RNA-seq data. The number of HM events at significantly hypermethylated genes in each tumor was determined as described above. Hypermethylation in each tumor was subsequently correlated to TET or DNMT gene expression in that tumor, correcting for hypoxia and proliferation status, using ANOVA. 5mC and 5hmC profiling using 450K arrays for 24 lung tumors Tumor samples Newly diagnosed and untreated non-small-cell lung cancer patients scheduled for curative-intent surgery were prospectively recruited. Included subjects had a smoking history of at least 15 pack-years. The study protocol was approved by the Ethics Committee of the University Hospital Gasthuisberg (Leuven, Belgium). All participants provided written informed consent. In the framework of a different project48, RNA-seq was performed on 39 tumors from these patients. Gene expression for these samples was clustered for their hypoxia metagene signature25. This yielded 2 clear clusters, containing respectively 24 and 15 normoxic and hypoxic tumors. Twelve samples were randomly selected from each cluster, in order to perform 5hmC and 5mC profiling. Illumina Infinium HumanMethylation450 BeadChips For Tet-assisted bisulfite (TAB)-chip, DNA was glycosylated and oxidized as described49, using the 5hmC TAB-Seq Kit (WiseGene, Chicago, USA). Subsequently, bisulfite conversion, DNA amplification and array hybridization were done following manufacturers instructions. Analysis of TAB-chip and BS-chip Data processing was largely as described50. In brief, intensity data files were read directly into R. Each sample was normalized using Subset-quantile within array normalization (SWAN) for Illumina Infinium HumanMethylation450 BeadChips49. Batch effects between chips and experiments were corrected using the runComBat function from the ChAMP bioconductor package51. For obtaining 5mC-specific beta values, TAB-chip generated normalized beta values were substracted from the standard 450K generated normalized beta values, exactly as described50. Murine cancer models All the experimental procedures were approved by the Institutional Animal Care and Research Advisory Committee of the KU Leuven. Hypoxia induction using sFlk1-overexpression For sFlk1-overexpression studies, male Tg(MMTV-PyMT) FVB mice were intercrossed with WT FVB female mice. Female pups of the Tg(MMTV-PyMT) genotype were retained, and tumors allowed to develop for 9 weeks. Subsequently, 2.5 µg of plasmid (Flk1-overexpressing or empty vector; randomly assigned within litter mates) per gram of mouse body weight was introduced in the blood stream using hydrodynamic tail vein injections52. Flk1 overexpression was monitored at 4 days after injection and at the day of sacrifice (18 days after the injection), by eye bleeds followed by an enzyme-linked immunosorbent assay for sFlk1 (R&D Systems, Abingdon, UK) in blood plasma. At 12 weeks of age, mice were sacrificed and mammary tumors harvested blinded for treatment. Hypoxia reduction using Phd2 haplodeficiency For the Phd2+/- experiments, male Tg(MMTV-PyMT) FVB mice were intercrossed with female Phd2-/+ mice, yielding litters of which half have either a Tg(MMTV-PyMT) genotype or a Tg(MMTV-PyMT);Phd2-/+ genotype. For the Phd2wt/fl experiments, male Tg(MMTV-PyMT) FVB mice were intercrossed with female Tie2-cre;Phd2wt/fl mice as described27, yielding litters of which half have either a Tie2-cre;Tg(MMTV-PyMT);Phd2wt/wt genotype or a Tie2-cre;Tg(MMTV-PyMT);Phd2-/+ genotype. At 16 weeks of age, female mice were sacrificed and mammary tumors harvested. qPCR analysis of expression of Tets and marker genes RNA was extracted from fresh-frozen tumors (stored at -80°C) using TRIzol (Life Technologies), and converted to cDNA and quantified as described for the cell lines. TaqMan probes and primers (IDT, Leuven, Belgium or Life technologies) are described under Supplementary table 12. TAB-sequencing (TAB-seq) of PyMT tumors TAB-seq libraries were prepared as described, using the 5hmC TAB-Seq Kit (WiseGene). DNA was bisulfite-converted using the EZ DNA Methylation-Lightning Kit (Zymo Research) and sequenced as described for SeqCapEpi experiments. Reads were mapped to the mouse genome (build Mm9) and further data processing was as for SeqCapEpi experiments. DNA from 3 independent tumors were selected per condition. TET oxidation efficiency was required to exceed 99.5% as estimated using a fully CG-methylated plasmid spike-in, 5hmC protection by glycosylation was 65% as estimated using a fully hydroxymethylated plasmid spike-in, bisulfite conversion efficiencies were estimated to exceed 99.8% based on nonCG methylation (=hmCpH %). Mapping statistics are summarized in Supplementary table 11. Targeted deep BS-seq As no murine capture kit was available for targeted BS-seq, a specific ampliconBS was developed for a set of 15 tumor suppressor gene promoters and 5 oncogene promoters. More specifically, DNA was bisulfite-converted using the Imprint DNA modification kit and amplified using the MegaMix Gold 2× mastermix and validated primer pairs. Per sample, PCR products were mixed to equimolar concentrations, converted into sequencing libraries using the NEBNext DNA library prep master mix set and sequenced to a depth of ~500×. Mapping and quantification were done as described for SeqCapEpi. The average and variance of methylation level M values in normal mammary glands were used as baseline, and amplicons displaying over 3 standard deviations more methylation (FDR-adjusted P<0.05) than this baseline were called as hypermethylated. At least 9 different tumors, each from different animals, were profiled per genotype or treatment, and differences in HM frequencies between sets of tumors were assessed using Mann-Whitney’s U-test. Statistics Data entry and analysis was performed in a blinded fashion. Statistical significance was calculated by two-tailed unpaired t-test (Excel) or analysis of variance (R) when repeated measures were compared. Data were tested for normality using the D’Agostino–Pearson omnibus test (for n > 8) or the Kolmogorov–Smirnov test (for n ≤ 8) and variation within each experimental group was assessed. Data are presented as means ± standard error of mean. DNA methylation and RNA-seq gene expression data distributions were transformed to a normal distribution by conversion to M values and log2 transformation respectively. Sample sizes were chosen based on prior experience for in vitro and murine experiments, or on sample and data availability for human tumor analyses. Other statistical methods (mostly related to specific sequencing experiments) are described together with the experimental details in other sections of the methods. Extended Data Extended data figure 1 Hypoxia-induced changes in 5hmC, 5mC and TET expression. Global 5hmC/C and 5mC/C content of DNA, TET1, TET2 and TET3 mRNA expression and hypoxia marker gene expression in 15 cell lines grown for 24 h under control (21% O2, white) or hypoxic (0.5% O2, red) conditions. TET mRNA copy number is expressed relative to B2M for human cell lines (HepG2, HT-1080, MCF10A, H358, MCF7, Hep3B, A549, H1299, SK-N-Be2c and SHSY5Y), and to Hprt for murine cell lines (LLC, N2A, 4T1, ESC-wt and ESC-Tet1-KO). Shown are cell lines derived from liver cancer (HepG2 and Hep3B), lung cancer (H358, A549, H1299 and LLC), breast cancer (MCF7 and 4T1), fibrosarcoma (HT1080), neuroblastoma (SK-N-Be2c and SHSY5Y), normal breast epithelium (MCF10A) and the inner cell mass of blastocyst-stage mouse embryos (ESC-wt and ESC-Tet1-KO). ALDOA and BNIP3 are expected to be increased, and HIF1A to be decreased upon hypoxia. The global 5fC content of ES cells is depicted, but was undetectable in cancer cell lines. Bars represent the mean ± s.e.m. of 5 different replicate samples. DNA and RNA from these replicates was extracted from cells derived from the same stock vial but grown on different days. * P<0.05, ** P<0.01, *** P<0.001 by paired t-tests. Extended data figure 2 Impact of hypoxia on TET expression. a, Changes in Tet1, Tet2 and Tet3 expression in mouse cell lines, at the protein level (top row, n=6) and the mRNA level (bottom row, n=5). middle row: representative immunoblot images of Hif1a, Tet1, Tet2 and Tet3. α-Tubulin serves as loading control, and expression of the corresponding coding gene (Tuba) was used to normalize mRNA expression, enabling a direct comparison of relative protein and relative mRNA expression changes. For the same reason, mRNA expression was depicted relative to control conditions, in contrast to the absolute levels shown in Extended data figure 1. Changes in Tet mRNA and protein expression correlate strongly (Pearson’s R: 0.855, P=4 × 10-4). For example, both 4T1 and N2A cells displayed increased Tet2 expression at the protein and mRNA level. Likewise, ES cells showed no pronounced changes at the protein nor at mRNA level. The overall expression of Tet enzymes was moreover not altered in any of these cell lines. For gel source data, see Supplementary figure 1. b, HIF1β ChIP-seq at the promoters of TET1, TET2 and TET3 and at hypoxia markers genes (BNIP3 and ALDOA), with peaks or promoter regions highlighted using colored boxes. Green and red boxes correspond to respectively overexpression and no overexpression (specified in the figure panel) of the corresponding gene, as determined using TaqMan in Extended data figure 1. Scale: reads per million reads and per basepair. c, (left) TET2 expression in MCF7 cells transfected with control (white) or TET2-targeting (purple) siRNAs, and (right) corresponding 5hmC levels as determined using LC/MS. d, 5hmC levels as determined using LC/MS, in wild-type (white) and Tet1-knockout (purple) ES cells grown under 21% (left) and 0.5% (right) oxygen tensions. Bars in c and d represent the mean ± s.e.m. of 5 replicate samples from cells derived from the same stock vial but grown on different days. * P<0.05, ** P<0.01, *** P<0.001 by paired t-tests (a, c, d) Extended data figure 3 Effects secondary to hypoxia. a-e, ROS production and redox state of MCF7 cells cultured for 24 h under control (21% O2, white) or hypoxic (0.5% O2, red) conditions. Shown are (a) GC/MS quantifications of changes in the cellular energy state as represented by the adenylate energy charge (AEC) (calculated as [ATP + 0.5 × ADP]/[ATP + ADP + AMP]); the cell’s reducing equivalents as represented by the relative NADH and NADPH levels (calculated as NADH/[NAD+ + NADH] and NADPH/[NADP+ + NADPH]) and the cell’s reductive capacity as represented by the levels of glutathione (calculated as GSH/[GSH + GSSG × 2]). b-c, Quantication (b) and representative FACS intensity traces (c) of total ROS levels in MCF7 cells exposed to hypoxia or H2O2, as assessed using 2',7'-dichlorodihydro-fluorescein diacetate (DCF-DA). d, Nuclear ROS in MCF7 cells as assessed using the Nuclear Peroxy Emerald 1 probe (NucPE1)39. MCF7 cells were exposed to 21% (control) or 0.5% O2 (hypoxia) for 24 h, after which live cells were loaded with NucPE1 (5 µM) and Hoechst 33342 (10 µg/mL) in O2 pre-equilibrated PBS for 15 minutes. After washing, control cells were incubated with H2O2 (0.5 mM in PBS) as a positive control, or with water (control and hypoxia cells) in PBS for 20 minutes. Cells were washed again and immediately imaged by confocal microscopy. Representative images are shown; scale: 50 µm. e, The nuclear NucPE1 signal, averaged across >100 nuclei and expressed relative to control conditions. f, LC/MS quantification of 8-oxoguanine concentrations in DNA of cells lines cultured for 24 h under control (21% O2, white) and hypoxic (0.5% O2, red) conditions. 8-oxoguanine serves as a marker of nuclear ROS53. g-i, GC/MS quantification of changes in the indicated metabolite levels in mouse embryonic stem cells (g), MCF10A cells (h) and MCF7 cells (i) grown for 24 h under control (21% O2, white), hypoxic (0.5% O2, red) or glutamine-free conditions (21% O2, black). j, Quantities of α-ketoglutarate and 2-hydroxyglutarate in MCF7 cells, expressed relative to α-ketoglutarate levels in MCF7 cells grown under control conditions (21% O2). k, LC/MS quantification of 5hmC levels in response to hypoxia (0.5% O2) and glutamine-free culture conditions. l, Growth of cell lines cultured for 24 h under control (21% O2, white) and hypoxic (0.5% O2, red) conditions, as assessed using a sulforhodamine B colorimetric assay. Changes in cell density after 24 h are depicted relative to control conditions (21% O2). m, IOX2-induced changes in the global 5hmC content of DNA, in TET mRNA expression and in hypoxia marker gene expression of 5 cell lines treated for 24 h with DMSO (carrier, white) or IOX2 (50 µM, blue). n, 5mC hydroxylation activity of nuclear lysates from MCF7 cells grown for 24 h under 21% or 0.5% O2 (white or red). Bars represent the mean ± s.e.m. of 5 (b, k, m), 6 (a, e), 16 (g-j) or 24 (l) samples prepared on different days. * P<0.05, ** P<0.01, *** P<0.001 by t-test (b, e, h-m). Extended data figure 4 Genomic profiles of 5mC and 5hmC. Shown are results from DIP-seq of DNA from MCF7 cells cultured for 24 h under 21% or 0.5% O2 (control and hypoxia), with examples of 5hmC-DIP-seq (top) and 5mC-DIP-seq (bottom) read depths (FPM, fragments per basepair per million fragments) at regions surrounding the transcription start site of NSD1, FOXA1 and CDKN2A. These show 5hmC loss (FDR<5%) and a 5mC gain that is more subtle, perhaps because the resolution of 5mC-DIP-seq is limiting: regions rich in 5hmC tend to be poorer in 5mC54, and thus have less substrate available for pull-down. 5mC-DIP-seq moreover captures all methylated sites, so most of the 5mC-DIP-seq signal does not derive from sites that are actively turning over 5hmC. Extended data figure 5 Impact of hypoxia on hypermethylation frequency in tumors. a, Immunofluorescence analysis of patient-derived tumor xenografts, stained for pimonidazole (PIMO, white), 5hmC (red), DNA (propidium iodide, blue) and pan-cytokeratin (green). Shown are representative images of a breast and 2 endometrial tumor xenografts. The inset on the right shows boxplots illustrating the signal in normoxic pimonidazole-negative nuclei (blue), and in hypoxic pimonidazole-positive nuclei (red). b, Hypoxia marker gene expression clusters, with the first 3 clusters used to define normoxic, intermediate and hypoxic tumors. c, Unsupervised clustering of 1,000 CpGs showing the highest average methylation increase in tumor versus corresponding normal tissues. The first 3 clusters were used to define tumors of low, intermediate and high HM. The color bar above the clusters annotates each tumor as normoxic, intermediate or hypoxic, as determined in Extended data figure 5b. d, Boxplots showing the relative expression (z-score) of genes in tumors wherein they have either 0 or ≥1 hypermethylation (HM) event in their promoter, stratified into normoxic, intermediate and hypoxic tumors (resp. blue, grey and red). Diamonds indicate means, boxplot wedges indicate 2 times the standard error of the median. Genes having ≥1 HM events in their promoters have a lower average expression level (P<0.01 for each tumor type). e, Fraction of genes having a promoter that is rich, intermediate or poor in CpGs, out of all gene promoters that are assessed on the 450K array (450K), and out of all gene promoters that are frequently hypermethylated in the indicated tumor types. f, Fraction of 1,742 TET-wildtype tumors and 39 TET-mutant that are normoxic, intermediate and hypoxic. P>0.2 for all comparisons. g, Cell proliferation marker gene46 expression clusters, with the first 2 clusters used to define high-proliferative and low-proliferative tumors. h, HM frequencies in low and high-proliferative tumors, with asterisks representing P-values from linear models correcting for variables specified in Supplementary table 8. i, Partial correlation coefficient (partial R2) estimates of the relative contribution of tumor characteristics (annotated in TCGA) to the variance in HM observed in these tumors. Partial R2 values were obtained from linear model estimation using ordinary least squares, and expressed as a fraction of the total variance (i.e. total R2) explained by the model when taking into account all indicated variables, as indicated between brackets under each tumor type. * P<0.05, ** P<0.01, *** P<0.001 by t-test (a) or by generalized linear model (h); BLCA bladder carcinoma; BRCA breast carcinoma, COAD colorectal adenocarcinoma, HNSC head and neck squamous cell carcinoma, KIRC kidney renal clear cell carcinoma, LUAD lung adenocarcinoma, LUSC lung squamous cell carcinoma, UCEC uterine corpus endometrial carcinoma. Extended data figure 6 Functional annotation of genes more frequently hypermethylated in hypoxic tumors. a, Ontology terms enrichment analysis of genes that are more frequently hypermethylated at their gene promoters in hypoxic than normoxic tumors, for 8 tumor types characterized in the TCGA Pan-Cancer effort. A representative set of terms is displayed, selected from terms enriched in most tumor types. P-values as defined by the grey-scale insert. Enrichment calculated using topGO. b, Selected examples of HM frequencies in the promoter of key tumor suppressor genes (PTEN, CDKN1C, ATM) more frequently hypermethylated in normoxic than hypoxic tumors. c, HM frequency in the promoter of selected genes involved in the processes indicated. P<0.05 for all genes (asterisks are not displayed). Bars in b and c represent the HM frequency ± s.e.m. P-values in (a) by Fisher’s exact test. Extended data figure 7 Impact of hypoxia on TET activity in human tumors. a, The t-value of correlation between HM and expression of TET or DNMT genes across 3,141 tumors of 8 tumor types (bladder, breast, colorectal, head and neck, kidney, lung adeno, lung squamous, and uterine carcinoma) profiled in TCGA for gene expression and DNA methylation, while correcting for tumor type, hypoxia and proliferation. The dotted line represents P<0.05, negative t-values represent inverse correlations. b, Hypoxia metagene signature applied to 63 glioblastoma multiforme tumors from TCGA. c, Boxplots showing the relative expression (z-score) of genes in tumors wherein they have either 0 or ≥1 hypermethylation (HM) event in their promoter, stratified into IHD1WT tumors that are normoxic (n=19), intermediate (n=21) and hypoxic (n=17) (resp. blue, grey and red), and IDH1R138-mutated tumors (n=4, yellow). Diamonds indicate means, boxplot wedges indicate 2 times the standard error of the median. Genes having ≥1 HM events in their promoters have a lower average expression level. No HM events were detected in IHD1WT normoxic tumors. d Hypoxia metagene signature applied to 12 normoxic and 12 hypoxic non-small-cell lung tumors. * P<0.05, *** P<0.001 by t-test (c). Extended data figure 8 5hmC, hypoxia and TSG HM in murine breast tumors. a, Frequency of HM events in the promoters of all genes, all oncogenes and all tumor suppressor genes (TS genes) as annotated28, in 695 human breast tumors available through TCGA and stratified into normoxic, intermediate and hypoxic subsets. b-c, DNA was extracted from 53 tumors developing in MMTV-PyMT mice of the indicated ages (c) or weights (d) and sequenced to a depth of ~500x. Plotted are z-scores of HM (y axis, exponential) for 15 TSGs, relative to the tumors from 11-week-old mice. The dotted line represents the threshold for a Bonferroni-adjusted P<0.05, and bold darker dots are used for tumors displaying significantly increased HM events. d, DNA extracted from 20 normal mammary glands from 14-week-old mice, PCR-amplified for 15 TSGs and sequenced to a depth of ~500x. Plotted are z-scores of HM relative to 11-week-old tumors. e, Staining of PyMT tumors for 5hmC (red), DNA (propidium iodide, blue), pimonidazole (white) and PyMT (green), and fraction of PyMT-positive cells in normoxic and hypoxic areas. The area outlined corresponds to the hypoxic, pimonidazole-positive section, arrowheads point to PyMT-negative cells, scale: 25µm. The bar chart inset illustrates the relative number of PyMT-positive cells in normoxic and hypoxic areas (resp. grey and red; n=19). f, Ki67-positive cells in PyMT tumors: representative image of staining for DNA (propidium iodide, blue), Ki67 (red) and pimonidazole (green); scale: 50 µm. The bar chart inset illustrates the quantification of Ki67-positive cells in normoxic and hypoxic areas (resp. grey and red) across 6 tumors, analysing 3 field of view with over 150 cells per field of view. g, CD45-positive cells in PyMT tumors: representative image of staining for DNA (propidium iodide, blue), 5hmC (red), pimonidazole (green) and CD45 (white); scale: 100 µm. The bar chart inset illustrates the quantification of CD45-positive cells in normoxic and hypoxic areas (resp. white and red) of 11 tumors, capturing on average ~2,500 nuclei per analysis. *** P<0.001 in (a) by Fisher’s exact test, significance relative to all genes. Extended data figure 9 Manipulation of tumor oxygenation in murine breast tumors, and effects on 5hmC, TSG HM and confounders. a, Plasma sFlk1 concentrations at the indicated times after hydrodynamic injection with an empty (n=7) or sFlk1-overexpression plasmid (n=5) (resp. grey and red). b-cQuantification of tumor vessel number (b) and hypoxic areas (c) of tumors from tg(MMTV-PyMT) mice, hydrodynamically injected with an empty or sFlk1-overexpression plasmid, with representative images of blood vessels stained for CD31 (b) and hypoxic areas stained for pimonidazole adducts (c); scale: 100µm. d Changes in RNA expression of hypoxia marker genes that are known to be downregulated (Mrc1) or upregulated (Bnip3, Car9, Ddit4) in hypoxic conditions. e, 5hmC levels (y axis) across mouse chromosome 18 (x axis) in 400kb bins, with the location of RefSeq genes (middle), and differences in 5hmC levels (lower). 5hmC levels were determined using shallow TAB-seq, and chromosome 18 was selected because it has large stretches of gene deserts that illustrate the 5hmC depletion in these areas (n=3). 5hmC levels decrease by 12.4±3.5% after sFlk1 overexpression, although technical limitations of TAB-seq (incomplete 5hmC protection or bisulfite conversion) may partially obscure the magnitude of effects. f, HM in tumors developing in 12-week-old mice receiving hydrodynamic injection with an empty (n=19) or sFlk1-overexpressing plasmid (n=24) 3 weeks earlier. DNA was bisulfite converted, PCR-amplified for the indicated oncogenes, and sequenced to a depth of ~500x. Plotted are z-scores of HM (y axis, exponential), relative to the more normoxic tumors (i.e. empty). The dotted line represents the threshold at 5% FDR, and bold darker dots the tumors displaying significantly increased HM events. g-j,(g) Relative weights of tumors from tg(MMTV-PyMT) mice, hydrodynamically injected with an empty (grey, n=19) or sFlk1-overexpression plasmid (red, n=24), and corresponding RNA expression of Ptprc (the gene encoding CD45, n=5) (h), of Tet enzymes (i, n=15 for empty plasmid, n=12 for sFlk1-overexpressing plasmid) and of cell proliferation markers (j, n=5 for each). k-m, As in (d-f), but for 16-week old tg(MMTV-PyMT) mice of the indicated genotype. n=5 (k), n=3 for Phd2+/+; n=4 for Phd2+/- (l) and n=9 (m) n, as in (d), but for 16-week old Tie2-Cre;tg(MMTV-PyMT) mice of the indicated genotypes (n=5). o, DNA was extracted from 17 breast tumors developing in Tie2-Cre;Phd2fl/wt;tg(MMTV-PyMT) mice (blue) and 13 breast tumors developing in Tie2-Cre;Phd2wt/wt;tg(MMTV-PyMT) mice (grey), all 16 weeks old. DNA was bisulfite converted, PCR-amplified for the indicated TSGs (left) and oncogenes (middle) and sequenced to a depth of ~500x. Plotted are z-scores of HM (y axis, exponential), relative to the more normoxic, Phd2wt/fl, tumors. The dotted line represents the threshold for a Bonferroni-adjusted P<0.05, and bold darker dots the tumors displaying significantly increased HM events. (right) 5hmC levels ± s.e.m. across a metagene in tumors of 16-week-old mice with the indicated genotype (n=3 for Phd2fl/fl; n=4 for Phd2wt/fl). p-u, Relative weights of tumors from Phd2+/-;tg(MMTV-PyMT) mice and Phd2+/+;tg(MMTV-PyMT) mice (n=10 and 9 resp.) (p-r) and from Tie2-Cre;Phd2fl/wt;tg(MMTV-PyMT) and Tie2-Cre;Phd2wt/wt;tg(MMTV-PyMT) mice (n=17 and 13 resp.) (s-u), and the corresponding RNA expression of cell proliferation markers (n=5) (p, s), of Tet enzymes (n=5) (q, t) and of Ptprc (n=5), the gene encoding CD45 (r, u). # P<0.10, * P<0.05, ** P<0.01, *** P<0.001 by t-test. Supplementary Material Supplementary Information is linked to the online version of the paper at www.nature.com/nature. supp_tables Acknowledgements We thank Gilian Peuteman, Thomas Van Brussel, Jens Serneels and Kerstin Kurz for assistance, Christopher Chang for NucPE1, Guo-Liang Xu for Tet-TKO ESCs. HZ and BT hold FWO-F postdoctoral fellowships. This work was supported by a ERC consolidator grant (CHAMELEON- 617595) to D.L. Figure 1 Effect of hypoxia on 5hmC in vitro. a, Levels of 5hmC (upper), and overall TET expression (lower) in cell lines grown for 24 h under 21% or 0.5% O2. RNA expression is expressed relative to the combined estimated level of all 3 TET paralogues under 21% O2., b-c, 5hmC/C levels in MCF7 cells exposed to different O2 levels for 24 h (b), or 0.5% oxygen for indicated times (c). d, Correlation of changes in overall TET expression and 5hmC upon hypoxia. Each circle represents a cell line, the full line the correlation. e-f, Levels of 5hmC (e, f) and α-ketoglutarate (f) in MCF7 cells grown with ascorbate (e), water or dimethyl-α-ketoglutarate (f) under 21% or 0.5% O2 (white or red). α-ketoglutarate changes are relative to matching water controls. g, As (a), but for cells exposed to IOX2. h-i, Michaelis-Menten curve of Tet1 (h) and Tet2 (i, n=3) for O2. Error bars denote s.e.m., grey areas: 95% c.i., n = 5 replicates for panels (a-h), *P<0.05, **P<0.01, ***P<0.001 by t-test (b, c, e) or ANOVA with post-hoc Tukey HSD (f). Figure 2 Genomic profiles of 5(h)mC in MCF7 following hypoxia. a, Changes in 5hmC at 290,382 peaks detected using 5hmC-DIP-seq. Peaks gaining and loosing 5hmC (red and blue) are highlighted at P<0.05 and 5% FDR adjustment (lighter and darker). b, Observed/expected fraction of 5hmC peaks overlapping with chromHMM chromatin states and exhibiting hypoxia-associated 5hmC loss (n=10,001, blue) or not (n=280,381, grey). c-d, Changes in 5mC after 24 (c) or 48 (d) hours of 0.5% O2, assessed by 5mC-DIP-seq at 10,001 hypohydroxymethylated peaks upon hypoxia (c) or by BS-seq at 1,894 regions capture-selected using SeqCapEpi (d). e, Expression changes of genes in hypohydroxymethylated, and both hypohydroxymethylated and hypermethylated peaks. Plots depict 3 (a, e), 4 (c) or 5 (d) replicates, P-values by negative binomial generalized linear models (a, c), Fisher’s exact (d) or t-test (e, ***P<0.001). Figure 3 Impact of hypoxia on hypermethylation in TCGA. a, Observed/expected number of hypoxic versus normoxic tumors in 3 methylation clusters for 1,000 CpGs hypermethylated in tumor versus normal tissue. b, Percentage of HM events in promoters of frequently HM genes. n = 3,141 tumors, *P<0.05, **P<0.01, ***P<0.001 by Cochran-Armitage (a), generalized linear model per tumor type corrected for co-variates (Supplementary table 8) (b). Figure 4 Impact of hypoxia on TET activity in human tumors. a, HM in 19 normoxic (blue), 21 intermediate (grey), 17 hypoxic (red) and 6 IDH1R132-mutated (yellow) glioblastomas. b, Overlap between genes hypermethylated in hypoxic versus IDH1R132-mutated glioblastomas. c-d, (c) 5hmC measured across 485,000 CpGs in 12 normoxic versus 12 hypoxic non-small-cell lung tumors, and (d) changes in 5(h)mC for unaltered CpGs (grey), and CpGs altered in both 5mC and 5hmC (25% FDR, blue; P<0.01, red). ***P<0.001 by Fisher’s exact (a), **P<0.01 by t-test (c). Figure 5 Impact of vessel pruning and normalization on 5hmC and TSG HM. a-b, Immunofluorescence of breast tumors in tg(MMTV-PyMT) mice. (a) Representative image, scale: 50 μm. (b) Boxplot of 5hmC signal in >150 PyMT-positive nuclei from 8 tumors, stratified for pimonidazole (yes/no) and normalized to pimonidazole-negative nuclei. c, 5hmC levels ± s.e.m. across a metagene in tumors of 12-week-old mice receiving empty or sFlk1-overexpressing plasmid (left, n=3), or 16-week-old mice with the indicated genotype (right, n=3 for Phd2+/+; n=4 for Phd2+/-). d-e, HM in (d) tumors developing in 12-week-old mice receiving empty (n=19) or sFlk1-overexpressing plasmid (n=24) 3 weeks earlier, and in (e) tumors developing in 16-week-old Phd+/- (n=10) and Phd+/+(n=9) mice. Plotted are z-scores of HM, relative to normoxic tumors (empty and Phd2+/- for d and e). Dotted line: 5% FDR, darker dots: significant HM. Brca1 and Timp3: not shown (no HM event detected). Hypermethylated genes on average had 5.8% (d) and 4.7% (e) more methylation. *P<0.05, **P<0.01, ***P<0.001 by t-test. Contributions B.T. and D.L. conceived and supervised the project, designed experiments, wrote the manuscript. B.T. and F.D.A. performed in vitro experiments and analysed data, helped by L.V.D.; M.C. and A.P. analysed Tet Michaelis-Menten kinetics; animal models provided by E.H., F.A. (xenografts), M.M. (sFlk1), A.K. and P.C. (Phd2+/-); V.N.K. contributed ideas, L.S. and K.P.K. reagents; J.S. quantified nucleotides by LC/MS, supervised by T.C.; B.G. quantified metabolites. H.Z. analysed TCGA; B.T., H.Z. and B.B. performed bioinformatics and statistics. Author Information Microarray and sequencing data are available at GEO under accession GSE71403. Reprints and permissions information is available at www.nature.com/reprints. Readers are welcome to comment on the online version of the paper. The authors declare no competing financial interests. 1 Esteller M Epigenetics in cancer The New England journal of medicine 2008 358 1148 1159 10.1056/NEJMra072067 18337604 2 Struhl K Is DNA methylation of tumour suppressor genes epigenetic? eLife 2014 3 e02475 10.7554/eLife.02475 24623307 3 Weisenberger DJ CpG island methylator phenotype underlies sporadic microsatellite instability and is tightly associated with BRAF mutation in colorectal cancer Nature genetics 2006 38 787 793 10.1038/ng1834 16804544 4 Mack SC Epigenomic alterations define lethal CIMP-positive ependymomas of infancy Nature 2014 506 445 450 10.1038/nature13108 24553142 5 Tahiliani M Conversion of 5-methylcytosine to 5-hydroxymethylcytosine in mammalian DNA by MLL partner TET1 Science 2009 324 930 935 10.1126/science.1170116 19372391 6 Shen L Genome-wide analysis reveals TET- and TDG-dependent 5-methylcytosine oxidation dynamics Cell 2013 153 692 706 10.1016/j.cell.2013.04.002 23602152 7 Xiao M Inhibition of alpha-KG-dependent histone and DNA demethylases by fumarate and succinate that are accumulated in mutations of FH and SDH tumor suppressors Genes & development 2012 26 1326 1338 10.1101/gad.191056.112 22677546 8 Figueroa ME Leukemic IDH1 and IDH2 mutations result in a hypermethylation phenotype, disrupt TET2 function, and impair hematopoietic differentiation Cancer Cell 2010 18 553 567 10.1016/j.ccr.2010.11.015 21130701 9 Xu W Oncometabolite 2-hydroxyglutarate is a competitive inhibitor of α-ketoglutarate-dependent dioxygenases Cancer cell 2011 19 17 30 21251613 10 Yang H Tumor development is associated with decrease of TET gene expression and 5-methylcytosine hydroxylation Oncogene 2013 32 663 669 10.1038/onc.2012.67 22391558 11 Ploumakis A Coleman ML OH, the Places You'll Go! Hydroxylation, Gene Expression, and Cancer Molecular cell 2015 58 729 741 10.1016/j.molcel.2015.05.026 26046647 12 Schofield CJ Ratcliffe PJ Oxygen sensing by HIF hydroxylases Nature Reviews Molecular Cell Biology 2004 5 343 354 15122348 13 Hanahan D Folkman J Patterns and emerging mechanisms of the angiogenic switch during tumorigenesis Cell 1996 86 353 364 8756718 14 Vaupel P Hockel M Mayer A Detection and characterization of tumor hypoxia using pO2 histography Antioxid Redox Signal 2007 9 1221 1235 10.1089/ars.2007.1628 17536958 15 Williams K TET1 and hydroxymethylcytosine in transcription and DNA methylation fidelity Nature 2011 473 343 348 10.1038/nature10066 21490601 16 Ito S Tet proteins can convert 5-methylcytosine to 5-formylcytosine and 5-carboxylcytosine Science 2011 333 1300 1303 10.1126/science.1210597 21778364 17 Mariani CJ TET1-mediated hydroxymethylation facilitates hypoxic gene induction in neuroblastoma Cell reports 2014 7 1343 1352 10.1016/j.celrep.2014.04.040 24835990 18 Zhao B Redox-active quinones induces genome-wide DNA methylation changes by an iron-mediated and Tet-dependent mechanism Nucleic Acids Res 2013 gkt1090 19 Blaschke K Vitamin C induces Tet-dependent DNA demethylation and a blastocyst-like state in ES cells Nature 2013 500 222 226 10.1038/nature12362 23812591 20 Koivunen P Transformation by the (R)-enantiomer of 2-hydroxyglutarate linked to EGLN activation Nature 2012 483 484 488 22343896 21 Bachman M 5-Hydroxymethylcytosine is a predominantly stable DNA modification Nat Chem 2014 6 1049 1055 10.1038/nchem.2064 25411882 22 Chowdhury R Selective small molecule probes for the Hypoxia Inducible Factor (HIF) prolyl hydroxylases ACS chemical biology 2013 8 1488 149 23683440 23 Taberlay PC Statham AL Kelly TK Clark SJ Jones PA Reconfiguration of nucleosome-depleted regions at distal regulatory elements accompanies DNA methylation of enhancers and insulators in cancer Genome Res 2014 24 1421 1432 10.1101/gr.163485.113 24916973 24 Cancer Genome Atlas Research, N The Cancer Genome Atlas Pan-Cancer analysis project Nature genetics 2013 45 1113 1120 10.1038/ng.2764 24071849 25 Buffa FM Harris AL West CM Miller CJ Large meta-analysis of multiple cancers reveals a common, compact and highly prognostic hypoxia metagene Br J Cancer 2010 102 428 435 20087356 26 Feinberg AP Irizarry RA Stochastic epigenetic variation as a driving force of development, evolutionary adaptation, and disease Proceedings of the National Academy of Sciences 2010 107 1757 1764 27 Kuchnio A The Cancer Cell Oxygen Sensor PHD2 Promotes Metastasis via Activation of Cancer-Associated Fibroblasts Cell reports 2015 12 992 1005 10.1016/j.celrep.2015.07.010 26235614 28 Vogelstein B Cancer Genome Landscapes Science 2013 339 1546 1558 10.1126/science.1235122 23539594 29 Landan G Epigenetic polymorphism and the stochastic formation of differentially methylated regions in normal and cancerous tissues Nature genetics 2012 44 1207 1214 10.1038/ng.2442 23064413 30 Laukka T Fumarate and Succinate Regulate Expression of Hypoxia-inducible Genes via TET Enzymes J Biol Chem 2016 291 4256 4265 10.1074/jbc.M115.688762 26703470 31 Paez-Ribes M Antiangiogenic therapy elicits malignant progression of tumors to increased local invasion and distant metastasis Cancer Cell 2009 15 220 231 10.1016/j.ccr.2009.01.027 S1535-6108(09)00034-8 [pii] 19249680 32 Heist RS Improved tumor vascularization after anti-VEGF therapy with carboplatin and nab-paclitaxel associates with survival in lung cancer Proc Natl Acad Sci U S A 2015 112 1547 1552 10.1073/pnas.1424024112 25605928 33 Sermeus A Hypoxia induces protection against etoposide-induced apoptosis: molecular profiling of changes in gene expression and transcription factor activity Molecular cancer 2008 7 27 10.1186/1476-4598-7-27 18366759 34 Schmittgen TD Livak KJ Analyzing real-time PCR data by the comparative CT method Nature protocols 2008 3 1101 1108 18546601 35 Majmundar AJ Wong WJ Simon MC Hypoxia-inducible factors and the response to hypoxic stress Molecular cell 2010 40 294 309 10.1016/j.molcel.2010.09.022 20965423 36 Feng J Liu T Zhang Y Using MACS to identify peaks from ChIP-Seq data Current protocols in bioinformatics 2011 editoral board, Andreas D. Baxevanis … [et al.] Chapter 2, Unit 2 14 10.1002/0471250953.bi0214s34 37 Durand RE Raleigh JA Identification of nonproliferating but viable hypoxic tumor cells in vivo Cancer Res 1998 58 3547 3550 9721858 38 Lippert AR Van de Bittner GC Chang CJ Boronate oxidation as a bioorthogonal reaction approach for studying the chemistry of hydrogen peroxide in living systems Acc Chem Res 2011 44 793 804 10.1021/ar200126t 21834525 39 Dickinson BC Tang Y Chang Z Chang CJ A nuclear-localized fluorescent hydrogen peroxide probe for monitoring sirtuin-mediated oxidative stress responses in vivo Chem Biol 2011 18 943 948 10.1016/j.chembiol.2011.07.005 21867909 40 Vichai V Kirtikara K Sulforhodamine B colorimetric assay for cytotoxicity screening Nat Protoc 2006 1 1112 1116 10.1038/nprot.2006.179 17406391 41 He YF Tet-mediated formation of 5-carboxylcytosine and its excision by TDG in mammalian DNA Science 2011 333 1303 1307 10.1126/science.1210944 21817016 42 Taiwo O Methylome analysis using MeDIP-seq with low DNA concentrations Nat Protoc 2012 7 617 636 10.1038/nprot.2012.012 22402632 43 McCarthy DJ Chen Y Smyth GK Differential expression analysis of multifactor RNA-Seq experiments with respect to biological variation Nucleic Acids Res 2012 gks042 44 Parker JS Supervised risk predictor of breast cancer based on intrinsic subtypes Journal of clinical oncology 2009 27 1160 1167 19204204 45 Murtagh F Legendre P Ward’s hierarchical agglomerative clustering method: which algorithms implement Ward’s criterion? Journal of Classification 2014 31 274 295 46 Nielsen TO A comparison of PAM50 intrinsic subtyping with immunohistochemistry and clinical prognostic factors in tamoxifen-treated estrogen receptor-positive breast cancer Clin Cancer Res 2010 16 5222 5232 10.1158/1078-0432.CCR-10-1282 20837693 47 Du P Comparison of Beta-value and M-value methods for quantifying methylation levels by microarray analysis BMC bioinformatics 2010 11 587 10.1186/1471-2105-11-587 21118553 48 Wauters E DNA methylation profiling of non-small cell lung cancer reveals a COPD-driven immune-related signature Thorax 2015 70 1113 1122 10.1136/thoraxjnl-2015-207288 26349763 49 Yu M Tet-assisted bisulfite sequencing of 5-hydroxymethylcytosine Nat Protoc 2012 7 2159 2170 10.1038/nprot.2012.137 23196972 50 Nazor KL Application of a low cost array-based technique - TAB-Array - for quantifying and mapping both 5mC and 5hmC at single base resolution in human pluripotent stem cells Genomics 2014 104 358 367 10.1016/j.ygeno.2014.08.014 25179373 51 Morris TJ ChAMP: 450k Chip Analysis Methylation Pipeline Bioinformatics 2014 30 428 430 10.1093/bioinformatics/btt684 24336642 52 Liu F Song Y Liu D Hydrodynamics-based transfection in animals by systemic administration of plasmid DNA Gene therapy 1999 6 1258 1266 10.1038/sj.gt.3300947 10455434 53 Pelicano H Carney D Huang P ROS stress in cancer cells and therapeutic implications Drug Resistance Updates 2004 7 97 110 15158766 54 Booth MJ Quantitative sequencing of 5-methylcytosine and 5-hydroxymethylcytosine at single-base resolution Science 2012 336 934 937 10.1126/science.1220671 22539555

PMC005xxxxxx/PMC5133460.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 101125414 26807 Expert Opin Biol Ther Expert Opin Biol Ther Expert opinion on biological therapy 1471-2598 1744-7682 27434205 5133460 10.1080/14712598.2016.1214266 NIHMS828463 Article Cellular immunotherapy for malignant gliomas Lin Yi MD Visiting Scholar in Neurological Surgery at University of California San Francisco, Helen Diller Family Cancer Research Building, 1450 3rd Street, San Francisco, CA 94158 Okada Hideho MD, PhD. Kathleen M. Plant Distinguished Professor in Neurological Surgery at University of California San Francisco, Helen Diller Family Cancer Research Building, 1450 3rd Street, San Francisco, CA 94158; Phone: (415)476-1637; UCSF campus mail box number 0520 hideho.okada@ucsf.edu 8 11 2016 29 7 2016 10 2016 01 10 2017 16 10 12651275 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. Introduction Cancer immunotherapy has made much progress in recent years. Clinical trials evaluating a variety of immunotherapeutic approaches are underway in patients with malignant gliomas. Thanks to recent advancements in cell engineering technologies, infusion of ex vivo prepared immune cells have emerged as promising strategies of cancer immunotherapy. Areas covered Herein, the authors review recent and current studies using cellular immunotherapies for malignant gliomas. Specifically, they cover the following areas: a) cellular vaccine approaches using tumor cell-based or dendritic cell (DC)-based vaccines, and b) adoptive cell transfer (ACT) approaches, including lymphokine-activated killer (LAK) cells, γδ T cells, tumor-infiltrating lymphocytes (TIL), chimeric antigen receptor (CAR)-T cells and T-cell receptor (TCR) transduced T cells. Expert opinion While some of the recent studies have shown promising results, the ultimate success of cellular immunotherapy in brain tumor patients would require improvements in the following areas: 1) feasibility in producing cellular therapeutics; 2) identification and characterization of targetable antigens given the paucity and heterogeneity of tumor specific antigens; 3) the development of strategies to promote effector T-cell trafficking; 4) overcoming local and systemic immune suppression, and 5) proper interpretation of imaging data for brain tumor patients receiving immunotherapy. glioma cellular immunotherapy dendritic cell vaccine adoptive cell transfer CAR-T cell therapy 1. Introduction Malignant gliomas are the most common type of primary malignant brain tumor, with more than 18,000 new cases diagnosed each year in the United States1. Despite advancements of conventional therapies, including surgery, radiation therapy and chemotherapy, outcomes for these patients remain dismal. Glioblastoma (GBM) is the most common and the most malignant of the gliomas; patients with GBM have a median survival of approximately 15 months following treatment with a combination of chemotherapy (Temozolomide) with radiation therapy2, pointing to the urgent need to develop novel efficacious therapeutic modalities. Cancer immunotherapy is aimed at enhancing the systemic and selective immune response against tumor cells. Both innate and adaptive immune responses play complementary roles. In innate immunity, natural killer (NK) and myeloid cells recognize and destroy virally infected cells and a range of tumor cells in a major histocompatibility complex (MHC)-unrestricted manner. Adaptive immune responses are antigen-specific and initiated by presentation of tumor antigens by antigen presenting cells (APCs). Most potent APC are dendritic cells (DCs) which can develop from myeloid cells. DCs present tumor-derived epitope peptides as MHC-peptide complexes to T cells via TCR. Activated T cells clonally expand and then traffic to the tumor-involved organs. T cells recognize antigen epitopes via MHC/peptide complex on tumor cells through the TCR, leading to T cell activation and release of preformed cytotoxic molecules (granzyme and perforin). The field of cancer immunotherapy has made exciting breakthroughs in recent years. The United States Food and Drug Administration (FDA) approved monoclonal antibodies (mAb) to the inhibitory immune checkpoint molecules cytotoxic T lymphocyte-4 (CTLA-4; ipilimumab) and programmed death 1 (PD-1; pembroluzimab and nivolumab) for metastatic melanoma as well as non-small cell lung cancer (NSCLC)3–6. In regard to cellular immunotherapy, the FDA approved the first vaccine against non-viral cancers (sipuleucel-T)7. Furthermore, chimeric antigen receptor (CAR) engineered autologous T cells have induced durable remissions among leukemia patients refractory to conventional therapies including bone marrow transplantation8, 9. Extensive preclinical and clinical studies are being conducted to extend these successes to other types of cancer, especially, solid cancer. Cellular immunotherapeutics could be fundamentally classified into vaccine approaches and adoptive transfer of effector cell approaches. In this review, we discuss recent and current efforts in the field of cellular immunotherapy for malignant gliomas. 2. Cellular vaccine approaches 2.1 Tumor cell vaccines Tumor cell vaccines utilize either autologous or allogeneic tumor cells that are attenuated. Since the inherent immunogenicity of the tumor cells may be limited, they are often genetically engineered to express costimulatory molecules, cytokines chemokines, or those in combination10. Especially, to enhance the presentation of tumor antigens, cytokines that activate antigen-presenting cells (APCs), such as GM-CSF and IL-411–13, co-stimulatory molecules, such as CD8014 have been used. In preclinical studies, peripheral immunization of rats bearing 9L gliosacroma in the brain with IL-4 transfected 9L cells achieved the most potent therapeutic benefit compared to GM-CSF, IL-12 and IFN-α15, 16. IL-4 produced at the local vaccine site appears to promote a T-helper 1-type antitumor immune response17, and the observed therapeutic response was further enhanced in cooperation with local delivery of IFN-α in the intracranial tumor site18. A phase I clinical study evaluated safety and immunological activity of a vaccine with autologous tumor cells admixed autologous fibroblasts that are engineered to express IL-4 in patients with recurrent malignant glioma19. While only 2 of 6 enrolled participants received scheduled two vaccinations, both participants demonstrated encouraging immunological and clinical and radiological responses19. No significant side effects were observed. However, generating sufficient numbers of IL-4-transfected vaccine cells required 7 to 8 weeks. Most participants were withdrawn from the trial because of tumor progression prior to the first vaccination, which posed a major feasibility issue. A phase I/IIa study evaluating autologous formalin-fixed tumor vaccine in newly diagnosed GBM demonstrated feasibility, tolerability as well as encouraging median overall survival (OS) of 22.2 months and a median progression-free survival (PFS) of 8.2 months20. More recently, a phase I study was conducted demonstrating the safety and feasibility of vaccination with irradiated autologous glioma cells mixed with irradiated GM-CSF-transduced allogeneic K562 cells in patients with recurrent malignant glioma 21. 2.2 Dendritic cell vaccines Dendritic cells (DCs) are the most potent APCs, and establishment of methods to culture DCs from peripheral blood-derived monocytes facilitated developments of DC-based vaccines in a variety of cancer types22. DCs can be coupled with a variety of tumor antigen sources, such as, synthetic peptides, autologous glioma lysate or acid-eluted glioma peptides. DCs can be directly fused with tumor cells or transfected with tumor RNA, cDNA, or viral vectors23. With the use of whole autologous tumor as the antigen source, DCs can present a wide array of possible tumor antigens to the host immune system24. Because DCs present tumor-antigens on their MHC molecules, it seems to make the most logical sense if DCs are loaded with acid-eluted peptides derived from autologous tumor cell surface MHC class I molecules. Vaccinations using DCs loaded with autologous acid-eluted peptides were safe25, and elicited detectable systemic toxicity and intracranial T-cell infiltration26. This method, however, requires 108–109 tumor cells to derive peptides for loading sufficient numbers of DCs, posing a major feasibility challenge. Use of autologous, whole glioma cell lysate can alleviate this concern as peptides from proteins in the lysate can still be presented by MHC. Wheeler et al. reported the immune and clinical responses from a phase II trial treating 44 patients (34 GBM) with autologous DC pulsed with tumor lysate. Fifty-three percent of GBM patients exhibited ≥1.5 fold vaccine-enhanced cytokine responses. Vaccine responders exhibited significantly longer survival relative to nonresponders, with 41% of vaccine responders survived at least 2 years compared with 7% of vaccine nonresponders27. In another trial, IL-4-transfected fibroblasts admixed with DCs loaded with tumor lysate were given intradermally in five newly diagnosed GBM patients19. The median time to progression (TTP) after surgical resection was 6 months. A more recent phase I/II study evaluating a tumor lysate-loaded DC-based vaccine in 77 patients with newly diagnosed GBM showed the feasibility of integrating this treatment in the standard-of-care treatment with surgery, radiotherapy, and chemotherapy. Median OS was 18.3 months since leukapheresis28. DCVax is an autologous DC vaccine pulsed with tumor lysate antigen for the treatment of GBM. Autologous tumor lysate–pulsed DC vaccination in conjunction with TLR agonists was evaluated for safety in newly diagnosed and recurrent glioblastoma patients29. In addition to encouraging survival data, patients whose tumors had mesenchymal gene expression signatures exhibited increased survival compared with historic controls of the same genetic subtype. Tumor samples with a mesenchymal gene expression signature had a higher number of CD3+ and CD8+ TILs compared with GBMs of other gene expression signatures, suggesting that GBM with the mesenchymal gene expression profile may be more responsive to immune-based therapies. None of these whole-GBM antigen-loaded DC vaccines demonstrated autoimmune encephalitis. Use of synthetic peptides encoding tumor antigen epitopes provides “off-the-shelf” feasibility thanks to unlimited availability of synthetic peptides. Furthermore, compared with the whole tumor cell antigen-based approaches, use of synthetic peptides targeting tumor-specific (mostly tumor-specific mutation-derived) or tumor-associated (non-mutated but expressed at higher levels in tumor cells vs. normal cells) antigens may reduce the risk of autoimmunity, although selection of antigens is crucial. There are recent excellent reviews on antigens targeted in immunotherapy of gliomas30–33. Several phase I/II trials employing synthetic peptide antigens have been conducted. We have evaluated novel α-type 1 polarizing DCs (αDC1), which were manufactured by maturation of monocyte-derived immature DCs with IL-1β, tumor necrosis factor-α, interferon (IFN)-α, IFN-γ and poly-I:C. αDC1 were loaded with synthetic peptides for glioma-associated antigen (GAAs) epitopes and administered in combination with polyinosinic-polycytidylic acid [poly(I:C)] stabilized by lysine and carboxymethylcellulose (poly-ICLC) in HLA-A2(+) patients with recurrent malignant gliomas34. GAAs for these peptides are EphA2, interleukin (IL)-13 receptor-alpha2, YKL-40, and gp100. In 22 recurrent high-grade glioma patients who received at least one vaccine, nine patients remained free from progression for at least 12 months. One patient with a GBM had a complete response, and IL-12 production levels by αDC1 positively correlated with time to progression34. In another trial, 19 GBM and one brainstem glioma patient received DCs pulsed with tumor-associated antigens (TAA: HER2, TRP-2, gp100, MAGE-1, IL-13Rα2 and AIM-2)35. The median OS was 38.4 months for patients with newly diagnosed GBM. OS was positively correlated with quantitative expression of MAGE-1, AIM-2, gp100 and HER2 in patient tumor samples. Some of recent DC vaccine studies evaluated combination with immunoadjuvants. These include adjuvant cytokine administration (GM-CSF)36, and toll-like receptor (TLR) agonists29, 34. Chemotherapy or antiangiogenic therapy may also potentiate of DC-based immunotherapy37, 38. However, a very carefully designed study has demonstrated that clinically relevant dosages of standard alkylating chemotherapies, such as temozolomide and cyclophosphamide, profoundly inhibit B and T cell responses to vaccines 39, calling for our cautions designing vaccine studies with concurrent chemotherapy. Viral antigens may represent particularly attractive targets for immunotherapy because they are foreign to the host immune system and thus are inherently immunogenic. Malignant brain tumors have not been shown to be virally induced, but studies have demonstrated frequent detection of low-level expression of human cytomegalovirus (CMV) genes within malignant gliomas40, 41. While the role of CMV in the biology of these tumors is a continued area of study 40, a recent study of a DC vaccine targeting CMV epitopes in GBM demonstrated promising results, especially in combination with the vaccine-site conditioning with tetanus-toxioid42. A number of clinical trials are conducted in malignant gliomas (Table 1). DC-based vaccines for brain tumors appear to be safe and can induce anti-tumor immune response. However, objective clinical benefits (objective anti-tumor response and/or extension of survival) remains to be determined. 3. Adoptive cell therapy (ACT) of effector cells In ACT of effector cells, large numbers (typically 1 × 106–9 orders) of immune effector cells are prepared ex vivo and infused to patients. In brain tumor patients, these cells have been administered locally in the brain tumor site or systemically via i.v. In the past, ex vivo prepared cells with undefined, broad antigen-specificity were mainly used, such as lymphokine-activated killer (LAK) cells. Recently, antigen-targeted approaches have been developed, such as the use of CAR and TCR-transduced cells (Table 2). Even though some of these approaches are quite successful in other cancer types, it is important to address unique challenges that arise when these approaches are applied for brain tumors. 3.1 LAK cells and NK cells LAK cells are autologous peripheral blood lymphocytes stimulated with IL-2 in vitro 43. Natural killer (NK) cells are the major effector population in LAK cells. They recognize cancer cells in a non-MHC-restricted fashion. LAK cells may represent a primitive immune surveillance system capable of recognizing and destroying altered cells. A number of clinical studies have been conducted treating GBM or high-grade glioma patients with local injection of LAK cells44. These studies demonstrated the safety of infusing autologous leukocytes into the tumor resection cavity. Some of them also have shown promising results in prolonging disease free survival. However, comparison of LAK cell therapy and IL-2 with IL-2 alone showed no significant difference in response rates in patients with renal cell carcinoma45. Also high-dose IL-2 may lead to capillary leak syndrome, including hypotension, oliguria, pulmonary edema and dyspnea, discouraging further study of the approach. Thus, a randomized phase II or III clinical study was never conducted. However, owing to recent advances in the field of NK cell biology46, there is a renewed interest in NK cell-based immunotherapy for cancer47. Recent preclinical advancements of NK cell immunotherapy include augmentation of antibody-dependent cellular cytotoxicity, manipulation of receptor-mediated activation, and adoptive immunotherapy with ex vivo-expanded, CAR-engineered NK cells (reviewed in 48). 3.2 γδ T cells In healthy donors, T cells bearing the γδ T cell receptor constitute 0.5–20% of CD3+ T lymphocytes in peripheral blood and in lymphoid tissues49. They can be isolated then expanded by IFN-γ, IL-2, monoclonal antibody against CD3, and IL-1α50. γδ T cells express natural cytotoxicity receptor natural killer p 44, and exert their cytolytic activity mainly via the non-MHC-restricted γδ TCR51. Activating NK cell receptors such as NKG2D and DNAM1 are also present on most γδ T cells, which recognize stress-induced ligands on tumor cells50, 52. Early phase clinical trials have been conducted in non-Hodgkin lymphoma, multiple myeloma, and metastatic solid tumors53. In brain tumors, preclinical studies also suggested that γδ T-cell depletion and impaired function occur prior to or concurrent with the growth of the brain tumor54. Expanded/activated γδ T-cells from both healthy controls and selected patients have significant cytotoxicity against primary GBM explants55. Also there are evidence that γδ T cell therapy may be safe for brain tumor patients who undergo standard cytotoxic therapies56, 57, opening a previously unexplored approach to cellular immunotherapy of brain tumors. 3.3 TIL transfer TILs are obtained from tumor tissue, draining lymph nodes or malignant effusions. They contain high numbers of tumor-specific T cells that presumably have already been selected for their ability to recognize and respond to the tumor antigens. While TILs may not possess sufficient antitumor activity in the highly immunosuppressive microenvironment established by tumors, activation and expansion of TILs ex vivo can overcome these immunosuppressive effects and allow for the generation of sufficient numbers of TILs for adoptive immunotherapy. These TILs are expanded ex vivo with high dose IL-2, then transferred back to the patient. Adoptive cell therapy with TILs in combination with lymphodepletion and high-dose IL-2 has mediated durable, complete regressions in patients with melanoma, with reproducible objective response rates of approximately 50% in patients with highly advanced, refractory metastatic melanoma, probably by targeting somatic mutations exclusive to each cancer58. However, in brain tumors only few attempts have been made59–61. This may be because obtaining and expanding enough numbers of TILs require highly immunogenic, large, and accessible tumors. For malignancies other than melanoma, it has been very difficult to expand TILs from tumor tissues62. Also T cells present at the tumor bed are often exhausted, limiting their functions and their proliferative capacity. To overcome this issue for gliomas, a clinical trial was performed first vaccinating patients with irradiated autologous tumor cells, then harvesting tumor-draining lymph node T cells, expanding them ex vivo with anti-CD3 antibody and bacterial superantigen Staphylococcal enterotoxin A, and systemically infusing these cells63, 64. Three out of ten patients with recurrent malignant gliomas63 and four out of ten patients with newly diagnosed malignant gliomas64 showed radiographic partial response. However, no study has proven prolongation of the survival of glioma patients. 3.4 Adoptive transfer of genetically engineered T-cells (CAR and TCR) 3.4.1 αβT-Cell Receptors The cDNAs for the α- and β-chains of the TCR are cloned from class I HLA-restricted TCRs of tumor-reactive cytotoxic T cells and transferred to fresh T cells. Several TCRs have been cloned for several HLA-restricted epitopes encoded by TAAs65–68. Genetic modification of T cells with α/β TCRs also requires high expression and correct pairing of two different receptor molecules from a single vector, which has proved problematic for transgenic α/β TCRs, especially because mispairing between transgene- or endogenous TCR-derived α and β chain can occur. A variety of gene-engineering technologies have been evaluated, such as small interfering RNA constructs that specifically down-regulate endogenous TCR;69 a disulfide bridge in the α/β constant (C) regions by the extra cysteine residues; substituting human with murine C regions; codon optimization to enhance protein synthesis; TCR chain leucine zipper fusions; and a single chain TCR (reviewed 70, 71). In the first reported trial to examine the in vivo efficacy of TCR-transduced T cells in patients with cancer, the adoptive transfer of autologous T cells that were transduced with a MART-1–reactive TCR lead to tumor regression in 2 of 15 treated patients with metastatic melanoma65. Another study using autologous T-cells transduced with TCR treated 36 patients with metastatic melanoma using high-avidity TCRs that recognized either the MART-1 or gp100 melanoma-melanocyte antigens67. Objective cancer regressions were observed in 30% and 19% of patients who received the MART-1 or gp100 TCR, respectively, but severe off-tumor, on-target toxicity was seen in the skin, eyes, and ears due to the presence of melanocytes in these organs. The use of a high-affinity TCR against the carcinoembryonic antigen (CEA) in patients with metastatic colorectal cancer that expressed high levels of this antigen72 was halted when all 3 patients experienced life-threatening colitis and colonic hemorrhage. Unexpected toxicities can also present when previously unknown cross-reactive targets are expressed in healthy vital organs. For example, while MAGE-A3 is not known to be expressed in any normal tissues, targeting an HLA-A2.1-restricted peptide in MAGE-A3 caused severe damage to brain gray matter, resulting in 2 deaths because this TCR homed to a different but related epitope expressed by MAGE-A12 at very low levels in the brain73. A TCR directed against NY-ESO-1, a cancer germline antigen expressed in a variety of solid cancers holds promise. Objective responses were observed in 11 of the 18 patients (61%) with synovial cell sarcoma and 11 of the 20 patients (55%) with melanoma who received autologous TCR-transduced cells. The total number of T cells and the number of antigen-reactive T cells administered to patients correlated with response to therapy. However, there was a lack of a correlation between clinical responsiveness and persistence of infused T cells, possibly resulted from a failure of the T cells to persist for longer time periods75. NY-ESO-1-specific TCR-engineered T cells also showed encouraging clinical response in 16 of 20 patients with myeloma, in which engineered T cells expanded, persisted, trafficked to bone marrow and exhibited a cytotoxic phenotype. Disease progression was associated with loss of T cell persistence or antigen escape76. For brain tumors, however, no clinical study with αβ TCR T cells has been initiated. 3.4.2 CAR-T cells CAR engineering involves transgene-expression of single chain variable fragment (scFv) of a monoclonal antibody (Ab), which is specific for a tumor cell surface protein, at the surface of T cells, allowing the T cells to recognize tumor directly and not through the MHC complex. The Ab is linked to the CD3ζ chain and other T cell activation pathways, allowing T cell activation and target cell killing77, 78. In addition, Abs bind antigens with much greater affinity than do TCRs, resulting in the formation of a more stable immunological synapse79. CARs have evolved over the last decade, with progressively increasing co-stimulatory activity. In addition to a single signaling unit derived from the CD3ζ chain or the high-affinity IgG receptor Fc1RIg80, second-generation CARs incorporate the intracellular domain of a co-stimulatory molecule, CD28. Subsequent incorporation of both CD28 and a tumor necrosis factor receptor family member CD137 (4-1BB), CD 27, CD134 (OX40), CD244, or ICOS has enhanced the ability of these receptors to stimulate cytokine secretion and T cell proliferation and persistence in preclinical studies81. Compared with TCR engineered T cells, CAR engineered T cells are applicable to all patients irrespective of their HLA alleles expressed, and circumvent tumor evasion through HLA down-regulation. CARs have been generated for the glioma cell surface antigens, including IL-13Rα 282, HER283, EphA284, and EGFRvIII85–87. It should also be noted that CARs can induce toxicity against self-antigens as well. Acute pulmonary toxicity resulting in death was observed after infusion of CAR-T cells specific for ERBB2, likely due to the recognition of low levels of this antigen on pulmonary epithelium74. These observations underscore the need for selecting tumor-specific antigens, such as tumor-specific mutation-derived antigens (i.e., neoantigens), for effective and safe ACT. Among these targets, only EGFRvIII is tumor-specific, while others are TAAs. In preclinical animal models, T cells expressing these EGFRvIII-specific CARs showed potent antitumor activity85, 86. Phase I clinical trials of CART cells that are engineered to target HER2 (NCT02442297) or EGFRvIII mutation (NCT01454596 and NCT02209376) in patients with GBM are ongoing. A phase I clinical study with T cells expressing IL13Rα2-specific CAR has demonstrated safety in patients with recurrent GBM88. Most individuals naturally have high-avidity T-cells against viral epitopes at high frequencies. Transduction of viral antigen-specific T-cells with CAR may allow re-stimulation of CAR-transduced T-cells via the endogenous viral antigen-specific TCR (e.g. CMV-specific TCR) and the corresponding epitope (bispecific T cells). A phase I clinical studies with CMV-specific “bispecific” T-cells transduced with anti-HER2 CAR has been initiated (NCT01109095). A study with T cells expressing HER2-specific CARs showed that these cells had potent antitumor activity against HER2-positive, CD133-positive glioma stem cells83. A similar bispecific T cell, anti-GD2 CAR EBV-specific T cells, is being conducted as a phase I study in patients with refractory/relapsed neuroblastoma (NCT00085930). EBV-specific T cells, expressing a GD2-ζ CAR, persisted significantly longer than control, non-viral specific GD2-ζ T cells. Infusion of GD2-specific T cells resulted in tumor necrosis or regression (including a complete remission) in four out of eight patients89. A number of ways have been explored to increase the specificity of CAR-T cells to achieve a more promising, safer targeting. CAR-T cells can be genetically modified to recognize two or more tumor- associated antigens, which can enhance discrimination between abnormal and healthy tissue. One can transfer two CARs90; split-signal CARs, which can limit full T cell activation to tumors expressing multiple antigens91, 92 ; tandem CARs (TanCARs), which contain ectodomains with two scFvs93, also limiting the risk of immune escape; or co-expression of inhibitory CARs (iCARs) directed against molecules in healthy organs together with their activating counterparts (reviewed with a schema in94). Furthermore, a novel approach has been developed to engineer the T cells with dual-receptor circuits, in which a synthetic Notch receptor for one antigen induces the expression of a CAR for a second antigen95. 4. Conclusion We discussed recent developments of cellular immunotherapy for malignant brain tumors. While some of novel cellular therapeutics, such as CAR therapy, have demonstrated remarkable successes in other cancer types, translation of those successes to brain tumors will not be achieved unless we gain in-depth understanding of the unique immunological environment of brain tumors and develop strategies that are adequate to overcome challenges associated with the environment. In the “Expert Opinion” section below, we will further discuss our perspective. 5. Expert Opinion After decades of efforts to revise the longstanding dogma that the brain and tumors arising therein are “immunologically privileged”, immunotherapy, including cellular immunotherapy, for brain tumors has been emerging as a promising approach. However, the ultimate success of cellular immunotherapy in brain tumor patients would require advancements in the following areas: 1) feasibility of timely production of cellular therapeutics; 2) paucity and heterogeneity of tumor specific antigens; 3) better strategies to promote antigen-presentation and effector T-cell trafficking; 4) local and systemic immune suppression, and 5) proper interpretation of imaging data for brain tumor patients receiving immunotherapy. 5.1 Feasibility of Producing Cellular Therapeutics Production of autologous cell products inevitably involves lengthy and intensive processes, such as leukapheresis, engineering, expansion, as well as quality-assurance tests and assays. There require substantial costs, infrastructure of the institution as well as invasive procedures for patients, such as leukapheresis. These are important issues to address, especially because we hope that cellular immunotherapy will become effective, standard-of-care therapy in future. Ongoing efforts are directed to development of efficient bioreactors and automated processing systems. Another way to solve the issue is to develop “off-the-shelf” allogeneic cell products that can be safely administered without being rejected by the host immune-system. While this requires multiple immuno-genetic engineering of cells, developments are underway in this direction. 5.2 Paucity of tumor-specific antigens and heterogeneity of antigen-expression Although the list of antigens that could be used for immunotherapy of brain tumors has expanded over the last decade31, there are not many truly brain tumor-specific antigens, except for those derived from EGFRvIII and mutant IDH196. Use of tumor-associated, but non-specific antigens (TAAs as referred in this manuscripts) can cause life-threatening and fatal events by on-target72 or off-target97 cross-reactivity of T-cells against normal cells. Furthermore, due to marked heterogeneity of genetics and protein expression in solid cancers, targeting a single antigen may result in the evolution of variants that lack the target antigen 98. These observations underscore the need for expanding the list of available tumor-specific antigens, such as mutation-derived antigens (i.e., neoantigens), for effective and safe immunotherapy. Extension of these approaches should foster broader availability of target antigens for immunotherapy of brain tumors. 5.3 Better Strategies to Promote Antigen-Presentation and Effector T-cell Trafficking While brain tumors are heavily infiltrated by myeloid cells, the vast majority of them are suppressive for effector T-cell functions but not effective APCs99. Efforts are being undertaken to modulate the function of these cells and promote their function as type-1 APCs. In regard to T-cell homing to brain tumors, although T-cells are able to traverse the blood-brain-barrier via chemokine axes and multistep adhesion processes, homing of effector CTL is weaker in brain tumors compared with cancer in other organs100, 101. To date, there have not been many immunotherapy regimens for brain tumors incorporating therapeutic agents that can facilitate T-cell homing to the brain tumor site. Our regimens using poly-ICLC have been among the first to address this issue and are expected to enhance T-cell homing to the glioma site 34, 100, 102. In other organ sites, Kershaw et al. have demonstrated that engineering the chemokine receptor CXCR2 into T cells enabled the T cells to efficiently migrate toward melanoma103. Transgenic co-expression of CCR4 improved the homing of CAR-CD30-modified T cells to CD30+ Hodgkin lymphoma that secreted CCL17 (the ligand for CCR4)104. Enhanced CCR2b expression from mesothelin-reactive CAR-T cells and CAR-GD2 T cells led to improved anti- tumor effects against malignant pleural mesothelioma and neuroblastoma105, 106. Some of these inventions may be applicable for brain tumors as well. 5.4 Local and Systemic Immune Suppression Brain tumors mediate a variety of immunosuppressive mechanisms to escape from immunological attacks. These include expression of check-point molecules and immunosuppressive cytokines as well as recruitment of regulatory T-cells and immunosuppressive myeloid cells. Furthermore, it is important to recognize that significant levels of systemic immunosuppression are likely caused by treatments for these patients. These include chemotherapy, such as temozolomide39 as standard-of-care, corticosteroids as well as radiation therapy. Grossman et al have suggested that lymphocyte counts alone are predictive of prognosis, with lower counts correlating with shorter survival in patients with GBM107. It is important to address how we can minimize the impact of treatment-induced immunosuppression by the time the patient receives immunotherapy, although for adoptive transfer of T-cells, lymphopenic conditions induced by prior treatments may serve as a proper “conditioning”, thereby promoting post-infusion expansion of T-cells. In CAR T cells, post-infusion in vivo activity is mainly supported through addition of costimulatory molecules in the CAR construct. Recently, TRUCK T cells, also called fourth-generation CARs, were developed involving two separate transgenes, with the CAR gene and a T cell activation responsive promoter linked to a cytokine108. Studies have shown that therapy with T cells engineered to express IL-12 could change the tumor microenvironment and enhance anti-tumor function109, 110. IL-12 secretion by engineered T cells expressing CARs resulted in the destruction of antigen negative cancer cells that may escape from T cell therapy111. Also antigen-specific T cells expressing dnTGF-βRII were resistant to the anti-proliferative effects of TGF-β and retained their effector function in vivo112. On the other hand, CAR-T cells also express PD1, and are susceptible to PD1/PDL1 interaction-mediated suppression113. It has been shown that blocking PD1 immunosuppression can boost CAR-T cell therapy, likely representing a fruitful area for future study114, 115. 5.5 Proper interpretation of imaging data for brain tumor patients receiving immunotherapy Early phase immunotherapy clinical trials in brain tumor patients have revealed unique challenges associated with assessment of radiological changes reflecting delayed responses or therapy-induced inflammation116. Neuroimaging often reveals temporary worsening of abnormal findings and even appearance of new lesions. Clinical benefit, including long-term survival and tumor regression, can still occur following initial apparent progression. A multinational and multidisciplinary panel of neuro-oncology immunotherapy experts recently described immunotherapy response assessment for neuro-oncology (iRANO) criteria 117 that are based on guidance for determination of tumor progression outlined by the immune-related response criteria (irRC)118 and the response assessment in neuro-oncology (RANO) working group 119. The iRANO guidelines specifically address interpretation of initial progressive imaging findings in the context of neuro-oncology patients with a goal of decreasing the likelihood of premature discontinuation of potentially beneficial therapies while ensuring maximum patient safety. Prospective evaluation of the iRANO criteria in brain tumor immunotherapy trials for neuro-oncology patients will be required to improve their ultimate clinical utility. To address above discussed issues, it is apparent that our ultimate success will largely hinge upon effective collaboration across multiple disciplines. Scientifically, we have to integrate cutting edge progresses in both cancer immunology and central nervous system immunology. To implement novel combination strategies, it is essential to promote effective collaboration across companies and regulatory authorities. Encouraged by recent success in cancer immunotherapy for other cancer types, we believe that we are on the right direction and hope that we will develop truly effective immunotherapies for patients with malignant brain tumors. Funding: This work is supported by NIH grant R21NS083171 through H Okada. Table 1 Open Studies of cellular vaccine therapy in patients with primary brain tumors (as of April 5, 2016 in clinicaltrials.gov) NCT# Study Name Responsible Party type of cells used Disease phase NCT01792505 Dendritic Cell Vaccine With Imiquimod for Patients With Malignant Glioma Cedars-Sinai Medical Center Tumor lysate loaded DCs Malignant Glioma I NCT02010606 Phase I Study of a Dendritic Cell Vaccine for Patients With Either Newly Diagnosed or Recurrent Glioblastoma Cedars-Sinai Medical Center DCs Pulsed With Lysate Derived From an Allogeneic Glioblastoma Stem-like Cell Line Glioblastoma I NCT00639639 Vaccine Therapy in Treating Patients With Newly Diagnosed Glioblastoma Multiforme (ATTAC) Duke University Medical Center Human CMV pp65-LAMP mRNA-pulsed DCs Newly diagnosed glioblastoma I NCT00890032 Vaccine Therapy in Treating Patients Undergoing Surgery for Recurrent Glioblastoma Multiforme Duke University Medical Center BTSC (brain tumor stem cells) mRNA-loaded DC vaccine Recurrent Glioblastoma I NCT02529072 Nivolumab With dendritic cells Vaccines for Recurrent Brain Tumors (AVERT) Duke University Medical Center Human CMV pp65-LAMP mRNA-pulsed DCs, in combination with anti-PD-1 Monoclonal Antibody (Nivolumab) Recurrent high grade glioma I NCT02709616 Personalized Cellular Vaccine for Glioblastoma (PERCELLVAC) Guangdong 999 Brain Hospital Tumor antigen mRNA-pulsed DCs Glioblastoma I, II NCT01567202 Study of dendritic cells Vaccination Against Glioblastoma Huashan Hospital DCs Loaded With autogeneic glioma stem-like cells (A2B5+) Glioblastoma II NCT01204684 Dendritic Cell Vaccine for Patients With Brain Tumors Jonson Comprehensive Cancer Center Tumor lysate loaded DCs high grade glioma II NCT01635283 Vaccine for Patients With Newly Diagnosed or Recurrent Low-Grade Glioma Jonson Comprehensive Cancer Center Tumor lysate loaded DCs low grade glioma II NCT01759810 Proteome-based Personalized Immunotherapy of Glioblastoma NeuroVita Clinic Hematopoietic stem cells (HSCs), dendritic vaccine (DV) and cytotoxic lymphocytes (CTLs) Glioblastoma II, III NCT00045968 [DCVax®-L] to Treat Newly Diagnosed GBM Brain Cancer (GBM) Northwest Biotherapeutics Tumor lysate loaded DCs Glioblastoma III NCT02649582 Adjuvant Dendritic Cell-immunotherapy Plus Temozolomide in Glioblastoma Patients (ADDIT-GLIO) University Hospital, Antwerp Autologous Wilms’ tumor 1 (WT1) messenger (m)RNA-loaded DCs Newly diagnosed glioblastoma I, II NCT01326104 Vaccine Immunotherapy for Recurrent Medulloblastoma and Primitive Neuroectodermal Tumor (Re-MATCH) University of Florida Total tumor RNA (TTRNA)-loaded DCs (dendritic cells), ex vivo expanded Autologous Lymphocyte Recurrent Medulloblastoma and Primitive Neuroectodermal Tumor I, II NCT02465268 Vaccine Therapy for the Treatment of Newly Diagnosed Glioblastoma Multiforme (ATTAC-II) University of Florida Human CMV pp65-LAMP mRNA-pulsed DCs Newly diagnosed glioblastoma II NCT01902771 Dendritic Cell Vaccine Therapy With In Situ Maturation in Pediatric Brain Tumors University of Miami Tumor lysate loaded DCs Pediatric Brain Tumors I Table 2 Open Studies of adoptive cell transfer therapy in patients with primary brain tumors (as of April 5, 2016 in clinicaltrials.gov) NCT# Study Name Responsible Party type of cells used Disease phase NCT02100891 Phase 2 STIR Trial: Haploidentical Transplant and Donor Natural Killer Cells for Solid Tumors Medical College of Wisconsin Donor NK cells Solid tumors, including high risk malignant brain tumors II NCT01759810 Proteome-based Personalized Immunotherapy of Glioblastoma NeuroVita Clinic (Russia) Dendritic vaccine, hematopoietic stem cells, cytotoxic lymphocytes Recurrent glioblastoma II, III NCT01326104 Vaccine Immunotherapy for Recurrent Medulloblastoma and Primitive Neuroectodermal Tumor University of Florida Total tumor RNA (TTRNA)-loaded DC + ex vivo expanded TTRNA loaded autologous lymphocyte transfer Recurrent Medulloblastoma and Primitive Neuroectodermal Tumor I, II NCT02661282 Phase II Autologous Cytomegalovirus (CMV)-Specific Cytotoxic T Cells for Glioblastoma (GBM) Patients M.D. Anderson Cancer Center Autologous CMV-Specific Cytotoxic T Cells Glioblastoma I, II NCT01109095 CMV-specific Cytotoxic T Lymphocytes Expressing Chimeric Antigen Receptors(CAR) Targeting HER2 in Patients With GBM (HERT-GBM) Baylor College of Medicine Anti-HER2 CAR transduced autologous CMV-specific cytotoxic T-lymphocytes (CTL) Glioblastoma I NCT02442297 T Cells Expressing HER2-specific CAR for Patients With Glioblastoma (iCAR) Baylor College of Medicine Anti-HER2 CAR transduced T Cells Recurrent or refractory GBM I NCT02541370 Treatment of Relapsed and/or Chemotherapy Refractory Advanced Malignancies by CART133 Chinese PLA General Hospital Anti-CD133-CAR modified T Cells Relapsed and/or Chemotherapy Refractory Malignancies I NCT02208362 Genetically Modified T-cells in Treating Patients With Recurrent or Refractory Malignant Glioma City of Hope Medical Center Anti-IL13Rα2 CAR and a truncated CD19 transduced central memory enriched T cells Recurrent/Refractory Malignant Glioma I NCT02209376 Pilot Study of Autologous T Cells Redirected to EGFRVIII-With a Chimeric Antigen Receptor in Patients With EGFRVIII+ Glioblastoma University of Pennsylvania in collaboration with University of California, San Francisco Autologous T cells transduced with a lentiviral vector to express a CAR specific for EGFRvIII Newly diagnosed and recurrent glioblastoma NCT02664363 EGFRvIII CAR T Cells for Newly-Diagnosed GBM (ExCeL) Duke University Medical Center Anti-EGFRvIII CAR transduced T Cells Newly diagnosed glioblastoma I NCT02575261 CAR-T Cell Immunotherapy for EphA2 Positive Malignant Glioma Patients Fuda Cancer Hospital, Guangzhou Anti-EphA2 CAR transduced T cells EphA2 Positive Recurrent and Metastatic Malignant Glioma I, II NCT01454596 CAR T Cell Receptor Immunotherapy Targeting EGFRvIII for Patients With Malignant Gliomas Expressing EGFRvIII National Cancer Institute (NCI) Anti-EGFRvIII CAR transduced PBL Malignant Gliomas Expressing EGFRvIII I, II NCT02617134 CAR-T Cell Immunotherapy in MUC1 Positive Solid Tumor PersonGen Biomedicine (Suzhou) Co., Ltd. Anti-MUC1 CAR transduced T cells MUC1 Positive Advanced Refractory Solid Tumor I, II NCT02331693 CAR T Cells in Treating Patients With Malignant Gliomas Overexpressing EGFR RenJi Hospital Anti-EGFR CAR transduced T cells EGFR-overexpressing malignant glioma I NCT02713984 A Clinical Research of CAR T Cells Targeting HER2 Positive Cancer Southwest Hospital, China Anti-HER2 CAR transduced T Cells HER2 positive cancer I, II Article Highlights Box Cellular immunotherapeutics could be classified into vaccine approaches and adoptive transfer of effector cell approaches. Cellular vaccine approaches utilize tumor cells and/or antigen presenting cells, such as dendritic cells. Adoptive cell transfer (ACT) approaches can utilize a variety of effector cell types, including lymphokine-activated killer (LAK) cells, γδ T cells, tumor-infiltrating lymphocytes (TIL), chimeric antigen receptor (CAR)-T cells and T-cell receptor (TCR) transduced T cells. The ultimate success of cellular immunotherapy in brain tumor patients would require improvements in the areas including feasibility in producing cellular therapeutics as well as strategies to promote effector T-cell trafficking to the tumor site and to overcome local and systemic immune suppression. With advances of technologies allowing antigen-specific targeting of ACTs, it is critical to expand the list of glioma-speicific antigens that can be safely targeted in future immunotherapies. Declaration of Interest H Okada is an inventor in the U.S. Patent Application No. 60,611, 797 (Utility Patent Application) “Identification of an IL-13 Receptor Alpha2 Peptide Analogue Capable of Enhancing Stimulation of Glioma-Specific CTL Response”. An exclusive licensing agreement has been completed on this application between University of Pittsburgh and Stemline, Inc. In this manuscript, the authors discussed a publication using this peptide, but the data interpretation was done by the entire study team and not by Dr Okada himself. Due to the potential conflicts of interest, Hideho Okada did not solely interpret any data in the current manuscript. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed. 1 Ostrom QT Gittleman H Fulop J CBTRUS Statistical Report: Primary Brain and Central Nervous System Tumors Diagnosed in the United States in 2008–2012 Neuro Oncol 2015 10 17 Suppl 4 iv1 iv62 26511214 2 Stupp R Mason WP van den Bent MJ Radiotherapy plus concomitant and adjuvant temozolomide for glioblastoma NEnglJMed 2005 352 10 987 96 3 Hodi FS O’Day SJ McDermott DF Improved survival with ipilimumab in patients with metastatic melanoma The New England journal of medicine 2010 8 19 363 8 711 23 20525992 4 Robert C Long GV Brady B Nivolumab in previously untreated melanoma without BRAF mutation N Engl J Med 2015 1 22 372 4 320 30 25399552 5 Robert C Ribas A Wolchok JD Anti-programmed-death-receptor-1 treatment with pembrolizumab in ipilimumab-refractory advanced melanoma: a randomised dose-comparison cohort of a phase 1 trial Lancet 2014 9 20 384 9948 1109 17 25034862 6 Sundar R Cho BC Brahmer JR Nivolumab in NSCLC: latest evidence and clinical potential Ther Adv Med Oncol 2015 3 7 2 85 96 25755681 7 Kantoff PW Higano CS Shore ND Sipuleucel-T immunotherapy for castration-resistant prostate cancer The New England journal of medicine 2010 7 29 363 5 411 22 20818862 8 Maude SL Frey N Shaw PA Chimeric antigen receptor T cells for sustained remissions in leukemia N Engl J Med 2014 10 16 371 16 1507 17 25317870 9 Grupp SA Kalos M Barrett D Chimeric antigen receptor-modified T cells for acute lymphoid leukemia N Engl J Med 2013 4 18 368 16 1509 18 23527958 10 Dilloo D Bacon K Holden W Combined chemokine and cytokine gene transfer enhances antitumor immunity Nat Med 1996 10 2 10 1090 5 8837606 11 Okada H Lieberman FS Walter KA Autologous glioma cell vaccine admixed with interleukin-4 gene transfected fibroblasts in the treatment of patients with malignant gliomas Journal of translational medicine 2007 5 67 18093335 12 Kusumoto M Umeda S Ikubo A Phase 1 clinical trial of irradiated autologous melanoma cells adenovirally transduced with human GM-CSF gene Cancer Immunol Immunother 2001 9 50 7 373 81 11676397 13 Dranoff G Jaffee E Lazenby A Vaccination with irradiated tumor cells engineered to secrete murine granulocyte-macrophage colony-stimulating factor stimulates potent, specific, and long-lasting anti-tumor immunity Proc Natl Acad Sci U S A 1993 4 15 90 8 3539 43 8097319 14 Dols A Smith JW 2nd Meijer SL Vaccination of women with metastatic breast cancer, using a costimulatory gene (CD80)-modified, HLA-A2-matched, allogeneic, breast cancer cell line: clinical and immunological results Human gene therapy 2003 7 20 14 11 1117 23 12885350 15 Okada H Giezeman-Smits KM Tahara H Effective cytokine gene therapy against an intracranial glioma using a retrovirally transduced IL-4 plus HSVtk tumor vaccine Gene Ther 1999 2 6 2 219 26 10435106 16 Okada H Villa L Attanucci J Cytokine gene therapy of gliomas: effective induction of therapeutic immunity to intracranial tumors by peripheral immunization with interleukin-4 transduced glioma cells Gene Ther 2001 8 8 15 1157 66 11509946 17 Eguchi J Hiroishi K Ishii S Interleukin-4 gene transduced tumor cells promote a potent tumor-specific Th1-type response in cooperation with interferon-alpha transduction Gene Ther 2005 12 733 41 15772692 18 Kuwashima N Nishimura F Eguchi J Delivery of Dendritic Cells Engineered to Secrete IFN-{alpha} into Central Nervous System Tumors Enhances the Efficacy of Peripheral Tumor Cell Vaccines: Dependence on Apoptotic Pathways J Immunol 2005 175 4 2730 40 16081851 19 Okada H Lieberman FS Walter KA Autologous glioma cell vaccine admixed with interleukin-4 gene transfected fibroblasts in the treatment of patients with malignant gliomas J Transl Med 2007 5 1 67 18093335 20 Ishikawa E Muragaki Y Yamamoto T Phase I/IIa trial of fractionated radiotherapy, temozolomide, and autologous formalin-fixed tumor vaccine for newly diagnosed glioblastoma J Neurosurg 2014 9 121 3 543 53 24995786 21 Curry WT Jr Gorrepati R Piesche M Vaccination with Irradiated Autologous Tumor Cells Mixed with Irradiated GM-K562 Cells Stimulates Antitumor Immunity and T Lymphocyte Activation in Patients with Recurrent Malignant Glioma Clinical cancer research : an official journal of the American Association for Cancer Research 2016 2 12 22 Kalinski P Edington H Zeh HJ Dendritic cells in cancer immunotherapy: vaccines or autologous transplants? Immunologic research 2011 8 50 2–3 235 47 21717071 23 Ajay D Sanchez-Perez L Choi BD Immunotherapy with tumor vaccines for the treatment of malignant gliomas Curr Drug Discov Technol 2012 12 9 4 237 55 22339070 24 De Vleeschouwer S Van Gool SW Van Calenbergh F Immunotherapy for malignant gliomas: emphasis on strategies of active specific immunotherapy using autologous dendritic cells Childs Nerv Syst 2005 1 21 1 7 18 15452731 25 Liau LM Black KL Martin NA Treatment of a patient by vaccination with autologous dendritic cells pulsed with allogeneic major histocompatibility complex class I-matched tumor peptides. Case Report Neurosurg Focus 2000 9 6 e8 26 Yu JS Wheeler CJ Zeltzer PM Vaccination of malignant glioma patients with peptide-pulsed dendritic cells elicits systemic cytotoxicity and intracranial T-cell infiltration Cancer research 2001 2 1 61 3 842 7 11221866 27 Wheeler CJ Black KL Liu G Vaccination elicits correlated immune and clinical responses in glioblastoma multiforme patients Cancer research 2008 7 15 68 14 5955 64 18632651 28 Ardon H Van Gool SW Verschuere T Integration of autologous dendritic cell-based immunotherapy in the standard of care treatment for patients with newly diagnosed glioblastoma: results of the HGG-2006 phase I/II trial Cancer Immunol Immunother 2012 11 61 11 2033 44 22527250 29 Prins RM Soto H Konkankit V Gene expression profile correlates with T-cell infiltration and relative survival in glioblastoma patients vaccinated with dendritic cell immunotherapy Clin Cancer Res 2011 3 15 17 6 1603 15 21135147 30 Okada H Brain tumor immunotherapy with type-1 polarizing strategies Ann N Y Acad Sci 2009 9 1174 18 23 19769732 31 Sampson JH Mitchell DA Vaccination strategies for neuro-oncology Neuro Oncol 2015 11 17 Suppl 7 vii15 vii25 26516221 32 Reardon DA Freeman G Wu C Immunotherapy advances for glioblastoma Neuro-oncology 2014 11 16 11 1441 58 25190673 33* Reardon DA Wucherpfennig KW Freeman G An update on vaccine therapy and other immunotherapeutic approaches for glioblastoma Expert Rev Vaccines 2013 6 12 6 597 615 Excellent review providing an updated and comprehensive information on recent immunotherapy approaches for glioblastomas 23750791 34 Okada H Kalinski P Ueda R Induction of CD8+ T-Cell Responses Against Novel Glioma-Associated Antigen Peptides and Clinical Activity by Vaccinations With {alpha}-Type 1 Polarized Dendritic Cells and Polyinosinic-Polycytidylic Acid Stabilized by Lysine and Carboxymethylcellulose in Patients With Recurrent Malignant Glioma J Clin Oncol 2011 1 20 29 3 330 36 21149657 35 Phuphanich S Wheeler CJ Rudnick JD Phase I trial of a multi-epitope-pulsed dendritic cell vaccine for patients with newly diagnosed glioblastoma Cancer Immunol Immunother 2013 1 62 1 125 35 22847020 36 Slingluff CL Jr Petroni GR Yamshchikov GV Clinical and immunologic results of a randomized phase II trial of vaccination using four melanoma peptides either administered in granulocyte-macrophage colony-stimulating factor in adjuvant or pulsed on dendritic cells J Clin Oncol 2003 11 1 21 21 4016 26 14581425 37 Wheeler CJ Das A Liu G Clinical responsiveness of glioblastoma multiforme to chemotherapy after vaccination Clinical cancer research : an official journal of the American Association for Cancer Research 2004 8 15 10 16 5316 26 15328167 38 Bloch O Crane CA Fuks Y Heat-shock protein peptide complex-96 vaccination for recurrent glioblastoma: a phase II, single-arm trial Neuro Oncol 2014 1 16 2 274 9 24335700 39* Litterman AJ Zellmer DM Grinnen KL Profound impairment of adaptive immune responses by alkylating chemotherapy Journal of immunology 2013 6 15 190 12 6259 68 A very well designed and considerate study evaluating impacts of chemotherapy on vaccine-reactive T cells in preclinical glioma vaccine models 40 Dziurzynski K Chang SM Heimberger AB Consensus on the role of human cytomegalovirus in glioblastoma Neuro-Oncology 3 1 2012 14 3 246 55 22319219 41 Cobbs CS Harkins L Samanta M Human cytomegalovirus infection and expression in human malignant glioma Cancer Res 2002 62 12 3347 50 12067971 42* Mitchell DA Batich KA Gunn MD Tetanus toxoid and CCL3 improve dendritic cell vaccines in mice and glioblastoma patients Nature 2015 3 19 519 7543 366 9 An important clinical study demonstrating the role of vaccine-site conditioning by Tetanus Toxoid 25762141 43 Merchant RE Ellison MD Young HF Immunotherapy for malignant glioma using human recombinant interleukin-2 and activated autologous lymphocytes. A review of pre-clinical and clinical investigations Journal of neuro-oncology 1990 4 8 2 173 88 2193121 44 Ishikawa E Takano S Ohno T Adoptive cell transfer therapy for malignant gliomas Adv Exp Med Biol 2012 746 109 20 22639163 45 Law TM Motzer RJ Mazumdar M Phase III randomized trial of interleukin-2 with or without lymphokine-activated killer cells in the treatment of patients with advanced renal cell carcinoma Cancer 1995 9 1 76 5 824 32 8625186 46 Fujisaki H Kakuda H Shimasaki N Expansion of highly cytotoxic human natural killer cells for cancer cell therapy Cancer research 2009 5 1 69 9 4010 7 19383914 47 Barkholt L Alici E Conrad R Safety analysis of ex vivo-expanded NK and NK-like T cells administered to cancer patients: a phase I clinical study Immunotherapy 2009 9 1 5 753 64 20636021 48 Rezvani K Rouce RH The Application of Natural Killer Cell Immunotherapy for the Treatment of Cancer Front Immunol 2015 6 578 26635792 49 Groh V Porcelli S Fabbi M Human lymphocytes bearing T cell receptor gamma/delta are phenotypically diverse and evenly distributed throughout the lymphoid system J Exp Med 1989 4 1 169 4 1277 94 2564416 50 von Lilienfeld-Toal M Nattermann J Feldmann G Activated gammadelta T cells express the natural cytotoxicity receptor natural killer p 44 and show cytotoxic activity against myeloma cells Clin Exp Immunol 2006 6 144 3 528 33 16734623 51 Girardi M Oppenheim DE Steele CR Regulation of cutaneous malignancy by gammadelta T cells Science 2001 10 19 294 5542 605 9 11567106 52 Deniger DC Maiti SN Mi T Activating and propagating polyclonal gamma delta T cells with broad specificity for malignancies Clin Cancer Res 2014 11 15 20 22 5708 19 24833662 53 Nicol AJ Tokuyama H Mattarollo SR Clinical evaluation of autologous gamma delta T cell-based immunotherapy for metastatic solid tumours British journal of cancer 2011 9 6 105 6 778 86 21847128 54 Lamb LS Jr Gammadelta T cells as immune effectors against high-grade gliomas Immunol Res 2009 45 1 85 95 19711198 55 Bryant NL Gillespie GY Lopez RD Preclinical evaluation of ex vivo expanded/activated gammadelta T cells for immunotherapy of glioblastoma multiforme Journal of neuro-oncology 2011 1 101 2 179 88 20532954 56 Pereboeva L Harkins L Wong S The safety of allogeneic innate lymphocyte therapy for glioma patients with prior cranial irradiation Cancer Immunol Immunother 2015 5 64 5 551 62 25676710 57 Lamb LS Jr Bowersock J Dasgupta A Engineered drug resistant gammadelta T cells kill glioblastoma cell lines during a chemotherapy challenge: a strategy for combining chemo- and immunotherapy PLoS One 2013 8 1 e51805 23326319 58** Rosenberg SA Restifo NP Adoptive cell transfer as personalized immunotherapy for human cancer Science 2015 4 3 348 6230 62 8 An updated, comprehensive and insightful review on adoptive cell transfer approach for cancer 25838374 59 Saris SC Spiess P Lieberman DM Treatment of murine primary brain tumors with systemic interleukin-2 and tumor-infiltrating lymphocytes J Neurosurg 1992 3 76 3 513 9 1738033 60 Quattrocchi KB Miller CH Cush S Pilot study of local autologous tumor infiltrating lymphocytes for the treatment of recurrent malignant gliomas Journal of neuro-oncology 1999 45 2 141 57 10778730 61 Glitza IC Haymaker C Bernatchez C Intrathecal Administration of Tumor-Infiltrating Lymphocytes Is Well Tolerated in a Patient with Leptomeningeal Disease from Metastatic Melanoma: A Case Report Cancer Immunol Res 2015 11 3 11 1201 6 26216417 62 Balch CM Riley LB Bae YJ Patterns of human tumor-infiltrating lymphocytes in 120 human cancers Arch Surg 1990 2 125 2 200 5 1689143 63 Plautz GE Barnett GH Miller DW Systemic T cell adoptive immunotherapy of malignant gliomas J Neurosurg 1998 7 89 1 42 51 9647171 64 Plautz GE Miller DW Barnett GH T cell adoptive immunotherapy of newly diagnosed gliomas Clin Cancer Res 2000 6 6 6 2209 18 10873070 65 Morgan RA Dudley ME Wunderlich JR Cancer Regression in Patients After Transfer of Genetically Engineered Lymphocytes Science 2006 314 5796 126 29 16946036 66 Stanislawski T Voss RH Lotz C Circumventing tolerance to a human MDM2-derived tumor antigen by TCR gene transfer Nat Immunol 2001 10 2 10 962 70 11577350 67 Johnson LA Morgan RA Dudley ME Gene therapy with human and mouse T-cell receptors mediates cancer regression and targets normal tissues expressing cognate antigen Blood 2009 7 16 114 3 535 46 19451549 68 Xue SA Gao L Thomas S Development of a Wilms’ tumor antigen-specific T-cell receptor for clinical trials: engineered patient’s T cells can eliminate autologous leukemia blasts in NOD/SCID mice Haematologica 2010 1 95 1 126 34 19679884 69 Okamoto S Mineno J Ikeda H Improved expression and reactivity of transduced tumor-specific TCRs in human lymphocytes by specific silencing of endogenous TCR Cancer Res 2009 12 1 69 23 9003 11 19903853 70 Govers C Sebestyen Z Coccoris M T cell receptor gene therapy: strategies for optimizing transgenic TCR pairing Trends Mol Med 2010 2 16 2 77 87 20122868 71 Zhang L Morgan RA Genetic engineering with T cell receptors Adv Drug Deliv Rev 2012 6 1 64 8 756 62 22178904 72 Parkhurst MR Yang JC Langan RC T cells targeting carcinoembryonic antigen can mediate regression of metastatic colorectal cancer but induce severe transient colitis Mol Ther 2011 3 19 3 620 6 21157437 73** Morgan RA Chinnasamy N Abate-Daga D Cancer regression and neurological toxicity following anti-MAGE-A3 TCR gene therapy Journal of immunotherapy 2013 2 36 2 133 51 A very important report demonstrating the importance of selecting appropriate antigens as well as previously unrecognized levels of susceptibility of the brain to T cell responses 23377668 74 Morgan RA Yang JC Kitano M Case Report of a Serious Adverse Event Following the Administration of T Cells Transduced With a Chimeric Antigen Receptor Recognizing ERBB2 Mol Ther 2010 18 4 843 51 20179677 75 Robbins PF Kassim SH Tran TL A pilot trial using lymphocytes genetically engineered with an NY-ESO-1-reactive T-cell receptor: long-term follow-up and correlates with response Clin Cancer Res 2015 3 1 21 5 1019 27 25538264 76 Rapoport AP Stadtmauer EA Binder-Scholl GK NY-ESO-1-specific TCR-engineered T cells mediate sustained antigen-specific antitumor effects in myeloma Nat Med 2015 8 21 8 914 21 26193344 77 Ngo MC Rooney CM Howard JM Ex vivo gene transfer for improved adoptive immunotherapy of cancer Human Molecular Genetics 4 15 2011 20 R1 R93 R99 21415041 78 Moeller M Haynes NM Trapani JA A functional role for CD28 costimulation in tumor recognition by single-chain receptor-modified T cells Cancer Gene Ther 2004 11 5 371 79 15060573 79 Beckman RA Weiner LM Davis HM Antibody constructs in cancer therapy - Protein engineering strategies to improve exposure in solid tumors Cancer 2007 1 109 2 170 79 17154393 80 Weijtens M Willemsen R Valerio D Single chain Ig/gamma gene-redirected human T lymphocytes produce cytokines, specifically lyse tumor cells, and recycle lytic capacity The Journal of Immunology 7 15 1996 157 2 836 43 8752936 81 Maus MV June CH Making Better Chimeric Antigen Receptors for Adoptive T-cell Therapy Clin Cancer Res 2016 4 15 22 8 1875 84 27084741 82 Kahlon KS Brown C Cooper LJ Specific recognition and killing of glioblastoma multiforme by interleukin 13-zetakine redirected cytolytic T cells Cancer research 2004 12 15 64 24 9160 6 15604287 83 Ahmed N Salsman VS Kew Y HER2-specific T cells target primary glioblastoma stem cells and induce regression of autologous experimental tumors Clin Cancer Res 2010 1 15 16 2 474 85 20068073 84 Chow KK Naik S Kakarla S T cells redirected to EphA2 for the immunotherapy of glioblastoma Mol Ther 2013 3 21 3 629 37 23070117 85 Ohno M Ohkuri T Kosaka A Expression of miR-17–92 enhances anti-tumor activity of T-cells transduced with the anti-EGFRvIII chimeric antigen receptor in mice bearing human GBM xenografts Journal for immunotherapy of cancer 2013 1 21 24829757 86 Johnson LA Scholler J Ohkuri T Rational development and characterization of humanized anti-EGFR variant III chimeric antigen receptor T cells for glioblastoma Sci Transl Med 2015 2 18 7 275 275ra22 87 Bullain SS Sahin A Szentirmai O Genetically engineered T cells to target EGFRvIII expressing glioblastoma Journal of neuro-oncology 2009 9 94 3 373 82 19387557 88 Brown CE Badie B Barish ME Bioactivity and Safety of IL13Ralpha2-Redirected Chimeric Antigen Receptor CD8+ T Cells in Patients with Recurrent Glioblastoma Clin Cancer Res 2015 9 15 21 18 4062 72 26059190 89 Pule MA Savoldo B Myers GD Virus-specific T cells engineered to coexpress tumor-specific receptors: persistence and antitumor activity in individuals with neuroblastoma Nat Med 2008 11 14 11 1264 70 18978797 90 Hegde M Corder A Chow KK Combinational targeting offsets antigen escape and enhances effector functions of adoptively transferred T cells in glioblastoma Mol Ther 2013 11 21 11 2087 101 23939024 91 Kloss CC Condomines M Cartellieri M Combinatorial antigen recognition with balanced signaling promotes selective tumor eradication by engineered T cells Nat Biotechnol 2013 1 31 1 71 5 23242161 92 Wilkie S van Schalkwyk MC Hobbs S Dual targeting of ErbB2 and MUC1 in breast cancer using chimeric antigen receptors engineered to provide complementary signaling J Clin Immunol 2012 10 32 5 1059 70 22526592 93 Grada Z Hegde M Byrd T TanCAR: A Novel Bispecific Chimeric Antigen Receptor for Cancer Immunotherapy Mol Ther Nucleic Acids 2013 2 e105 23839099 94 Minagawa K Zhou X Mineishi S Seatbelts in CAR therapy: How Safe Are CARS? Pharmaceuticals (Basel) 2015 8 2 230 49 26110321 95 Roybal KT Rupp LJ Morsut L Precision Tumor Recognition by T Cells With Combinatorial Antigen-Sensing Circuits Cell 2016 2 11 164 4 770 9 26830879 96 Schumacher T Bunse L Pusch S A vaccine targeting mutant IDH1 induces antitumour immunity Nature 2014 06/25/online;advance online publication 97 Cameron BJ Gerry AB Dukes J Identification of a Titin-derived HLA-A1-presented peptide as a cross-reactive target for engineered MAGE A3-directed T cells Sci Transl Med 2013 8 7 5 197 197ra03 98 Sampson JH Heimberger AB Archer GE Immunologic Escape After Prolonged Progression-Free Survival With Epidermal Growth Factor Receptor Variant III Peptide Vaccination in Patients With Newly Diagnosed Glioblastoma Journal of Clinical Oncology 11 1 2010 28 31 4722 29 20921459 99 Kohanbash G Okada H Myeloid-derived suppressor cells in gliomas and glioma-development Immunol Invest 2012 41 6–7 658 79 23017140 100 Okada H Butterfield LH Hamilton RL Induction of Robust Type-I CD8+ T-cell Responses in WHO Grade 2 Low-Grade Glioma Patients Receiving Peptide-Based Vaccines in Combination with Poly-ICLC Clinical cancer research : an official journal of the American Association for Cancer Research 2015 1 15 21 2 286 94 25424847 101 Okada H Tsugawa T Sato H Delivery of Interferon-alpha transfected dendritic cells into central nervous system tumors enhances the antitumor efficacy of peripheral peptide-based vaccines Cancer Res 2004 64 16 5830 38 15313927 102 Pollack IF Jakacki RI Butterfield LH Antigen-specific immune responses and clinical outcome after vaccination with glioma-associated antigen peptides and polyinosinic-polycytidylic acid stabilized by lysine and carboxymethylcellulose in children with newly diagnosed malignant brainstem and nonbrainstem gliomas J Clin Oncol 2014 7 1 32 19 2050 8 24888813 103 Kershaw MH Wang G Westwood JA Redirecting migration of T cells to chemokine secreted from tumors by genetic modification with CXCR2 Human gene therapy 2002 11 1 13 16 1971 80 12427307 104 Di Stasi A De Angelis B Rooney CM T lymphocytes coexpressing CCR4 and a chimeric antigen receptor targeting CD30 have improved homing and antitumor activity in a Hodgkin tumor model Blood 2009 6 18 113 25 6392 402 19377047 105 Craddock JA Lu A Bear A Enhanced tumor trafficking of GD2 chimeric antigen receptor T cells by expression of the chemokine receptor CCR2b J Immunother 2010 10 33 8 780 8 20842059 106 Moon EK Carpenito C Sun J Expression of a functional CCR2 receptor enhances tumor localization and tumor eradication by retargeted human T cells expressing a mesothelin-specific chimeric antibody receptor Clin Cancer Res 2011 7 15 17 14 4719 30 21610146 107 Grossman SA Ye X Lesser G Immunosuppression in Patients with High-Grade Gliomas Treated with Radiation and Temozolomide Clinical Cancer Research 8 15 2011 17 16 5473 80 21737504 108 Chmielewski M Abken H TRUCKs: the fourth generation of CARs Expert opinion on biological therapy 2015 15 8 1145 54 25985798 109 Kerkar SP Muranski P Kaiser A Tumor-specific CD8+ T cells expressing interleukin-12 eradicate established cancers in lymphodepleted hosts Cancer research 2010 9 1 70 17 6725 34 20647327 110 Zhang L Kerkar SP Yu Z Improving adoptive T cell therapy by targeting and controlling IL-12 expression to the tumor environment Mol Ther 2011 4 19 4 751 9 21285960 111 Chmielewski M Kopecky C Hombach AA IL-12 release by engineered T cells expressing chimeric antigen receptors can effectively Muster an antigen-independent macrophage response on tumor cells that have shut down tumor antigen expression Cancer research 2011 9 1 71 17 5697 706 21742772 112 Bollard CM Rossig C Calonge MJ Adapting a transforming growth factor beta-related tumor protection strategy to enhance antitumor immunity Blood 2002 5 1 99 9 3179 87 11964281 113 Abate-Daga D Hanada K Davis JL Expression profiling of TCR-engineered T cells demonstrates overexpression of multiple inhibitory receptors in persisting lymphocytes Blood 2013 8 22 122 8 1399 410 23861247 114 John LB Devaud C Duong CP Anti-PD-1 antibody therapy potently enhances the eradication of established tumors by gene-modified T cells Clin Cancer Res 2013 10 15 19 20 5636 46 23873688 115 John LB Kershaw MH Darcy PK Blockade of PD-1 immunosuppression boosts CAR T-cell therapy Oncoimmunology 2013 10 1 2 10 e26286 24353912 116 Okada H Pollack IF Do we need novel radiologic response criteria for brain tumor immunotherapy? Expert Review of Neurotherapeutics 2011 5 01 11 5 619 22 21539483 117* Okada H Weller M Huang R Immunotherapy Response Assessment in Neuro-Oncology (iRANO): A Report of the RANO Working Group Lancet Oncol 2015 15 534 42 This proposes a novel response criteria specifically designed for immunotherapy of brain tumors 118 Wolchok JD Hoos A O’Day S Guidelines for the Evaluation of Immune Therapy Activity in Solid Tumors: Immune-Related Response Criteria Clinical Cancer Research 12 1 2009 15 23 7412 20 19934295 119 Wen PY Macdonald DR Reardon DA Updated Response Assessment Criteria for High-Grade Gliomas: Response Assessment in Neuro-Oncology Working Group Journal of Clinical Oncology 4 10 2010 28 11 1963 72 20231676

PMC005xxxxxx/PMC5133694.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 101564430 39699 Eur J Prev Cardiol Eur J Prev Cardiol European journal of preventive cardiology 2047-4873 2047-4881 26025448 5133694 10.1177/2047487315588758 NIHMS829575 Article Association between fine particulate matter exposure and subclinical atherosclerosis: A meta-analysis Akintoye Emmanuel 1 Shi Liuhua 2 Obaitan Itegbemie 1 Olusunmade Mayowa 1 Wang Yan 2 Newman Jonathan D 3 Dodson John A 3 1 Master of Public Health Program, School of Public Health, Harvard University, Boston, USA 2 Department of Environmental Health – Exposure, Epidemiology and Risk Program, School of Public Health, Harvard University, Boston, USA 3 Leon H Charney Division of Cardiology, Department of Medicine, New York University School of Medicine, New York, USA Corresponding author: John A Dodson, Leon H Charney Division of Cardiology, Department of Medicine, New York University School of Medicine, 227 East 30th Street, TRB 851, New York, NY 10016, USA., John.Dodson@nyumc.org, Twitter: @emmassontweet 13 11 2016 29 5 2015 4 2016 02 12 2016 23 6 602612 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. Background Epidemiological studies in humans that have evaluated the association between fine particulate matter (PM2.5) and atherosclerosis have yielded mixed results. Design In order to further investigate this relationship, we conducted a comprehensive search for studies published through May 2014 and performed a meta-analysis of all available observational studies that investigated the association between PM2.5 and three noninvasive measures of clinical and subclinical atherosclerosis: carotid intima media thickness, arterial calcification, and ankle-brachial index. Methods and results Five reviewers selected studies based on predefined inclusion criteria. Pooled mean change estimates and 95% confidence intervals were calculated using random-effects models. Assessment of between-study heterogeneity was performed where the number of studies was adequate. Our pooled sample included 11,947 subjects for carotid intima media thickness estimates, 10,750 for arterial calcification estimates, and 6497 for ankle-brachial index estimates. Per 10 μg/m3 increase in PM2.5 exposure, carotid intima media thickness increased by 22.52 μm but this did not reach statistical significance (p = 0.06). We did not find similar associations for arterial calcification (p = 0.44) or ankle-brachial index (p = 0.85). Conclusion Our meta-analysis supports a relationship between PM2.5 and subclinical atherosclerosis measured by carotid intima media thickness. We did not find a similar relationship between PM2.5 and arterial calcification or ankle-brachial index, although the number of studies was small. Particulate matter air pollution tunica intima vascular calcification ankle-brachial index Introduction Exposure to fine particulate matter with aerodynamic diamete ≤2.5 μm (PM2.5) has been shown to have adverse health effects on multiple organ systems.1,2 Inhaled PM2.5 can be deposited deep in alveoli and is hypothesized to enhance inflammation and oxidative stress and alter cardiac autonomic activity.3–5 Though earlier studies primarily focused on respiratory health outcomes, there is evidence that PM2.5 is a risk factor for cardiovascular disease (CVD) events5,6 including hypertension,7,8 cardiovascular mortality,1,9 and increased hospital admissions for CVD.10 Experimental animal studies have reported more rapid progression of atherosclerosis with long-term ambient particulate matter exposure compared with filtered air.4,11 However, human studies, which cannot be performed in a controlled manner, are limited to observational cohorts that have yielded mixed results.12–14 In addition, these studies have assessed different measures of clinical and subclinical atherosclerosis, including carotid intima media thickness (CIMT), arterial calcification (coronary aortic calcification (CAC); abdominal aortic calcification (AAC); or thoracic aortic calcification (TAC)), and ankle-brachial index (ABI). In light of the prior inconclusive associations between PM2.5 and atherosclerosis, as well as the potential heterogeneity in study methodologies and outcomes, we therefore conducted a systematic review and meta-analysis of studies published to date. Methods Search strategy We followed the recommendations of the Meta-analysis of Observational Studies in Epidemiology (MOOSE) Group in the design, implementation and reporting of our study.15 We conducted a comprehensive literature search of four databases –MEDLINE, EMBASE, Web of Science and Environmental Index – to identify relevant articles that were published through May 2014. Our search queries combined the exposure (PM2.5) with atherosclerosis and surrogate markers of atherosclerosis. Search terms for MEDLINE were (“Particulate Matter”[mesh] OR “Air Pollution”[mesh] OR air pollution[tiab] OR particulate*[ tiab] OR fine partic*[tiab] OR pm2.5[tiab] OR pm 2.5[tiab]) AND (“Arteriosclerosis”[mesh] OR “Tunica Intima”[Mesh] OR intima[tiab] OR “Vascular Calcification”[Mesh] OR “coronary artery calcification” OR “ankle brachial index” [Mesh] OR arterioscleros*[tiab] OR atheroscleros*[tiab] OR atherogen*[ tiab] OR arterial disease*[tiab] OR arterial occlus*[tiab]) NOT (“animals”[mesh] NOT “humans”[mesh]). Full details of our search strategies for the other databases are available in Supplementary Material online (Supplemental Table). Study selection We included cross-sectional and longitudinal cohort studies evaluating associations between PM2.5 and clinical or subclinical atherosclerosis as assessed by CIMT, arterial calcification, or ABI. All languages were included in our search. For the meta-analysis, we excluded non-human studies, studies reporting environmental exposures other than PM2.5, and studies reporting estimates other than absolute change in outcome per change in the level of PM2.5. In addition, for overlapping studies from the same cohort, we included only the most comprehensive or updated study with the most extensive method of covariate adjustment. Five authors (EA, LS, IO, OM, JAD) independently evaluated non-duplicate abstracts found in the four databases (N = 1505) using our search algorithm. Articles deemed relevant to our study (N = 220) were then selected for independent review of full text and references by two separate authors (Figure 1). We applied our inclusion and exclusion criteria to determine articles for final inclusion; disagreements between two authors were resolved by a third author. Reference lists of all relevant articles (including review articles) were scanned to identify publications that were potentially missed by our initial literature search. Data extraction Data from the final selected manuscripts were independently extracted by two authors and compared to ensure accuracy. Information extracted included citation data, authors’ names, publication year, data source, country, sample size, age distribution, sex distribution, year of data collection, study design, baseline exposure level, outcome measure, effect estimate, and standard error of effect estimate. For studies that reported multiple effect estimates, we extracted the estimate from the main model or model that reflected the greatest degree of control for potential confounders. For each included manuscript, we extracted mean change in CIMT, relative risk for arterial calcification, or mean change in ABI, as applicable. Statistical analysis In order to ensure uniformity of exposure across studies, all estimates were standardized to per 10 μg/m3 increase in PM2.5. Effect measures were pooled using the random effect model of DerSimonian and Laird to account for between-study variation.16 Heterogeneity between studies was explored by visual inspection of the forest plot, Cochran Q statistic (p<0.05), and I-squared (I2) statistic. Consistent with prior thresholds we considered an I2 statistic ≥50% to represent substantial heterogeneity and ≥75% to represent considerable heterogeneity.17 We assessed potential sources of heterogeneity such as year of publication, country, study design, sample size, and baseline level of exposure by using meta-regression. We did not perform an assessment for publication bias given the small number of studies for each endpoint.18 All statistical tests were two-sided and p values less than 0.05 were considered to be statistically significant. Analyses were conducted with STATA Version 13. Results Of the 12 manuscripts considered for data extraction, four were from the Multi-ethnic Study of Atherosclerosis (MESA; providing three CIMT, three arterial calcification, and one ABI estimate), four were from the German Heinz Nixdorf Recall Study (HNRS; providing one CIMT, two arterial calcification, and one ABI estimate), and two (both CIMT) were from the same author (Kunzli et al.). After retaining only one article per cohort for each endpoint, our final sample included eight manuscripts, from which we extracted five CIMT, two CAC, and two ABI estimates. We analyzed a total of 18,590 subjects (mean age 58 years, 52% female). As some studies reported more than one type of atherosclerotic marker (e.g. all markers were evaluated in the MESA cohort), 11,947 subjects contributed to the CIMT estimates, 10,750 contributed to the arterial calcification estimate and 6497 contributed to the ABI estimates. The mean level of PM2.5 exposure among studies ranged from 13.66 μg/m3 to 22.8 μg/m3. Other study characteristics are summarized in Table 1.12–14,19–27 Although there were some differences in exposure assessment, overall the methods of assessment were comparable across studies and within each endpoint (Table 2). CIMT Meta-analysis of the five studies evaluating the outcome of CIMT demonstrated that CIMT increased by 22.52 μm for every 10 μg/m3 increase in PM2.5 but this association did not reach statistical significance (p = 0.06) (Figure 2). There was considerable heterogeneity between studies (I2 = 83%, p<0.01), although exploration using meta-regression showed that year of publication (p = 0.61), country (p = 0.23), study design (p = 0.52), sample size (p = 0.50), and baseline level of exposure (p = 0.97) did not explain this heterogeneity. Arterial calcification We found five manuscripts reporting on the association between PM2.5 exposure and three subtypes of arterial calcification: CAC (three), AAC (one), and TAC (one), but three were from the MESA cohort while the remaining two were from the HNRS cohort. After excluding overlapping studies from the same cohort, only two manuscripts reporting on CAC were retained for the final analysis, which yielded a non-significant positive association (relative risk = 1.35 per 10 μg/m3 increase in PM2.5, p = 0.44) (Figure 3). Given the small number of studies we did not test for heterogeneity. ABI Only two manuscripts reported on the association between PM2.5 exposure and peripheral arterial disease by ABI. Our meta-analysis of these two studies yielded a non-significant association between PM2.5 and ABI (change in ABI per 10 μg/m3 increase in PM2.5 = – 0.001, p = 0.85) (Figure 4). As with arterial calcification, we did not consider a test for heterogeneity given the low number of studies. Discussion Human studies that have investigated the association between PM2.5 and clinical and subclinical atherosclerosis have yielded mixed results. In addition, a variety of different outcome measures have been used. In order to summarize the available evidence in the literature, we conducted a meta-analysis among eight studies comprising 18,590 subjects. We found marginal evidence to support the association between PM2.5 exposure and CIMT. While there was considerable heterogeneity (I2 = 83%) among CIMT studies, the positive association between PM2.5 and CIMT appeared to be consistent across all but one study. Though not statistically significant, our findings demonstrated that for every 10 μg/m3 increase in PM2.5, CIMT increased by 22.52 μm (95% confidence interval (CI) –1.26, 46.29 μm). This estimate is within the range of change in CIMT that has been associated with CVD events.28 While the magnitude of effect of PM2.5 exposure may appear relatively small, PM2.5 exposure is common with a wide range in world-wide exposures. It can vary from as low as a US Environmental Protection Agency recommended level of ≤35 μg/m3 to as high as over 200 μg/m3 in countries such as China,29,30 with levels that may exceed 500 μg/ m3. Therefore, a 200 μg/m3 increase in PM2.5 would be expected to translate into roughly a 450 μm increase in CIMT, an estimation that would be of significant clinical impact given that average CIMT in the general population is around 800 μm.31 Also of note, potential interactions were evaluated in three of the five CIMT studies, and two of these (Lenters et al.,22 Kunzli et al.21) suggested that the association between PM2.5 and CIMT was stronger in females than in males. Kunzli et al. also reported a significant interaction with age, indicating a stronger association in participants ≥60 years compared with participants<60 years. Accordingly, there may be subgroups that are at particularly high risk of adverse effects from PM2.5. Though the exact mechanism of the association between PM2.5 and cardiovascular disease remains uncertain, experimental animal studies have suggested some mechanistic links between PM2.5 and increased CIMT. PM2.5 may provoke an inflammatory response and cytokine release from the pulmonary vascular bed, altering vasomotor tone and lipid peroxidation.4,11,32,33 These studies have emphasized the relationship between PM2.5 exposure and pro-oxidant and pro-inflammatory mediators important in the pathogenesis of atherosclerosis. However, most of these experiments involve concurrent administration of a high fat diet in order to accelerate atheroma formation, thereby indicating that dietary factors may be an important modifier of the effect of PM2.5. We did not find evidence of an association between PM2.5 and either CAC (estimated based on Agatston score34) or ABI (measured by Doppler ultrasound). For both measures we found a low number of studies and thereforewemay have been underpowered to find a meaningful association. Alternatively, it is possible that CIMT may identify areas of increased thickness and nonocclusive atherosclerotic plaque, which may represent earlier stages of arterial injury or atherosclerosis than measures of arterial calcification orABI.28 It is also possible that the atherosclerotic mechanism of PM2.5 directly alters CIMT with little or no influence on arterial calcification or ABI. Our meta-analysis has several strengths, including a protocol-driven approach in order to limit bias in study selection, as well as a broad population represented, including a wide age range that was studied across three different countries (United States, Germany, and Netherlands). However, there are potential limitations that deserve consideration. First, the number of studies, particularly for arterial calcification and ABI, was small, limiting our ability to derive strong conclusions from these analyses and to explore for potential sources of heterogeneity. Second, we found evidence of significant heterogeneity among CIMT estimates, thereby limiting the generalizability of our results. Third, most studies were cross-sectional, which limits estimations of causality given the lack of temporality between exposure and outcome. However, these cross-sectional studies did adjust for major cardiovascular risk factors, demographic information, as well as socio-economic status, which could potentially confound the association between PM2.5 and atherosclerosis. Lastly, despite a rigorous methodology within each study, there is the potential for measurement error in the assessment of exposure and/or outcome that may have biased study estimates towards no-association. In conclusion, we found a positive association across multiple studies between PM2.5 and subclinical atherosclerosis as measured by CIMT, although this did not reach statistical significance. There was considerable heterogeneity among studies, which may limit the generalizability of this finding. We did not find similar associations between PM2.5 and other surrogate markers of atherosclerosis (arterial calcification or ABI), which may be due to lower sensitivity of these indices or lack of a sufficient number of studies. More studies may be needed to explore potential sources of heterogeneity among CIMT estimates, and to further assess the association between PM2.5 and the other surrogate markers. We thank Paul Bain PhD at Harvard University’s Countway Library for providing guidance for our literature search. We also thank Chung-Cheng Hsieh ScD and Julie Goodman PhD at Harvard School of Public Health for providing guidance in our methods and statistical analysis. Funding Dr. Newman is supported by a grant from the National Institute for Diabetes and Digestive and Kidney Diseases (U24DK076169-09, subcontract 25732-60). Dr. Dodson is supported by a grant from the National Institute of Aging (R03AG045067) and a T Franklin Williams Scholarship Award (funding provided by: Atlantic Philanthropies, Inc, the John A. Hartford Foundation, the Alliance for Academic Internal Medicine-Association of Specialty Professors, and the American College of Cardiology). None of the remaining authors receive any specific grant from any funding agency in the public, commercial, or not for profit sectors. Figure 1 Screening and selection process. Of 1505 non-duplicate articles found, 220 were retained after screening titles and abstracts. These articles underwent full-text review by two separate investigators; there were 12 studies that met our predefined inclusion criteria; after removing overlapping cohorts, our final sample included eight studies. Figure 2 Meta-analysis of mean change in carotid intima media thickness (CIMT) per 10 μg/m3 increase in PM2.5. Mean change in CIMT from each study was standardized to per 10 μg/m3 increase in PM2.5. ES: standardized estimate; CI: confidence interval; PM2.5: particulate matter with aerodynamic diameter ≤2.5 μm Figure 3 Meta-analysis of relative risk of arterial calcification per 10 μg/m3 increase in PM2.5. Relative risk estimates from each study were standardized to per 10μg/m3 increase in PM2.5. ES: standardized estimate; CI: confidence interval; PM2.5: particulate matter with aerodynamic diameter ≤2.5 μm Figure 4 Meta-analysis of mean change in ankle-brachial index (ABI) per 10μg/m3 increase in PM2.5. Mean change in ABI from each study was standardized to per 10 μg/m3 increase in PM2.5. ES: standardized estimate; CI: confidence interval; PM2.5: particulate matter with aerodynamic diameter ≤2.5 μm Table 1 Demographic characteristics of the 12 studies that evaluated the association between PM2.5 and atherosclerosis. First author and year Outcome Country Study design Data collection N Age, years Mean (SD/range) % female Data source Baseline exposure level (mg/m3) Adar 201312 a CIMT USA Longitudinal 2000–2005 5362 62 (10) 52 MESA 16.6 Bauer 201013 CIMT Germany Cross-sectional 2000–2003 3380 60 (7.7) 48 HNRS 16.8 Breton 201214 CIMT USA Cross-sectional 2007–2009 768 20 (1.5) 59 TROY 15.7 Sun 201319 CIMT USA Cross-sectional 2000–2002 6256 62 (45–84) 52 MESA 13.66 Diez Roux 2008 20a CIMT USA Longitudinal 2000–2002 5037 62 (45–84) 53 MESA 16.7 Künzli 200521 CIMT USA Cross-sectional 1998–2003 798 59 (9.8) 45 VEAPS & BVAIT 20.3 Lenters 201022 CIMT Netherlands Longitudinal 1999–2000 745 28 (0.9) 53 Not reported 20.7 Künzli 201023a CIMT USA Longitudinal 1995–2007 1438 59 (9.6) 63 Five trialsb 20.79 Allen 200924 a AAC USA Cross-sectional 2000–2002 1147 66 (9.4) 50 MESA 15.8 Kälsch 201425 a TAC Germany Cross-sectional 2000–2003 4238 60 (7.8) 50 HNRS 16.62 Sun 201319 CAC USA Cross-sectional 2000–2002 6256 62 (45–84) 52 MESA 13.66 Diez Roux 2008 20a CAC USA Cross-sectional 200–2002 2149 62 (45–84) 53 MESA 16.7 Hoffmann 200726 CAC Germany Cross-sectional 200–2003 4494 60 (7.8) 51 HNRS 22.8 Hoffmann 200927 ABI Germany Cross-sectional 200–2003 4348 60 (7.8) 51 HNRS 22.8 Diez Roux 2008 20 ABI USA Cross-Sectional 2000–2002 2149 62 (45–84) 52 MESA 16.7 a Study was excluded from the meta-analysis due to cohort overlap. b The five trials included in Kunzli 2010 are: B-vitamin Atherosclerosis Intervention Trial (BVAIT), Vitamin E Atherosclerosis Prevention Study (VEAPS), Estrogen in the Prevention of Atherosclerosis Trial (EPAT), Troglitazone Atherosclerosis Regression Trial (TART), Women’s Estrogen-Progestin Lipid-Lowering Hormone Atherosclerosis Regression Trial (WELL-HART). PM2.5: particulate matter with aerodynamic diameter ≤ 2.5 μm; CIMT: carotid intima media thickness; CAC: coronary artery calcification; ABI: ankle-brachial index; HNRS: Heinz Nixdorf Recall Study; MESA: Multi-Ethnic Study of Atherosclerosis; TROY: Testing Responses On Youth study Table 2 Methodological details in eight studies included in the meta-analysis. First author and year Location Exposure Method of exposure measurement Outcome Method of outcome measurement Confounders adjusted for Bauer 201013 Essen, Mulheim and Bochum PM2.5 Average of the previous 365 days of daily surface concentrations of PM2.5 were taken for each participant. A chemistry transport model was used with input data from emission inventories, meteorology, and regional topography. CIMT B-mode ultrasound. The mean of all 10 manual measurements on both was used as the outcome variable. City, area of residence, age, sex, education, economic activity, smoking variables, environmental tobacco smoke, alcohol consumption, physical activity, BMI, diabetes, LDL-C, HDL-C, intake of statins. Breton 201214 Southern CA PM2.5 Residential addresses geocoded. Spatial interpolation of ambient air quality data from four stations used along with geocoded residential address of participants. Air quality data was interpolated using inverse distance squared weighting. Air pollutant estimates were from EPA’s Air Quality System database. CIMT The mean of 70–100 measurements of the right common carotid artery was used. Instrument was a high resolution B-mode ultrasound attached to a 10MHz linear array transducer. Age, sex, race/ethnicity, BMI, systolic blood pressure, secondhand smoke in childhood, current secondhand smoke, hsCRP, LDL-C, HDL-C. No effect modification by any variable was found on analysis Lenters 201022 Utrecht, Netherlands PM2.5 Overall concentrations of PM2.5 for the year 2000 at the residential address were assessed regionally via interpolation of regional background concentrations. The urban component was assessed with regression models using data on 10 categories of land use in 100-m grids, population density and land use predictors. CIMT High resolution B-mode ultrasound of the right and left common carotid arteries using a 7.5MHz linear array transducer. Both the mean and the maximum CIMT were assessed. Age, sex, BMI, pack years of active smoking, exposure to secondhand smoke in childhood, alcohol intake, highest education, highest profession, diabetes mellitus, neighborhood income, hypertension, HDL-C, LDL-C, family history of CVD. Non-significance evidence of effect modification by sex (stronger association in women), smoking status (stronger association with smokers), and education (stronger association in the less educated). Künzli 200521 Los Angeles, CA PM2.5 Geostatistical model derived for mean home outdoor PM2.5 level, data source was year 2000 data obtained from 23 state and local district monitoring stations. Residential geocoding used. CIMT High-resolution far wall B-mode ultrasound images of the right common carotid artery. Age, sex, education, income, current secondhand smoke, current personal smoking, former personal smoking, blood pressure, LDL-C, antihypertensive medications, lipid lowering medications. Effect modification by age and sex was found, with the association strongest among elderly women aged ≥60 years. Sun 201319 Los Angeles County, CA; Chicago, IL; Baltimore MD; St Paul, MN; Forsyth County, NC; and New York, NY PM2.5 Residential addresses geocoded. Three different approaches were used: the annual average concentration of the two week measurement at the monitor nearest to each study participant’s residence, inverse distance weighting of all annual average monitor concentrations in each area relative to each subject’s residence, and city wide average concentrations based on all monitors within each area. CIMT CAC High resolution B-mode ultrasound. Mean far wall thickness of the right common carotid retrospectively gated to end-diastole was used. Two chest CT scans per participant. Mean Agatston score34 of the scans used for analysis. Age, gender, race and ethnicity, total cholesterol, LDL-C, smoking status, hypertension, lipid lowering medications, level of education, waist circumference, family income, body surface area, BMI, squared BMI, diabetes, HDL-C, triglycerides. Effect modification by variables not assessed. Hoffmann 200726 The three cities: Essen, Mulheim and Bochum (Germany) of the Ruhr area in Germany PM2.5 Residential geocoding. EURAD modeling of daily mean values for PM2.5 for the year 2002 with input data from official emission inventories, meteorological information, and regional topographical data. Annual average calculated for each grid and assigned to each participant living in that grid. CAC Use of chest CTs for each participant. CAC score calculated by the Agatston score. Final CAC score was summation of CAC scores of all foci in the epicardial coronary system. Age, sex, city of residence, area of residence, education, smoking, physical inactivity, waist to hip ratio, diabetes, blood pressure and lipids. No effect modification was carried out although a subgroup analysis of elderly patients was carried out. Hoffmann 200927 The three cities: Essen, Mulheim and Bochum (Germany) of the Ruhr area in Germany PM2.5 Residential geocoding. EURAD dispersion and chemistry transport modeling of daily mean values for PM2.5 for the year 2002 with input data from official emission inventories, meteorological information, and regional topographical data. Annual average calculated for each grid and assigned to each participant living in that grid. ABI 8MHz Doppler transducer used. The index was calculated as the ratio of: highest ankle artery pressures measured either in the posterior tibial or the dorsalis pedis artery and the highest systolic brachial pressure measured in the right and left arm. Age, sex, city of residence, area of residence, education, smoking, physical inactivity, waist to hip ratio, diabetes, BMI, socioeconomic status, lipid lowering medication, antihypertensive medication, blood pressure and lipids. Non-significant effect modification by age was found with a stronger association among enrolled subjects ≥60 years of age. Diez Roux 200820 Los Angeles County, CA; Chicago, IL; Baltimore MD; St Paul, MN; Forsyth County, NC; and New York, NY PM2.5 Spatiotemporal modeling of the monthly mean PM2.5 measures for the prior 20 years with data obtained from the US EPA’s aerometric information retrieval service database. Residential geocoding used. ABI 5 MHz probe on a hand held Doppler instrument. For each lower extremity, ABI numerator was the highest pressure (dorsalis pedis or posterior tibial from that leg) obtained. ABI denominator was the averaged brachial artery blood pressure except if there was a difference of 10mmHg or more, in which case the highest systolic blood pressure was used. Ratios were calculated separately for the left and right sides and the minimum value was used for analyses. Age, Sex, race, socioeconomic factors, BMI, hypertension, HDL-C, LDL-C, smoking, diabetes, diet and physical activities. The following variables were explored for effect modification: age, sex, lipid levels, site, education, race/ethnicity, diabetes, BMI, smoking. No effect modification by these variables was found. PM2.5: particulate matter with aerodynamic diameter ≤ 2.5 μm; CIMT: carotid intima media thickness; BMI: body mass index; LDL-C: low-density lipoprotein cholesterol; HDL-C: high-density lipoprotein cholesterol; hsCRP: high-sensitivity C-reactive protein; EPA: Environmental Protection Agency; CVD: cardiovascular disease; CAC: coronary aortic calcification; CT: computed tomography; EURAD: European Air Pollution Dispersion; ABI: ankle-brachial index Conflict of interest The authors declare that there is no conflict of interest. 1 Laden F Schwartz J Speizer F Reduction in fine particulate air pollution and mortality: Extended follow-up of the Harvard Six Cities study Am J Respir Crit Care Med 2006 173 667 16424447 2 Kloog I Melly SJ Ridgway WL Using new satellite based exposure methods to study the association between pregnancy pm2. 5 exposure, premature birth and birth weight in Massachusetts Environ Health 2012 11 1 8 22236490 3 Nelin TD Joseph AM Gorr MW Direct and indirect effects of particulate matter on the cardiovascular system Toxicol Lett 2012 208 293 299 22119171 4 Suwa T Hogg JC Quinlan KB Particulate air pollution induces progression of atherosclerosis J Am Coll Cardiol 2002 39 935 942 11897432 5 Brook RD Rajagopalan S Pope CA Particulate matter air pollution and cardiovascular disease an update to the scientific statement from the American Heart Association Circulation 2010 121 2331 2378 20458016 6 Newby DE Mannucci PM Tell GS on behalf of ESC Working Group on Thrombosis, European Association for Cardiovascular Prevention and Rehabilitation and ESC Heart Failure Association Expert position paper on air pollution and cardiovascular disease Eur Heart J 2015 36 2 83b 93b 25492627 7 Kubesch N De Nazelle A Guerra S Arterial blood pressure responses to short-term exposure to low and high traffic-related air pollution with and without moderate physical activity Eur J Prev Cardiol 2015 22 548 557 25326542 8 Bilenko N van Rossem L Brunekreef B Trafficrelated air pollution and noise and children’s blood pressure: Results from the PIAMA birth cohort study Eur J Prev Cardiol 2015 22 4 12 24047569 9 Pope CA Burnett RT Thurston GD Cardiovascular mortality and long-term exposure to particulate air pollution epidemiological evidence of general pathophysiological pathways of disease Circulation 2004 109 71 77 14676145 10 Schwartz J Morris R Air pollution and hospital admissions for cardiovascular disease in Detroit, Michigan Am J Epidemiol 1995 142 23 35 7785670 11 Soares SR Carvalho-Oliveira R Ramos-Sanchez E Air pollution and antibodies against modified lipoproteins are associated with atherosclerosis and vascular remodeling in hyperlipemic mice Atherosclerosis 2009 207 368 373 19486979 12 Adar SD Sheppard L Vedal S Fine particulate air pollution and the progression of carotid intima-medial thickness: A prospective cohort study from the multiethnic study of atherosclerosis and air pollution PLoS Med 2013 10 e1001430 23637576 13 Bauer M Moebus S Möhlenkamp S Study Investigative Group HNR Urban particulate matter air pollution is associated with subclinical atherosclerosis: Results from the HNR (Heinz Nixdorf Recall) study J Am Coll Cardiol 2010 56 1803 1808 21087707 14 Breton CV Wang X Mack WJ Childhood air pollutant exposure and carotid artery intima-media thickness in young adults Circulation 2012 126 1614 1620 22896588 15 Stroup DF Berlin JA Morton SC Meta-analysis of observational studies in epidemiology: A proposal for reporting. Meta-analysis Of Observational Studies in Epidemiology (MOOSE) group JAMA 2000 283 2008 2012 10789670 16 DerSimonian R Laird N Meta-analysis in clinical trials Control Clin Trials 1986 7 177 188 3802833 17 Deeks JJ Higgins JPT Altman DG Higgins JPT Green S Chapter 9: Analysing data and undertaking meta-analyses Cochrane Handbook for Systematic Reviews of Interventions Version 5.1.0 (updated March 2011) The Cochrane Collaboration www.cochrane-handbook.org 2011 accessed 1 October 2014 18 Lau J Ioannidis JP Terrin N The case of the mis-leading funnel plot BMJ 2006 333 597 600 16974018 19 Sun M Kaufman JD Kim SY Particulate matter components and subclinical atherosclerosis: Common approaches to estimating exposure in a Multi-Ethnic Study of Atherosclerosis cross-sectional study Environ Health 2013 12 39 23641873 20 Diez Roux AV Auchincloss AH Franklin TG Long-term exposure to ambient particulate matter and prevalence of subclinical atherosclerosis in the Multi- Ethnic Study of Atherosclerosis Am J Epidemiol 2008 167 667 675 18227099 21 Künzli N Jerrett M Mack W Ambient air pollution and atherosclerosis in Los Angeles Environ Health Perspect 2005 113 201 206 15687058 22 Lenters V Uiterwaal CS Beelen R Long-term exposure to air pollution and vascular damage in young adults Epidemiology 2010 21 512 520 20407379 23 Künzli N Jerrett M Garcia-Esteban R Ambient air pollution and the progression of atherosclerosis in adults PLoS One 2010 5 e9096 20161713 24 Allen RW Criqui MH Diez Roux AV Fine particulate matter air pollution, proximity to traffic, and aortic atherosclerosis Epidemiology 2009 20 254 264 19129730 25 Kälsch H Hennig F Moebus S Are air pollution and traffic noise independently associated with atherosclerosis: The Heinz Nixdorf Recall Study; Heinz Nixdorf Recall Study Investigative Group Eur Heart J 2014 35 853 860 24194529 26 Hoffmann B Moebus S Möhlenkamp S Heinz Nixdorf Recall Study Investigative Group Residential exposure to traffic is associated with coronary atherosclerosis Circulation 2007 116 489 496 17638927 27 Hoffmann B Moebus S Kröger K Residential exposure to urban air pollution, ankle-brachial index, and peripheral arterial disease Epidemiology 2009 20 280 288 19194299 28 Stein JH Korcarz CE Hurst RT American Society of Echocardiography Carotid Intima-Media Thickness Task Force Use of carotid ultrasound to identify subclinical vascular disease and evaluate cardiovascular disease risk: A consensus statement from the American Society of Echocardiography Carotid Intima-Media Thickness Task Force Endorsed by the Society for Vascular Medicine J Am Soc Echocardiogr 2008 21 93 111 18261694 29 United States Environmental Protection Agency. Office of Air and Radiation, Office of Air Quality Planning and Standards, Fact Sheet Revised air quality standards for particle pollution and updates to the Air Quality Index (AQI) US EPA North Carolina, USA 30 Wang X Bi X Sheng G Hospital indoor PM10/ PM2. 5 and associated trace elements in Guangzhou, China Sci Total Environ 2006 366 124 135 16197981 31 Den Ruijter HM Peters SA Anderson TJ Common carotid intima-media thickness measurements in cardiovascular risk prediction: A meta-analysis JAMA 2012 308 796 803 22910757 32 Chen T Jia G Wei Y Beijing ambient particle exposure accelerates atherosclerosis in ApoE knockout mice Toxicol Lett 2013 223 146 153 24045146 33 Li R Navab M Pakbin P Ambient ultrafine particles alter lipid metabolism and HDL anti-oxidant capacity in LDRL-null mice J Lipid Res 2013 54 1608 1615 23564731 34 Agatston AS Janowitz WR Hildner FJ Quantification of coronary artery calcium using ultrafast computed tomography J Am Coll Cardiol 1990 15 827 832 2407762

PMC005xxxxxx/PMC5133699.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 0372762 3389 Dev Biol Dev. Biol. Developmental biology 0012-1606 1095-564X 20727874 5133699 10.1016/j.ydbio.2010.08.013 NIHMS236869 Article Fetal and post-natal lung defects reveal a novel and required role for Fgf8 in lung development Yu Shibin a* Poe Bryan a* Schwarz Margaret b Elliot Sarah c Albertine Kurt H. c Fenton Stephen d Garg Vidu b Moon Anne M. acef a Department of Pediatrics, University of Utah, Salt Lake City, UT 84112 b Department of Pediatrics, University of Texas Southwestern Medical Center at Dallas, Dallas, TX 75390 c Department of Neurobiology and Anatomy, University of Utah, Salt Lake City, UT 84112 d Department of Surgery, University of Utah, Salt Lake City, UT 84112 e Department of Human Genetics, University of Utah, Salt Lake City, UT 84112 f Program in Molecular Medicine, University of Utah, Salt Lake City, UT 84112 Corresponding author: Anne Moon MD, PhD, Associate Professor of Pediatrics, Neurobiology and Anatomy, and Human Genetics, University of Utah, 15 North, 2030 East, Room 4160B, Salt Lake City, UT USA 84112, phone: 1-801-585-0711 fax: 1-801-585-0701, anne.moon@genetics.utah.edu * co-first authors 1 11 2016 19 8 2010 1 11 2010 02 12 2016 347 1 92108 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. The fibroblast growth factor, FGF8, has been shown to be essential for vertebrate cardiovascular, craniofacial, brain and limb development. Here we report that Fgf8 function is required for normal progression through the late fetal stages of lung development that culminate in alveolar formation. Budding, lobation and branching morphogenesis are unaffected in early stage Fgf8 hypomorphic and conditional mutant lungs. Excess proliferation during fetal development disrupts distal airspace formation, mesenchymal and vascular remodeling, and Type I epithelial cell differentiation resulting in postnatal respiratory failure and death. Our findings reveal a previously unknown, critical role for Fgf8 function in fetal lung development and suggest that this factor may also contribute to postnatal alveologenesis. Given the high number of premature infants with alveolar dysgenesis and lung dysplasia, and the accumulating evidence that short-term benefits of available therapies may be outweighed by long term detrimental effects on postnatal alveologenesis, the therapeutic implications of identifying a factor or pathway that can be targeted to stimulate normal alveolar development are profound. lung development FGF alveologenesis pulmonary vascular development Type I alveolar cell Type II alveolar cell mouse model proliferation differentiation INTRODUCTION We previously reported that 100% of Fgf8 deficient newborn mice die in the first postnatal day with cyanosis and respiratory failure (Frank et al., 2002). Since only 40% of these mutants have predictably lethal cardiovascular defects, our observations led us to hypothesize that pulmonary dysfunction due to abnormal prenatal lung development may cause respiratory failure and death in these animals, and here we report our studies confirming this hypothesis. Correct progression through embryonic, pseudoglandular, canalicular, and saccular stages of prenatal lung development culminates in alveolar formation after birth, and these stages of lung development are conserved in many vertebrates (Alcorn et al., 1981; Davies et al., 1988; Docimo et al., 1991). During the embryonic stage, lung development is initiated when the primitive foregut endoderm (epithelium) is induced to invade the surrounding splanchnic mesoderm (mesenchyme) and form the trachea and the bronchi at embryonic day (E) 9.5 in the mouse. The proximal airways form by branching morphogenesis during the embryonic (~E10.5–14.5) and pseudoglandular stages (E14–16.5 in mouse) (Metzger et al., 2008). The canalicular stage (E16–17.5) is characterized by creation of the pulmonary acinus (air sac), and the multiplication and canalization of capillaries. During the saccular stage (E17–19), peripheral airways form terminal clusters of potential air spaces, and the capillary network and septal cores are remodeled in preparation for gas exchange at birth. In humans, the alveolar phase of lung development normally begins in utero during the final weeks of development and continues postnatally (Boyden, 1974; Davies and Reid, 1970; Emery and Wilcock, 1966; Langston et al., 1984), whereas in the mouse, alveolarization is predominantly a post-natal event. Alveolar formation is manifest structurally by the protrusion of partitions from saccular walls (Burri, 1997; Schittny et al., 1998) and subsequent elongation of those partitions into secondary septa with accompanying capillaries. Although septal thinning has been attributed to prenatal mesenchymal apoptosis in rabbits and rats (Bruce et al., 1999), attenuation of the septal mesenchymal core in mice appears to be primarily due to elongation and remodeling of the septa prenatally and via apoptosis postnatally, because very little apoptosis is present in the mouse lung at fetal stages (Bird et al., 2007; Muglia et al., 1999), our unpublished observations. The mechanisms that regulate initial budding, branching morphogenesis and proximal airway development have been intensely studied. However, late initiation and prolonged duration of alveolar formation pose significant challenges to identifying the genetic and molecular mechanisms that regulate the late stage lung development (Boyden, 1974; Burri, 1997; Davies and Reid, 1970; Dunnill, 1962; Emery and Wilcock, 1966; Langston et al., 1984). Mouse knockout models have identified many molecules that regulate embryonic lung development (Colvin et al., 2001; Lindahl et al., 1997), but this approach often profoundly affects the earliest stages, and pulmonary development ceases and/or death occurs prior to the initiation of the saccular and alveolar stages. FGF signaling proteins regulate multiple morphogenetic processes during vertebrate organogenesis (Szebenyi and Fallon, 1999). FGF receptors (FgfRs) are transmembrane tyrosine kinases. Ligand binding induces receptor dimerization, autophosphorylation, intracellular signaling cascades and ultimately, altered gene expression and cell behavior. Several Fgfs and all known FgfRs (FgfR1-5) are expressed in the lung (Bellusci et al., 1997; Colvin et al., 1999; Hu et al., 1998; Powell et al., 1998; Sannes et al., 1992). The temporospatial expression of FGF ligands and receptors varies with developmental stage. Fgf10 and Fgf7 are expressed in lung mesenchyme while Fgf1, Fgf9, and Fgf18 are expressed in both mesenchyme and epithelia. Fgf8 transcripts have been detected in embryonic mouse and adult rat lung (Lin et al., 2002; Schmitt et al., 1996); however expression of Fgf8 in fetal and postnatal mouse lung has not previously been described. Germline ablation of some Fgf ligands and their receptors has established a central role of reciprocal epithelial/mesenchymal FGF signaling during the embryonic and pseudoglandular stages (Arman et al., 1999; Bellusci et al., 1997; Cardoso, 2001; Warburton et al., 1999), but no FGFs have yet been discovered to function during the fetal stages. Targeted deletion of either Fgf10 or its receptor, FgfR2b, inhibits budding of the lung from the trachea (De Moerlooze et al., 2000; Min, 1998; Sekine et al., 1999) whereas later embryonic ablation disrupts branching morphogenesis (Abler et al., 2009). FGF9 stimulates mesenchymal proliferation and propagation of branching in part via stimulating Fgf10 (Colvin et al., 2001). FGF7 stimulates proliferation and differentiation of distal lung in culture (Cardoso, 2001), but Fgf7-null mice have normal lung morphology (Guo et al., 1996). Fgf18 null mutants have lung hypoplasia and decreased fibroblast proliferation during the canalicular stage but normal distal epithelial differentiation (Usui et al., 2004). Although data are somewhat conflicting (Celli et al., 1998; Hokuto et al., 2003; Weinstein et al., 1998; Yi et al., 2006), the overall body of literature suggests that the fetal saccular to alveolar transition and post-natal lung development are FGF-dependent processes, however, no FGF ligands have previously been tested by conditional loss-of-function sufficiently late in lung development to establish which, if any, are required for saccular or alveolar development. Here we report that Fgf8 has unique required functions to support fetal lung development and may play a role in post-natal alveologenesis. Our data indicate that FGF8 is required to regulate proliferation in the fetal stages; in the absence of adequate Fgf8 function, there is pronounced excess proliferation in both the epithelial and mesenchymal compartments from E16.5–18.5 and subsequent disruption of epithelial differentiation and abnormal septal and vascular remodeling. MATERIALS AND METHODS Mice The Fgf8 null (Moon and Capecchi, 2000), hypomorphic (Frank et al., 2002) and conditional (Park et al., 2006) alleles are as previously described and are maintained in 50% each C57Bl6 and SV129 background. The Isl1Cre allele (Park et al., 2006) is in a 50% C57Bl6 and 25% each BlSw and Sv129 background, and the MesP1Cre allele (Saga et al., 1999) is 50% C57Bl6 and 25% each ICR and SV129. Lung wet/dry weight ratios Lung wet/dry weight ratios were determined by use of the microwave drying technique (Peterson et al., 1982). Briefly, the lungs were freshly excised and the wet weight determined.. The tissues were dried in a microwave oven at 200W for 1 hour and then weighed; they were returned to the oven at 5 minute intervals and re-weighed to determine when dessication was complete and the final measurement used as the dry weight. Preparation of RNA and cDNA for microarray and microarray data analyses E18.5 lung lobes were dissected away from large airways and vessels and stored in RLT buffer (Qiagen) at −80°C. Three specimens of each genotype (Fgf8+/+, controls; Fgf8H/−, hypomorphic mutants) were pooled to generate each sample. Total RNA was extracted from four samples (RNeasy Micro Kit, Qiagen). This experiment was run in quadruplicate on Agilent mouse whole-genome expression arrays. Agilent two-color LRILAK labeling, the Agilent two-color GE hybridization/wash protocol, and the Agilent 5-micron XDR scanning protocols were carried out by the University of Utah Microarray Core Facility. The array image data were quantitated using Agilent Feature Extraction software (version 9.5.1.1). Subtle intensity-dependent bias was corrected with LOWESS normalization, with no background subtraction. The raw and normalized data sets have been submitted to GEO (accession number: gse22893). Spots with intensity below background were removed prior to statistical analysis. Statistical analysis of normalized log-transformed data was performed in GeneSifter (www.genesifter.net). Differentially expressed transcripts were defined (adjusted for multiple testing using the Benjamini and Hochberg method) as P<0.05. Data were hierarchically clustered with Spotfire (TIBCO) and heat maps for selected genes were generated. For qRT-PCR validation, total RNA from each of the initial four pools (100 ng) was reverse transcribed to cDNA using the SuperScript III First-Strand Synthesis System (Invitrogen). Quantitative RT-PCR was performed with iQ SYBR Green Supermix on the iCycler system (Bio-Rad) and normalization was to hprt, gapdh and βactin. Primer sequences will be provided on request. Real-time quantitative RT-PCR of Fgf8 and Fgf10 transcripts in wild type lungs Total RNA was isolated from 5 to 12 pooled lung homogenates from E11.5 to P5 mice using TriZol reagent (Invitrogen, 15596-026). After DNase treatment (Qiagen, 79254), RNA was further purified using the RNAeasy MinElute Cleanup kit (Qiagen, 74204). cDNA templates for PCR amplification were synthesized from 2 μg of total RNA using a SuperScript® III First-Strand Synthesis Kit (Invitrogen, 18080-400) in the presence of Oligo dT. qPCR and quantitation were performed in the Bio-Rad iCycler IQ™ Multicolor Real-Time PCR Detection System using 25 μl IQ™ SYBR® Green Supermix (Bio-Rad, 170-8882), 250 nM of each primer and 1 μl of cDNA. Each sample was run in triplicate. The amplification ramp included initial hold of 5 minutes at 94 °C, followed by a three step cycle consisting of denaturation at 94 °C (30 seconds), annealing at 57 °C (30 seconds) and extension at 72 °C (30 seconds) and the amplification fluorescence was read at the end of the cycle. The quantitative expression value for each specific transcript was normalized to that of the GAPDH. Ratios of Fgf8 and Fgf10 transcripts at each time point relative to that at E14.5 are shown from E11.5 to P5. Isolation of lung mesenchymal and epithelial cells E15.5 embryonic lungs were isolated and the tissue minced in ice cold PBS. A single cell suspension was obtained by 0.5% collagenase (Sigma, C0130) digestion at 37 °C for 45 minutes and then incubated with RBC lysis buffer (Sigma, R7757) at RT for 15 minutes to remove red blood cells. The cells were filtered through a 100-μm cell strainer and resuspended in DMEM with 10% fetal calf serum. Mesenchymal cells and epithelial cells were isolated based on their differential adherence to culture flask plastic over time. Adherent cells after the first 45 minutes in culture were mesenchymal cells; the unattached cells at this 45 minute point were isolated and cultured for an additional 1 hour yielding a second batch of attached cells which is a mixture of epithelial cells and fibroblasts; these were discarded. The persistently unattached cells are epithelial cells, which were then centrifuged at low speed (120g × 3min) for isolation. Isolated epithelial and mesenchymal cells were fixed in 4% PFA for 20 min on ice and processed immediately. 105 cells in 250 μl PBS buffer were added to a cytospin chamber and the specimens were cytocentrifuged at 750 rpm for 3 minutes, leaving a button of concentrated cells on the slide. For qRT-PCR assay of these isolated populations, cell total RNA sample was prepared using RNAeasy micro Kit (Qiagen, 74004) with the protocol provided by the manufacture. cDNA synthesis and quantitative PCR procedures were the same as described above. FGF8 stimulation of lung mesenchymal and epithelial cells and analysis of BrdU incorporation E15.5 mesenchymal and epithelial cells were isolated as above and 4 X105 (epithelial or mesenchymal) cells were plated per well containing a sterilized 12mm round coverslip in a 24 well plate. Each experimental treatment was repeated in triplicate. Cells were allowed to attach either in the presence or absence of serum, with the simultaneous addition of 20 ng/ml FGF8 with 100ng/ml heparin, or heparin alone. The cells were incubated for 24 or 48 hours and BrdU was added to a final concentration of 10uM, cells incubated an additional 12 hours and then rinsed in 1XPBS and fixed in −20 °C 100% methanol for 30 minutes. Coverslips were removed from the wells, washed in PBS, immersed in 2N HCL for 20 minutes at room temperature, washed in 0.1M sodioum borate (ph 8.5) for 2 minutes at room temperature. After washing for 5 minutes in PBS, they were washed 3 times in a blocking solution containing PBS+ 0.2% triton x100 + 3% BSA for 10 minutes each and then incubated with BrdU primary antibody (Abcam, ab6326) in block for 2 hours. Antibody was removed by 3 X10 minute washes in block and then Alexa 594 goat anti-mouse secondary was added for I hour at room temperature. After final washing in block X3, the coverslips were mounted in Vectashield with DAPI. 3 randomly selected 200X fields were counted on each coverslip (3 coverslips per experimental treatment, 9 random fields counted in total per experimental treatment group): the total nuclei in living cells per field was counted, as was the subset with BrdU + nuclei. The percent positive nuclei is expressed as (BrdU nuclei/total nuclei) × 100. Quantitation of overall cell adherence, growth and survival (Figure 5C, D) was performed based on initial plating of 4 × 105 cells and each random 200X field counted represents 1/67th of the total surface area of the 12 mm coverslip. Immunofluorescence For isolated mesenchymal and epithelial cells, the cells were permeabilized with 0.1% Triton X100 in PBS for 15 min and then blocked with 5% bovine serum in PBS containing 0.1% Triton X100. For sectioned lung, the lung was removed and fixed routinely, embedded in OCT, and sectioned at 10 μm. Immunohistochemical analysis was performed using the following antibodies: anti-FoxA2 (1:1000; Abcam, ab40874), anti-PECAM (1:200; BD Pharmingen, 550274), Anti-Nkx2.1 (1:100, GeneTex, 8G7G3/1), anti-SMA (1:100, Sigma, A2547), anti- Aquaporin-5 (1:100; CalBiochem, 178615), and anti-SPC (1:200; Upstate 07-647). Trypsin antigen retrieval was used to unmask the antigen for PECAM detection. Blocking was performed using 5% goat or donkey serum in PBST. Antibodies were incubated O/N at 4 C. Secondary antibodies conjugated with Alexa 488, Alexa 594 or Cy3 (1:1000; Molecular Probes) were incubated for 1 hour at RT. Slides were mounted using VECTASHIELD mounting medium with DAPI (Vector Laboratories, H-1200). Images were obtained using a Zeiss Axio microscope and Zeiss AxioVision software. The anti-SPC signal was imaged on an IX81 Olympus microscope and images captured with a Hamamatsu Orca digital camera with a DSU spinning confocal unit using Slidebook software. BrdU immunohistochemical staining of mouse lung sections Mice received intraperitoneal injection of 50 mg/gm body weight 5-bromo-2′-deoxyuridine (Invitrogen, 00-0103) 2 hours prior to sacrifice. The lung was removed and processed routinely, embedded in OCT Compound, and sectioned at 10 μm. Microwave antigen retrieval was performed for 20 min in Antigen Unmasking Solution (Vector Laboratories, H-3300). Endogenous peroxidases were quenched by incubation at room temperature in 1% H2O2 for 20 min followed by rinsing in PBST (PBS with 0.1% Tween 20). Blocking was performed using 5% donkey serum in PBST. Rat anti-BrdU primary antibody (Abcam, ab6326) in 5% donkey serum was added to the tissue sections at a 1:1000 dilution. Sections were incubated overnight at 4 C and then rinsed in PBST. Biotinylated specific staining was detected with biotinylated donkey anti-rat secondary antibody (Jackson ImmunoResearch, 712-065-153) horseradish peroxidase complexes using VECTASTAIN Elite ABC kits (Vector Laboratories, PK-6100). Antigen-antibody complexes were visualized using DAB Substrate (Vector Laboratories, SK-4105). The labeling index was expressed as the number of BrdU-positive ratio and counted by Image-Pro Plus v. 6.0software (Media Cybernetics, Inc). Results were expressed as means ± SE of at least 3 mice per group and p values were calculated using the Student’s two tailed t-test. β-galactosidase Staining MesP1 and Islet1 lineage analysis was performed with a standard breeding strategy in which the females contained the conditional Rosa26lacz reporter allele (Soriano, 1999) and the males carried the MesP1Cre or Islet1Cre alleles. E11.5 embryos were stained for β-galactosidase activity using a standard protocol. All embryos from a litter were fixed in 0.2% glutaraldehyde in PBS, 0.02% NP40 (PBN) on ice for 15 min, washed in PBN, and placed in X-gal staining solution (5 mM potassium ferricyanide, 5 mM potassium ferrocyanide, 2 mM MgCl2, 1 mg/ml X-gal, in PBS, pH 7.3). Staining was carried out overnight at room temperature while rocking. After staining, embryos were washed in PBN and post-fixed in 4% paraformaldehyde overnight at 4°C. Embryos were then rinsed in PBS and cleared in glycerol for photographs, embedded in paraffin, sectioned transversely, and counterstained with nuclear Fast Red. RESULTS Decreased Fgf8 function disrupts lung development Fgf8 expression is required in multiple temporospatial expression domains during vertebrate development, and null embryos die early in embryogenesis (Meyers et al., 1998b; Moon and Capecchi, 2000; Sun et al., 1999). The phenotypes of Fgf8 hypomorphs (produce less than 50% of the normal amount of Fgf8 mRNA) revealed the essential functions of this protein in cardiovascular, nervous system, limb and pharyngeal development and left-right patterning (Abu-Issa et al., 2002; Frank et al., 2002; Meyers et al., 1998b; Moon, 2006). Subsequent conditional mutagenesis experiments have been instrumental in further dissecting the requirements for Fgf8 activity at different times and locations (Ilagan et al., 2006; Lewandoski et al., 2000; Macatee et al., 2003; Moon, 2006; Moon and Capecchi, 2000; Park et al., 2006; Trumpp et al., 1999). We previously reported that 100% of Fgf8 deficient hypomorphs (Fgf8H/−, animals bearing a hypomorphic allele and a null allele of Fgf8) die in the first postnatal day with cyanosis; however, only 40% of these mutants have predictably lethal cardiovascular defects (Frank et al., 2002). We observed these animals after birth and noted that in spite of significant respiratory effort, they uniformly became cyanotic and died within hours of birth. Thus, we questioned whether pulmonary defects contribute to their postnatal death and so collected lungs from Fgf8H/− hypomorphs, and their wild type, null heterozygote (Fgf8+/−) or hypomorph heterozygote (Fgf8H/+) littermates at E18.5 and immediately after birth (P0.5). On gross pathology at E18.5, the lungs of the Fgf8H/− mutants are much larger relative to total body size than controls (Fig. 1A–D, fetuses are photographed side-by-side; longitudinal black bars superimposed over the lungs are the same length, and the lungs are photographed side-by-side). This correlates with a significantly increased lung to body weight ratio in the mutants (6.9 +/− 0.7 % in Fgf8 hypomorphs versus 4.9 +− 0.8 % in controls, p=0.04, N=4) that is not attributable to edema, because mutants also have increased dry to wet lung weight ratio (12.9 +− 1.9% in Fgf8 mutants versus 9.1 +− 0.3% wild type, p=0.02). Thus the mutant lungs are hyperplastic. In fact, at P0.5 when all controls have generated large air spaces, thin septa (Fig. 1E, E′ control red arrows), and a well-organized capillary network, the mutants have dramatically hypercellular septa that appear not to have progressed beyond the early saccular stage; the capillary network (as visualized by the bright pink red blood cells in Figure 1 F′) is largely remote from the airspaces embedded in thick septal walls (Fig. 1F, F′, mutant red arrows). Thickness of the septal walls in mutants is 150% that of controls (4.9 ± 0.7 μm controls versus 7.6 ± 1.5 μm mutants, p<0.05, N=6). Fgf8 is expressed at low levels in the developing mouse lung Fgf8 mRNA transcripts have previously been detected in E12.5 mouse and adult rat lung (Lin et al., 2002; Schmitt et al., 1996); however the expression of Fgf8 at other stages has not been characterized. We assayed Fgf8 transcripts in whole lung lobes from wild type mice (all proximal airways and vessels were removed) by quantitative RT-PCR; Figure 2A shows the expression level at each stage relative to the level at E14.5 (data shown are normalized to gapdh). We detected expression at all stages of prenatal lung development, with a marked increase at postnatal day 5 (P5). Normalization to βactin also revealed a marked increase in the relative level of Fgf8 expression postnatally (not shown). No Fgf8 transcripts were detected in RNA prepared from Fgf8 null embryo negative controls. The absolute cycle threshold (Ct) values for Fgf8 in the lung ranged from 29–33, indicating a low number of transcripts and consistent with our inability to detect the mRNA signal above background by in situ hybridization with digoxigenin- or 35S-labeled antisense riboprobes (not shown). In contrast, Fgf10 transcripts were detected with absolute Ct values ranging from 15–19 (Fig. 2B) and levels increased steadily over the course of prenatal lung development, as has been reported (Bellusci et al., 1997). To determine whether Fgf8 is expressed in both the mesenchymal and epithelial cellular compartments of the developing lung, we dissociated E15.5 lung tissue and isolated epithelial from mesenchymal cells using a protocol adapted from (Schuger et al., 1993), based on the differential adherence of dissociated cells types over time to standard tissue culture flasks. This method provides excellent separation of epithelial and mesenchymal cells as demonstrated by analysis of Cytospin preparations of the two populations with mesenchymal (PECAM, SMA; Fig. 3A, A′, B, B′) and epithelial (Nkx2.1, FoxA2; C, C′, D, D′) protein markers. When we assayed these isolated populations for transcript levels using qRT-PCR, and compared the ratio of transcripts detected in the mesenchymal versus epithelial compartments, we found that Fgf8 transcripts are relatively enriched epithelial cells (6.5 fold), along with transcripts for Foxa2 and Shh, while Pecam and Fgf10 transcripts were enriched in mesenchymal cells, as expected, (Fig. 3E). The source of Fgf8 transcripts in lung is not red blood cells or another circulating cell type, as these experiments were performed after red blood cell lysis, and we have specifically found that Fgf8 transcripts are undetectable in e15.5–17.5 fetal blood (data not shown). All four Fgf receptors are expressed in the lung (Powell et al., 1998) and we quantitated the relative levels of FGF receptor transcripts in these isolated cell populations (Supplemental Table 1). We obtained the expected ratios for FgfR2b and FgfR2c (enriched in epithelia and mesenchyme, respectively (Arman et al., 1999)) and detected expression of FgfR1b and c, FgfR3, and FgfR 4. While these data are not absolute transcript numbers, we have optimized efficiency of all primer sets and thus the dCt values (Ct of the gene of interest less that of the housekeeping gene) can be compared for relative abundance of the different FgfR amplicons. These data indicate that FgfR1 isoform b and c transcripts are present at comparable levels to that of FgfR2 b and c in the epithelium and mesenchyme, respectively. We were unable to generate reliable isoform specific primers for FgfR3; our data confirm that FgfR3 and FgfR4 transcripts are expressed in both the epithelial and mesenchymal compartments. Fgf8 deficiency causes excess proliferation and disrupts the fetal phase of lung development Given the markedly thickened septal walls observed in the hypomorphs, we tested whether excess proliferation was a contributing mechanism. We labeled S-phase cells with an intraperitoneal pulse of BrdU administered to the mother 2 hours prior to fetal harvest and performed anti-BrdU immunohistochemistry at multiple stages of lung development. While the percentage of labeled cells was not different between hypomorphs and controls at the late embryonic and early fetal stages (E13.5–15.5; Fig. 4 A–C), by E16.5, proliferation in the mutants both visually and statistically exceeded that in controls by more than 2 fold in both the distal epithelium and mesenchyme (Fig. 4 A, B, C). Furthermore, although the total percentage of proliferating epithelial and mesenchymal cells decreases in the hypomorphs at E17.5 and E18.5 (as it does in controls), it remains abnormally high (Fig. 4 A, D). We assayed for apoptosis using the TUNEL assay in these specimens and found only rare TUNEL+ cells in both mutants and controls (data not shown). FGFs generally stimulate cell proliferation and FGF9, 10 and 18 all do so in the lung in vivo (Bellusci et al., 1997; Colvin et al., 2001; Usui et al., 2004). Although the cell proliferation and and survival responses to FGF8 are dosage and context dependent (Storm et al., 2006; Storm et al., 2003), the hyperproliferative lung phenotype in FGF8 deficient mutants is nonetheless surprising. To better understand whether excess proliferation of both mesenchymal and epithelial cels in the mutants represents a direct or secondary consequence of FGF8 deficiency, we assayed whether FGF8 stimulates or or represses proliferation of lung epithelial and mesenchymal cells at E15.5. Isolated cells were cultured for 24 and 48 hours in the presence or absence of serum and recombinant FGF8b and then pulsed with BrdU and immunostained. As shown in Figure 5A, by 24 hours FGF8 had significantly increased the percentage of BrdU positive mesenchymal cells both in the absence (columns a and b, 2.3 fold increase, p<0.001) and presence (columns c and d, 1.2 fold increase, p < 0.001) of serum, and at 48 hours in the absence of serum (columns e and f, 2.7 fold, p<0.001). The cumulative effects of adherence to substrate, survival and growth of the cultured mesenchymal cells on total cell number at the time of counting is shown in Figure 5C. At 24 and 48 hours in the absence of serum, approximately 120% of the number of mesenchymal cells initially plated were present (compare columns a and e) and this was not significantly affected by FGF8 (columns b and f); the neglible increase in mesenchymal cell number at 48 hours in the serum-free groups is likely attributable to their already being nearly confluent at 24 hours. Significantly fewer mesenchymal cells were present at 24 hours in the serum-treated groups (columns a versus c, 1.4 fold decrease, p < 0.0001; columns b versus d, 1.2 fold decrease, p=0.02); the reason for this effect is unclear, but may be due to decreased attachment to the culture substrate. The lower cell number/lower confluence at 24 hours in the serum-treated cells appears to have permitted cell proliferation over the next 24 hours, as evident by the significantly increased number of cells present at 48 hours (columns c versus g, 1.5 fold increase, p<0.0001; columns d versus h, 1.5 fold increase p < 0.0001). The results obtained with cultured lung epithelial cells were dramatically different. First, the number of cells surviving as a percent of the number plated was much lower than for mesenchymal cells in all groups (Figure 5D). Unlike mesenchymal cells, the epithelial cells were dependent on serum for adherence, growth and survival because at 24 hours in the absence of serum, there were fewer epithelial cells per well than in serum-treated wells (Figure 5D, columns a versus c, 5.2 fold increase with serum, p <0.0001; columns b versus d, 3.8 fold increase, p < 0.001; columns f versus h, 2.7 fold increase p <0.001). The serum-free cells did not increase BrdU incorporation in response to FGF8 (Figure 5B, columns a and b). Most remarkably, in the presence of serum, FGF8 treatment decreased the percentage of proliferating epithelial cells at 24 (Figure 5B, columns c versus d, 1.4 fold decrease, p< 0.001) and 48 hours (Figure 5B, columns g versus h, 1.5 fold decrease, p < 0.001) After 48 hours, the majority of epithelial cells cultured in the absence of serum had pyknotic nuclei, but we scored those with normal nuclear morphology and none had incorporated BrdU either in the presence or absence of FGF8 (Figure 5B, columns e, f). Although the presence of serum permitted better survival of epithelial cells to 48 hours (Figure 5D, columns f versus h), FGF8 nonetheless decreased BrdU incorporation in these cells (Figure 5B, columns g and h). In total, these data indicate that FGF8 at the dosage employed here has opposite effects on BrdU incorporation in epithelial versus mesenchymal cell types. In order to obtain a picture of the molecular pathways that were dysregulated in association with the excess proliferation seen in the fetal mutant lung, we performed a genome-wide expression analysis comparing E18.5 Fgf8 hypomorphic lungs to wild type controls. The direction and magnitude of the changes were very consistent between the 4 arrays, conferring high statistical power to our analysis. Furthermore, we analyzed expression of 46 genes of interest by qRT-PCR and in 43, the findings validated those of the microarray, a specificity of 93% (Supplemental Table 2). Microarrays are less sensitive than qRT-PCR and in this experiment the fold change reported by the microarray usually underestimated the magnitude of the changes in transcript levels (or did not detect differences) detected by qRT-PCR. In addition to examining absolute fold-changes, we used pathway-based analysis in order to detect modest, yet potentially functionally relevant changes in the context of multiple “hits” to a pathway (Curtis et al., 2005; Park et al., 2008). KEGG analysis revealed that the most overrepresented category of dysregulated genes was “Cell cycle” (60 upregulated, z-score 10.66, heat map not shown due to large number of genes); the changes were overwhelmingly in the direction of excess proliferation, consistent with our BrdU results. Overrepresented downregulated genes in the “Focal adhesion” pathway (53, z = 7.03) and “ECM-receptor interaction” pathway (31, z = 7.43) suggests that disrupted contact inhibition may contribute to the hyperproliferative state. Wnt ligands, receptors and canonical target gene transcripts were increased, while those of numerous Wnt inhibitors were decreased (Figure 6). Of the FGF ligands expressed in lung, Fgf8 (the level of transcripts in mutant lung assayed by qRT-PCR was 7.8 fold lower than in controls), Fgf10 (2.3 fold increased) and Fgf9 (5.7 fold increased) were dysregulated. Expression of FgfR3 and FgfR4 was decreased, while that of FgfRs 1 and 2 were unchanged (Supplementary Table 2). Other findings relevant to lung morphogenesis and the Fgf8 mutant fetal/postnatal phenotypes are shown in the heat maps in Figure 6. Results are expressed as a fold change in mRNA level (up-red, down-green) in mutants vs. wild type. Critical mediators of VEGF signaling and vascular maturation were decreased. The profile of many glucocorticoid target genes is similar to that reported in glucocorticoid receptor null mutants, which also have a hyperproliferation phenotype (Bird et al., 2007). ECM structural, remodeling, and signal modulating genes that are aberrantly expressed include several encoding metalloproteases, ctgf (connective tissue growth factor), igfbp3 (IGF binding protein 3), ibsp (Integrin binding sialoprotein) and loxl2 (lysl oxidase 2). In total, the gene expression changes and other phenotypic features support the hypothesis that decreased FGF8 signaling causes dysfunction of pathways that regulate proliferation, cell-cell and cell-ECM interactions, and differentiation in the distal fetal lung. The presence of Fgf8 transcripts at multiple developmental stages led us to hypothesize that the pulmonary phenotype in Fgf8 hypomorphs reflects a specific requirement for FGF8 in the developing lung; however, widespread defects in organogenesis and function, including variable cardiovascular defects in these hypomorphs could theoretically have secondary effects on lung development. Furthermore, the hypomorphic allele contains a neomycin phosphotransferase gene that can adversely affect function of both adjacent and distant genes (Frank et al., 2002; Olson et al., 1996) raising the possibility that the lung phenotype results from dysfunction of a different locus (note that this would predict that the phenotype would be present in hypomorph heterozygotes, which is not the case: Fig. 1A, C, E). To test our hypothesis and rule out possible confounding effects of other loci, we used a nonhypomorphic conditional allele of Fgf8 (Park et al., 2006) and two Cre drivers with different temporospatial activities to specifically test the effects of Fgf8 loss-of-function on lung development. Isl1Cre is active in pharyngeal endoderm, including pulmonary epithelial precursors, and in a subset of mesosderm from E8.5 (Park et al., 2006). MesP1Cre activity is strictly restricted to anterior mesoderm (Park et al., 2006; Saga et al., 1999), including pulmonary mesenchymal precursors. The blue β-galactosidase staining in panels E, F of Figure 7 indicates regions of Cre-mediated recombination of the ROSA26lacZ Cre reporter gene (Soriano, 1999). By E10.5, MesP1Cre has already recombined the reporter throughout the lung mesenchyme (Me, Fig. 7E) and Isl1Cre has recombined the reporter in the lung epithelia (Ep, Fig. 7F) but only a subset of the lung mesenchyme. Remarkably, E18.5, Fgf8;Isl1Cre mutants phenocopy the lung defects of Fgf8 hypomorphs (compare Fig. 7 C, C′ to D, D′): the mutants survive to birth, die with cyanosis postnatally and their lungs have hypercellular septa and a paucity of capillaries (Fig. 7 D, D′, and Fig. 9). Fgf8;Isl1Cre mutants exhibit the same excess proliferation from E16.5 seen in hypomorphs (Fig. 8). This is in spite of the fact that proliferation in mesenchyme of controls at E14-5–16.5 in this genetic background (Bl Swiss/SV129/C57Bl6) is slightly higher than that seen in the background in which the hypomorphs are maintained (SV129/C57Bl6) as shown in Figures 4A, B versus 8A, B. In notable contrast to the variable cardiovascular phenotypes of Fgf8 hypomorphs (Frank et al., 2002), 100% of Fgf8;Isl1Cre mutants have the cardiac outflow tract defect, persistent truncus arteriosus (Park et al., 2006). Ablation of Fgf8 only in mesodermal lung precursors using MesP1Cre (Fig. 7 B, B′, E) does not recapitulate the mutant phenotype even though we have shown that Fgf8;MesP1Cre mutants also have vascular and outflow tract defects (Park et al., 2006; Watanabe et al., 2009). The different lung phenotypes in Fgf8;Isl1Cre versus Fgf8;MesP1Cre mutants indicates that cardiovascular defects are not the cause of the lung phenotype in Fgf8;Isl1Cre or Fgf8 hypomorphs. We have also confirmed the independence of the lung and cardiac outflow tract phenotypes in Fgf8;Hoxa3IresCre mutants (Macatee et al., 2003); these mutants have pancellular CRE activity caudal to the second pharyngeal arch from ~E8.5 and although they only rarely have outflow tract defects (Macatee et al., 2003), they also have the lung phenotype (not shown). These data cumulatively indicate that the pulmonary and cardiovascular phenotypes reflect separate requirements for Fgf8 function in development of the lung versus the heart/great vessels, respectively. The conditional ablation experiments rule out secondary effects or dysfunction of other loci as a cause of pulmonary defects in Fgf8 hypomorphic and Fgf8;Isl1Cre conditional mutants, and support our hypothesis that FGF8 signaling is specifically required for lung development. The different tissue specificities of Isl1Cre and MesP1Cre suggest that Fgf8 function is required in endodermal precursors of lung epithelia or later in their derivatives to support late fetal lung development. Lung development occurs normally through the embryonic and pseudoglandular stages in Fgf8;Isl1Cre and hypomorphic mutants, but is disrupted in the canalicular stage by excess proliferation, failure of differentiation, and abnormal septal and vascular remodeling The Fgf8 hypomorphs revealed a great deal about the importance of proper Fgf8 function for normal lung development, but to avoid potential confounding pleiotropic effects on lung development in Fgf8 hypomorphs, and to focus our efforts using a defined loss-of-function mutant with invariant phenotypes, we performed further immunohistochemical exploration of the Fgf8 mutant phenotype in Fgf8;Isl1Cre conditional mutants. All experiments were repeated on a minimum of 3 mutants (from each class) and littermate control lung samples at each stage; control and mutant specimens were processed on the same slide and the results are highly reproducible. Expansion of air spaces and thinning of primary septa occur in the late fetal period concurrent with vascular remodeling. Efforts to dissect the regulation of late fetal and postnatal alveolar development have been hindered by the fact that lung development in most mouse models of disrupted lung development is already abnormal at embryonic stages. However, at the embryonic and early pseudoglandular stages (E13.5, E14.5), Fgf8;Isl1Cre and Fgf8 hypomorphic mutants mutant lungs have normal FoxA2, Nkx2.1, PECAM and smooth muscle actin (SMA) immunoreactivity (Supplemental Figures 1, 2, 3). Lung morphology and proliferation were also normal through the pseudoglandular and early canalicular stages in Fgf8;Isl1Cre and Fgf8 hypomorphic mutants (Figs. 4, 8 and Supplemental Figures 1–3). To better characterize the abnormalities present at the late fetal stages we examined numerous immunohistochemical markers in Fgf8;Isl1Cre mutants (Fig. 9) and Fgf8 hypomorphs (Fig. 10) compared to littermate controls. At E16.5, production of FoxA2 in the distal airway epithelia and of PECAM in endothelial cells appears comparable in Fgf8;Isl1Cre mutants and controls (Fig. 9 A, A′). However, at E17.5, control lungs have progressed to the saccular stage and the number of FoxA2+ cells present in the distal lung is decreasing (Fig. 9B, white arrowheads) as immature Type II cells differentiate into Type I cells. In contrast, the Fgf8;Isl1Cre mutant lung has persistent cuboidal FoxA2+ cells surrounding rare, small airspaces (Fig. 9B′, yellow arrowheads). The pattern of PECAM reactivity still appears normal in Fgf8;Isl1Cre mutants at this stage (Fig. 9 E, E′). By E18.5, the control lung has many large airspaces and an adjacent, well organized capillary network encompassing the thin septa in readiness for gas exchange at birth (Fig. 9 C, F). However, Fgf8;Isl1Cre mutant septa are thickened and hypercellular, and many cuboidal FoxA2+ cells still line the saccules (Fig. 9C′, yellow arrowheads); the vessels are intraseptal with decreased PECAM immunoreactivity (Fig. 9F′, yellow arrowheads). Consistent with the FoxA2 staining pattern, controls at this stage display very few cuboidal, Nkx2.1+ cells (Fig. 9G, blue nuclei, green Nkx2.1, white arrowheads) at the airspace interface however, many undifferentiated Nkx2.1+ cells line the small airspaces in the mutants. The fraction of total cells that are Nkx2.1+ in wild type lungs at E18.5 was 0.15 +/− 0.02, while that in Fgf8;Isl1Cre mutants was 0.29 +/− 0.05, p= 0.01. Increased immunoreactivity for surfactant protein C in the mutant epithelia (Fig. 9 J) supports the hypothesis that the excess Nkx2.1/FoxA2+ epithelial population represent Type II alveolar cells that have failed to differentiate into Type I cells. Fgf8 hypomorphs display the same immunohistochemical abnormalities at E18.5 (Fig. 10). Abnormalities in the septa, airway epithelium and capillary network are even more severe in Fgf8;Isl1Cre mutants just prior to birth (E19, Fig 9K–L′). Compared with controls, fewer PECAM+ capillaries are present in Fgf8;Isl1Cre mutant lungs and the network is hypoplastic and aberrantly located within the thickened septa (yellow, Fig. 9L, L′). The fraction of total cells that are PECAM+ in wild type lungs at E18.5 was 0.39 +/− 0.04, while that in Fgf8;Isl1Cre mutants was 0.19 +/− 0.01, p< 0.01. Aquaporin 5+ Type I cells are easily detected outlining the normal elongated plasma membrane at the air interface in controls (Fig. 9M, white staining and arrowheads). In Fgf8;Isl1Cre mutants (Fig. 9N), and Fgf8 hypomorphs (Fig. 10E′), minimal Aquaporin 5+ signal is present and the rare, faintly staining cells have failed to develop a normal squamous morphology; this is consistent with the failed differentiation of Type II cells. Thus, numerous crucial events that normally occur at the late saccular/early alveolar stages in preparation for gas exchange at birth fail to occur in Fgf8 mutants. DISCUSSION The onset and progression of the Fgf8 mutant lung phenotypes during fetal lung development indicates that sufficient FGF8 is produced in Fgf8;Isl1Cre mutants and hypomorphs to achieve a threshold required for the the embryonic and pseudoglandular stages of lung development, or that FGF8 is functionally redundant with other FGFs during these early periods. FGF8 and FGF18 are the most closely related ligands expressed in the lung epithelium, but the reported phenotype of Fgf18 null mutants is quite distinct from that of Fgf8;Isl1Cre and hypomorphic mutants: Fgf18 null mutants have hypoplastic lungs with decreased cell proliferation and preserved epithelial differentiation (Usui et al., 2004), while Fgf8 mutants have a hyperplastic phenotype with distal epithelial and mesenchymal overproliferation, and disrupted distal epithelial differentiation. Thus FGF8 and FGF18 are not functionally redundant at fetal stages, and the phenotypes of both classes of Fgf8 mutant indicate that no other FGF can compensate for the loss of FGF8 during fetal lung development. Furthermore, expression of FGF10 and FGF9 is increased in FGF8 hypomorphs at E18.5. These ligands stimulate epithelial and mesenchymal proliferation, respectively, in vivo; if increased gene expression also results in increased protein levels, these ligands may be contributing to the excess proliferation observed in mutants. The accepted paradigm is that FGFs signal in a paracrine fashion to effect epithelial-mesenchymal communication in the developing embryo. Specific ligand/receptor and receptor isoform binding preferences demonstrated by in vitro binding and mitogenic assays support this idea (Zhang et al., 2006), and the results of these studies are frequently extrapolated to infer what the functionally relevant interactions are in vivo. However, overlapping receptor binding capacities and expression patterns as well as gene ablation studies in mice clearly indicate that functional redundancy and “unfavorable” ligand/receptor interactions not only occur, but have required functions in vivo (Park et al., 2008; Powers et al., 2000). For example, no FGF8 isoforms bound or activated the “epithelial’ splice forms of FGFR1-3 in vitro (MacArthur et al., 1995) yet studies by us and others clearly indicate that FGF8 does signal to some epithelia via FGFR1 in autocrine loops (Deng et al., 1994; Park et al., 2008; Trokovic et al., 2003a; Trokovic et al., 2003b), and Moon, unpublished. Published in vitro data indicate that FGFR3c mediates the greatest response to FGF8, but FGF8 also activated FGFR1 and 2 “c” isoforms and FGFR4 in these mitogenic assays (Zhang et al., 2006). Work from our lab and many others in vivo suggest that FGFR1 and 2 are crucial mediators of FGF8 signaling in different regions of the developing embryo; this ligand seems to interact with different receptors in a highly context dependent manner and the ultimate cellular effects are context and dosage sensitive (Frank et al., 2002; Meyers et al., 1998a; Storm et al., 2006; Storm et al., 2003). The ability of a given FGF ligand to bind a particular cell in vivo, and the subsequent effects on the cell are determined by selective ligand/receptor binding and extracellular matrix (ECM) composition, particularly heparan-sulfate proteoglycans because these molecules are functional co-receptors for FGFs. Thus it is not possible to predict a priori what receptor(s) or isoforms are functional receptors for FGF8 in the lung. Temporospatially regulated ablation of candidate receptors will be required to address this question. Because expression of a soluble dominant-negative form of FgfR2, isoform 2b (the epithelially-expressed isoform) did not disrupt postnatal lung development, it has been argued that alveolar development is FGF-independent (Hokuto et al., 2003). However, this engineered receptor does not block interactions of all ligands that may interact with FGFR2b, nor with FGFR2c or other FGF receptors (Celli et al., 1998). Importantly, deletion of both FgfR3/FgfR4 permits normal prenatal lung development but disrupts postnatal alveolar formation (Weinstein et al., 1998). More recently, additional evidence indicates that FGFR3 and 4 play important roles in regulating ECM production, secondary septation, and other aspects of both normal and pathogenic alveolar development, such as seen in infants with bronchopulmonary dysplasia (Boucherat et al., 2008; Park et al., 2007; Srisuma et al., 2010). It is unknown if FGFR1 or 2 have roles in fetal lung development, but antagonism of FGFR1 function in postnatal rats disrupts alveologenesis (Yi et al., 2006). We are performing postnatal ablation of Fgf8 to determine if this ligand contributes to postnatal alveologenesis or and if post-natal ablation phenocopies abnormalities seen in FgfR3/4 deficient mice. The potential roles of FGFR1 and 2 in alveologenesis also merit further investigation. The level of Fgf8 transcripts detected in lung epithelial cells suggests that either the entire epithelium expresses at very low levels, or that only a subset of epithelial cells express Fgf8. Furthermore, the increased level of mesenchymal proliferation in hypomorphs relative to Fgf8;Isl1Cre mutants (in which only a subset of mesenchymal cells have Cre activity, see Figs. 4 and 8) suggests that that there may be additive Fgf8 functions in lung epithelia and mesenchyme or that the higher baseline proliferation in the genetic background of the Isl1Cre mutants obscures some aspects of the phenotype. Since ablation of Fgf8 restricted to the mesodermal precursors of the lung using MesP1Cre does not recapitulate the pulmonary phenotype seen in hypomorphs or Fgf8;IslCre mutants, it is likely that Fgf8 function is required in either the endodermal precursors of the lung epithelia or, at later stages in the lung epithelia derived from this endoderm. If Fgf8 function is required in the epithelial precursors, our data suggest that the affected cells are not “symptomatic” during the embryonic or pseudoglandular stages. We are currently investigating the location and fate of Fgf8-expressing cells in the developing lung using an Fgf8-lineage tracer gene-targeted line. We hypothesize that excess proliferation due to loss of autocrine FGF8 signaling causes failure of epithelial differentiation and secondary abnormal mesenchymal and endothelial cell behavior, initiating a cycle that progressively impairs ongoing epithelial/mesenchymal differentiation and vascular remodeling. That FGF8 protein has important functions in tissues in which its transcript levels are low and difficult to detect by standard methods is not unprecedented; for example, loss of Fgf8 function in the E8.5 pharyngeal mesoderm containing the second heart field, where its expression is very difficult to detect by RNA in situ hybridization, disrupts outflow tract and pharyngeal vascular development (Park et al., 2006; Illagen et al., 2006; Watanabe et al., 2009). Unfortunately, available antibodies are also insufficiently sensitive for detection of FGF8 protein in low-level producing tissues and thus were not useful for our attempts to localize FGF8 producing cells in the lung (Moon unpublished, (Sun et al., 2002)). Our finding of markedly increased proliferation as the first detectable histologic abnormality in Fgf8 mutants suggest that some targets of the transcriptional effectors of FGF8 in the lung reside high in the regulatory hierarchy controlling proliferation. The results in Figure 5 suggest that FGF8 produced by the epithelium may have different autocrine (anti-proliferative) and paracrine (pro-proliferative) effects in vivo. Our data suggest that excess mesenchymal proliferation seen in the mutants may not be a primary result of FGF8 deficiency and that FGF8 does not directly repress proliferation in these cells but rather, regulates downstream pathways such that the excess proliferation is a secondary event. An important caveat to this hypothesis is that in vivo, cellular responses such as proliferation, differentiation or cell death in different developmental contexts are very sensitive to the dosage of FGF8 sensitive (Frank et al., 2002; Meyers et al., 1998a; Storm et al., 2006; Storm et al., 2003). An outstanding question is whether the different effects we observed on BrdU incorporation, adherence, growth and survival in cultured cells are the result of differential receptor activation in mesenchymal versus epithelial cells (Supplemental Table 1), or attributable to relative dosage effects: very few epithelial cells attach and survive over the initial culture period and thus the remaining cells may be exposed to a higher ongoing effective dose of FGF8 than in the mesenchymal cultures. Clarifying the in vivo effects of FGF8 on proliferation of these different cells types will require precise temporospatial gene ablation and inducible transgenic rescue experiments. The E18.5 transcriptional profile (Fig. 6) represents an advanced stage of the Fgf8 mutant phenotype and thus does not reveal the most proximal molecular defects downstream of loss of FGF8. The alterations in Wnt pathway gene expression are particularly provocative because the profile suggests canonical pathway activation (including upregulation of Wnt7b (Rajagopal et al., 2008)) which would stimulate proliferation (as would the increases in Shh, Bmp and FGF activity suggested by the microarray and qPCR expression data, Figure 6 and Supplemental Table 2). However, the profile with regard to Wnt signaling is complicated by increased levels of Wnt5a transcripts; this ligand has been shown to decrease proliferation in lung and overexpression stimulated Fgf10 in transgenic mice (Li et al., 2005; Li et al., 2002; Tai et al., 2009). It will be necessary to determine which pathways are primarily disrupted at the onset of the phenotype to identify those that “trigger” the phenotype and whether the mesenchyme or epithelium is the primary FGF8 target. Expression analyses of isolated mesenchymal and epithelial cell populations isolated from earlier stage wild type and mutant lungs are needed to identify these most proximal molecular events. The fact that many regulators of ECM composition are dysregulated may provide important insight into the origin and/or progression of the phenotype. ECM constitution modulates bioavailability and signal transduction of numerous growth factors and morphogens. Several regulatory targets of the Ets transcription factors Pea3 and Erm that are regulated by, and mediate transcriptional effects of FGF8 are ECM remodeling enzymes (Brent and Tabin, 2004; Firnberg and Neubuser, 2002; Moon et al., 2006; Park et al., 2006; Raible and Brand, 2001; Roehl and Nusslein-Volhard, 2001; Crawford et al., 2001; Crawford and Matrisian, 1996). Both of these Ets factors are widely expressed in the lung, and expression of a dominant negative Erm/Engrailed fusion protein in airway epithelia resulted in excess Type II and absence of Type I cells (Liu et al., 2003), remarkably similar to what we observe in Fgf8 mutants. Interrogation of earlier molecular abnormalities in Fgf8 mutants may reveal that the proximal event that stimulates excess proliferation and ultimately disrupts numerous signaling cascades is altered ECM composition and intercellular interactions. For example, CTGF modulates signaling of diverse pathways via effects on the ECM. IBSP is regulated by Tgfβ and interacts with both matrix metalloproteinases and integrins to influence cell adhesion, epithelial-mesenchymal interactions and ECM deposition. Loxl2 is a lysyl oxidase that catalyzes crosslinking of elastin and collagens, and influences cell adhesion and mobility (Peinado et al., 2008). The IGF pathway regulates proliferation in fetal and postnatal lung and is normally downregulated as the distal epithelia matures (Batchelor et al., 1995; Schuller et al., 1995; Boucherat et al., 2008). IGF ligands, binding proteins (IGFBP) and their receptors are widely expressed in fetal lung, but it is IGFBPs that are most dynamically regulated with stage. IGFBP3 inhibits proliferation and causes apoptosis in some cell lines independently of IGF receptor function and decreases cell proliferation in the lung (Jones and Clemmons, 1995) and its expression is markedly decreased in the hyperproliferative lungs of Fgf8 mutants. Another potentially important link between abnormal expression of Igfbp3 and Loxl2, increased cell density, and the progressively dysplastic capillary network in Fgf8 mutant lungs is raised by a recent report that these genes are regulated by elements containing Ets binding motifs adjacent to hypoxia-response elements (Aprelikova et al., 2006; Salnikow et al., 2008). This suggests that progression of some aspects of the Fgf8 mutant phenotype may be driven by metabolic demand of the hyperplastic mutant lung exceeding its blood supply. Mice born after deficient fetal glucocorticoid signaling die from respiratory failure; their distal airspaces are abnormally remodeled and have not progressed past the canalicular stage. They have reduced numbers of Type I, but not Type II cells, and reductions in surfactant protein production (Cole et al., 1995; Cole et al., 2004; Muglia et al., 1995). Recent studies have shown that a major defect in Glucocorticoid Receptor null mutants is an excess of proliferating cells in both the mesenchymal and epithelial compartments of the distal lung at e18.5 (Bird et al., 2007). Although CNS malformations in Fgf8 mutants could affect function of the hypothalamic/pituitary/adrenal axis and fetal glucocorticoid production, it has been established that lung maturation does not depend on fetally synthesized glucocorticoids, but rather on those produced by the mother (Muglia et al., 1995; Muglia et al., 1999). Relatively little is known about the factors that direct pulmonary vascular development and coordinate it with that of the airways (Perl et al., 2002; Roman et al., 1998; Warburton et al., 2000). Our observation of progressive loss of PECAM staining and an attenuated vascular network in Fgf8 mutants will provide an important model in which to study these molecular and cellular events. Bidirectional epithelial mesenchymal signaling between respiratory epithelium and endothelial cells is required for differentiation and maturation of both populations (Bhatt et al., 2000). Loss of even a single copy of Vegfa results in early embryonic lethality due to widespread defects in vasculogenesis and angiogenesis (Carmeliet et al., 1996); Endothelial cell proliferation reflects the relative expression levels of the Vegf receptors Kdr and Flt-1 throughout the stages of lung development and vascular reorganization; these receptors mediate opposing responses that are critical for both vascular and airway morphogenesis (Galambos et al., 2002). Kdr expression depends on VEGF signaling and loss of VEGF and the resultant decrease in Kdr expression disrupts endothelial and primary saccular septal development (Yamamoto et al., 2007), as well as postnatal alveolar formation. Endothelial PECAM function is also required for airway development independent of its effects on endothelial cell proliferation and survival (DeLisser et al., 2006). Thus the decrease in Vegfa transcripts seen on the micorarray (and confirmed by qRT-PCR) combined with decreased expression of Flt1, Kdr and Pecam are likely highly significant for both the vascular and epithelial aspects of the Fgf8 mutant phenotype. Fgf8 mutants have fewer Type I cells and aberrant persistence of Nkx2.1/SPC positive pneumocytes lining the distal airways. Nkx2.1 (also called Thyroid transcription factor-1) encodes an Nkx family member that activates expression of several lung-specific genes and is required for normal airway branching and epithelial development (Kimura et al., 1996). Persistent Nkx2.1 expression in the distal lung epithelia causes Type II cell hyperplasia and inhibits alveolarization (Wert et al., 2002), similar to what we observe in of Fgf8 mutants. This supports our hypothesis that FGF8 signaling is essential for the function of specific pathways that regulate differentiation and progression from the saccular to the alveolar stage. CONCLUSION Our studies reveal that Fgf8 function is required to correctly regulate proliferation and support remodeling and differentiation during the saccular to alveolar transition. The relatively high level of Fgf8 expression in postnatal lung (Fig. 2, P1 and P5) suggests this factor may also play a role in alveologenesis. Our future experiments will employ temporspatial-specific conditional loss-of function to define the window of Fgf8 activity required for the saccular to alveolar transition and test whether this factor is also required for postnatal lung morphogenesis. Therapies for alveolar dysplasia, associated surfactant deficiency and progressive lung dysfunction in premature newborns are largely supportive; even in term infants, alveolar dysgenesis and lung dysplasias represent serious immediate threats to life and cause long-term morbidity. Evidence is accumulating that the short-term benefits of available therapies may be outweighed by long term detrimental effects on postnatal alveologenesis. Thus, the therapeutic implications of identifying a growth factor or molecular pathway that can be targeted to stimulate normal saccular and alveolar development are profound. Supplementary Material 01 Supplemental Figure 1. Embryonic lung development appears normal in Fgf8;Isl1Cre mutants Transverse sections of E13.5 control and Fgf8;Isl1Cre lungs immunostained for the markers labeled at top. Nuclei are stained blue with DAPI in all panels. No differences in airway branching detected with anti-FoxA2 (green A, A′) and anti-Nkx2.1 (pink, B, B′), proliferation (anti-pHH3, green B, B′), endothelial/vascular development (anti-PECAM, red C, C′) or peri-airway smooth muscle cell (anti-SMA, green D, D′) number or location are detectable. Red blood cell autofluorescence is apparent in all panels. Ao, aorta; E, esophagus; Ep, airway epithelia; Me, lung mesenchyme; R, right lung; L, left lung. 02 Supplemental Figure 2. Early pseudoglandular lung morphology and marker expression appears normal in Fgf8;Isl1Cre mutants Transverse sections of E14.5 control and Fgf8;Isl1Cre lungs immunostained for the markers labeled at bottom. Nuclei are stained blue with DAPI in all panels. No differences in airway branching are detected with anti-FoxA2 (pink A–B′) and anti-Nkx2.1 (yellow, C–D′) and intensity of the markers is very comparable between mutants and controls. Endothelial/vascular staining appears normal in mutants (anti-PECAM, green F, F′) as is peri-proximal airway and vascular smooth muscle cell actin intensity and location (anti-SMA, green H, H′). 03 Supplemental Figure 3. Lung development through the pseudoglandular stage appears normal in Fgf8 hypomorphic mutants Transverse sections of E14.5 (A–H′) and E15.5 (I–L′) control and Fgf8 hypomorphic lungs immunostained for the markers labeled at bottom. A–H′) Nuclei are stained blue with DAPI in all panels. No differences in airway branching are detected with anti-FoxA2 (pink B, B′) and anti-Nkx2.1 (yellow, D, D′) and intensity of these markers is very comparable between mutants and controls. Endothelial/vascular staining is normal in mutants (anti-PECAM, green F, F′) as is peri-proximal airway and vascular smooth muscle cell (anti-SMA, green H, H′) staining. I–J′) The location, pattern and intensity of FoxA2 staining (brown, DAB) in paraffin sectioned lung is comparable between mutants and controls at E15.5. K–L′) PECAM staining intensity (brown, DAB) and pattern in paraffin sectioned lung are indistinguishable between the control and mutants at E15.5; these sections were counterstained with methyl green (blue). 04 Supplemental Table 1. Relative Abundance of Fgf Receptor mRNAs in E15.5 Isolated Lung Epithelial and Mesenchymal Cells Assayed by qRT-PCR 05 Supplemental Table 2. qRT-PCR Analyses of Gene Expression Changes in E18.5 Fgf8 Hypomorphic Versus Control Lungs We thank Kandis Carter, Amy Larson, Ashna Wyckoff, Zheng Ming Wang and Asha Acharya for technical assistance, Brian Dalley and Brett Milash for their expertise with performing and analyses of the microarrays, Yumiko Saga and Sylvia Evans for mouse lines, and the Departments of Pediatric and General Surgery and Primary Children’s Medical Center Foundation for supporting Dr. Fenton. We appreciate Kirk Thomas and members of the Moon lab for critical reading and editing of the manuscript. This work was supported by an American Lung Association Career Investigator Award to AMM and in part by NIH grants HL-60061 and HL-75764 to MAS. Figure 1 Fgf8 hypomorphic mutants die at birth with dysplastic, hypercellular lungs A–D) Whole mount E18.5 control (Fgf8 H/+) and hypomorph (Fgf8 H/−) littermate mice (A, B) and lungs (C, D) photographed side by side; note that in spite of the overall decrease in body size of the mutant, the lungs are as large as those in controls (vertical black lines in C and D are the same size). E–F′) Low (E, F) and high (E′, F′) magnification photographs of H&E stained paraffin sections from P0.5 inflation-fixed lung preparations. Note the hypercellular, thickened septal walls (compare red arrows in E′ and F′) and location of many red blood cells (pink stained cells, located in capillaries) embedded in the septa whereas in the control the red blood cells are immediately adjacent to the airspaces. TRU; terminal respiratory unit Figure 2 Fgf8 is expressed at low levels throughout prenatal lung development and increases markedly postnatally Quantitative reverse transcription-PCR was used to assay Fgf8 (A) and Fgf10 (B) transcript levels in lung tissue prepared from E11.5-P5 specimens. Fold expression is shown relative to that at E14.5 and normalized to gapdh. Fgf10 quantitation is shown for comparison and increases gradually over this time period. Ct; cycle threshold value Figure 3 Separation of epithelial and mesenchymal cells from E15.5 lungs reveals enrichment of Fgf8 transcripts in the epithelial compartment Immunocytochemistry using the antibodies/markers listed at top to assay isolated mesenchymal A–D) and epithelial (A′–D′) lung cells. Note that SMA and PECAM are detected as expected in mesenchymal cells (A and B, respectively) but not in epithelial cells (A′, B′) while Nkx2.1 and FoxA2 label epithelial cells. These staining patterns validate the stringency and specificity of the cell isolation method. E) Quantitative reverse transcription-PCR was used to compare transcript levels in the isolated mesenchymal and epithelial cells; Fgf8 transcripts were 6.5 fold greater in the epithelial cells. Other marker transcripts used as controls were present in the expected cellular compartments. Figure 4 Excess proliferation occurs in lungs of Fgf8 hypomorphic mutants after E15.5 A) Immunohistochemical detection of BrdU-labeled nuclei (brown staining) in lung sections at the ages noted at top (E13.5–18.5). Control specimens are shown in the top row and Fgf8 hypomorphic specimens in the bottom row. There is a dramatic increase in the number of labeled cells in the mesenchyme and epithelia from E16.5 onward (red boxed panels). Note that the morphology of the mutant lung also appears normal prior to E16.5. B) Quantitation of distal mesenchymal proliferation confirms statistically significant increases in number of BrdU labeled nuclei in mutants at E16.5. ** p<0.005 C) Quantitation of epithelial cell proliferation confirms statistically significant increases in number of BrdU labeled nuclei in mutants at E16.5. ** p<0.005 D) Fraction of total (mesenchymal + epithelial) BrdU labeled nuclei is significantly increased in mutants at both E17.5 and E18.5. ** p<0.005 Figure 5 FGF8 stimulates BrdU incorporation by E15.5 lung mesenchymal cells in culture but represses it in epithelial cells A, B) Bar graphs representing percentage of BrdU positive nuclei under the culture conditions listed. A) FGF8 stimulates BrdU incorporation by cultured E15.5 lung mesenchymal cells in the presence or absence of serum (a, b; c,d; e, f). ** p < 0.001 B) FGF8 represses BrdU incorporation by cultured E15.5 lung epithelial cells in the presence of serum (c, d and g, h). ** p < 0.001, * p < 0.01 C, D) Overall adherence, growth and survival of cultured E15.5 lung cells as reflected by the number of total cells present at the time of assay represented as a percent of the number initially plated. C) Fewer mesenchymal cells are present at 24 hours when cultured in the presence of serum regardless of the presence of FGF8 (compare a to c, and b to d) but these cells then proliferate over the next 24 hours to increase the total number of cells by 1.5 fold (g, h). D) Few epithelial cells adhere or survive to 24 hours in the absence of serum and FGF8 treatment slightly improves viability in the first 24 hours in the absence of serum (a versus, b). Serum dramatically improves the number of cells present at 24 hours although these numbers are still < 50% of that achieved by mesenchymal cells and FGF8 has no additive effect to serum (c, d). There are significantly fewer cells in the FGF8 treated serum-free group at 48 hours (e versus f) and this is remedied by treatment with serum (h) ** p < 0.001, * p < 0.01 Figure 6 Pathway-based heat maps of gene expression changes in E18.5 hypomorphic lung tissue relative to controls Green symbolizes decreased expression in mutant and red, increased. The four columns in each heat map represent the results from each of the four different experiments and show that the expression changes detected were highly reproducible from array to array. The glucocorticoid map shows genes that have been previously shown to be dysregulated in glucocorticoid receptor null mutant lungs (Bird et al., 2007)). ECM, extracellular matrix; CAM, cell adhesion molecule Figure 7 Loss of Fgf8 function in the epithelia of Fgf8;Isl1Cre conditional mutants phenocopies the lung phenotypes of Fgf8 hypomorphs but ablation in the lung mesenchyme of Fgf8;MesP1Cre mutants does not A–D′) H&E stained paraffin sections from E18.5 lung specimens obtained from mutants with the genotypes listed at top. Low and high magnification views are shown and were obtained at similar peripheral locations in the lung. B, B′) Ablation of Fgf8 only in the mesenchyme with MesP1Cre permits normal lung development through fetal development. C, C′) Fgf8 hypomorphic mutant lung phenotype. D, D′) Fgf8;Isl1Cre conditional mutant lung has thickened septa and few airspaces, phenocopying the hypomorphic lung. E) The ROSA26 lacz Cre reporter shows that MesP1Cre is active and recombines in all lung mesenchymal (Me) precursors by E11.5 but no evidence of recombination is detectable in the lung epithelia (Ep). F) The ROSA26 lacz Cre reporter shows that Isl1Cre is active and recombines broadly in lung epithelial precursors by E11.5, but only in a subset of lung mesenchyme. Figure 8 Excess proliferation occurs in the lungs of Fgf8;Isl1Cre conditional mutants after E15.5 A) Immunohistochemical detection of BrdU-labeled nuclei (brown DAB staining) in lung sections at the ages noted at top (E14.5–18.5). Control specimens are shown in the top row and Fgf8;Isl1Cre conditional mutant specimens in the bottom row. There is an excess in the number of labeled cells in the mesenchyme and epithelia from E16.5 onward (red boxed sections). Note morphology of the mutant lung also appears normal prior to E16.5. B) Quantitation of distal epithelial proliferation confirms statistically significant increases in number of BrdU labeled nuclei in mutants at E16.5. Note that proliferation of both cell types is increased in the controls in this background as compared to that of the controls in Figure 4. * p< 0.05 C) Quantitation of mesenchymal cell proliferation confirms statistically significant increases in number of BrdU labeled nuclei in mutants at E16.5. ** p<0.005 D) Fraction of total (mesenchymal + epithelial) BrdU labeled nuclei is significantly increased in mutants at both E17.5 and E18.5. ** p<0.005 Figure 9 Differentiation and vascular remodeling are disrupted at the saccular and early alveolar stages in Fgf8;Isl1Cre conditional mutants All panels show immunohistochemical stains of cryosectioned lung specimens at the ages stated. A–C′) Staining for FoxA2 protein in the distal epithelia (red) from E16.5–e18.5. There are few FoxA2+ cells lining the forming airspaces in E17.5 and E18.5 controls (B, C white arrowheads) and a normal pattern in mutants at E16.5 (A′). At E17.5 (B′) and 18.5 (C′) mutants have a paucity of airspaces and persistent high numbers of FoxA2+ cells (yellow arrowheads) in the developing airspaces. D–F′) Staining of the developing pulmonary vasculature for PECAM immunoreactivity (green) reveals grossly normal vascular development in the mutants until E18.5; at E18.5 there is less detectable staining (F′, yellow arrowheads). G–H) Consistent with the FoxA2 data, double staining for Nkx2.1 (green) and nuclei (blue) shows a few Nkx2.1+ cells located in the septa, but not lining the airspaces of controls (white arrowheads) while there is abnormal persistence of many Nkx2.1+ cells lining the small airspaces of the mutants (red arrowheads). I, J) Double staining for surfactant protein C (SPC, red) and nuclei (blue) shows increased number of SPC+ Type II alveolar cells or a more primitive precursor cell lining the airspaces. K–N) Analysis at E19. Double staining for PECAM (yellow/green) and nuclei (red) reveals that, relative to controls (K, K′), there is a paucity of distal vessels and vessels are aberrantly located with the thickened septa of the mutants (L, L′) rather than adjacent to the airspaces. Aquaporin 5 staining reveals paucity of Type I cells lining the airspaces and their aberrant cuboidal morphology in mutants (N, arrowheads) as compared to their larger surface area and squamous morphology which gives a fine reticular pattern in controls (M, arrowheads). Figure 10 Differentiation and vascular remodeling are disrupted at the saccular and early alveolar stages in Fgf8 hypomorphic mutants All panels show immunohistochemical stains of cryosectioned lung specimens at the E18.5. A–A′) Staining of the developing pulmonary vasculature for PECAM immunoreactivity (red) reveals is less detectable staining in the mutants. B, B′) Double staining for surfactant protein C (SPC, red) and nuclei (blue) shows increased number of SPC+ cells lining the airspaces. C, C′) Double staining for Nkx2.1 (green) and nuclei (blue) shows rare Nkx2.1+ cells in the airspaces of controls (white arrowheads) while there is abnormal persistence of many Nkx2.1+ cells lining the small airspaces of the mutants. D, D′) Staining for FoxA2 protein in the distal epithelia (pink). There are few FoxA2+ cells lining the forming airspaces in controls while mutants have a paucity of airspaces and persistent high numbers of FoxA2+ cells lining the small airspaces. E, E′) Aquaporin 5 staining reveals paucity of Type I cells in mutants. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. REFERENCES CITED Abler LL Mansour SL Sun X 2009 Conditional gene inactivation reveals roles for Fgf10 and Fgfr2 in establishing a normal pattern of epithelial branching in the mouse lung Dev Dyn 238 1999 2013 19618463 Abu-Issa R Smyth G Smoak I Yamamura K Meyers EN 2002 Fgf8 is required for pharyngeal arch and cardiovascular development in the mouse Development 129 4613 4625 12223417 Alcorn DG Adamson TM Maloney JE Robinson PM 1981 A morphologic and morphometric analysis of fetal lung development in the sheep Anatomical Record 201 655 667 7340570 Aprelikova O Wood M Tackett S Chandramouli GV Barrett JC 2006 Role of ETS transcription factors in the hypoxia-inducible factor-2 target gene selection Cancer Res 66 5641 5647 16740701 Arman E Haffner-Krausz R Gorivodsky M Lonai P 1999 Fgfr2 is required for limb outgrowth and lung-branching morphogenesis Proc Natl Acad Sci U S A 96 11895 11899 10518547 Batchelor DC Hutchins AM Klempt M Skinner SJ 1995 Developmental changes in the expression patterns of IGFs, type 1 IGF receptor and IGF-binding proteins-2 and -4 in perinatal rat lung J Mol Endocrinol 15 105 115 8800636 Bellusci S Grindley J Emoto H Itah N Hogan BL 1997 Fibroblast growth factor 10 (FGF10) and branching morphogenesis in the embryonic mouse lung Development 124 4867 4878 9428423 Bhatt AJ Amin SB Chess PR Watkins RH Maniscalco WM 2000 Expression of vascular endothelial growth factor and Flk-1 in developing and glucocorticoid-treated mouse lung Pediatr Res 47 606 613 10813585 Bird AD Tan KH Olsson PF Zieba M Flecknoe SJ Liddicoat DR Mollard R Hooper SB Cole TJ 2007 Identification of glucocorticoid-regulated genes that control cell proliferation during murine respiratory development J Physiol 585 187 201 17901120 Boucherat O Franco-Montoya M Thibault C Incitti R Chailley-Heu B Delacourt C Bourbon JR 2007 Gene expression profiling in lung fibroblasts reveals new players in alveolarization Physiological Genomics 32 128 141 17911382 Boyden EA 1974 The mode of origin of pulmonary acini and respiratory bronchioles in the fetal lung American Journal of Anatomy 141 317 328 4614663 Brent AE Tabin CJ 2004 FGF acts directly on the somitic tendon progenitors through the Ets transcription factors Pea3 and Erm to regulate scleraxis expression Development 131 3885 3896 15253939 Bruce MC Honaker CE Cross RJ 1999 Lung fibroblasts undergo apoptopsis following alveolarization American Journal of Respiratory Cell and Molecular Biology 20 228 236 9922213 Burri PH 1997 Structural aspects of prenatal and postnatal development and growth of the lung McDonald JA Lung Growth and Development Marcel Dekker, Inc New York 1 35 Cardoso WV 2001 Molecular regulation of lung development Annual Review of Physiology 63 471 494 Carmeliet P Ferreira V Breier G Pollefeyt S Kieckens L Gertsenstein M Fahrig M Vandenhoeck A Harpal K Eberhardt C Declercq C Pawling J Moons L Collen D Risau W Nagy A 1996 Abnormal blood vessel development and lethality in embryos lacking a single VEGF allele Nature 380 435 439 8602241 Celli G LaRochelle WJ Mackem S Sharp R Merlino G 1998 Soluble dominant-negative receptor uncovers essential roles for fibroblast growth factors in multi-organ induction and patterning Embo J 17 1642 1655 9501086 Cole TJ Blendy JA Monaghan AP Krieglstein K Schmid W Aguzzi A Fantuzzi G Hummler E Unsicker K Schutz G 1995 Targeted disruption of the glucocorticoid receptor gene blocks adrenergic chromaffin cell development and severely retards lung maturation Genes Dev 9 1608 1621 7628695 Cole TJ Solomon NM Van Driel R Monk JA Bird D Richardson SJ Dilley RJ Hooper SB 2004 Altered epithelial cell proportions in the fetal lung of glucocorticoid receptor null mice Am J Respir Cell Mol Biol 30 613 619 14578211 Colvin JS Feldman B Nadeau JH Goldfarb M Ornitz DM 1999 Genomic organization and embryonic expression of the mouse fibroblast growth factor 9 gene Dev Dyn 216 72 88 10474167 Colvin JS White AC Pratt SJ Ornitz DM 2001 Lung hypoplasia and neonatal death in Fgf9-null mice identify this gene as an essential regulator of lung mesenchyme Development 128 2095 2106 11493531 Crawford HC Fingleton B Gustavson MD Kurpios N Wagenaar RA Hassell JA Matrisian LM 2001 The PEA3 subfamily of Ets transcription factors synergizes with beta-catenin-LEF-1 to activate matrilysin transcription in intestinal tumors Mol Cell Biol 21 1370 1383 11158322 Crawford HC Matrisian LM 1996 Mechanisms controlling the transcription of matrix metalloproteinase genes in normal and neoplastic cells Enzyme Protein 49 20 37 8796995 Curtis RK Oresic M Vidal-Puig A 2005 Pathways to the analysis of microarray data Trends Biotechnol 23 429 435 15950303 Davies G Reid L 1970 Growth of the alveoli and pulmonary arteries in childhood Thorax 25 669 681 5533319 Davies P Reid L Lister G Pitt B 1988 Postnatal growth of the sheep lung: a morphometric study Anatomical Record 220 281 286 3364753 De Moerlooze L Spencer-Dene B Revest J Hajihosseini M Rosewell I Dickson C 2000 An important role for the IIIb isoform of fibroblast growth factor receptor 2 (FGFR2) in mesenchymal-epithelial signalling during mouse organogenesis Development 127 483 492 10631169 DeLisser HM Helmke BP Cao G Egan PM Taichman D Fehrenbach M Zaman A Cui Z Mohan GS Baldwin HS Davies PF Savani RC 2006 Loss of PECAM-1 function impairs alveolarization J Biol Chem 281 8724 8731 16377626 Deng CX Wynshaw-Boris A Shen MM Daugherty C Ornitz DM Leder P 1994 Murine FGFR-1 is required for early postimplantation growth and axial organization Genes Dev 8 p3045 3057 Docimo SG Crone RK Davies P Reid L Retik AB Mandell J 1991 Pulmonary development in the fetal lamb: Quantitative histologic study of the alveolar phase Anatomical Record 229 495 498 2048754 Dunnill MS 1962 Postnatal growth of the lung Thorax 17 329 333 Emery JL Wilcock PF 1966 The postnatal development of the lung Acta Anatomica 65 10 29 6007548 Firnberg N Neubuser A 2002 FGF signaling regulates expression of Tbx2, Erm, Pea3, and Pax3 in the early nasal region Dev Biol 247 237 250 12086464 Frank DU Fotheringham LK Brewer JA Muglia LJ Tristani-Firouzi M Capecchi MR Moon AM 2002 An Fgf8 mouse mutant phenocopies human 22q11 deletion syndrome Development 129 4591 4603 12223415 Galambos C Ng YS Ali A Nogushi A Lovejoy S D’Amore PA deMello DE 2002 Defective pulmonary development in the absence of heparin-binding vascular endothelial growth factor isoforms American Journal of Respiratory Cell and Molecular Biology 27 194 203 12151311 Guo L Degenstein L Fuchs E 1996 Keratinocyte growth factor is required for hair development but not for wound healing Genes Dev 10 165 175 8566750 Hokuto I Perl AK Whitsett JA 2003 Prenatal, but not postnatal, inhibition of fibroblast growth factor receptor signaling causes emphysema J Biol Chem 278 415 421 12399466 Hu MC Qui WR Wang YP Hill D Ring BD Scully S Bolon B DeRose M Luethy R Simonet WS Arakawa T Danilenko DM 1998 FGF-18, a novel member of the fibroblast growth factor family, stimulates hepatic and intestinal proliferation Molecular and Cellular Biology 18 6063 6074 9742123 Ilagan R Abu-Issa R Brown D Yang YP Jiao K Schwartz RJ Klingensmith J Meyers EN 2006 Fgf8 is required for anterior heart field development Development 133 2435 2445 16720880 Jones JI Clemmons DR 1995 Insulin-like growth factors and their binding proteins: biological actions Endocr Rev 16 3 34 7758431 Kimura S Hara Y Pineau T Fernandez-Salguero P Fox CH Ward JM Gonzalez FJ 1996 The T/ebp null mouse: thyroid-specific enhancer-binding protein is essential for the organogenesis of the thyroid, lung, ventral forebrain, and pituitary Genes Dev 10 60 69 8557195 Langston C Kida K Reed M Thurlbeck WM 1984 Human lung growth in late gestation and in the neonate American Review of Respiratory Disease 129 607 613 6538770 Lewandoski M Sun X Martin GR 2000 Fgf8 signalling from the AER is essential for normal limb development Nat Genet 26 460 463 11101846 Li C Hu L Xiao J Chen H Li JT Bellusci S Delanghe S Minoo P 2005 Wnt5a regulates Shh and Fgf10 signaling during lung development Dev Biol 287 86 97 16169547 Li C Xiao J Hormi K Borok Z Minoo P 2002 Wnt5a participates in distal lung morphogenesis Dev Biol 248 68 81 12142021 Lin W Furthauer M Thisse B Thisse C Jing N Ang SL 2002 Cloning of the mouse Sef gene and comparative analysis of its expression with Fgf8 and Spry2 during embryogenesis Mech Dev 113 163 168 11960706 Lindahl P Karlsson L Hellstrom M Gebre-Medhin S Willetts K Heath JK Betsholtz C 1997 Alveogenesis failure in PDGF-A-deficient mice is coupled to lack of distal spreading of alveolar smooth muscle cell progenitors during lung development Development 124 3943 3953 9374392 Liu Y Jiang H Crawford HC Hogan BL 2003 Role for ETS domain transcription factors Pea3/Erm in mouse lung development Dev Biol 261 10 24 12941618 MacArthur CA Lawshe A Xu J Santos-Ocampo S Heikinheimo M Chellaiah AT Ornitz DM 1995 FGF-8 isoforms activate receptor splice forms that are expressed in mesenchymal regions of mouse development Development 121 3603 3613 8582274 Macatee TL Hammond BP Arenkiel BR Francis L Frank DU Moon AM 2003 Ablation of specific expression domains reveals discrete functions of ectoderm- and endoderm-derived FGF8 during cardiovascular and pharyngeal development Development 130 6361 6374 14623825 Metzger RJ Klein OD Martin GR Krasnow MA 2008 The branching programme of mouse lung development Nature 453 745 750 18463632 Meyers EN Lewandoski M Martin GR 1998a An Fgf8 allelic mutant allelic series generated by Cre- and Flp-mediated recombination Nature Genetics 18 136 141 9462741 Meyers EN Lewandoski M Martin GR 1998b An Fgf8 mutant allelic series generated by Cre- and Flp-mediated recombination Nat Genet 18 136 141 9462741 Min H Danilenko DM Scully SA Bolon B Ring BD Tarpley JE DeRose M Simonet SW 1998 Fgf-10 is required for both limb and lung development and exhibits striking functional similarity to Drosophila branchless Genes and Development 12 3156 3161 9784490 Moon AM 2006 Mouse models for investigating the developmental basis of human birth defects Pediatr Res 59 749 755 16641221 Moon AM Capecchi MR 2000 FGF8 is required for outgrowth and patterning of the limbs Nature Genetics 26 455 459 11101845 Moon AM Guris DL Seo JH Li L Hammond J Talbot A Imamoto A 2006 Crkl deficiency disrupts fgf8 signaling in a mouse model of 22q11 deletion syndromes Dev Cell 10 71 80 16399079 Muglia L Jacobson L Dikkes P Majzoub JA 1995 Corticotropin-releasing hormone deficiency reveals major fetal but not adult glucocorticoid need Nature 373 427 432 7830793 Muglia LJ Bae DS Brown TT Vogt SK Alvarez JG Sunday ME Majzoub JA 1999 Proliferation and differentiation defects during lung development in corticotropin-releasing hormone-deficient mice Am J Respir Cell Mol Biol 20 181 188 9922208 Olson EN Arnold HH Rigby PW Wold BJ 1996 Know your neighbors: three phenotypes in null mutants of the myogenic bHLH gene MRF4 Cell 85 1 4 8620528 Park EJ Ogden LA Talbot A Evans S Cai CL Black BL Frank DU Moon AM 2006 Required, tissue-specific roles for Fgf8 in outflow tract formation and remodeling Development 133 2419 2433 16720879 Park EJ Watanabe Y Smyth G Miyagawa-Tomita S Meyers E Klingensmith J Camenisch T Buckingham M Moon AM 2008 An FGF autocrine loop initiated in second heart field mesoderm regulates morphogenesis at the arterial pole of the heart Development 135 3599 3610 18832392 Park M Rieger-Fackeldey E Schanbacher BL Cook AC Bauer JA Rogers LK Hansen TN Welty SE Smith CV 2007 Altered expressions of fibroblast growth factor receptors and alveolarization in neonatal mice exposed to 85% oxygen Pediatr Res 62 6 652 7 17957151 Peinado H Moreno-Bueno G Hardisson D Perez-Gomez E Santos V Mendiola M de Diego JI Nistal M Quintanilla M Portillo F Cano A 2008 Lysyl oxidase-like 2 as a new poor prognosis marker of squamous cell carcinomas Cancer Res 68 4541 4550 18559498 Perl AK Wert SE Nagy A Lobe CG Whitsett JA 2002 Early restriction of peripheral and proximal cell lineages during formation of the lung Proc Natl Acad Sci U S A 99 10482 10487 12145322 Powell PP Wang CC Horinouchi H Shepherd K Jacobson M Lipson M Jones R 1998 Differential expression of fibroblast growth factor receptors 1 to 4 and ligand genes in late fetal and early postnatal rat lung Am J Respir Cell Mol Biol 19 563 572 9761752 Powers CJ McLeskey SW Wellstein A 2000 Fibroblast growth factors, their receptors and signaling Endocr Relat Cancer 7 165 197 11021964 Raible F Brand M 2001 Tight transcriptional control of the ETS domain factors Erm and Pea3 by Fgf signaling during early zebrafish development Mech Dev 107 105 117 11520667 Rajagopal J Carroll TJ Guseh JS Bores SA Blank LJ Anderson WJ Yu J Zhou Q McMahon AP Melton DA 2008 Wnt7b stimulates embryonic lung growth by coordinately increasing the replication of epithelium and mesenchyme Development 135 1625 1634 18367557 Roehl H Nusslein-Volhard C 2001 Zebrafish pea3 and erm are general targets of FGF8 signaling Curr Biol 11 503 507 11413000 Roman J Schuyler W McDonald JA Roser S 1998 Heparin inhibits lung branching morphogenesis: potential role of smooth muscle cells in cleft formation Am J Med Sci 316 368 378 9856690 Saga Y Miyagawa-Tomita S Takagi A Kitajima S Miyazaki J Inoue T 1999 MesP1 is expressed in the heart precursor cells and required for the formation of a single heart tube Development 126 3437 3447 10393122 Salnikow K Aprelikova O Ivanov S Tackett S Kaczmarek M Karaczyn A Yee H Kasprzak KS Niederhuber J 2008 Regulation of hypoxia-inducible genes by ETS1 transcription factor Carcinogenesis 29 1493 1499 18381358 Sannes PL Burch KK Khosia J 1992 Immunohistochemical localization of epidermal growth factor and acidic and basic fibroblast growth factors in postnatal developing and adult rat lung American Journal of Respiratory Cell and Molecular Biology 7 230 237 1379823 Schittny JC Djonov V Fine A Burri PH 1998 Programmed cell death contributes to postnatal lung development Am J Respir Cell Mol Biol 18 786 793 9618383 Schmitt JF Hearn MT Risbridger GP 1996 Expression of fibroblast growth factor-8 in adult rat tissues and human prostate carcinoma cells J Steroid Biochem Mol Biol 57 173 178 8645626 Schuger L Varani J Mitra R Jr Gilbride K 1993 Retinoic acid stimulates mouse lung development by a mechanism involving epithelial-mesenchymal interaction and regulation of epidermal growth factor receptors Dev Biol 159 462 473 8405671 Schuller AG van Neck JW Beukenholdt RW Zwarthoff EC Drop SL 1995 IGF, type I IGF receptor and IGF-binding protein mRNA expression in the developing mouse lung J Mol Endocrinol 14 349 355 7545402 Sekine K Ohuchi H Fujiwara M Yamasaki M Yoshizawa T Soto T Yagishita N Matsui D Koga Y Itoh N Kato S 1999 Fgf10 is essential for limb and lung formation Nature Genetics 21 138 141 9916808 Soriano P 1999 Generalized lacZ expression with the ROSA26 Cre reporter strain Nat Genet 21 70 71 9916792 Srisuma S Bhattacharya S Simon DM Solleti SK Tyagi S Starcher B Mariani TJ 2010 Fibroblast growth factor receptors control epithelial-mesenchymal interactions necessary for alveolar elastogenesis Am J Resp Crit Care Med 181 8 838 50 20093646 Storm EE Garel S Borello U Hebert JM Martinez S McConnell SK Martin GR Rubenstein JL 2006 Dose-dependent functions of Fgf8 in regulating telencephalic patterning centers Development 133 1831 1844 16613831 Storm EE Rubenstein JL Martin GR 2003 Dosage of Fgf8 determines whether cell survival is positively or negatively regulated in the developing forebrain Proc Natl Acad Sci U S A 100 1757 1762 12574514 Sun X Mariani FV Martin GR 2002 Functions of FGF signalling from the apical ectodermal ridge in limb development Nature 418 501 508 12152071 Sun X Meyers EN Lewandoski M Martin GR 1999 Targeted disruption of Fgf8 causes failure of cell migration in the gastrulating mouse embryo Genes Dev 13 1834 1846 10421635 Szebenyi G Fallon JF 1999 Fibroblast growth factors as multifunctional signaling factors Int Rev Cytol 185 45 106 9750265 Tai CC Sala FG Ford HR Wang KS Li C Minoo P Grikscheit TC Bellusci S 2009 Wnt5a knock-out mouse as a new model of anorectal malformation J Surg Res 156 278 282 19577771 Trokovic N Trokovic R Mai P Partanen J 2003a Fgfr1 regulates patterning of the pharyngeal region Genes Dev 17 141 153 12514106 Trokovic R Trokovic N Hernesniemi S Pirvola U Vogt Weisenhorn DM Rossant J McMahon AP Wurst W Partanen J 2003b FGFR1 is independently required in both developing mid- and hindbrain for sustained response to isthmic signals Embo J 22 1811 1823 12682014 Trumpp A Depew MJ Rubenstein JL Bishop JM Martin GR 1999 Cre-mediated gene inactivation demonstrates that FGF8 is required for cell survival and patterning of the first branchial arch Genes Dev 13 3136 3148 10601039 Usui H Shibayama M Ohbayashi N Konishi M Takada S Itoh N 2004 Fgf18 is required for embryonic lung alveolar development Biochem Biophys Res Commun 322 887 892 15336546 Warburton D Schwarz M Tefft D Flores-Delgado G Anderson KD Cardoso WV 2000 The molecular basis of lung morphogenesis Mech Dev 92 55 81 10704888 Warburton D Zhao J Berberich MA Bernfield M 1999 Molecular embryology of the lung: then, now, and in the future American Journal of Physiology Lung Cellular Molecular Physiology 276 L697 L704 Watanabe Y Miyagawa-Tomita S Vincent SD Kelly RG Moon AM Buckingham ME 2009 Role of Mesodermal FGF8 and FGF10 Overlaps in the Development of the Arterial Pole of the Heart and Pharyngeal Arch Arteries Circ Res Weinstein M Xu X Ohyama K Deng CX 1998 FGFR-3 and FGFR-4 function cooperatively to direct alveogenesis in the murine lung Development 125 3615 3623 9716527 Wert SE Dey CR Blair PA Kimura S Whitsett JA 2002 Increased expression of thyroid transcription factor-1 (TTF-1) in respiratory epithelial cells inhibits alveolarization and causes pulmonary inflammation Dev Biol 242 75 87 11820807 Yamamoto H Yun EJ Gerber HP Ferrara N Whitsett JA Vu TH 2007 Epithelial-vascular cross talk mediated by VEGF-A and HGF signaling directs primary septa formation during distal lung morphogenesis Dev Biol 308 44 53 17583691 Yi M Belcastro R Shek S Luo D Post M Tanswell AK 2006 Fibroblast growth factor-2 and receptor-1alpha(IIIc) regulate postnatal rat lung cell apoptosis Am J Respir Crit Care Med 174 581 589 16728710 Zhang X Ibrahimi OA Olsen SK Umemori H Mohammadi M Ornitz DM 2006 Receptor specificity of the fibroblast growth factor family. The complete mammalian FGF family J Biol Chem 281 15694 15700 16597617

PMC005xxxxxx/PMC5134256.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 0410462 6011 Nature Nature Nature 0028-0836 1476-4687 26863189 5134256 10.1038/nature16995 NIHMS827774 Article Structural basis for promiscuous PAM recognition in Type I-E Cascade from E. coli Hayes Robert P. 1 Xiao Yibei 1 Ding Fran 1 van Erp Paul B. G. 4 Rajashankar Kanagalaghatta 2 Bailey Scott 3 Wiedenheft Blake 4 Ke Ailong 1* 1 Department of Molecular Biology and Genetics, Cornell University, 253 Biotechnology Building, Ithaca, NY 14853, USA 2 Department of Chemistry and Chemical Biology, Cornell University, NE-CAT, Advanced Photon Source, Argonne National Laboratory, Argonne, IL 60439, USA 3 Department of Biochemistry and Molecular Biology, Johns Hopkins University, Bloomberg School of Public Health, Baltimore, MD 21205, USA 4 Department of Microbiology and Immunology, Montana State University, Bozeman, MT 59717, USA * Correspondence to ailong.ke@cornell.edu; ak425@cornell.edu 4 11 2016 10 2 2016 25 2 2016 02 12 2016 530 7591 499503 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. Clusters of regularly interspaced short palindromic repeats (CRISPRs) and cas (CRISPR-associated) operon form an RNA-based adaptive immune system against foreign genetic elements in prokaryotes1. Type I account for 95% of CRISPR systems, and have been utilized to control gene expression and cell fate2,3. During CRISPR RNA (crRNA)-guided interference, Cascade (CRISPR-associated complex for antiviral defense) facilitates crRNA-guided invasion of double-stranded DNA (dsDNA) for complementary base-pairing with the target DNA strand, while displacing the non-target strand, forming an R-loop4,5. Cas3 nuclease/helicase is recruited subsequently to degrade two DNA strands4,6,7. Protospacer adjacent motif (PAM) flanking target DNA is crucial for self vs. foreign discrimination4,8–16. Here we present a 2.45 Å crystal structure of E. coli Cascade bound to a foreign dsDNA target. The 5′-ATG PAM is recognized in double-stranded form, from the minor groove side, by three structural features in Cse1. The promiscuity inherent to minor groove DNA recognition rationalizes the puzzling observation that a single Cascade can respond to several distinct PAM sequences. Optimal PAM recognition coincides with a wedge insertion, initiating the directional target DNA strand unwinding for segmented base-pairing with crRNA. The non-target strand is guided along a parallel path 25 Å apart, and the R-loop structure is further stabilized by locking this strand behind Cse2 dimer. These observations provide the structure basis for understanding the PAM-dependent directional R-loop formation process17,18. Differentiating between ‘self’ and ‘non-self’ antigens is critical in CRISPR systems, as foreign target sequences (protospacers) are identical to sequences recorded in the host CRISPR locus (spacers). In Type I and II CRISPR systems, foreign DNA detection relies on protein-mediated PAM recognition4,8–16. Whereas PAM recognition in Cas9-based Type II systems has been elucidated based on major-groove DNA contact19,20, it remains unclear as to whether Cascade recognizes PAM from DNA major or minor groove12, in ss- or ds-form4,21. A particularly puzzling observation is the promiscuity in PAM recognition. Five PAM sequences (5′-ATG, AAG, AGG, GAG, and TAG reading from the non-target strand) are capable of triggering robust CRISPR interference via E. coli Cascade4,11,21,22. Crystal structures of E. coli Cascade in free- and ssDNA-bound forms revealed multiple conformational states, and provided valuable insights about the crRNA-guided ssDNA recognition mechanism21,23,24, however, the mechanisms for dsDNA entry, PAM recognition, and R-loop formation remain poorly defined. To understand the PAM-dependent foreign DNA recognition mechanism, we determined the 2.45 Å crystal structure of E. coli Cascade bound to dsDNA that forms a partial R-loop (Fig. 1a–c, Extended Data Fig. 1); such DNA substrates were efficiently bound by T. fusca Type I-E Cascade, and the resulting complex specifically recruited Cas325. This structure agrees well with the cryo-EM reconstruction of the full R-loop/Cascade complex, underlining its validity in explaining the R-loop formation process21 (Extended Data Fig. 2, Supplementary Video 1). Comparison with high-resolution crystal structures suggests the R-loop formation requires both the sliding of Cse1-CTD (C-terminal domain)-Cse2.1-Cse2.2, as seen in the ssDNA-bound structure, and the engagement of Cse1-NTD (N-terminal domain) as seen in the free-Cascade structure21,23,24 (Extended Data Fig. 3, Supplementary Videos 2, 3). In addition, a localized conformational change takes place near the putative Cas3 binding site21, which is only observed in this R-loop bound structure, therefore may play a role in Cas3 recruitment (Extended Data Fig. 4). dsDNA enters Cascade between Cas7.5 and Cas7.6, contacted by the lysine-rich helices16,26 (Fig. 1d–e, Extended Data Fig. 5). DNA bifurcates underneath PAM. The entire target DNA strand flips to form the segmented DNA-crRNA duplex27. The 10-nt non-target strand is guided to a parallel path 25 Å apart by sequence nonspecific contacts, an active mechanism to stabilize R-loop (Fig. 1d–e). Modeling dsDNA beyond PAM projects it across Cse1-CTD without severe steric clashes (Fig. 1f), illustrating a possible PAM-searching scenario21. The 5′-ATG PAM sequence is recognized in the double-stranded form, from the minor groove side, by the Cse1 subunit of the Cascade (Fig. 2a–b). This rather surprising mode of recognition strongly biases towards the target DNA strand, which rationalizes the previous observations that mismatched PAMs could be tolerated, provided the target strand sequence was optimal4,21. Three structural features in Cse1 are involved in PAM recognition: the glutamine-wedge, the glycine-loop and a lysine-finger (Fig. 2a–b). Only CT-1−GNT-1 is tolerated at the −1 PAM position (PAM-1) in E. coli5,21, although recent analyses revealed spacer-dependent tolerance of alternative base-pairs at PAM-128. In our structure (Fig. 2a–d), the amide of A355 in the glutamine-wedge donates a H-bond to O2 of CT-1, specifying a pyrimidine in the target strand. The carbonyl of G157 in the glycine-loop accepts a water-mediated H-bond from N2 of GNT-1. GNT-1-to-inosine substitution assayed by Electrophoretic Mobility Shift Assay (EMSA) suggests this contact only provides minor discrimination against ANT-1(Extended Data Fig. 6a). Cascade’s affinity for different PAM-1 base-pair combinations corroborate well with the structure observation that a target strand pyrimidine is strongly specified (Extended Data Fig. 6b). The 10-fold differences in Kd, however, do not support an absolute CT-1−GNT-1 specification at PAM-1. Indeed, Cas3 was able to cleave alternative PAM-1 targets, provided Cascade concentration is above the corresponding Kd (Extended Data Fig. 6c). These results echo the recent finding in suggesting that PAM-1 readout is further complicated by Cascade/Cas3 expression level and spacer content28. Importantly, PAM recognition coincides with the glutamine-wedge insertion into dsDNA underneath PAM (Fig. 2a, d). The tip residue Q354 stacks underneath CT-1 and, together with N353, sterically displaces the first two protospacer nucleotides in the target strand, forcing them to rotate outwards. Given its strategic location, it is rather surprising that this wedge is not highly conserved in sequence (Extended Data Fig. 7). Indeed, tip residue substitutions (Q354A, N353A, and Q354A/N353A) had little consequence in DNA-binding. In contrast, trimming this wedge (NNQAS352-356/GG) led to 100-fold DNA binding defect, suggesting the wedge functions through a steric interference mechanism, to nucleate the target strand displacement upon PAM recognition (Fig. 2e–f). A serine-to-phosphate ‘lock’ is essential in initiating the target strand flipping in Cas919. T125 is in a similar location in our structure but contacts the +1 bridging oxygen instead, and T125A substitution had negligible defect (Fig. 2e). Recognition at PAM-2 is promiscuous by E. coli Cascade. Only GT-2-CNT-2 is rejected at this position; the other three combinations lead to efficient interference22. Here the glycine-loop residues (159–161) assume a lip-like structure, introduces DNA bending at AT-2-TNT-2, and ‘bites’ onto PAM-1 base-pair in conjunction with the glutamine-wedge underneath; TNT-2 retreats backwards and tilts upwards (Fig. 2b, d). The rim of the glycine-loop explores shape-complementarity to AT-2, and donates a weak H-bond to N3 of AT-2. G160A substitution disrupts the shape complementarity, and reduced Cascade-binding affinity by ~100-fold and Cas3-cleavage to baseline level (Fig. 2e–f). Rejection of GT-2-CNT-2 at PAM-2 is rationalized by the fact that N2 amine of GT-2 would introduce steric clashes against the glycine loop; whereas TT-2-ANT-2 or CT-2-GNT-2 would not, based on modeling (Extended Data Fig. 8a). Indeed, removal of this amine in inosine substitution largely rescued the DNA-binding defect, confirming that N2 of GT-2 is the anti-determinant for PAM-2 specificity (Extended Data Fig. 8b). An equivalent glycine-rich loop is present in all known Cse1 structures; they likely play a similar minor groove DNA recognition function (Extended Data Fig. 7). PAM-3 is typically specified as a pyrimidineT-purineNT pair by E. coli Cascade4,11,21. 5′-TAG PAM also leads to interference, but GT-3-CNT-3 containing PAMs fail to22,28. Here a favorable electrostatic interaction from a lysine-finger (K268) to O2 of TT-3 is observed, which rationalizes the strong preference for pyrimidine at this position (Fig. 2b, d). K268A mutation reduced Cascade-binding and Cas3-cleavage by >8-fold and >10-fold, respectively, emphasizing its positive contribution to PAM-3 recognition (Fig. 2e–f). Interestingly, K268A mutant still retained wild-type level discrimination against 5′-CTG-PAM (Extended Data Fig. 8c). GT-3–to-inosine substitution ultimately proved that N2 of GT-3 also serves as a strong anti-determinant in the rejection of GT-3–containing PAMs (Extended Data Fig. 8d). Interestingly, K268 makes an electrostatic interaction to CT-4 (Fig. 2b, d), implying certain level of sequence discrimination at PAM-4 as well. Detailed structure dissection also helps rationalizing the self-avoidance mechanism. All spacers in E. coli CRISPR loci are ‘protected’ by a 5′-CCG-PAM. This PAM is the combination of the least-preferred nucleotides at each position (3′-GT-3GT-2CT-1), which would strongly disfavor Cascade-mediated R-loop formation, despite a perfect spacer match. The non-target strand is guided sequence nonspecifically along the Cascade surface, 20–25 Å away from the target strand (Fig. 3a). The sugar-phosphates of nucleotides 1–3 are contacted by K163, G169, and K296 of Cse1-NTD, nucleotides 6–9 by a string of positive charges(R393/K394/K488/H489/R491) across Cse1-CTD (Fig. 3a). The redundant interactions were not disrupted by a point mutation (Y397A). A double mutant (488–489 KH/AA) neutralizing a positive patch, however, led to a 4-fold binding defect (Fig. 3b). The 10th/last nucleotide rests at an intersection between Cse1-CTD and Cse2.1. To investigate whether the following non-target residues travels on the surface or backside of the Cse2 dimer, we further determined a 3.2 Å structure where the non-target protospacer is 22-nt longer. Although most of the additional residues remain unresolved, density clearly reveals that the non-target strand residues take the downward trench route towards the backside of the Cse2.1 dimer, attracted by the favorable electrostatic environment therein (Fig. 3c–e). The non-target strand sequestration likely corresponds to the extra “locking” step after most R-loop forms, as a mechanism to prevent the R-loop collapse17. In summary, our structural analysis provides important insights about the PAM-dependent directional R-loop formation process17 (Fig. 4). Recognition of an interference PAM by Cascade coincides with the wedge-mediated displacement of the first two target strand nucleotides, initiating DNA unwinding. The directional DNA melting ensures the ordered guidance of the non-target DNA strand ~25Å away from the target strand, as a mechanism to stabilize the seed bubble. Further R-loop propagation leads to non-target strand sequestration behind Cse2 dimer, locking the R-loop in place. Conformational changes accompany the process and reorganize the Cse1 surface, paving the way for Cas3 binding. The active guidance of the non-target DNA strand is a theme not observed in Cas9-DNA structures19,20. It rationalizes the observation that the Cascade-bound R-loop is significantly more stable than that by Cas917. Besides five interference PAMs, Twenty-one other PAMs stimulate the “primed adaptation” in E. coli22, where Cascade and Cas3 actively recruit the Cas1/Cas2 spacer acquisition complex29,30. Priming PAMs may lead to suboptimal Cascade contacts and non-canonical R-loops. Such R-loops are difficult to form and may not be completely unwound18, requiring higher Cascade concentration30 and favorable DNA torque17. They also fail to directly recruit Cas330, which may indicate that the non-target DNA strand is misguided, as Cas3 recruitment is contingent upon non-target strand contact as well25. Overall, our work provides a framework for future studies to better understand the interference and primed adaptation mechanisms in Type I CRISPR-Cas systems. METHODS Expression and purification of Cascade, Cascade-dsDNA complex and Cas3 Sequence information for primers and the synthetic CRISPR expression cassette can be found in Extended Data Table 1. Cascade expression was similar to26. Briefly, cse1 was PCR amplified from E. coli K12 genomic DNA and cloned into pRSF-Duet-ORF1 vector (KanR), between Ncol and NotI restriction sites (Extended Data Table 1). The cse2-cas7-cas5e-cas6e sub-operon was cloned into pET52b (AmpR) between Ncol and NotI; as a Precission cleavable His6 fusion at the N-terminal of cse2. The pre-crRNA expression cassette was synthesized by Life Technologies and cloned into the pHSG-398 vector (CamR) (Extended Data Table 1). E. col BL21 (DE3) star cells containing the 3 plasmids were grown in LB medium at 37 °C to O.D. 600 of 0.6. Cascade expression was induced by the addition of 0.5 mM isopropyl-β-D-1-thiogalactopyranoside (IPTG) and cells were further cultured at 20 °C for 12 hr. The cells ware disrupted by sonication in buffer A (50mM Tris pH 8.0, 20mM imidazole, 300 mM NaCl), loaded to Ni-NTA column, and eluted with buffer A supplemented with 300 mM imidazole. The His6 tag was cleaved by Precision protease, and back-adsorbed with a second Ni-NTA column binding step. Cascade was concentrated and buffer exchanged into buffer B (20 mM Tris pH 7.5, 150 mM NaCl, 2 mM DTT), and further purified on Superdex 200 prep grade column (GE healthcare). Free-Cascade containing fractions were pooled, concentrated to 15 mg/mL, flash-frozen, and stored at −80 °C. Cascade-dsDNA complex was prepared by mixing free-Cascade with dsDNA R-loop mimicking substrate. DNA substrates were chemically synthesized from IDT (Extended Data Table 1). The non-target strand was annealed with the target strand at a 1.5:1 molar ratio. The resulting R-loop mimicking substrate was mixed with Cascade at a 2:1 molar ratio, incubated at room temperature for 30 minutes, and re-purified on Superdex 200. Cascade-dsDNA complex fractions were pooled and utilized in crystallization trials. The Cascade-dsDNA complex containing the 32-nt non-target strand overhang was also obtained using the above protocol, except the His tag was not cleaved. Cascade mutants were constructed with site-directed mutagenesis, and purified using the same method as free-Cascade, except that the N-terminal His­­6 tag was left intact. Cascade integrity was checked using SDS-PAGE (Extended Data Fig. 4b). The Cas3 gene was amplified from E. coli K12 genomic DNA and cloned between BamHI and Xhol into the pET28a-SUMO plasmid. E. coli BL21 (DE3) star cells were grown in LB medium at 20 °C to O.D.600 of 0.3, induced with 0.2 mM IPTG, and further cultured at 20 °C for 12 hr. The Ni-NTA and SEC purification procedures were similar to the procedure mentioned above. The monomeric SUMO-Cas3 fractions were pooled, concentrated to 2 mg/mL, flash-frozen and stored at -80 °C until usage in biochemical assays. Crystallization and structure determination of Cascade/partial R-loop complex Cascade complex crystals were grown using the hanging drop vapor-diffusion method by mixing 2 μL of purified Cascade-ds-DNA complex (15 mg/mL) with 2 μL of mother liquor (1.6 M Na/K Phosphate pH 6.2) at 18 °C. Initial crystals appeared after 2 weeks and grew to full size after ~6 weeks. Crystals were cryoprotected in motherliquor supplemented with 20% ethylene glycol and flash frozen in liquid nitrogen. Diffraction data were collected at Advanced Photon Source – NECAT beamline 24-ID-C and were processed with HKL200031. The structure was solved by molecular replacement with PDB: 4QYZ as the search model. Iterative model building and refinement was conducted with COOT32 and PHENIX33. A summary of the diffraction and refinement statistics can be found in Extended Data Table 2. The Ramachandran plot for the Cascade-dsDNA 10-nt overhang structure indicated 96.65% of residues in the favored region, 3.10% allowed, and 0.25% outliers. The Ramachandran plot for the Cascade ds-DNA 32-nt overhang structure indicated 94.56% of residues in the favored region, 4.81% allowed, and 0.63% outliers. Figures were generated using Pymol34 and CCP4mg35. Electrophoretic mobility shift assay Fluorescent dsDNA target substrates were generated for biochemical assays. The crRNA-matching targets with varied PAMs were cloned into pCDF-Duet between the Pstl and Ncol sites Extended Data Table 1. The dsDNA1 and dsDNA2 substrates were PCR amplified from the plasmid using the indicated fluorescent oligonucleotides (5′ 6-FAM for the non-target strand and 5′ Cy5 for the target strand). The dsDNA3 substrates were prepared by oligonucleotide annealing. All dsDNA substrates were subsequently gel-purified. The dsDNA1 substrate (5′ATG-PAM) was used for all main-text EMSA and Cas3 cleavage assays. The dsDNA2 substrates were used for the experiments shown in Extended Data Figures 6b, 6c, and 8c. The dsDNA3 substrates were used in the experiments shown in Extended Data Figures 6a, 8b, 8d. DNA binding was conducted in 20 mM Tris pH 7.5, 150 mM NaCl, 5% glycerol. The dsDNA substrate concentration was held constant at 3 nM and Cascade concentration was titrated as indicated. The Cascade and dsDNA were incubated at 37 °C for 30 minutes in 20 mM Tris pH 7.5, 150 mM NaCl, 5% glycerol. EMSA was carried out at 4 °C on 2% agarose gels. Fluorescent signals were scanned using a Typhoon 9200 scanner. Cascade-mediated Cas3 DNA cleavage assay Cascade-R-loops were pre-formed by mixing 40 nM Cascade or Cascade mutants with 6 nM fluorescent dsDNA target in binding buffer (5 mM HEPES pH 7.5, 60 mM KCl) at 37 °C for 30 minutes. Cascade-R-loops were then mixed with 500 nM SUMO-Cas3 in DNA cleavage RXN buffer (5 mM HEPES pH 7.5, 60 mM KCl, 10 mM MgCl2, 10 μM CoCl2). Either 2 mM ATP or 2 mM AMPPNP was added and the reaction was incubated at 37 °C for 30 minutes. Cy5 and 6-FAM fluorescent signals were recorded by Typhoon 9200 scanner. The wild-type E. coli Cascade specifically nicked the non-target DNA strand ~10–12 nt into the R-loop region in the presence of a non-hydrolyzable ATP analog, AMPPNP. Addition of ATP triggered processive degradation of the non-target DNA strand and distributive degradation of the target strand upstream of the R-loop. These results are consistent with previous studies11,21. Extended Data Extended Data Figure 1 Electron density of nucleic acids in Cascade-dsDNA complex structure 2FO-FC electron density map (1.0 σ). Nucleic acid strands are shown as sticks and colored as previously indicated. Extended Data Figure 2 Comparison between the Cascade-partial R-loop crystal structure and the Cascade-full R-loop EM reconstruction Rigid body docking of the partial R-loop Cascade crystal structure into the EM reconstruction (EMD-5929) of the full R-loop Cascade illustrates a similar overall conformation, with a correlation value of 0.83. Extended Data Figure 3 Conformational changes in Cascade upon partial R-loop formation Comparison of free- (PDB: 4TVX; in a darker shade) and partial R-loop bound Cascade (this study, in lighter shade). Arrows indicate the direction of the movement. The C-terminal domain (CTD) of Cse1 pivots 30° about a hinge at the Cse1 NTD-CTD interface. The amplified motion at the tip of the Cse1-CTD slides with the Cse2 dimer ~12 Å relative to the Cas7 scaffold, and protrudes upwards into the R-loop. Cse1 NTD remains in the docked position, stabilized by the clamping of Cse1 L1 loop onto the exposed tri-nucleotide motif on crRNA 5′-handle. Extended Data Figure 4 Local conformational rearrangement in Cse1 at the proposed Cas3 binding site a, Vector map showing global and local conformational changes in Cse1. Cse1-NTD (in cyan) undergoes moderate movement. Cse1-CTD (in green) swings about a pivoting point (W307) as part of the global conformational changes, The Cse1 NTD/CTD interface (in orange) undergoes significant local conformational rearrangement upon partial R-loop formation. b, SDS-PAGE of wild type and mutant Cascades used in mutagenesis. All mutants contained stoichiometric amounts of Cse1, except W307A, in which Cse1 failed to assemble. c, Detailed structure rearrangement at Cse1 NTD/CTD interface, involving amino acids 300–326. Different colors are used to differentiate structural elements (amino acid numbers tabulated to the right side). Partial R-loop and free-Cascade structures are rendered in solid and semi-transparent cartoons, respectively. d, W307 from Cse1-NTD rotates inside a hydrophobic socket at Cse1-CTD during the conformational change (free-Cascade-CTD rendered in a darker shade of green). Disruption of this interaction (W307A) caused Cse1 dissociation from Cascade. e, Docking of our structure into the cryo-EM reconstruction of the crosslinked Cascade-dsDNA-Cas3 complex (EMD-5930). Cas3 density and the location of the local rearrangement in Cse1 are circled to highlight their proximity. Extended Data Figure 5 Guided entry of dsDNA into Cascade A series of positively charged residues from Cas7.5, Cas7.6, Cas5e and Cse1 (dark blue surfaces) guide the dsDNA into Cascade and towards PAM recognition elements in Cse1-NTD. Extended Data Figure 6 Influence of PAM-1 base-pair composition on Cascade binding affinity and Cas3 cleavage efficiency a, Inosine substitution of GNT-1 led to a mild 2-fold binding defect, suggesting the N2 amine of GNT-1 is a minor determinant of specificity. b, Binding Kd of E. coli Cascade for all four base-pair combinations at PAM-1. Kd for 5′-ATG, ATA, ATC, and ATT PAMs were determined to be ~10, 20, 40, and 100 nM, respectively. c, Cas3 cleavage efficiency was governed by Cascade’s affinity for the corresponding PAM-containing target (5′-ATG>ATA≫ATT≈ATC). Cascade concentration above the Kd led to efficient Cas3 cleavage. Extended Data Figure 7 Conservation of PAM recognition elements a, Sequence alignment of Cse1 proteins with known structures. Identical residues are highlighted in red, conserved residues in red text. DNA-contacting residues in the Cse1-NTD and CTD are marked by cyan and green asterisks, respectively. The glycine loop and glutamine wedge are in blue and orange boxes, respectively. Residues mediating the W307 ball-and-socket interaction are marked with black asterisks. b, Structural alignment of Cse1 PAM recognition elements. Apo Cse1 from T. thermophillus (PDB: 4AN8), A. ferrooxidans (PDB: 4H3T) and T. fusca (PDB: 3WVO) were superimposed with the Cse1 in our partial R-loop forming E. coli Cascade structure. Despite sequence variation, a glycine-rich loop is present in each Cse1 structure, and likely plays a similar function to recognize PAM from the minor groove (left inset). The glutamine-wedge protrusion is highly conserved in 3-D. Each wedge features a long side chain at the tip (right inset), which likely stacks underneath PAM in a similar fashion. Extended Data Figure 8 Rationalization of PAM-2 and PAM-3 specificity using nucleotide substitution and modeling a, Modeling of alternative base-pairs at PAM-2 suggests that only the N2 amine of GT-2 would cause steric clashes with Cα of G160 (lower right quadrant), this amine therefore may serve as the anti-determinant for the rejection of GT-CNT at PAM-2. b, EMSAs demonstrating that removal of this amine in inosine substitution rescued the Cascade binding defect. c. Whereas K268A contained reduced affinity for the correct PAM, it still possessed strong discrimination against GT-CNT at PAM-3 (5′-CTG), suggesting the further presence of a mechanism to reject GT-3. d, Inosine substitution of GT-3 restored the Cascade binding Kd to ~40 nM, leading to the conclusion that the N2 amine of GT-3 is a minor determinant of specificity. Extended Data Table 1 Oligonucleotides used in this study. Oligonucleotide Sequence (5′−3′) Notes Cse1 (forward) GCGCGCCATGGCTAATTTGCTTATTGATAACTGGATCC Ncol + Cse1 fragment (pRSF-Duet ORF1) Cse1 (reverse) GGCCCGCGGCCGCTCAGCCATTTGATGGCCCTCC Cse1 fragment + NotI (pRSF-Duet ORF1) Cse2-Cas7-Cas5e-Cas6e (forward) GCGCGGGTACCAGATGGCTGATGAAATTGATGCA Kpnl + Cse2-Cas7-Cas5e-Cas6e fragment (pET52b) Cse2-Cas7-Cas5e-Cas6e (reverse) CGCGCGCGGCCGCTCACAGTGGAGCCAAAGATAG Cse2-Cas7-Cas5e-Cas6e fragment + NotI (pET-52b) Cse2-Cas7-Cas5e-Cas6e (forward) GCGCGCCATGGGTCATCACCACCATCATCACGGTGCACTTGAAGTCCTCTTTC Ncol + 6XHIS + Precission (pET-52b) R-loop mimic (target strand) CTGTTGGCAAGCCAGGATCTGAACAATACCGTCATCGAGCACTGCACAGA R-loop mimic (non-target strand) TCTGTGCAGTGCTCGATGTTTTATTTAT SUMO-Cas3 (forward) GCGCCGCGGATCCATGGAACCTTTTAAATATATATGCCAT BamHI + Cas3 fragment (pET28b-SUMO vector) SUMO-Cas3 (reverse) GCGCCGCCTCGAGTTATTTGGGATTTGCAGGGAT Cas3 fragment + Xhol (pET28b-SUMO vector) Target plasmid construction (non-target strand) CATGG ATGACGGTATTGTTCAGATCCTGGCTTGCCAACAGCTGCA Ncol overhang + Target sequence + Pstl overhang (pCDF-Duet1) PAM altered as indicated in text Target plasmid construction (target strand) GCTGTTGGCAAGCCAGGATCTGAACAATACCGT CATC Pstl + non-target sequence + Ncol (pCDF-Duet1) PAM altered as indicated in text. dsDNA1 Fluorescent ATG PAM substrate (forward) AACTTTAATAAGGAGATATACCATGGATG 5′–6-FAM label, PCR product used in main text EMSA and Cas3 cleavage assays. dsDNA1 Fluorescent ATG PAM substrate (reverse) GCGGCCGCAAGCTTGTC 5′- Cy5 label, PCR product used in main text EMSA and Cas3 cleavage assays. dsDNA2 Fluorescent substrate (forward) TAATACGACTCACTATAGGG T7 Forward 5′–6-FAM label. Used with dsDNA1 (reverse) primer to amplify set of 5′-ATA, ATT, ATC, ATG, AGG and CTG PAM substrates from target plasmids. dsDNA3 (non-target strand) AGATATACATGG ATGACGGTATTGTTCAGATCCTGGCTTGCCAACAGCTGCAGGTCGACA 5′ 6-FAM or HEX label, PAM altered (base mutation or inosine substitution) as indicated in text. dsDNA3 (target strand) TGTCGACCTGCAGCTGTTGGCAAGCCAGGATCTGAACAATACCGT CATCCATGTATATCT PAM altered (base mutation or inosine substitution) as indicated in text. crRNA expression cassette GCGCCGGGAATTCCCTGCATTAGG TAATACGACTCACTATAGG GAGTTCCCCGCGCCAGCGGGGATAAACCGACGGTATTGTTCAGATCCTGGCTTGCCAACAG GAGTTCCCCGCGCCAGCGGGGATAAACCGACGGTATTGTTCAGATCCTGGCTTGCCAACAG GAGTTCCCCGCGCCAGCGGGGATAAACCGACGGTATTGTTCAGATCCTGGCTTGCCAACAG GAGTTCCCCGCGCCAGCGGGGATAAACCGACGGTATTGTTCAGATCCTGGCTTGCCAACAG GAGTTCCCCGCGCCAGCGGGGATAAACCGACGGTATTGTTCAGATCCTGGCTTGCCAACAG GAGTTCCCCGCGCCAGCGGGGATAAACCGACGGTATTGTTCAGATCCTGGCTTGCCAACAG GAGTTCCCCGCGCCAGCGGGGATTTTGCC CTAGCATAACCCCTTGGGGCCTCTAAACGGGTCTTGAGGGGTTTTTTGGACCGGATCCCCG The crRNA expression cassette containing six identical repeat-spacer units was de novo synthesized and inserted to the pHSG-398 vector. The T7 promoter, T7 terminator, and the repeats are highlighted green, red, and yellow, respectively. Extended Data Table 2 Data collection and refinement statistics. Cascade-dsDNA Cascade-dsDNA (32-nt non-target spacer) Data collection Space group P2 2121 P 2 21 21 Cell dimensions  a, b, c (Å) 92.98, 150.06,400.55 92.81,149.83,404.10  α, β, γ (°) 90, 90, 90 90, 90, 90 Resolution (Å) 50.0–2.45 (2.49−2.45) 50.0–3.20 (3.26−3.21)* R merge 0.169 (1.288) 0.265 (1.066) Rpim 0.050 (0.531) 0.143 (0.667) I/σI 12.9 (1.1) 5.8 (1.5) Completeness (%) 99.7 (97.2) 96.8 (84.1) Redundancy 12.1 (6.5) 3.6 (3.0) CC(l/2) 0.997 (0.496) 0.967 (0.422) Refinement Resolution (Å) 50.0–2.45 50.0–3.21 No. reflections 205,381 90,215 Rwork/Rfree 0.2058/0.2327 0.2087/0.2470 No. atoms  Macromolecule 27896 27237  Ligand/ion 1(Zn) 1(Zn)  Water 1051 0 B-factors  Macromolecule 53.60 62.10  Ligand/ion 51.90 61.50  Water 45.90 N/A R.m.s. deviations  Bond lengths (Å) 0.013 0.018  Bond angles (°) 0.840 1.010 * One crystal was used for each structure. * Values in parentheses are for highest-resolution shell. Supplementary Material Extended_video_1 Extended_video_2 Extended_video_3 This work is supported by NIH grants GM102543 and GM086766 to A.K, GM097330 to S.B. and GM108888 to B.W. NE-CAT beamlines were supported by NIH grants P41 GM103403 and S10 RR029205. We thank Gerald Feigenson and John Mallon for technical help, Ilya Finkelstein and Ian Price for helpful discussions. Figure 1 Foreign DNA bound Cascade structure a, Arrangement of I-E CRISPR-cas locus in the E. coli K12 genome. The color schemes are preserved throughout all figures. b, Nucleic acid sequences and c, overall views of foreign DNA-bound Cascade structure. d, Entry of dsDNA between Cas7.5 and Cas 7.6, PAM recognition by three structural elements in Cse1-NTD (magenta), and partial R-loop underneath PAM. e, Schematic of Cascade-DNA contacts around PAM regions. Hydrogen bonds and electrostatic contacts as dashed lines, hydrophobic interactions as solid lines, and waters as blue circles. f, Modeling of Cascade sampling B-form dsDNA for PAM. Figure 2 PAM recognition by Cse1 subunit of Cascade a, b, Detailed views featuring PAM recognition by the glutamine-wedge (and its involvement in target DNA strand displacement), glycine-loop, and lysine-finger of Cse1. The 2FO-FC electron density map is displayed at 1.5 σ. c, Summarization of the five interference PAMs in E. coli Type I-E CRISPR system, with the observed Cascade contacts marked. d, Top-down views of PAM recognition at each base-pair. Left: stringent recognition of PAM-1. Middle: shape complementarity to PAM-2. Right: Electrostatic contacts to PAM-3 and -4. e, Mutagenesis assayed by Cascade-binding (EMSA) and f, Cas3-mediated DNA cleavage to evaluate the observed PAM contacts. Figure 3 Active guidance of the non-target DNA strand by Cascade a, Sequence nonspecific electrostatic contacts guide the 10-nt non-target strand (red sticks) protospacer along Cse1-NTD surface (cyan), and across a binding site on Cse1-CTD (green). Y397 and L481 from Cse1-CTD further interdigitate between di-nucleotide stacks. b, EMSA evaluating non-target contacts by Cascade. c, d, Nucleic acid sequence and Zoom-in view of the 3.2 Å structure of Cascade programmed with a 22-nt longer non-target strand. The 2FO-FC electron density map at 1.0 σ clearly reveals that the longer non-target DNA strand takes the trench route. Inset illustrates the favorable electrostatic surface. Figure 4 Model for PAM-dependent directional R-loop formation in Type I CRISPR-Cas system a, Cascade samples dsDNA minor groove for various sequence combinations (yellow sticks). b, Left: interference and priming PAMs lead to longer dwell time and local DNA bending. Middle: Interference PAMs allow optimal minor groove interaction, which is coupled with the glutamine wedge insertion and disruption of the first 2-bp of protospacer. Right: Directional DNA melting leads to segmented DNA/crRNA duplex formation at the target strand side, and favorable sugar-phosphate contacts to the non-target side, leading to seed bubble stabilization. c, Further DNA unwinding leads to the non-target strand sequestration to the backside of Cse2 dimer, locking R-loop in place. d. R-loop formation is accompanied by a pivoting motion in Cse1-CTD and a sliding motion in Cse2 dimer. Local rearrangement occurs in Cse1 (depicted as a flag), licensing Cas3 recruitment. Supplementary Information Three supplementary movies. Author Contributions R.P.H. and A.K. designed the research, R.P.H. determined the structure and performed biochemical analyses, Y.X., F.D., and P.E. contributed to biochemical analysis, K.R. assisted with diffraction data collection and processing, B.W. and S.B. contributed to assay setup. R.P.H., B.W., and A.K. wrote the manuscript. The structure factors and coordinates for Cascade/partial R-loop structures with 10-nt and 35-nt non-target protospacer have been deposited in the Protein Data Bank under accession numbers 5H9F and 5H9E, respectively. The authors declare no competing financial interests. 1 van der Oost J Westra ER Jackson RN Wiedenheft B Unravelling the structural and mechanistic basis of CRISPR-Cas systems Nature reviews. Microbiology 12 479 492 2014 24909109 2 Luo ML Mullis AS Leenay RT Beisel CL Repurposing endogenous type I CRISPR-Cas systems for programmable gene repression Nucleic acids research 43 674 681 2015 25326321 3 Caliando BJ Voigt CA Targeted DNA degradation using a CRISPR device stably carried in the host genome Nature communications 6 6989 2015 4 Westra ER CRISPR immunity relies on the consecutive binding and degradation of negatively supercoiled invader DNA by Cascade and Cas3 Molecular cell 46 595 605 2012 22521689 5 Jore MM Structural basis for CRISPR RNA-guided DNA recognition by Cascade Nature structural & molecular biology 18 529 536 2011 6 Brouns SJ Small CRISPR RNAs guide antiviral defense in prokaryotes Science 321 960 964 2008 18703739 7 Wiedenheft B Structures of the RNA-guided surveillance complex from a bacterial immune system Nature 477 486 489 2011 21938068 8 Westra ER Type I-E CRISPR-cas systems discriminate target from non-target DNA through base pairing-independent PAM recognition PLoS genetics 9 e1003742 2013 24039596 9 Marraffini LA Sontheimer EJ Self versus non-self discrimination during CRISPR RNA-directed immunity Nature 463 568 571 2010 20072129 10 Mojica FJ Diez-Villasenor C Garcia-Martinez J Almendros C Short motif sequences determine the targets of the prokaryotic CRISPR defence system Microbiology 155 733 740 2009 19246744 11 Mulepati S Bailey S In vitro reconstitution of an Escherichia coli RNA-guided immune system reveals unidirectional, ATP-dependent degradation of DNA target The Journal of biological chemistry 288 22184 22192 2013 23760266 12 Rollins MF Schuman JT Paulus K Bukhari HS Wiedenheft B Mechanism of foreign DNA recognition by a CRISPR RNA-guided surveillance complex from Pseudomonas aeruginosa Nucleic acids research 43 2216 2222 2015 25662606 13 Sashital Dipali G Wiedenheft B Doudna Jennifer A Mechanism of Foreign DNA Selection in a Bacterial Adaptive Immune System Molecular cell 46 606 615 2012 22521690 14 Sinkunas T In vitro reconstitution of Cascade-mediated CRISPR immunity in Streptococcus thermophilus The EMBO journal 32 385 394 2013 23334296 15 Sternberg SH Redding S Jinek M Greene EC Doudna JA DNA interrogation by the CRISPR RNA-guided endonuclease Cas9 Nature 507 62 67 2014 24476820 16 van Erp PB Mechanism of CRISPR-RNA guided recognition of DNA targets in Escherichia coli Nucleic acids research 2015 17 Rutkauskas M Directional R-Loop Formation by the CRISPR-Cas Surveillance Complex Cascade Provides Efficient Off-Target Site Rejection Cell reports 2015 18 Blosser TR Two distinct DNA binding modes guide dual roles of a CRISPR-Cas protein complex Molecular cell 58 60 70 2015 25752578 19 Anders C Niewoehner O Duerst A Jinek M Structural basis of PAM-dependent target DNA recognition by the Cas9 endonuclease Nature 513 569 573 2014 25079318 20 Nishimasu H Crystal Structure of Staphylococcus aureus Cas9 Cell 162 1113 1126 2015 26317473 21 Hochstrasser ML CasA mediates Cas3-catalyzed target degradation during CRISPR RNA-guided interference Proceedings of the National Academy of Sciences of the United States of America 111 6618 6623 2014 24748111 22 Fineran PC Degenerate target sites mediate rapid primed CRISPR adaptation Proceedings of the National Academy of Sciences of the United States of America 111 E1629 1638 2014 24711427 23 Zhao H Crystal structure of the RNA-guided immune surveillance Cascade complex in Escherichia coli Nature 515 147 150 2014 25118175 24 Jackson RN Structural biology. Crystal structure of the CRISPR RNA-guided surveillance complex from Escherichia coli Science 345 1473 1479 2014 25103409 25 Huo Y Structures of CRISPR Cas3 offer mechanistic insights into Cascade-activated DNA unwinding and degradation Nature structural & molecular biology 21 771 777 2014 26 Jackson RN Crystal structure of the CRISPR RNA-guided surveillance complex from Escherichia coli Science 345 1473 1479 2014 25103409 27 Mulepati S Heroux A Bailey S Crystal structure of a CRISPR RNA-guided surveillance complex bound to a ssDNA target Science 345 1479 1484 2014 25123481 28 Xue C CRISPR interference and priming varies with individual spacer sequences Nucleic acids research 2015 29 Datsenko KA Molecular memory of prior infections activates the CRISPR/Cas adaptive bacterial immunity system Nature communications 3 945 2012 30 Redding S Surveillance and Processing of Foreign DNA by the Escherichia coli CRISPR-Cas System Cell 2015 31 Otwinowski Z Minor W Processing of X-ray Diffraction Data Collected in Oscillation Mode Methods in Enzymology 276 307 326 1997 32 Emsley P Cowtan K Coot: model-building tools for molecular graphics Acta crystallographica. Section D, Biological crystallography 60 2126 2132 2004 15572765 33 Adams PD PHENIX: a comprehensive Python-based system for macromolecular structure solution Acta crystallographica. Section D, Biological crystallography 66 213 221 2010 20124702 34 Schrodinger LLC The PyMOL Molecular Graphics System, Version 1.3r1 2010 35 Collaborative Computational Project, N The CCP4 suite: programs for protein crystallography Acta crystallographica. Section D, Biological crystallography 50 760 763 1994 15299374

PMC005xxxxxx/PMC5134422.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 101465514 34788 Mol Plant Mol Plant Molecular plant 1674-2052 1752-9867 26002145 5134422 10.1016/j.molp.2015.05.006 NIHMS741613 Article Live Cell Imaging with R-GECO1 Sheds Light on flg22- and Chitin-Induced Transient [Ca2+]cyt Patterns in Arabidopsis Keinath Nana F. 14 Waadt Rainer 134 Brugman Rik 2 Schroeder Julian I. 3 Grossmann Guido 2 Schumacher Karin 1 Krebs Melanie 1* 1 Centre for Organismal Studies, Plant Developmental Biology, University of Heidelberg, Im Neuenheimer Feld 230, 69120 Heidelberg, Germany 2 Centre for Organismal Studies, University of Heidelberg, 69120 Heidelberg, Germany 3 Division of Biological Sciences, Cell and Developmental Biology Section, University of California San Diego, 92093 La Jolla, USA * Correspondence: Melanie Krebs (melanie.krebs@cos.uni-heidelberg.de) 4 These authors contributed equally to this article. 28 11 2016 19 5 2015 8 2015 02 12 2016 8 8 11881200 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. Intracellular Ca2+ transients are an integral part of the signaling cascade during pathogen-associated molecular pattern (PAMP)-triggered immunity in plants. Yet, our knowledge about the spatial distribution of PAMP-induced Ca2+ signals is limited. Investigation of cell- and tissue-specific properties of Ca2+-dependent signaling processes requires versatile Ca2+ reporters that are able to extract spatial information from cellular and subcellular structures, as well as from whole tissues over time periods from seconds to hours. Fluorescence-based reporters cover both a broad spatial and temporal range, which makes them ideally suited to study Ca2+ signaling in living cells. In this study, we compared two fluorescence-based Ca2+ sensors: the Förster resonance energy transfer (FRET)-based reporter yellow cameleon NES-YC3.6 and the intensity-based sensor R-GECO1. We demonstrate that R-GECO1 exhibits a significantly increased signal change compared with ratiometric NES-YC3.6 in response to several stimuli. Due to its superior sensitivity, R-GECO1 is able to report flg22- and chitin-induced Ca2+ signals on a cellular scale, which allowed identification of defined [Ca2+]cyt oscillations in epidermal and guard cells in response to the fungal elicitor chitin. Moreover, we discovered that flg22- and chitin-induced Ca2+ signals in the root initiate from the elongation zone. calcium imaging R-GECO1 flg22 chitin sensor Arabidopsis Introduction Ca2+ as a second messenger is of fundamental importance for all eukaryotic organisms. In plants, Ca2+-mediated signaling participates in the regulation of abiotic stress responses as well as in signal transduction during interaction with symbiotic microorganisms and pathogens (Dodd et al., 2010). Ca2+ signaling is characterized by a transient increase of cytosolic Ca2+ ([Ca2+]cyt), which is accomplished by the activity of transporters and channels that allow influx of Ca2+ from intra- and extracellular stores (Dodd et al., 2010). These transient [Ca2+]cyt changes are then sensed and decoded by a complex network of Ca2+-binding proteins and interacting kinases that transduce the information downstream to regulate effector proteins and gene expression (Luan, 2009; Boudsocq and Sheen, 2010; Romeis and Herde, 2014; Steinhorst and Kudla, 2014). It is thought that specificity in Ca2+ signaling is primarily conferred through spatial and temporal qualities of the Ca2+ signal, which differs in frequency, amplitude, and duration in a stimulus-dependent manner (McAinsh and Pittman, 2009). Genetically encoded indicators (GECIs) are ideal tools to study transient [Ca2+]cyt patterns, as they enable non-invasive monitoring of [Ca2+]cyt dynamics in living cells (Monshausen, 2012; Pérez Koldenkova and Nagai, 2013). In plants, luminescent aequorin-based Ca2+ reporters have been widely used to investigate the role of Ca2+ signaling during abiotic stress responses, such as cold, osmotic, and salt stress (Knight et al., 1991, 1996; Kiegle et al., 2000; Martí et al., 2013), and to study Ca2+ signaling during plant defense reactions (Blume et al., 2000; Lecourieux et al., 2002; Kwaaitaal et al., 2011; Ranf et al., 2011). Despite its high sensitivity, aequorin has a low quantum yield, which limits its use for live cell imaging (Brini, 2008). To overcome the limitation of low brightness, aequorin was fused to green fluorescent protein (GFP) to create a bioluminescence resonance energy transfer (BRET) sensor. Here, the luminescence of aequorin is transferred via FRET to GFP, which in turn emits light of higher quantum yield (Baubet et al., 2000). Recently, this concept has been applied in Arabidopsis to measure [Ca2+]cyt dynamics in entire seedlings and leaves (Xiong et al., 2014). However, there are two issues that remain limiting when using aequorin Ca2+ sensors. First, there is a need for co-factor loading, which is time consuming and might therefore interfere with the timing of experiments. Second, commonly available devices for photon counting usually do not resolve cellular and subcellular structures, resulting in compromised spatial resolution. However, the analyses of molecular mechanisms behind Ca2+ signaling networks requires versatile Ca2+ reporter tools that cover a broad range of resolution in time and space. Because of their higher quantum yield, fluorescence-based GECIs are able to fulfill both criteria. Imaging periods from seconds to hours that cover from subcellular structures to whole organisms are feasible. Fluorescence-based GECIs can be categorized in two classes: ratiometric FRET-based reporters and intensiometric reporters. Yellow cameleon (YC) Ca2+ sensors belong to the group of ratiometric FRET-based indicators (Miyawaki et al., 1997). They consist of an enhanced cyan fluorescent protein (ECFP) variant as FRET donor and a yellow fluorescent protein (YFP) variant as acceptor. Both fluorophores are linked via a Ca2+ sensory domain that is composed of calmodulin (CaM), a short linker, and the M13 peptide, which binds to CaM in a Ca2+-dependent manner (Ikura et al., 1992; Porumb et al., 1994). Upon Ca2+ binding, the sensory domain structurally rearranges, thereby changing the distance and orientation of the fluorescent proteins to increase FRET efficiency (Miyawaki et al., 1997). Yellow cameleons have been extensively engineered to improve brightness and dynamic range (Nagai et al., 2004), minimize interference with endogenous Ca2+ signaling (Palmer and Tsien, 2006), and modulate Ca2+ affinities (Horikawa et al., 2010). Additional FRET-based Ca2+ reporters that use the troponin C Ca2+-binding moiety as sensory domain have been developed (Heim and Griesbeck, 2004; Mank et al., 2008; Thestrup et al., 2014). In plants, FRET-based Ca2+ reporters have advanced our understanding of Ca2+ signaling on the cellular and even subcellular scale (Choi et al., 2012; Monshausen, 2012 and references therein). Currently, yellow cameleon YC3.6 (Nagai et al., 2004) is the most commonly used fluorescent Ca2+ reporter in plants (Choi et al., 2012). Intensiometric Ca2+ reporters with Pericam and GCaMPas prototypes are designed with a single circularly permutated green fluorescent protein (cpGFP) that is flanked by an N-terminal CaM-binding M13 peptide and by a C-terminal CaM (Nagai et al., 2001; Nakai et al., 2001). Upon Ca2+ binding, the CaM domain folds around the M13 peptide, thereby altering the chemical environment of the chromophore, leading to enhanced fluorescence emission (Akerboom et al., 2009). These reporters have also been extensively engineered to improve stability, sensitivity, and dynamic range resulting in various generations of GCaMPs and the development of G-GECOs (Akerboom et al., 2009; Muto et al., 2011; Zhao et al., 2011; Akerboom et al., 2012). To enable multi-color Ca2+ imaging, blue and yellow shifted intensity-based Ca2+-reporters have been developed by mutagenesis of G-GECO1.1 and GCaMP3 (Zhao et al., 2011; Akerboom et al., 2013). Red and orange emitting reporters were generated by replacing the cpGFP with cpmApple to generate R- and O-GECOs (Zhao et al., 2011; Wu et al., 2013) or with cpmRuby to generate RCaMPs (Akerboom et al., 2013). Among these reporters, R-GECO1 was the first red-shifted intensity-based Ca2+ reporter; it was combined with FRET-based reporters to perform multi-compartment and multi-parameter imaging (Zhao et al., 2011). In Arabidopsis, YC3.6 was used in combination with R-GECO1 to resolve cell-specific Ca2+ responses in pollen and synergid cells during pollen tube perception, demonstrating the potential of multi-color Ca2+ imaging (Ngo et al., 2014). When compared with ratiometric FRET-based Ca2+ reporters, intensity-based Ca2+ reporters are characterized by a superior dynamic range but are more sensitive to fluctuations in sensor expression and distribution, instrument noise, and motion artifacts (O'Connor and Silver, 2013; Pérez Koldenkova and Nagai, 2013). Therefore, the appropriate Ca2+ reporter should be selected according to the experimental needs. In order to investigate whether R-GECO1 would be useful to facilitate Ca2+ imaging in plants, we created dual expression lines of the fluorescent Ca2+ sensors R-GECO1 and NES-YC3.6, and compared their performance under in vivo conditions. We found that R-GECO1 shows significantly higher signal changes compared with NES-YC3.6 in response to extracellular ATP and plasma membrane hyperpolarization. To make use of the higher sensitivity of R-GECO1, we studied cell- and tissue-specific characteristics of Ca2+ signals elicited by the microbe-derived molecules flg22 and chitin. Flg22 and chitin are conserved PAMPs that are recognized cell autonomously by pattern recognition receptors (Boller and Felix, 2009 and references therein). Within seconds to minutes after PAMP perception, plants respond with a transient increase in [Ca2+]cyt (Blume et al., 2000; Lecourieux et al., 2002; Kwaaitaal et al., 2011; Ranf et al., 2011). Although it is well established that Ca2+ signals are an integral part of the signaling cascade during PAMP-triggered immunity (Boller and Felix, 2009), detailed spatial and temporal information of signal onset and propagation is missing at the cellular level. Here, we report that R-GECO1 is able to resolve flg22 and chitin-induced Ca2+ signals on a cellular scale. We were able to observe defined transient [Ca2+]cyt patterns in leaves as well as in roots of Arabidopsis seedlings. We identified differences in guard cell Ca2+ signaling in response to flg22 that suggest a cell-autonomous perception of flg22 in guard cells. In roots, we identified that flg22 and chitin-induced Ca2+ signals initiate from the elongation zone. Taken together, imaging PAMP-induced Ca2+ signals with R-GECO1 allowed the identification of cell-specific differences in Ca2+ signaling in the leaf, as well as the observation of tissue-specific responses in the root. Results and Discussion In Vivo Comparison of Ca2+ Sensors NES-YC3.6 and R-GECO1 In order to compare in vivo properties of FRET-based and intensity-based Ca2+ reporters, we generated stable transgenic Arabidopsis lines expressing both cytosolic localized NES-YC3.6 (Nagai et al., 2004; Krebs et al., 2012) and cytosolic and nuclear localized R-GECO1 (Zhao et al., 2011). For comparable expression levels, both reporters were expressed under control of the UBQ10 promoter (Norris et al., 1993; Grefen et al., 2010). Fluorescence microscopy revealed that NES-YC3.6 and R-GECO1 were evenly expressed throughout the whole plant including guard cells and pollen (Supplemental Figure 1 and Figure 2). Due to the uniform expression pattern and their distinct spectral properties, we were able to perform comparative in vivo analyses of both Ca2+ reporters. R-GECO1 Is More Sensitive Toward Changes of [Ca2+]cyt than NES-YC3.6 To evaluate sensor performance with respect to dynamic signal change, we determined the maximum signal change and signal-to-noise ratio (SNR) of NES-YC3.6 and R-GECO1 in response to 1 mM ATP (Figure 1) and to plasma membrane hyperpolarization (Supplemental Figure 2). Both stimuli are known to induce fast and robust changes of [Ca2+]cyt (Allen et al., 2000; Tanaka et al., 2010; Choi et al., 2014). To compare the readout of the intensity-based sensor R-GECO1 with the readout of the ratiometric FRET-based sensor NES-YC3.6, fractional fluorescence changes (ΔF/F) and fractional ratio changes (ΔR/R) were calculated respectively. [Ca2+]cyt dynamics were recorded from the root hair zone of 6- to 8-day-old seedlings (Figure 1A). Fluorescence intensity values of a representative measurement are shown in Figure 1B and the respective fractional signal changes of NES-YC3.6 (ΔR/R) and R-GECO1 (ΔF/F) are shown in Figure 1C. We observed that, in response to 1 mM ATP, R-GECO1 showed a 17-fold increased maximum signal change compared with NES-YC3.6 (Figure 1C and 1D). Similarly, when we induced cytosolic Ca2+ transients by plasma membrane hyperpolarization, the maximum signal change for R-GECO1 was 11 times higher compared with NES-YC3.6 (Supplemental Figure 2A, 2B, and 2E). To exclude the possibility that the higher signal changes observed for R-GECO1 are due to an overrepresentation of nuclear Ca2+ changes in R-GECO1 transgenic lines, we excluded plant cell nuclei for data analyses (Supplemental Figure 2C and 2D) and compared the maximum signal change of cytosol and nucleus versus cytosol only (Supplemental Figure 2E). Importantly, the maximum signal change of R-GECO1 after plasma membrane hyperpolarization was not changed (Supplemental Figure 2E), no matter whether nuclei were excluded from the analyses or not. From this we conclude that the improved sensor performance of R-GECO1 is not due to an overrepresentation of nuclear Ca2+ signal changes and it is therefore valid to compare cytosolic localized NES-YC3.6 with cytosolic and nuclear localized R-GECO1. The increased maximum signal amplitude observed for R-GECO1 results in significantly higher SNRs for R-GECO1 compared with NES-YC3.6 (Figure 1E and Supplemental Figure 2F). The baseline noise, indicated by the SD of the baseline, was found to be in a similar range for R-GECO1 and NES-YC3.6 (Figure 1F and Supplemental Figure 2G). Yet, it is known that under certain imaging conditions, intensiometric sensors are more prone to noise, as changes in sensor localization as well as motion artifacts will be directly converted into altered levels of fluorescence intensity that are not correlated with changes of [Ca2+]cyt. At lower magnifications, where sample movement and focus shifts are minimally effective, intensity-based reporters might outperform ratiometric FRET-based reporters because of their higher dynamic range. However, at high magnifications, the use of ratiometric reporters is an advantage as they are intrinsically normalized. Therefore, the design of ratiometric R-GECO1 by fusion to a second fluorescent protein would provide a strategy to further increase the quality of live cell Ca2+ imaging, in particular under unfavorable imaging conditions. Our results show that under in vivo conditions R-GECO1 is 11–17 times more sensitive to [Ca2+]cyt changes compared with the FRET-based Ca2+ sensor NES-YC3.6 (Figure 1 and Supplemental Figure 2), demonstrating its huge potential for Ca2+ imaging in plants. Individual Sensor Performance of NES-YC3.6 and R-GECO1 Is Not Impaired in Dual Expression Lines The two Ca2+ sensors NES-YC3.6 (Kd = 250 nM; Nagai et al., 2004) and R-GECO1 (Kd = 480 nM; Zhao et al., 2011) differ to some extent in their affinity for Ca2+. To rule out that sensor readout is biased in dual expression lines due to a competition for cytosolic Ca2+, we directly compared the Ca2+-dependent signal change of NES-YC3.6 and R-GECO1 in response to plasma membrane hyperpolarization in dual expression and single expression lines (Supplemental Figure 3). A sequence of Ca2+ transients was induced by alternate application of depolarization (Depol) and hyperpolarization buffer (Hyper; Supplemental Figure 3B, 3D, and 3F). We found that the Ca2+-dependent signal change was in a similar range for NES-YC3.6 in the dual and the single expression line (Supplemental Figure 3B and 3D). Similarly, fractional fluorescence changes of R-GECO1 are comparable between the dual and the single expression line (Supplemental Figure 3B and 3F). Quantification of the mean maximum Ca2+-dependent signal change of NES-YC3.6 and R-GECO1 indicates that there is no significant difference between dual and single expression lines (Supplemental Figure 3G). These results demonstrate that the sensor performance of NES-YC3.6 and R-GECO1 is not affected in dual expression lines and therefore results obtained from experiments with dual expression lines are valid and reflect the in vivo properties of the respective sensors. Visualization of Tip-Localized Ca2+ Gradients Root hairs and pollen tubes are experimental systems in which polar cell growth occurs. In both systems, the tip-focused Ca2+ gradient is essential to establish and maintain polar cell growth (Reiss and Herth, 1978; Bibikova et al., 1997; Monshausen et al., 2008; Steinhorst and Kudla, 2013). We acquired ratiometric and fluorescence images of growing pollen tubes (Figure 2A) and root hairs (Figure 2B) expressing NES-YC3.6 and R-GECO1 to visualize the Ca2+ distributions in these cells. Both sensors report the expected [Ca2+]cyt gradients at the tips of pollen tubes (Figure 2A) and root hairs (Figure 2B). To investigate dynamic changes of [Ca2+]cyt, we performed time-lapse imaging in growing root hairs and were able to resolve cytosolic Ca2+ oscillations at the tip of the root hair (Figure 2C and 2D and Supplemental Movie 1). Maxima of [Ca2+]cyt were observed every 24.3 s (SD ± 6.4 s) over a time period of 360 s (Figure 2D). These results are in accordance with a previous study in which the maximum cytosolic Ca2+ signal at the root hair tip was found to oscillate with a frequency of two to four peaks per minute (Monshausen et al., 2008). As R-GECO1 is an intensity-based indicator, Ca2+-induced signal changes might be influenced by sensor distribution or pH oscillations (Bibikova et al., 1998; Feijó et al., 1999). Although the accuracy of intensity-based sensors might not be ideally suited to visualize cellular Ca2+ distributions, our data clearly show that R-GECO1 is able to report tip-localized Ca2+ gradients as well as Ca2+ oscillations during tip growth, which reveals its potential for the investigation of Ca2+-related phenotypes associated with pollen tube growth or root hair development. To increase the precision of R-GECO1, the aforementioned generation of ratiometric versions of R-GECO1 would be desirable. R-GECO1 Is More Sensitive Toward Changes of pH than NES-YC3.6 Wild-type Aequorea victoria GFP and several of its derivates are sensitive to changes in pH (Kneen et al., 1998; Young et al., 2010). Also R-GECO1, derived from circular permutated mApple (Shaner et al., 2008), is pH sensitive in vitro (Zhao et al., 2011). Within a physiological pH range of 6.8–8.0, R-GECO1 fluorescence emission changes by a factor of 4.0 for the Ca2+-free form and by a factor of 1.4 for the Ca2+-bound form, which resembles a change in the dynamic range between 8 and 16 (Zhao et al., 2011). Therefore, we wanted to investigate to what extent intracellular pH changes could bias Ca2+ measurements in Arabidopsis. To compare the pH sensitivity of R-GECO1 and NES-YC3.6, we performed a pH clamp experiment in which we controlled the intracellular pH of seedlings, using pH equilibration buffers (Yoshida, 1994; Krebs et al., 2010), within a physiologically relevant range of pH 6.8–8.0 (Figure 3). In order to distinguish between pH- and Ca2+-induced signal changes, the experiment was carried out in the presence of the Ca2+ channel blocker lanthanum chloride (LaCl3), the extracellular Ca2+ chelater ethylene glycol tetraacetic acid (EGTA), and the intracellular, membrane-permeant Ca2+ chelater 1,2-bis(o-aminophenoxy)ethane-N,N, N′,N′-tetraacetic acid (acetoxymethyl ester) (BAPTA-AM), which have been previously shown to block or attenuate cytosolic Ca2+ transients in Arabidopsis (Knight et al., 1996; Young et al., 2006; Kwaaitaal et al., 2011; Ranf et al., 2011). To demonstrate that pH treatments were effective in planta, we monitored changes in cytosolic pH (pHcyt) in Arabidopsis seedlings expressing pHGFP (Moseyko and Feldman, 2001; Fendrych et al., 2014; Figure 3A). As indicated, ΔR/R of pHGFP increased stepwise with increasing pH values of the equilibration buffers (Figure 3B). We also found that the fluorescence intensity of both Ca2+ sensors was sensitive to intracellular pH changes (Figure 3D). Fluorescence intensities of the NES-YC3.6 fluorophores ECFP and cpVenus increased by a factor of 1.3 from pH 6.8 to 8.0, whereas the emission intensity of R-GECO1 increased by a factor of 2.1 (Figure 3D). Since emission intensities of ECFP and cpVenus increase by the same factor, the resulting emission ratio is not affected by pH fluctuations (Figure 3E). In contrast, the pH-induced increase in fluorescence emission of R-GECO1 is directly translated into increased ΔF/F values (Figure 3E). Under conditions that might involve changes in pHcyt, the pH sensitivity of R-GECO1 has to be taken into account for data interpretation. Based on in vitro data (Zhao et al., 2011), cytoplasmic alkalinization could dampen [Ca2+]cyt responses of R-GECO1 through an increased baseline fluorescence emission (Figure 3E). To distinguish between pH- and Ca2+-induced signal changes, a pH reporter such as pHGFP could be investigated in parallel with R-GECO1. The spectral properties of pHGFP and R-GECO1 allow simultaneous imaging of both reporters. Therefore, dual reporter lines that simultaneously express pHGFP and R-GECO1 would be ideal for the collection and interpretation of data of stimulus-induced [Ca2+]cyt changes that also involve changes in pHcyt. Live Cell Imaging Using R-GECO1 Resolves PAMP-Triggered Single-Cell Ca2+ Transients in Leaves Even though it is well documented that transient increases in [Ca2+]cyt are among the earliest responses that take place after PAMP perception (Blume et al., 2000; Lecourieux et al., 2002; Kwaaitaal et al., 2011; Ranf et al., 2011), little is known about signal dynamics and propagation in different tissues and cell types at the single-cell level. To demonstrate that increased sensitivity of R-GECO1 would be instrumental to visualize PAMP-triggered single-cell Ca2+ signals, we challenged Arabidopsis seedlings with two well-described elicitors: the bacterial-derived peptide flg22 (Felix et al., 1999; Jeworutzki et al., 2010) and the fungal elicitor chitin (Mithöfer et al., 1999). We performed live cell imaging with R-GECO1 and NES-YC3.6 in detached true leaves of 14- to 16-day-old Arabidopsis seedlings. At indicated time points, we applied either 100 nM flg22 (Figure 4A and 4B) or 100 μg/ml chitin (Figure 4C and 4D). In the case of flg22, we detected defined oscillations of [Ca2+]cyt that lasted for approximately 30 min (Figure 4A and 4B and Supplemental Movie 2). We observed similar responses for chitin-treated leaves in which cytosolic Ca2+ oscillations lasted for approximately 20 min (Figure 4C and 4D and Supplemental Movie 3). As observed for stimulation with ATP and plasma membrane hyperpolarization (Figure 1 and Supplemental Figure 2), the signal amplitude after flg22 and chitin application is several times higher for R-GECO1 than for NES-YC3.6 (Figure 4B and 4D). As a result, R-GECO1 is able to visualize flg22- and chitin-induced transient [Ca2+]cyt patterns with higher sensitivity and resolution than NES-YC3.6 (Figure 4B and 4D). Previous studies using aequorin as Ca2+ reporter detected similar signal durations in response to flg22 and chitin in whole seedlings. However, the Ca2+ response measured by aequorin was monophasic with a rapid increase followed by an exponential decay (Kwaaitaal et al., 2011; Ranf et al., 2011; Maintz et al., 2014). Mathematical simulations showed that different shapes of Ca2+ response curves simply reflect the number of measured cells (Dodd et al., 2006). To further illustrate the relation between the number of measured cells and the shape of the Ca2+ response curve, we analyzed regions of different size. Ca2+ signals extracted from a subcellular domain of two or three neighboring cells (Figure 4B region of interest [ROI] 1, Figure 4D, ROI3) exhibited oscillatory patterns. However, Ca2+ signals of single cells do no oscillate in phase (Supplemental Movies 2 and 3). Therefore, the oscillatory Ca2+ response in regions that cover several cells (Figure 4B ROI2, Figure 4D, ROI4) was less pronounced. PAMP-Induced Signal Changes of R-GECO1 Are Ca2+-Dependent and Specific for flg22 or Chitin Next we wanted to demonstrate that the flg22- and chitin-induced signal changes of R-GECO1 are both Ca2+ dependent and PAMP specific. For this, we performed control experiments in leaves in which we either inhibited Ca2+ transients by application of the plasma membrane Ca2+ channel blocker LaCl3 (Nathan et al., 1988; Demidchik et al., 2002) or we used inactive forms of the respective PAMPs. As shown before, treatment with 1 μM flg22 or 100 μg/ml chitin induced a clear transient increase of [Ca2+]cyt (Supplemental Figure 4A and 4B). However, if leaves where pre-treated for 30 min with 1 mM LaCl3, no significant increase of [Ca2+]cyt could be observed after application of 1 μM flg22 or 100 μg/ml chitin (Supplemental Figure 4C and 4D). These results are in line with previous observations (Kwaaitaal et al., 2011; Ranf et al., 2011) and demonstrate that flg22- and chitin-induced signal changes of R-GECO1 are truly Ca2+ dependent. Similarly, no transient increase of [Ca2+]cyt could be observed in leaves treated with 1 μM inactive flg15Δ5 (Felix et al., 1999; Ranf et al., 2011) or 1 μM inactive hexameric chitin (ch6; Zhang et al., 2002; Ranf et al., 2011; Supplemental Figure 4E and 4F), whereas application of inactive PAMPs did not evoke considerable changes in [Ca2+]cyt, treatment with 1 mM ATP-induced cytosolic Ca2+ transients, which demonstrates the competence of leaves to respond to external stimulations (Supplemental Figure 4E and 4F). Overall, these experiments verified that the PAMP-induced signal changes of R-GECO1 are Ca2+ dependent and that the changes in cytosolic Ca2+ are specifically induced by the bacterial elicitor flg22 and the fungal elicitor chitin. Differences in Guard Cell Ca2+ Signaling in Response to flg22 and Chitin Since stomatal pores represent putative entry points for bacteria (Melotto et al., 2008) and fungi (Guimarães and Stotz, 2004), we decided to study guard cell Ca2+ dynamics after bacterial and fungal elicitor treatment in more detail (Figure 5). For this reason, two different imaging setups were used: The lower leaf surface (abaxial site) of detached first true leaves was either imaged from the top (Supplemental Figure 5A) or from the bottom (Supplemental Figure 5B). For top imaging, the upper (adaxial) site of the leaf was facing the glass slide, whereas the abaxial site was oriented toward the objective and the site of PAMP application (Supplemental Figure 5A). For bottom imaging, the abaxial site of the leaf was facing the cover slip and oriented toward the objective, whereas the adaxial site was oriented toward the site of PAMP application (Supplemental Figure 5B). At indicated time points, we treated true leaves of 14- to 16-day-old Arabidopsis seedlings with 100 nM flg22 (Figure 5A, 5C, 5E, and 5F) or with 100 μg/ml chitin (Figure 5B, 5D, 5G, and 5H). To compare Ca2+ signals from guard cells and epidermal cells, we selected the central part of the stomatal pore and a nearby region of an epidermal cell for data analyses (Figure 5A–5D). With both imaging setups we were able to observe defined [Ca2+]cyt oscillations in epidermal leaf cells after application of 100 nM flg22 or 100 μg/ml chitin (Figure 5A – 5H ROI2, ROI4, ROI6, ROI8). When the site of PAMP application was identical to the site of imaging (top imaging, Supplemental Figure 5A), 66% of the guard cells (n = 66) responded to flg22 with [Ca2+]cyt oscillations (Figure 5A ROI1, 5E ROI1, 5I). Correspondingly, 61% of the guard cells (n = 74) showed [Ca2+]cyt oscillations after chitin treatment (Figure 5B ROI3, 5G ROI3, 5I) when the elicitor was applied to the site of imaging. However, during bottom imaging (Supplemental Figure 5B), guard cell [Ca2+]cyt oscillations were observed after chitin application (Figure 5D ROI7, 5H ROI7, and Supplemental Movie 5), but only rarely after treatment with flg22 (Figure 5C ROI5, 5F ROI5, and Supplemental Movie 4). In this case only, 19% (n = 34) of the stomatal guard cells responded to flg22, whereas 85% of the guard cells (n = 37) responded to chitin (Figure 5J). The fact that guard cells imaged in the bottom mode (Supplemental Figure 5B and Figure 5J) respond less frequently to flg22 than guard cells imaged in the top mode (Supplemental Figure 5A and Figure 5I) most likely indicates that PAMP accessibility to the tissue is restricted in the bottom mode since the leaf surface is in tight contact with the cover glass (Supplemental Figure 5B). The extended lag time observed between PAMP application and onset of Ca2+ signaling supports this idea. Ca2+ signaling was initiated 7.0 ± 2.6 min (n = 11) after flg22 and 8.2 ± 2.6 min (n = 6) after chitin treatment in cells that were facing the cover glass (Figure 5F and 5H). Cells that were in direct contact with the bathing solution and therefore closer to the site of PAMP application responded much faster to application of flg22 (1.6 ± 0.5 min, n = 8) and chitin (2.6 ± 1.0 min, n = 8; Figure 5E and 5G). Our results suggest that guard cells have to perceive flg22 cell autonomously, since they are symplastically isolated and signaling epidermal cells are not sufficient to elicit a Ca2+ response in an adjacent guard cell (Figure 5C and Supplemental Movie 4). The observed differences between guard cell Ca2+ signaling in response to flg22 and chitin could be due to different diffusion rates of the respective PAMPs or due to differences in Ca2+ signal propagation. The latter would suggest that additional signaling mechanisms, such as production of reactive oxygen species (ROS) in the apoplast and electrical signaling might be required to communicate signals between epidermal and guard cells (Romeis and Herde, 2014; Gilroy et al., 2014; Steinhorst and Kudla, 2014). These data show that R-GECO1 is well suited to study cell-specific Ca2+ responses in intact leaves and that chitin as shown before flg22 (Thor and Peiter, 2014), induces defined [Ca2+]cyt oscillations in epidermal and guard cells. Furthermore, our results suggest that guard cells perceive flg22 in a cell-autonomous way. Flg22- and Chitin-Induced Ca2+ Signals in the Root Initiate from the Elongation Zone Tissue specificity with respect to pathogen attack is not very well investigated. Most studies on PAMP-induced immunity have focused on responses associated with the aerial parts of the plant, most likely due to the fact that in most experimental systems, roots have been more difficult to access than leaves. A quantitative approach in roots and shoots of Arabidopsis seedlings demonstrated that the amplitude of Ca2+ signals in response to PAMPs is dependent on the tissue context. It was shown that the majority of the Ca2+ response induced by flg22 could be attributed to the shoot part of the seedling, whereas, in response to the fungal elicitor N-acetylchitooctaose (ch8), the root contributed to a higher extent to the total Ca2+ response of the seedling (Ranf et al., 2011). To study PAMP-induced Ca2+ dynamics in roots in more detail, we performed Ca2+ imaging of 6- to 7-day-old seedlings that had been challenged with flg22 or chitin. Seedlings were grown and imaged in RootChip16, a microfluidic platform that allows reversible and non-invasive application of elicitors via micro-perfusion (Grossmann et al., 2011, 2012; Jones et al., 2014). Since roots have been reported to be less sensitive to flg22 (Ranf et al., 2011), we used 1 μM flg22 for elicitation. In agreement with previous work (Ranf et al., 2011), we found that all the roots tested were clearly responsive to flg22 treatment. We detected a clear increase in [Ca2+]cyt after application of 1 μM flg22 (Figure 6A, 6C and 6E). Treatment with 100 μg/ml chitin (Figure 6B, 6D, and 6F) also elicited a clear increase in [Ca2+]cyt. However, the frequency of Ca2+-responsive roots was lower since three of nine chitin-treated roots did not show a clear Ca2+ response (Figure 6F, root 3). Kymograph analyses revealed two interesting facts (Figure 6C and 6D). The Ca2+ signal initiated in the root elongation zone from where it spread toward the root tip and base and the signal amplitude was at maximum within the elongation zone and decreased as the signal spread (Figure 6A, 6B, 6E, 6F and Supplemental Movies 6, 7). These observations raise several interesting questions about tissue and cell specificity of immune responses but also about directionality and propagation of Ca2+ signals. Remarkably, the spatial onset of the Ca2+ response in the elongation zone correlated with the activation of defense genes. Several immune responsive genes are specifically upregulated in the elongation zone within 3–5 h after flg22 treatment (Millet et al., 2010). In addition, flg22-induced callose deposition was found to be restricted to the elongation zone (Millet et al., 2010). In contrast, no such correlation between initiation of Ca2+ signaling and gene expression in the elongation zone was observed for chitin. It has been speculated that tissue-specific immune responses of different PAMPs could reflect different infection strategies (Millet et al., 2010). Our data show that root Ca2+ imaging with R-GECO1 allowed the identification of tissue-specific Ca2+ responses. Flg22 as well as chitin induced Ca2+ transients in the root elongation zone, which spread toward the root tip and base. Whether such local Ca2+ signals are able to trigger systemic immune responses in other parts of the root or even in the shoot remains to be determined. In this study we established Ca2+ measurements in Arabidopsis using the intensity-based reporter R-GECO1. Comparative in vivo analysis demonstrated that R-GECO1 shows significantly increased Ca2+-dependent signal changes compared with NES-YC3.6. The increased sensitivity of R-GECO1 enabled visualization of flg22- and chitin-induced Ca2+ signals on a cellular scale. We have proved that flg22- and chitin-induced signal changes of R-GECO1 are Ca2+-dependent and specific for the respective elicitors. We identified that chitin induces defined [Ca2+]cyt oscillations in guard cells and suggest that guard cells perceive flg22 in a cell-autonomous way. Moreover, we found that flg22- and chitin-induced Ca2+ signals in the root initiate from the elongation zone. Overall, our data show that R-GECO1 is a useful tool that greatly facilitates Ca2+ imaging in plants. Due to the single-fluorophore nature of GECO-based indicators, they hold enormous potential for future applications such as multi-compartment and multi-parameter imaging. Experimental Procedures Cloning Procedure R-GECO1 (Zhao et al., 2011) was amplified from pTor-PE-R-GECO1 (Addgene plasmid 32465) using oligonucleotides R-GECO1-SpeF (5′-tttactagtatggtcgactcttcacgtc-3′) and R-GECO1-XmaR (5′-tttcccgggc tacttcgctgtcatcatttg-3′). The resulting fragment was inserted via SpeI/XmaI into a modified pUC19 plasmid pUC-pUBQ10 (Waadt et al., 2014) and subcloned into the barII-UT plasmid (Waadt et al., 2014), which contains a glufosinate resistance cassette for plant herbicide selection and an expression cassette for R-GECO1 consisting of the UBQ10 promoter (Norris et al., 1993) and the HSP18.2 terminator (Nagaya et al., 2010). Plant Materials and Growth Conditions for Arabidopsis To generate transgenic Arabidopsis thaliana (Col-0) lines that express two Ca2+ sensors simultaneously, the cytosolic yellow cameleon reporter line NES-YC3.6 (Krebs et al., 2012) was transformed with the binary vector barII-UT-R-GECO1, according to standard procedures (Hellens et al., 2000). Transgenic lines were selected for BASTA resistance on plates containing 10 μg/ml BASTA. For in vitro culture, seeds were surface sterilized using EtOH followed by stratification for 48 h at 4°C. Seedlings were grown at 22°C with cycles of 16 h light and 8 h darkness on plates containing half strength Murashige and Skoog (MS) basal salt mixture (Duchefa; www.duchefa-biochemie.com) supplemented with 0.5% sucrose. Medium pH was set to 5.8 using KOH and medium was solidified using 0.5% phytoagar (Duchefa). Pollen germination was performed as described previously (Hicks et al., 2004). Chemicals, Buffers, and Elicitors Stock solutions of 100 mM MgATP (pH 7.0 KOH; Sigma-Aldrich, www.sigmaaldrich.com), 100 μM flg22 (EZBiolab, www.ezbiolab.com), 1 mM flg15Δ5, 1 mM ch6, 200 mg/ml chitin, 1 M LaCl3 (Sigma-Aldrich), and 100 mM EGTA (Sigma-Aldrich) were prepared in water. The chitin stock solution was freshly prepared for each experiment by grinding chitin powder from shrimp cells (Sigma-Aldrich) for 10 min with a mortar and a pestle. The inactive PAMPs flg15Δ5 and ch6 were a kind gift from Justin Lee (IPB Halle, Germany; Ranf et al., 2011). The stock solution of 10 mM BAPTA-AM (Life Technologies, www.lifetechnologies.com) was dissolved in DMSO. The pH equilibration buffers (Yoshida, 1994; Krebs et al., 2010) and the de- and hyperpolarization buffers (Allen et al., 2000) were prepared as described previously. Ca2+ and pH Imaging Confocal laser scanning microscopy was performed on a Leica SP5II equipped with a DMI6000 inverted stand (Leica Microsystems, www.leica-microsystems.com). NES-YC3.6 was excited with 458 nm and fluorescence emission was detected between 465 and 505 nm (ECFP) and between 530 nm and 570 (cpVenus). R-GECO1 was excited with 561 nm and its emission was detected between 620 and 650 nm. pHGFP was sequentially excited with 405 and 488 nm and fluorescence emission was detected between 500 and 530 nm. Laser and gain settings were adjusted individually to have comparable baseline intensity values for each experiment. Images for NES-YC3.6, R-GECO1 and pHGFP were recorded using HyD detectors with a frame rate of 5 s. Samples were mounted for either top or bottom imaging (Supplemental Figure 5). For top imaging, the imaging chamber was formed from modeling clay (Supplemental Figure 5A). A thin film of medical adhesive was used to fix detached true leaves with the upper leaf surface facing the glass slide and the lower leaf surface facing the bathing solution. Leaves were covered with liquid half strength MS medium. If not stated otherwise, samples were imaged in the bottom imaging mode (Supplemental Figure 5B) and sample mounting was performed as described previously with minor modifications (Krebs and Schumacher, 2013). Instead of cotton, rock wool was used as spacer between the seedling and the imaging chamber. For imaging, seedlings were placed in liquid half strength MS medium. For ATP, flg22, and chitin treatments, two-fold concentrations of the respective agent were prepared in liquid half strength MS medium and added in a 1:1 volume ratio to the imaging chamber to achieve rapid concentration equilibrium. For plasma membrane hyperpolarization and treatments with pH equilibration buffers, a peristaltic pump was connected to the imaging chamber to perfuse the seedlings with the respective bathing solutions. Ca2+ Imaging on Roots in the RootChip16 Epifluorescent imaging was performed on a Nikon Eclipse Ti microscope equipped with a 20× 0.75 NA multi-immersion objective (Nikon, www.nikon.com) and an Andor iXon plus electron multiplying charge coupled device (Andor, www.andor.com) camera. R-GECO1 was excited using an Obis 561-50 DPSS diode laser (Omicron Laserage, www.omicron-laser.com). Fluorescence emission was detected between 570 and 640 nm. Images were recorded with a time interval of 1.5 s. RootChip16 sample preparation was done as described previously (Grossmann et al., 2011, 2012), using the advanced RootChip16 (Jones et al., 2014). In brief, seeds were surface sterilized and germinated on cut pipette tips, prefilled with solidified Hoagland's media (Sigma-Aldrich), and plugged into the chip. Seven-day-old roots grown inside the RootChip16's observation chambers were imaged and perfused with half strength liquid Hoagland's medium. Treatments were done with 5-min square pulses of flg22 or chitin dissolved in half strength liquid Hoagland's medium. Image Processing and Data Analysis For data analyses, fluorescence intensity values of each channel were extracted from the ROIs indicated using ImageJ (imagej.nih.gov/ij/). When ROIs were not specified, the entire image frame was used for data analyses. For NES-YC3.6 and pHGFP, fractional ratio changes (ΔR/R) were calculated from background corrected intensity values as (R – R0)/R0, where R0 is the average ratio of the baseline (25 frames, 2 min) of a measurement. Correspondingly, fractional fluorescence changes (ΔF/F) for R-GECO1 were calculated from background corrected intensity values of R-GECO1 as (F – F0)/F0, where F0 represents the average fluorescence intensity of the baseline (25 frames, 2 min) of a measurement. To compare the performance of NES-YC3.6 with R-GECO1, different sensor parameters were evaluated. The SD of the baseline was calculated from 25 frames, representing a 2-min time period. The maximum signal change represents the mean of three individual frames that cover the maximum peak amplitude (15-s time period). Accordingly, the SNR was calculated by dividing the mean maximum peak amplitude by the SD of the baseline. Ratiometric images were calculated as described previously (Kardash et al., 2011). For quantification of stomata that show a significant change in [Ca2+]cyt after elicitor treatment, the following criteria were applied. Stomata were considered to respond if the Ca2+ signal after elicitor treatment was three times higher than the SD of the baseline. The SD of the baseline was calculated from 360 frames that represent 30 min prior to elicitor application. Supplementary Material Supplemental The authors would like to thank Beate Schöfer and Mónica Fajardo-López for excellent technical assistance. We are grateful to Justin Lee (IPB Halle, Germany) for the inactive PAMPs flg15Δ5 and ch6. Funding: This work was funded by grants from Deutsche Forschungsgemeinschaft to N.K. (KE 1719/2-1) and K.S. (FOR964) and by grants from National Institute of Health (GM060396-ES010337) and National Science Foundation (MCB1414339) to J.I.S. Figure 1 R-GECO1 Exhibits Enhanced Ca2+-Dependent Signal Change Compared with NES-YC3.6 Ca2+-dependent signal changes in response to 1 mM ATP in roots of 6- to 8-day-old seedlings expressing NES-YC3.6 and R-GECO1. (A) Fluorescence images of ECFP, cpVenus, R-GECO1, and corresponding bright field image. Scale bar represents 50 μm. (B) Time-dependent fluorescence intensities. (C) Time-dependent normalized NES-YC3.6 emission ratios (ΔR/R) and normalized R-GECO1 fluorescence intensities (ΔF/F). (D–F) Maximum signal change (D), signal-to-noise ratios (E), and SD of the baseline (F). Error bars represent SD of three independent experiments. Figure 2 Visualization of Tip-Localized [Ca2+]cyt Gradients (A and B) Tip-localized [Ca2+]cyt gradients were visualized in germinating pollen tubes (A) and growing root hair cells (B) expressing NES-YC3.6 and R-GECO1. Shown are ratiometric images for NES-YC3.6, fluorescence images for R-GECO1, and the corresponding bright field images. (C) Time-dependent [Ca2+]cyt dynamics in a growing root hair. Fluorescence images of R-GECO1 at different time points. Time format, sss. (D) Time-dependent normalized R-GECO1 fluorescence intensities (ΔF/F) were extracted from the apex of a growing root hair indicated by the circled area in (C). Arrowheads in (D) correspond to the images shown in (C). Scale bars in (A–C) represent 20 μm. Figure 3 R-GECO1 Is More Sensitive toward Changes of pH than NES-YC3.6 pH-dependent signal changes in response to pH equilibration buffers in roots of 6- to 8-day-old seedlings expressing pHGFP or NES-YC3.6 and R-GECO1. For cytosolic pH adjustment, pH equilibration buffers ranging from pH 6.8–8.0 were applied sequentially to the seedlings. To suppress Ca2+ ion fluxes during the pH treatments, indicated concentrations of LaCl3, EGTA, and BAPTA-AM were applied simultaneously. (A) Fluorescence image of pHGFP and the corresponding bright field image. (B) Time-dependent normalized pHGFP ratio changes (ΔR/R). (C) Fluorescence images of ECFP, cpVenus, R-GECO1, and corresponding bright field image. (D) Time-dependent fluorescence intensities of ECFP, cpVenus, and R-GECO1. (E) Time-dependent normalized NES-YC3.6 emission ratios (ΔR/R) and normalized R-GECO1 fluorescence intensities (ΔF/F). Data are representative of n = 5 measurements. Scale bars in (A) and (C) represent 50 μm. Figure 4 R-GECO1 Detects PAMP-Triggered [Ca2+]cyt Oscillations in Intact Leaves Ca2+-dependent signal changes in response to flg22 (A, B) and chitin (C, D) in true leaves of 14- to 16-day-old seedlings expressing NES-YC3.6 and R-GECO1. Fluorescence images of R-GECO1 5:00 min after flg22 (A) and 11:17 min after chitin application (C). (B, D) Time-dependent normalized NES-YC3.6 emission ratios (ΔR/R) and normalized R-GECO1 fluorescence intensities (ΔF/F) calculated from ROIs 1–4 indicated in (A) and (C). Shown are representative experiments with n ≥ 6. Scale bars in (A) and (C) represent 50 μm and scale bars in (B) and (D) indicate ΔF/F (R-GECO1) and DR/R (NES-YC3.6) with y = 1. Figure 5 Ca2+ Signaling in Response to flg22 and Chitin in Guard Cells and Epidermal Cells Ca2+-dependent signal changes in response to flg22 and chitin in true leaves of 14- to 16-day-old seedlings were measured with different experimental setups. (A–D) Fluorescence images of R-GECO1. (A) Top imaging 3:10 min after flg22 application. (B) Top imaging 4:15 min after chitin application. (C) Bottom imaging 12:25 min after flg22 application. (D) Bottom imaging 9:52 min after chitin application. (E–H) Graphs show time-dependent normalized R-GECO1 fluorescence intensities (ΔF/F) calculated from ROIs 1–8 outlined in (A–D). Different treatments are indicated by gray boxed areas. (E) Top imaging 100 nM flg22. (F) Bottom imaging 100 nM flg22. (G) Top imaging 100 μg/ml chitin. (H) Bottom imaging 100 μg/ml chitin. (I) Percentage of guard cells that exhibited significant [Ca2+]cyt elevations in response to 100 nM flg22 (n = 66) or 100 μg/ml chitin (n = 74) as revealed in the top imaging setup. (J) Percentage of guard cells that exhibited significant [Ca2+]cyt elevations in response to 100 nM flg22 (n = 37) or 100 μg/ml chitin (n = 34) as revealed in the bottom imaging setup. (A–H) Shown are representative experiments with n ≥ 6. Scale bars in (A–D) represent 15 μm and scale bars in (E–H) indicate ΔF/F with y = 1. Figure 6 Flg22- and Chitin-Induced [Ca2+]cyt Transients in the Root Originate from the Elongation Zone Ca2+-dependent signal changes in response to 1 μM flg22 (A, C, E) and 100 μg/ml chitin (B, D, F) in roots of 6- to 7-day-old seedlings. Ca2+ imaging was performed in the RootChip16. (A and B) Time series of normalized R-GECO1 fluorescence images (ΔF/F). (C and D) Kymographs were extracted along three pixel-wide dashed lines indicated in (A) and (B). (E and F) Normalized R-GECO1 fluorescence intensities (ΔF/F) were measured from ROIs in the elongation and the root hair zone indicated by the boxed areas in (A) and (B). Shown are [Ca2+]cyt dynamics in response to flg22 and chitin of three independent roots. Arrowheads correspond to the images shown in (A) and (B). Gray boxes in (C–F) indicate a 5-min square pulse of flg22 or chitin, respectively. Scale bars in (A) and (B) indicate 200 μm and scale bars in (E) and (F) indicate ΔF/F with y = 0.1. Time format, m:ss. Note that the apparent increase in signal at the root cap in (A) and (B) does not indicate high levels of [Ca2+]cyt. It is a result of image calculation due to the growth of the root. This area was not included in the analysis. Supplemental Information: Supplemental Information is available at Molecular Plant Online. No conflict of interest declared. Akerboom J Rivera JDV Rodríguez Guilbe MM Malavé ECA Hernandez HH Tian L Hires SA Marvin JS Looger LL Schreiter ER 2009 Crystal structures of the GCaMP calcium sensor reveal the mechanism of fluorescence signal change and aid rational design J Biol Chem 284 6455 6464 19098007 Akerboom J Chen TW Wardill TJ Tian L Marvin JS Mutlu S Calderón NC Esposti F Borghuis BG Sun XR 2012 Optimization of a GCaMP calcium indicator for neural activity imaging J Neurosci 32 13819 13840 23035093 Akerboom J Carreras Calderón N Tian L Wabnig S Prigge M Tolö J Gordus A Orger MB Severi KE Macklin JJ 2013 Genetically encoded calcium indicators for multi-color neural activity imaging and combination with optogenetics Front Mol Neurosci 6 2 23459413 Allen GJ Chu SP Schumacher K Shimazaki CT Vafeados D Kemper A Hawke SD Tallman G Tsien RY Harper JF 2000 Alteration of stimulus-specific guard cell calcium oscillations and stomatal closing in Arabidopsis det3 mutant Science 289 2338 2342 11009417 Baubet V Le Mouellic H Campbell AK Lucas-Meunier E Fossier P Brúlet P 2000 Chimeric green fluorescent protein-aequorin as bioluminescent Ca2+ reporters at the single-cell level Proc Natl Acad Sci USA 97 7260 7265 10860991 Bibikova TN Zhigilei A Gilroy S 1997 Root hair growth in Arabidopsis thaliana is directed by calcium and an endogenous polarity Planta 203 495 505 9421933 Bibikova TN Jacob T Dahse I Gilroy S 1998 Localized changes in apoplastic and cytoplasmic pH are associated with root hair development in Arabidopsis thaliana Development 125 2925 2934 9655814 Blume B Nürnberger T Nass N Scheel D 2000 Receptor-mediated increase in cytoplasmic free calcium required for activation of pathogen defense in parsley Plant Cell 12 1425 1440 10948260 Boller T Felix G 2009 A renaissance of elicitors: perception of microbe-associated molecular patterns and danger signals by pattern-recognition receptors Annu Rev Plant Biol 60 379 406 19400727 Boudsocq M Sheen J 2010 Stress signaling II: calcium sensing and signaling Abiotic Stress Adaptation in Plants SE - 4 Pareek A Sopory SK Bohnert HJ Netherlands Springer 75 90 Brini M 2008 Calcium-sensitive photoproteins Methods 46 160 166 18848993 Choi WG Swanson SJ Gilroy S 2012 High-resolution imaging of Ca2+, redox status, ROS and pH using GFP biosensors Plant J 70 118 128 22449047 Choi J Tanaka K Cao Y Qi Y Qiu J Liang Y Lee SY Stacey G 2014 Identification of a plant receptor for extracellular ATP Science 343 290 294 24436418 Demidchik V Davenport RJ Tester M 2002 Nonselective cation channels in plants Annu Rev Plant Biol 53 67 107 12221989 Dodd AN Jakobsen MK Baker AJ Telzerow A Hou SW Laplaze L Barrot L Scott Poethig R Haseloff J Webb A 2006 Time of day modulates low-temperature Ca2+ signals in Arabidopsis Plant J 48 962 973 17227550 Dodd AN Kudla J Sanders D 2010 The language of calcium signaling Annu Rev Plant Biol 61 593 620 20192754 Feijó JA Sainhas J Hackett GR Kunkel JG Hepler PK 1999 Growing pollen tubes possess a constitutive alkaline band in the clear zone and a growth-dependent acidic tip J Cell Biol 144 483 496 9971743 Felix G Duran JD Volko S Boller T 1999 Plants have a sensitive perception system for the most conserved domain of bacterial flagellin Plant J 18 265 276 10377992 Fendrych M Van Hautegem T Van Durme M Olvera-Carrillo Y Huysmans M Karimi M Lippens S Guérin CJ Krebs M Schumacher K 2014 Programmed cell death controlled by ANAC033/SOMBRERO determines root cap organ size in Arabidopsis Curr Biol 24 931 940 24726156 Gilroy S Suzuki N Miller G Choi WG Toyota M Devireddy A Mittler R 2014 A tidal wave of signals: calcium and ROS at the forefront of rapid systemic signaling Trends in Plant Science 19 623 630 25088679 Grefen C Donald N Hashimoto K Kudla J Schumacher K Blatt MR 2010 A ubiquitin-10 promoter-based vector set for fluorescent protein tagging facilitates temporal stability and native protein distribution in transient and stable expression studies Plant J 64 355 365 20735773 Grossmann G Guo WJ Ehrhardt DW Frommer WB Sit RV Quake SR Meier M 2011 The RootChip: an integrated microfluidic chip for plant science Plant Cell 23 4234 4240 22186371 Grossmann G Meier M Cartwright HN Sosso D Quake SR Ehrhardt DW Frommer WB 2012 Time-lapse fluorescence imaging of Arabidopsis root growth with rapid manipulation of the root environment using the RootChip J Vis Exp http://dx.doi.org/10.3791/4290 Guimarães RL Stotz HU 2004 Oxalate production by Sclerotinia sclerotiorum deregulates guard cells during infection Plant Physiol 136 3703 3711 15502012 Heim N Griesbeck O 2004 Genetically encoded indicators of cellular calcium dynamics based on troponin C and green fluorescent protein J Biol Chem 279 14280 14286 14742421 Hellens R Mullineaux P Klee H 2000 A guide to Agrobacterium binary Ti vectors Trends Plant Sci 5 446 451 11044722 Hicks GR Rojo E Hong S Carter DG Raikhel NV 2004 Geminating pollen has tubular vacuoles, displays highly dynamic vacuole biogenesis, and requires VACUOLESS1 for proper function Plant Physiol 134 1227 1239 14988481 Horikawa K Yamada Y Matsuda T Kobayashi K Hashimoto M Matsura T Miyawaki A Michikawa T Mikoshiba K Nagai T 2010 Spontaneous network activity visualized by ultrasensitive Ca2+ indicators, yellow Cameleon-Nano Nat Methods 7 729 732 20693999 Ikura M Clore GM Gronenborn AM Zhu G Klee CB Bax A 1992 Solution structure of a calmodulin-target peptide complex by multidimensional NMR Science 256 632 638 1585175 Jeworutzki E Roelfsema MRG Anschütz U Krol E Elzenga JTM Felix G Boller T Hedrich R Becker D 2010 Early signaling through the Arabidopsis pattern recognition receptors FLS2 and EFR involves Ca2+-associated opening of plasma membrane anion channels Plant J 62 367 378 20113440 Jones AM Danielson JÅH Manojkumar SN Lanquar V Grossmann G Frommer WB 2014 Abscisic acid dynamics in roots detected with genetically encoded FRET sensors Elife 3 e01741 24737862 Kardash E Bandemer J Raz E 2011 Imaging protein activity in live embryos using fluorescence resonance energy transfer biosensors Nat Protoc 6 1835 1846 22051797 Kiegle E Moore CA Haseloff J Tester MA Knight MR 2000 Cell-type-specific calcium responses to drought, salt and cold in the Arabidopsis root Plant J 23 267 278 10929120 Kneen M Farinas J Li Y Verkman AS 1998 Green fluorescent protein as a noninvasive intracellular pH indicator Biophys J 74 1591 1599 9512054 Knight MR Campbell AK Smith SM Trewavas AJ 1991 Transgenic plant aequorin reports the effects of touch and cold-shock and elicitors on cytoplasmic calcium Nature 352 524 526 1865907 Knight H Trewavas AJ Knight MR 1996 Cold calcium signaling in Arabidopsis involves two cellular pools and a change in calcium signature after acclimation Plant Cell 8 489 503 8721751 Krebs M Schumacher K 2013 Live cell imaging of cytoplasmic and nuclear Ca2+ dynamics in Arabidopsis roots Cold Spring Harb Protoc 2013 776 780 23906910 Krebs M Beyhl D Gö rlich E Al-Rasheid KAS Marten I Stierhof YD Hedrich R Schumacher K 2010 Arabidopsis V-ATPase activity at the tonoplast is required for efficient nutrient storage but not for sodium accumulation Proc Natl Acad Sci USA 107 3251 3256 20133698 Krebs M Held K Binder A Hashimoto K Den Herder G Parniske M Kudla J Schumacher K 2012 FRET-based genetically encoded sensors allow high-resolution live cell imaging of Ca2+ dynamics Plant J 69 181 192 21910770 Kwaaitaal M Huisman R Maintz J Reinstadler A Panstruga R 2011 Ionotropic glutamate receptor (iGluR)-like channels mediate MAMP-induced calcium influx in Arabidopsis thaliana Biochem J 440 355 365 21848515 Lecourieux D Mazars C Pauly N Ranjeva R Pugin A 2002 Analysis and effects of cytosolic free calcium increases in response to elicitors in Nicotiana plumbaginifolia cells Plant Cell 14 2627 2641 12368509 Luan S 2009 The CBL-CIPK network in plant calcium signaling Trends Plant Sci 14 37 42 19054707 Maintz J Cavdar M Tamborski J Kwaaitaal M Huisman R Meesters C Kombrink E Panstruga R 2014 Comparative analysis of MAMP-induced calcium influx in arabidopsis seedlings and protoplasts Plant Cell Physiol 55 1813 1825 25231962 Mank M Santos AF Direnberger S Mrsic-Flogel TD Hofer SB Stein V Hendel T Reiff DF Levelt C Borst A 2008 A genetically encoded calcium indicator for chronic in vivo two-photon imaging Nat Methods 5 805 811 19160515 Martí MC Stancombe MA Webb AAR 2013 Cell- and stimulus type-specific intracellular free Ca2+ signals in Arabidopsis Plant Physiol 163 625 634 24027243 McAinsh MR Pittman JK 2009 Shaping the calcium signature New Phytol 181 275 294 19121028 Melotto M Underwood W He SY 2008 Role of stomata in plant innate immunity and foliar bacterial diseases Annu Rev Phytopathol 46 101 122 18422426 Millet YA Danna CH Clay NK Songnuan W Simon MD Werck-Reichhart D Ausubel FM 2010 Innate immune responses activated in Arabidopsis roots by microbe-associated molecular patterns Plant Cell 22 973 990 20348432 Mithöfer A Ebel J Bhagwat AA Boller T Neuhaus-Url G 1999 Transgenic aequorin monitors cytosolic calcium transients in soybean cells challenged with β-glucan or chitin elicitors Planta 207 566 574 Miyawaki A Llopis J Heim R McCaffery JM Adams JA Ikura M Tsien RY 1997 Fluorescent indicators for Ca2+ based on green fluorescent proteins and calmodulin Nature 388 882 887 9278050 Monshausen GB 2012 Visualizing Ca2+ signatures in plants Curr Opin Plant Biol 15 1 6 22217656 Monshausen GB Messerli MA Gilroy S 2008 Imaging of the yellow cameleon 3.6 indicator reveals that elevations in cytosolic Ca2+ follow oscillating increases in growth in root hairs of Arabidopsis Plant Physiol 147 1690 1698 18583529 Moseyko N Feldman LJ 2001 Expression of pH-sensitive green fluorescent protein in Arabidopsis thaliana Plant Cell Environ 24 557 563 11706851 Muto A Ohkura M Kotani T Higashijima S Nakai J Kawakami K 2011 Genetic visualization with an improved GCaMP calcium indicator reveals spatiotemporal activation of the spinal motor neurons in zebrafish Proc Natl Acad Sci USA 108 5425 5430 21383146 Nagai T Sawano A Park ES Miyawaki A 2001 Circularly permuted green fluorescent proteins engineered to sense Ca2+ Proc Natl Acad Sci USA 98 3197 3202 11248055 Nagai T Yamada S Tominaga T Ichikawa M Miyawaki A 2004 Expanded dynamic range of fluorescent indicators for Ca2+ by circularly permuted yellow fluorescent proteins Proc Natl Acad Sci USA 101 10554 10559 15247428 Nagaya S Kawamura K Shinmyo A Kato K 2010 The HSP terminator of Arabidopsis thaliana increases gene expression in plant cells Plant Cell Physiol 51 328 332 20040586 Nakai J Ohkura M Imoto K 2001 A high signal-to-noise Ca2+ probe composed of a single green fluorescent protein Nat Biotechnol 19 137 141 11175727 Nathan RD Kanai K Clark RB Giles W 1988 Selective block of calcium current by lanthanum in single bullfrog atrial cells J Gen Physiol 91 549 572 2455767 Ngo QA Vogler H Lituiev DS Nestorova A Grossniklaus U 2014 A calcium dialog mediated by the FERONIA signal transduction pathway controls plant sperm delivery Dev Cell 29 491 500 24814317 Norris SR Meyer SE Callis J 1993 The intron of Arabidopsis thaliana polyubiquitin genes is conserved in location and is a quantitative determinant of chimeric gene expression Plant Mol Biol 21 895 906 8385509 O'Connor N Silver RB 2013 Ratio imaging: practical considerations for measuring intracellular Ca2+ and pH in living cells Methods Cell Biol 114 387 406 23931515 Palmer AE Tsien RY 2006 Measuring calcium signaling using genetically targetable fluorescent indicators Nat Protoc 1 1057 1065 17406387 Pérez Koldenkova V Nagai T 2013 Genetically encoded Ca2+ indicators: properties and evaluation Biochim Biophys Acta 1833 1787 1797 23352808 Porumb T Yau P Harvey TS Ikura M 1994 A calmodulin-target peptide hybrid molecule with unique calcium-binding properties Protein Eng 7 109 115 8140087 Ranf S Eschen-Lippold L Pecher P Lee J Scheel D 2011 Interplay between calcium signalling and early signalling elements during defence responses to microbe- or damage-associated molecular patterns Plant J 68 100 113 21668535 Reiss HD Herth W 1978 Visualization of the Ca2+-gradient in growing pollen tubes of Lilium longiflorum with chlorotetracycline fluorescence Protoplasma 97 373 377 Romeis T Herde M 2014 From local to global: CDPKs in systemic defense signaling upon microbial and herbivore attack Curr Opin Plant Biol 20 1 10 24681995 Shaner NC Lin MZ McKeown MR Steinbach PA Hazelwood KL Davidson MW Tsien RY 2008 Improving the photostability of bright monomeric orange and red fluorescent proteins Nat Methods 5 545 551 18454154 Steinhorst L Kudla J 2013 Calcium - a central regulator of pollen germination and tube growth Biochim Biophys Acta 1833 1573 1581 23072967 Steinhorst L Kudla J 2014 Signaling in cells and organisms -calcium holds the line Curr Opin Plant Biol 22 14 21 25195171 Tanaka K Swanson SJ Gilroy S Stacey G 2010 Extracellular nucleotides elicit cytosolic free calcium oscillations in Arabidopsis Plant Physiol 154 705 719 20671112 Thestrup T Litzlbauer J Bartholomäus I Mues M Russo L Dana H Kovalchuk Y Liang Y Kalamakis G Laukat Y 2014 Optimized ratiometric calcium sensors for functional in vivo imaging of neurons and T lymphocytes Nat Methods 11 175 182 24390440 Thor K Peiter E 2014 Cytosolic calcium signals elicited by the pathogen-associated molecular pattern flg22 in stomatal guard cells are of an oscillatory nature New Phytologist 204 873 881 25243759 Waadt R Hitomi K Nishimura N Hitomi C Adams SR Getzoff ED Schroeder JI 2014 FRET-based reporters for the direct visualization of abscisic acid concentration changes and distribution in Arabidopsis Elife 3 1 28 Wu J Liu L Matsuda T Zhao Y Rebane A Drobizhev M Chang YF Araki S Arai Y March K 2013 Improved orange and red Ca2+ indicators and photophysical considerations for optogenetic applications ACS Chem Neurosci 4 963 972 23452507 Xiong TC Ronzier E Sanchez F Corratgé-Faillie C Mazars C Thibaud JB 2014 Imaging long distance propagating calcium signals in intact plant leaves with the BRET-based GFP-aequorin reporter Front Plant Sci 5 1 13 Yoshida S 1994 Low temperature-induced cytoplasmic acidosis in cultured Mung Bean (Vigna radiata [L.] Wilczek) cells Plant Physiol 104 1131 1138 12232153 Young JJ Mehta S Israelsson M Godoski J Grill E Schroeder JI 2006 CO2 signaling in guard cells: calcium sensitivity response modulation, a Ca2+-independent phase, and CO2 insensitivity of the gca2 mutant Proc Natl Acad Sci USA 103 7506 7511 16651523 Young B Wightman R Blanvillain R Purcel SB Gallois P 2010 pH-sensitivity of YFP provides an intracellular indicator of programmed cell death Plant Methods 6 27 21118545 Zhang B Ramonell K Somerville S Stacey G 2002 Characterization of early, chitin-induced gene expression in Arabidopsis Mol Plant Microbe Interact 15 963 970 12236603 Zhao Y Araki S Wu J Teramoto T Chang YF Nakano M Abdelfattah AS Fujiwara M Ishihara T Nagai T 2011 An expanded palette of genetically encoded Ca2+ indicators Science 333 1888 1891 21903779

PMC005xxxxxx/PMC5134739.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 0404511 7473 Science Science Science (New York, N.Y.) 0036-8075 1095-9203 27846599 5134739 10.1126/science.aag1322 NIHMS829096 Article Potent Protection against H5N1 and H7N9 Influenza via Childhood Hemagglutinin Imprinting* Gostic Katelyn M. 1 Ambrose Monique 1 Worobey Michael 2** Lloyd-Smith James O. 13** 1 Department of Ecology and Evolutionary Biology, University of California, Los Angeles, Los Angeles, CA, 90095, USA 2 Department of Ecology and Evolutionary Biology, University of Arizona, Tucson, AZ, 85721, USA 3 Fogarty International Center, National Institutes of Health, Bethesda, MD, 20892, USA ** Correspondence and requests for materials should be addressed to worobey@email.arizona.edu or jlloydsmith@ucla.edu 12 11 2016 11 11 2016 11 5 2017 354 6313 722726 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. Two zoonotic influenza A viruses (IAV) of global concern, H5N1 and H7N9, exhibit unexplained differences in age distribution of human cases. Using data from all known human cases of these viruses, we show that an individual’s first IAV infection confers lifelong protection against severe disease from novel hemagglutinin (HA) subtypes in the same phylogenetic group. Statistical modeling shows protective HA imprinting is the crucial explanatory factor, providing 75% protection against severe infection and 80% protection against death for both H5N1 and H7N9. Our results enable us to predict age distributions of severe disease for future pandemics and demonstrate that a novel strain’s pandemic potential increases yearly when a group-mismatched HA subtype dominates seasonal influenza circulation. These findings open new frontiers for rational pandemic risk assessment. The spillover of novel influenza A viruses (IAV) is a persistent threat to global health. H5N1 and H7N9 are particularly concerning avian-origin IAVs, each having caused hundreds of severe or fatal human cases (1). Despite commonalities in their reservoir hosts and epidemiology, these viruses show puzzling differences in age distribution of observed human cases (1,2). Existing explanations, including possible protection against H5N1 among older birth-year cohorts exposed to the neuraminidase of H1N1 as children (3,4) or age biases in exposure to infected poultry (5–7), cannot fully explain these opposing patterns of severe disease and mortality. Another idea is that severity of H5N1 and H7N9 differs by age, leading to case ascertainment biases (1), but no explanatory mechanism has been proposed. The key antigenic determinants for IAV susceptibility are the virus’s two surface glycoproteins, hemagglutinin (HA) and neuraminidase (NA), where different numbered subtypes canonically indicate no cross-immunity. However, recent experiments have revealed that broadly-protective immune responses can provide cross-immunity between different HA subtypes, particularly subtypes in the same phylogenetic group (8–14). (HA group 1 contains (human seasonal) subtypes H1, H2 and avian H5, while group 2 contains seasonal H3 and avian H7; Fig. 1A, Fig. S1). Combining these insights into heterosubtypic immunity with the concept of ‘original antigenic sin’ (15) or ‘antigenic seniority’ (16), we hypothesized that individuals imprint on the HA group of their first IAV exposure and thereby experience a reduced risk of severe disease from novel IAVs within that same phylogenetic group. This hypothesis predicts that the 1968 pandemic, which marked the transition from an era of group 1 HA circulation (1918–1968) to a group 2-dominated one (1968-present) (Fig. 1B), caused a major shift in population susceptibility that would explain why H5N1 cases are generally detected in younger people than H7N9 (2,17–19). Our analysis of human cases of H5N1 and H7N9 revealed strong evidence that childhood HA imprinting indeed provides profound, lifelong protection against severe infection and death from these viruses. These findings allowed us to develop new approaches for IAV pandemic risk assessment, preparedness and response, but also raise possible challenges for future vaccination strategies. Reconstructing IAV exposure history by birth year To investigate whether an individual’s initial childhood exposure to IAV influences later susceptibility to H5 and H7 viruses, we estimated the fraction of each birth-year cohort from 1918 to 2015 with first exposure to H1, H2, or H3 – or the fraction still naïve – for each country in our study (China, Egypt, Cambodia, Indonesia, Thailand, Vietnam). We estimated the annual probability of IAV infection in children using published age-seroprevalence data (20,21) and then rescaled this baseline attack rate to account for year-to-year variability in IAV circulation intensity (Supplementary Text). One resulting country-specific reconstruction of HA history is depicted in Fig. 1C. While H3N2 has dominated since 1968, a non-negligible fraction of many birth-year cohorts from the 1970s onwards was exposed first to H1N1 viruses, with notable peaks near the re-emergence of H1N1 in 1977 and the 2009 pandemic. H5N1 and H7N9 cases track HA imprinting patterns Next, we compiled data on all known human cases of H5N1 and H7N9 with reported patient age (Fig. 2A,B). These data encompass mostly clinically severe and fatal cases; total incidence remains unknown. Thus, our analysis focused on the determinants of severe cases. The possible existence of many undetected mild cases, as hypothesized for H7N9 (1), is consistent with HA imprinting since broadly-protective immune responses are expected to provide partial protection (8,14), i.e., reduce severity without preventing infection altogether (4,12,22–25). The preponderance of observed H7N9 cases among older cohorts, and H5N1 cases among younger cohorts, is clear (Fig. 2A,B). These patterns reflect birth year, not age: H5N1 cases occurred over 19 years from 1997–2015, yet cases from all years exhibit similar dependence on birth year. Analysis of 361 H5N1 cases in Egypt, the one country with many cases across the last decade, shows no trend in case birth years through time, while case age increased steadily (p=0.0003, Spearman’s correlation; Fig. S2). So, on average, the same birth cohorts remained at high risk of severe infection, even as members grew ten years older. Fig. 2C and D depict the case data normalized to demographic age distributions in affected countries. (If all birth cohorts had equal risk of severe infection, case incidence would be proportional to age distribution.) Bars above the midline thus represent birth years showing excess risk, while bars below indicate a shortfall. This normalization highlights two points: first, excess incidence and mortality data for H5N1 and H7N9 are near-mirror images of each other. Second, the group 1 to group 2 HA transition in 1968 is the key inflection point, with those born before the emergence of H3N2 showing protection against severe cases of H5N1 but not H7N9, and those born after 1968 showing the opposite pattern. For H7N9 severe case incidence also spikes in birth years around 1977 and 2009, when resurgent H1N1 circulation would have caused considerable mismatched imprinting. One-sided binomial exact tests showed excess H5N1 incidence had a lower probability of occurring in cohorts born before 1968 (p<1e−10), while excess H7N9 incidence was more probable in these same cohorts (p<1e−9). The same pattern held for excess mortality (Supplementary Text). These patterns suggest that the immune system imprints on conserved HA epitopes from the first-ever exposure to IAV, resulting in heterosubtypic (but within-group) protection against severe infection. Even more striking is the tight correspondence of observed H5N1 and H7N9 incidence and mortality with a priori predictions based on HA imprinting patterns and demographic age distributions (Fig. 2). We emphasize that the black lines in Fig. 2 are not fitted to the case data, but are independent predictions (Fig. 1C). Differences between the predictions and data are remarkably small—some noise arises from generalization across time and countries (e.g. attack rates for the reconstruction came from German data, but focal populations are largely Asian), and from small case numbers. Incorporating additional epidemiological factors and estimating the protective efficacy of imprinting further improved correspondence between predictions and data (Fig. S3). In contrast, NA imprinting patterns (which fully capture patterns of childhood exposure to N1) are a poor fit to H5N1 and H7N9 data from 1957–1968 cohorts (Fig. S4), and NA-mediated protection is not supported by statistical modeling. HA imprinting explains age distributions To formally assess the HA imprinting hypothesis alongside previous explanations (1,3–7) for observed H5N1 and H7N9 age distributions, we developed a set of multinomial models. These models related the probability that a case occurred in a given birth cohort to country- and year-specific demography, and risk factors including age-based risk of exposure to poultry, age-based risk of severe disease or case ascertainment, and reconstructed patterns of first exposure (and hence potential immunological imprinting) to HA or NA subtypes (Table S1). Model comparison showed HA imprinting was the dominant explanatory factor for observed incidence and mortality patterns for both H5N1 and H7N9. It was the only tested factor included in all plausible models for both viruses (i.e. all models with Akaike weights greater than 4e−5). The best models also included age-based risk of severe disease, echoing patterns known from seasonal influenza epidemiology. Age-based poultry exposure risk (estimated based on contact data from China (6, 7)) was included for H7N9 but not H5N1, perhaps reflecting that age-specific poultry exposures vary across the multiple countries affected by H5N1 or that humans interact differently with ill (H5N1-infected) versus asymptomatic (H7N9-infected) poultry. In models including HA imprinting, NA imprinting never showed any significant effect (Table S2). Remarkably, despite differences between the viruses and age cohorts involved, the estimated protective effects of HA imprinting were nearly identical for H7N9 and H5N1. In all models, protective HA imprinting reduced the risk of severe infection with H5N1 or H7N9 by ~75%, and the risk of death by ~80% (Table 1, Figs. S5–S7, Table S2). Antigenic seniority across influenza subtypes Most individuals born before the emergence of H3N2 in 1968, and exposed first to group 1 HA antigens (Fig. 1), have also been exposed to H3N2 after 1968—probably multiple times. Yet these seasonal group 2 exposures later in life evidently fail to override group 1 HA imprinting from childhood (Fig. 2). The birth-year specific protection seen for human H5N1 and H7N9 thus clearly indicates that clinically relevant antigenic seniority— preferential recall of immunological reactivities to antigens encountered earlier in life upon later exposure to cross-reactive antigens (16)—can act across HA subtypes of the same HA group, not just within subtypes as often assumed. While the precise mechanism underlying antigenic seniority in this context remains to be determined, antibodies directed against conserved HA epitopes provide a plausible explanation for protection at the level of HA groups. For example, research following the 2009 H1N1 pandemic drew attention to the fact that primary exposure to a novel IAV can preferentially boost broadly-protective antibodies that bind conserved HA head or stem epitopes shared by different HA subtypes (8–14), even though immune memory against more variable epitopes on the novel HA head may be absent. This absence may in fact enable robust expression of otherwise subdominant, broadly-protective responses to conserved epitopes such as those on the HA stem (8). In particular, primary exposure to H5N1 or H7N9 can activate HA stem-specific reactivities induced by previous infection by H1 or H3, respectively (12,13,26). Indeed, others have suggested that heterosubtypic antibodies might attenuate disease from other IAV strains and may be imprinted to some degree by childhood exposure, though their serological assays provided no ability to detect or predict actual patterns of protection relevant to H5N1 and H7N9 in human populations (27). Given the immunodominant nature of HA head reactivities (13,14,26,28), conserved HA head epitopes shared within, but not between, HA groups (29) may play a role in these patterns of protection. Cross-reactive HA-specific CD4+ or CD8+ T cell responses should also be investigated, since they too are likely to be disproportionately shared within HA groups (given the sequence similarities within each clade) and might be especially capable of facilitating the sort of long-term immunity indicated by the data. Nevertheless, current data, including the high degree of sequence conservation of stem domains within each HA group (Fig. 1A, Fig. S1), seem most consistent with a stem-directed mechanism for the antigenic seniority acting at the HA-group level (13). Divergence in HA stem amino acid sequences within each phylogenetic group is comparable to divergence in globular head sequences within a single HA subtype (i.e. the scale at which antigenic seniority is already known to act (16); Fig. S1), but stem divergences between the two HA groups are markedly higher. Notably, H3 and H7 are as divergent as any pair of group 2 HAs; since H3 childhood exposure provides protection against H7 it may thus protect as well or better against the other group 2 HAs (H4, H10, H14, H15), but perhaps not at all against more divergent group 1 HAs (Fig. S1C). Similarly, the joint consideration of protein sequence conservation patterns (Fig. 1A, Fig. S1) along with immunological and epidemiological data suggests that H1 or H2 childhood exposure may protect generally against zoonotic group 1—but not group 2—HAs. The putative generality in HA imprinting protection patterns for novel HA subtypes other than H5N1 or H7N9 is tentatively supported by the preponderance of HA group-mismatched childhood exposures among the small number of clinically significant human cases detected to date: pooling data from 28 human cases of H5N6, H6N1, H7N7, H9N2, H10N7 and H10N8, age patterns are consistent with HA imprinting (p=0.019; see Supplementary Text), but case numbers are insufficient to investigate particular subtypes. Immunological experiments (e.g. using chimeric HA proteins (12)) are needed to systematically map HA cross-protection patterns across all HA subtypes, both within and between HA groups. Rational projections of future pandemic risk For any new pandemic IAV strain capable of efficient human-to-human transmission, HA imprinting patterns would combine with age-based mixing patterns (30–32) to determine the epidemiological impacts of the first pandemic wave. We created projections for a putative pandemic-capable strain of subtype H5 or H7—such as a gain-of-function strain or a natural variant with mutations increasing human-to-human transmissibility. The data on observed H7N9 and H5N1 cases enabled us to quantify how matched HA imprinting reduces the probability of developing a severe infection, but not how matched imprinting affects an individual’s probability of acquiring a milder infection or the infectivity of such mild infections. People who become infected despite prior immunity likely transmit influenza at reduced rates owing to diminished viral titers and viral shedding, as observed in humans and in animal models (4,12,22–25). We thus assumed, conservatively, that imprinting does not change the probability of acquiring infection upon exposure, but can reduce severity and infectivity in individuals with protective HA imprinting. Fig. 3A illustrates the projected age-structured attack rate of severe cases for hypothetical pandemics of H5 or H7 IAV occurring in 2015 in the United Kingdom. The projected risk profiles for severe infection are shaped strongly by HA imprinting, including the prediction that individuals above 50 years of age (i.e. born well before 1968 and first exposed to a group 1 HA) would experience much lower morbidity than younger age groups in an H5 pandemic. Similar projections for China and Vietnam reveal the influence of demographic differences between countries (Fig. S8). The qualitative patterns in projected age impacts are robust to a wide range of assumptions about how seasonal influenza vaccination might affect imprinting (Fig. 3A), and to the assumed infectivity of mild cases arising in individuals with protective HA imprinting (Fig. S8A). Projections for pandemics occurring a decade from now highlight predictable shifts in severe disease risk patterns as the imprinted population ages, with the key pivot point around birth years near 1968 shifted to older ages (Fig. S8). Impacts in the youngest age groups would depend on patterns of IAV circulation in the coming decade. All pandemic projections that account for HA imprinting exhibit markedly lower severe attack rates than projections assuming no imprinting protection (Fig. 3A, Fig. S8). Total attack rates (including mild and subclinical cases) would be higher and more evenly distributed across age groups than the severe attack rates shown here. Over any prolonged period when IAV circulation is dominated by one HA group, imprinting generates growing herd immunity against zoonotic IAV strains from that group. Conversely, zoonotic strains from the mismatched HA group benefit from the rising proportion of humans without protection. So long as mild cases arising in people with group-matched imprinting contribute any less to transmission than unprotected cases, or if some fraction of infection events is prevented by imprinting-derived immunity, imprinting will alter the transmissibility of zoonotic IAV strains in the human population. This is summarized by the effective reproductive number, Reff, which quantifies transmission in a partially immune population (Fig. 3B). Crucially, a zoonotic strain that is initially subcritical (i.e. with Reff < 1 and therefore unable to spread sustainably) could—due solely to susceptibility changes in the human population— emerge as supercritical, and hence as a pandemic threat, if the mismatched HA group dominates IAV circulation for a sufficient period (Fig. 3B). Our work implies that we have never seen a true ‘virgin soil’ influenza pandemic, and that all prior estimates of R0 for pandemic IAVs are systematic underestimates since they do not account for protection induced by HA imprinting. Conversely, we see that imprinting raises the threshold R0 necessary for a novel subtype to invade. Interestingly, the co-circulation of group 1 and 2 HAs since 1977 has balanced herd immunity in a way that increases the inherent transmissibility needed for a novel subtype from either HA group to invade. As a generality, Reff for zoonotic influenza strains will change through time depending on seasonal influenza patterns and demographic background, and the magnitude of change will depend on the infectivity of imprinting-protected cases (Fig. S9). Discussion Our findings show that major patterns in zoonotic IAV epidemiology, previously attributed to patient age, are in fact driven by birth year. IAV strains circulating during an individual’s childhood confer long-term protection against novel HA subtypes from the same phylogenetic group. Hence, antigenic seniority extends across IAV subtypes, introducing previously unrecognized generational structure to influenza epidemiology. These immune imprinting effects have implications for public health and highlight that influenza virulence represents a joint phenotype between virus and host—even for strains not yet adapted to the human population. These findings support the hypothesis that the unusually high mortality in young adults during the 1918 H1N1 (group 1) pandemic may have arisen primarily from mismatched H3 (group 2) imprinting in the cohort born between ~1880 and 1900 (19). This same cohort was strongly affected during the (group 1) 1957 pandemic (33); yet they suffered no excess mortality when they were even older, during the (group 2) 1968 pandemic (34). The possibility that mismatched HA imprinting currently contributes to the greater health impacts of seasonal H3N2 (relative to H1N1) in today’s older age classes is worth investigating. And a diagnostic assay to determine whether an individual was imprinted on a group 1 or group 2 HA may be useful for individualized clinical care and vaccine design strategies, both for pandemic and seasonal IAVs. Our findings raise questions about whether seasonal influenza vaccination might boost broadly-protective anti-HA responses (27) or alter imprinting from natural infection in IAV-naïve children. By exposing IAV-naïve children simultaneously to group 1 (H1N1) and group 2 (H3N2) antigens, vaccination might confer dual imprinting to both HA groups, or prevent strong imprinting against either HA group—or it could have no effect beyond delaying the age of imprinting via the first natural infection. Our sensitivity analyses demonstrated that, given the low IAV vaccination coverage in H5N1- and H7N9-affected countries, none of these effects would change our study’s conclusions (Fig. S7). However, to properly inform early childhood vaccine policy, future research must determine which, if any, of these effects occur. HA group imprinting also might complicate ‘universal’ vaccination approaches targeting conserved HA epitopes. Our findings indicate potent, long-lasting cross-protection between subtypes, putatively based on such responses. However, universal vaccination may have to outperform natural infection in its ability to induce broad immunity in the face of previous imprinting. The persistence of group 1 imprinting in older adults, despite decades of natural exposure to H3N2 after 1968 (Fig. 2), and the relative weakness of group 2 anti-HA stem reactivities in these older cohorts (11), suggest HA exposures later in life do not readily alter broadly-protective responses in individuals already imprinted to a particular HA group. To be effective, would bivalent (group 1 and group 2 HA stem) universal vaccines need to be delivered to infants prior to natural IAV infection? Or, might universal vaccines even impair natural, long-term protection of the sort we have detected against H5N1/H7N9 if received prior to an individual’s first natural IAV infection? Our findings are consistent with the known potential for repeated infection by seasonal IAV subtypes. Group-matched imprinting, like other broadly-protective IAV immune responses, is expected to protect against severe disease but not necessarily against infection (8,12,14). This parallels the reduced severity observed for repeat infections with seasonal strains (22,23,25). Furthermore, re-exposure to a seasonal subtype typically elicits memory responses against the immunodominant HA head, which mask subdominant broadly-protective responses involved in group-level imprinting (26). For any country with suitable contact and demographic data, the methods shown here can provide rolling estimates of which age groups would be at highest risk for severe disease, should particular novel HA subtypes emerge. Such projections could guide cohort- or region-specific prevention, preparation, or control. Quantitative projections of changes in Reff, and hence pandemic risk, will require further research into the protection arising from matched imprinting: is some fraction of cases prevented entirely, and by what factor is infectivity reduced in mild cases arising in protected individuals? Our findings show that emergence risk cannot be considered in isolation, even for ‘novel’ pathogens that have not circulated in humans before. These pathogens are commonly assumed to have a blank slate of immunologically naïve humans to infect, but cross-protection from related pathogens can generate substantial population immunity. When this community of related pathogens undergoes major shifts, as during influenza pandemics, the landscape of population immunity changes accordingly. Thus emergence of novel pathogens can be governed by bottom-up control, with population immunity acting in an important and predictable manner to modulate the widely-recognized effects of virological and ecological risk factors. This perspective opens new frontiers for quantitative and mechanistic analysis of emergence risk. Supplementary Material Supplementary Text Supplementary data: Code for model fitting Table S1: H5N1 Line List Table S2: H7N9 Line List We thank the Lloyd-Smith lab and the Worobey lab for helpful comments, C. Viboud for providing insight into historic influenza data, T. Mega and S. Wu for assistance compiling data, B. Cowling for sharing poultry exposure data, and P. Horby for sharing Vietnam contact data. K.G. is supported by the National Institute of General Medical Sciences of the National Institutes of Health (T32GM008185). M.A. is supported by the National Science Foundation Graduate Research Fellowship (DGE-1144087). M.W. is supported by the David and Lucile Packard Foundation. J.O.L-S. is supported by the National Science Foundation (EF-0928690), the Research and Policy for Infectious Disease Dynamics (RAPIDD) program of the Science and Technology Directorate, Department of Homeland Security, and Fogarty International Center, National Institutes of Health. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. Case data and code for model fitting are available as supplementary data files. Fig. 1 HA groups and reconstruction of 20th century HA imprinting. (A) HA groups 1 and 2, and pairwise amino acid similarities in the HA stem region. Darker colored cells indicate higher similarity (see Fig. S1). Each within-group subtype pair is more similar (83.2%–97.8%) than any between-group pair (75.9%–81.7%). (B) History of seasonal IAV circulation, and (C) estimated fraction of each birth cohort in China with initial exposure to each subtype. Estimated patterns in other countries (not shown) are identical up to 1977, and very similar thereafter. Pandemic years are marked on the horizontal axis. Blue represents group 1 HA viruses, red represents group 2, and grey represents naïve children who have not yet experienced an IAV infection. Fig. 2 H7N9 and H5N1 observed cases and deaths by birth year (bars). Black lines show a priori prediction based on demographic age distribution and reconstructed patterns of HA imprinting. (A) 680 H7N9 cases, from China, 2013–2015. (B) 835 H5N1 cases, from Cambodia, China, Egypt, Indonesia, Thailand and Vietnam, 1997–2015. (C, D) Case data normalized to demographic age distribution from appropriate countries and case observation years. Fig. 3 Projected effects of HA imprinting on future pandemics. (A) Attack rate of severe cases, by age group, for hypothetical H5 (blue) and H7 (red) IAV pandemics in 2015 (R0=2.5, relative infectiousness of imprinting-protected individuals (α)=0.5), assuming UK demography and age-structured mixing (Supplementary Text). Lines show the average of 100 simulated outcomes, and shaded regions show the central 95%. Three vaccination scenarios explored: vaccination of IAV-naïve children could cause dual imprinting to both HA groups (dashed lines), prevent imprinting to both groups (dotted lines), or have no effect on imprinting (solid lines). (B) Projected change in Reff, for hypothetical H5 (blue) or H7 (red) IAV with R0=1.2 and α=0.5, if group 1 IAVs make up 100% or 75% of seasonal circulation after 2015. Table 1 Estimated protection from HA imprinting. Factors in model HA imprinting protection (95% CI) ΔAIC Akaike weight H5N1 DAH 0.75 (0.65–0.82) 0.00 0.9994 DEAH 0.83 (0.76–0.88) 15.35 4.65E-4 DEANH 0.83 (0.73–0.88) 17.32 1.74E-4 DH 0.80 (0.71–0.85) 69.18 9.50E-16 DEH 0.87 (0.80–0.90) 103.31 3.69E-23 DENH 0.86 (0.78–0.90) 105.29 1.37E-23 H7N9 DEAH 0.76 (0.67–0.82) 0.00 1.00 DAH 0.81 (0.74–0.87) 42.87 4.09E-10 DEH 0.84 (0.78–0.88) 61.59 4.23E-14 DENH 0.83 (0.75–0.88) 62.26 3.02E-14 DH 0.88 (0.84–0.92) 138.40 8.83E-31 D=Demography, E=Exposure to poultry, A=High-risk age groups, H=HA imprinting, N=NA imprinting (see Methods, Table S1). * This manuscript has been accepted for publication in Science. This version has not undergone final editing. Please refer to the complete version of record at http://www.sciencemag.org/. The manuscript may not be reproduced or used in any manner that does not fall within the fair use provisions of the Copyright Act without the prior, written permission of AAAS. The authors declare no competing financial interests. Figures S1–S12 Tables S1–S2 Databases S1–S3 References (35–74) References and Notes 1 Qin Y Differences in the Epidemiology of Human Cases of Avian Influenza A(H7N9) and A(H5N1) Viruses Infection Clin Infect Dis 61 563 71 2015 25940354 2 Cowling BJ Comparative epidemiology of human infections with avian influenza A H7N9 and H5N1 viruses in China: a population-based study of laboratory-confirmed cases Lancet 382 129 37 2013 23803488 3 Kucharski AJ Edmunds WJ Cross-immunity and age patterns of influenza A(H5N1) infection Epidemiol Infect 143 1119 24 2015 25115493 4 Sandbulte MR Cross-reactive neuraminidase antibodies afford partial protection against H5N1 in mice and are present in unexposed humans PLoS Med 4 e59 2007 17298168 5 Rivers C Lum K Lewis B Eubank S Estimating Human Cases of Avian Influenza A(H7N9) from Poultry Exposure PLoS Curr 2013 10.1371/currents.outbreaks.264e737b489bef383fbcbaba60daf928 6 Cowling BJ Preliminary inferences on the age-specific seriousness of human disease caused by avian influenza A(H7N9) infections in China, March to April 2013 Euro Surveill 18 20475 2013 23725807 7 Wang L Human exposure to live poultry and psychological and behavioral responses to influenza A(H7N9), China Emerg Infect Dis 20 2014 8 Wrammert J Broadly cross-reactive antibodies dominate the human B cell response against 2009 pandemic H1N1 influenza virus infection J Exp Med 208 181 93 2011 21220454 9 Pica N Hemagglutinin stalk antibodies elicited by the 2009 pandemic influenza virus as a mechanism for the extinction of seasonal H1N1 viruses Proc Natl Acad Sci U S A 109 2573 8 2012 22308500 10 Li GM Pandemic H1N1 influenza vaccine induces a recall response in humans that favors broadly cross-reactive memory B cells Proc Natl Acad Sci U S A 109 9047 52 2012 22615367 11 Miller MS Neutralizing antibodies against previously encountered influenza virus strains increase over time: a longitudinal analysis Sci Transl Med 5 198ra107 2013 12 Krammer F H3 stalk-based chimeric hemagglutinin influenza virus constructs protect mice from H7N9 challenge J Virol 88 2340 3 2014 24307585 13 Ellebedy AH Induction of broadly cross-reactive antibody responses to the influenza HA stem region following H5N1 vaccination in humans Proc Natl Acad Sci U S A 111 13133 8 2014 25157133 14 Palese P Wang TT Why do influenza virus subtypes die out? A hypothesis MBio 2 2011 10.1128/mBio.00150-11 15 Francis T On the Doctrine of Original Antigenic Sin Proc Am Philos Soc 104 572 578 1960 16 Lessler J Evidence for antigenic seniority in influenza A (H3N2) antibody responses in southern China PLoS Pathog 8 e1002802 2012 22829765 17 Smallman-Raynor M Cliff AD Avian Influenza A (H5N1) Age Distribution in Humans Emerg Infect Dis 13 510 512 2007 17552119 18 Terajima M Co MDT Ennis FA Age and different influenza viruses Lancet Infect Dis 14 101 2014 24457169 19 Worobey M Han GZ Rambaut A Genesis and pathogenesis of the 1918 pandemic H1N1 influenza A virus Proc Natl Acad Sci U S A 111 8107 8112 2014 24778238 20 Sauerbrei A Schmidt-Ott R Hoyer H Wutzler P Seroprevalence of influenza A and B in German infants and adolescents Med Microbiol Immunol 198 93 101 2009 19194722 21 Sauerbrei A Prevalence of antibodies against influenza A and B viruses in children in Germany, 2008 to 2010 Eurosurveillance 19 1 8 2014 22 Laurie KL Multiple infections with seasonal influenza A virus induce cross-protective immunity against A(H1N1) pandemic influenza virus in a ferret model J Infect Dis 202 1011 20 2010 20715930 23 Schulman JL Kilbourne ED Induction of partial specific hetertypic immunity in mice by a single infection with influenza A virus J Bacteriol 89 170 4 1965 14255658 24 Houser KV Pearce MB Katz JM Tumpey TM Impact of prior seasonal H3N2 influenza vaccination or infection on protection and transmission of emerging variants of influenza A(H3N2)v virus in ferrets J Virol 87 13480 9 2013 24089569 25 Wilkinson TM Preexisting influenza-specific CD4+ T cells correlate with disease protection against influenza challenge in humans Nat Med 18 274 280 2012 22286307 26 Andrews SF Immune history profoundly affects broadly protective B cell responses to influenza Sci Transl Med 7 316ra192 316ra192 2015 27 Kohler I Prevalence and predictors for homo- and heterosubtypic antibodies against influenza A virus Clin Infect Dis 59 1386 1393 2014 25139962 28 Sui J Structural and functional bases for broad-spectrum neutralization of avian and human influenza A viruses Nat Struct Mol Biol 16 265 273 2009 19234466 29 Henry Dunand CJ Preexisting human antibodies neutralize recently emerged H7N9 influenza strains 125 1 14 2015 30 Mossong J Social contacts and mixing patterns relevant to the spread of infectious diseases PLoS Med 5 e74 2008 18366252 31 Read JM Social mixing patterns in rural and urban areas of southern China Proc Biol Sci 281 20140268 2014 24789897 32 Horby P Social contact patterns in Vietnam and implications for the control of infectious diseases PLoS One 6 e16965 2011 21347264 33 Dauer CC Serfling RE Mortality from Influenza Am Rev Respir Dis 83 15 28 1961 34 Simonsen L Reichert TA Miller MA The virtues of antigenic sin: consequences of pandemic recycling on influenza-associated mortality Int Congr Ser 1263 791 794 2004 35 Chan PKS Outbreak of Avian Influenza A (H5N1) Virus Infection in Hong Kong in 1997 Clin Infect Dis 34 S58 S64 2002 11938498 36 Fiebig L Avian influenza A(H5N1) in humans: New insights from a line list of World Health Organization confirmed cases, September 2006 to August 2010 Eurosurveillance 16 1 10 2011 37 Kucharski A Data from: Distinguishing between reservoir exposure and human-to-human transmission for emerging pathogens using case onset data PLoS Curr Outbreaks 2014 10.5061/dryad.2g43n 38 Burnham KP Anderson DR Model Selection and Multimodel Inference: A Practical Information-Theoretic Approach Springer New York 2 2002 39 US Census Bureau International Programs, International Data Base 2015 [available at http://www.census.gov/population/international/data/idb/region.php?N=Results&T=10&A=separate&RT=0&Y=2013&R=−1&C=ID] 40 Coates BM Staricha KL Wiese KM Ridge KM Influenza A Virus Infection, Innate Immunity, and Childhood JAMA Pediatr 169 956 63 2015 26237589 41 Cromer D The burden of influenza in England by age and clinical risk group: A statistical analysis to inform vaccine policy J Infect 68 363 71 2014 24291062 42 Widdowson MA Monto AS Epidemiology of Influenza in Textbook of influenza Webster RG Monto AS Braciale TJ Lamb RA John Wiley & Sons Ltd Oxford, UK 2 2013 250 265 43 Bodewes R Prevalence of antibodies against seasonal influenza A and B viruses in children in Netherlands Clin Vaccine Immunol 18 469 476 2011 21209157 44 Neuzil KM Burden of interpandemic influenza in children younger than 5 years: a 25-year prospective study J Infect Dis 185 147 52 2002 11807687 45 Glezen WP Emerging infections: pandemic influenza Epidemiol Rev 18 64 76 1996 8877331 46 Global Influenza Surveillance and Response System FluNet 2015 [available at www.who/int/flunet] 47 Thompson WW Mortality associated with influenza and respiratory syncytial virus in the United States JAMA 289 179 186 2003 12517228 48 Palache A Oriol-Mathieu V Fino M Xydia-Charmanta M Seasonal influenza vaccine dose distribution in 195 countries (2004–2013): Little progress in estimated global vaccination coverage Vaccine 33 5598 605 2015 26368399 49 Kearse M Geneious Basic: An integrated and extendable desktop software platform for the organization and analysis of sequence data Bioinformatics 28 1647 1649 2012 22543367 50 Stamatakis A RAxML-VI-HPC: maximum likelihood-based phylogenetic analyses with thousands of taxa and mixed models Bioinformatics 22 2688 2690 2006 16928733 51 Fraser C Cummings DAT Klinkenberg D Burke DS Ferguson NM Influenza Transmission in Households During the 1918 Pandemic Am J Epidemiol 174 505 514 2011 21749971 52 Mills CE Robins JM Lipsitch M Transmissibility of 1918 pandemic influenza Nature 432 904 906 2004 15602562 53 Fraser C Pandemic Potential of a Strain of Influenza A (H1N1): Early Findings Science 324 1557 1561 2009 19433588 54 Viboud C Transmissibility and mortality impact of epidemic and pandemic influenza, with emphasis on the unusually deadly 1951 epidemic Vaccine 24 6701 6707 2006 16806596 55 White LF Estimation of the reproductive number and the serial interval in early phase of the 2009 influenza A/H1N1 pandemic in the USA Influenza Other Respi Viruses 3 267 276 2009 56 Yang Y The Transmissibility and Control of Pandemic Influenza A (H1N1) Virus Science 326 729 733 2009 19745114 57 Cowling BJ Fang VJ Riley S Peiris JSM Leung GM Estimation of the serial interval of influenza Epidemiology 20 344 347 2009 19279492 58 Public Health England Influenza: the Green Book, chapter 19 Version 10 2015 [found at https://www.gov.uk/government/publications/influenza-the-green-book-chapter-19] 59 Public Health England The National Childhood Flu Immunisation Programme 2016/17: Information for healthcare practitioners 2016 [found at https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/540515/Childhood_flu_programme_information_for_healthcare_practitioners.pdf] 60 Public Health England Seasonal influenza vaccine uptake amongst GP patients in England Provisional monthly data for 1 September 2014 to 31 January 2015 2015 [found at https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/407946/2903322_SeasonalFlu_GP_Jan2015_acc2.pdf] 61 Blank Patricia R Schwenkglenks Matthias Szucs Thomas D 2008 Influenza Vaccination Coverage Rates in Five European Countries during Season 2006/07 and Trends over Six Consecutive Seasons BMC Public Health 8 272 10.1186/1471-2458-8-272 18673545 62 Public Health England Seasonal flu vaccine uptake in children of primary school age: winter season 2015 to 2016 2016 [found at https://www.gov.uk/government/statistics/seasonal-flu-vaccine-uptake-in-children-of-primary-school-age-winter-season-2015-to-2016] 63 Owusu JT Seasonal influenza vaccine coverage among high-risk populations in Thailand, 2010–2012 Vaccine 33 742 7 2015 25454853 64 Kittikraisak W Influenza vaccination coverage and effectiveness in young children in Thailand, 2011–2013 Influenza Other Respi Viruses 9 85 93 2015 65 Zhou L Seasonal influenza vaccination coverage rate of target groups in selected cities and provinces in China by season (2009/10 to 2011/12) PLoS One 8 e73724 2013 24040041 66 Lau JTF Mo PKH Cai YS Tsui HY Choi KC Coverage and parental perceptions of influenza vaccination among parents of children aged 6 to 23 months in Hong Kong BMC Public Health 13 1026 2013 24171947 67 Centers for Disease Control and Prevention Influenza Vaccination Coverage among Children Aged 6–23 Months – United states, 2008–09 Influenza Season 2010 [available at http://www.cdc.gov/flu/fluvaxview/coverage_6-23months.htm] 68 Allison MA Influenza vaccine effectiveness in healthy 6- to 21-month-old children during the 2003–2004 season J Pediatr 149 755 762.e1 2006 17137887 69 Housworth J Langmuir AD Excess mortality from epidemic influenza, 1957–1966 Am J Epidemiol 100 40 48 1974 4858301 70 Olson DR Simonsen L Edelson PJ Morse SS Epidemiological evidence of an early wave of the 1918 influenza pandemic in New York City Proc Natl Acad Sci USA 102 11059 11063 2005 16046546 71 Noble GR Basic and Applied Influenza Research Beare AS CRC Press Boca Raton, FL 1982 11 50 72 Gover M Influenza and Pneumonia Mortality in a Group of 90 Cities in the United States, August 1935–March 1943, with a Summary for August 1920–March 1943 Public Health Rep 58 1033 1061 1943 19315929 73 Collins SD Lehmann J Trends and Epidemics of Influenza and Pneumonia Public Health Rep 66 1487 1517 1951 14875911 74 Ailling DW Blackwelder WC Stuart-Harris CH A study of excess mortality during influenza epidemics in the United States, 1968–1976 Am J Epidemiol 113 30 43 1981 7457477

PMC005xxxxxx/PMC5134742.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 101632887 42600 Microb Cell Microb Cell Microbial cell (Graz, Austria) 2311-2638 27917388 5134742 10.15698/mic2016.08.516 NIHMS818202 Article Similar environments but diverse fates: Responses of budding yeast to nutrient deprivation Honigberg Saul M. * Division of Cell Biology and Biophysics, University of Missouri-Kansas City, 5007 Rockhill Rd, Kansas City MO 64110, USA * Corresponding Author: Saul M. Honigberg, PhD, Tel: +1 816 235 2578; honigbergs@umkc.edu 22 9 2016 8 2016 02 12 2016 3 8 302328 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. Diploid budding yeast (Saccharomyces cerevisiae) can adopt one of several alternative differentiation fates in response to nutrient limitation, and each of these fates provides distinct biological functions. When different strain backgrounds are taken into account, these various fates occur in response to similar environmental cues, are regulated by the same signal transduction pathways, and share many of the same master regulators. I propose that the relationships between fate choice, environmental cues and signaling pathways are not Boolean, but involve graded levels of signals, pathway activation and master-regulator activity. In the absence of large differences between environmental cues, small differences in the concentration of cues may be reinforced by cell-to-cell signals. These signals are particularly essential for fate determination within communities, such as colonies and biofilms, where fate choice varies dramatically from one region of the community to another. The lack of Boolean relationships between cues, signaling pathways, master regulators and cell fates may allow yeast communities to respond appropriately to the wide range of environments they encounter in nature. pseudohyphal growth sporulation meiosis quiescence Boolean logic cell-cell signals yeast communities INTRODUCTION A) Alternative cell fates: Biology ain't always Boolean Most tissues contain multipotent stem cells-- i.e. cells able to differentiate into one or more cell types. The choice between fates depends largely on stimuli from the environment/niche of the cell. Often a given fate choice depends on multiple signals - some that promote and some that inhibit a particular fate. These observations suggest that Boolean logic may apply to cell-fate choice—i.e. a particular fate is adopted in response to the presence or absence of a particular combination (or combinations) of signals. For example, the lactose operon is active only when lactose is present and glucose is absent [1]. However, many fate choices do not fit easily into the framework of Boolean logic. Given the facile genetics and unmatched gene annotation in Saccharomyces cerevisiae [2], this yeast has served for many years as a model for the regulation of differentiation [3, 4]. The present review focuses on fate choices in diploid cells of the budding yeast, S. cerevisiae (Baker's yeast). I propose that many aspects of this choice are non-Boolean in nature. Diploid yeast can differentiate in multiple ways (Fig. 1). In particular, as nutrients become depleted, these cells differentiate in at least three distinct ways: 1) they can sporulate to form haploid spores (reviewed in [5, 6]), 2) they can switch into “pseudohyphal growth” (phg) to grow as elongated chains of cells (reviewed in [7–9]), or 3) they can enter a stable non-proliferative state known as “quiescence” or “stationary phase” where they age and eventually undergo programmed cell death (reviewed in [10–12]. The current review focuses on the mechanisms by which S. cerevisiae chooses between these several “nutrient-deprivation” fates and the biological functions of each choice. Because in nature individual yeast cells typically proliferate, differentiate, age and die all within the context of multicellular communities such as colonies and biofilms, a particular focus of this review is how cell-fate decisions occur within these communities. B) Central hypothesis: Similar environment - different fates The central hypothesis presented in this review is that the choice of cell fate of S. cerevisiae is determined by relatively small differences in nutrient environment, which are then reinforced by cell-cell signals. I term this central hypothesis the “similar environment, different fate (SEDF)” hypothesis. The SEDF hypothesis contrasts with a view in which each cell fate responds to discrete differences in environmental cues. Cell-fate decisions determined by discrete differences in cues can be expressed a Boolean relationship between these cues and a given cell fate. An example of a Boolean relationship between inputs and outputs is shown in Fig. 2A. Boolean logic requires that there are two states for each input (e.g. “1” and “0”) with respect to environmental cues. For example, if a response is linked to a threshold level (e.g. if a given fate requires the presence of a nutrient above a certain concentration), that would also be considered Boolean, since there are effectively only two states. An example of a non-Boolean relationship between input and output is shown in Fig. 2B. In this example the range of concentrations of a given cue that activate a particular cell fate depends on the concentration (not simply the presence or absence) of a second cue. Thus, the key feature of SEDF is that the relative level of multiple cues determines cells fate, not just their presence or absence. Either a Boolean or non-Boolean model is consistent with the observation that each fate occurs most frequently in some environments than others. However, in a Boolean model, as mentioned above, the environments that promote one fate are clearly discrete from the environments that promote a different fate. A Boolean relationship is represented by a theoretical landscape (Fig. 2Ci). The discrete red and blue peaks in this figure represent two discrete differentiation responses; the two axes represent increasing intensity of two environmental cues (e.g. increasing concentration oxygen and nitrogen). In contrast, in a non-Boolean model the environments that promote each fate can overlap (Fig. 2Ci). The landscape corresponding to the “red fate” in Fig. 2Ci shows an additional feature not allowed in Boolean models. Unlike the blue peak, in the red peak the two signals interact such that the response peak is not symmetrical relative to the axes. In other words, the optimal level for one cue is different, depending on the level of the second cue. There are three main reasons to propose yeast cell fate follows an SEDF (non-Boolean) rather than a Boolean model, as discussed throughout this review. First, all 3 types of diploid differentiation occur in very similar environments, so fate choice is probably determined not by the presence or absence of one or more extracellular signals, but by the relative amount of these signals, i.e. fate choice cannot be represented by Boolean logic. Second, the transcription factors and signal transduction pathways that regulate yeast cell differentiation are not, either alone or in combination, specific for only one form of differentiation. Indeed, not only is the relationship between fate choice and environmental cues not Boolean, neither is the relationship between fate choice and the activity of most regulators. Third, yeast differentiation fates, despite mechanisms ensuring their stability, are remarkably flexible without dramatic changes in environment. For example, many communities of yeast are partitioned into populations undergoing different fates. The main ramification of the SEDF hypothesis is to underline the importance of cell-cell signals in the context of yeast communities. As proposed below, the key mechanism that allows SEDF is cell-cell signaling that reinforces fate choice. Within these communities, the combination of environmental and cell-cell signals allows cell fate choice to be coordinated both temporally and spatially. In particular, I propose that relatively modest quantitative differences in nutrient environment (and signal pathways activity) are sufficient to efficiently specify a single cell fate because these small differences are reinforced by cell-to-cell signals. In the context of the theoretical landscapes described above, one can consider that these cell-cell signals allow discrete fate peaks even in a very similar environment—the effectively “sharpen” these peaks (Fig. 2Ciii). As a result, SEDF allows different regions of the community to adopt complementary fates within a relatively uniform environment. Furthermore, this type of coordination provides biological functions to the community that are not available to individual cells. As a result, these signals provide functions that may echo the origins of communication on this planet. LESSONS FROM THE WILD Before considering the evidence for SEDF, it is useful to regard possible differences between yeast found in nature (in which fate choices evolved) and yeast in the laboratory (in which fate choice can be studied). Recently, it has become clear that natural populations of S. cerevisiae exist throughout the world [13–15] and that these natural populations are different from domesticated yeast populations such as industrial or vineyard yeast as well as from clinical yeast isolates [16, 17]. In particular, the ability to isolate, genotype, and in many cases sequence the genome of these natural isolates of S. cerevisiae has revealed at least two important features of the evolution of cell fate choice in wild yeast - the stability of the diploid state and the diversity of ecological niches in which these yeast can be found. A) Natural S. cerevisiae populations are homozygous diploids Based on genome analysis of many wild yeast isolates, it is clear that the ploidy of S. cerevisiae isolated from natural environments is diploid rather than haploid or polyploid [18–20]. By way of contrast, in the lab S. cerevisiae readily proliferates as haploids, diploids or polyploids. Indeed, polyploidy (particularly tetraploidy) is very common among industrial and food-processing S. cerevisiae [21, 22], and polyploidy and even aneuploidy may occur relatively frequently in response to environmental stress (reviewed in [23]). The fact that haploid yeast have not been isolated from the wild does not necessarily mean that sporulation is rare in natural environments. Indeed, yeast strains isolated from nature generally sporulate and mate efficiently in the laboratory under a range of nutrient conditions [24–26]. However, any spores isolated from nature might not be recognized as such, since they would probably germinate, mate and form diploids as soon as first cultured in the lab. There are several reasons that haploids are likely short-lived intermediates both in nature and when first cultured in the lab. In the first place, after sporulation is complete, the four haploid products of sporulation, two of each mating type, are held tightly together in an ascal sac - the remnant of the cell wall of the parent cell. As a result, once nutrients are restored, haploids of opposite mating type efficiently undergo mating with other spores from the same ascus (intra-ascal mating) to restore the diploid state. Furthermore, any haploids that fail to mate with their sister spores would likely mate soon after beginning to proliferate, because most wild yeast are homothallic, meaning that a dividing cell produces a daughter cell of the opposite mating-type from the mother cell, allowing subsequent mating between mother and daughter. Finally, diploid cells proliferate more rapidly than haploid cells under most conditions [27], allowing diploids to out-compete haploids in the wild. The greater stability of the diploid state in S. cerevisiae contrasts with other yeast species such as S. pombe and Candida lusitaniae, which are both more stable in the haploid state. In these species, meiosis closely follows mating, rather than the reverse as in S. cerevisiae [28, 29]. One implication of intra-ascal mating in wild yeast is that inbreeding between sister spores is much more frequent than outbreeding of unrelated yeast [30]. Also contributing to the low rates of outcrossing is the fact that yeast grows and sporulates primarily in clonal communities, i.e. starting from a single cell. Consistent with high levels of inbreeding in natural populations, analysis of molecular polymorphisms including whole-genome sequencing of natural isolates reveal them to be largely homozygous. Indeed, natural yeast have significantly less heterozygosity than clinical or vineyard isolates [31, 32] perhaps because of less selection pressure from human associations on the natural yeast. Although inbreeding may be the rule for wild yeast, genome-wide sequencing indicates that outbreeding does occur [30] including between wild and domesticated yeast [13]. For example, outbreeding may occur in the gut of insect vectors such as wasps [33]. In fact, it has been suggested that the ratio of inbreeding to outbreeding can be regulated in wild yeast [34]. Haploids may be rare in the wild, but their ability to proliferate in the laboratory is extremely useful. Haploids can be easier to work with than diploids; for example, only a single allele must be deleted to eliminate the gene product. Furthermore haploids undergo many of the same fate choices as diploid cells. For example, haploids can switch from budding to filamentous growth, and haploids can enter quiescence and age. Although wild-type haploids are not normally able to initiate meiosis & sporulation, introduction of certain mutant alleles can bypass these constraints [35, 36]. To return to the evolution of fate choice, since yeast found in the wild are almost always largely homozygous diploids, it is clear that the nexus for yeast differentiation is the diploid cell as nutrients become depleted. Below I describe how both yeast's metabolism and its natural habitats provide the environmental cues that trigger differentiation. B) Natural Saccharomyces habitats are both diverse and changeable It is useful to consider the natural habitats/ecological niches of yeast in the wild with respect to the biology of fate choice. In the lab, yeast differentiate when nutrients becoming limited, not only because nutrient depletion slows or halts proliferation, but also because it directly promotes differentiation. As discussed in this section, it is particularly the changes in metabolism and hence nutrient environment accompanying late stages of growth that promote differentiation. The crux of S. cerevisiae metabolism is the “Crabtree effect” [37]. The Crabtree effect describes the observation that when glucose is available, yeast will metabolize (ferment) this sugar completely to ethanol, acetate and other non-fermentable carbon sources (NFCs), and these non-fermentable products are only themselves efficiently metabolized further (via respiration) once glucose is completely exhausted. This switch from fermentation to respiration, termed the diauxic shift, occurs even when oxygen remains plentiful throughout. Not surprisingly, while glucose is plentiful, multiple signal transduction pathways respond to this glucose to repress respiratory enzymes and metabolism of alternative carbon sources (reviewed in [38]). Thus, S. cerevisiae is primarily adapted to proliferate on glucose and convert it to non-fermentable carbon. Many yeast species do not display the Crabtree effect, and it has been suggested that this effect evolved in S. cerevisiae and other “Crabtree yeast” as a mechanism to out-compete surrounding microbial species because the ethanol produced by fermentation and secreted by the yeast inhibits the growth of these other species [39]. The Crabtree effect is also central to the economic power of S. cerevisiae because it allows the production of high levels of ethanol during fermentation and high levels of CO2 during respiration. Wild yeast have been observed to proliferate in nature primarily on plant matter rich in sugars, such as tree exudates or rotting fruit [14, 26]. In contrast, recent isolation of S. cerevisiae from sites throughout the world such as soil demonstrates that this species is capable of inhabiting a wide range of ecological niches that are not rich in glucose [15, 40]. It seems likely that in these latter niches, yeast is not proliferating and may exist primarily as spores. Nutrient environment changes rapidly both during late stages of growth and during differentiation itself. During late stages of growth, fermentable carbon sources, nitrogen and other nutrients become depleted whereas NFCs can remain relatively plentiful (Fig. 3). Indeed, both the intracellular and extracellular metabolome studies reveal dramatic changes in the concentrations of intermediary metabolites and amino acids during the diauxic shift [41, 42]. These changes continue as differentiation progresses. For example, analysis of sporulation in transcriptome [38, 43], proteome [44, 45], and metabolome [46, 47] studies indicate that the expression and activity of metabolic enzymes continues to fluctuate as sporulation progresses. In summary, the biology and natural ecology of yeast indicate that yeast is distributed widely, but that it may primarily proliferate on fermentable carbon sources (FCs). The Crabtree effect ensures that as cells begin to exhaust nutrients and slow or cease growth, NFCs remain relatively plentiful. As discussed in the Section “Shared environmental cues & distinct fates”, this nutrient environment is optimal for differentiation regardless of the particular fate (consistent with SEDF). Thus, both the metabolism and the ecology of yeast suggest that yeast evolved such that late stages of growth provide the environmental cues necessary for differentiation. C) A caveat regarding comparing laboratory and natural strains Regulation of cell differentiation by nutrient environment has been studied in many different laboratory strains of yeast. It is now evident that these strain backgrounds can vary significantly with respect to the relationship between environmental cues and differentiation fates. These results raise the question of how well fate choice in lab strains reflects the fate choices that evolved in the wild. Indeed, strain variants with altered differentiation responses may have been selected in early laboratory strains, descendants of which are now used in most modern laboratories [48]. The effect of strain backgrounds on differentiation responses is a critical consideration when synthesizing data from different studies done in different strain backgrounds. One example of phenotypic variation between common laboratory strain backgrounds is sporulation efficiency. Moreover, this variation extends to different isolates of industrial, clinical and wild yeast. Among some of these strains, variation in sporulation efficiency has been traced (e.g. by QTL analysis) to allele differences at a relatively few loci, e.g. the transcription factor Rme1 [32, 49, 50]. Furthermore, sporulation efficiency under a single condition likely underestimates the variation in sporulation capacity between strains. As one example, several common lab strain backgrounds (e.g. W303 and SK1) sporulate very efficiently under optimal conditions but sporulate much less efficiently than natural isolates on low concentrations of glucose [25]. Phg efficiency also varies considerably between strain backgrounds - both among laboratory strain backgrounds and among natural isolates of yeast [51]. In the strains that have been compared, this variation is again largely attributable to one or a few loci; for example, the transcription factor Flo8, a master regulator of phg [52–54]. In fact, several common laboratory strain backgrounds of yeast that lack a functional allele of Flo8 are completely unable to undergo phg, but can partially recover this ability when Flo8 is restored [55]. As with sporulation, variation in phg extends beyond efficiency under optimal conditions. For example, in some laboratory strain backgrounds (e.g. Ʃ1278b) phg occurs most efficiently in medium containing glucose, whereas in other strain backgrounds (e.g. SK1), phg occurs most efficiently in NFCs [56–59]. Finally, as with phg and sporulation, the rate of ageing in quiescent cells varies significantly between different laboratory strains [60]. In addition to allele variation for master regulatory genes such as RME1 and FLO8, the responsiveness of signal transduction pathways to environmental cues varies significantly between laboratory strains (reviewed in [61]). Many experiments connecting signaling and differentiation pathways have been done in only a single strain background, so it is wise to be circumspect in synthesizing results based in different strain backgrounds. An equally important point is that it is unlikely that any single laboratory strain background represents the “real” evolved response; indeed, the variation between laboratory strains is mirrored by variation between natural isolates. Obviously, experiments with lab yeast strains have driven and will continue to drive most of what we understand about fate choice, as they do for most yeast biology. However, in considering the implications of these experiments to fate choice, it is useful to remember possible differences between these lab strains and the natural yeast strains in which fate choice evolved. VARYING FATES, VARYING FUNCTIONS From a functional viewpoint, a key aspect of the choice between cell fates is that each fate has a different biological role. In this section, these roles are discussed, especially in the context of the nutrient limitation that triggers diploid yeast differentiation. A) Fate #1: sporulation - sex, food and energy The primary function of sporulation is to produce cells (haploid spores) that are more resistant to environmental stresses than the vegetative cells from which they derive. For example, spores resist antimicrobials, high temperatures and prolonged starvation to a much greater extent than vegetative cells [62–64]. This resistance derives in large part from the thick walls encasing each spore [65, 66]. Another aspect of spore resistance is that spores survive the insect gut much better than vegetative or quiescent cells [62]. Fruit flies and other insects are thought to be a major vector by which yeast spread from one ecological niche to another and outbreed to less related strains [67, 68]. Indeed, it has been proposed that the environment of the insect gut is the primary site at which the spore wall protects viability [5]. Another presumed function of sporulation is to increase genetic diversity in a yeast population as a result of meiotic recombination coupled with independent assortment of chromosomes. Indeed, meiotic recombination in yeast may also increase genetic variation as a result of increased mutation rates near meiotic recombination sites [69]. However, as mentioned earlier, most wild yeast communities are clonal populations of largely homozygous strains. Yeast meiosis could be important in generating diversity after relatively rare out-breeding or mutation events by stimulating loss-of-heterozygosity at new alleles. In any case, yeast meiosis likely provides a relatively minor selective advantage relative to the increased environmental resistance of spores. Sporulation as a response to limiting nutrients is particularly interesting in the context of the large energy requirement for the sporulation program. For example, energy is required to express hundreds of gene products required for meiosis and spore wall formation including many, such as spore wall proteins, that are produced to very high levels [44, 45, 70–72]. In addition, many of the cellular processes required in sporulation have additional energy requirements—e.g. DNA replication and chromosome segregation. This abundant expenditure of energy in response to nutrient limitation has been termed the sporulation “energy paradox” [73, 74]. It is likely that this apparent paradox is resolved through a combination of several mechanisms. In the first place, storage carbohydrates, for example glycogen, accumulate during late stages of growth and are subsequently utilized for energy during sporulation (reviewed in [75]). Indeed, mutants defective in accumulating these storage carbohydrates fail to sporulate [76, 77]. In the second place, deprivation for nitrogen (or other essential nutrients) when NFC is still abundant allows abundant energy production/respiration in the absence of cell division. Thus, yeast build environment-resistant spore walls by using both internal and external energy sources generated during growth. B) Fate #2: Quiescence - dormancy, ageing & death 1) Not dead, just quiet Quiescence is a differentiated state in which yeast cease growth (reviewed in [12]) and undergo genome-wide changes in transcriptional expression and chromosome topography [78–80], cytoskeletal organization [81] and cytosolic fluidity [82, 83]. Thus, like sporulation, quiescence is a response to nutrient deprivation that is likely to require an energy investment. One of the major functions of quiescence is the same as that of sporulation - resistance to environmental stresses. For example, activation of the Mpk1 cell wall integrity pathway in quiescent (Q) cells induces the induction of cell-wall repair genes [84, 85], and activation of Rim15 kinase in these same cells induces stress-resistance genes [86–88]. Quiescence in yeast has been studied primarily in haploids but is equally available to diploid cells. Interestingly, Q diploids are more resistant to environmental stress than growing cells but less resistant than spores to environmental stress [5]. Q cells do not remain viable indefinitely. As time passes and Q cells age, their viability diminishes. Thus, quiescence, aging, and eventual death can be considered progressive stages in a single differentiation pathway. 2) The universal fate choice— getting older vs. the alternative As stated above, cell death occurs naturally as cells age. For example, in suspended yeast cultures most cells have reached the end of their lifespan approximately 1-3 weeks after they have ceased growth [63, 89, 90]. Lifespan is ended through a programmed cell death (PCD) (reviewed in [91–93]). PCD in yeast displays many of the same cellular landmarks as apoptosis in higher organisms, such as DNA fragmentation, cell surface changes, and involvement of mitochondria and reactive oxygen species (ROS) [94, 95]. For this reason, yeast PCD is often referred to as “yeast apoptosis”. However, yeast PCD does not utilize all of the same regulators as mammalian apoptosis nor the extensive family of caspases typical of metazoan apoptosis [96, 97]. In this review, to avoid semantic distinctions, I refer to yeast apoptosis as PCD. Yeast lifespan is limited not only by the period of time that elapses after growth ceases (chronological ageing), but also by the number of times a mother cell can divide before it dies (replicative ageing) (reviewed in [98]). These two types of ageing are regulated by many of the same pathways (reviewed in [99, 100]), and they are also linked in the sense that chronologically aged yeast have shorter replicative lifespans than chronologically young cells [101]. Nevertheless, the two types of age are not interchangeable; for example, the Sir2 histone deacetylase inhibits replicative ageing, but Sir2 actually stimulates chronological ageing in some strain backgrounds and conditions (reviewed in [102]). In addition to ageing, yeast PCD is also triggered by a wide variety of other environmental stresses [103–106]. A common feature between ageing and most other triggers for PCD in yeast is the accumulation of oxidative and other cellular damage. According to one view of ageing, accumulation of damage over time eventually triggers PCD (reviewed in [107]). What is the relationship of nutrient environment to ageing and subsequent cell death? A common feature of replicative lifespan control from yeast to metazoans is that lifespan is increased when nutrients are limited, i.e. calorie restriction [108, 109]. Although the role of calorie restriction in lifespan extension is still a matter of debate, one idea is that when metabolism is limited, ROS and hence oxidative damage are also limited, and as a result lifespan is extended [110–113]. A corollary to this hypothesis is that inducing repair of oxidative stress also extends lifespan [114, 115]. Regardless of the of PCD trigger, its function in yeast and other single-celled organisms has been the subject of debate. One idea is that because yeast growth is largely clonal, programmed death in one cell could benefit other cells with the same genotype, thus providing a selective advantage for the genotype. For example, suspended cultures accumulate ROS after prolonged incubation, triggering PCD in most of the culture and presumably releasing enough nutrients from the dying cells to allow a subpopulation of still viable cells to continue growth and acquire adaptive mutations [116]. More generally, several investigators have proposed mechanisms by which programmed cell death benefits the overall (or average) survival of a clonal community in S. cerevisiae [92, 117] and other microorganisms [118–120]. The specific case of cell fate choice (including PCD) within communities is discussed further in the section “Shared communities – coordinated fates”. In summary, quiescence, ageing and PCD can be considered a single progressive pathway with different functions at earlier stages (e.g. resistance to environmental stress) than at later stages (e.g. possible re-distribution of nutrients). C) Fate # 3: Phg- the life of the forager 1) Overview looking for a better neighborhood Phg, like both quiescence and sporulation, is a response to diminished nutrients. However, phg is unique among yeast diploid differentiation fates in that it is also a means of cell proliferation. One likely function of phg is foraging. Specifically, phg allows yeast communities such as colonies to expand and access distant nutrients more efficiently than is possible when yeast divide through its standard (ovoid) budding patterning. Indeed, in many strain backgrounds phg occurs only in communities, not in suspended cultures [121]. However, in other strain backgrounds pseudohyphae will form under certain conditions even in suspended cultures [122, 123]. At least two types of foraging are associated with pseudohyphae: i) extension of chains of elongated yeast cells along the surface of the underlying substrate, and ii) invasion of these chains into the underlying substrate. 2) Exploring the surface The first type of foraging is closely related to the structure of pseudohyphae as chains of elongated cells. These chains radiate out from the perimeter of a colony or biofilm along the surface of the agar or other hard surface like plastic on which these communities grow. Limiting phg to the fringe of the community may be the most efficient mechanism to access distant sources of food. 3) Exploring below the surface As yeast colonies mature, they sometimes grow into (or “invade”) the underlying agar medium. Although invasive growth is often studied in haploids, diploids colonies are equally capable of invading agar. By the same reasoning as above, invasive growth potentially allows access to distal nutrients. In addition, invasive growth may provide benefits by anchoring a yeast community to its underlying substrate, and invasive growth of yeast into fruit and other natural substrates have been observed [124]. By analogy, the much longer hyphal and pseudohyphal filaments formed by the pathogenic yeast Candida albicans are necessary for these yeast to invade host tissue and, hence, for pathogenicity (reviewed in [125]). In theory, the elongated-chain geometry of pseudohyphae could provide the force necessary for invasion, but in fact the connection between pseudohyphae and invasive growth is not straightforward. Most laboratory strains that can form pseudohyphae can also invade agar [57, 59, 126]. However, many genes have been implicated in one program but not the other [127, 128]. Furthermore, in some wild and laboratory yeast grown on agar plates, the region where colonies invade the agar is not associated with extensive phg [25]. 4) Phg and phg-spectrum phenotypes Invasiveness is only one of a spectrum of wild-type phenotypes that require many of the same genes as phg but which do not always directly require phg. In particular, these “phg-spectrum” phenotypes all depend on expression of flocculins, which are a class of lectin-type proteins involved in both cell adhesion and cell signaling [129]. The diversity of flocculin-dependent phenotypes reflects the variety of communities that can form in this species. For example, the FLO11 flocculin is required to form “structured colonies”, which have with a distinctive lacey appearance [130, 131], “mats”, which are large thin colonies formed on low agar (high moisture) plates [132], “flors”, which are thin colonies of yeast that form on the top of liquid cultures [133], “flocs”, which are large clumps of cells that form suspended in cultures [58, 134], and “minicolonies” which are biofilm like structures that adhere to plastic surfaces submerged in medium [56, 133]. Both lab and natural yeast isolates vary considerably in their ability to undergo these phg-spectrum phenotypes [20, 131] D) Summary - distinct functions for discrete fates As discussed in the Section “Lessons from the wild”, all three differentiation fates are a response to a limited-nutrient environment, but as we saw in the Section “Varying fates, varying functions”, each fate has distinct functions and costs from the other two cell fates. For example, both quiescence and sporulation result in cells that are more resistant to the environment, but the higher resistance of spores relative to quiescence comes at the expense of a sizeable energy investment. Phg, a response to less severe nutrient depletion than the other two fates, functions more for foraging than for resistance to stress. Because fate choices are functionally quite distinct, it is striking, as discussed below, that they all respond to many of the same environmental cues, signaling pathways, and master regulators. OVERLAPPING REGULATORS BUT DISTINCT FATES A) Introduction to the master regulators of differentiation At one level, differentiation fate is determined by expression of master regulator(s). Master regulators are gene products that are required to initiate differentiation; indeed by one definition, a true master regulator is sufficient to trigger differentiation when expressed ectopically under conditions where that fate is normally suppressed. In a Boolean logic system, each fate would be defined by the presence or absence of one or more master regulators. Indeed, each differentiation pathway in yeast requires one or more master regulators and subsequent expression changes in hundreds of genes. However, as described below several of these master regulators activate more than one differentiation pathway, repress more than one pathway, or activate one pathway while repressing another (Fig. 4). Phg is regulated by a set of transcription factors including Flo8, Ste12, Tec1 and Nrg1, all of which activate transcription of the flocculin gene FLO11 (reviewed in [135]). Sporulation is initiated by expression of the Ime4 RNA methylase, the Ime1 transcription factor and the Ime2 protein kinase (reviewed in [5]). Quiescence is activated by expression of the Xbp1 transcription factor [78, 79]. B) Activators of meiosis (IME1, IME2) also activate phg Surprisingly, two key master regulators of sporulation, the Ime1 transcription factor and Ime2 kinase, are also required for phg in some laboratory strain backgrounds. In particular, these genes are required for phg in the SK1 but not the Σ1278b background [57]. This result is consistent with an observation described above—in SK1, phg (like IME1 and IME2 expression) is induced by NFC, but in Σ1278b phg is induced by glucose, which strongly inhibits IME1 and IME2. The roles of Ime2 in activating both phg and sporulation in SK1 is broadened further when Ime2 homologs in other fungal species are considered. For example, Ime2p homologs in Aspergillus nidulans and Neurospora crassa activate the development of sexual structures, and Ime2 homologs in Ustilago maydis and Cryptococcus neoformans activate mating and filamentous growth (reviewed in [136]). Thus, the idea that a master regulator can control more than a single differentiation program extends across species. C) Rme1 and Ime4 activity and the sporulation/phg choice Rme1 is a transcriptional activator that inhibits sporulation and stimulates phg. Rme1 prevents sporulation in haploid cells, a critical function since haploid meiosis is inevitably lethal. Rme1, which is expressed at high levels in haploids exposed to sporulation conditions, prevents meiosis by activating transcription of IRE1 a non-coding gene approximately 1 kb upstream of IME1. Transcription of IRE1 through the IME1 promoter prevents IME1 transcription [137]. At the same time that Rme1 blocks IME1 transcription in haploids, it also activates FLO11 transcription to promote filamentous growth [138]. Rme1 is also expressed in diploids (though at lower levels), and hence may balance sporulation and pseudohyphal fates in diploids. For example, wild yeast isolates with relatively high Rme1 expression tend to have low sporulation and high phg, whereas strains with relatively low Rme1 expression have the reverse tendency [32]. Not surprisingly, both IME1 and FLO11 are regulated by a number of other transcription factors in addition to Rme1. Indeed, IME1 and FLO11 have among the largest upstream intergenic distance of any yeast gene - consistent with these genes having especially complex promoters [139]. Multiple nutrient signals are integrated in regulating FLO11 not only through the transcription factors that bind its complex promoter but also through regulating activity of the Msb2/MAPK pathway that activates some of these transcription factors [135]. The expression pattern of IME4 is opposite to that of RME1 — IME4 is expressed to higher levels in diploids than in haploids [140]. Like IME1, IME4 is regulated by an overlapping long noncoding transcript [141, 142]. Furthermore, Ime4 also acts opposite to Rme1 in balancing sporulation with phg. For example, Ime4 inhibits FLO11 transcript accumulation while promoting Ime1 transcript accumulation [141, 143]. Ime4 is an RNA N-6 adenosine methyltransferase that likely acts on many hundreds of RNAs [144]. Interestingly, N-6 adenosine methylation also regulates expression of large gene sets in plants and animals during differentiation and development (reviewed in [145]). Other proteins that (like Ime4) promote sporulation and inhibit phg include Bir1, which is homologous to IAP (inducer of apoptosis protein) [146], and Spo21, a 67 aa protein that localizes to the prospore [147]. D) Quiescence and sporulation: shared regulator & shared properties Xbp1 represses transcription of many genes by targeting the Rpd3 histone deacetylase to these genes [79]. Interestingly, Xbp1 is not only required for quiescence but also for sporulation, at least in part because Xbp1 represses transcription of G1 cyclins [148, 149]. These cyclins not only trigger the G1 to S transition, they also repress IME1 transcription and inhibit Ime1p nuclear localization [150, 151]. It is not known whether Xbp1 is required for all forms of quiescence. Most meiotic genes are not induced in quiescent cells, but at least some of the same metabolic enzymes and stress resistant enzymes are induced in both programs [46, 152]. For example, trehalose synthesis is required for both initiation of meiosis and maintenance of quiescent cell viability [76, 153]. Similarly, many proteins required for heat shock resistance are induced during both quiescence/ageing and sporulation [44, 154, 155]. It is possible that cells first enter quiescence and only then may sometimes also initiate sporulation. Alternatively, quiescence and sporulation may be mutually exclusive fates that share certain common regulators and target genes. E) Summary Master regulators of diploid cell fate do not display Boolean-logic relationships to fate choice, i.e. there is not a particular combination(s) of known master regulators that specify each fate (Fig. 4). Instead, the relationship between master regulators and cell fate is complex. Indeed, some regulators, such as Flo11 regulate only a single fate, whereas other regulators (Xbp1, Ime1, and Ime2) activate more than one fate. Finally, a third class of regulators (Rme1 and Ime4), activate one fate while repressing another. How these master regulators together determine fate choice is unknown, but evidence so far indicates that their relationship to fate choice is non-Boolean. SHARED ENVIRONMENTAL CUES & DISTINCT FATES A) Introduction: a common environment stimulates each fate Given that all three diploid differentiation programs occur as nutrients become limiting, how does a cell choose between programs? It is possible that a small difference in the concentration of one or more nutrients under these conditions would result in passing a concentration threshold required for activation of a single fate choice. This would be an example of a cue/fate relationship that could be represented a Boolean logic since there are only two states relative to the threshold. This section addresses the question of whether the evidence allows for a Boolean relationship between cues and fate. A summary of the relationships between environmental cues and differentiation pathways is shown in Fig. 3. Sporulation occurs under the specific condition of active respiration (NFCs being much higher than FCs), high pH, and depleted nitrogen and/or other nutrients (reviewed in [156, 157]). Quiescence, like sporulation, is triggered by the absence of at least one essential growth nutrient and alkaline pH, but (unlike sporulation) quiescence does not have a requirement for respiration (reviewed in [10, 158]. Finally, pseudohyphal differentiation occurs at intermediate-to-low nitrogen concentrations (reviewed in [135]), can also be induced by other cues such as fusel alcohols, and also responds to other cues in some strain backgrounds (reviewed in [159]). B) Carbon source Yeast can metabolize many different carbon sources, but from the point of view of cell differentiation, there are two main types of carbon source. The first type is the FCs, in particular glucose, which is fermented through glycolysis to produce the second class - the NFCs. During late stages of growth, NFCs, particularly ethanol and acetate, are metabolized to carbon dioxide, and the balance between glucose and NFC during late stages of growth is a critical determinant of cell fate (Fig. 3). 1) Carbon source, quiescence and sporulation Glucose even in low concentrations effectively inhibits the initiation of meiosis in most laboratory strains. Presumably, sporulation has evolved such that it only initiates after the favored carbon source for growth, glucose, is fully metabolized to NFCs. For sporulation to initiate, a second cue besides the absence of glucose is the presence NFCs. Indeed, sporulation requires not only the energy provided by respiration but also the specific presence of a NFC [160], and the continued presence of NFCs is required even at late stages of sporulation [161]. Thus, as the ratio of NFCs to glucose (or other FCs) increases at late stages of growth, the environment becomes increasingly optimized for sporulation (Fig. 3). Quiescence is classically defined as occurring after the complete depletion of both NFCs and FCs [12]. However, a core set of genes is induced during quiescence regardless of which nutrient (C, N or PO4) is limiting [162, 163]. Although quiescence can be induced in a range of environments, extracellular environment strongly influences the properties of Q cells. For example, metabolic and other biological properties of Q cells vary depending on which nutrient is limiting (reviewed in [10, 164]). These different properties suggest the existence of multiple types of Q cells depending on the presence or absence of particular environmental cues. For example, ethanol and acetic acid accelerate chronological ageing in Q cells [60, 99, 165, 166]. 2) Carbon & phg A change in the ratio between NFCs and glucose likely regulates phg in all strain backgrounds, but the optimal NFC/glucose ratio may vary between backgrounds. For example, phg occurs efficiently in the Σ1278b background when grown on glucose medium, but is much more efficient in the SK1 background when grown on acetate medium [56, 57]. Several results may help to explain this difference. In the first place, even in Σ1278b, high concentrations of glucose (2%) inhibit haploid filamentous and invasive growth [167, 168]. In the second place, assays for phg in glucose medium generally involve cells dividing for many generations before N limitation triggers phg (termed the dimorphic switch) [169]. Thus, at least some glucose has been converted to NFCs by the time phg initiates. In this respect, increased NFC/glucose ratio during late stages of growth may stimulate phg as well as sporulation; though the optimal ratio may be lower in Σ1278b than in SK1. C) Nitrogen supply 1) Nitrogen availability, growth and differentiation It is tempting to neatly classify fate choice in yeast as driven by the presence or absence of two or three nutrients; yet, in nature yeast must make cell fate decisions across a range of different environments. The relative nitrogen availability from different sources illustrates this point. Yeast is capable of assimilating nitrogen from many different sources; however, the efficiency of utilizing nitrogen varies considerably depending on the source [170]. For example, ammonium and glutamine are very good nitrogen sources for growth, whereas urea and tryptophan are poor sources. As discussed below, nitrogen “quality”, besides affecting growth rate, also affects fate choice. 2) Nitrogen and fate choice In laboratory cultures, nitrogen is often the nutrient that becomes limiting during late stages of growth. In addition to directly regulating differentiation (see Section “Same signal paths – different fates”), nitrogen limitation prevents sufficient G1 cyclin from accumulating to activate the G1-to-S transition (START). The START transition blocks both sporulation and quiescence, both of which can only initiate from G1. Furthermore, cyclin expression inhibits both transcription of IME1 [150, 151, 171] and nuclear localization of Ime1 [172]. Similarly, cyclin expression also blocks the establishment and maintenance of quiescence [79]. Finally, once cells do enter quiescence, higher quality N sources accelerate ageing relative to lower quality sources [173]. An intermediate level of nitrogen is required both for phg and many phg-spectrum phenotypes [59]. For example, FLO11 is induced when low concentrations of nitrogen are present in the medium but repressed both at high nitrogen concentrations and when nitrogen is completely absent [174]. Similarly, both FLO11 induction and phg require intermediate G1 cyclin levels [128]. Thus, phg requires both detecting and utilizing intermediate N concentrations [175–177]. In summary, nitrogen level/availability regulates cell fate, not just its presence or absence or its concentration relative to a single threshold. Thus, the role of nitrogen in regulating differentiation cannot be represented by a Boolean “1” vs. “0” relationship between this cue and cell fate. D) The simplest of signals: pH 1) Extracellular pH fluctuates dramatically during growth Changes in extracellular pH coincide with nutrient depletion and contribute to fate choice. Proliferation is much less sensitive to external pH than is differentiation - yeast grows efficiently through a broad pH range (pH 3–8) - in part because acid pumps provide a relatively uniform intracellular pH regardless of external pH [178]. However, extracellular pH does change dramatically during both proliferation and differentiation. During fermentative growth, the secretion of organic acids such as acetate and pyruvate decreases external pH ≤ 4.0 [166]. In contrast, in media containing plentiful nitrogen and other nutrients, subsequent respiration of these organic acids during the diauxic shift converts these acids to C02, which is either released as a gas or solubilized as bicarbonate. As a result, during the diauxic shift, extracellular pH increases to ≥ 8.0. In this respect, extracellular pH reflects the ratio of FCs to NFCs [179]. Note that if the ratio of glucose to other essential nutrients is high enough in the chosen growth medium, then glucose will never be exhausted, and cells will not undergo either the diauxic shift or the second pH transition. 2) Extracellular pH has complex effects on cell differentiation Increased extracellular pH during late stages of growth stimulates quiescence [180]. Likewise, when Q cells are exposed to low pH, their viability (chronological lifespan) is strongly diminished [166, 181]. Increased extracellular pH also activates sporulation, and this pH continues to rise as sporulation progresses [182, 183]. Finally, pH has varying effects on phg-spectrum of phenotypes. For example, flocculation is stimulated by acidic pH [58, 184, 185], whereas, invasive growth is stimulated by alkaline pH [186]. E) Is it really just all about the food? In natural environments, yeast must adapt to temperature fluctuations over the course of the day and the course of the year. In general, diploid differentiation is more sensitive to temperature than is proliferation. For example, most laboratory strains are unable to sporulate at moderately high temperatures (> 34°C) even though growth is still efficient at temperatures exceeding 37°C [187]. Similarly, many clinical yeast isolates are able to grow as ovoid cells at much higher temperatures than as pseudohyphae [188, 189]. Finally, as cells enter quiescence, resistance to heat and other stress initially increases and then declines as cells age [190, 191]. Another aspect of cellular environment is cell-cell contacts. The role of cell-cell contacts in cell-fate decisions is most clear for phg, where cell adhesion molecules like the flocculin, Flo11, and the mucin, Msb2, are required for many phg-spectrum phenotypes (reviewed in [192, 193]). Sporulation in minicolonies might also be regulated by cell-surface contacts, since only the pseudohyphae projecting from minicolonies, not cells in the core, are capable of sporulating [56]. F) Summary Given that all three types of diploid differentiation occur in the nutrient-depleted environment characteristic of late stages of growth, it is not surprising that the environmental cues that regulate them are similar, but there are two points worth emphasizing. First, although each pathway occurs independently, there is no single combination of the presence or absence of cues that unambiguously specifies a single fate (Fig. 3). Second, the level or relative concentration of an environmental cue (e.g. ratio of NFCs to FCs or low, intermediate or high nitrogen) often correlates better with fate choice than simply the presence of that cue above some single threshold concentration. Both findings support the SEDF hypothesis and argue against a Boolean relationship between cues and fates. SAME SIGNAL PATHS - DIFFERENT FATES A) Overview It is striking that all three diploid differentiation pathways not only occur in similar nutrient environments but are also regulated by the same four major nutrient signal transduction pathways (PKA, TorC1, Snf1 and Rim101). As with master regulators and environmental cues, in the case of signaling pathways, one can imagine pathways in only one of two states, ON or OFF, relative to pathway targets including differentiation programs. Alternatively, there could be multiple or even continuous (graded) pathway activity levels. As one example, variable numbers of receptors could result in nearly continuous activity levels. The mechanisms and components of signal transduction have been discussed in several excellent recent reviews [3, 61, 194, 195], so this section of the viewpoint focuses only on the role of these pathways in regulating fate choice and the possibility of a Boolean relationship between signal pathways and fate choice (Fig. 4). B) Complex relationship between signal paths and nutrient signals Study of most yeast nutrient signaling pathways initiated with a focus on the relationship between a single nutrient and a single pathway. For example, the TorC1 pathway is primarily activated by nitrogen, the Ras/PKA pathway is primarily activated by glucose and the Snf1 pathway is primarily repressed by glucose. However, it is now apparent that most nutrient-responsive pathways relay information about more than one type of nutrient. For example, multiple receptors responding to different classes of ligands can converge to regulate the same signaling pathway. The Snf1 pathway exemplifies the ability of a single pathway to respond to diverse cues. Although this pathway was identified and characterized as active when glucose is absent, and is often referred to as the “glucose repression pathway”, it also responds to other types of cellular stress (reviewed in [196]). For example, in the absence of glucose, this pathway responds instead to nitrogen levels [197]. Similarly, the TorC1 pathway is not only activated by abundant nitrogen, but also by glucose, by osmotic stress, and by other types of cellular stress [198]. Finally, the PKA pathway, which is activated in high glucose through the Ras branch of the pathway, is also regulated through other branches that respond to carbon source, ammonium, amino acids and phosphate (reviewed in [199]). In fact, even the Ras branch of the PKA pathway is not regulated solely by glucose but is sensitive to acetate levels [200] and may respond more generally to cytosolic pH [201, 202]. C) PKA inhibits all three forms of differentiation PKA is active during rapid growth in FCs and has reduced activity in Q and sporulating cells. In yeast (and metazoans) PKA activates both the expression and activity of metabolic enzymes required for rapid growth. In addition, PKA represses genes required in non-proliferating, slow growing, and respiring cells (e.g. glycogen storage). Thus, PKA activity is low during all three types of cell differentiation. PKA inhibits differentiation through both general mechanisms, which repress all three differentiation programs, and specific mechanisms, which repress only one (or two) of the three differentiation choices. One general mechanism for inhibition occurs when PKA phosphorylates Whi3 protein [203]. Once phosphorylated, Whi3 releases CLN3 mRNA, allowing progression of the cell cycle from G1 to S (START) and corresponding inhibition of quiescence and sporulation [203–205]. Another general mechanism by which PKA inhibits all three forms of differentiation is by phosphorylating and inactivating Rim15 kinase [206]. Rim15 is required for the metabolic changes that accompany quiescence [207] and also activates the endosulfine, Igo1, which helps maintain mRNA populations during quiescence [208–210]. Rim15 also activates the Msn2/Msn4 transcription factors, which directly activate stress response genes necessary for quiescence [86, 211], and Rim15 is also required for the lifespan extension caused by calorie restriction [115, 212]. In addition to its role in quiescence, Rim15 is required for both the expression and activity of Ime1 [213–215]. Thus, Rim15 activates sporulation in part through different mechanisms from those it uses to activate quiescence [216]. Finally, Rim15 is required for at least some phg-spectrum phenotypes, such as the formation of structured (lacey) colonies [131]. PKA also inhibits differentiation through regulators specific to just one or two differentiation fates. For example, PKA phosphorylates and activates the Sok2 transcriptional repressor [217, 218]. Active Sok2 directly represses IME1 transcription, hence blocking sporulation [219] Sok2 also indirectly represses FLO11 and other genes required for phg [220, 221]. PKA also specifically inhibits spore morphogenesis by inhibiting Smk1 MAPK, which activates this late stage of sporulation [222]. Finally, PKA also specifically inhibits both quiescence and phg by directly (and indirectly) inactivating Yak1 kinase [223, 224]. Yak1 is a key activator of proteins required for quiescence [225, 226] and phg [227]. Thus, PKA represses all three forms of diploid cell differentiation through an array of general and specific mechanisms. D) TorC1 pathway inhibits sporulation and quiescence/ageing, but activates phg TorC1 is one of two Tor complexes in yeast with largely separate roles. When nutrients (especially N) are readily available, the TorC1 pathway activates cellular processes needed for rapid growth, such as protein translation. At the same time, this pathway represses processes induced during nutrient limitation such as utilization of poor nitrogen sources, autophagy, and the stress response. Fully activated TorC1 directly represses both sporulation and entry into quiescence. TorC1 represses these programs through activating Sch9 kinase, which in turn inhibits Rim15 activity/nuclear localization [228]. Thus, the TorC11 pathway acts in parallel to PKA in repressing Rim15 (reviewed in [229]). TorC1 also specifically represses sporulation by preventing nuclear localization/activation of Ime1p [172]. Interestingly, a low level of TorC1 is required for IME1 expression; thus sporulation only initiates with moderate TorC1 activity [230]. In addition, once cells become quiescent, activation of TorC1 diminishes their viability (i.e. promotes ageing) [89]. Indeed, calorie restriction slows ageing by decreasing Sch9 activity [229, 231, 232]. In contrast to its role in inhibiting these other differentiation programs, TorC1 stimulates phg through a second branch of the TorC1 pathway, which activates the Tap42 phosphatase [233]. Tap42 in turn stabilizes the Tec1 transcription factor, which binds FLO11 and activates its transcription [234]. Reflecting that optimal phg requires intermediate N concentrations (see Section “Shared environmental cues & distinct fates”), phg (like sporulation) also requires an intermediate level of TorC1 activity [174]. E) Snf1 kinase activates all three forms of differentiation The Crabtree effect depends on the ability of extracellular glucose to repress enzymes required to metabolize other carbon sources. As levels of glucose and other nutrients decline during late stages of growth, repression of Snf1 kinase activity is released. Activated Snf1 in turn triggers the expression and/or activation of enzymes required for respiration and resistance to stress [235–238], and at the same time stimulates transcription of genes required for differentiation. For example, Snf1 activates FLO11 transcription by phosphorylating Nrg1/Nrg2 transcription factors [167, 239, 240]. Similarly, Snf1 is required for IME2 transcription even in the absence of respiration [150]. Finally, either hyper-activating or deleting Snf1 shortens chronological lifespan [115, 241, 242]. Thus, maintaining viability during quiescence may require intermediate levels of Snf1 activity, similar to the intermediate TorC1 activity required for phg and sporulation. F) The Rim101 pathway activates both filamentous and meiotic fates The Rim101 pathway is primarily activated by alkaline pH (reviewed in [243], and as cells undergo the diauxic shift and extracellular pH increases, this pathway stimulates both pseudohyphal differentiation and sporulation [186]. For example, both FLO11 induction and IME1 induction require Rim101 [182, 244–246]. Indeed, a RIM101 polymorphism underlies many of the differences between Σ1278b and S288C transcriptomes during growth [247]. G) Other pathways and pathway interactions The above signaling pathways are unlikely to be the only ones regulating differentiation. For example, the Rgt2/glucose induction pathway, which is activated by glucose, represses Ime2 protein stability [248, 249], the Hog1 MAPK pathway, which responds to high osmolarity, inhibits pseudohyphal differentiation [250, 251], and the cell wall integrity (protein kinase C) pathway, which responds to cell-wall stress, is required both to maintain quiescent cells [252, 253] and for pseudohyphal differentiation [254–256]. Finally, the Hac1 (unfolded protein response) pathway, which responds to nitrogen starvation, mediates induction of both sporulation and phg [257]. Not only does each of the nutrient-signaling pathways in yeast respond to multiple environmental cues, but also these pathways are not insulated from one another. Instead, the PKA, Snf1, TorC1 and other pathways are densely intertwined by virtue of shared components and shared targets (reviewed in [3, 61, 194]). As just one example, the TorC1 pathway stimulates pseudohyphal differentiation in part by inhibiting Snf1 kinase [258], whereas the Rim101 and Snf1 pathways converge in both inhibiting Nrg1, a transcriptional activator of FLO11 [259]. Thus, nutrient controls on differentiation occur through an intertwined network of pathways. H) Summary: signaling pathways and fate choice As can be seen from Fig. 4, there is not a Boolean relationship between the four major nutrient signaling pathways and fate choice. In other words, fate choice is not determined by the ON or OFF state of one or more pathways (even relative to a threshold level of activity). Indeed, most of these pathways act in the same way on multiple fates. One exception is the TorC1 pathway, which activates phg while interfering with the other two pathways. However, even TorC1 does not act in a strictly Boolean manner, since several lines of evidence suggest an intermediate level of TorC1 activity, rather than either the fully active or fully inactive state is what activates both phg and sporulation. More generally, because these pathways regulate initiation of all 3 fates, the level of activation of each pathway, rather than a Boolean state, determines fate choice. Further work is required to ask if the activity level of one pathway relative to another contributes to fate choice. FLEXIBILITY & STABILITY IN DIFFERENTIATION CHOICE A) Fate choice stability Another aspect of the relationship between fate choice and environment is the stability and reversibility of this choice. At one level, all three types of differentiated cells are reversible, in the sense that if sufficient nutrients are restored, pseudohyphae, Q cells and asci all re-enter standard cell proliferation. Furthermore, under some circumstances yeast can switch from one fate to another without intervening growth as undifferentiated cells. In this section, the SEDF hypothesis will be considered in the context of the stability /reversibility of fate choice. B) Epigenetic mechanisms stabilize cell fate Each fate is in part stabilized by genome-wide epigenetic changes. For example, although pseudohyphae can produce ovoid cells when restored to a very nutrient-rich environment, phg is relatively stable to small changes in environment [59, 159]. One mechanism for this stability is that the FLO11 promoter is bi-stable, i.e. it stays in either the ON or OFF state for many cell divisions before switching, and this switch depends on chromatin structure and in particular on histone modifying enzymes [260, 261] and long non-coding RNA [262–264]. FLO11 expression is also stabilized by a prion-like form of the transcription factor, Mot3. Soluble Mot3 represses FL011 transcript, and this repression is released when Mot3 forms heritable prions [265]. Both quiescence and sporulation are also characterized by chromatin modifications. For example, genome-wide alterations in histone modification accompany both progression through sporulation [266, 267], and transcriptional inactivation after sporulation is complete [268]. Similarly, quiescence is characterized by both global changes in histone modification [78, 269], and positioning of RNAPII upstream of many genes poised for induction when nutrients are restored [270]. Once cells initiate sporulation, it is only during the early stages of this program that the meiotic fate is directly reversible. Reversibility in this context means that cells in early stages of meiosis can directly re-enter proliferation if resupplied with nutrients (“return to growth”). The loss of this reversibility, termed “commitment to meiosis”, may also have its basis in epigenetics. More specifically, when cells in these early stages are resupplied with growth nutrients, they exit sporulation and resume cell division; however, at about the time of the first meiotic division, cells become irreversibly committed to meiosis, meaning that they complete sporulation even when resupplied with growth nutrients [271, 272]. Commitment to meiosis may involve a positive feedback loop regulating transcription of NDT80 [273, 274], and as with FLO11, the silenced state of NDT80 requires a histone deacetylase [222, 275]. Commitment to meiosis may also involve a second epigenetic mechanism — Rim4 forms an amyloid-like protein that binds transcripts during sporulation to delay their translation until late stages of the program [276], and it may be that binding of the amyloid to these transcripts also protects them from degradation after nutrients are resupplied [277]. C) When yeast change their fate In some circumstances, cells can switch from one differentiation fate to another. For example, the SK1 laboratory strain background can switch from meiotic development to pseudohyphal development in response to a changing environment [143]. Even committed meiotic cells, if arrested at late stages of meiosis, can eventually re-enter the cell division cycle as pseudohyphae [278, 279]. Even without shifting media or environment, biofilm-like communities of the SK1 strain background form pseudohyphae at their surface that subsequently undergo a further differentiation to form linear asci [56, 57]. Similarly, as noted in the Section “Overlapping regulators but distinct fates”, the shared expression pattern between quiescence and sporulation suggest that cells first enter quiescence and then proceed to sporulation, again without being transferred to a new environment. These several lines of evidence indicate that each of three fates can be temporally coordinated with at least one other fate. D) Summary A variety of epigenetic mechanisms likely provide stability to yeast differentiation fates, but it is interesting that both phg and sporulation fates are stabilized by amyloid/prion proteins as well as by chromatin modification. Despite these mechanisms, differentiating cells can return to undifferentiated proliferation in response to environmental changes - in some cases even before differentiation is complete. Moreover, yeast can switch from one differentiation pathway to another even in the absence of dramatic environmental changes. The reversibility and flexibility of fate choice in yeast is another argument that distinct fates can occur in a similar environment (SEDF), and against a Boolean mechanism of fate choice. SHARED COMMUNITIES - COORDINATED FATES A) Introduction Yeast grow in nature as communities, not as suspended cultures. Given that different fates respond to the same cues and regulators and that these fates sometimes inter-convert, one might expect that in the same population and community, some cells might adopt one fate and some cells another. In fact, as described below, this is exactly what happens both in communities and suspended cultures. Thus, the question of whether different fates are triggered by discrete cues and regulators can be asked in the context of yeast populations. As described below, communities that contain cells adapting different fates are particularly relevant to the SEDF hypothesis. B) Fate partitioning within populations 1) Fate choice in cultures Before discussing fate choice within communities, it is worth considering fate choice in suspended cultures. Microenvironment can vary between different regions of a community, whereas suspended cultures provide the opportunity to measure fate choice within a population with both uniform genotype and uniform environment. One aspect revealed in suspended cultures is the probabilistic nature of fate choice. That is, depending on the conditions of the cultures, a given fraction of the culture adapts a particular fate. As described below, the probabilistic aspect of fate choice is clearest in cultures in which the population in the culture divides into two alternative differentiation fates. Sporulation is particularly useful in comparing cell fates. For most laboratory strains suspended sporulation cultures (e.g. media containing only 2% potassium acetate) contains both sporulated and unsporulated cells. The unsporulated sub-population likely corresponds to Q cells, since this subpopulation can be further subdivided, by cytological markers into viable cells, those undergoing PCD, and those already dead [280]. Similar to diploid cultures, haploid suspended cultures grown for extended periods form at least two populations of cells, designated Q and non-quiescent (NQ) cells. Q cells differ from NQ cells in many properties, including density, stress resistance, lifespan, transcriptome and proteome [152, 281, 282]. 2) Examples of the spatial organization of cell fate within communities As they proliferate, yeast naturally form tightly packed multi-cellular communities such as colonies and biofilms (reviewed in [283, 284]). Strikingly, differentiation is not homogenously distributed throughout the community, but instead occurs in specific regions. Below are four examples of yeast communities in which cell fate is spatially organized in communities either with respect to undifferentiated cells or with respect to cells adapting a different fate. a) Sporulation patterns in diploid colonies: Patterns of sporulated cells in colonies are easily visible in embedded sections of colonies [182, 285]. Specifically, asci are found throughout the top half of mature colonies and also in a thin layer of cells near the agar surface. In contrast, asci are almost completely absent throughout the broad central layer of the colony. Furthermore, the boundaries between sporulating and non-sporulating cells are sharply defined, and this same pattern is observed in a range of laboratory and natural yeast strains. Indeed, strains newly isolated from the wild form this pattern on both FCs and NFCs and on both synthetic and rich media [25]. The underlying layer of unsporulated cells can be considered a type of quiescent cell, and are termed “feeder cells” because they remain viable for many days, increase in permeability and stimulate sporulation in the overlying cell layer [74]. As discussed in the Section “Cell-cell signals in communities determine cell fate”, this stimulation is thought to occur because feeder cells provide nutrients to overlying cells, with the increased permeability dependent on induction of the CWI pathway. b) Upper (U) and Lower (L) cell layers in haploid colonies: Haploid colonies form a similar pattern of differentiation as the diploid colonies described above, although of course they do not sporulate. Haploid colonies form two sharply defined layers of cells-- termed U and L cells [286, 287]. U cells differ from L cells with respect to size, morphology, metabolism, gene expression, and viability. c) Sexual reproduction on the fringe: Both phg and sporulation occur specifically at the edge of some communities. For example, small colonies forming on agar media limiting for nitrogen (microcolonies) form a core of (typically-shaped) ovoid yeast with only the outer surface of the colony containing chains of pseudohyphae projecting from the colony [59]. Similarly, minicolonies, biofilm-like communities growing on plastic surfaces submerged in medium, switch from typical yeast divisions at the early stages of cell growth to phg at the periphery as the growth begins to slow [56]. Biofilms formed by pathogenic yeast such as C. albicans form a similar pattern in that the underlying layer of cells is comprised of ovoid cells, whereas the top layer consists of hyphae [288, 289]. As described above, the SK1 strain background both sporulates and undergoes phg very efficiently. As mentioned in the Section “Flexibility & stability in differentiation choice”, the pseudohyphae at the periphery of SK1 microcolonies and minicolonies subsequently sporulate to form linear 2-, 3, and 4-spore asci [56, 57]. In minicolonies, the timing of the transition from phg to sporulation has been followed by time-lapse microscopy. In these communities sporulation occurs synchronously around the colony edge but never in the interior of the colony. Furthermore, if phg is blocked, sporulation is also prevented [56]. Thus in these communities, phg and meiotic differentiation are sequential, rather than alternative, fates. d) PCD patterning in colonies: Like sporulation and pseudohyphal differentiation, PCD does not occur uniformly throughout communities. In spot colonies growing on a NFC for 1-3 weeks, apoptosis occurs mostly in the cells that form the core of the colony. In contrast, cells on the rim and outer surface of the colony maintain high viability [290]. It is possible that the high levels of PCD in the core increase the survival or growth among cells at the rim of the same colony, perhaps by providing scarce resources from the PCD region to the viable region [291]. 3) Sharp boundaries as a common characteristic of community patterning A striking characteristic of pattern formation in both colonies and biofilms is the sharply defined boundary surrounding regions of differentiation (reviewed in [284]). These boundaries can be visualized not only with respect to cytological markers like apoptosis, spore formation, and phg as described above, but also by patterns of gene expression [74, 182, 286, 287]. Indeed, on one side of a boundary, cells differentiate very efficiently to one fate, whereas neighboring cells in contact with these cells but on the other side of the boundary differentiate equally efficiently but to a different fate. C) Summary The SEDF hypothesis, that cells can adopt different fates in similar environments, is strongly supported by the sharp boundaries observed between neighboring regions of a yeast community. That is, cells on either side of a boundary share roughly the same environment but adopt distinct fates. In contrast, these boundaries are not easy to reconcile with a Boolean model of fate choice. In the next section we discuss one mechanism allowing SEDF - cell-to-cell signals. CELL-CELL SIGNALS IN COMMUNITIES DETERMINE CELL FATE A) Introduction: role of cell-cell signals in cell fate and patterning As described above, pseudohyphal differentiation, quiescence/aging and sporulation occur under similar conditions (SEDF), are regulated by many of the same signaling pathways and master regulators, and often occur together in the same community. Taken together these results argue against a Boolean relationship between input and cell fate. Given this, how do neighboring cells along a boundary efficiently adopt different fates, and how has yeast evolved to make the correct fate choice across the wide spectrum of environments found in nature? As discussed below, one possible answer to both of these questions is cell-to-cell signaling. Signals between individual yeast cells within communities are typically small diffusible molecules (reviewed in [284]). Below I consider two broad classes of cell-cell signals in yeast: 1) “enlistment” signals, which are produced by differentiating cells within one region of a community to stimulate cells within the same region to adopt the same fate, and 2) “diplomatic signals”, which are produced by differentiating cells in one region of the community to influence cells in an adjoining region to adopt a different fate. B) Enlistment signals—intercellular feedback stimulates differentiation & patterning Enlistment signals, by reinforcing the same fate choice in neighboring cells, contribute to the spatial organization of communities such that many of the cells within one region of the community adopt the same fate. 1) Alkali signals and sporulation patterning As described in the Section “Shared environmental cues & distinct fates” above, respiration of organic acids during late stages of growth and during sporulation leads to an increase in extracellular pH, which in turn stimulates sporulation in other cells of the same population. In colonies, after a narrow layer of cells near the center initiates sporulation, the resulting alkalization progressively drives expansion of the sporulating region upward to eventually include the entire top half of the colony [182]. This wave of sporulation depends on the Rim101 pathway both to sense and to produce alkali—suggesting an intercellular positive feedback loop involving pH signals. 2) Aromatic alcohols and pseudohyphal patterning Regulation of the dimorphic switch in microcolonies also involves metabolites and an intercellular positive feedback loop [292]. In the case of phg, these signals include the aromatic alcohols phenylethanol and tryptophol. These alcohols are produced and secreted during phg, and in turn stimulate phg in surrounding cells. Moreover, extracellular tryptophol induces further synthesis of tryptophol within pseudohyphal cells, which further amplifies the feedback loop [292]. 3) Role of ammonium in cycles of proliferation & quiescence Another example of a cell-cell signal that operates in yeast communities is the ammonium produced by haploid colonies after ≥ 14 d of growth. Cells on the surface of these aged colonies produce ammonium to much higher levels than cells in the core of these colonies [290, 293]. At the same time, surface cells induce expression of the ATO1 ammonium exporter [294]. Ammonium contributes to the survival of surface cells and to a new cycle of proliferation at the edge of the colony [295]. Because surface cells both produce and respond to ammonium, ammonium is another example of enlistment signals acting through an intercellular positive feedback loop. Indeed, ammonium can diffuse from one colony to its neighbors, synchronizing these colonies with respect to their growth/quiescence cycles [296]. In summary, enlistment signals such as alkali, aromatic alcohols or ammonium may all act through a common mechanism—an intercellular positive-feedback loop. The role of these signals varies; of the three signals discussed above, one activates sporulation, another activates phg, and a third activates proliferation. These feedback loops amplify relatively small differences in nutrient microenvironment to generate larger environmental differences. In this respect, intercellular positive feedback loops contribute to community pattern formation by localizing cell-cell signals to specific regions of a community. C) Diplomatic signals—crossing boundaries to influence behavior 1) The Rlm1 paradigm and the DPEB hypothesis In contrast to enlistment signals, diplomatic signals occur between community regions adopting different fates. An example of this second type of cell-to-cell signal occurs in sporulating colonies, where activation of the Rlm1 transcription factor in feeder cells stimulates sporulation in an overlying layer of cells through a cell non-autonomous mechanism [74]. Thus, sporulating colonies contain both recruitment and diplomatic signals (Fig. 5). This type of diplomatic signal may allow yeast colonies to sporulate in a wider range of environments than possible for individual cells, a hypothesis termed, “differential partitioning provides environmental buffering” or DPEB [73, 74]. According to this hypothesis, under optimal sporulation conditions, the colony is partitioned such that there are relatively few cells allocated to the feeder cell fate. In contrast, under suboptimal sporulation conditions, a greater portion of the colony is given over to the feeder cell fate, and overall colony sporulation is highly dependent on the nutrients and/or signals provided by these feeder cells. Thus, increasing the proportion of feeder cells buffers sporulation efficiency in suboptimal environments. Differential partitioning is a form of phenotypic plasticity within communities that is related in some ways to task allocation in social organisms such as ant, termite and bee species [297, 298]. 2) Last gasp diplomatic signal Another example of signals between different regions of the same colony may occur in the haploid colonies undergoing age-triggered PCD described in the Section “Shared communities – coordinated fates”. In these colonies, PCD in the colony core is postulated to provide nutrients to continued cell proliferation at the colony's rim [290]. Because intracellular ROS levels, including H2O2, which is relatively stable in an extracellular environment, rise in core cells even at relatively early stages of development [299, 300], it is possible that peroxide is a diplomatic signal. Thus diplomatic signals, either from feeder cells or PCD cells, may contribute to forming and maintaining the sharp boundaries between regions undergoing alternative fates. D) Summary: Boolean logic and cell-cell signals In several types of yeast communities, closely neighboring cells on either side of a boundary adopt alternative fates consistent with the SEDF hypothesis but not Boolean models. For yeast to adapt distinct fates in the very similar nutrient environments on either side of a boundary, cell-cell signals are likely essential. The cell-cell signals so far identified may represent only a fraction of the signals operating in yeast communities. Enlistment and diplomatic cell-cell signals cooperate to reinforce small differences in the nutrient microenvironment and are likely important to both establish and maintain boundaries. For example, enlistment signals within one region of a community can amplify differences via intercellular positive feedback loops; whereas, diplomatic signals between neighboring regions can help enforce the sharp boundary between regions. Similar mechanisms are critical to forming boundaries between tissues during metazoan development [301, 302]. Finally, cell-cell signals can provide biological function not available to individual cells, such as buffering the efficiency of a differentiation program against unfavorable environments (DPEB). ORIGINS OF COMMUNICATION & OF MULTICELLULARITY A) Looking backwards: ancient communities & the birth of communication What are the broader implications of the non-Boolean SEDF hypothesis and the biological function of pattern formation in yeast communities? Organized patterns within microbial communities date back to the earliest life on earth. For example, stromatolite fossils from billions (109) of years ago provide evidence of organization within cyanobacteria communities (reviewed in [303, 304]). The selection pressure for communication to evolve in microbes can be placed in the context of the more complicated forms of communication that evolved in complex organisms. In this respect, it is relevant that some of the most successful metazoan species on earth as judged by total biomass (e.g. humans, termites and ants), are those with highly developed modes of communication. It has been suggested that the “simple multicellularity” that exists in microbial communities differs from the “complex multicellularity” characteristic of plants and animals because of two properties present only in complex organisms: 1) many cells in complex systems do not make contact with the external environment, and 2) high levels of cell-cell communication [305, 306]. However, a close look at the biology of yeast communities calls these distinctions into question. B) The advantages of communication & community Cell signaling between yeast cells hints at the advantages gained from the evolution of communication. In this respect, most of the signaling molecules discussed above share two common characteristics. First, they are important in regulating cell differentiation. Second, they act within the context of a community of yeast, often to partition this community into different regions adopting distinct cell fates. One of the main advantages to yeast of growing within a community may lie in the ability of a community to partition into regions adopting different fates; these regions may cooperate to allow greater biological function than is possible for individual cells. For example, limiting pseudohyphal and meiotic differentiation to the edge of minicolonies and microcolonies maximizes the ability of spores to disperse from the colony [56, 57]. As a second example, in colonies the underlying layer of differentiated feeder cells presumably provides nutrients necessary for sporulation in the overlying cell layer, which again is more optimal for spore dispersal. Finally, PCD in the core of a haploid colony may provide the nutrients that allow proliferation at the colony's rim. Note that because natural yeast communities are largely clonal, the survival of the genotype depends on the overall survival of the colony, rather than competition between S. cerevisiae genotypes in the same community. To extrapolate further, the likely clonal nature of early microbial communities may have led not only to the evolution of the first cell-cell signals on earth but also to the first pattern formation. C) Model for co-evolution of signaling and multicellularity 1) From response to communication The ability of organisms to respond to their environment is expected even in the earliest life forms. Cell-to-cell signals in modern yeast may yield insight into these ancient signals. In particular, many yeast cell-cell signals are simply metabolites produced by nearby organisms such as alkali generated during respiration or ammonium produced by Q cells. Thus, the earliest form of communication between organisms may have evolved as metabolic byproducts coordinating growth with differentiation. 2) From communication to organization Cell-cell signals in yeast not only regulate cell-fate, they also organize communities such that different regions adopt different fates. By analogy, the response to primordial cell-cell signals may have evolved such that variation in signal concentration across a community contributed to the organization of this community and hence increased fitness. 3) From organization to multicellularity The patterning of cell types in microbial communities may presage cell type patterning within complex multicellular organisms. Indeed, it is conceivable that the earliest multicellular organisms evolved from microbial species with relatively sophisticated cell type patterns. Conversely, patterning in modern microorganisms may provide a useful model for some of the fundamental principles guiding pattern formation in all organisms. CONCLUSIONS A) Biology ain't always Boolean Similar environmental cues promote each of three alternative differentiation fates (SEDF), and these cues act through many of the same signal transduction pathways and master regulators to control fate choice. Thus, a Boolean representation tracking the presence or absence of a given cue or cues (or a discrete threshold) is not sufficient to describe the relationship between environmental cues and fate choice. Similarly, the relationship between fate choice and signal pathway/master regulator activity also cannot be accurately represented by Boolean logic. B) Graded and specialized responses So how is fate chosen? In the first place, it is likely that signal transduction pathways are not simply ON or OFF, but have a graded range of activities. By this view, the relative level of activity of signaling pathways determines cells fate as is seen for the TorC1 pathway. For example, the number of activated receptors may vary depending on the concentration of ligand. Indeed, the same signaling pathway might regulate different targets depending on its level of activity. In any case, the activity of a given signal transduction pathway likely also depends on interactions with other signaling pathways. C) Cell-cell signals reinforce differences and provide flexibility A second aspect to the choice of differentiation fates is provided by cell-cell signals. Relatively small differences in the microenvironment around cells can be reinforced or amplified by both recruitment and diplomatic cell-cell signals. These signals organize yeast communities into cooperative assemblies such that programs such as differentiation or proliferation occur more efficiently than is possible for individual cells. Of particular note is the ability of communities to adjust the allocation of cells to different fates within the community depending on the environment (DPEB). In microorganisms as in more complex organisms, cell-to-cell signals are fundamental to life and may be nearly as ancient. ACKNOWLEDGMENTS Preparation of this review was supported by funding from the National Institutes of General Medical Sciences of the NIH under award number R15GM094770. I thank Dr. David Spade (Univ. Missouri-Kansas City, UMKC) and Dr. Deendayal Dinakarpandian (UMKC) for helpful discussions, and I am grateful to Dr. Dinakarpandian for preparing Fig. 2C. Abbreviatons DPEB differential patitioning provides environmental buffering FC fermentable carbon source NFC non-fermentable carbon source PCD programmed cell death phg pseudohyphal growth Q quiescent ROS reactive oxygen species SEDF similar environment, different fate FIGURE 1 Alternative fates for diploid yeast S. cerevisiae typically have an ovoid shape when proliferating (1), and can differentiate to form chains of elongated pseudohyphal cells (2), rounded quiescent cells that subsequently age and succumb to programmed cell death (3), or tetrad asci, i.e. four haploid spores held together in an ascal sac (4). FIGURE 2 Boolean and non-Boolean relationships between input and output (A) Boolean truth table that represents the relationship between all combinations of the presence (1) or absence (0) of two possible inputs (A and B) and the occurrence of a given output. With respect to differentiation choices, examples of inputs could be the presence/absence of particular environmental cues or the activation/ inactivation of particular signaling pathways, and examples of outputs would be the occurrence (1) or not (0) of a particular type of differentiation. In an authentic Boolean truth table the response (as well as signal) would be only “1” or “0”, but for the example given, three alternative fates (F1–F3) are indicated for conciseness. As a result, this table can be considered a collapsed stack of truth tables, with one truth table for each possible fate. (B) Example of non-Boolean relationship between input and output. Rather than a given input being present or absent, the amount of input affects the output. In the context of differentiation choices, the amount of input could reflect the concentration of a particular environmental cue or the level of activation of a given signaling pathway. Note that in the contrived example shown, when the amount of input A is constant, output depends on the amount of input B not merely its presence or absence (compare row 2 and 3). (C) Environmental landscape graphs showing theoretical relationship between the efficiency/probability of cell fate (Z-axis) and two environmental variables (X- and Y-axes). The red and blue peaks represent two different cell fates. (i) In a Boolean landscape, fates are discrete, they never occur in the same environment, also Boolean response peaks are symmetric relative to the axes, so the blue peak is Boolean and the red peak is not. (ii) SEDF model is not Boolean since the two fate response peaks overlap. (iii) Even in the SEDF model, fates can be made discrete by reinforcing small differences in environment by cell-cell signaling. FIGURE 3 Environmental cues determine cell fate All three differentiation fates occur as nutrients become depleted, and this depletion provides at least three environmental cues that control differentiation fate as follows. (i) The ratio of non-fermentable carbon sources to fermentable carbon sources (NFCs/FCs) affects fate, with higher levels of NFCs stimulating sporulation (Sp). The arrow + bar shown linking NFCs/FCs to pseudohyphal growth (phg) reflects that in some laboratory strain backgrounds phg occurs efficiently when a FC source (glucose), is present, whereas in other strain backgrounds this differentiation occurs more efficiently in a NFC (acetate). (ii) Alkali increases in the environment during late stages of growth, and this alkali promotes both quiescence (Q) and sporulation. (iii) Nitrogen and possibly other essential nutrients (N) inhibit both, sporulation and quiescence. The arrow bar connecting N to phg represents that phg occurs most efficiently when intermediate levels of N are present, i.e. Phg is inefficient at either high N or in the absence of N. FIGURE 4 Roles of signal transduction pathways and master regulators on cell fate Known relationships between signal transduction pathways (filled rectangles), master regulators (open rectangles), and cell fate (abbreviations as in Fig. 3). The dotted arrow connecting the Ime1 and Ime2 master regulators (Ime1, 2) and phg represents that these master regulators are required for phg in some strain backgrounds but not in others. It should be noted that this diagram is meant as a working model of these relationships, other connections between pathways and regulators are likely. FIGURE 5 Recruitment and diplomatic signals combine to partition colonies into regions adopting different fates (i) After growth of colony is complete, cells in the underlying cell layer (beige) differentiate into a type of quiescent cells termed “feeder cells”. Feeder cells (designated by a chef's hat) remain viable but have increased permeability, allowing them to provide nutrients and/or other signals (red arrows) to the overlying layer of cells (light blue). (ii) These “diplomatic signals” between regions of the yeast community promote sporulation in upper layer cells near the border (tetrad asci are shown in magenta). As these first cells sporulate, continued respiration of acetate in these cells causes alkalization of the microenvironment. The alkali produced by sporulating cells is a “recruitment signal” (blue arrows) to trigger sporulation in surrounding cells in this layer. (iii) This intercellular positive feedback loop involving sporulation and alkalization results in an upward wave of sporulation eventually including the entire upper region of the colony. CONFLICT OF INTEREST I declare no conflict of interest with the content of my article entitled “Similar environments but diverse fates: Responses of budding yeast to nutrient deprivation”. REFERENCES 1 Veliz-Cuba A Stigler B Boolean models can explain bistability in the lac operon J Comput Biol 2011 18 6 783 794 21563979 2 Botstein D Fink GR Yeast: an experimental organism for 21st Century biology Genetics 2011 189 3 695 704 22084421 3 Zaman S Lippman SI Zhao X Broach JR How Saccharomyces responds to nutrients Annu Rev Genet 2008 42 27 81 18303986 4 Schneper L Duvel K Broach JR Sense and sensibility: nutritional response and signal integration in yeast Curr Opin Microbiol 2004 7 6 624 630 15556035 5 Neiman AM Sporulation in the budding yeast Saccharomyces cerevisiae Genetics 2011 189 3 737 765 22084423 6 Kupiec M Byers B Esposito RE Mitchell AP Pringle JR Broach JR Jones EW Meiosis and sporulation in Saccharomyces cerevisiae The molecular and cellular biology of the yeast Saccharomyces: Cell cycle and cell biology 1997 889 1036 Cold Spring Harbor Laboratory Press Cold Spring Harbor, NY 7 Dickinson JR Filament formation in Saccharomyces cerevisiae--a review Folia Microbiol (Praha) 2008 53 1 3 14 18481212 8 Gagiano M Bauer FF Pretorius IS The sensing of nutritional status and the relationship to filamentous growth in Saccharomyces cerevisiae FEMS Yeast Res 2002 2 4 433 470 12702263 9 Gancedo JM Control of pseudohyphae formation in Saccharomyces cerevisiae FEMS Microbiol Rev 2001 25 1 107 123 11152942 10 Palkova Z Wilkinson D Vachova L Aging and differentiation in yeast populations: elders with different properties and functions FEMS Yeast Res 2014 14 1 96 108 24119061 11 De Virgilio C The essence of yeast quiescence FEMS Microbiol Rev 2012 36 2 306 339 21658086 12 Gray JV Petsko GA Johnston GC Ringe D Singer RA Werner-Washburne M “Sleeping beauty”: quiescence in Saccharomyces cerevisiae Microbiology and molecular biology reviews : MMBR 2004 68 2 187 206 15187181 13 Liti G Carter DM Moses AM Warringer J Parts L James SA Davey RP Roberts IN Burt A Koufopanou V Tsai IJ Bergman CM Bensasson D O'Kelly MJ van Oudenaarden A Barton DB Bailes E Nguyen AN Jones M Quail MA Goodhead I Sims S Smith F Blomberg A Durbin R Louis EJ Population genomics of domestic and wild yeasts Nature 2009 458 7236 337 341 19212322 14 Naumov GI Naumova ES Sniegowski PD Saccharomyces paradoxus and Saccharomyces cerevisiae are associated with exudates of North American oaks Can J Microbiol 1998 44 11 1045 1050 doi 10029999 15 Wang QM Liu WQ Liti G Wang SA Bai FY Surprisingly diverged populations of Saccharomyces cerevisiae in natural environments remote from human activity Mol Ecol 2012 21 22 5404 5417 22913817 16 Aa E Townsend JP Adams RI Nielsen KM Taylor JW Population structure and gene evolution in Saccharomyces cerevisiae FEMS Yeast Res 2006 6 5 702 715 16879422 17 Fay JC Benavides JA Evidence for domesticated and wild populations of Saccharomyces cerevisiae PLoS Genet 2005 1 1 66 71 16103919 18 Hose J Yong CM Sardi M Wang Z Newton MA Gasch AP Dosage compensation can buffer copy-number variation in wild yeast Elife 2015 4 19 Cubillos FA Louis EJ Liti G Generation of a large set of genetically tractable haploid and diploid Saccharomyces strains FEMS Yeast Res 2009 9 8 1217 1225 19840116 20 Hope EA Dunham MJ Ploidy-regulated variation in biofilm-related phenotypes in natural isolates of Saccharomyces cerevisiae G3 (Bethesda) 2014 4 9 1773 1786 25060625 21 Albertin W Marullo P Polyploidy in fungi: evolution after whole-genome duplication Proceedings Biological sciences / The Royal Society 2012 279 1738 2497 2509 22 Sicard D Legras JL Bread, beer and wine: yeast domestication in the Saccharomyces sensu stricto complex C R Biol 2011 334 3 229 236 21377618 23 Storchova Z Ploidy changes and genome stability in yeast Yeast 2014 31 11 421 430 25155743 24 Gerke JP Chen CT Cohen BA Natural isolates of Saccharomyces cerevisiae display complex genetic variation in sporulation efficiency Genetics 2006 174 2 985 997 16951083 25 Piccirillo S Honigberg SM Sporulation patterning and invasive growth in wild and domesticated yeast colonies Res Microbiol 2010 161 5 390 398 20420901 26 Sniegowski PD Dombrowski PG Fingerman E Saccharomyces cerevisiae and Saccharomyces paradoxus coexist in a natural woodland site in North America and display different levels of reproductive isolation from European conspecifics FEMS Yeast Res 2002 1 4 299 306 12702333 27 Zorgo E Chwialkowska K Gjuvsland AB Garre E Sunnerhagen P Liti G Blomberg A Omholt SW Warringer J Ancient evolutionary trade-offs between yeast ploidy states PLoS Genet 2013 9 3 e1003388 23555297 28 Sherwood RK Scaduto CM Torres SE Bennett RJ Convergent evolution of a fused sexual cycle promotes the haploid lifestyle Nature 2014 506 7488 387 390 24390351 29 Merlini L Dudin O Martin SG Mate and fuse: how yeast cells do it Open Biol 2013 3 3 130008 23466674 30 Ruderfer DM Pratt SC Seidel HS Kruglyak L Population genomic analysis of outcrossing and recombination in yeast Nat Genet 2006 38 9 1077 1081 16892060 31 Diezmann S Dietrich FS Saccharomyces cerevisiae: population divergence and resistance to oxidative stress in clinical, domesticated and wild isolates PLoS One 2009 4 4 e5317 19390633 32 Magwene PM Kayikci O Granek JA Reininga JM Scholl Z Murray D Outcrossing, mitotic recombination, and life-history tradeoffs shape genome evolution in Saccharomyces cerevisiae Proceedings of the National Academy of Sciences of the United States of America 2011 108 5 1987 1992 21245305 33 Stefanini I Dapporto L Berna L Polsinelli M Turillazzi S Cavalieri D Social wasps are a Saccharomyces mating nest Proc Natl Acad Sci U S A 2016 113 8 2247 2251 26787874 34 Knop M Evolution of the hemiascomycete yeasts: on life styles and the importance of inbreeding Bioessays 2006 28 7 696 708 16929561 35 Mitchell AP Herskowitz I Activation of meiosis and sporulation by repression of the RME1 product in yeast Nature 1986 319 738 742 3513021 36 Wagstaff JE Klapholz S Esposito RE Meiosis in haploid yeast Proc Natl Acad Sci U S A 1982 79 9 2986 2990 7045878 37 van Dijken JP Weusthuis RA Pronk JT Kinetics of growth and sugar consumption in yeasts Antonie Van Leeuwenhoek 1993 63 3–4 343 352 8279829 38 Kim JH Roy A Jouandot D 2nd Cho KH The glucose signaling network in yeast Biochim Biophys Acta 2013 1830 11 5204 5210 23911748 39 Dashko S Zhou N Compagno C Piskur J Why, when, and how did yeast evolve alcoholic fermentation? FEMS Yeast Res 2014 14 6 826 832 24824836 40 Goddard MR Greig D Saccharomyces cerevisiae: a nomadic yeast with no niche? FEMS Yeast Res 2015 15 3 41 Paczia N Nilgen A Lehmann T Gatgens J Wiechert W Noack S Extensive exometabolome analysis reveals extended overflow metabolism in various microorganisms Microb Cell Fact 2012 11 122 22963408 42 Zampar GG Kummel A Ewald J Jol S Niebel B Picotti P Aebersold R Sauer U Zamboni N Heinemann M Temporal system-level organization of the switch from glycolytic to gluconeogenic operation in yeast Mol Syst Biol 2013 9 651 23549479 43 Schuller HJ Transcriptional control of nonfermentative metabolism in the yeast Saccharomyces cerevisiae Curr Genet 2003 43 3 139 160 12715202 44 Kumar R Dhali S Srikanth R Ghosh SK Srivastava S Comparative proteomics of mitosis and meiosis in Saccharomyces cerevisiae Journal of proteomics 2014 109 1 15 24972321 45 Grassl J Scaife C Polden J Daly CN Iacovella MG Dunn MJ Clyne RK Analysis of the budding yeast pH 4–7 proteome in meiosis Proteomics 2010 10 3 506 519 20029842 46 Walther T Letisse F Peyriga L Alkim C Liu Y Lardenois A Martin-Yken H Portais JC Primig M Francois J Developmental stage dependent metabolic regulation during meiotic differentiation in budding yeast BMC biology 2014 12 60 25178389 47 Ray D Ye P Characterization of the metabolic requirements in yeast meiosis PLoS One 2013 8 5 e63707 23675502 48 Mortimer RK Johnston JR Genealogy of principal strains of the yeast genetic stock center Genetics 1986 113 1 35 43 3519363 49 Deutschbauer AM Davis RW Quantitative trait loci mapped to single-nucleotide resolution in yeast Nat Genet 2005 37 12 1333 1340 16273108 50 Ben-Ari G Zenvirth D Sherman A David L Klutstein M Lavi U Hillel J Simchen G Four linked genes participate in controlling sporulation efficiency in budding yeast PLoS Genet 2006 2 11 e195 17112318 51 Casalone E Barberio C Cappellini L Polsinelli M Characterization of Saccharomyces cerevisiae natural populations for pseudohyphal growth and colony morphology Res Microbiol 2005 156 2 191 200 15748984 52 Song Q Johnson C Wilson TE Kumar A Pooled segregant sequencing reveals genetic determinants of yeast pseudohyphal growth PLoS Genet 2014 10 8 e1004570 25144783 53 Liu H Styles CA Fink GR Saccharomyces cerevisiae S288C has a mutation in FLO8, a gene required for filamentous growth Genetics 1996 144 3 967 978 8913742 54 Jubany S Tomasco I Ponce de Leon I Medina K Carrau F Arrambide N Naya H Gaggero C Toward a global database for the molecular typing of Saccharomyces cerevisiae strains FEMS Yeast Res 2008 8 3 472 484 18294198 55 Van Mulders SE Christianen E Saerens SM Daenen L Verbelen PJ Willaert R Verstrepen KJ Delvaux FR Phenotypic diversity of Flo protein family-mediated adhesion in Saccharomyces cerevisiae FEMS Yeast Res 2009 9 2 178 190 19087208 56 White MG Piccirillo S Dusevich V Law DJ Kapros T Honigberg SM Flo11p adhesin required for meiotic differentiation in Saccharomyces cerevisiae minicolonies grown on plastic surfaces FEMS Yeast Res 2011 11 2 223 232 21205160 57 Strudwick N Brown M Parmar VM Schroder M Ime1 and Ime2 are required for pseudohyphal growth of Saccharomyces cerevisiae on nonfermentable carbon sources Mol Cell Biol 2010 30 23 5514 5530 20876298 58 Bayly JC Douglas LM Pretorius IS Bauer FF Dranginis AM Characteristics of Flo11-dependent flocculation in Saccharomyces cerevisiae FEMS Yeast Res 2005 5 12 1151 1156 16043420 59 Gimeno CJ Ljungdahl PO Styles CA Fink GR Unipolar cell divisions in the yeast S. cerevisiae lead to filamentous growth: regulation by starvation and RAS Cell 1992 68 6 1077 1090 1547504 60 Fabrizio P Gattazzo C Battistella L Wei M Cheng C McGrew K Longo VD Sir2 blocks extreme life-span extension Cell 2005 123 4 655 667 16286010 61 Engelberg D Perlman R Levitzki A Transmembrane signaling in Saccharomyces cerevisiae as a model for signaling in metazoans: state of the art after 25 years Cellular signalling 2014 26 12 2865 2878 25218923 62 Coluccio AE Rodriguez RK Kernan MJ Neiman AM The yeast spore wall enables spores to survive passage through the digestive tract of Drosophila PLoS ONE 2008 3 8 e2873 18682732 63 Granot D Snyder M Glucose induces cAMP-independent growth-related changes in stationary-phase cells of Saccharomyces cerevisiae Proc Natl Acad Sci U S A 1991 88 13 5724 5728 1648229 64 Romano P Suzzi G Sensitivity of Saccharomyces cerevisiae vegetative cells and spores to antimicrobial compounds J Appl Bacteriol 1985 59 4 299 302 3905741 65 Briza P Breitenbach M Ellinger A Segall J Isolation of two developmentally regulated genes involved in spore wall maturation in Saccharomyces cerevisiae Genes and Development 1990 4 1775 1789 2249774 66 Muthukumar G Suhng SH Magee PT Jewell RD Primerano DA The Saccharomyces cerevisiae SPR1 gene encodes a sporulation-specific exo-1,3-beta-glucanase which contributes to ascospore thermoresistance J Bacteriol 1993 175 2 386 394 8419289 67 Reuter M Bell G Greig D Increased outbreeding in yeast in response to dispersal by an insect vector Curr Biol 2007 17 3 R81 83 17276903 68 Lachance MA Gilbert DG Starmer WT Yeast communities associated with Drosophila species and related flies in an eastern oak-pine forest: a comparison with western communities J Ind Microbiol 1995 14 6 484 494 7662291 69 Rattray A Santoyo G Shafer B Strathern JN Elevated mutation rate during meiosis in Saccharomyces cerevisiae PLoS Genet 2015 11 1 e1004910 25569256 70 Primig M Williams RM Winzeler EA Tevzadze GG Conway AR Hwang SY Davis RW Esposito RE The core meiotic transcriptome in budding yeasts Nat Genet 2000 26 4 415 423 11101837 71 Chu S DeRisi J Eisen M Mulholland J Botstein D Brown PO Herskowitz I The transcriptional program of sporulation in budding yeast Science 1998 282 5389 699 705 9784122 72 Becker E Liu Y Lardenois A Walther T Horecka J Stuparevic I Law MJ Lavigne R Evrard B Demougin P Riffle M Strich R Davis RW Pineau C Primig M Integrated RNA- and protein profiling of fermentation and respiration in diploid budding yeast provides insight into nutrient control of cell growth and development Journal of proteomics 2015 119 30 44 25662576 73 Piccirillo S Kapros T Honigberg SM Phenotypic plasticity within yeast colonies: differential partitioning of cell fates Curr Genet 2016 62 2 467 473 26743103 74 Piccirillo S Morales R White MG Smith K Kapros T Honigberg SM Cell Differentiation and Spatial Organization in Yeast Colonies: Role of Cell-Wall Integrity Pathway Genetics 2015 201 4 1427 1438 26510787 75 Esposito RE Klapholz S Meiosis and ascospore development. In: Strathern JN, Jones EW, Broach JR, editors. Molecular biology of the yeast Saccharomyces: Life cycle and inheritance 1981 211 287 Cold Spring Harbor Laboratory Cold Spring Harbor, NY 76 De Silva-Udawatta MN Cannon JF Roles of trehalose phosphate synthase in yeast glycogen metabolism and sporulation Molecular microbiology 2001 40 6 1345 1356 11442833 77 Ramaswamy NT Li L Khalil M Cannon JF Regulation of yeast glycogen metabolism and sporulation by Glc7p protein phosphatase Genetics 1998 149 1 57 72 9584086 78 McKnight JN Boerma JW Breeden LL Tsukiyama T Global Promoter Targeting of a Conserved Lysine Deacetylase for Transcriptional Shutoff during Quiescence Entry Mol Cell 2015 59 5 732 743 26300265 79 Miles S Li L Davison J Breeden LL Xbp1 directs global repression of budding yeast transcription during the transition to quiescence and is important for the longevity and reversibility of the quiescent state PLoS Genet 2013 9 10 e1003854 24204289 80 Rutledge MT Russo M Belton JM Dekker J Broach JR The yeast genome undergoes significant topological reorganization in quiescence Nucleic Acids Res 2015 43 17 8299 8313 26202961 81 Laporte D Courtout F Salin B Ceschin J Sagot I An array of nuclear microtubules reorganizes the budding yeast nucleus during quiescence J Cell Biol 2013 203 4 585 594 24247429 82 Munder MC Midtvedt D Franzmann T Nuske E Otto O Herbig M Ulbricht E Muller P Taubenberger A Maharana S Malinovska L Richter D Guck J Zaburdaev V Alberti S A pH-driven transition of the cytoplasm from a fluid- to a solid-like state promotes entry into dormancy Elife 2016 5 83 Joyner RP Tang JH Helenius J Dultz E Brune C Holt LJ Huet S Muller DJ Weis K A glucose-starvation response regulates the diffusion of macromolecules Elife 2016 5 84 Li L Miles S Breeden LL A Genetic Screen for Saccharomyces cerevisiae Mutants That Fail to Enter Quiescence G3 (Bethesda) 2015 5 8 1783 1795 26068574 85 Torres J Di Como CJ Herrero E De La Torre-Ruiz MA Regulation of the cell integrity pathway by rapamycin-sensitive TOR function in budding yeast J Biol Chem 2002 277 45 43495 43504 12171921 86 Cameroni E Hulo N Roosen J Winderickx J De Virgilio C The novel yeast PAS kinase Rim 15 orchestrates G0-associated antioxidant defense mechanisms Cell Cycle 2004 3 4 462 468 15300954 87 Zhang N Wu J Oliver SG Gis1 is required for transcriptional reprogramming of carbon metabolism and the stress response during transition into stationary phase in yeast Microbiology 2009 155 Pt 5 1690 1698 19383711 88 Roosen J Engelen K Marchal K Mathys J Griffioen G Cameroni E Thevelein JM De Virgilio C De Moor B Winderickx J PKA and Sch9 control a molecular switch important for the proper adaptation to nutrient availability Mol Microbiol 2005 55 3 862 880 15661010 89 Powers RW 3rd Kaeberlein M Caldwell SD Kennedy BK Fields S Extension of chronological life span in yeast by decreased TOR pathway signaling Genes & development 2006 20 2 174 184 16418483 90 Fabrizio P Pozza F Pletcher SD Gendron CM Longo VD Regulation of longevity and stress resistance by Sch9 in yeast Science 2001 292 5515 288 290 11292860 91 Carmona-Gutierrez D Buttner S The many ways to age for a single yeast cell Yeast 2014 31 8 289 298 24842537 92 Ruckenstuhl C Carmona-Gutierrez D Madeo F The sweet taste of death: glucose triggers apoptosis during yeast chronological aging Aging (Albany NY) 2010 2 10 643 649 21076182 93 Fabrizio P Longo VD Chronological aging-induced apoptosis in yeast Biochim Biophys Acta 2008 1783 7 1280 1285 18445486 94 Wloch-Salamon DM Bem AE Types of cell death and methods of their detection in yeast Saccharomyces cerevisiae J Appl Microbiol 2012 95 Strich R Programmed Cell Death Initiation and Execution in Budding Yeast Genetics 2015 200 4 1003 1014 26272996 96 Greenwood MT Ludovico P Expressing and functional analysis of mammalian apoptotic regulators in yeast Cell Death Differ 2010 17 5 737 745 19911009 97 Vachova L Palkova Z Caspases in yeast apoptosis-like death: facts and artefacts FEMS Yeast Res 2007 7 1 12 21 17311581 98 Sinclair D Mills K Guarente L Aging in Saccharomyces cerevisiae Annu Rev Microbiol 1998 52 533 560 9891807 99 Longo VD Shadel GS Kaeberlein M Kennedy B Replicative and chronological aging in Saccharomyces cerevisiae Cell Metab 2012 16 1 18 31 22768836 100 Kaeberlein M McVey M Guarente L Using yeast to discover the fountain of youth Sci Aging Knowledge Environ 2001 2001 1 pe1 14602950 101 Ashrafi K Sinclair D Gordon JI Guarente L Passage through stationary phase advances replicative aging in Saccharomyces cerevisiae Proc Natl Acad Sci U S A 1999 96 16 9100 9105 10430902 102 Wierman MB Smith JS Yeast sirtuins and the regulation of aging FEMS Yeast Res 2014 14 1 73 88 24164855 103 Gao Q Ren Q Liou LC Bao X Zhang Z Mitochondrial DNA protects against salt stress-induced cytochrome c-mediated apoptosis in yeast FEBS letters 2011 585 15 2507 2512 21740907 104 Sapienza K Bannister W Balzan R Mitochondrial involvement in aspirin-induced apoptosis in yeast Microbiology 2008 154 Pt 9 2740 2747 18757807 105 Ludovico P Sousa MJ Silva MT Leao C Corte-Real M Saccharomyces cerevisiae commits to a programmed cell death process in response to acetic acid Microbiology 2001 147 Pt 9 2409 2415 11535781 106 Madeo F Frohlich E Ligr M Grey M Sigrist SJ Wolf DH Frohlich KU Oxygen stress: a regulator of apoptosis in yeast J Cell Biol 1999 145 4 757 767 10330404 107 Longo VD Fabrizio P Chronological Aging in Saccharomyces cerevisiae Subcell Biochem 2012 57 101 121 22094419 108 Bishop NA Guarente L Genetic links between diet and lifespan: shared mechanisms from yeast to humans Nat Rev Genet 2007 8 11 835 844 17909538 109 Lin SJ Defossez PA Guarente L Requirement of NAD and SIR2 for life-span extension by calorie restriction in Saccharomyces cerevisiae Science 2000 289 5487 2126 2128 11000115 110 Ayer A Gourlay CW Dawes IW Cellular redox homeostasis, reactive oxygen species and replicative ageing in Saccharomyces cerevisiae FEMS Yeast Res 2014 14 1 60 72 24164795 111 Finkel T Holbrook NJ Oxidants, oxidative stress and the biology of ageing Nature 2000 408 6809 239 247 11089981 112 Sohal RS Weindruch R Oxidative stress, caloric restriction, and aging Science 1996 273 5271 59 63 8658196 113 Ring J Sommer C Carmona-Gutierrez D Ruckenstuhl C, Eisenberg T, Madeo F (2012). The metabolism beyond programmed cell death in yeast Exp Cell Res 318 11 1193 1200 22480867 114 Ocampo A Liu J Schroeder EA Shadel GS Barrientos A Mitochondrial respiratory thresholds regulate yeast chronological life span and its extension by caloric restriction Cell Metab 2012 16 1 55 67 22768839 115 Weinberger M Mesquita A Caroll T Marks L Yang H Zhang Z Ludovico P Burhans WC Growth signaling promotes chronological aging in budding yeast by inducing superoxide anions that inhibit quiescence Aging (Albany NY) 2010 2 10 709 726 21076178 116 Fabrizio P Battistella L Vardavas R Gattazzo C Liou LL Diaspro A Dossen JW Gralla EB Longo VD Sxuperoxide is a mediator of an altruistic aging program in Saccharomyces cerevisiae The Journal of cell biology 2004 166 7 1055 1067 15452146 117 Severin FF Meer MV Smirnova EA Knorre DA Skulachev VP Natural causes of programmed death of yeast Saccharomyces cerevisiae Biochimica et biophysica acta 2008 1783 7 1350 1353 18319066 118 Allocati N Masulli M Di Ilio C De Laurenzi V Die for the community: an overview of programmed cell death in bacteria Cell Death Dis 2015 6 e1609 25611384 119 Lewis K Programmed death in bacteria Microbiology and molecular biology reviews : MMBR 2000 64 3 503 514 10974124 120 Ameisen JC The origin of programmed cell death Science 1996 272 5266 1278 1279 8650538 121 Kron SJ Styles CA Fink GR Symmetric cell division in pseudohyphae of the yeast Saccharomyces cerevisiae Mol Biol Cell 1994 5 9 1003 1022 7841518 122 Dickinson JR Irreversible formation of pseudohyphae by haploid Saccharomyces cerevisiae FEMS Microbiol Lett 1994 119 1–2 99 103 8039677 123 Kuriyama H Slaughter JC Control of cell morphology of the yeast Saccharomyces cerevisiae by nutrient limitation in continuous culture Lett Appl Microbiol 1995 20 1 37 40 doi 7765865 124 Pitoniak A Birkaya B Dionne HM Vadaie N Cullen PJ The signaling mucins Msb2 and Hkr1 differentially regulate the filamentation mitogen-activated protein kinase pathway and contribute to a multimodal response Mol Biol Cell 2009 20 13 3101 3114 19439450 125 Martin R Wachtler B Schaller M Wilson D Hube B Host-pathogen interactions and virulence-associated genes during Candida albicans oral infections Int J Med Microbiol 2011 301 5 417 422 21555244 126 Laxman S Tu BP Multiple TORC1-associated proteins regulate nitrogen starvation-dependent cellular differentiation in Saccharomyces cerevisiae PLoS One 2011 6 10 e26081 22043304 127 Ryan O Shapiro RS Kurat CF Mayhew D Baryshnikova A Chin B Lin ZY Cox MJ Vizeacoumar F Cheung D Bahr S Tsui K Tebbji F Sellam A Istel F Schwarzmuller T Reynolds TB Kuchler K Gifford DK Whiteway M Giaever G Nislow C Costanzo M Gingras AC Mitra RD Andrews B Fink GR Cowen LE Boone C Global gene deletion analysis exploring yeast filamentous growth Science 2012 337 6100 1353 1356 22984072 128 Loeb JD Kerentseva TA Pan T Sepulveda-Becerra M Liu H Saccharomyces cerevisiae G1 cyclins are differentially involved in invasive and pseudohyphal growth independent of the filamentation mitogen-activated protein kinase pathway Genetics 1999 153 4 1535 1546 10581264 129 Lo WS Dranginis AM The cell surface flocculin Flo11 is required for pseudohyphae formation and invasion by Saccharomyces cerevisiae Mol Biol Cell 1998 9 1 161 171 9436998 130 Voordeckers K De Maeyer D van der Zande E Vinces MD Meert W Cloots L Ryan O Marchal K Verstrepen KJ Identification of a complex genetic network underlying Saccharomyces cerevisiae colony morphology Mol Microbiol 2012 86 1 225 239 22882838 131 Granek JA Magwene PM Environmental and genetic determinants of colony morphology in yeast PLoS Genet 2010 6 1 e1000823 20107600 132 Reynolds TB Fink GR Bakers' yeast, a model for fungal biofilm formation Science 2001 291 5505 878 881 11157168 133 Zara G Zara S Pinna C Marceddu S Budroni M FLO11 gene length and transcriptional level affect biofilm-forming ability of wild flor strains of Saccharomyces cerevisiae Microbiology 2009 155 Pt 12 3838 3846 19729408 134 Douglas LM Li L Yang Y Dranginis AM Expression and characterization of the flocculin Flo11/Muc1, a Saccharomyces cerevisiae mannoprotein with homotypic properties of adhesion Eukaryot Cell 2007 6 12 2214 2221 17921350 135 Cullen PJ Sprague GF Jr. The regulation of filamentous growth in yeast Genetics 2012 190 1 23 49 22219507 136 Irniger S The Ime2 protein kinase family in fungi: more duties than just meiosis Molecular microbiology 2011 80 1 1 13 21306447 137 van Werven FJ Neuert G Hendrick N Lardenois A Buratowski S van Oudenaarden A Primig M Amon A Transcription of two long noncoding RNAs mediates mating-type control of gametogenesis in budding yeast Cell 2012 150 6 1170 1181 22959267 138 van Dyk D Hansson G Pretorius IS Bauer FF Cellular differentiation in response to nutrient availability: The repressor of meiosis, Rme1p, positively regulates invasive growth in Saccharomyces cerevisiae Genetics 2003 165 3 1045 1058 14668363 139 Rupp S Summers E Lo HJ Madhani H Fink G MAP kinase and cAMP filamentation signaling pathways converge on the unusually large promoter of the yeast FLO11 gene EMBO J 1999 18 5 1257 1269 10064592 140 Shah JC Clancy MJ IME4, a gene that mediates MAT and nutritional control of meiosis in Saccharomyces cerevisiae Mol Cell Biol 1992 12 3 1078 1086 1545790 141 Hongay CF Grisafi PL Galitski T Fink GR Antisense transcription controls cell fate in Saccharomyces cerevisiae Cell 2006 127 4 735 745 17110333 142 Gelfand B Mead J Bruning A Apostolopoulos N Tadigotla V Nagaraj V Sengupta AM Vershon AK Regulated antisense transcription controls expression of cell-type-specific genes in yeast Mol Cell Biol 2011 31 8 1701 1709 21300780 143 Agarwala SD Blitzblau HG Hochwagen A Fink GR RNA methylation by the MIS complex regulates a cell fate decision in yeast PLoS genetics 2012 8 6 e1002732 22685417 144 Schwartz S Agarwala SD Mumbach MR Jovanovic M Mertins P Shishkin A Tabach Y Mikkelsen TS Satija R Ruvkun G Carr SA Lander ES Fink GR Regev A High-resolution mapping reveals a conserved, widespread, dynamic mRNA methylation program in yeast meiosis Cell 2013 155 6 1409 1421 24269006 145 Yue Y Liu J He C RNA N6-methyladenosine methylation in post-transcriptional gene expression regulation Genes Dev 2015 29 13 1343 1355 26159994 146 Uren AG Beilharz T O'Connell MJ Bugg SJ van Driel R Vaux DL Lithgow T Role for yeast inhibitor of apoptosis (IAP)-like proteins in cell division Proc Natl Acad Sci U S A 1999 96 18 10170 10175 10468581 147 Hurtado S Kim Guisbert KS Sontheimer EJ SPO24 is a transcriptionally dynamic, small ORF-encoding locus required for efficient sporulation in Saccharomyces cerevisiae PLoS One 2014 9 8 e105058 25127041 148 Mai B Breeden LL Identification of target genes of a yeast transcriptional repressor Methods Mol Biol 2006 317 267 277 16264235 149 Mai B Breeden L CLN1 and its repression by Xbp1 are important for efficient sporulation in budding yeast Molecular and cellular biology 2000 20 2 478 487 10611226 150 Purnapatre K Piccirillo S Schneider BL Honigberg SM The CLN3/SWI6/CLN2 pathway and SNF1 act sequentially to regulate meiotic initiation in Saccharomyces cerevisiae Genes Cells 2002 7 7 675 691 12081645 151 Colomina N Gari E Gallego C Herrero E Aldea M G1 cyclins block the Ime1 pathway to make mitosis and meiosis incompatible in budding yeast The EMBO journal 1999 18 2 320 329 9889189 152 Aragon AD Rodriguez AL Meirelles O Roy S Davidson GS Tapia PH Allen C Joe R Benn D Werner-Washburne M Characterization of differentiated quiescent and nonquiescent cells in yeast stationary-phase cultures Mol Biol Cell 2008 19 3 1271 1280 18199684 153 Shi L Sutter BM Ye X Tu BP Trehalose is a key determinant of the quiescent metabolic state that fuels cell cycle progression upon return to growth Mol Biol Cell 2010 21 12 1982 1990 20427572 154 Tapia H Morano KA Hsp90 nuclear accumulation in quiescence is linked to chaperone function and spore development in yeast Mol Biol Cell 2010 21 1 63 72 19889838 155 Kurtz S Rossi J Petko L Lindquist S An ancient developmental induction: heat-shock proteins induced in sporulation and oogenesis Science 1986 231 4742 1154 1157 3511530 156 Honigberg SM Purnapatre K Signal pathway integration in the switch from the mitotic cell cycle to meiosis in yeast J Cell Sci 2003 116 Pt 11 2137 2147 12730290 157 van Werven FJ Amon A Regulation of entry into gametogenesis Philos Trans R Soc Lond B Biol Sci 2011 366 1584 3521 3531 22084379 158 Werner-Washburne M Roy S Davidson GS Aging and the survival of quiescent and non-quiescent cells in yeast stationary-phase cultures Subcell Biochem 2012 57 123 143 22094420 159 Ceccato-Antonini SR Sudbery PE Filamentous growth in Saccharomyces cerevisiae Braz J Microbiol 2004 35 173 181 160 Jambhekar A Amon A Control of meiosis by respiration Curr Biol 2008 18 13 969 975 18595705 161 Lee RH Honigberg SM Nutritional regulation of late meiotic events in Saccharomyces cerevisiae through a pathway distinct from initiation Mol Cell Biol 1996 16 6 3222 3232 8649433 162 Conway MK Grunwald D Heideman W Glucose, nitrogen, and phosphate repletion in Saccharomyces cerevisiae: common transcriptional responses to different nutrient signals G3 (Bethesda) 2012 2 9 1003 1017 22973537 163 Galdieri L Mehrotra S Yu S Vancura A Transcriptional regulation in yeast during diauxic shift and stationary phase Omics 2010 14 6 629 638 20863251 164 Klosinska MM Crutchfield CA Bradley PH Rabinowitz JD Broach JR Yeast cells can access distinct quiescent states Genes Dev 2011 25 4 336 349 21289062 165 Orlandi I Ronzulli R Casatta N Vai M Ethanol and acetate acting as carbon/energy sources negatively affect yeast chronological aging Oxid Med Cell Longev 2013 2013 802870 24062879 166 Burtner CR Murakami J Kennedy BK Kaeberlein M A molecular mechanism of chronological aging in yeast Cell Cycle 2009 8 8 1256 1270 19305133 167 Palecek SP Parikh AS Huh JH Kron SJ Depression of Saccharomyces cerevisiae invasive growth on non-glucose carbon sources requires the Snf1 kinase Molecular microbiology 2002 45 2 453 469 12123456 168 Cullen PJ Sprague GF Jr The roles of bud-site-selection proteins during haploid invasive growth in yeast Molecular biology of the cell 2002 13 9 2990 3004 12221111 169 Cullen PJ Evaluating yeast filamentous growth at the single-cell level Cold Spring Harbor protocols 2015 2015 3 272 275 25734069 170 Godard P Urrestarazu A Vissers S Kontos K Bontempi G van Helden J Andre B Effect of 21 different nitrogen sources on global gene expression in the yeast Saccharomyces cerevisiae Mol Cell Biol 2007 27 8 3065 3086 17308034 171 Day A Markwardt J Delaguila R Zhang J Purnapatre K Honigberg SM Schneider BL Cell size and Cln-Cdc28 complexes mediate entry into meiosis by modulating cell growth Cell Cycle 2004 3 11 1433 1439 15611626 172 Colomina N Liu Y Aldea M Gari E TOR regulates the subcellular localization of Ime1, a transcriptional activator of meiotic development in budding yeast Mol Cell Biol 2003 23 20 7415 7424 14517308 173 Santos J Leitao-Correia F Sousa MJ Leao C Ammonium is a key determinant on the dietary restriction of yeast chronological aging in culture medium Oncotarget 2015 6 9 6511 6523 25576917 174 Vinod PK Sengupta N Bhat PJ Venkatesh KV Integration of global signaling pathways, cAMP-PKA, MAPK and TOR in the regulation of FLO11 PLoS One 2008 3 2 e1663 18301741 175 Boeckstaens M Andre B Marini AM The yeast ammonium transport protein Mep2 and its positive regulator, the Npr1 kinase, play an important role in normal and pseudohyphal growth on various nitrogen media through retrieval of excreted ammonium Mol Microbiol 2007 64 2 534 546 17493133 176 Murray LE Rowley N Dawes IW Johnston GC Singer RA A yeast glutamine tRNA signals nitrogen status for regulation of dimorphic growth and sporulation Proc Natl Acad Sci U S A 1998 95 15 8619 8624 9671727 177 Lorenz MC Heitman J The MEP2 ammonium permease regulates pseudohyphal differentiation in Saccharomyces cerevisiae EMBO J 1998 17 5 1236 1247 9482721 178 Orij R Postmus J Ter Beek A Brul S Smits GJ In vivo measurement of cytosolic and mitochondrial pH using a pH-sensitive GFP derivative in Saccharomyces cerevisiae reveals a relation between intracellular pH and growth Microbiology 2009 155 Pt 1 268 278 19118367 179 Murakami C Delaney JR Chou A Carr D Schleit J Sutphin GL An EH Castanza AS Fletcher M Goswami S Higgins S Holmberg M Hui J Jelic M Jeong KS Kim JR Klum S Liao E Lin MS Lo W Miller H Moller R Peng ZJ Pollard T Pradeep P Pruett D Rai D Ros V Schuster A Singh M pH neutralization protects against reduction in replicative lifespan following chronological aging in yeast Cell Cycle 2012 11 16 3087 3096 22871733 180 Hayashi M Ohkuni K Yamashita I Control of division arrest and entry into meiosis by extracellular alkalization in Saccharomyces cerevisiae Yeast 1998 14 10 905 913 9717236 181 Mirisola MG Longo VD Acetic acid and acidification accelerate chronological and replicative aging in yeast Cell Cycle 2012 11 19 3532 3533 22951542 182 Piccirillo S White MG Murphy JC Law DJ Honigberg SM The Rim101p/PacC pathway and alkaline pH regulate pattern formation in yeast colonies Genetics 2010 184 3 707 716 20038633 183 Ohkuni K Hayashi M Yamashita I Bicarbonate-mediated social communication stimulates meiosis and sporulation of Saccharomyces cerevisiae Yeast 1998 14 7 623 631 9639309 184 Soares EV Flocculation in Saccharomyces cerevisiae: a review J Appl Microbiol 2011 110 1 1 18 21114594 185 Stratford M Induction of flocculation in brewing yeasts by change in pH value FEMS Microbiol Lett 1996 136 1 13 18 8919449 186 Li W Mitchell AP Proteolytic activation of Rim1p, a positive regulator of yeast sporulation and invasive growth Genetics 1997 145 1 63 73 9017390 187 Esposito MS Esposito RE The genetic control of sporulation in Saccharomyces. I. The isolation of temperature-sensitive sporulation-deficient mutants Genetics 1969 61 1 79 89 5802566 188 McCusker JH Clemons KV Stevens DA Davis RW Saccharomyces cerevisiae virulence phenotype as determined with CD-1 mice is associated with the ability to grow at 42 degrees C and form pseudohyphae Infect Immun 1994 62 12 5447 5455 7960125 189 McCusker JH Clemons KV Stevens DA Davis RW Genetic characterization of pathogenic Saccharomyces cerevisiae isolates Genetics 1994 136 4 1261 1269 8013903 190 Nussbaum I Weindling E Jubran R Cohen A Bar-Nun S Deteriorated stress response in stationary-phase yeast: Sir2 and Yap1 are essential for Hsf1 activation by heat shock and oxidative stress, respectively PLoS One 2014 9 10 e111505 25356557 191 Kourtis N Tavernarakis N Cellular stress response pathways and ageing: intricate molecular relationships EMBO J 2011 30 13 2520 2531 21587205 192 Verstrepen KJ Fink GR Genetic and epigenetic mechanisms underlying cell-surface variability in protozoa and fungi Annu Rev Genet 2009 43 1 24 19640229 193 Cullen PJ Signaling mucins: the new kids on the MAPK block Crit Rev Eukaryot Gene Expr 2007 17 3 241 257 17725491 194 Conrad M Schothorst J Kankipati HN Van Zeebroeck G Rubio-Texeira M Thevelein JM Nutrient sensing and signaling in the yeast Saccharomyces cerevisiae FEMS Microbiol Rev 2014 38 2 254 299 24483210 195 Broach JR Nutritional control of growth and development in yeast Genetics 2012 192 1 73 105 22964838 196 Hedbacker K Carlson M SNF1/AMPK pathways in yeast Front Biosci 2008 13 2408 2420 17981722 197 Orlova M Ozcetin H Barrett L Kuchin S Roles of the Snf1-activating kinases during nitrogen limitation and pseudohyphal differentiation in Saccharomyces cerevisiae Eukaryot Cell 2010 9 1 208 214 19880754 198 Hughes Hallett JE Luo X Capaldi AP State transitions in the TORC1 signaling pathway and information processing in Saccharomyces cerevisiae Genetics 2014 198 2 773 786 25085507 199 Thevelein JM Bonini BM Castermans D Haesendonckx S Kriel J Louwet W Thayumanavan P Popova Y Rubio-Texeira M Schepers W Vandormael P Van Zeebroeck G Verhaert P Versele M Voordeckers K Novel mechanisms in nutrient activation of the yeast protein kinase A pathway Acta Microbiol Immunol Hung 2008 55 2 75 89 18595314 200 Jungbluth M Mosch HU Taxis C Acetate Regulation of Spore Formation Is under the Control of the Ras/Cyclic AMP/Protein Kinase A Pathway and Carbon Dioxide in Saccharomyces cerevisiae Eukaryotic cell 2012 11 8 1021 1032 22660623 201 Dechant R Saad S Ibanez AJ Peter M Cytosolic pH regulates cell growth through distinct GTPases, Arf1 and Gtr1, to promote Ras/PKA and TORC1 activity Mol Cell 2014 55 3 409 421 25002144 202 Dechant R Binda M Lee SS Pelet S Winderickx J Peter M Cytosolic pH is a second messenger for glucose and regulates the PKA pathway through V-ATPase EMBO J 2010 29 15 2515 2526 20581803 203 Mizunuma M Tsubakiyama R Ogawa T Shitamukai A Kobayashi Y Inai T Kume K Hirata D Ras/cAMP-dependent protein kinase (PKA) regulates multiple aspects of cellular events by phosphorylating the Whi3 cell cycle regulator in budding yeast J Biol Chem 2013 288 15 10558 10566 23471970 204 Burtner CR Murakami CJ Olsen B Kennedy BK Kaeberlein M A genomic analysis of chronological longevity factors in budding yeast Cell Cycle 2011 10 9 1385 1396 21447998 205 Mosch HU Fink GR Dissection of filamentous growth by transposon mutagenesis in Saccharomyces cerevisiae Genetics 1997 145 3 671 684 9055077 206 Reinders A Burckert N Boller T Wiemken A De Virgilio C Saccharomyces cerevisiae cAMP-dependent protein kinase controls entry into stationary phase through the Rim15p protein kinase Genes & development 1998 12 18 2943 2955 9744870 207 Watanabe D Zhou Y Hirata A Sugimoto Y Takagi K Akao T Ohya Y Takagi H Shimoi H Inhibitory Role of Greatwall-Like Protein Kinase Rim15p in Alcoholic Fermentation via Upregulating the UDP-Glucose Synthesis Pathway in Saccharomyces cerevisiae Appl Environ Microbiol 2015 82 1 340 351 26497456 208 Bontron S Jaquenoud M Vaga S Talarek N Bodenmiller B Aebersold R De Virgilio C Yeast endosulfines control entry into quiescence and chronological life span by inhibiting protein phosphatase 2A Cell Rep 2013 3 1 16 22 23273919 209 Luo X Talarek N Virgilio CD Initiation of the yeast G 0 program requires Igo1 and Igo2, which antagonize activation of decapping of specific nutrient-regulated mRNAs RNA Biol 2011 8 1 210 Talarek N Cameroni E Jaquenoud M Luo X Bontron S Lippman S Devgan G Snyder M Broach JR De Virgilio C Initiation of the TORC1-regulated G0 program requires Igo1/2, which license specific mRNAs to evade degradation via the 5'–3' mRNA decay pathway Mol Cell 2010 38 3 345 355 20471941 211 Lee P Kim MS Paik SM Choi SH Cho BR Hahn JS Rim15-dependent activation of Hsf1 and Msn2/4 transcription factors by direct phosphorylation in Saccharomyces cerevisiae FEBS Lett 2013 587 22 3648 3655 24140345 212 Wei M Fabrizio P Hu J Ge H Cheng C Li L Longo VD Life span extension by calorie restriction depends on Rim15 and transcription factors downstream of Ras/PKA, Tor, and Sch9 PLoS genetics 2008 4 1 e13 18225956 213 Vidan S Mitchell AP Stimulation of yeast meiotic gene expression by the glucose-repressible protein kinase Rim15p Mol Cell Biol 1997 17 5 2688 2697 9111339 214 Quan Z Cao L Tang Y Yan Y Oliver SG Zhang N The Yeast GSK-3 Homologue Mck1 Is a Key Controller of Quiescence Entry and Chronological Lifespan PLoS Genet 2015 11 6 e1005282 26103122 215 Swinnen E Wanke V Roosen J Smets B Dubouloz F Pedruzzi I Cameroni E De Virgilio C Winderickx J Rim15 and the crossroads of nutrient signalling pathways in Saccharomyces cerevisiae Cell Div 2006 1 3 16759348 216 Sarkar S Dalgaard JZ Millar JB Arumugam P The Rim15-endosulfine-PP2ACdc55 signalling module regulates entry into gametogenesis and quiescence via distinct mechanisms in budding yeast PLoS Genet 2014 10 6 e1004456 24968058 217 Ward MP Gimeno CJ Fink GR Garrett S SOK2 may regulate cyclic AMP-dependent protein kinase-stimulated growth and pseudohyphal development by repressing transcription Mol Cell Biol 1995 15 12 6854 6863 8524252 218 Ward MP Garrett S Suppression of a yeast cyclic AMP-dependent protein kinase defect by overexpression of SOK1, a yeast gene exhibiting sequence similarity to a developmentally regulated mouse gene Mol Cell Biol 1994 14 9 5619 5627 8065298 219 Shenhar G Kassir Y A positive regulator of mitosis, Sok2, functions as a negative regulator of meiosis in Saccharomyces cerevisiae Mol Cell Biol 2001 21 5 1603 1612 11238897 220 Pan X Heitman J Sok2 regulates yeast pseudohyphal differentiation via a transcription factor cascade that regulates cell-cell adhesion Mol Cell Biol 2000 20 22 8364 8372 11046133 221 Borneman AR Leigh-Bell JA Yu H Bertone P Gerstein M Snyder M Target hub proteins serve as master regulators of development in yeast Genes Dev 2006 20 4 435 448 16449570 222 McDonald CM Wagner M Dunham MJ Shin ME Ahmed NT Winter E The Ras/cAMP pathway and the CDK-like kinase Ime2 regulate the MAPK Smk1 and spore morphogenesis in Saccharomyces cerevisiae Genetics 2009 181 2 511 523 19087957 223 Lee P Paik SM Shin CS Huh WK Hahn JS Regulation of yeast Yak1 kinase by PKA and autophosphorylation-dependent 14-3-3 binding Mol Microbiol 2011 79 3 633 646 21255108 224 Lee P Cho BR Joo HS Hahn JS Yeast Yak1 kinase, a bridge between PKA and stress-responsive transcription factors, Hsf1 and Msn2/Msn4 Mol Microbiol 2008 70 4 882 895 18793336 225 Cardona F Del Olmo ML Aranda A Phylogenetic origin and transcriptional regulation at the post-diauxic phase of SPI1, in Saccharomyces cerevisiae Cell Mol Biol Lett 2012 17 3 393 407 22610976 226 Werner-Washburne M Brown D Braun E Bcy1, the regulatory subunit of cAMP-dependent protein kinase in yeast, is differentially modified in response to the physiological status of the cell J Biol Chem 1991 266 29 19704 19709 1655793 227 Malcher M Schladebeck S Mosch HU The Yak1 protein kinase lies at the center of a regulatory cascade affecting adhesive growth and stress resistance in Saccharomyces cerevisiae Genetics 2011 187 3 717 730 21149646 228 Mirisola MG Taormina G Fabrizio P Wei M Hu J Longo VD Serine- and threonine/valine-dependent activation of PDK and Tor orthologs converge on Sch9 to promote aging PLoS Genet 2014 10 2 e1004113 24516402 229 Swinnen E Ghillebert R Wilms T Winderickx J Molecular mechanisms linking the evolutionary conserved TORC1-Sch9 nutrient signalling branch to lifespan regulation in Saccharomyces cerevisiae FEMS Yeast Res 2014 14 1 17 32 24102693 230 Weidberg H Moretto F Spedale G Amon A van Werven FJ Nutrient Control of Yeast Gametogenesis Is Mediated by TORC1, PKA and Energy Availability PLoS Genet 2016 12 6 e1006075. 27272508 231 Pan Y Schroeder EA Ocampo A Barrientos A Shadel GS Regulation of yeast chronological life span by TORC1 via adaptive mitochondrial ROS signaling Cell Metab 2011 13 6 668 678 21641548 232 Pan Y Shadel GS Extension of chronological life span by reduced TOR signaling requires down-regulation of Sch9p and involves increased mitochondrial OXPHOS complex density Aging (Albany NY) 2009 1 1 131 145 20157595 233 Cutler NS Pan X Heitman J Cardenas ME The TOR signal transduction cascade controls cellular differentiation in response to nutrients Molecular biology of the cell 2001 12 12 4103 4113 11739804 234 Bruckner S Kern S Birke R Saugar I Ulrich HD Mosch HU The TEA transcription factor Tec1 links TOR and MAPK pathways to coordinate yeast development Genetics 2011 189 2 479 494 21840851 235 Humston EM Dombek KM Tu BP Young ET Synovec RE Toward a global analysis of metabolites in regulatory mutants of yeast Anal Bioanal Chem 2011 401 8 2387 2402 21416166 236 Usaite R Jewett MC Oliveira AP Yates JR 3rd Olsson L Nielsen J Reconstruction of the yeast Snf1 kinase regulatory network reveals its role as a global energy regulator Mol Syst Biol 2009 5 319 19888214 237 Wang Z Wilson WA Fujino MA Roach PJ Antagonistic controls of autophagy and glycogen accumulation by Snf1p, the yeast homolog of AMP-activated protein kinase, and the cyclin-dependent kinase Pho85p Mol Cell Biol 2001 21 17 5742 5752 11486014 238 Enjalbert B Parrou JL Teste MA Francois J Combinatorial control by the protein kinases PKA, PHO85 and SNF1 of transcriptional induction of the Saccharomyces cerevisiae GSY2 gene at the diauxic shift Mol Genet Genomics 2004 271 6 697 708 15221454 239 Van de Velde S Thevelein JM Cyclic AMP-protein kinase A and Snf1 signaling mechanisms underlie the superior potency of sucrose for induction of filamentation in Saccharomyces cerevisiae Eukaryot Cell 2008 7 2 286 293 17890371 240 Kuchin S Vyas VK Carlson M Snf1 protein kinase and the repressors Nrg1 and Nrg2 regulate FLO11, haploid invasive growth, and diploid pseudohyphal differentiation Mol Cell Biol 2002 22 12 3994 4000 12024013 241 Lorenz DR Cantor CR Collins JJ A network biology approach to aging in yeast Proc Natl Acad Sci U S A 2009 106 4 1145 1150 19164565 242 Ashrafi K Lin SS Manchester JK Gordon JI Sip2p and its partner snf1p kinase affect aging in S. cerevisiae Genes & development 2000 14 15 1872 1885 10921902 243 Maeda T The signaling mechanism of ambient pH sensing and adaptation in yeast and fungi Febs J 2012 279 8 1407 1413 22360598 244 Scherz K Andersen Bojsen R Gro L Rejkjaer Sorensen Weiss M Nielsen Lisby M Folkesson A Regenberg B Genetic basis for Saccharomyces cerevisiae biofilm in liquid medium G3 (Bethesda) 2014 4 9 1671 1680 25009170 245 Chavel CA Dionne HM Birkaya B Joshi J Cullen PJ Multiple signals converge on a differentiation MAPK pathway PLoS Genet 2010 6 3 e1000883 20333241 246 Lamb TM Mitchell AP The transcription factor Rim101p governs ion tolerance and cell differentiation by direct repression of the regulatory genes NRG1 and SMP1 in Saccharomyces cerevisiae Mol Cell Biol 2003 23 2 677 686 12509465 247 Read T Richmond PA Dowell RD A trans-acting Variant within the Transcription Factor RIM101 Interacts with Genetic Background to Determine its Regulatory Capacity PLoS Genet 2016 12 1 e1005746 26751950 248 Gray M Piccirillo S Purnapatre K Schneider BL Honigberg SM Glucose induction pathway regulates meiosis in Saccharomyces cerevisiae in part by controlling turnover of Ime2p meiotic kinase FEMS Yeast Res 2008 8 5 676 684 18616605 249 Purnapatre K Gray M Piccirillo S Honigberg SM Glucose inhibits meiotic DNA replication through SCFGrr1p-dependent destruction of Ime2p kinase Molecular and cellular biology 2005 25 1 440 450 15601864 250 Shively CA Eckwahl MJ Dobry CJ Mellacheruvu D Nesvizhskii A Kumar A Genetic networks inducing invasive growth in Saccharomyces cerevisiae identified through systematic genome-wide overexpression Genetics 2013 193 4 1297 1310 23410832 251 Furukawa K Furukawa T Hohmann S Efficient construction of homozygous diploid strains identifies genes required for the hyper-filamentous phenotype in Saccharomyces cerevisiae PLoS One 2011 6 10 e26584 22039512 252 Krause SA Gray JV The protein kinase C pathway is required for viability in quiescence in Saccharomyces cerevisiae Current biology : CB 2002 12 7 588 593 11937029 253 Sundaram V Petkova MI Pujol-Carrion N Boada J de la Torre-Ruiz MA Tor1, Sch9 and PKA downregulation in quiescence rely on Mtl1 to preserve mitochondrial integrity and cell survival Mol Microbiol 2015 97 1 93 109 25855033 254 Vancetto GT Ceccato-Antonini SR MPK1 gene is required for filamentous growth induced by isoamyl alcohol in Saccharomyces cerevisiae strains from the alcoholic fermentation Appl Microbiol Biotechnol 2007 75 1 111 115 17245577 255 Martinez de Maranon I Marechal PA Gervais P Passive response of Saccharomyces cerevisiae to osmotic shifts: cell volume variations depending on the physiological state Biochem Biophys Res Commun 1996 227 2 519 523 8878546 256 de Llanos R Hernandez-Haro C Barrio E Querol A Fernandez-Espinar MT Molina M Differences in activation of MAP kinases and variability in the polyglutamine tract of Slt2 in clinical and non-clinical isolates of Saccharomyces cerevisiae Yeast 2010 27 8 549 561 20586115 257 Schroder M Clark R Liu CY Kaufman RJ The unfolded protein response represses differentiation through the RPD3-SIN3 histone deacetylase EMBO J 2004 23 11 2281 2292 15141165 258 Orlova M Kanter E Krakovich D Kuchin S Nitrogen availability and TOR regulate the Snf1 protein kinase in Saccharomyces cerevisiae Eukaryot Cell 2006 5 11 1831 1837 16980405 259 Nishizawa M Tanigawa M Hayashi M Maeda T Yazaki Y Saeki Y Toh-e A Pho85 kinase, a cyclin-dependent kinase, regulates nuclear accumulation of the Rim101 transcription factor in the stress response of Saccharomyces cerevisiae Eukaryot Cell 2010 9 6 943 951 20382759 260 Law MJ Ciccaglione K Fine-tuning of histone H3 Lys4 methylation during pseudohyphal differentiation by the CDK submodule of RNA polymerase II Genetics 2015 199 2 435 453 25467068 261 Halme A Bumgarner S Styles C Fink GR Genetic and epigenetic regulation of the FLO gene family generates cell-surface variation in yeast Cell 2004 116 3 405 415 15016375 262 Wang LC Montalvo-Munoz F Tsai YC Liang CY Chang CC Lo WS The Histone Acetyltransferase Gcn5 Regulates ncRNA-ICR1 and FLO11 Expression during Pseudohyphal Development in Saccharomyces cerevisiae Biomed Res Int 2015 2015 284692 25922832 263 Bumgarner SL Neuert G Voight BF Symbor-Nagrabska A Grisafi P van Oudenaarden A Fink GR Single-cell analysis reveals that noncoding RNAs contribute to clonal heterogeneity by modulating transcription factor recruitment Mol Cell 2012 45 4 470 482 22264825 264 Bumgarner SL Dowell RD Grisafi P Gifford DK Fink GR Toggle involving cis-interfering noncoding RNAs controls variegated gene expression in yeast Proc Natl Acad Sci U S A 2009 106 43 18321 18326 19805129 265 Holmes DL Lancaster AK Lindquist S Halfmann R Heritable remodeling of yeast multicellularity by an environmentally responsive prion Cell 2013 153 1 153 165 23540696 266 Lardenois A Stuparevic I Liu Y Law MJ Becker E Smagulova F Waern K Guilleux MH Horecka J Chu A Kervarrec C Strich R Snyder M Davis RW Steinmetz LM Primig M The conserved histone deacetylase Rpd3 and its DNA binding subunit Ume6 control dynamic transcript architecture during mitotic growth and meiotic development Nucleic Acids Res 2015 43 1 115 128 25477386 267 Krishnamoorthy T Chen X Govin J Cheung WL Dorsey J Schindler K Winter E Allis CD Guacci V Khochbin S Fuller MT Berger SL Phosphorylation of histone H4 Ser1 regulates sporulation in yeast and is conserved in fly and mouse spermatogenesis Genes Dev 2006 20 18 2580 2592 16980586 268 Xu M Soloveychik M Ranger M Schertzberg M Shah Z Raisner R Venkatasubrahmanyan S Tsui K Gebbia M Hughes T van Bakel H Nislow C Madhani HD Meneghini MD Timing of transcriptional quiescence during gametogenesis is controlled by global histone H3K4 demethylation Dev Cell 2012 23 5 1059 1071 23123093 269 McKnight JN Tsukiyama T The conserved HDAC Rpd3 drives transcriptional quiescence in S. cerevisiae Genom Data 2015 6 245 248 26697386 270 Radonjic M Andrau JC Lijnzaad P Kemmeren P Kockelkorn TT van Leenen D van Berkum NL Holstege FC Genome-wide analyses reveal RNA polymerase II located upstream of genes poised for rapid response upon S. cerevisiae stationary phase exit Mol Cell 2005 18 2 171 183 15837421 271 Ganesan AT Holter H Roberts C Some observations on sporulation in Saccharomyces CR Lab Carlsberg 1958 13 1 6 272 Kirsop BH Studies in yeast sporulation. I. Some factors influencing sporulation J Inst Brew 1954 60 393 273 Tsuchiya D Yang Y Lacefield S Positive feedback of NDT80 expression ensures irreversible meiotic commitment in budding yeast PLoS Genet 2014 10 6 e1004398 24901499 274 Winter E The Sum1/Ndt80 transcriptional switch and commitment to meiosis in Saccharomyces cerevisiae Microbiology and molecular biology reviews : MMBR 2012 76 1 1 15 22390969 275 Corbi D Sunder S Weinreich M Skokotas A Johnson ES Winter E Multisite phosphorylation of the Sum1 transcriptional repressor by S-phase kinases controls exit from meiotic prophase in yeast Mol Cell Biol 2014 34 12 2249 2263 24710277 276 Berchowitz LE Kabachinski G Walker MR Carlile TM Gilbert WV Schwartz TU Amon A Regulated Formation of an Amyloid-like Translational Repressor Governs Gametogenesis Cell 2015 163 2 406 418 26411291 277 Jin L Zhang K Xu Y Sternglanz R Neiman AM Sequestration of mRNAs Modulates the Timing of Translation during Meiosis in Budding Yeast Mol Cell Biol 2015 35 20 3448 3458 26217015 278 Honigberg SM Conicella C Espositio RE Commitment to meiosis in Saccharomyces cerevisiae: involvement of the SPO14 gene Genetics 1992 130 4 703 716 1582554 279 Honigberg SM Esposito RE Reversal of cell determination in yeast meiosis: postcommitment arrest allows return to mitotic growth Proceedings of the National Academy of Sciences of the United States of America 1994 91 14 6559 6563 8022820 280 Yang H Ren Q Zhang Z Chromosome or chromatin condensation leads to meiosis or apoptosis in stationary yeast (Saccharomyces cerevisiae) cells FEMS yeast research 2006 6 8 1254 1263 17156022 281 Davidson GS Joe RM Roy S Meirelles O Allen CP Wilson MR Tapia PH Manzanilla EE Dodson AE Chakraborty S Carter M Young S Edwards B Sklar L Werner-Washburne M The proteomics of quiescent and nonquiescent cell differentiation in yeast stationary-phase cultures Molecular biology of the cell 2011 22 7 988 998 21289090 282 Allen C Buttner S Aragon AD Thomas JA Meirelles O Jaetao JE Benn D Ruby SW Veenhuis M Madeo F Werner-Washburne M Isolation of quiescent and nonquiescent cells from yeast stationary-phase cultures J Cell Biol 2006 174 1 89 100 16818721 283 Palkova Z Vachova L Yeast cell differentiation: Lessons from pathogenic and non-pathogenic yeasts Semin Cell Dev Biol 2016 284 Honigberg SM Cell signals, cell contacts, and the organization of yeast communities Eukaryot Cell 2011 10 4 466 473 21296916 285 Piccirillo S Honigberg SM Yeast colony embedding method J Vis Exp 2011 49 286 Vachova L Hatakova L Cap M Pokorna M Palkova Z Rapidly developing yeast microcolonies differentiate in a similar way to aging giant colonies Oxid Med Cell Longev 2013 2013 102485 23970946 287 Cap M Stepanek L Harant K Vachova L Palkova Z Cell differentiation within a yeast colony: metabolic and regulatory parallels with a tumor-affected organism Molecular cell 2012 46 4 436 448 22560924 288 Baillie GS Douglas LJ Role of dimorphism in the development of Candida albicans biofilms J Med Microbiol 1999 48 7 671 679 10403418 289 Chandra J Kuhn DM Mukherjee PK Hoyer LL McCormick T Ghannoum MA Biofilm formation by the fungal pathogen Candida albicans: development, architecture, and drug resistance J Bacteriol 2001 183 18 5385 5394 11514524 290 Vachova L Palkova Z Physiological regulation of yeast cell death in multicellular colonies is triggered by ammonia J Cell Biol 2005 169 5 711 717 15939758 291 Vachova L Palkova Z Aging and longevity of yeast colony populations: metabolic adaptation and differentiation Biochemical Society transactions 2011 39 5 1471 1475 21936836 292 Chen H Fink GR Feedback control of morphogenesis in fungi by aromatic alcohols Genes Dev 2006 20 9 1150 1161 16618799 293 Vachova L Kucerova H Devaux F Ulehlova M Palkova Z Metabolic diversification of cells during the development of yeast colonies Environ Microbiol 2009 11 2 494 504 19196279 294 Vachova L Chernyavskiy O Strachotova D Bianchini P Burdikova Z Fercikova I Kubinova L Palkova Z Architecture of developing multicellular yeast colony: spatio-temporal expression of Ato1p ammonium exporter Environ Microbiol 2009 11 7 1866 1877 19302539 295 Palkova Z Devaux F Icicova M Minarikova L Le Crom S Jacq C Ammonia pulses and metabolic oscillations guide yeast colony development Mol Biol Cell 2002 13 11 3901 3914 12429834 296 Palkova Z Forstova J Yeast colonies synchronise their growth and development J Cell Sci 2000 113 Pt 11 1923 1928 10806103 297 Gordon DM The rewards of restraint in the collective regulation of foraging by harvester ant colonies Nature 2013 498 7452 91 93 23676676 298 Wilson EO Holldobler B Dense heterarchies and mass communication as the basis of organization in ant colonies Trends Ecol Evol 1988 3 3 65 68 21227094 299 Cap M Vachova L Palkova Z Reactive oxygen species in the signaling and adaptation of multicellular microbial communities Oxid Med Cell Longev 2012 2012 976753 22829965 300 Cap M Vachova L Palkova Z Yeast colony survival depends on metabolic adaptation and cell differentiation rather than on stress defense J Biol Chem 2009 284 47 32572 32581 19801643 301 St Johnston D Sanson B Epithelial polarity and morphogenesis Curr Opin Cell Biol 2011 23 5 540 546 21807488 302 Irvine KD Rauskolb C Boundaries in development: formation and function Annu Rev Cell Dev Biol 2001 17 189 214 doi 11687488 303 Rokas A The origins of multicellularity and the early history of the genetic toolkit for animal development Annu Rev Genet 2008 42 235 251 18983257 304 Hall-Stoodley L Costerton JW Stoodley P Bacterial biofilms: from the natural environment to infectious diseases Nat Rev Microbiol 2004 2 2 95 108 15040259 305 Niklas KJ Newman SA The origins of multicellular organisms Evol Dev 2013 15 1 41 52 23331916 306 Knoll AH The multiple origins of complex multicellularity Annu Rev Earth Planet Sci 2011 39 211 217 239

PMC005xxxxxx/PMC5134913.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 9106748 8519 Psychiatr Genet Psychiatr. Genet. Psychiatric genetics 0955-8829 1473-5873 27606929 5134913 10.1097/YPG.0000000000000148 NIHMS808672 Article Rapporteur Summaries of Plenary, Symposia, and Oral sessions from the XXIIIrd World Congress of Psychiatric Genetics Meeting in Toronto, Canada, October 16-20, 2015 Zai Gwyneth 12345* Alberry Bonnie 6 Arloth Janine 78 Bánlaki Zsófia 9 Bares Cristina 10 Boot Erik 311121314 Camilo Caroline 15 Chadha Kartikay 25 Chen Qi 16 Cole Christopher B. 2517 Cost Katherine Tombeau 118 Crow Megan 19 Ekpor Ibene 20 Fischer Sascha B. 21 Flatau Laura 22 Gagliano Sarah 48 Kirli Umut 24 Kukshal Prachi 25 Labrie Viviane 326 Lang Maren 27 Lett Tristram A. 28 Maffioletti Elisabetta 29 Maier Robert 30 Mihaljevic Marina 31 Mittal Kirti 12 Monson Eric T. 32 O'Brien Niamh L. 33 Østergaard Søren Dinesen 3435 Ovenden Ellen 36 Patel Sejal 25 Peterson Roseann E. 37 Pouget Jennie G. 235 Rovaris Diego Luiz 3839 Seaman Lauren 40 Shankarappa Bhagya 41 Tsetsos Fotis 42 Vereczkei Andrea 9 Wang Chenyao 43 Xulu Khethelo 44 Yuen Ryan K. C. 45 Zhao Jingjing 4647 Zai Clement C. 123 Kennedy James L. 1235* 1 Neurogenetics Section, Centre for Addiction and Mental Health, Toronto, ON, Canada 2 Campbell Family Mental Health Research Institute, Centre for Addiction and Mental Health, Toronto, ON, Canada 3 Department of Psychiatry, University of Toronto, ON, Canada 4 Frederick W. Thompson Anxiety Disorders Centre, Department of Psychiatry, Sunnybrook Health Sciences Centre 5 Institute of Medical Science, University of Toronto, Toronto, ON, Canada 6 Molecular Genetics Unit, Department of Biology, University of Western Ontario, London, ON, Canada 7 Max Planck Institute of Psychiatry, Munich, Germany 8 Department of Translational Research in Psychiatry, Institute of Computational Biology, Helmholtz Zentrum München, Germany 9 Department of Medical Chemistry, Molecular Biology and Pathobiochemistry, Semmelweis University, Budapest, Hungary 10 School of Social Work, University of Michigan, Ann Arbor, MI, USA 11 The Dalglish Family 22q Clinic, Toronto, ON, Canada 12 University Health Network, Toronto, ON, Canada 13 Clinical Genetics Research Program, Centre for Addiction and Mental Health, Toronto, ON, Canada 14 Department of Nuclear Medicine, Academic Medical Centre, Amsterdam, The Netherlands 15 Institute and Department of Psychiatry, University of São Paulo Medical School, São Paulo, Brazil 16 Department of Medical Epidemiology and Biostatistics, Karolinska Institutet, Stockholm, Sweden 17 Biomedical Sciences Division, Department of Biology, University of Ottawa, Ottawa, ON, Canada 18 Department of Psychology, University of Toronto, Toronto, ON, Canada 19 Stanley Institute for Cognitive Genomics, Cold Spring Harbor Laboratory, Woodbury, NY, USA 20 Department of Psychiatry, University of Calabar Teaching Hospital, Calabar, Nigeria 21 Human Genomics Research Group, Department of Biomedicine, University of Basel, Basel, Switzerland 22 Institute of Psychiatric Phenomics and Genomics, University of Munich, Munich, Germany 23 Molecular Biology Laboratory, National Center of Medical Genetic of Cuba, Cuba 24 Department of Psychiatry, Ege University School of Medicine, Izmir, Turkey 25 Department of Genetics, University of Delhi, South Campus, New Delhi, India 26 Krembil Family Epigenetics Laboratory, Centre for Addiction and Mental Health, Toronto, ON, Canada 27 Department of Genetic Epidemiology in Psychiatry, Central Institute of Mental Health, Medical Faculty Mannheim, University of Heidelberg, Heidelberg, Germany 28 Charité Universitätsmedizin Berlin, Germany 29 Genetics Unit, IRCCS Centro S. Giovanni di Dio, Fatebenefratelli, Brescia, Italy 30 Queensland Brain Institute, University of Queensland, St. Lucia, Australia 31 Clinic for Psychiatry, Clinical Center of Serbia, Serbia 32 Department of Psychiatry, University of Iowa, Iowa City, IA, USA 33 Molecular Psychiatric Laboratory, Division of Psychiatry, University College London, London, UK 34 Department of Clinical Medicine, Aarhus University, Aarhus, Denmark 35 Psychosis Research Unit, Aarhus University Hospital, Risskov, Denmark 36 Human Genetics Lab, Department of Genetics, Stellenbosch University, South Africa 37 Virginia Institute for Psychiatric and Behavioral Genetics, Virginia Commonwealth University, Richmond, Virginia, USA 38 Department of Genetics, Instituto de Biociências, Federal University of Rio Grande do Sul, Brazil 39 ADHD Outpatient Clinic, Hospital de Clínicas de Porto Alegre, Federal University of Rio Grande do Sul, Brazil 40 Department of Chemistry and Biochemistry, Department of Psychiatry and Biobehavioral Sciences, Semel Institute for Neuroscience and Human Behavior, University of California, Los Angeles, Los Angeles, CA, USA 41 Molecular Genetics Lab, Department of Psychiatry, National Institute of Mental Health and Neurosciences, Bangalore, Karnataka, India 42 Department of Molecular Biology and Genetics, Democritus University of Thrace, Alexandroupolis, Greece 43 Nagoya University, Japan 44 Department of Psychiatry, Stellenbosch University, South Africa 45 The Centre for Applied Genomics, Genetics and Genome Biology, The Hospital for Sick Children, Toronto, ON, Canada 46 School of Psychology, Shaanxi Normal University, Xi'an, China 47 School of Psychology, National University of Ireland, Galway, Ireland 48 Department of Biostatistics and Center for Statistical Genetics, University of Michigan, Ann Arbor, Michigan, USA Corresponding authors: Dr. James L. Kennedy, MD FRCPC, MSc, 250 College Street, Toronto, ON M5T 1R8, Canada; Tel: (416) 979-4987; Fax: (416) 979-4666; jim.kennedy@camh.ca; Dr. Gwyneth Zai, MD FRCPC, PhD, 250 College Street, Toronto, ON M5T 1R8, Canada; Tel: (416) 535-8501 ext. 30145; Fax: (416) 979-4666; gwyneth.zai@camh.ca * Both authors are co-senior authors. 10 8 2016 12 2016 01 12 2017 26 6 229257 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. The XXIIIrd World Congress of Psychiatric Genetics (WCPG) meeting, sponsored by the International Society of Psychiatric Genetics (ISPG), was held in Toronto, ON, Canada, on October 16-20, 2015. Approximately 700 participants attended to discuss the latest state-of-the-art findings in this rapidly advancing and evolving field. The following report was written by trainee travel awardees. Each was assigned one session as a rapporteur. This manuscript represents the highlights and topics that were covered in the plenary sessions, symposia, and oral sessions during the conference, and contains major notable and new findings. International Society of Psychiatric Genetics (ISPG) DNA genome-wide association study (GWAS) mood disorders psychiatric genetics schizophrenia World Congress of Psychiatric Genetics (WCPG) Introduction The International Society of Psychiatric Genetics (ISPG) was founded in 1992 with a mission to facilitate research in the genetics of psychiatric disorders and related traits and to promote education in psychiatric genetics. It sponsors an annual meeting, which is held in alternating cities between North American and European countries. The XXIIIrd World Congress of Psychiatric Genetics (WCPG) took place in Toronto, ON, Canada from October 16-20, 2015. Over 650 attendees in psychiatry, psychology, genetics, and other related fields had the opportunity to attend 65 scientific sessions. This meeting provided early investigator travel awards to 34 international and 11 local trainees to present their work at this meeting. One of the goals of this conference is to expand our reach to and involve other developing countries. Of the 32 international awardees who attended the meeting, nine (28%) presented work from their developing countries including Brazil, Cuba, India, Nigeria, Serbia, and South Africa. The 2015 congress was chaired by Dr. James L. Kennedy and the WCPG Rapporteur Program were chaired and organized by both Dr. Gwyneth Zai and Dr. Kennedy. Rapporteurs for the 65 conference sessions were early investigator awardees, each with a task to summarize individual session in addition to relevant discussions. This has been the tradition to summarize the conference sessions into a publication since 2007 in New York. The following sections are organized based on the date of the sessions, followed by sub-section of plenary sessions, symposia, and oral sessions. Friday October 16, 2015 Keynote Lecture Data Integration for Disease Gene Identifications: Genome × Transcriptome × EMR (reported by Robert Maier) Professor Nancy Cox (Vanderbilt University, USA) presented trait mapping results using PrediXcan (Gamazon et al., 2015), a gene-based association method that utilizes genetic and transcriptome data to understand the molecular mechanisms of disease phenotypes. In the first step, the Genotype-Tissue Expression (GTeX) database was used to train tissue-specific genetic predictors of gene expression levels for those 20% - 40% of genes whose transcript levels are at least moderately heritable. The least absolute shrinkage and selection operator (LASSO) regression analysis resulted in predictors of transcript levels that are based on approximately 60 – 80 cis expression quantitative trait loci (eQTLs) per gene. Half of those predictors have a prediction R2 greater than 0.2. Using predictors that are based only on the genetically determined part of expression has the advantage of bypassing reverse causation of phenotype on transcript levels. These predictors of transcript levels were then applied to the large BioVU dataset, the Vanderbilt's biorepository of DNA that has been extracted from discarded blood collected during routine clinical testing and are linked to de-identified medical records. The goal of the study was to perform a phenome wide association study (PheWAS) in which associations between disorders and predictors of gene expression are identified. The BioVU repository consists of electronic medical records (EMR) for more than two million individuals, 20,000 of which have been genotyped to date. Results from the gene-disease association tests based on approximately 5,000 BioVU subjects with heart tissue expression predictors of approximately 500 genes were then presented. One of the most interesting examples was a significant association between reduced predicted expression of the glutamate receptor, ionotropic kainate 5 (GRIK5) gene and various eye related disorders. A clustered regularly-interspaced short palindromic repeats (CRISPR) zebrafish knockout model subsequently validated the role of GRIK5 in eye development. Examples of genes associated with neurological and psychiatric phenotypes (which Professor Cox termed the “quintessential human phenotypes”) include the beta-1,4-N-acetyl-galactosaminyl transferase 4 ((B4GALNT4) gene with mood disorders, the direct IAP (inhibitor-of-apoptosis)-binding protein with low pI (DIABLO) gene with psychosis, and the cytohesin 2 (CYTH2), synaptic vesicle glycoprotein 2A (SV2A), chymotrypsin-like elastase family, member 2A (CELA2A), and prostate and testis expressed 2 (PATE2) genes with addiction, alcohol disorders and “failure to thrive”, respectively. Future plans include extension of the analysis to larger sample and additional genes, with particular focus on: a) genes related to Mendelian diseases and drug targets; b) experimental validation of current significant associations; c) prediction and association of up-regulated (as opposed to down-regulated) gene expression; and d) comparisons of PrediXcan predictions and polygenic risk score (PRS) predictions. The first question in the Q&A session was regarding the lack of significant associations of genes which have previously been associated with psychiatric disorders. Dr. Cox explained that she presented preliminary results based on only 500 genes, which have been analyzed to date. Dr. Mark Daley (Broad Institute, USA) inquired about statistical significance after multiple testing with many gene-phenotype combinations. Dr. Cox pointed out that the complex correlation structure in the EMR complicates corrections for multiple testing and that the use of eigenphenotypes could be helpful. Saturday October 17, 2015 Plenary Sessions International Initiatives in Cancer Genomics and Big Data (reported by Chris Cole) As the efficiency and accuracy of human genome sequencing increases, a new field of care has emerged. Precision medicine (formerly known as personalized medicine), tailoring therapies to the individuals based on their genetics, has gone from science fiction to scientific reality. Cancer medicine especially has become the vanguard of this rapidly advancing field, says Dr. Thomas Hudson, the president and scientific director of the Ontario Institute for Cancer Research (OICR) in Canada. Genetic information, only readily attainable after the recent 100,000-fold decrease in sequencing costs, is used to predict individual risk, optimize screening programs, and identify disease at earlier stages. Diagnostic genetics can lead to more precise diagnosis and more accurate prognostic interventions. The potential benefits of individualized therapy are large, as can be attested by the successes of several early interventions such as Gleevec and human epidermal growth factor receptor 2 (HER2) therapies, and the rate of discovery and clinical approval has been accelerating. Such discoveries, however, require global collaboration of a large scale, which is not often achievable. Dr Hudson has been at the heart of several efforts to centralize and standardize the collection and utilization of genomic data for healthcare. With 85 standardized projects deployed across the world, some of the preexisting mysteries of cancer are becoming clearer. From discovering 24 unique carcinogenic patterns of mutations to identifying Aristolochic acid in traditional medicine as carcinogenic, the approximately $20 million investment per center is starting to pay dividends. With tremendous amount of data being gathered, OICR has become the hub to deal with data privacy and availability. On the clinical side, Dr. Hudson has started a feasibility study with five sites in Ontario. The study, testing protocols as well as outcomes, examines the effect of personalized therapy on actionable genes in a population of patients beyond the standard of care. Though certain genes may be involved in disease pathways, the individual variants are often novel. The fact that many patients may have a unique mutation encourages the sharing of crucial data between physicians and researchers. To this end, Dr. Hudson and other international colleagues have established the Global Alliance for Genomics and Health, an international collaboration to accelerate progress in human health research and standardize procedures. Similar to the World Wide Web consortium assigning a standardized IP address, the Global Alliance allows researchers from around the world to speak the same language and quickly integrate their data. With 350 members in 35 counties and four separate working groups, the Global Alliance has tackled some of the most pressing issues of the genomic era. From a novel application program interface (API) for interacting with genomic data to a framework for the responsible sharing of genomic and health related data, the consortium has utilized expertise from clinical medicine, genetics, and computer science. Their current projects include the Beacon project, which allows physicians to light a “beacon” when a particular mutation is observed, and the breast cancer (BRCA) Challenge, where physicians can obtain a standardized answer to whether their patient's mutation is clinically significant. With emerging and growing new and exciting data, Dr. Hudson reminds us that individuals are keys to creating tools, and organizations are the best ways to gather and incorporate experiences from around the globe. The recent advances in sequencing technologies, cancer genomics, and targeted therapies have created the perfect platform for personalized medicine. It's up to us to capture and transform this potential into clinical practice. The Regulome in Psychiatric Therapy: Integrating Chromosomal Architecture, Genetic Variation, Epistasis, and Evolution (reported by Eric Monson) Dr. Wolfgang Sadee (The Ohio State University, USA) began his talk on the point that we are nearing the post-genome-wide association study (GWAS) era. GWAS findings have yielded a wealth of information but many results remain with unclear clinical significance, particularly because greater than 90% of these results reside within intergenic and intronic genomic regions (Welter et al., 2014). If further explored, these variants may offer critical insight to disease etiology, risk, and therapies. Dr. Sadee explained that the clinical significance of variants may depend on evolution, the three-dimensional architecture of human deoxyribonucleic acid (DNA), and/or epistatic interactions. Variants may be deleterious (typically rare variants) or provide fitness benefits except when combined with certain environmental stressors and/or epistatic effects (typically common variants). Such risk factors may remain hidden in GWAS analyses (Sadee et al., 2014). Variants may affect well-conserved but undescribed regulatory networks leading to broad effects not readily detectable in single nucleotide polymorphism (SNP) GWAS associations (Stergachis et al., 2014). Dr. Sadee cautioned that, due to these complexities, analysis focus must be balanced to capture only the information needed to describe causative variant effects and to avoid noise from surrogate markers and overlapping/competing regulatory systems in broad examinations (Sadee et al., 2014). Such noise may explain the recent lack of detected epistasis in GWAS assessments (Schizophrenia Working Group of the Psychiatric Genomics, 2014). Dr. Sadee then described methods, which are useful for the exploration of functional regulatory variant effects. Allelic expression imbalance (AEI) (Johnson et al., 2008) can identify variants that perturb the transcription, splicing, and translation of proteins. Broadening the scope of an initially narrow investigation can also help identify epistatic interactions. For example, Dr. Sadee's team examined the cytochrome P450 2D6 (CYP2D6) variant rs16947 (the CYP2D6*2 allele), described to have no effect on expression levels, but shown to have inconsistent behavior. They identified that rs16947 reduces CYP2D6 expression if present alone; however, if the high linkage disequilibrium (LD) variant, rs5758550, which is located 100kb away from CYP2D6, interacts with its promoter by DNA looping, increased expression is observed. The net result is normalized expression of CYP2D6, indicating the need to include both variants in clinical metabolism panels (Wang et al., 2014). Dr. Wang in Dr. Sadee's research group has also detected previously unknown regulatory networks between SNPs within/near the CYP3A family of genes via the circularized chromosome conformation capture (4C) analysis (unpublished results), which can identify potentially distant DNA regions that interact with a known site through chromosome conformation changes. Finally, Dr. Sadee's team found that the dopamine D2 receptor gene (DRD2) SNP rs2514218 is associated with schizophrenia and resides largely in the opposite haplotype to two SNPs (rs1076560 and rs2283265) that were found to disrupt splicing (Zhang et al., 2007). It was further found that the DRD2 SNP rs1076560 interacts with several dopamine transporter gene (SLC6A3) variants and environmental stress, which drastically increases the risk of death associated with heavy cocaine abuse (Sullivan et al., 2013). These findings demonstrated that future efforts to identify the function of disease-associated variants should thoughtfully utilize tools and evolutionary understanding to unravel potentially complex regulatory systems. Successes can offer important insights into the underlying basis of disease and offer appropriate targets for clinical applications. Worldwide Opportunities in Psychiatric Genetics Research (reported by Zoe Robaina) Dr. Lin He (Shanghai Jiao Tong University, China) reported the current developments in China, the value of special populations, and opportunities for international collaborations. He showed the significant progress in the identification of candidate genes for schizophrenia and other mental disorders by analyzing the genetic structure, GWAS and CNV in the Han Chinese population, as well as results of his team's investigations in a Chinese schizophrenia sample using various genetic approaches including GWAS, epigenetics, pharmacogenetics study, and knock-out mouse model study. Dr. Jingjing Zhao (Shaanxi Normal University, China and National University of Ireland, Ireland) commented that Dr. He's work represented a good example for worldwide opportunities in psychiatric genetic research and to foster international collaborations. Dr. Zhao also agreed with Dr. He's opinion that the ISPG board of directors should include representatives from China and other developing countries and that one of the future WCPG annual meetings should be held in China in order to promote worldwide opportunities in psychiatric genetics. Dr. Thelma B.K. (University of Delhi South Campus, New Delhi, India) highlighted the utility of studying populations of different ethnicities to unravel the genetic basis of both complex as well as monogenic disorders in humans using the contemporary genome-wide SNP arrays and whole exome sequencing tools. Drawing examples from complex traits in the genetically distinct Indian population that her group has been working on, she demonstrated: a) the differences in the genome architecture of the Indian populations in comparison to the Caucasian and other HapMap populations; b) consequent limited replication of Caucasian meta-analysis findings in Indian case-control cohort studies in rheumatoid arthritis (RA) and ulcerative colitis (UC); c) discovery of novel susceptibility loci from GWASs of Indian RA and UC cohorts; and d) the contribution of such a population to the international consortium on celiac disease for example. She further shared her team's exciting findings of novel disease causal variants in Mendelian forms of X-linked intellectual disability, Parkinson's disease, and schizophrenia. She elaborated her work on schizophrenia using exome sequencing technique. This study sample consisted of 17 families of Indian origin with at least two or more members having a diagnosis of schizophrenia. Novel variants including compound heterozygotes in a few biologically/pharmacologically relevant genes have been found to segregate with disease in some of the families. Her team's recent discovery of a mutation in the trace amine associated receptor 1 (TAAR1) gene in a family with autosomal dominant form of schizophrenia has provided a strong genetic evidence for the role of this gene, of potential pharmacological relevance in disease etiology. Dr. Homero Vallada (University of Sao Paulo Medical School, Brazil) spoke about the Brazilian population admixture, which is generally more diverse than the Caucasian population. The observed diversity in the Brazilians is in part due to the large geographical landscape and the migration of several different ancestral origins in Brazil throughout history. The population distribution within the large country gives raise to isolated or semi-isolated groups, which offer good platforms for genetic investigation in general and psychiatric genetic research. Differences in genetic profile and exposure to specific environments may result in different phenotypes including potential psychopathologies. Dr. Vallada presented his work on the molecular genetics investigations of crack cocaine addiction and significant association was detected for genetic variations in the butyrylcholinesterase (BChE) gene and the risk of crack cocaine addiction. He also reported that crack cocaine appeared to be more addictive than the powder form of cocaine. Dr. Chunyu Liu (University of Chicago, USA) discussed the ISPG Global Diversity Task Force with the goals to increase global efforts in psychiatric genetics research and to reduce barriers for global research and education. Therefore, a workshop in South Africa (2015) and two annual meetings in China, the first and second “Summit on Chinese Psychiatric Genetics” (2014 and 2015), were organized to address these aims. During the Chinese summits, investigators were given the opportunity to present their latest research and discuss the current state and future directions of psychiatric genetics. In line with the Task Force's mission, the participation of early career investigators was strongly encouraged. This informal research organization is steadily growing with more than 30 participants representing researchers from various countries. It will be spearheading initiatives to promote collaborations and data sharing in China. This project will serve as a blue print for similar activities to be held in Eastern Europe, Latin America, India, and Africa in the future. Challenges in Genetic Testing and Counselling (reported by Erik Boot) In this plenary panel session, Dr. Francis McMahon (Johns Hopkins University and NIMH, USA) started by presenting a general overview on “genetic testing and precision medicine in psychiatry”. He first discussed potential uses of genetic testing, including the formulation of differential diagnosis, the prediction of treatment outcomes in terms of response and adverse events, and the identification of high-risk individuals. He continued speaking on several key questions related to genetic testing in clinical psychiatric practice. The first question that he raised was whether a certain genetic marker can be genotyped reliably. Another question was how valid the association is between the genetic marker and psychiatric disease. Finally, he raised the question whether the test result has any clinical utility. Dr. McMahon noted that genetic testing has already been utilized in psychiatry in terms of commercial panels marketed to psychiatrists and psychologists and direct-to-customer tests for patients, their relatives, and other individuals. He provided several examples of promising genetic and pharmacogenetic testing in addition to tests currently in use. He noted that the best studied predictive factors to date are not from genetics, but are based on diagnosis, clinical features, family history, treatment adherence, comorbidity, and other biomarkers. Dr. McMahon raised the issue of incidental and secondary findings that can arise from any GWASs. He stated that there is currently no consensus protocol in place to deal with this concern of identifying, reporting and counselling based on unanticipated findings. He mentioned that the American College of Medical Genetics published recommendations for reporting incidental findings in clinical whole exome sequencing findings that should be reported back to the patients; however, guidelines are not yet in place to interpret them. Dr. McMahon discussed that individuals considering genetic testing should receive genetic counselling prior to testing in order to discuss the impact of anticipated and incidental results. Finally, he stressed the importance of providing further education to clinicians and patients, and the need for additional research. Dr. Jehannine Austin (University of British Columbia, Canada) led a case discussion on practical and psychosocial issues that can emerge from genetic testing for psychiatric disorders. Subjects of discussion included appreciation of the importance of exploring and explaining in lay language the etiology of mental illness to patients and their family members, in addition to reviewing how to address psychosocial issues associated with genetic counseling and genetic testing for mental illness. Dr. Austin presented a simplified version of the additive model of risk of developing a psychiatric disorder using the “mental illness jar” analogy. Psychiatric disorders are likely caused by a combination of genetic and environmental factors (i.e., if the jar becomes full with factors depicted as shapes in the jar [crosses the threshold of normal behaviour], the individual experiences an episode of mental illness). Finally, she emphasized that genetic tests will not be able to 100% predict whether a person will or will not develop a mental illness; however these tests may provide important contributions to clinical practice in psychiatry. Oral Sessions ADHD/Child Behaviour (reported by Qi Chen) Dr. Andrea Johansson Capusan (Linkoping University, Sweden) described findings from a population based twin study of 18,000 adult twins. The study aimed to investigate the extent to which the association between childhood maltreatment and symptoms of Attention Deficit Hyperactivity Disorder (ADHD) in adults can be explained by familial confounding (i.e. familial factors that are shared by siblings within the same family but different between families) and whether or not it is consistent with a causal interpretation. The results showed that childhood maltreatment was significantly associated with higher self-rated DSM-IV ADHD symptom scores in adults. Within twin pair analysis showed decreasing but significant estimates for dizygotic (DZ) twins and monozygotic (MZ) twins, indicating that the association is in part explained by familial confounding, but is likely to be causal. Dr. Qi Chen (Karolinska Institutet, Sweden) shared findings from a population based family study on the familial aggregation of ADHD in over eight million relative pairs consisting of twins, full siblings, maternal and paternal half siblings, full cousins, half cousins. Significant associations measured by hazard ratios (HRs) were observed in all subgroups of relative pairs. The magnitude of HRs was reduced with decreasing genetic relatedness. The study found no obvious etiological difference in ADHD between males and females. If family members were affected by ADHD persistent into adulthood, the familial aggregation appeared to be even stronger, indicating such families could be considered a high-risk group and may require diagnostic screening. Dr. Ditte Demontis (Aarhus University, Denmark) presented findings from a meta-analysis of GWASs of ADHD based on the largest ADHD data freeze to date, consisting of 18,000 ADHD cases and 34,000 controls. The study revealed 10 genome-wide significantly associated loci with ADHD and served as an important step leading towards future research in dissecting the genetic architecture of ADHD. Dr. Beate St Pourcain (University of Bristol, UK) presented a study in which social-communication difficulties were found to be genetically correlated with ADHD traits and clinical ADHD. The genetic correlations varied with age, with stronger correlation being observed before age 10 and after age 12 for ADHD traits. The findings supported that there are shared genetic influences between social-communication difficulties and ADHD traits in the general population, as well as clinically-diagnosed ADHD, which may depend on developmental stage. Dr. Evie Stergiakouli (University of Bristol, UK) presented a study investigating the association between ADHD and smoking status and alcohol consumption during pregnancy and breastfeeding. Polygenetic risk score analysis was used to disentangle the genetic effects from prenatal environmental risks. Higher polygenic score of ADHD was associated with higher odds of smoking but not for alcohol before pregnancy and in non-breastfeeding mothers. The findings confirmed that shared genetic effects may play a role in the association between ADHD and smoking during pregnancy and breastfeeding. Dr. Christie Burton (University of Toronto, Canada) presented a hypothesis-driven genome-wide association study (GWAS-HD) of a quantitative obsessive-compulsive (OC) trait in youth from the community. Two SNPs in an intron of protein tyrosine phosphatase receptor type D (PTPRD) gene reached genome-wide significance for the OC traits. SNPs in neuronal PAS domain protein 2 (NPAS2) and the central nervous system (CNS) developmental gene set and the CNS development gene-set as a whole were also associated with OC traits, supporting the hypothesis that genetic variants with functional implication in brain development may be involved in obsessive-compulsive disorder (OCD). This session emphasized the power of using GWAS-HD approach and the importance of using quantitative trait in the general population to boost statistical power for future psychiatric genetic research. Bipolar and Mood Disorders (reported by Sascha Fischer) Dr. Melvin McInnis (University of Michigan, USA) presented results from a gene expression study in induced pluripotent stem cells reprogrammed to neurons and glial cells, from individuals affected with Bipolar Disorder (BD) and controls. They found a total of 82 differentially expressed microRNAs (miRNAs). Differences in neuronal lineage allocation were also observed: whereas BD neurons prefer ventral medial ganglionic eminence derivatives, control neurons prefer dorsal cortical precursors. In addition to these results, differences in calcium signaling were detected in BD neurons. BD neurons were more active than control neurons but displayed reduced calcium signaling with lithium pre-treatment. Dr. Niamh O'Brien (University College London, UK) reported study results from a High Resolution Melting (HRM) analysis of four calcium channel genes in 1,098 patients affected with BD. Two non-synonymous CACNG4 variants were associated with mental illness (rs371128228, p=1.05×10-4, OR=4.39 and 17:65026851 (C/T), p=0.0005, OR=9.52). The rs371128228 marker was associated with reduced glutamate receptor AMPA 1 level at the cell surface. Based on a replicated GWAS finding in the CACNA1C gene, data from 99 whole-genome sequenced BD individuals were analyzed. Two variants associated with BD (p=0.015, OR=1.15) were detected in the third intron of CACNA1C. These variants were associated with significantly decreased gene expression. Ms. Niamh Mullins (King's College London, UK) reported on her GWAS and PRS results of suicide attempts in mood disorders, mainly BD and Major Depressive Disorder (MDD) from PGC data. They analyzed 1,075 suicide attempters and 7,081 non-attempters with MDD, 1,852 suicide attempters and 3,285 non-attempters with BD, as well as 18,771 controls in two ways: within-cases analysis (attempters versus non-attempters) and attempters versus controls; separately for each cohort and between cohorts. In suicide-attempters with MDD vs. controls, one genome-wide significant finding was identified on chromosome 14 (rs8013144, P=8.60×10-11, OR=2.2). Dr. Andreas J. Forstner (University of Bonn, Germany) reported on his findings of shared risk loci and pathways between schizophrenia and BD. Association testing was conducted for the 128 schizophrenia-associated SNPs (Schizophrenia Working Group of the Psychiatric Genomic Consortium, 2014) in a large GWAS dataset of BD comprising 9,747 patients and 14,278 controls (Mühleisen et al., 2014). After reimputation and correction for control overlap, 22 schizophrenia-SNPs showed nominally significant p values in the BD GWAS. The strongest associated SNP was located near the tetratricopeptide repeat and ankyrin repeat containing 1 (TRANK1) gene (p=8.8×10-8). Pathway analysis using INRICH and Ingenuity pathway analysis revealed 25 nominally significant canonical pathways including calcium and glutamate signaling. Dr. Fernando Goes (Johns Hopkins University, USA) presented findings of a whole-exome sequencing study on a BD family sample. Four to five affected individuals from each of eight multiplex families were exome sequenced and analyzed for rare variants (minor allele frequency [MAF] <1%). Eighty-four rare damaging, segregating variants in 82 genes were detected and association testing was conducted in independent samples with a total of 3,541 BD cases and 4,774 controls. No significant association for genes or variants remained after correction for multiple testing. The detected risk genes in BD families displayed an overlap with recently identified genes for autism and schizophrenia. Ms. Monika Budde (Medical Center of the University of Munich, Germany) presented a study on the genetic basis of functional outcome in BD. 2,957 LD-based regions were tested for their association with the Global Assessment of Functioning (GAF) score, a measure of social, occupational, and psychological functioning. In a joint analysis of linkage disequilibrium (LD) blocks with putative functional pertinence across 511 German and 1,081 US BD patients, one LD block on chromosome 15 was significantly associated with GAF (kernel score test: p=1.29×10-5 metric GAF; p=5.64×10-6 GAF-extremes). Schizophrenia: Pathways, RNA and CNVs (reported by Marina Mihaljevic) Mr. Aswin Sekar (Harvard Medical School, Boston, USA) reported on complex structural variation in the Major Histocompatibility Complex (MHC) locus as underlying the association of schizophrenia to the MHC region (Schizophrenia Working Group of the Psychiatric Genomics Consortium, 2014). Using novel methods, he characterized various structural forms of the complement component 4 (C4) gene and showed that these structural forms affect expression of C4 in human brain tissue and are associated with schizophrenia risk in proportion to their effect on C4 expression. He also presented data suggesting a role for C4 in synaptic pruning in mice and concluded that these findings could potentially help explain the pathological finding of synapse loss in schizophrenia brains. Mr. Mads Engel Hauberg (Aarhus University, Denmark) further explored the potential role of miRNAs in the etiology of schizophrenia. He presented a statistical ‘gene set association’ approach to find miRNAs that are regulators of schizophrenia genes and functional genetic variants relating to miRNA. He highlighted that miR-9-5p and miR-137 are regulators of common variant schizophrenia risk genes and are themselves also risk genes. Ms. Jeannie Pouget (Centre for Addiction and Mental Health, Canada) presented the first comprehensive evaluation of genetic overlap between schizophrenia and 18 autoimmune diseases, according to their epidemiological associations (Benros et al., 2014). She systematically analyzed genome-wide significant autoimmune SNPs with the Psychiatric Genomic Consortium (PGC) genotype data. Results showed no evidence of genetic overlap between schizophrenia and any of the 18 autoimmune diseases and no support for autoimmune-driven subsets of schizophrenia. Further research will include SNPs with more liberal thresholds for association with autoimmune diseases. Dr. Peter Holmans (Cardiff University, UK) investigated extensive pathway analysis of the largest PGC schizophrenia dataset. He combined results from seven pathways analysis methods which had been applied to 9,016 pathways from large generic pathway sets and 183 candidate pathways regarding particular biological hypotheses. This multiple analysis confirmed significant enrichment of pathways related to dopaminergic synapse, postsynaptic density, seizures, calcium channels and FMRP targets, with considerable genes overlapping among the aforementioned pathways, and suggested further study of their biological mechanisms. Dr. Daniel Howrigan (Massachusetts General Hospital, Boston, USA) presented novel methods for the analysis of rare copy number variants (CNVs) in schizophrenia, applied to cohort from PGC study of schizophrenia. CNV association testing was controlled for genotyping platforms, ancestry and CNV calling metrics. Results confirmed increased CNV burden in schizophrenia. Deletions were significantly enriched among gene sets related to synaptic function and activity-regulated cytoskeleton-associated (ARC) protein complex. Duplications showed enrichment in N-methyl-D-aspartate (NMDA) receptor complex. He presented evidence for Xq28 to emerge as a schizophrenia CNV ‘hotspot’. Dr. Jacob Vorstman (Rudolf Magnus Institute, Utrecht, Netherlands) discussed new data on the cumulative burden of genetic double hits in schizophrenia. He combined concurrent CNV and SNP data in large Dutch cohort recruited from the Genetic Risk and Outcome in Psychosis (GROUP) Consortium. Preliminary results showed increased burden of deleterious impact inferred by double hits in deleted sequence in schizophrenia and difference between cases and controls driven by higher number of and higher degree of deleteriousness of the disease-associated SNPs (dSNPs: functional SNPs in genes affected by CNVs). These dSNP effects were not detected in duplicated sequence. He concluded that deletions co-occurring with a functional SNP on the remaining allele could be an additional mechanism involved in etiology of schizophrenia. Symposia Sessions Genetic Aspects of Behavioural Addictions: New Insights from Human and Pre-Clinical Methods (reported by Cristina Bares and Fotis Tsetsos) Dr. Daniela Lobo (Centre for Addiction and Mental Health, Canada) spoke about pathological gambling and described a study in which addiction related genes were selected from previous studies and their own research in the KARG (Knowledgebase for Addiction Related Genes) database. In their study in humans, Dr. Lobo observed an association between pathological gambling and the rs167771 single-nucleotide polymorphism in the dopamine receptor (DRD3) gene, after correction for age. When they corrected for sex, they found an association with the calcium/calmodulin-dependent protein kinase 2 delta (CAMK2D) rs3815072 marker (Lobo et al., 2014). The DRD3 functional marker Ser9Gly has been previously associated with addiction (Kreek et al., 2005), but Dr. Lobo did not find an association in her study (Mulert et al., 2006). Dr. Fiona Zeeb (Center for Addiction and Mental Health, Canada) focused on the environmental factor of gambling disorder. As dopamine sensitization is present in pathological gamblers, Dr. Zeeb examined whether repeated exposure to gambling opportunities caused dopamine sensitization and possibly contribute to problem gambling. Using the rat gambling task (rGT), developed by Zeeb and colleagues, she found that rats exposed to repeated sessions of uncertainty (akin to chronic gambling scenarios in human patients) resulted in dopamine sensitization. This uncertainty exposure also increased risky decision-making on the rGT. Furthermore, increased risky decision-making also enhanced sensitization. Dr. Jose Nobrega (Centre for Addiction and Mental Health, Canada) used the rGT to examine possible brain changes by in situ hybridization (ISH) in the genes identified by Dr. Lobo. The ratio of high vs low risk choices was analyzed for correlations with the ISH. A significant correlation was observed between the levels of DRD3 in the islands of Calleja and high-risk options. He also investigated the link between impulsivity and deep brain stimulation (DBS) in rats, with inconclusive results. Lastly, by using the rGT in a depression model, he reported that escapable stress might have beneficial effects to impulsivity, but inescapable stress may worsen the condition. Mr. Michael Barrus (University of British Columbia, Canada) talked about the gambling models that they have developed, the cued version of the rGT, the rodent slot machine task, the rodent betting task and the loss chasing, and their applicability in their research. He reports that the use of all models provides insight into different biological aspects of gambling, such as the dopamine D4 receptor in the anterior cingulate cortex. The discussion, which was led by Dr. Vincenzo de Luca (University of Toronto, Canada), focused on the validity of what is measured in the animal models, how the measurements in rats map to human behavior. Other topics of discussion included: the variability of the animals in terms of age and strain and the validity of the time out negative reinforcer. It was mentioned that the negative reinforcer used in the rGT and negative reinforcers used by other groups cannot fully capitulate the losses experienced by problem gamblers. However, the use of timeout periods detracted from the main reward in the rGT. Therefore, the negative reinforcement is somewhat similar to what human gamblers experience. It was acknowledged that the way by which loss is modeled is a limitation of the paradigm. Polygenic Score Methodology in Psychiatric Genetics (reported by Janine Arloth and Lauren Seaman) Dr. Frank Dudbridge (London School of Hygiene and Tropical Medicine, UK) presented an enlightening overview of the theory and applications of PRSs. He described the technique as a vital component to examine the missing heritability of a multitude of complex psychiatric disorders since risk prediction for these phenotypes is typically challenging. He provided information on previously reported study design parameters to help researchers who are interested in using this informative analysis. (Dudbridge, 2013) In brief, he ended with a discussion of novel software, AVENGEME, which can investigate “chip heritability” (i.e., the heritability explained by SNPs on a specific genotyping array), genetic correlations, and the effect size of SNPs to the risk of developing the examined trait or disorder. Overall, the field aims to move from gene discovery to optimizing phenotypic prediction, as well as to address the entire genetic risk of these enigmatic diseases. Mr. Jack Euesden (King's College London, UK), introduced a single command line tool to measure PRSs called “PRSice” (Eusden et al., 2015). It provides the best-fit PRS for all calculated and tested PRSs of different SNP sets at different p-value thresholds. He discussed the importance of controlling for variants in LD when performing PRSs. PRSice handles this problem by using the PLINK software command “clump” (Chang et al., 2015). Furthermore, he discussed the issue of causal variants, which are more likely to reside in functional regions. He compared the performance of PRSice to penalized regression models (LASSO and elastic-net models) and found that PRSice outperforms these latter models. Finally he showed a new PRS method, called “PRSlice”, to identify biomarkers/PRSs for a phenotype without having GWAS data for this phenotype available. Mr. Robert Maier (University of Queensland, Australia) presented his work on multivariate PRSs, which is based on genotype summary statistics. Standard PRS methods do not account for LD structure and thereby losing information by simply excluding SNPs based on a certain LD measure and p-value threshold. He showed two methods to measure PRS without excluding any SNP and without having the full genotype data available. At first, he showed how to use approximate Best Linear Unbiased Prediction (BLUP) to estimate effects from GWAS. Such SNP-BLUP models intrinsically account for LD between SNPs. The second method he showed was the multi-trait BLUP that evaluates risk across multiple disorders by combining single trait BLUP into multi-trait BLUP of random effects. Finally, he showed an application of both methods using the PGC data for schizophrenia and BD. (Maier et al., 2015) He identified a small decrease in prediction accuracy when using summary statistics (single-trait BLUP) in comparison to using samples with full genotype data. Furthermore, by combining SNP effects from different traits (multi-trait BLUP for two traits: schizophrenia and BD), the prediction accuracy was further improved. Ms. Hilary Finucane (Massachusetts Institute of Technology, USA) discussed how to employ GWAS summary statistics to partition heritability by functional categories. This approach can shed new light into statistical models for quantitative phenotypes or endophenotypes, especially in large psychiatry samples, since some of these categories can disproportionately contribute to the observed heritability. She spoke about the concern that, while there is much information to be extracted from large meta-analyses, variance components methods are intractable with the increased sizes as well as requiring complete genotypic data. Her group's proposed method is to utilize summary statistics (i.e., LD and stratified LD score regressions) to calculate partitioned heritability (Finucane et al., 2015). Insights into the Genetic Architecture and Molecular Markers of Major Depression from the CONVERGE Project (reported by Diego Rovaris and Khethelo Xulu) Dr. Kenneth Kendler (Virginia Commonwealth University, USA) opened the symposium by introducing the CONVERGE (China Oxford and VCU Experimental Research on Genetic Epidemiology) project. Dr. Kendler explained the main purpose of the CONVERGE study, emphasizing a large sample size (N = 12,000). The CONVERGE project aims to identify molecular markers conferring susceptibility to the development of MDD. To reduce genetic heterogeneity, it was designed to include only Chinese Han women and exclude cases with depression related to substance abuse. To date, it is the largest single study consisting of one single consistent phenotype. The CONVERGE project consists of 59 participating hospitals from 45 cities of 21 provinces in China. To reduce the likelihood of misclassification of controls, all control participants were personally interviewed. In addition, the CONVERGE project has information regarding environmental risk factors for both cases and control participants. This allowed for the modeling of genome wide gene-environment interactions. Researchers from the CONVERGE study presented results across a variety of completed or in-progress analyses. Dr. Tim Bernard Bigdeli (Virginia Commonwealth University, USA) started by reporting on the progress made in understanding the genetic architecture of MDD of Chinese Han women. The project has identified two genome-wide significant variants contributing to the risk of MDD development (CONVERGE Consortium, 2015). These two loci are located on chromosome 10, one in the 5′ region of the sirtuin1 (SIRT1) gene (P = 2.53 × 10−10) and another in an intron of the phospholysinephosphohistidine inorganic pyrophosphate phosphatase (LHPP) gene (P = 6.45 × 10−12). When the analysis of 4,509 cases was restricted to a severe subtype of MDD, melancholia, there was an increase in the effect size and significance of the signal at the SIRT1 locus. The CONVERGE project attributed their success to the recruitment of a homogeneous cohort with severe illness. Results were replicated in a sample of Chinese Han men and women but were not replicated in the PGC MDD samples of European descent, which is perhaps due to differences in allele and haplotype frequencies. Dr. Roseann Peterson (Virginia Commonwealth University, USA) talked about gene-environmental (G×E) interactions in the CONVERGE project. Significant main effects of childhood sexual abuse (CSA) and stressful life events (SLE) on MDD were found and accounted for upwards of 11% of the variance in MDD, as well as interesting G×E interactions between variants in the SIRT1 gene and CSA (P = 0.008), and variants in the LHPP gene and SLE (P = 0.0002). Dr. Peterson also showed that environmental risk factors can change GWAS results: When individuals of high environmental exposure were removed from genetic analyses additional genetic variants were implicated in MDD risk including variants in the mitochondrial iron transporter (SLC25A37), lysophosphatidylglycerol acyltransferase 1(LPGAT1), and the putative uncharacterized protein Clorfl95/inositol-trisphosphate 3-kinase B (Clorfl95/ITPKB) genes. Ms. Na Cai (Oxford University, UK) presented results showing molecular changes and potential molecular markers of MDD from the CONVERGE study (Cai et al., 2015). Here, they followed up on the findings from the human studies by using animal models to investigate any changes of mitochondrial DNA (mtDNA) and telomere length, using stressed mice versus controls. Stressed mice have been found to have more mtDNA in comparison to controls. Furthermore, telomere length in stressed mice was shortened when compared to controls, corroborating the results found in humans. In addition, to test whether the hypothalamicpituitary-adrenal axis plays a role, mice were injected with corticosterone. Mice that were injected with corticosterone were found to have decreased telomere length in comparison to controls. The series of findings suggested that the molecular changes might be a consequence of MDD. Dr. Bradley Todd Webb (Virginia Commonwealth University, USA) spoke about associations between oral microbiome and MDD in the CONVERGE study. He showed that the oral microbiome is robustly associated with MDD and these differences between cases and controls can be shown quantitatively and qualitatively. Moreover, this association may be partly influenced by the use of medication. Dr. Bradley pointed out that these results come from an exploratory study, which does not allow a clear distinction between correlation and causation. Finally, Dr. Douglas Levinson (Stanford University, USA) briefly discussed the findings obtained in the CONVERGE study. He recognized the effort to collect a large and homogeneous sample and also spoke on the SNP-heritability results found in the CONVERGE GWAS, which was one of the greatest successes in MDD genetic research to date. Sunday October 18, 2015 Plenary Sessions The Notorious Past and Bright Future of Psychiatry (reported by Katherine Tombeau Cost) Dr. Jeffrey Lieberman (New York State Psychiatric Institute and Columbia University, USA) presented a plenary session on the mystery of mental illness and psychiatry's notorious efforts to solve it. Dr. Lieberman's comments were largely based on his recently published book, SHRINKS: The Untold Story of Psychiatry (Lieberman, 2015) http://www.jeffreyliebermanmd.com/index.html. He began by noting that psychiatry was the only specialty in all of medicine to have a specific movement opposed to it. The “anti-psychiatry” initiative was started about 50 years ago, by Thomas Szasz, a psychiatrist at State University of New York in collaboration with L. Ron Hubbard, a science fiction author and founder of the Church of Scientology. Dr. Szasz's motivation stemmed from a desire to be an academic provocateur, while Mr. Hubbard's desire to discredit psychiatry derived from an economic and competitive market share interests to convince potential converts of the value of his Dianetics philosophy and the Scientology approach over psychiatric medicine. The anti-psychiatry movement gained steam in the cultural turmoil of the 1960's and evolved into an aggressive, pernicious, and persistent effort to deny the existence of mental illness and the ability of psychiatry to understand and treat it. According to Dr. Lieberman, this disaffection with psychiatry was not entirely unfounded, and contributed to by the historical missteps of the profession. Until the latter twentieth century, psychiatry was not scientifically driven and had largely been unable to accurately explain the bases of mental disorders, and had produced minimal effective treatments to alleviate the symptoms and ease the suffering of patients. Although psychiatric illnesses have been documented for centuries, it was not until relatively recently that more accurate diagnoses and effective treatments became available. Dr. Lieberman described several notable milestones in the history of psychiatry. In 1844, psychiatrists formed the first medical specialty professional association called the Association of Medical Superintendents of American Institutions for the Insane, which was a precursor to the American Psychiatric Association (APA). At the time, the prevailing scientific approach to understanding human disease was to examine anatomical pathology, and this was more difficult and less fruitful in psychiatry. Therefore, mental illness was often ascribed to meta-physical causes, which often resulted in ineffective, silly, inhumane, and often harmful “cures”. Philippe Pinel (1745-1826) was heralded for releasing asylum patients from their chains and creating humane environments where “moral therapy” was practiced. But still, from the late 18th century to the mid-20th century, over millions of patients were held in institutions under deplorable conditions. During this time, the theories of Francis Galton on eugenics, Sigmund Freud on psychoanalysis, and Walter Freeman on lobotomies flourished. In the 1970's the APA commissioned Robert Spitzer to revamp the nosology of psychiatry, in attempt to make diagnoses more empirically based and less arbitrary. Spitzer worked with many groups and professionals to develop consensus on the conditions listed in the DSM-III, famously declassifying homosexuality as a mental illness and, in collaboration with Dr. Nancy Andreason, to formally classify Post Traumatic Stress Disorder (PTSD). At this same time, effective psychopharmacology (including antipsychotic, antidepressant, mood stabilizing and anxiolytic drugs) and psychosocial treatments (such as Dr. Aaron Beck's Cognitive Behaviour Therapy [CBT] and Gerald Klerman and Myrna Weissman's Interpersonal Therapy [IPT]) were developed and experimentally verified to reduce suffering. Psychiatry has finally become a scientifically based and clinically competent medical specialty that is able to benefit from progress through research. Consequently, the previous “stepchild of medicine” is now able to meet the challenges of mental illness and mental health care including reducing stigma, dysfunctional and inequitable health care policy and financing, inadequate infrastructure, services, and workforce needs. Epigenetics of Psychiatric Disease: Progress, Problems and Perspectives (reported by Bonnie Alberry) Dr. Art Petronis (Centre for Addiction and Mental Health, Canada) discussed epigenetics in psychiatric disease. He introduced epigenetics as instructions – how DNA should be read. He highlighted that a perfect genome could be ruined with erroneous epigenetics. Dr. Petronis outlined epigenetic relevance to disease using three postulates. First, epigenetic factors contribute to phenotype, evidenced by the agouti mouse phenotype (Morgan et al., 1999). Second, there is partial stability, whereby marks are modified by developmental programs via environmental or stochastic events. Partial stability is exemplified through the ten-eleven translocation (TET) enzymes, which actively demethylate cytosines. Third, epigenetics are a secondary mechanism of heritability. Epigenetics were initially considered only heritable in somatic cells. Due to two large epigenetic reprogramming events –in primordial germ cells and in zygotes – transgenerational inheritance was thought to be impossible. Many exceptions have since been found, including the agouti mouse model (Morgan et al., 1999). As the zygotic reprogramming event is less harsh, epigenetic recombination occurs at fertilization, underlying uniqueness of zygotes. Dr. Petronis explained epigenetics as responsible for disease etiology (Petronis, 2010). MZ twins have DNA modification differences, due to environmental or stochastic factors. Meanwhile, DZ twins have greater differences (Kaminsky et al., 2009). Dr. Petronis suggested epigenetic differences in DZ twins are due to zygote epigenetic diversity. The question of how to identify DNA-independent zygotic epigenetic heritability was then explored. Dr. Petronis and his team employed a model using inbred mice to generate artificial MZ twins, gestating genetically identical offspring and a MZ twin pair (Gartner and Baunack, 1981). In mice, Gartner and Baunack (1981) found MZ twins had greater similarity than polyzygotic littermates, and intangible variation was not explained by genetics or environment. Dr. Petronis suggested DNA modifications as a candidate to explain heritability through zygotic epigenomes. Dr. Petronis introduced work investigating SNPs exhibiting allele-specific DNA modification (ASM-SNPs). Brain ASM-SNPs were significantly enriched in schizophrenia patients in GWAS. The distribution of ASP-SNPs was skewed towards the most significant GWAS SNP p-values. ASM-SNPs were most common in functional sites, stressing the importance of DNA modifications in regulatory regions. Lastly, Dr. Petronis used epigenetic studies of lactose intolerance to model the development of schizophrenia, emphasizing temporal dimension. Dr. Petronis suggested psychiatric disease behaves like multiple, age-dependent, ‘lactose-intolerance’-like epigenetic situations. As time passes, DNA modifications accumulate at schizophrenia risk SNPs, leading to symptom peaks. In discussion, Dr. Petronis addressed histone modifications also playing an important, epigenetic role. However, he suggested that while relevant, less is known in disease context, DNA modifications are more stable to investigate than histone modifications. Dr. Petronis added that while DNA methylation changes with age, there are also fluctuations that may contribute to the episodic nature of psychiatric illnesses. Identifying Illness and Treatment Biological Markers through Transcranial Magnetic Stimulation (reported by Viviane Labrie) Dr. Zafiris Jeffrey Daskalakis from the Centre for Addiction and Mental Health presented a plenary lecture on the benefits of transcranial magnetic stimulation (TMS) in treating major depression and as a method to probe neurophysiological function in psychiatric disorders. He first presented data showing that GABA neurotransmission deficits in psychiatric disorders can be detected using a TMS-based motor inhibition paradigm. Inhibitory neurotransmission mediated by the GABA system can be activated by TMS resulting in a cortical silent period – a suppression of motor response. Several psychiatric disorders have deficits in the cortical silent period, though patterns of deficits differ between disease types (Radhu et al., 2013). The atypical antipsychotic clozapine was found reverse the impaired cortical silent period in schizophrenia, suggesting that clozapine may mediate symptomatic relief through the GABA pathway. Dr. Daskalakis also demonstrated that TMS can be applied to assess GABA-mediated cortical inhibition in the prefrontal cortex, a brain area of considerable importance to psychiatric illness. Interestingly, prefrontal cortical inhibition was shown have some degree of heritability, where deficits in cortical inhibition were significantly higher among healthy relatives of patients with schizophrenia than in unrelated controls. This demonstrated evidence that cortical inhibition could be a useful biomarker to help identify psychiatric diseases like schizophrenia. Dr. Daskalakis completed his talk by demonstrating the applicability of TMS for medication-resistant depression. Induction of therapeutic seizures by magnetic stimulation was found to be a useful alternative to electroconvulsive therapy for depression, as the seizures could be better localized to the affected neural tissues, which minimized side effects while significantly improving symptoms in treatment-resistant depression. Symposia Sessions Sequencing, Direct-To-Consumer-Testing, Biobanking: The Explosion of Ethical Challenges in Psychiatric Genetics (reported by Laura Flatau and Prachi Kukshal) Dr. Jehannine Austin (University of British Columbia, Canada) gave a talk on how to apply genetic counselling to problems arise in adolescent psychiatry. The major concerns in this area include counselling families with an affected child or parent and the impact of psychiatric disorders on the child or adolescent, family dynamics, and social stigma. Dr. Austin reported that the process of counselling with family members is more important than disclosing the exact risk of developing an illness. She recalled times during her genetic counselling practice when it was crucial for her to handle the problems mentioned above empathetically. She presented several case examples and illustrated the need for thoughtful and tailored counselling to help patients to deal with their family dynamics and to discuss a well-rounded approach in explaining the genetics and environmental risk of psychiatric illnesses. Ms. Rosa Spencer Tansley (Bournemouth University, UK) presented a study on the quantitative and qualitative methods focusing on the responses of patients and their families to psychiatric genetic counselling. She reported that the perception and expectations towards genetic counselling influence the patient's engagement with the service and patient outcome. The data (57 patients and 29 family members) that she presented suggested that although many perceived psychiatric genetic counselling as beneficial, misconceptions about the service and ethical considerations in regard to its delivery were noted, indicating an urgent need to educate the public regarding genetics, gene-environment interaction, genetic counselling as a discipline, and its application in psychiatry. Her study showed that there is a strong demand for psychiatric genetic counselling but public awareness is relatively low and therefore, there is a need to resolve misconceptions by educating the public. Ms. Laura Flatau (Ludwig-Maximilians-University Munich, Germany) talked about the ‘Right Not To Know’ especially in the context of incidental findings. She presented the results of a quantitative survey study with 536 participants including the general population, patients and medical healthcare professionals. Her findings suggested that although the majority of individuals (∼80%) would like to receive information about an incidental finding, there are specific cases (i.e., hereditary cancer) in which 25% of the participants would choose their ‘Right Not To Know’. Comparing the attitudes between different groups, individuals with a higher education level tended to be more critical towards genetic testing, and they were more likely to choose their ‘Right Not To Know’. Attitude towards wanting information versus the ‘Right Not to Know’was found to be affected by the way the question was asked (i.e., concrete scenarios vs. simple questions) and the individual to whom was asked (i.e., general population or health care professionals). Mr Fuji Nagami (Tohoku University, Japan) presented data from the Tohoku Medical Megabank project (ToMMo). It is an ongoing project to reconstruct and establish the public health systems in a community of 150,000 participants who have been affected by the 2011 Tohoku earthquake and tsunami in Japan. The aim of the project was to use research findings of common diseases (i.e., cancers, cardiovascular diseases, strokes, diabetes, and mental diseases) with gene-environment interaction for the constructive regeneration of such disastrous events. By addressing the various ethical issues related to psychiatric problems arising from such stressful situation, the project aimed to build an integrated biobank. This biobank contains bio-specimens, questionnaire data, and physiological survey data from cohort studies and analytical datasets, including genomic and other omics data from a subset of the total sample. The examination of various aspects of psychological well-being including the occurrence of mental health problems (i.e., posttraumatic stress reaction [PTSR], anxious state, and depressive state) showed a negative impact of natural disasters on mental health. Individuals who were affected by the earthquake had almost double the national average rate of mental health problems including PTSR, anxious and depressive state. Mr. Nagami identified several ethical issues (i.e., biobank by genome cohort studies, return of results, mental health research in area affected by disaster, and data sharing) related to the setup of the Tohoku Medical Megabank Project, including the collection of large amount of data while protecting the privacy of individuals. Dr. Marcella Rietschel (Clinical Institute of Mental Health, Germany) was the moderator and Dr. Thomas G. Schulze (University of Munich, Germany) was the chair for the session. Dr. Austin and others stressed the need to improve the education of medical trainees and psychiatrists in patient counselling besides prescribing drugs. Counselling should be tailored to each individual on a case-by-case basis using clinical judgment and at the same time, respecting the individual's autonomy if one chooses the ‘Right Not To Know’. Furthermore, discussion was focused on the extent of psychiatric genetic counselling and the differences between general and psychiatric genetic counselling given that such distinction may lead to further stigmatization of psychiatric illnesses. Dissecting the Genetic Contribution to Depression: Progress at Last (reported by Elisabetta Maffioletti and Roseann Peterson) Dr. Douglas F. Levinson (Stanford University, USA) opened the symposium with a discussion of the difficulty faced in the identification of specific genetic variants predisposing to MDD. Despite considerable heritability, as demonstrated by twin and family studies (Sullivan et al., 2000), earlier efforts by the PGC showed no genome-wide significant results, even with sample sizes of over 9,000 MDD cases and 9,000 controls (Ripke et al., 2013a). Dr. Levinson suggested that the lack of findings may be due to the need for even larger sample sizes to reach the ‘inflection point’ at which sample size and significance of variants increases proportionately (Ripke et al., 2013a). The heterogeneity of MDD may require the identification of homogeneous subgroups in which statistical power to detect the modest effect sizes expected is maximized. He concluded by emphasizing that screening of controls (to reduce the probability of MDD in the control sample), stricter definition of case status, as well as limiting analyses to more severe forms of MDD (i.e., recurrent depression subtype) will likely aid gene finding efforts. Dr. Naomi R. Wray (University of Queensland, Australia) examined potential sources of heterogeneity across studies leading to differences in SNP-based heritability estimates for MDD between PGC-MDD1 (18%) and PGC-MDD2 (9%). First, she examined the genetic correlation (rG) between males and females, finding estimates near unity, indicating that it was premature to conclude that there was lower rG between the sexes for MDD when compared to other psychiatric disorders. Dr. Wray then highlighted that there were significant differences in SNP-based heritability by cohort, indicating unknown sources of heterogeneity across samples, and also reported lower rG among individual MDD cohorts when compared to schizophrenia and BD samples. She suggested that this heterogeneity may be due to potential different environmental factors across studies, loose definitions between cases and controls, and broad ascertainment biases. Dr. Cathryn M. Lewis (King's College London, UK) presented recent genome-wide association meta-analyses of MDD conducted by the PGC, using an expanded sample size of over 16,000 cases. In the current PGC-MDD ‘data freeze’ of 29 cohorts, no genome-wide significant findings were detected. However, when examining results by sex, significant associations were identified for females only in nitric oxide synthase 1 (NOS1) (rs76821249, p=2.2×10-8) and for males only in leucine rich repeat and fibronectin type III domain containing 5 (LRFN5) (rs8016327, p = 5.5×10-8). Adding to PGC-MDD29, the Kaiser Permanente Genetic Epidemiology Research on Adult Health and Aging (GERA) cohort (7,162 cases, 38,307 controls), and 23andMe (14,906 cases, 41,465 controls) which comprised a total sample size of 38,991 cases and 105,404 controls yielded a significant hit on chromosome 5 (hg19 position: 103903810; p=3.8×10-8). When stratifying the sample by sex, significant associations of a marker in the Major Histocompatibility Complex, Class I, human leukocyte antigen B (HLA-B)region (p=2.9×10-8) in females and a locus in the Huntingtin gene (HTT) (p=1.1×10-8) in males were found. When further meta-analyzed with the CONVERGE sample, a variant in the LRFN5 gene, which was previously significant in males of PGC-MDD29 sample, was also associated with MDD in the combined analysis (p=4.5×10-8). Dr. Kenneth S. Kendler (Virginia Commonwealth University, USA) presented results from the CONVERGE study, a whole-genome sequencing study of 5,303 Han Chinese women with recurrent MDD and 5,337 screened controls. The data were collected from 59 hospitals across China and represented one of the largest and most homogeneous MDD cohorts with the following inclusion criteria: (a) recruiting females only, (b) cases with severe form of recurrent major depressive episodes through clinical interview, and (c) screened controls past the age of typical MDD onset. Dr. Kendler reported that they successfully detected and replicated two common variants contributing to MDD risk on chromosome 10q: upstream of SIRT1 and in an intron of LHPP (Converge Consortium, 2015). He also commented on the genetic architecture of MDD, reporting that (1) genome-wide SNP-based heritability was estimated as 21-28%, (2) the heritability in MDD explained by each chromosome was proportional to its length (r=0.680) thus supporting a highly polygenic etiology, (3) the variance explained was distributed across the allelic frequency spectrum, (4) partitioning by genic annotation indicated a greater contribution of SNPs in coding regions and within the 3′-UTR regions, and (5) that DNase hypersensitive sites in many cell types including brain-related cells were enriched for associations with MDD (Peterson et al., submitted to Mol Psychiatry). Dr. Patrick F. Sullivan (University of North Carolina at Chapel Hill, USA and Karolinska Institutet, Sweden) presented evidence for shared genetic contributions between MDD and both psychiatric traits and physical characteristics, using a GWAS summary statistics approach (Bulik-Sullivan et al., 2015a). Dr. Sullivan reported significant genetic correlations between MDD and schizophrenia (rG=0.396), BD (rG=0.407), ADHD (rG=0.505), depression symptoms (rG=1.0), neuroticism (rG=0.831), smoking status (rG=0.286), early onset stroke (rG=0.312), migraine without aura (rG=0.169), and cardiovascular disease (rG=0.188). He also noted several limitations of the study including (1) limited power of the studied samples included, (2) the use of summary statistics as opposed to using full raw information, (3) the rG approach (Bulik-Sullivan et al., 2015b) applied has not been designed for analysis across multiple ancestry groups, (4) inability to rule out confounding genetic effects, and (5) potential sampling bias. Dr. Sullivan concluded by stating that this approach may be useful for interconnections of psychiatric disorders and to highlight common genetic architecture across complex disorders. Mitochondria Genetics and Function in Psychosis (reported by ZsófiaBánlaki) The chairs of the symposium, Dr. Vanessa Goncalves and Dr. James L. Kennedy (Centre for Addiction and Mental Health, Canada) introduced the session highlighting that the investigation of mtDNA variants is a promising but technically challenging, and yet under-explored field in psychiatric genetics. One reason can be explained by the variable mtDNA copy number, which can reach 1,000 per cell and the presence of heteroplasmy. Wildtype and mutant mtDNA proportions are highly variable. Thus, as Dr. Goncalves described, although deficit in the oxidative phosphorylation (OXPHOS) of mitochondria has been implicated in schizophrenia, the recent PGC GWAS did not support a role of mitochondrial function in schizophrenia (Ripke et al., 2013b; Schizophrenia Working Group of the Psychiatric Genomics, 2014); however, this GWAS did not investigate genes within the circular mtDNA genome. Very few studies focusing on mtDNA variants showed that somatic mutation rates vary with tissue types and are higher in certain brain regions of schizophrenia patients when compared to healthy individuals (Rollins et al., 2009; Sequeira et al., 2012; Sequeira et al., 2015). The present study analyzed 42 common and 167 rare single nucleotide polymorphisms (SNPs) in 4,778 cases and 5,819 controls. A rare and six common variants reached nominal significance, but they did not survive testing for multiple comparisons. Haplogroup analysis detected a higher rate of schizophrenia in the J-T group characteristic to the European Caucasian population. The mitochondrially encoded cytochrome b (MT-CYB) rs3088309 marker was the top hit for association with schizophrenia. The fact that rs3088309 is a missense variant with potential functional relevance (unpublished data) further supports its role in the pathogenesis of schizophrenia. As it was remarked during the discussion period, maternal inheritance in schizophrenia could provide additional evidence for the relevance of mtDNA variants, but this has not been analyzed in the present study and literature data are controversial. Dr. Marquis Vawter (University of California, USA) reported on the findings of mitochondrial hypofunction in schizophrenia and the genetic background of schizophrenia. Previous literature data have consistently implicated mitochondrial dysfunction in the pathophysiology of schizophrenia; however, it is difficult to differentiate between the cause and effect of this dysfunction. Dendritic spine loss, reduction in mitochondria copy number, and decreased expression of mitochondria encoded transcripts are all characteristics of schizophrenia. Evidence suggests that epistasis between genes from the nuclear and mtDNA may play an important role in the etiology of schizophrenia. eQTL analysis showed a strong enrichment of approximately 1,000 autosomal mitochondrial genes in the cortex (Kim et al., 2014), and common mtDNA variants were found to contribute to the risk of several common complex diseases including schizophrenia (Hudson et al., 2014, Sequeira et al., 2012). Dr. Vawter presented preliminary analysis of some large recent GWAS results showing a modest over-representation of nuclear encoded mitochondria genes in schizophrenia. Preliminary data showed an increase in the rate of non-synonymous mtDNA mutations. The exact localization and copy number of mitochondria within dendrites and axons using a case-control study design is currently in progress. A question was raised regarding the issue of clonal expansion and Dr. Vawter discussed that although heteroplasmic mutations are generally not tissue-specific, certain types of mutation can accumulate at specific sites, such as large deletions in dopamine innervated regions. This may be related to mtDNA dynamics, stability, and non-homologous recombination. Thus, it was recommended that large GWAS studies should incorporate mtDNA variants along with nuclear SNPs for epistatic interactions between both genomes. Dr. Dost Ongur (McLean Hospital/Harvard Medical School, USA) presented the results on his magnetic resonance spectroscopy (MRS) studies of bioenergetic abnormalities in psychosis. Since gamma oscillation producing cells such as inhibitory GABAergic interneurons consume high level of energy as shown by their enrichment with mitochondria, these cells are believed to be critical in the development of cognitive disorders when energy supply is depleted (Kann et al., 2014). MRS has previously been shown to be a useful tool for assessing the levels of the rapidly mobilizable energy reserve phosphocreatine (PCr) and the immediate energy source adenosine triphosphate (ATP) in vivo (Du et al., 2012). In both chronic and first-episode schizophrenia patients, marked reduction was observed in the PCr peak, providing evidence for a reduced enzymatic reaction rate for creatine kinase (Du et al., 2014). The pH scale also became more acidic in chronic patients when compared to first-episode patients, suggesting enhanced anaerobic glycolysis. Correlation analysis between ATP/PCr levels and pH is currently underway. In contrast to patients with schizophrenia, bipolar I disorder patients detected normal PCr/ATP level and pH. However, upon photic stimulation, ATP but not PCr level was reduced in the visual cortex of patients with bipolar I disorder, whereas the pattern was reversed in healthy individuals, indicating an inability to decrease the PCr level in bipolar I disorder patients at high energy demand (Yuksel et al., 2015). Investigation of creatine kinase function is in progress. These findings implicated that schizophrenia may be characterized by a severe and pervasive bioenergetic failure and bipolar I disorder may require brain activation to unmask abnormality in a compensated bioenergetics system at rest. This further suggested bioenergetic dysfunction in response to environmental factors in bipolar I disorder. Compromised bioenergetics thus may lead to abnormal brain function in psychotic disorders. This may potentially reveal novel drug targets related to the mitochondria. Dr. Dorit Ben-Shachar (Rambam Health Care Campus and Technion-Israel Institute of Technology) provided evidence for a multifaceted mitochondrial dysfunction in peripheral cells and postmortem brains. Dr. Ben-Shachar reported that the enzymatic activity of complex I of the OXPHOS system, nicotinamide adenine dinucleotide (NADH) dehydrogenase, was found to be higher in both medicated and non-medicated schizophrenia patients at the acute state, while reduced at the residual state, as compared to major depression and bipolar I disorder patients and healthy controls. This activity was accompanied by altered expression of three nuclear encoded genes, NADH dehydrogenase (ubiquinone) flavoprotein 1, 51kDa (NDUFV1), NADH dehydrogenase (ubiquinone) flavoprotein 2, 24kDa (NDUFV2) and NADH dehydrogenase (ubiquinone) Fe-S protein 1, 75kDa (NDUFS1), encoding different subunits of complex I. These abnormalities were accompanied with reduced synthesis rate of complex I, pathological interaction between dopamine and the complex and impaired cell respiration. Dr. Ben-Shachar also reported that neuronal differentiation of induced pluripotent stem cells (iPSCs) reprogrammed from schizophrenia hair follicle keratinocytes showed that differentiation into dopaminergic neurons was severely impaired, while glutamatergic neurons failed to mature. These impairments were associated with various deficits in mitochondria, similar to those previously observed in schizophrenia (Robicsek et al., 2013). Transfer of isolated active normal mitochondria into schizophrenia cells restored respiratory function, and reduced dopamine toxicity, while only partially restored mitochondrial network dynamics. These significant positive effects lasted for about three weeks, and then gradually faded. In addition, transfer of healthy mitochondria improved differentiation of schizophrenia derived iPSC into glutamatergic neurons. The presented data pinpoint mitochondria as an additional pathological factor in schizophrenia and suggest a role for mitochondria in neuronal differentiation. Mitochondria transfer may lead to new treatment approaches for brain diseases with developmental connectivity and bioenergetics abnormalities such as schizophrenia. During the discussion, a comment was provided regarding mitochondria haplogroup effects, which could also be considered at transfer and that the impact of complex I inhibitors in animal models would be intriguing to investigate. Oral Sessions Genome-wide Approaches in Other Disorders (reported by Andrea Vereczkei) Dr. Erin Dunn (Harvard Medical School, USA) reported on a GWAS conducted on a Hispanic sample with generalized anxiety disorder (GAD). Since the Hispanic population is highly underrepresented in psychiatric genetic research it was important to see which genetic variants are common among this population. GWAS was carried out on GAD symptoms and a SNP, rs78602344 in the thrombospondin 2 gene (THBS2) reached genome-wide significance. However, efforts to replicate this finding in three independent Hispanic samples did not confirm this result. Since the disease prevalence in Hispanic population is approximately half of the European population, larger replication sample sizes are needed for future studies in anxiety disorders of Hispanics. Dr. Sandra Meier (Center for Register-based Research, Denmark) presented the associations of anxiety disorders and depression with increased mortality. Clinical anxiety represents a core symptom of several different anxiety disorders, which is highly heterogeneous because patients with an anxiety disorder often present with comorbid psychiatric conditions. The goal of the study was to compare the mortality rate between different anxiety disorders including GAD, social anxiety disorder, agoraphobia, specific phobia, panic disorder, OCD, acute stress reaction and posttraumatic stress disorder (PTSD) in 50,000 patients with an anxiety disorder followed between 2002 and 2011. The results showed no familial confounding factors. Patients with anxiety disorders had higher rate of natural and unnatural causes of death. Individuals with comorbid depression were particularly more likely to die by unnatural causes. Mr. Eric Monson (University of Iowa, USA) discussed the results of a whole-exome sequencing study of BD patients who attempted suicide. Suicidal behavior is the most severe outcome of psychiatric disorders, and has a heritability of 30-50%. Primarily candidate gene studies and GWASs have been used to examine common variation in suicidal behavior to date. This study examined rare functional variations within suicidal behavior. 387 BD patients with a history of suicide attempts and 631 BD patients with no past suicide attempts were enrolled in the study. Mr. Monson analyzed over 800k genetic variations and no genome-wide significance was identified. Top hit genes with p<0.01 were chosen for further analyses. Within these analyses, a significant enrichment score of synapse associated genes was detected. Dr. Chia-Yen Chen (Massachusetts General Hospital, USA) presented findings from a GWAS on army soldiers with a history of trauma exposure in the United States. Trauma exposure is an essential diagnostic criterion for PTSD and also poses an increased risk for depression, substance use disorders, and anxiety. Twin studies also showed a 47-60% heritability for trauma exposure. In the Army Study To Assess Risk and Resilience in Service members (Army STARRS) sample, life-time cumulated trauma exposure was analyzed in GWAS. Two cohorts were included in the study: new soldiers and soldiers deployed to Afghanistan, with a total of approximately 18,000 samples with genotype data. Association study was carried out based on different ethnic groups. In the European American samples, a locus in the low-density lipoprotein receptor class A domain-containing protein 4 (LDLRAD4) gene on chromosome 18 was implicated in suggestive association with trauma. This gene was previously found to be associated with BD and schizophrenia. In the African American population, a locus in the leucine rich repeat containing 4C (LRRC4C) gene, which was previously found to be associated with BD, was significantly associated with trauma. Both findings were not replicated in other populations. Dr. Laura Bierut (Washington University School of Medicine, USA) discussed the role of the cholinergic receptor, nicotinic, alpha 5 (CHRNA5) gene in nicotine dependence, smoking status, and lung cancer. Smoking behaviour and lung cancer have been linked to markers on chromosome 15. The present study shows evidence for a complex relation amongst rs16969968 SNP, also known as Mr. Big of CHRNA5 and smoking quantity, smoking cessation, as well as lung cancer risk. However, the allele frequency of this marker varies across different populations (35% in Central European, 6% in African-American, 3% in Asian), although the odds ratios remain similar. This study also showed that exhaled carbon-monoxide is a stronger predictor of lung cancer than self-reported smoking status. Dr. Bierut concluded that the rs16969968's low- and high-risk genotypes may be associated with a 4-year delay in smoking cessation. This may in turn provide earlier detection of lung cancer in these patients. Dr. Laramie Duncan (Harvard Medical School, USA) reported GWAS results on anorexia nervosa. Female adolescents are amongst the highest risk of developing anorexia nervosa. It is characterized by preoccupation with weight, body image, and food. It also has the highest mortality rate of all psychiatric disorders. Approximately 4,000 cases were analyzed in the present GWAS and one single SNP reached the genome-wide significance level: rs11174203 in the family with sequence similarity 19 (chemokine [C-C motif]-like), member A2 (FAM19A2) gene on chromosome 12. Heritability for anorexia was calculated using LD score regression (LDSC), and the point estimate of 0.23 is comparable to other psychiatric disorders. Genetic correlations were positive and significant with schizophrenia and BD, but negative (and yet significant) with body mass index (BMI). Epigenetics and Other Approaches (reported by Ryan K. C. Yuen) Dr. Therese Murphy (University of Exeter, UK) presented a study of DNA methylation profiling in the brains of MDD suicide completers. The DNA methylation profiles between 20 depressed suicide completer cases and 20 non-psychiatric, sudden-death controls in two brain regions (Brodmann Area 11 [BA11] and Brodmann Area 25 [BA25]) were compared. They identified a region at an immune-related non-coding gene, psoriasis susceptibility 1 candidate 3 gene (PSORSIC3), which was significantly hypomethylated in cases compared to controls. Dr Murphy further identified a co-methylated module, which was significantly correlated with both MDD and a Suicide attempt polygenic risk score. Dr. Eilis Hannon (University of Exeter, UK) investigated the correlation of DNA methylation between blood and brain to determine if blood sample can be used as a surrogate for DNA methylation studies of the brain. Comparing the inter-individual variation of DNA methylation in blood, prefrontal cortex, entorhinal cortex, superior temporal gyrus and cerebellum from 75 individuals, she found that the predictive power of blood for the brain was low, only less than 20% of the variance can be explained. Sites with positive correlation were found, but much of the correlation was due to genetic influence on DNA methylation. Dr. Carolin Purmann (Stanford University, USA) presented a novel approach called Combined Long-Insert Paired-End and Capture (CLIP-Cap) sequencing to resolve complex genomic rearrangements. With the use of average ∼9kb insert size paired-end sequencing targeting on chromosome 22q, she showed that CLIP-Cap was capable of determining the heterozygous terminal 22q13.3 deletion and the isodicentric breakpoints. She further showed that the assay was able to detect other balanced structural variations, such as the Philadelphia translocation. She suggested that this approach can potentially detect all the structural variants in the captured reads as long as the target region is known. Dr. Gail Davies (University of Edinburgh, UK) reported a large-scale genome-wide association study on verbal-numerical reasoning (n=36,035), memory (n=112,067), and reaction time (n=111,483). Using a customized Affymetrix array targeting on common SNPs, she reported genome-wide significant regions on chromosomes 7, 14 and 22 for verbal-numerical reasoning, chromosomes 2 and 12 for reaction time, but no significant region was found for memory. Mr. Tarjinder Singh (Wellcome Trust Sanger Institute, UK) presented a meta-analysis of whole exome sequencing studies in schizophrenia including data from the UK10K consortium. By analyzing the de novo and rare (MAF<0.1%) loss-of-function (LoF) variants in a total of 4,264 cases and 9,343 controls, they found that the LoF variants in the KMT2F gene coding for SET domain containing 1A were significantly associated with schizophrenia (P=3.3E-9). There were 3 de novo LoF variants from trio families and 7 LoF variants identified from case-control samples. KMT2F is a member of a family of genes where disruptive variants result in dominant Mendelian disorders of histone machinery. Substance Abuse (reported by Ibene Ekpor) Dr. Andrew Bergen (SRI International, USA) introduced the ‘SmokeScreen Genotyping Array’ as a genome-wide array designed for addiction studies. He presented his work on Modeling Tobacco Exposures including the role of Nicotine Metabolic Enzymes. He explained that available data from the Total Exposure Study (TES) was analyzed including evaluation of nucleic acid quality, bio specimens and clinical chemistries. The results were correlated with existing data. The findings were that multivariate analysis participants with banked bio specimens were significantly more likely to self-identify as white, to be older, to have increased total nicotine equivalents per cigarette and decreased serum cotinine. In an analysis of three existing nicotine metabolism studies with participants of three continental ancestries using the smoke screen array, Dr. Bergen and collaborators identified genome-wide significant association of common variants at CYP2A6. They estimated that the top ranked SNP accounts for 12 – 27% of nicotine metabolite ratio (NMR) variation. Dr. Bergen disclosed that they were able to identify individual SNPs at nicotine metabolic enzymes in nicotine metabolism that can be used to model nicotine metabolism and increase the power of models. Dr. Ian Gizer (University of Missouri, USA) presented the result of their research work on the whole genome sequence analysis of cannabis dependence across two independent cohorts. He explained that qualitative genetic studies have established a genetic etiology of cannabis use disorder. He reported that their study was focused on gene-and path-way-based analysis of both common and rare variants obtained via whole-genome sequencing from two cohorts of predominantly European ancestry and predominantly Native American ancestry. The participants (n = 2,529) were people who met DSM-IV cannabis dependence based on the Semi-Structured Assessment for the Genetic of Alcoholism (SSAGA). All the participants whole genome sequence data was analyzed, while gene-based analysis of rare variants were conducted using the optimized sequence Kernel association test (SKAT-O). The result showed that gene-based analysis of rare coding variants (MAF<0.02) yielded significant evidence of association for a single gene with cannabis dependence (C1ORF110), and a suggestive evidence of an association with a second gene (microfibrillar associated protein 3 [MFAP3]). In addition, pathway analyses revealed significant evidence for the enrichment of genes related to potassium ion transport. He suggested that the results require replication with large samples. Mr. Eric Diehl (University of Western Ontario, Canada) described changes in the hippocampus in a mouse model of fetal alcohol spectrum disorder (FASD). Diehl explained that epigenetic dysregulation of genetic programs in the brain are involved in FASD. Diehl's laboratory's model of FASD shows learning and memory impairment and persistent changes in the brain gene expression into adulthood in a mouse model of FASD. 70 days old Mouse pups injected with saline or ethanol at post-natal days 4 and 7 had their hippocampus isolated and used for gene and miRNA expression microarray, methylated DNA immunoprecipitation microarray and histone H33 lysine 41 trimethylation and H33 lysine 27 trimethylation ChIP-chip. The results were dozens of gene and miRNA expression changes in the hippocampus of adult mice exposed to ethanol during development and hundreds of epigenetic methylation changes. These genes were predominantly oxidative stress-related. One of the ways alcohol induces oxidative stress of the developing brain is to reduce antioxidant levels and increase reactive oxygen species. The observed oxidative stress footprint may persist into adulthood hence identification of this mechanism may provide potential diagnostic targets or therapeutic approaches to help those affected by FASD. Dr. Jennifer Ware (University of Bristol, UK) explained the relevance of using biomarkers to carry out objective assessment of the various behavioural phenotypes of tobacco users. She discussed the results of a GWAS meta-analysis of levels of cotinine, the primary metabolite of nicotine based on 4,548 daily smokers of European ancestry, and identified variants in two genomic regions, 15q25.1 and 4q13.2, to be associated with cotinine levels. Furthermore, she discussed the limitation and benefits of GWAS employing alternative tobacco use biomarkers such as exhaled carbon monoxide levels. Dr. Jack Euesden (Kings College London, UK) commented on the importance of looking at smoking behaviour as a relevant phenotype in further studies. Ms. Bonnie Alberry (University of Western Ontario, Canada) discussed the result of the effect of continuous prenatal alcohol exposure (PAE) and post-natal maternal separation in mouse behavior as well as gene expression in the hippocampus. Behavioral tests showed learning deficit due to PAE and postnatal maternal separation. The expression of a large number of genes was also altered as a result of PAE with or without postnatal maternal separation. Ms. Alberry drew the following conclusions: the experimental model they used represents a realistic model; independent and comprehensive assessment of array gene expression as well as RNA sequencing will yield a highly reliable list of altered genes than relying on qPCR for confirmation of a few select genes. Monday October 19, 2015 Plenary Session Mitochondria and their Potential Role in Neuropsychiatric Disorders (reported by Maren Lang) A pressing question in biomedical science today, according to Dr. Douglas Wallace of the Center for Mitochondrial and Epigenomic Medicine at the Children's Hospital of Philadelphia, USA is why can't we understand and cure the common “complex” disorders. He postulates that our lack of success in addressing these crucial clinical concerns is the inadequacy of the underlying assumptions upon which we have based our investigations. The prevailing conceptual frameworks (paradigms) of western medicine are that diseases are anatomically based and that all genes are located on chromosomes and thus inherited according to the laws of Mendel. Indeed, all anatomical genes are chromosomal and Mendelian. However, to be alive requires not only anatomy but also the energy which animates us and this energy is generated primarily by the mitochondrial oxidation of our food with the oxygen that we breathe via mitochondrial oxidative phosphorylation (OXPHOS). The most important OXPHOS energetic genes are coded by a DNA located within the mitochondrion, the mitochondrial DNA (mtDNA), while all of the anatomical genes for the mitochondrion are located in the nuclear DNA (nDNA). The mtDNA is maternally inherited and present in hundreds to thousands of copies per cell. The high mtDNA copy number means that cells can contain mixtures of mutant and normal mtDNAs (heteroplasmy) which randomly segregate during mitosis and meiosis to give variable energetic defects. Different organs rely on mitochondrial energy to different extents. The brain has the highest mitochondrial energy demand, representing 2% of our body weight yet using 20% of our oxygen, so mild, systemic, mitochondrial, energy defects preferentially affect the brain. Hence, Wallace proposes that mild mitochondrial defects are the primary cause of neuropsychiatric disease. Mitochondrial energy defects can result from alterations in mtDNA or nDNA coded mitochondrial genes or from aberrant interactions between the two sets of mitochondrial genes. Mitochondria also communicate the cellular energetic status to the nucleus through mitochondrially-generated high energy intermediates that modulate the cellular signal transduction pathways and the epigenome. For example, the mtDNA tRNALeu(UUR) mutation at nucleotide (nt) 3243A>G is associated with autism and diabetes at 10-30% 3243G mutant, neuromuscular disease at 50-90% mutant, and lethal perinatal disease at 100% mutant. The phase-like changes in phenotype in response to continuous changes in 3243G heteroplasmy are the result of abrupt changes in the nDNA transcriptional profile, presumable reflecting epigenomic transitions. The mtDNA also accumulated mutations along the maternal lineages as women populated Africa. After only two mtDNA successfully left Africa, the mtDNA accumulated additional mutations as women migrate to Eurasia and the Americas. A subset of these mutations caused functional changes in OXPHOS which permitted regional adaption to local environments giving rise to groups of descendent haplotypes known as haplogroups. The physiological differences between regional haplogroups have been found to predispose to a wide range of neurologic and psychologic disorders including Alzheimer and Parkinson Disease, stroke, macular degeneration, deafness, and depression. Mutations in nDNA coded mitochondrial genes also cause disease. The severity of the nDNA mutation phenotypes can be further modulated by the mtDNA variation. To establish the causal role of mitochondrial variation in neuropsychological disease, Wallace reported the generation of a series of mouse lines with genetic alterations in nDNA and/or mtDNA mitochondrial mutations. Creation of the analogous mouse nDNA and mtDNA mutation combinations found in humans resulted in the same phenotypic manifestations. Mixing two normal mouse mtDNAs within the female mouse germline resulted in mice with a bipolar-like phenotype associated with a severe memory defect. Mice harbouring various nDNA or mtDNA mutations were found to show strikingly different responses to stress. Mild mitochondrial defects were also found to impaired embryonic migration of interneurons, which Wallace hypothesized cause of the excitation-inhibition imbalance associated with attention deficient-hyperactivity syndrome, compulsive behaviour, autism, and schizophrenia. Wallace concluded that mitochondrial energetics and associated high energy mitochondrial intermediates are the mediators between environmental energy availability and demands and the genome. If energetics is in balance with the cellular and environmental demands then this is health. However, if there is a chronic energy deficit this leads to disease and ultimately death. Symposia Sessions Genetic Architecture Insights from Joint Investigators of Rare CNVs and Common SNPs (reported by Sarah Gagliano and Kirti Mittal) All of the speakers in this symposium were female, which is inspirational. Dr. Lea Davis (University of Chicago, USA) presented her work testing the hypothesis that an individual may develop disease by surpassing either a polygenic or a variant liability threshold. Given this hypothesis, one would expect there to be a negative correlation between the polygenic burden (genomic risk scores) and rare variant burden (genic CNVs >500kb in less than 1% of samples) among cases. For proof of principle, Type 1 Diabetes, which has a known risk locus in the HLA region with large effects, was examined. She then presented results from three psychiatric disorders, Tourette syndrome, OCD, and autism spectrum disorders (ASD). Results showed a modest but significant negative association among cases between scores and rare CNVs in analyses for the childhood-onset disorders (Tourette syndrome and ASD), but not for OCD. Ms. Lingxue Zhu (Carnegie Mellon University, USA) noted that although each common variation tends to have smaller effects than rare variation, the former accounts for a large portion of liability (50%). She presented two models for predicting ASD risk from common SNPs: Genomic-BLUP (G-BLUP) and linear mixed model (LMM). G-BLUP is a random effects model assuming all small effects. LMM measures fixed effects. To select SNPs that have large fixed effects, weighted lasso was applied, resulting in 50, 250, or 1,100 SNPs to include into the LMM. The G-BLUP model (LMM with no SNPs having fixed effects) performed best (area under the curve = 0.74). When additional fixed effects were included, accuracy decreased. Ms. Zhu presented her work on a related trait, head circumference deviation. Those predictions were more accurate when parental head size was included, but common variants did not add much. Ms. Niamh Mullins (King's College London, UK) presented her work done at deCODE Genetics (Iceland), investigating selection pressures on genetic variants for psychiatric disorders in the general Icelandic population. PRS for five psychiatric disorders were used to predict fecundity (number of children), using linear mixed effects models. PRS for autism was associated with reduced fecundity in the population, excluding patients. This effect was specific to males. PRS for the other disorders were not significantly associated with fecundity. Neuropsychiatric CNVs implicated in autism and schizophrenia, were associated with reduced fecundity, with larger effects in males. The results from this population suggest that, with the exception of autism, selection pressures may operate on some but not all components of the genetic architecture of psychiatric disorders. Dr. Sarah Bergen (Karolinska Institute, Sweden) discussed the contribution of CNVs and SNPs to schizophrenia risk in the Psychiatric Genomics Consortium samples. Polygenic scores were compared for carriers and non-carriers of implicated CNV risk loci (individually and in aggregate), large CNV deletions, and in terms of total genomic CNV burden. Cases with implicated CNVs and large deletions had lower polygenic scores than other cases, and an inverse relationship with total CNV burden was also significant. These relationships were not observed in controls. These results converged to broadly support a liability threshold model of genetic risk for schizophrenia. The session finished with a discussion led by Dr. Naomi Wray (University of Queensland, Australia) who concluded that despite limited power, the results from the speakers suggested that PRS do tend to be lower for individuals who carry rare CNVs of large effect. If one has rare CNVs, then there is a lower threshold of polygenic risk disease burden, and it also seems that such CNVs decrease fecundity. The Genetic Dissection of Bipolar Disorder: From Common to Rare Risk Variation (reported by Niamh O'Brien) Dr. John Kelsoe (University College San Diego, USA) reported the findings from the Psychiatric Genomics Consortium Bipolar Disorder (PGC2-BIP32) genome wide association analysis. The case-control sample for this study consists of 20,352 BD cases and 31,358 controls. The analysis identified 19 BD associated loci, 12 of which are novel and provides refinement of known BD associated loci such as TRANK1 (p=5.54×10-14) (Chen et al., 2013). Subphenotype analysis identified six new genes associated with BD-I and three new genes for a combined analysis of BD-II and schizoaffective disorder. A z-score mixture model suggested that BD is more polygenic than schizophrenia. Data-driven Expression-Prioritized Integration for Complex Traits (DEPICT) pathway analysis implicated brain-related pathways including the calcium and potassium ion transporters and glutamatergic signalling in the pathophysiology of BD (Pers et al., 2015). Dr. Tadafumi Kato (Riken Brain Science Institute, Japan) reported on sequencing analysis looking at de novo point mutations in BD. The study focused on 79 probands with BD. Seventy de novo point mutations were found, 64 single nucleotide variants (SNVs) and 6 insertion/deletions (indels). Global enrichment analysis showed an enrichment of de novo loss of function and protein-altering mutations in individuals with BD-I and schizoaffective disorder. BD probands with protein-altering de novo changes showed significantly earlier age-of-onset. Genes hit by de novo or protein altering variants are significantly enriched for intolerant genes. Intolerant genes are depleted for protein-altering mutations as determined by a Residual Variation Intolerance Score (RVIS) (Petrovski et al., 2013). A gene encoding microtubule-actin crosslinking factor 1 (MACF1) is the most intolerant gene reported in this analysis and is hit by a frameshift variant. Dr. Peter Zandi (Johns Hopkins, Bloomberg School of Public Health, USA) reported on data from the Bipolar Sequencing Consortium (BSC). The goal of this study is to identify rare genetic variants that influence the risk of BD. The founding cohorts consist of 4,733 BD cases and 9,246 controls. The preliminary analysis consists of 3,633 BD cases and 4,992 controls. Dr. Zandi reports that the MAF did not differ across study groups despite the different platforms used for exome sequencing. A gene-wise burden test showed 10,043 genes with disruptive variants, 5,050 of these genes harbour less than one variant. Neither the gene-wise burden test nor single variant analysis showed significant results. Dr. Seth Ament (Institute for Systems Biology, USA) reported on family data from the BSC. The study consisted of a uniform analysis pipeline with ANNOtate VARiation (ANNOVAR) (Wang et al., 2010); focused on protein altering variants that are present in two or more affected individuals and have a MAF of less than 0.01 in the 1,000 genome project. Dr. Ament reported 143 pedigrees from 652 pedigrees that contained 526 loss of function SNVs and 11,856 rare coding SNVs. The top ontology enriched pathways for rare coding variants in the BSC pedigrees highlighted pathways different to those previously reported such as DNA binding and DNA strand elongation. There was an excess of genes in which a rare SNV segregates with BD in multiple pedigrees, such as two loss of function variant in the gamma-aminobutyric acid A receptor, alpha 6 gene (GABRA6). Currently there are no genome wide significant hits but aggregation of individual level data and case control cohorts will help elucidate the effects of rare variants in family study of BD. Current Approaches to Genetic/Genomic Studies on Alcoholism (reported by Caroline Camilo and Bhagya Shankarappa) Dr. Dayne Mayfield (University of Texas at Austin, USA) was the chair of this session and introduced the current approaches to genetic studies on alcoholism. Dr. Howard Edenberg (Indiana University School of Medicine, USA) began the session describing the complex trait of alcoholism, which is likely caused by a combination of multiple genes and environmental factors. He stated that alcoholism runs in families and has a high rate of psychiatric comorbidity. He reported the importance of identifying genetic and epigenetic modifications that may contribute to the risk of alcoholism. Dr. Edenberg showed that common and rare variants require different strategies to investigate the risk of disease given that common variants tend to have small effects on risk whereas rare variants have larger effects. He cited several approaches such as family-based GWAS, exome sequencing of rare variants, genomic studies of lymphoblastoid cell lines (LCLs), iPSCs, and brain tissue in addition to epigenetics and prospective studies of adolescents to identify genes that may contribute to the risk of developing alcoholism and the important interactions between phenotype, environment, and genetics in alcoholism. Dr. Sean Farris (University of Texas at Austin, USA) reported his study on the neurogenomic networks that are involved in alcohol use disorder (AUD). He showed a variant-driven gene network with strong interaction between the genes in the human prefrontal cortex. He presented his data on the gene network for lifetime alcohol consumption and dysregulation of gene expression including epigenomics networks. He discussed that datasets continue to grow in size and complexity. He also stated that gene networks support disease-gene associations and show system-wide perturbations related to alcohol dependence. Dr. Farris concluded by stating that there is converging evidence for multiple candidate genes and epigenetics involvement implicated in alcohol dependence. Dr. Subhash Pandey (University of Illinois at Chicago, USA) spoke about adolescent alcohol exposure and epigenetic mechanisms, explaining the interaction between neurobiological and behavioral changes in addition to epigenetic factors (i.e., histone and DNA modifications) in adolescence with alcohol consumption. Evidence suggests that these changes can alter gene expression. He presented his study on adolescent intermittent ethanol (AIE) exposure paradigm in an alcohol binge-drinking model in rats. Particularly, Dr. Pandey and his colleagues investigated brain-derived neurotrophic factor (BDNF) gene expression and also examined histone acetylation (H3K9&14) of BDNF exons I and IV promoter regions in the amygdala of adolescent intermittent saline (AIS) and AIE rats in adulthood. They found a decrease in BDNF gene expression in the amygdala after AIE in adulthood. This appears to be due to AIE-induced decrease in histone acetylation of BDNF in the amygdala in adulthood. He also discussed about the effect of AIE on changes in expression of the lysine specific demethylase 1 (LSD1) and neuron specific LSD1+8a enzymes. The expression of LSD1 and LSD1+8a were decreased in the amygdala of AIE compared to AIS in adulthood, which in turn increases the methylation of histone H3K9 dimethylation (me2) without producing any change in the levels of H3K4me2, leading to increased anxiety and alcohol consumption in adulthood. Dr. Shizhong Han (University of Iowa, USA) discussed the importance of GWAS in AUD, describing the polygenic nature of AUD. He presented his data on the integrated GWAS and protein-protein interaction network analysis in AUD. He utilized the GWAS of AUD and tissue-specific gene expression data to examine the relationship of AUD risk genes in brain and non-brain tissues. The results showed that AUD risk genes are highly connected in brain regions, but not in other non-brain tissues. Furthermore, he spoke about his approaches of constructing a brain-specific network for gene prioritization. He summarized his presentation by discussing that the nominally significant findings of genes are functionally related in human brain tissues, and form networks that underline relevant biological mechanism. One example is the suppressor of cytokine signaling 6 (SOCS6) gene, which plays an important role in AUD. Altered gene expression and increased cytokines have been reported in human postmortem AUD brain tissues. He concluded that brain-specific gene networks may help to prioritize AUD risk genes for future studies. Dr. Abbas Parsian (National Institute of Health/NIAA, USA) closed the session by summarizing the results and discussing the interactions between the genetics and environmental factors in alcoholism. Tracking the Descent to Mental Illness – Insights into the Trajectory to Illness from Studies of Youth at High Risk of Bipolar Disorder (reported by Søren Dinesen Østergaard) Improving the possibilities for early identification of mental disorder has been a priority in psychiatry for many years (Akiskal et al., 1983; Goldberg et al., 2001; Ostergaard et al., 2014). However, early detection of mental disorders remains challenging due to the absence of strong biological and psychopathological predictors. This symposium focused on initiatives aiming at identifying such predictors based on studies of high-risk individuals. Dr. Uher (Dalhousie University, Canada) showed preliminary results from the Families Overcoming Risks and Building Opportunities for Well-being (FORBOW) study (Uher et al., 2014) in which children of parents with severe mental illness are recruited and followed over time. In the FORBOW study, the cohort members and their parents undergo detailed structured interviews. The preliminary results indicated that there is a same sex-specific parent of origin effect in anxiety (i.e., mood disorders in mothers predict anxiety in daughters), while there is an opposite sex-specific parent of origin effect in psychosis (i.e., severe mental illness in mothers predict psychosis in sons). Furthermore, the results indicate that early psychopathological antecedents are associated with later development of severe mental illness. Dr. John I. Nurnberger (Indiana University School of Medicine) showed results from the Bipolar High Risk Study Group and the Bipolar Disorder Genome Study (BiGS) (Nurnberger et al., 2011; Monahan, Stump et al., 2015). From the prediction perspective, the key findings were that anxiety and externalizing disorders predicted development of major affective disorder (BD or recurrent major depression) in individuals at high-risk. Dr. Janice M. Fullerton (Neuroscience Research Australia, Australia) described results from analyses of neuroimaging and genetic data from the Bipolar Kids & Sibs Study. This initiative recruited young individuals with BD or with a first-degree relative with BD. Based on the magnetic resonance imaging (MRI) data, it was demonstrated that individuals with high familial risk for BD have reduced interhemispheric connectivity. Furthermore, these individuals also have a higher genetic load for BD (as quantified by PRS) when compared to controls (Fullerton et al., 2015). Dr. Andrew M. McIntosh (University of Edinburgh, UK) presented outcomes from the Scottish Bipolar Family Study. Individuals with first-degree family history of BD and healthy controls were recruited for a structured psychiatric interview and MRI scanning at baseline and follow-up. The results indicated that increased activation of the insula cortex at study baseline was associated with an increased risk of developing major depression during follow-up (Whalley et al., 2015). Dr. Philip Mitchell (University of New South Wales, Australia)raised the issue of potential lack of power in the individual studies described above, and suggested forming a consortium, which can utilize a meta-analytical approach with the gathered data to predict the risk of BD in high-risk individuals. Genetics of Research Domain Criteria (RDoC) (reported by Tristram Lett) Dr. Paul Arnold (University of Calgary, Canada) and Stephen Glatt (SUNY Upstate Medical University, USA) introduced the session and key issues on the next steps in genetic research on RDoC domains and constructs. Dr. Sarah Morris (NIMH, USA) gave a brief overview of RDoC including: (a) the RDoC initiative as a National Institute of Mental Health (NIMH)-led effort to change how patients (and non-patients) are characterized for research purposes, and (b) the RDoC framework for classifying subjects to neurobehavioral constructs based on our understanding of brain and behaviour. She stated that these homogenous subgroups potentially capture more subthreshold (subclinical) individuals based on DSM-5 or ICD-10 diagnostic categories alone. Furthermore, that RDoC is a dynamic framework that will evolve with new research. Dr. Joan Kaufman (Kennedy Krieger Institute/Johns Hopkins, USA) described the genetics of childhood trauma related to psychiatric disorders. In an ongoing study of 400 maltreated children of which 125 subjects had undergone functional magnetic resonance imaging (fMRI), dimensional measures of child maltreatment predicted hippocampal activation and functional connectivity to regions involved in fear response. Moreover, the effect of trauma on hippocampal sensitivity decreased with social support. These studies demonstrated the advantages of the RDoC framework by identifying an interacting stress by social support mechanism on clinical intermediate phenotypes in a high risk group with diverse psychiatric outcome. Dr. Paul Arnold discussed dimensionality and heritability of OCD in a community-based study of 16,718 children (6-18 years) collected at the Ontario Science Centre in Canada. The children were administered the Toronto Obsessive Compulsive Scale. In a subset of 220 twin pairs, a consecutive heritability analysis was undertaken. He reported a high heritability of obsessive-compulsive dimensions varying between 30-77%. The results of this study applying a dimensional approach supported the use of RDoC in OCD patients. Dr. Yanli Zhang-James (SUNY Upstate Medical University, USA) reviewed four types of genetics studies of aggression including human twin and GWAS studies, rodent knock-out models and candidate genes, rare genetic disorders with antisocial/aggressive behavior from the Online Mendelian Inheritance in Man database (OMIM), and transcriptomics of rodent models. Among OMIM genes with antisocial behavior, nominal GWAS findings, rodent knock-out models, and aggresso-type candidate genes, several common pathways regulating synaptic transmission emerged including serotonergic, dopaminergic, and GABAergic signaling. There was further evidence implicating mitochondrial dysfunction and MAPK (mitogen-activated protein kinases, originally called ERK, extracellular signal-regulated kinases) signaling. Dr. Kristin Nicodemus (University of Edinburgh, UK) focused on the RDoC language construct. She used latent semantic analysis (LSA) to derive variables in free speech data in individuals at high-risk for psychosis. Semantic coherence, phrase length, and use of determiners was 100% accurate at predicting transition to psychosis. In a subsequent candidate language gene study of schizophrenia patients, healthy siblings and controls, the disrupted in schizophrenia (DISC1) rs12133766 variant was associated with vector length; however, this association was not observed using standard measures of verbal fluency. She concluded that using this RDoC framework for a broader definition of language can provide novel understanding of the genetic and neurobiological mechanisms of language dysfunction. Genetics of Comorbidity between Substance Use Disorders and Other Severe Mental Illness (reported by Jennie Pouget) Many patients with mental illness suffer from more than one disease, and substance use disorders are particularly prevalent comorbidities. The underlying reasons for substance use comorbidities are not clear. In the genomic era, we are reaching a point where we can articulate hypotheses about comorbidity across psychiatric disorders and test them with reasonable statistical power. Dr. Nelson Freimer (UCLA, USA) gave an overview of a large study comprising pedigrees ascertained for severe bipolar disorder from founder populations of Colombia and Costa Rica. These pedigrees have provided insights into the genetic relationships between bipolar disorder and cognitive and neuroimaging endophenotypes, identifying 53 heritable endophenotypes associated with bipolar disorder including cortical thickness in prefrontal and temporal regions (Fears et al., 2014). Currently, these families are being revisited for detailed phenotyping of substance use disorders, which will help uncover genetic factors underlying substance use comorbidities in bipolar disorder. Dr. Sarah Hartz (Washington University in Saint Louis, USA) presented genetic data evaluating the comorbidity between nicotine dependence and schizophrenia. Dr. Hartz identified 16 genetic variants previously associated with schizophrenia that were also associated with nicotine dependence (p < 0.05) in a recent GWAS of 17,074 ever smokers (Hancock et al., 2015). Most notable was rs8042374, an intronic variant of the gene encoding the neuronal nicotinic acetylcholine receptor α3 subunit (CHRNA3), which is the first variant to reach genome-wide significance in two psychiatric disorders. Dr. Kerry Ressler (Emory University, USA) provided a thought-provoking overview of insights obtained from a sample of highly traumatized patients ascertained in the inner city of Atlanta (Khoury et al., 2010). In this cohort, the level of substance use strongly correlated with childhood abuse and current PTSD symptoms. Accumulating evidence suggests that the neuro-circuitry of addiction and PTSD may be shared, with communication between the amygdala and cortex playing an important role in both of these disorders. One salient example is variant rs1433375 in the gene encoding sodium channel and clathrin linker 1 (SCLT1), which was associated with comorbid alcohol use (measured by the by the Alcohol Use Disorders Identification Test) in this highly traumatized cohort. SLCT1 is highly expressed in the cerebellum, and carriers of the A risk allele for rs1433375 showed less dorsolateral prefrontal cortex connectivity to the cerebellum than patients with the G allele in a follow-up imaging study. As discussant, Dr. Patrick Sullivan (UNC Chapel Hill, USA) emphasized that the dissection of psychiatric comorbidity – including substance use disorders – may be the most important emerging problem in psychiatric research because it has been largely neglected up until this point. He challenged the field to focus on this issue, with a particular emphasis on the utility of prospective longitudinal studies. Oral Sessions Schizophrenia (reported by Chenyao Wang) Mr. Jonathan Hess (SUNY Upstate Medical University, USA) reported that they have succeeded in providing a framework by which to integrate single-nucleotide polymorphisms emerging from GWAS with multi-omic datasets. There is a critical gap in our understanding of the functional consequences of psychiatric disorder-associated variants in context of gene-expression regulation. Particular splicing-factor motifs were altered by schizophrenia- or bipolar disorder-associated variants more often than expected by chance, in genes such as CUG triplet repeat, RNA binding protein (CUGBP), elav-like family members 1 and 4 (CELF1 and CELF4) for schizophrenia, and epithelial splicing regulatory protein 1 (ERSP1), and serine/arginine-rich splicing factor 5 (SRSF5) for BD. Their research team implicated several risk variants in abnormal splice site binding with predictive methods, and linked these observations to gene expression levels in brain tissue. Dr. Pippa Thomson (Institute of Genetics and Molecular Medicine, UK) presented results from their clinical and genetic re-evaluation of the Scottish t(1:11) family in which a translocation disrupts DISC1 and the DISC1 fusion partner 1 (DISC1FP1) gene. The t(1;11) family presented with a broad spectrum of psychiatric diagnoses including schizophrenia, BD and recurrent MDD. Genome-wide significant linkage to major psychiatric illness was identified between broad peaks across both translocation breakpoints; with a LOD (logarithm [base 10] of odds) score of 6.1 for translocation status. Additional linkage peaks with LOD scores greater than 3 were identified on chromosomes 3q and 5q. PRS derived from the PGC schizophrenia and BD GWASs also predicted illness within the family. These results confirm the linkage of the translocation with major mental illness in this family and identify additional loci which may explain the variable presentation of illness. Dr. George Kirov (Cardiff University, UK) clarified the role of maternal and paternal duplications at 15q11-q13 in neuropsychiatric disorders. Maternal duplications are highly pathogenic, resulting in neurodevelopmental disorders in around 75% of carriers. Individuals with paternal duplications have an increased risk of developing autistic spectrum disorder, developmental delay, or multiple congenital anomalies, but not schizophrenia. About 60% of duplications are de novo. Despite their lower pathogenicity, paternal duplications are less frequent in the general population, possibly due to reduced fecundity of carriers and survival of embryos. Dr. Douglas Ruderfer (Icahn School of Medicine at Mount Sinai, USA) clarified CNVs in intolerant genes would be more likely to have deleterious effects using a large sample and an empirical approach they calculated frequency and tolerability of CNV at the gene level. While directly using the Exome Aggregation Consortium (ExAC) CNV data as a convenience control sample runs a high risk of bias, they demonstrated improved power to detect schizophrenia loci when considered along with an appropriate matched control sample. Dr. Menachem Fromer (Icahn School of Medicine at Mount Sinai, USA) presented RNA-seq data of dorsolateral prefrontal cortex and anterior cingulate cortex from post-mortem brain of schizophrenia patients and controls. They overlaid the resulting expression quantitative trait loci with the 108 common variant loci associated with schizophrenia (Schizophrenia Working Group of the Psychiatric Genomics Consortium, 2014) and found significant overlaps in the genes between the two datasets. Dr. Evangelos Vassos (King's College London, UK) estimated the predictive power of PRS in discriminating case-control status in first episode psychosis and to predict the development of schizophrenia as opposed to other psychoses. PRS was a powerful predictor of case-control status in Europeans, even though half of the cases did not have an established diagnosis of schizophrenia at the time of assessment. The PRS also showed some ability to distinguish between those first-episode psychosis cases who developed schizophrenia from those who did not. Advances in Autism (reported by Megan Crow) Dr. Jakob Grove (Aarhus University, Denmark) presented results from the Lundbeck Foundation Initiative for Integrative Psychiatric Research (iPSYCH) Autistic Spectrum Disorder (ASD) GWAS, focusing only on the strict European cluster (11,661 ASD and 21,427 controls) combined with data from the PGC-ASD GWAS. Six loci with genome-wide significance were found and LD score regression against the PGC-ASD GWAS showed significant genetic correlation (∼76%, p=2.9×10-13). An analysis of the ASD results at the PGC-schizophrenia loci showed that 96/128 indices had the same direction of effect in ASD as in schizophrenia (p<5×10-8). LD score regression provided evidence for widespread overlap between ASD and schizophrenia (∼23%, p=2.8×10-6), and a positive genetic correlation with educational attainment (∼20%) and childhood intelligence (∼30%). Mr. Jack Kosmicki (Harvard University, USA) presented his work studying ASD-related de novo variants using the ExAC database. Mr. Kosmicki found that approximately one third of previously identified ASD-related de novo single-nucleotide variants were present in other individuals in ExAC. De novo protein truncating variants (PTVs) absent from ExAC (‘non-ExAC’) and those in likely haplo-insufficient genes were enriched in cases (odds ratio [OR]=1.98 for all non-ExACde novo PTVs, OR=3.4 for non-ExAC likely haplo-insufficient de novo PTVs), and the non-ExAC de novo PTV rate predicted intelligence quotient (IQ) (p=5×10-4). Similar trends were observed for inherited PTVs (OR=1.4 for likely haplo-insufficient non-ExAC variants). In ASD cases, a reduced 3:1 male:female bias in de novo rate was observed with non-ExAC likely haplo-insufficient de novo variants, whereas a 6:1 male:female bias was observed with all other de novo PTVs, indicating that females are more likely to have rare de novo PTVs in putative haplo-insufficient genes. Dr. Elise Robinson (Massachusetts General Hospital, USA) presented an analysis of the heritability of continuous social and communication traits using data from iPSYCH-ASD, PGC-ASD, the Avon Longitudinal Study of Parents and Children (ALSPAC), the Simons Simplex Collection (SSC) and ExAC. Using LD score regression Robinson found that ∼25% of ASD common variation (with PGC-ASD and iPSYCH-ASD being considered separately) is shared with common variation that influences the social and communication disorders checklist in the ALSPAC cohort. This was also the case for de novo variants in that the rate of non-ExAC de novo loss of function and predicted damaging missense variants in the SSC cohort linearly predicted impairment measured by the Vineland Scales of Adaptive Behavior (p<0.01 for both cases and controls). Dr. Janita Bralten, PhD (Radboud University, Netherlands) presented the results of a GWAS of autistic traits in the general population. Dr. Bralten validated a self-report questionnaire, then tested the association between genotypes in an ASD candidate gene set (146 genes) and trait scores across 5 sub-categories in a population sample (n=1981). An association was observed between “rigidity” and ASD candidates in a competitive gene set analysis test (p=0.005), which was primarily driven by genes associated with “neurite outgrowth” (p=0.003); a SNP in the MET proto-oncogene, receptor tyrosine kinase (MET) gene was statistically significant after controlling for the family-wise error rate (p=1.4×10-4). Dr. Ryan Yuen (The Hospital for Sick Children, Canada) presented his work studying de novo variation in 200 ASD simplex families and 258 control families. Yuen found that ∼70% of single-nucleotide variants and insertion-deletions were paternally derived, and that the number of de novo variants correlated with paternal age. The somatic mutation rate was 3.6 per genome in ASD, and the sequence context of these mutations differed from germline mutations. Damaging variants were enriched in cases, and ASD de novo variants were enriched for functions related to synaptic transmission, chromatin modification and translation. Dr. Katri Kantojärvi (National Institute for Health and Welfare, Finland) presented an association study on nine previously identified psychiatric-related calcium channel, voltage-dependent, L type, alpha 1C subunit (CACNA1C) SNPs in infant sleep regulation using a cohort of 1,017 Finnish eight-month-old babies. Four SNPs were associated with parent-reported sleep latency overall (p<0.05), and some sex differences were observed. In a subset of 60 babies, an association was found between one SNP and three polysomnographically measured sleep traits (p<0.05). Neuroimaging and Alternate Phenotypes (reported by Sejal Patel) Dr. Derrek Hibar (University of Southern California, USA) discussed the use of brain imaging data from the Enhancing NeuroImaging Genetics through Meta-Analysis (ENIGMA) consortium and investigating potential intermediate phenotypes for psychiatric disorders such as OCD. Genetic variations from genome wide association studies under the international collaboration, ENIGMA-OCD working group was examined in eight substructure brain volumes. There was no evidence of pleiotropy; however, in the test for concordance, there was significant association between genetics variants and an increase in nucleus accumbens and putamen volumes in addition to an increase risk for OCD. Dr. Ida Sønderby (Norwegian Centre for Mental Disorders Research, Oslo) presented the ENIGMA-CNV project that aims to associate CNVs with brain imaging phenotypes. Approximately 12,000 individuals from 16 cohorts worldwide with both genetics and neuroimaging data have been collected for analysis. Preliminary analysis on a specific CNV supports previous findings. More cohorts were encouraged to join. Dr. Arash Nazeri (Centre for Addiction and Mental Health, Canada) presented a genome wide interaction study, investigating interaction effects between genetic variants and serum urate on striatal dopamine transporter density binding ratio (as indexed by DaTscan striatal binding ratio) in people with Parkinson's disease. The interaction of gene variant and serum urate on MRI-derived regional brain volumes (voxel-based morphometry) and clinical status were also investigated. The inositol polyphosphate 5-phosphatase K (INPP5K) rs1109303 (T>G) variant showed a significant interaction effect on striatal dopamine transporter density with the association between serum-urate level and striatal dopamine transporter being positive in G-allele carriers and negative in TT genotype carriers. Similar interaction effect was observed on prefrontal cortex volume and clinical severity of Parkinson's disease. In conclusion, INPP5K rs1109303 genotype could inform pharmaco-therapeutic approaches targeting urate pathway in Parkinson's disease. Dr. Daniel Felsky (Centre for Addiction and Mental Health, Canada) presented the interaction between the sortilin-like receptor (SORL1) gene and BDNF. Post-mortem brains were used to quantify 13 SORL1 transcripts isoforms. The T allele of the rs12364988 on the transcript isoform SORL1-005 reduced the expression of SORL1 in the BDNF Val/Val homozygotes and increased the expression of SORL1 in BDNF Met carriers. This result demonstrated a novel interaction between SORL1 and BDNF, which may play a role in SORL1 alternative splicing. Dr. Danielle Posthuma (Vrije Universiteit University, Netherlands) discussed the importance of gene and environment interaction on psychiatric traits. A meta-analysis was done on 3,306,594 twin pairs investigating the genetic and environmental factors in 47 distinct psychiatric traits. The heritability of all psychiatric traits was 0.462, with a higher heritability (0.518) in young age group (0-11) when compared to other age groups. Results demonstrated that most psychiatric traits are influenced by additive genetic variance. Dr. Margaret Maciukiewicz (Centre for Addiction and Mental Health, Canada) presented her study on MDD in relation to response to a serotonin-norepinephrine reuptake inhibitor (SNRI), duloxetine. Nine gene variants (imputed and genotyped) were selected for Lasso regression. In support vector machine (SVM) models, the accuracy was 61.75%. When non-genetic predictors were added to the model, the accuracy increased to 80.29% in SVM but further refinement is needed for clinical settings. Tuesday October 20, 2015 Plenary Session Dopamine, Schizophrenia, and the Process of Discovery in the Brain Sciences (reported by Jingjing Zhao) Professor Arvid Carlsson's discovery of dopamine as an important neurotransmitter has contributed considerably toward the genetics and neurobiology of various diseases as well as drug discovery and treatments to clinical patients. This plenary session started with watching a recorded video interview of Professor Carlsson, followed by a live skype call with Professor Carlsson. In the video interview, Professor Carlsson first described his scientific career development in the 1950s and his early experiments that led to his discovery of dopamine as an important neurotransmitter. He commented on the challenges for himself to choose a direction that was different from his supervisor and appreciated that his supervisor did not oppose him to proceed on his own direction even though it challenged the scientific opinions of the time. He spoke about his experience when he was notified that he won the Nobel Prize in 2000. Surprisingly, the first question that he asked when he received the phone call from the Nobel notifier was: “How do you formulate the reason to give me the prize?” Professor Carlsson especially pointed out the importance of considering the negative and side effects of long-term treatment in drug development and suggested that new medications should be careful in stabilization, balancing the “brake” and “accelerator”, and should keep the plasticity of brain at an optimum level. Regarding the recent financial cut-back in Europe for neuroimaging studies, Professor Carlsson was in favour of studying the brain as a promising direction and believed that a lack of harmony of the brain were coupled with many diseases. Finally, Professor Carlsson summarized the challenges for research without hypothesis such as the GWAS of schizophrenia and commented on the disadvantages of the current classification of disorders. In the skype call, Professor Carlsson answered questions from the audience with various backgrounds. Professor Carlsson provided advice to both young researchers and old scientists as to how to proceed in the field respectively. He suggested young researchers to start with a simple project to gathers better motivation to do research. For senior scientists, he rather encouraged them to fulfil their early scientific dreams that they did not have the chance to reach in their early academic career. Professor Carlsson answered a question from Dr Chunyu Liu (University of Illinois, USA) about new types of neurotransmitters other than dopamine and agreed that compounds having signaling properties may all have important functions in diseases. Professor Carlsson also commented on the release of dopamine, a question laid out by Professor Robin Murray (King's College London, UK). He believes that both pre- and post-synaptic components of dopaminergic transmission play a role in schizophrenia. In the skype meeting, Professor Carlsson emphasized again the importance of balanced functions of a new drug and highlighted that it would be a mistake for not taking side effect into account when inventing new drugs given how vulnerable the brain is and how important plasticity of brain would play a role at early stage of life and for the entire life. Finally, Professor Carlsson completed his skype call by answering a question about the opportunity of female scientists raised by a female postdoctoral researcher from the University of Cambridge. Professor Carlsson acknowledged the tremendous role of women in his academic career. He admitted that although a lot of development and progress for providing equal opportunity to female scientists have achieved as decades passed, the final goal was still not reached and females still did not have the same opportunities to the top positions as males. Oral Sessions Dissecting the Schizophrenia Phenotype (reported by Umut Kirli) Dr. Daniel Howrigan (Massachusetts General Hospital, Boston, USA) discussed the contribution of de novo coding mutations to schizophrenia risk. He presented findings from analysis of exome sequencing data on 1,697 schizophrenia trios. While an emerging pattern of de novo risk is evident among well-characterized gene sets and an excess of genes with recurrent damaging mutations, the increased liability toward schizophrenia due to de novo mutations comprises only a modest fraction of the overall genetic liability and to date no single gene has been established as a putative de novo schizophrenia risk factor. Dr. Tristram Lett (Charite University Hospital, Berlin, Germany) presented a study investigating the influence of the functional rs3749034 variant in the glutamic acid decarboxylase 1 (GAD1) gene on brain structure and working memory performance in schizophrenia patients and healthy controls. The effect of this variant on long-interval cortical inhibition (LICI) in the dorsolateral prefrontal cortex (DLPFC) was subsequently examined using TMS with electroencephalogram (TMS-EEG). He discussed the findings indicating that genetic variation in GAD1 may affect white matter fractional anisotropy, GABAergic inhibitory neurotransmission in the DLPFC and working memory performance. Dr. Alexander Richards (Cardiff University, UK) presented preliminary data from EU-GEI (EUropean network of national schizophrenia networks studying Gene-Environment Interactions), a cohort of ultrahigh risk and frank psychosis cases in UK, Netherlands, Italy, France, Turkey, Spain, Serbia, Ireland and Brazil. The research is focusing on non- affective psychosis (not only schizophrenia); cognitive scales, social and environmental risk variables are available to examine interactions with genetic risk. Mr. Ahmed Al Amri (University of Leeds, UK) presented an autozygosity mapping in combination with whole-exome sequencing study conducted in a first-cousin consanguineous family, in which two out of eight siblings were affected with psychosis. He reported a missense mutation, c.C1348T:p.R450C, in the deafness, autosomal recessive 31 (DFNB31) gene at 9q32, which was predicted by all mutation prediction packages to be pathogenic and co-segregated with psychosis in the family in a manner consistent with recessive inheritance. This variant was suggested to impair the interaction of DFNB31 with UBR4 (ubiquitin protein ligase E3 component N-recognin 4), which is known to have roles in neurogenesis, neuronal migration and neuronal signaling. Dr. Giulio Genovese (Broad Institute, Cambridge, USA) presented a schizophrenia case-control cohort investigating rare disruptive mutations in constrained genes (that harbor the expected amount of synonymous variations but significantly under-represented missense variations). He reported that overall 24% of schizophrenia cases (and just 17% of controls) harbored private disruptive mutations in the most constrained genes. Dr. Emma Dempster (University of Exeter, UK) presented a study examining the role of epigenetic variation in schizophrenia, focusing on DNA methylation differences in disease-discordant MZ twins. She reported that the most significant differentially methylated position was located in the histone deacetylase 4 (HDAC4) gene, encoding a histone deacetylase implicated in synaptic plasticity and memory formation and a differentially methylated region (DMR) was identified in the HLA region which had been implicated in previous GWASs of schizophrenia. Biostatistics and Bioinformatics (reported by Kartikay Chadha) Dr. Megan Crow (Cold Spring Harbor Laboratory, USA) presented her research exploring cell-type specific co-expression of genes with recurrent loss-of-function de novo mutations in ASD. Dr. Crow built co-expression networks for six genetically targeted adult mouse inhibitory interneuron types and analyzed their functional connectivity using a neighbor voting algorithm in cross-validation. This enabled her to conclude that ASD candidate genes are strongly co-expressed in inhibitory interneuron networks, with further investigation indicating that this is primarily driven by high expression of these genes. Dr. Raymond Walters (Massachusetts General Hospital/Broad Institute, USA) suggested a hypothesis that “GWAS of continuous traits in population samples can be used to improve power to detect the loci for psychiatric phenotypes”. Dr. Walters and his team demonstrated efficient power enrichment of transforming dichotomous phenotypes to continuous latent liability variables, and the effect of genetic covariance on the relationship between the latent liability variables and the continuous phenotypes by varying genetic architectures through simulation studies before applying the proposed approach to studies of ADHD with the EArly Genetics & Lifecourse Epidemiology (EAGLE) Consortium and the PGC. Mr. Christaan de Leeuw (Vrije Universiteit Amsterdam, the Netherlands) presented his work to investigate the self-contained and competitive gene-set analysis methods of the GWAS data. The simulation studies showed a high false-reporting rate for the self-contained approach for the analysis of a polygenic phenotype, particularly in large gene sets and increasing sample sizes. Christaan concluded that self-contained analysis doesn't provide reliable results, and the alternative competitive methods may have biases as well. He added, “obtaining higher statistical power is difficult for strongly heritable traits, and that power doesn't improve significantly with increasing sample size”. Dr. Verneri Anttila (Massachusetts General Hospital/Broad Institute, USA) spoke about his research on a joint analysis of 23 brain diseases to reveal novel patterns in the genetic background of psychiatric and neurological diseases via a cross-disorder heritability analysis, using the LD score regression approach for all GWAS data. His research showed a general trend in psychiatric diseases to have considerable risk-increasing co-morbidity with a variety of other psychiatric diseases, notably with schizophrenia and major depression, showing considerable co-morbidity with most of the studied psychiatric phenotypes. Dr. Sarah Gagliano (Centre for Addiction and Mental Health, Canada) presented her research of prioritizing genetic risk variants for psychiatric disorders based on functional genomic information using a machine learning approach. She trained an elastic net model using 14 different functional annotations including splice sites, nonsynonymous SNPs, and DNase I hypersensitive sites. The data was divided into training and test sets, and the resulting model had reasonable accuracy (with area under the receiver operating characteristic curve of around 0.7). She then presented a comparison of statistical learning methods using different combinations of three previously published annotation sets with three algorithms (Gagliano et al., 2015). Dr. Wim Verleyen (Cold Spring Harbor Laboratory, USA) introduced the audience to a tool for customized network analysis called SAPLING (sapling.cshl.edu). SAPLING is a web application which utilizes heterogeneous data resources for in-depth analysis; existing tools, for example, GENEMANIA (Warde-Farley et al., 2010) and DAPPLE (Rossin et al., 2011), lack these properties. He reported examples of using SAPLING in the context of psychiatric genetics (autism, synaptic interactions, and A) where the downstream analysis was customized with data and algorithms using the tool to produce results showing that aggregation across more network data and brain-related data improves performance while condition-specificity within the underlying data appeared to be difficult. He concluded that customized network analysis might be needed to handle functional interpretation of gene lists related to psychiatric disorders. Pharmacogenetics of Response and Side Effects (reported by Ellen Ovenden) Dr. Douglas Ruderfer (Mount Sinai School of Medicine, USA) opened the session by discussing his research on the genetic overlap between schizophrenia susceptibility and antipsychotic treatment response. Known and predicted drug target genes were investigated for enrichment for schizophrenia susceptibility loci. The majority of significantly enriched loci fell within novel predicted antipsychotic target genes (277 of 347 total genes; P = 0.019). Additionally, Dr. Ruderfer found that there is an enrichment for rare mutations within drug targets when assessing treatment resistant patients. Dr. Raquel Iniesta (King's College London, UK) presented a machine learning approach to antidepressant treatment response. Her hypothesis was that utilizing a combination of clinical and genetic variables could more accurately predict treatment outcome. Dr. Iniesta collected various clinical and genetic information from patients (N = 430). The machine learning approach used a training (N = 280) and testing (N = 150) dataset to predict future outcomes using the collected information. Dr. Iniesta observed that accuracy was improved by combining genetic and clinical variables for both nortriptyline (R2 = 16%) and citalopram (R2 = 15%) subgroups. Dr. Arun Tiwari (Centre for Addiction and Mental Health, Canada) discussed his study on the orexin receptors and antipsychotic-induced weight gain (AIWG). Several polymorphisms in the human copper transporter 2 (HCTR2) gene were nominally associated (P ∼ 5×10-3) with AIWG. Dr. Tiwari pointed out that these variants fall in a region that has been predicted to have weak enhancer activity (The ENCODE Project Consortium, 2012). Dr. Tiwari and his colleagues also constructed a preliminary risk model for AIWG that predicted 67% of the variance. Ms. Sophie Legge (Cardiff University, UK) reported on her exploration of genetic factors associated with clozapine-induced neutropenia. The patient sample included patients with clozapine-induced neutropenia from the CLOZUK and CardiffCOGS cohorts (defined by Rees et al., 2013), and clozapine-treated controls (without clozapine-induced neutropenia). For the GWAS findings, two intergenic variants reached genome-wide significance. After replication, one variant affecting both solute carrier organic anion transporter family members 1B3 (SLCO1B3) and 1B7 (SLCO1B7) transcripts was significant. This is a novel finding for clozapine research, although the SLCO genes have previously been associated with adverse drug reactions (SEARCH Collaborative Group, 2008). Dr. Joanna Biernacka (Mayo Clinic, USA) reported on her results of a pharmacogenomic GWAS on antidepressant-induced weight gain. The aim was to identify genetic variants that predict weight gain or loss during the course of treatment with citalopram or escitalopram. Although baseline weight was available, weight was not measured at follow-up visits, and therefore retrospective recall data derived from the Quick Inventory of Depressive Symptomatology (QIDS) was used to define weight change after initiation of treatment. At week 4, one variant close to the complexin 1 gene reached genome-wide significance for weight loss, and at week 8, a different variant within the aldo-keto reductase family 1 member C2 (AKR1C2) gene was significantly associated with weight loss. Dr. Biernacka pointed out that both genes are candidates for antidepressant-induced weight gain/loss based on prior evidence of their impact on insulin exocytosis and adipogenesis, respectively. Dr. Todd Lencz (Zucker Hillside Hospital, USA) discussed the pharmacogenetics of antipsychotic-naïve patients. His study made use of a subset of the Malhotra et al. (2012) cohort to investigate risperidone-induced hyperprolactinemia and/or weight gain. Both of the top hits occurred within the CDK5 regulatory subunit associated protein 1-like 1 (CDKAL1) gene with the first SNP associated with increased prolactin and the second with increased weight gain. The mechanism involved leads to abberant proinsulin accumulation (Wei et al., 2011). Dr. Lencz announced that the Phase II data from 1,000 first episode psychosis patients will be presented during the next meeting in 2016. We would like to acknowledge the funding source for the Early Career Investigator Program (ECIP), NIAAA 5R13AA017055-08, Nurnberger, John I., Conference Support for World Congress on Psychiatric Genetics, and the University of Toronto McLaughlin Centre. Each summary is the subjective understanding of the rapporteur for each session. The data reported are as heard during the presentation and where possible; all statements have been sent to the speakers for approval for accuracy. However, the speakers are not responsible for any of the information contained in this report. We therefore would like to thank the speakers, WCPG organizers and committee members. We would like to acknowledge one of our rapporteurs, Dr. Zoe Robaina Jimenez for her contribution to this summary. Abbreviations 4C Circularized chromosome conformation capture ADHD Attention deficit hyperactivity disorder AEI Allelic expression imbalance AIE Adolescent intermittent ethanol AIS Adolescent intermittent saline AIWG Antipsychotic-induced weight gain AKR1C2 Aldo-keto reductase family 1 member C2 gene ALSPAC Avon Longitudinal Study of Parents and Children ANNOVAR ANNOtate VARiation APA American Psychiatric Association API Application program interface ARC Activity-regulated cytoskeleton-associated Army STARRS Army Study To Assess Risk and Resilence in Service members ASD Autistic spectrum disorders ASM Allele-specific DNA modification ASM-SNPs SNPs exhibiting allele-specific DNA modification ATP Adenosine triphosphate AUD Alcohol use disorder B4GALNT4 Beta-1,4-N-acetyl-galactosaminyl transferase 4 gene BA Brodmann Area BChE Butyrylcholinesterase gene BD Bipolar disorder BDNF Brain-derived neurotrophic factor gene BiGS Bipolar Disorder Genome Study BLUP Best Linear Unbiased Prediction BMI Body mass index BSC Bipolar Sequencing Consortium C4 Complement component 4 gene CACNA1C Calcium channel, voltage-dependent, L type, alpha 1C subunit gene CAMK2D Calcium/calmodulin-dependent protein kinase 2 delta gene CBT Cognitive behavioural therapy CDKAL1 CDK5 regulatory subunit associated protein 1-like 1 gene CELA2A Chymotrypsin-like elastase family, member 2A gene CELF1 Elav-like family member 1 gene CELF4 Elav-like family member 4 gene CHRNA3 Cholinergic receptor, nicotinic, alpha 3 gene CHRNA5 Cholinergic receptor, nicotinic, alpha 5 gene CLIP-Cap Combined Long-Insert Paired-End and Capture Clorfl95/ITPKB Clorfl95/inositol-trisphosphate 3-kinase B gene CNS Central nervous system CNVs Copy number variants CONVERGE China Oxford and VCU Experimental Research on Genetic Epidemiology CRISPR Clustered regularly-interspaced short palindromic repeats CSA Childhood sexual abuse CUGBP CUG triplet repeat, RNA binding protein gene CYP2D6 Cytochrome P450 2D6 gene CYTH2 Cytohesin 2 gene DBS Deep brain stimulation DEPICT Data-driven Expression-Prioritized Integration for Complex Traits DFNB31 Deafness, autosomal recessive 31 gene DISC1 Disrupted in schizophrenia gene DISC1FP1 Disrupted in schizophrenia fusion partner 1 gene DLPFC Dorsolateral prefrontal cortex DMR Differentially methylated region DNA Deoxyribonucleic acid DRD2 Dopamine D2 receptor gene DRD3 Dopamine D3 receptor gene dSNPs Disease-associated single nucleotide polymorphisms DZ Dizygotic EAGLE EArly Genetics &Lifecourse Epidemiology EMR Electronic medical records ENIGMA Enhancing NeuroImaging Genetics through Meta-Analysis eQTLs Expression quantitative trait loci ERSP1 Epithelial splicing regulatory protein 1 gene ExAC Exome Aggregation Consortium EU-GEI EUropean network of national schizophrenia networks studying Gene-Environment Interactions FAM19A2 Family with sequence similarity 19 (chemokine [C-C motif]-like), member A2 gene FASD Fetal alcohol spectrum disorder fMRI Functional magnetic resonance imaging FORBOW Families Overcoming Risks and Building Opportunities for Well-being GABRA6 Gamma-aminobutyric acid A receptor, alpha 6 gene GAD Generalized anxiety disorder GAD1 Glutamic acid decarboxylase 1 gene GAF Global Assessment of Functioning G-BLUP Genomic-BLUP GERA Genetic Epidemiology Research on Adult Health and Aging GRIK5 Glutamate receptor, ionotropic kainate 5 gene GROUP Genetic Risk and Outcome in Psychosis GTeX Genotype-Tissue Expression GxE Gene-environmental GWAS Genome-wide association study GWAS-HD Hypothesis-driven genome-wide association study HCTR2 Human copper transporter 2 gene HDAC4 Histone deacetylase 4 gene HER2 Human epidermal growth factor receptor 2 HLA-B Major histocompatibility complex, class I, human leukocyte antigen B HRs Hazard ratios HRM High Resolution Melting HTT Huntingtin gene IAP Inhibitor-of-apoptosis indels Insertion/deletions INPP5K Inositol polyphosphate 5-phosphatase K gene iPSCs Induced pluripotent stem cells iPSYCH The Lundbeck Foundation Initiative for Integrative Psychiatric Research IPT Interpersonal therapy IQ Intelligence quotient ISPG International Society of Psychiatric Genetics KARG Knowledgebase for Addiction Related Genes KMT2F SET domain containing 1A gene LASSO Least absolute shrinkage and selection operator LCLs Lymphoblastoid cell lines LD Linkage disequilibrium LDSC Linkage disequilibrium score regression LHPP Phospholysinephosphohistidine inorganic pyrophosphate phosphatase gene LMM Linear mixed model LOD Logarithm (base 10) of odds LoF Loss-of-function LPGAT1 Lysophosphatidylglycerol acyltransferase 1 gene LRFN5 Leucine rich repeat and fibronectin type III domain containing 5 gene LRRC4C Leucine rich repeat containing 4C gene LSA Latent semantic analysis LSD1 Lysine specific demethylase 1 gene MACF1 Microtubule-actin crosslinking factor 1 gene MAF Minor allele frequency MAPK Mitogen-activated protein kinases MDD Major depressive disorder MET MET proto-oncogene, receptor tyrosine kinase gene MFAP3 Microfibrillar associated protein 3 gene MHC Major Histocompatibility Complex miRNAs MicroRNAs MRI Magnetic resonance imaging MRS Magnetic resonance spectroscopy MT-CYB Mitochondrially encoded cytochrome b gene mtDNA Mitochondrial DNA MZ Monozygotic NADH Nicotinamide adenine dinucleotide nDNA Nuclear DNA NDUFV1 NADH dehydrogenase (ubiquinone) flavoprotein 1, 51kDa NDUFV2 NADH dehydrogenase (ubiquinone) flavoprotein 2, 24kDa NDUFS1 NADH dehydrogenase (ubiquinone) Fe-S protein 1, 75kDa NIMH National Institute of Mental Health NMDA N-methyl-D-aspartate NMR Nicotine metabolite ratio NOS1 Nitric oxide synthase 1 gene NPAS2 Neuronal PAS domain protein 2 gene OC Obsessive-compulsive OCD Obsessive-compulsive disorder OICR Ontario Institute for Cancer Research OMIM Online Mendelian Inheritance in Man database OR Odds ratio OXPHOS Oxidative phosphorylation PAE Prenatal alcohol exposure PATE2 Prostate and testis expressed 2 gene PCr Reserve phosphocreatine PGC Psychiatric Genomic Consortium PGC2-BIP32 Psychiatric Genomic Consortium Bipolar Disorder PheWAS Phenome wide association study PRS Polygenic risk score PSORSIC3 Psoriasis susceptibility 1 candidate 3 gene PTPRD Protein tyrosine phosphatase receptor type D gene PTSD Posttraumatic stress disorder PTSR Posttraumatic stress reaction PTVs Protein truncating variants QIDS Quick Inventory of Depressive Symptomatology RA Rheumatoid arthritis RDoC Research Domain Criteria rG Genetic correlation rGT Rat gambling task RVIS Residual Variation Intolerance Score SCLT1 Sodium channel and clathrin linker 1 gene SIRT1 Sirtuin 1 gene SKAT-O Optimized sequence Kernel association test SLC25A37 Mitochondrial iron transporter gene SLC6A3 Dopamine transporter gene SLCO1B3 Solute carrier organic anion transporter family member 1B3 gene SLCO1B7 Solute carrier organic anion transporter family member 1B7 gene SLE Stressful life events SNP Single nucleotide polymorphism SNRI Serotonin-norepinephrine reuptake inhibitor SNVs Single nucleotide variants SOCS6 Suppressor of cytokine signaling 6 gene SORL1 Sortilin-like receptor 1 gene SRSF5 Serine/arginine-rich splicing factor 5 gene SSAGA Semi-Structured Assessment for the Genetic of Alcoholism SCC Simons Simplex Collection SV2A Synaptic vesicle glucoprotein 2A gene SVM Support vector machine TAAR1 Trace amine associated receptor 1 gene TES Total exposure study TET Ten-eleven translocation THBS2 Thrombospondin 2 gene TMS Transcranial magnetic stimulation TMS-EEG Transcranial magnetic stimulation with Electroencephalogram ToMMo Tohoku Medical Megabank project TRANK1 Tetratricopeptide repeat and ankyrin repeat containing 1 gene UBR4 Ubiquitin protein ligase E3 component N-recognin 4 gene UC Ulcerative colitis WCPG World Congress of Psychiatric Genetics Conflicts of Interest: There are no conflicts of interest for this report. Akiskal HS Walker P Puzantian VR King D Rosenthal TL Dranon M 1983 Bipolar outcome in the course of depressive illness. Phenomenologic, familial, and pharmacologic predictors J Affect Disord 5 2 115 128 6222091 Benros ME Pedersen MG Rasmussen H Eaton WW Nordentoft M Mortensen PB 2014 A nationwide study on the risk of autoimmune diseases in individuals with a personal or a family history of schizophrenia and related psychosis Am J Psychiatry 171 218 226 24129899 Bulik-Sullivan B Loh P Finucane HK Ripke S Yang J Patterson N Daly MJ Price AL Neale BM 2015a LD Score regression distinguishes confounding from polygenicity in genome-wide association studies Nat Genet 47 3 291 295 25642630 Bulik-Sullivan B Finucane HK Anttila V Gusev A Day FR Loh PR Duncan L Perry JR Patterson N Robinson EB Daly MJ Price AL Neale BM ReproGen Consortium; Psychiatric Genomics Consortium; Genetic Consortium for Anorexia Nervosa of the Wellcome Trust Case Control Consortium 3 2015b An atlas of genetic correlations across human diseases and traits Nat Genet 47 11 1236 1241 26414676 Cai N Chang S Li Y Li Q Hu J Liang J Song L Kretzschmar W Gan X Nicod J Rivera M Deng H Du B Li K Sang W Gao J Gao S Ha B Ho HY Hu C Hu J Hu Z Huang G Jiang G Jiang T Jin W Li G Li K Li Y Li Y Li Y Lin YT Liu L Liu T Liu Y Liu Y Lu Y Lv L Meng H Qian P Sang H Shen J Shi J Sun J Tao M Wang G Wang G Wang J Wang L Wang X Wang X Yang H Yang L Yin Y Zhang J Zhang K Sun N Zhang W Zhang X Zhang Z Zhong H Breen G Wang J Marchini J Chen Y Xu Q Xu X Mott R Huang GJ Kendler K Flint J 2015 Molecular signatures of major depression Curr Biol 25 1146 1156 25913401 Chang CC Chow CC Tellier LC Vattikuti S Purcell SM Lee JJ 2015 Second-generation PLINK: rising to the challenge of larger and richer datasets Gigascience 4 7 25722852 Chen DT Jiang X Akula N Shugart YY Wendland JR Steele CJ Kassem L Park JH Chatterjee N Jamain S Cheng A Leboyer M Muglia P Schulze TG Cichon S Nöthen MM Rietschel M McMahon FJ Farmer A Mcguffin P Craig I Lewis C Hosang G Cohen-Woods S Vincent JB Kennedy JL Strauss J BiGS 2013 Genome-Wide Association Study Meta-Analysis Of European And Asian-Ancestry Samples Identifies Three Novel Loci Associated With Bipolar Disorder Mol Psychiatry 18 2 195 205 22182935 CONVERGE Consortium 2015 Sparse whole-genome sequencing identifies two loci for major depressive disorder Nature 523 588 591 26176920 Du F Cooper AJ Thida T Sehovic S Lukas SE Cohen BM 2014 In vivo evidence for cerebral bioenergetic abnormalities in schizophrenia measured using 31P magnetization transfer spectroscopy JAMA Psychiatry 71 19 27 24196348 Du F Zhang Y Chen W 2012 Relayed magnetization transfer from nuclear Overhauser effect and chemical exchange observed by in vivo (3)(1)P MRS in rat brain Magn Reson Imaging 30 716 721 22459438 Dudbridge F 2013 Power and Predicitive Accuracy of Polygenic Risk Scores PLoS Genet 9 3 e1003348 23555274 Euesden J Lewis CM O'Reilly PF 2015 PRSice: Polygenic Risk Score software Bioinformatics 31 9 1466 1468 25550326 Fears SC Service SK Kremeyer B Araya C Araya X Bejarano J Ramirez M Castrillón G Gomez-Franco J Lopez MC Montoya G Montoya P Aldana I Teshiba TM Abaryan Z Al-Sharif NB Ericson M Jalbrzikowski M Luykx JJ Navarro L Tishler TA Altshuler L Bartzokis G Escobar J Glahn DC Ospina-Duque J Risch N Ruiz-Linares A Thompson PM Cantor RM Lopez-Jaramillo C Macaya G Molina J Reus VI Sabatti C Freimer NB Bearden CE 2014 Multisystem component phenotypes of bipolar disorder for genetic investigations of extended pedigrees JAMA Psychiatry 71 4 375 387 24522887 Finucane HK Bulik-Sullivan B Gusev A Trynka G Reshef Y Loh PR 2015 Paritioning heritability by functional annotation using genome-wide association summary statisitcs Nat Genet 47 11 1228 1135 26414678 Fullerton JM Koller DL Edenberg HJ Foroud T Liu H Glowinski AL McInnis MG Wilcox HC Frankland A Roberts G Schofield PR Mitchell PB Nurnberger JI Bipolar High Risk Study Group, BIGS Consortrium 2015 Assessment of first and second degree relatives of individuals with bipolar disorder shows increased genetic risk scores in both affected relatives and young At-Risk Individuals Am J Med Genet B Neuropsychiatr Genet 168 7 617 629 26178159 Gagliano SA Ravji R Barnes M Weale M Knight J 2015 Smoking Gun or Circumstantial Evidence? Comparison of Statistical Learning Methods using Functional Annotations for Prioritizing Risk Variants Sci Rep 5 13373 26300220 Gamazon ER Wheeler HE Shah KP Mozaffari SV Aquino-Michaels K Carroll RJ Eyler AE Denny JC Nicolae DL Cox NJ Im HK GTEx Consortium 2015 A gene-based association method for mapping traits using reference transcriptome data Nat Genet 47 9 1091 1098 26258848 Gartner K Baunack E 1981 Is the similarity of monozygotic twins due to genetic factors alone? Nature 292 5824 646 647 7254360 Goldberg JF Harrow M Whiteside JE 2001 Risk for bipolar illness in patients initially hospitalized for unipolar depression Am J Psychiatry 158 8 1265 1270 11481161 Hancock DB Reginsson GW Gaddis NC Chen X Saccone NL Lutz SM Qaiser B Sherva R Steinberg S Zink F Stacey SN Glasheen C Chen J Gu F Frederiksen BN Loukola A Gudbjartsson DF Brüske I Landi MT Bickeböller H Madden P Farrer L Kaprio J Kranzler HR Gelernter J Baker TB Kraft P Amos CI Caporaso NE Hokanson JE Bierut LJ Thorgeirsson TE Johnson EO Stefansson K 2015 Genome-wide meta-analysis reveals common splice site acceptor variant in CHRNA4 associated with nicotine dependence Transl Psychiatry 5 e651 26440539 Hudson G Gomez-Duran A Wilson IJ Chinnery PF 2014 Recent mitochondrial DNA mutations increase the risk of developing common late-onset human diseases PLoS Genet 10 e1004369 24852434 Johnson AD Zhang Y Papp AC Pinsonneault JK Lim JE Saffen D 2008 Polymorphisms affecting gene transcription and mRNA processing in pharmacogenetic candidate genes: detection through allelic expression imbalance in human target tissues Pharmacogenet Genomics 18 781 791 18698231 Kaminsky ZA Tang T Wang SC Ptak C Oh GH Wong AH 2009 DNA methylation profiles in monozygotic and dizygotic twins Nat Genet 41 2 240 245 19151718 Kann O Papageorgiou IE Draguhn A 2014 Highly energized inhibitory interneurons are a central element for information processing in cortical networks J Cereb Blood Flow Metab 34 1270 82 24896567 Khoury L Tang YL Bradley B Cubells JF Ressler KJ 2010 Substance use, childhood traumatic experience, and Posttraumatic Stress Disorder in an urban civilian population Depress Anxiety 27 12 1077 1086 21049532 Kim Y Xia K Tao R Guisti-Rodriguez P Vladimirov V van der Oord E 2014 A meta-analysis of gene expression quantitative trait loci in brain Transl Psych 4 e459 Kreek MJ Nielsen DA Butelman ER Laforge KS 2005 Genetic influences on impulsivity, risk taking, stress responsivity and vulnerability to drug abuse and addiction Nature Neurosci 8 1450 1457 16251987 Lieberman JA 2015 SHRINKS: The Untold Story of Psychiatry New York, NY Little, Brown and Company, Hachette Book Group Lieberman JA Stroup S McEvoy J Swartz MS Rosenheck RA Perkins DO 2005 Effectiveness of antipsychotic drugs in patients with chronic schizophrenia N Engl J Med 353 1209 1223 16172203 Lobo D Aleksandrova L Knight J Casey D el-Guebaly N Nobrega J Kennedy J 2014 Addiction-related genes in gambling disorders: new insights from parallel humand and pre-clinical models Mol Psychiatry 20 8 1002 1010 25266122 Maier R Moser G Chen GB Ripke S Coryell W Cross-Disorder Working Group of the Psychiatric Genomics Consortium 2015 Joint analysis of psychiatric disorders increases accuracy of risk prediction for schizophrenia, bipolar disorder, and major depressive disorder Am J Hum Genet 96 2 283 294 25640677 Malhotra AK Correll CU Chowdhury NI Müller DJ Gregersen PK Lee AT 2012 Association between common variants near the melanocortin 4 receptor gene and severe antipsychotic drug-induced weight gain Arch Gen Psychiatry 69 9 904 912 22566560 Monahan PO Stump T Coryell WH Harezlak J Marcoulides GA Liu H Steeger CM Mitchell PB Wilcox HC Hulvershorn LA Glowinski AL Iyer-Eimerbrink PA McInnis M Nurnberger JI 2015 Bipolar Disorder Genome Study (BiGS) Consortium Confirmatory test of two factors and four subtypes of bipolar disorder based on lifetime psychiatric co-morbidity Psychol Med 45 10 2181 2196 25823794 Morgan H Sutherland H Martin D Whitelaw E 1999 Epigenetic inheritance at the agouti locus in the mouse Nat Genet 23 3 314 318 10545949 Mulert C Juckel G Giegling I Pogarell O Leicht G Karch S Mavrogiorgou P Möller HJ Hegerl U Rujescu D 2006 A Ser9Gly polymorphism in the dopamine D3 receptor gene (DRD3) and event-related P300 potentials Neuropsychopharmacology 31 1335 1344 16395310 Mühleisen TW Leber M Schulze TG Strohmaier J Degenhardt F Treutlein J 2014 Genome-wide association study reveals two new risk loci for bipolar disorder Nature communications 5 3339 Nurnberger JI Jr McInnis M Reich W Kastelic E Wilcox HC Glowinski A Mitchell P Fisher C Erpe M Gershon ES Berrettini W Laite G Schweitzer R Rhoadarmer K Coleman VV Cai X Azzouz F Liu H Kamali M Brucksch C Monahan PO 2011 A high-risk study of bipolar disorder: childhood clinical phenotypes as precursors of major mood disorders Arch Gen Psychiatry 68 10 1012 1020 21969459 Ostergaard SD Straszek S Petrides G Skadhede S Jensen SO Munk-Jorgensen P Nielsen J 2014 Risk factors for conversion from unipolar psychotic depression to bipolar disorder Bipolar Disord 16 2 180 189 24215495 Pers TH Karjalainen JM Chan Y Westra HJ Wood AR Yang J Lui JC Vedantam S Gustafsson S Esko T Frayling T Speliotes EK Boehnke M Raychaudhuri S Fehrmann RS Hirschhorn JN Franke L Genetic Investigation of ANthropometric Traits (GIANT) Consortium 2015 Biological Interpretation Of Genome-Wide Association Studies Using Predicted Gene Functions Nat Commun 6 5890 25597830 Peterson RE Cai N Bigdeli TB Li Y Reimers M Nikulova A Webb BT Bacanu SA Riley BP Flint J Kendler KS 2016 The genetic architecture of major depression in Han Chinese women Mol Psychiatry submitted Petronis A 2010 Epigenetics as a unifying principle in the aetiology of complex traits and diseases Nature 465 721 727 20535201 Petrovski S Wang Q Heinzen EL Allen AS Goldstein DB 2013 Genic Intolerance To Functional Variation And The Interpretation Of Personal Genomes Plos Genet 9 e1003709 23990802 Radhu N de Jesus DR Ravindran LN Zanjani A Fitzgerald PB Daskalakis ZJ 2013 A meta-analysis of cortical inhibition and excitability using transcranial magnetic stimulation in psychiatric disorders Clin Neurophysiol 124 7 1309 1320 23485366 Rees E Walters JTR Georgieva L Isles AR Chambert KD Richards AL Mahoney-Davies G Legge SE Moran JL McCarroll SA O'Donovan MC O'wen MJ Kirov G 2013 Analysis of copy number variations at 15 schizophrenia-associated loci Br J Psychiatry 204 2 108 114 24311552 Ripke S Wray NR Lewis CM Hamilton SP Weissman MM Breen G Byrne EM Blackwood DHR Boomsma DI Cichon S Heath AC Holsboer F Lucae S Madden PAF Martin NG McGuffin P Muglia P Noethen MM Penninx BP Pergadia ML 2013a A mega-analysis of genome-wide association studies for major depressive disorder Mol Psychiatry 18 4 497 511 22472876 Ripke S O'Dushlaine C Chambert K Moran JL Kahler AK Akterin S 2013b Genome-wide association analysis identifies 13 new risk loci for schizophrenia Nat Genet 45 1150 1159 23974872 Robicsek O Karry R Petit I Salman-Kesner N Muller FJ Klein E 2013 Abnormal neuronal differentiation and mitochondrial dysfunction in hair follicle-derived induced pluripotent stem cells of schizophrenia patients Mol Psychiatry 18 1067 1076 23732879 Rollins B Martin MV Sequeira PA Moon EA Morgan LZ Watson SJ 2009 Mitochondrial variants in schizophrenia, bipolar disorder, and major depressive disorder PLoS One 4 e4913 19290059 Rossin EJ Lage K Raychaudhuri S Xavier RJ Tartar D Cotsapas C Daly MJ IIBDGC 2011 Proteins Encoded in Genomic Regions Associated with Immune-Mediated Disease Physically Interact and Suggest Underlying Biology PLoS Genet 7 1 e1001273 21249183 Sadee W Hartmann K Seweryn M Pietrzak M Handelman SK Rempala GA 2014 Missing heritability of common diseases and treatments outside the protein-coding exome Hum Genet 133 1199 215 25107510 Schizophrenia Working Group of the Psychiatric Genomics Consortium 2014 Biological insights from 108 schizophrenia-associated genetic loci Nature 511 7510 421 427 25056061 SEARCH Collaborative Group Link E Parish S Armitage J Bowman L Heath S Matsuda F Gut I Lathrop M Collins R 2008 SLCO1B1 variants and statin-induced myopathy – a genomewide study N Engl J Med 359 8 789 799 18650507 Sequeira A Rollins B Magnan C van Oven M Baldi P Myers RM 2015 Mitochondrial mutations in subjects with psychiatric disorders PLoS One 10 e0127280 26011537 Sequeira A Martin MV Rollins B Moon EA Bunney WE Macciardi F 2012 Mitochondrial mutations and polymorphisms in psychiatric disorders Front Genet 3 103 22723804 Stergachis AB Neph S Sandstrom R Haugen E Reynolds AP Zhang M 2014 Conservation of trans-acting circuitry during mammalian regulatory evolution Nature 515 365 370 25409825 Sullivan D Pinsonneault JK Papp AC Zhu H Lemeshow S Mash DC 2013 Dopamine transporter DAT and receptor DRD2 variants affect risk of lethal cocaine abuse: a gene-gene-environment interaction Transl Psychiatry 3 e222 23340505 Sullivan PF Neale MC Kendler KS 2000 Genetic epidemiology of major depression: review and meta-analysis Am J Psychiatry 157 10 1552 1562 11007705 The ENCODE Project Consortium 2012 An integrated encyclopedia of DNA elements in the human genome Nature 489 57 74 22955616 Uher R Cumby J MacKenzie LE Morash-Conway J Glover MJ Aylott A Propper L Abidi S Bagnell A Pavlova B Hajek T Lovas D Pajer K Gardner W Levy A Alda M 2014 A familial risk enriched cohort as a platform for testing early interventions to prevent severe mental illness BMC Psychiatry 14 344 25439055 Wallace DC 2015 Mitochondrial DNA Variation in Human Radiation and Disease Cell 163 1 33 38 26406369 Wang D Poi MJ Sun X Gaedigk A Leeder JS Sadee W 2014 Common CYP2D6 polymorphisms affecting alternative splicing and transcription: long-range haplotypes with two regulatory variants modulate CYP2D6 activity Hum Mol Genet 23 268 278 23985325 Wang K Li M Hakonarson H 2010 Annovar: Functional Annotation Of Genetic Variants From High-Throughput Sequencing Data Nucleic Acids Res 38 e164 20601685 Warde-Farley D Donaldson SL Comes O Zuberi K Badrawi R Chao P Franz M Grouios C Kazi F Lopes CT Maitland A Mostafavi S Montojo J Shao Q Wright G Bader GD Morris Q 2010 The GeneMANIA prediction server: biological network integration for gene prioritization and predicting gene function Nucleic Acids Res 38 Web Server issue W214 220 20576703 Wei FY Suzuki T Watanabe S Kimura S 2011 Deficit of tRNALys modification by Cdkal1 causes the development of type 2 diabetes in mice J Clin Invest 121 9 3598 3608 21841312 Welter D MacArthur J Morales J Burdett T Hall P Junkins H 2014 The NHGRI GWAS Catalog, a curated resource of SNP-trait associations Nucleic Acids Res 42 D1001 1006 24316577 Whalley HC Sussmann JE Romaniuk L Stewart T Kielty S Lawrie SM Hall J McIntosh AM 2015 Dysfunction of emotional brain systems in individuals at high risk of mood disorder with depression and predictive features prior to illness Psychol Med 45 6 1207 1218 25229638 Yuksel C Du F Ravichandran C Goldbach JR Thida T Lin P 2015 Abnormal high-energy phosphate molecule metabolism during regional brain activation in patients with bipolar disorder Mol Psychiatry 20 1079 1084 25754079 Zhang Y Bertolino A Fazio L Blasi G Rampino A Romano R 2007 Polymorphisms in human dopamine D2 receptor gene affect gene expression, splicing, and neuronal activity during working memory Proc Natl Acad Sci U S A 104 20552 20557 18077373

PMC005xxxxxx/PMC5134926.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 0404511 7473 Science Science Science (New York, N.Y.) 0036-8075 1095-9203 27846469 5134926 10.1126/science.aad6872 NIHMS829268 Article A nuclease that mediates cell death induced by DNA damage and poly(ADP-ribose) polymerase-1 Wang Yingfei 1234‡ An Ran 125 Umanah George K. 12 Park Hyejin 126 Nambiar Kalyani 12 Eacker Stephen M. 126 Kim BongWoo 3 Bao Lei 3 Harraz Maged M. 127 Chang Calvin 1 Chen Rong 12 Wang Jennifer E. 3 Kam Tae-In 126 Jeong Jun Seop 89 Xie Zhi 10* Neifert Stewart 126 Qian Jiang 10 Andrabi Shaida A. 12† Blackshaw Seth 7910 Zhu Heng 89 Song Hongjun 127 Ming Guo-li 127 Dawson Valina L. 126711‡ Dawson Ted M. 12678‡ 1 Neuroregeneration and Stem Cell Programs, Institute for Cell Engineering, Johns Hopkins University School of Medicine, Baltimore, MD 21205, USA 2 Department of Neurology, Johns Hopkins University School of Medicine, Baltimore, MD 21205, USA 3 Department of Pathology, University of Texas Southwestern Medical Center, Dallas, TX 75390, USA 4 Department of Neurology and Neurotherapeutics, University of Texas Southwestern Medical Center, Dallas, TX 75390, USA 5 Department of Neurology of Huashan Hospital, State Key Laboratory of Medical Neurobiology, Fudan University, Shanghai 200032, China 6 Adrienne Helis Malvin Medical Research Foundation, New Orleans, LA 70130-2685, USA 7 Solomon H. Snyder Department of Neuroscience, Johns Hopkins University School of Medicine, Baltimore, MD 21205, USA 8 Department of Pharmacology and Molecular Sciences, Johns Hopkins University School of Medicine, Baltimore, MD 21205, USA 9 Center for High-Throughput Biology, Johns Hopkins University School of Medicine, Baltimore, MD 21205, USA 10 Department of Ophthalmology, Johns Hopkins University School of Medicine, Baltimore, MD 21205, USA 11 Department of Physiology, Johns Hopkins University School of Medicine, Baltimore, MD 21205, USA * Present address: Zhongshan Ophthalmic Center, Sun Yat-sen University, Guangzhou, China. † Present address: Department of Pharmacology and Toxicology, University of Alabama at Birmingham, Birmingham, AL 35294, USA. ‡ Corresponding author. tdawson@jhmi.edu (T.M.D.); vdawson1@jhmi.edu (V.L.D.); yingfei.wang@utsouthwestern.edu (Y.W.) 11 11 2016 7 10 2016 02 12 2016 354 6308 aad6872This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. Inhibition or genetic deletion of poly(ADP-ribose) (PAR) polymerase-1 (PARP-1) is protective against toxic insults in many organ systems. The molecular mechanisms underlying PARP-1–dependent cell death involve release of mitochondrial apoptosis-inducing factor (AIF) and its translocation to the nucleus, which results in chromatinolysis. We identified macrophage migration inhibitory factor (MIF) as a PARP-1–dependent AIF-associated nuclease (PAAN). AIF was required for recruitment of MIF to the nucleus, where MIF cleaves genomic DNA into large fragments. Depletion of MIF, disruption of the AIF-MIF interaction, or mutation of glutamic acid at position 22 in the catalytic nuclease domain blocked MIF nuclease activity and inhibited chromatinolysis, cell death induced by glutamate excitotoxicity, and focal stroke. Inhibition of MIF's nuclease activity is a potential therapeutic target for diseases caused by excessive PARP-1 activation. Poly(ADP-ribose) (PAR) polymerase-1 (PARP-1) is a nuclear enzyme that is activated by DNA damage and facilitates DNA repair (1). Excessive activation of PARP-1 causes an intrinsic caspase-independent cell death program designated parthanatos (2, 3), which occurs after toxic insults in many organ systems (4, 5), including ischemia-reperfusion injury after stroke and myocardial infarction; inflammatory injury; reactive oxygen species–induced injury; glutamate excitotoxicity; and neurodegenerative diseases, such as Parkinson's disease and Alzheimer's disease (2, 4, 6). Consistent with the idea that PARP-1 is a key cell-death mediator, PARP inhibitors or genetic deletion of PARP-1 protect against such cellular injury in models of human disease (2, 4, 5, 7). During parthanatos, PAR causes release of apoptosis-inducing factor (AIF) from the mitochondria and its translocation to the nucleus, resulting in fragmentation of DNA into 20- to 50-kb fragments (2, 8–11). AIF itself has no obvious nuclease activity (2). Although it has been suggested that CED-3 protease suppressor (CPS)–6, an endonuclease G (EndoG) homolog in Caenorhabditis elegans, cooperates with the worm AIF homolog (WAH-1) to promote DNA degradation (12), mammalian EndoG does not seem to have an essential role in PARP-dependent chromatinolysis and cell death (13, 14) and after transient focal cerebral ischemia in mammals (15). Thus, the nuclease responsible for the chromatinolysis during parthanatos is not known. PARP-1–dependent cell death requires MIF To confirm that the EndoG is dispensable for parthanatos, the CRISPR-Cas9 system was used to deplete (knockout) EndoG from human neuroblastoma cell line (SH-SY5Y) cells (fig. S1A). We found that knockout of EndoG failed to block N-methyl-N′-nitro-N-nitrosoguanidine (MNNG)–induced parthanatos (fig. S1B) and large DNA fragmentation (fig. S1C); thus, EndoG is unlikely to be the main contributor to PARP-1–dependent large DNA fragmentation and MNNG-induced cell death (fig. S1) (13, 15). To identify a PARP-1–dependent AIF-associated nuclease (PAAN), we probed protein chips containing more than 16,000 and 5000 human recombinant proteins in duplicate along with several control proteins (16) with recombinant mouse AIF. The 160 strongest interacting proteins were depleted with small interfering RNA (siRNA) in cultured human HeLa cells to screen for modifiers of parthanatos induced by MNNG (2, 9, 11) (Fig. 1, A and B). We further tested whether depletion of these potential AIF-interacting proteins provided protection equivalent to that of depletion of PARP-1 and whether the proteins exhibited sequence and structure similarity consistent with possible nuclease activity. Depletion of AIF interactor 18 was as protective as depletion of PARP-1 (Fig. 1B). AIF interactor 18 is previously known under various synonyms (glycosylation-inhibiting factor, phenylpyruvate tautomerase, l-dopachrome tautomerase, l-dopachrome isomerase), and it is collectively known as macrophage migration inhibitory factor (MIF or MMIF) (17, 18). Effects of three different short hairpin RNA (shRNA) constructs against human and mouse MIF confirmed that depletion of MIF protected against parthanatos induced by MNNG toxicity in HeLa cells or N-methyl-d-aspartate (NMDA) excitotoxicity in mouse primary cortical neurons (fig. S2, A to F). To rule out off-target effects from the shRNA, we prepared MIF constructs that were resistant to shRNA 1 (RshRNA1) and 3 (RshRNA3). Cells expressing these constructs were impervious to effects of the shRNAs (fig. S2, G and H). MIF contains three PD-D/E(X)K superfamily motifs that are found in many nucleases (19–21) (Fig. 1, C and D) and are highly conserved across mammalian species (fig. S3A). It also contains a CxxCxxHx(n)C zinc finger domain (Fig. 1C and fig. S3B), which is commonly found in DNA damage-response proteins (20). MIF exists as a trimer (22–24). The core PD-D/E(X)K topology structure in the MIF trimer consists of four β strands next to two α strands (Fig. 1E and fig. S3, C to G, and supplementary text), which is similar to those of well-characterized nucleases, including Eco RI, Eco RV, Exo III, and Pvu II (fig. S3, H to O, and supplementary text). These sequence analyses and three-dimensional (3D) modeling results indicated that MIF belongs to the PD-D/E(X)K nuclease-like superfamily (25, 26). MIF is a nuclease To determine whether MIF has nuclease activity, we incubated a plasmid c–promoter DNA (pcDNA) vector with recombinant human MIF. Supercoiled pcDNA was cleaved by MIF—but not by its nuclease-deficient mutant MIF E22Q [in which glutamine (Q) replaces glutamic acid (E) at position 22] identified in the nuclease assays below—into an open circular form and, further, to a linear form (Fig. 2A). Moreover, MIF cleaved human genomic DNA in a concentration- and time-dependent manner (fig. S4, A and B). Addition of 10 mM Mg2+, 2 mM Ca2+, or 1 mM Mn2+ was required for MIF nuclease activity (fig. S4C), consistent with the divalent cation concentrations required for in vitro activity of other similar nucleases (27). EDTA blocked MIF's nuclease activity against human genomic DNA (Fig. 2B). In the absence of the divalent cation or with the cation at 2 to 10 μM, MIF had no nuclease activity (fig. S4C). Addition of 200 μM Zn2+ precipitated genomic DNA in the presence of MIF, whereas 2 μM Zn2+ had no effect. Na+ had no effect on MIF's nuclease activity (fig. S4C). Pulsed-field gel electrophoresis indicated that MIF cleaves human genomic DNA into large fragments comparable to those of DNA purified from HeLa cells treated with MNNG (Fig. 2B, lane 8). Depletion of MIF with shRNA prevented MNNG-induced DNA cleavage, which was similar to the effect of PARP inhibition by 3,4-dihydro-5[4-(1-piperindinyl)butoxy]-1(2H)-isoquinoline (DPQ) (Fig. 2C). Because MIF has been reported to have tautomerase activity, we tested the effects of the MIF tautomerase inhibitor ISO-1 (28). ISO-1 failed to prevent MNNG-induced DNA damage (Fig. 2C). Moreover, the MIF P2G (also known as the P1G) tautomerase mutant, which lacks tautomerase activity (29), had no effect on MIF's nuclease activity (fig. S4D). These data indicate that MIF is a nuclease that functions in PARP-1–dependent DNA fragmentation. To identify amino acid residues critical for MIF's nuclease activity, we mutated key aspartate, glutamate, and proline residues within the PD-D/E(X)K domains of MIF. E22Q, but not Glu replaced by Asp (E22D), inhibited MIF's nuclease activity, whereas replacement with Ala (E22A) partially reduced MIF's nuclease activity (Fig. 2D; fig. S4, E to H; and supplementary text). Thus, this glutamic acid residue (E22) in the first α helix of MIF is critical for its nuclease activity, which is consistent with reports that this glutamic acid in the first α helix of many exonuclease-endonuclease-phosphatase (EEP) domain superfamily nucleases is highly conserved and that it is the active site for nuclease activity (25, 26). MIF has both oxidoreductase and tautomerase activities (28, 30, 31). MIF active site mutants E22Q and E22A had no effect on MIF's oxidoreductase or tautomerase activities (fig. S5, A and B, and supplementary text). The lack of effect indicated that MIF nuclease activity is independent of its oxidoreductase and tautomerase activities. Moreover, MIF's protein conformation was unaffected by the E22Q and E22A mutations as determined by far-ultraviolet (UV) circular dichroism (CD) and near UV CD spectroscopy (fig. S5, C to M, and supplementary text). The purity of MIF proteins was confirmed by Coomassie blue staining, fast protein liquid chromatography (FPLC), and mass spectrometry (MS) assays (fig. S4G and fig. S5, C and D; Materials and methods; and supplementary text). No adventitious nuclease contamination was observed. MIF preferentially binds to stem-loop single-stranded DNA To determine the characteristics of DNA sequences bound by MIF in an unbiased manner, HeLa cells were treated with dimethyl sulfoxide (DMSO) or MNNG (50 μM, 15 min), followed by anti-MIF chromatin immunoprecipitation (ChIP) assays and deep sequencing (fig. S6 and supplementary text). We used the multiple Em for motif elicitation (MEME) program, which performs comprehensive motif analysis on large sets of nucleotide sequences (32), and we identified two classes of MIF-binding motifs (Fig. 3A). The first class (sequences 1 through 3) represents a highly related family of overlapping sequences (Fig. 3A and fig. S7A). The sequence features of this family are best captured in sequence 1 with 30 nucleotides and designated PS30, the most statistically significant motif identified, as determined by the MEME program (E-value = 1.4e–051) (Fig. 3A and fig. S7A). The second class identified was a poly(A) sequence. We performed 3D modeling to determine likely points of DNA interaction with MIF's PD-D/E(X)K motif. Within the PD-D/E(X)K motif, P16 and D17 on MIF are predicted to be positioned close to double-stranded DNA (dsDNA), whereas E22 is close to ssDNA, indicating MIF might bind single-stranded DNA (ssDNA), dsDNA, or both (fig. S7B). We examined both single-stranded and double-stranded forms of MIF DNA substrates for MIF binding and cleavage specificity. We synthesized the ssPS30 sequence with a 5′ biotin label and subjected it to an electrophoretic mobility shift assay (EMSA) (fig. S7C). MIF bound to the biotin-labeled ssPS30, forming one major complex in the presence of 10 mM Mg2+ (fig. S7C), which was completely disrupted by the addition of excess unlabeled DNA substrate (PS30) or a polyclonal antibody to MIF (fig. S7C). MIF E22Q, E22A, P16A, P17A, and P17Q mutants still formed MIF/ssPS30 complexes (fig. S7C). Because ssPS30 has the potential to form a stem-loop structure with unpaired bases at the 5′ and 3′ ends, we tested whether MIF binds to ssDNA with sequence or structure specificity. We used 5′ biotin-labeled ssPS30 and sequence-related substrates with different structures created by removing unpaired bases at the 5′ end, 3′ end, or both 5′ and 3′ ends, or by eliminating the stem loop in the EMSA (Fig. 3B and fig. S8). Completely removing the 3′ unpaired bases (5′bLF) had no effect on the DNA-MIF complex formation (Fig. 3B). In contrast, removing the 5′ unpaired bases (5′bRF) reduced, but did not abolish DNA-MIF binding. Similar results are observed when both 5′ and 3′ unpaired bases were removed (5′bSL). Thus MIF appears to mainly bind to 5′ unpaired bases in ssDNA with stem-loop structures. We also used a poly(A) sequence that has no stem loop (5′bPA30) and a short poly(A) sequence at the 5′ end of a stem-loop structure (5′b3F1) as the substrates. MIF failed to bind to 5′bPA30 but did bind to 5′b3F1. These results indicated that a stem loop is required for MIF-ssDNA binding (Fig. 3B and fig. S8). We also tested a substrate unrelated in sequence but that had a stem loop–like structure (5′bL3). MIF bound weakly to 5′bL3. But its binding efficiency was much lower than that of 5′bPS30. These data indicate that MIF preferentially binds to ssDNA with a stem loop and that its specificity is not entirely determined by the sequence. We also tested whether MIF bound to dsDNA with PS30; poly(A); substrates with sequence similarity to PS30 (5′bPS30, 5′bSL, 5′bLF, 5′bRF, 5′bPA30, and 5′bPA5E); and others with nonrelated sequences (PCS and 5′bL3) (Fig. 3B and fig. S8). MIF failed to bind to any of these double-stranded substrates (Fig. 3B). MIF cleaves 3′ unpaired bases of stem-loop ssDNA To determine whether MIF cleaves ssDNA or dsDNA, we added 35 random nucleotides to both the 5′ and 3′ ends of the PS30 DNA binding motif (designated PS100) and, under identical conditions, measured cleavage of ssDNA (ssPS100) or dsDNA (dsPS100). MIF cleaved ssPS100 and its complementary strand ssPS100R, but not dsPS100 (fig. S9, A and B). The MIF DNA binding motif identified by ChIP sequencing (PS30) appeared to be sufficient for MIF cleavage because MIF cleaved ssPS30 in a concentration-dependent manner (fig. S9C). MIF cleavage of ssPS30 required Mg2+, and MIF E22Q and E22A mutations blocked the cleavage of ssPS30 (fig. S9D). MIF cleaved ssPS30 in a time-dependent manner with a t1/2 of 12 min, and it cleaved ssPS30 in a concentration-dependent manner with an affinity for substrate (Km) of 2 μM and a maximum initial velocity (Vmax) of 41.7 nM/min (fig. S9, E to G). These kinetic properties are similar to those of other PD-D/E(X)K nucleases, such as Eco RI (27, 33). MIF also cleaved dsPS30 (fig. S9H), but required at least 4 times as high MIF concentrations and a four-fifths reduction in substrate concentration (compare lane 2 of fig. S9C to lane 2 of fig. S9H). MIF failed to cleave its related sequence dsRF, or the nonrelated sequence dsL3 (fig. S9H). MIF's preference for ssDNA is consistent with the 3D model of ssDNA binding to MIF's active site (fig. S7B) and our MIF-DNA binding assays (Fig. 3B). In the presence of AIF, MIF more efficiently cleaved genomic DNA and dsPS30 (fig. S10, A to C), which might be because of the observation that AIF enhanced MIF binding to dsDNA (fig. S10D). To determine whether MIF has sequence- or structure-specific endonuclease or exonuclease activity, we synthesized a series of variants labeled at the 5′ and 3′ ends with biotin, on the basis of the secondary structure of the DNA substrate ssPS30, and measured their cleavage by MIF (Fig. 3C and fig. S8). MIF had 3′ exonuclease activity and preferentially recognized and degraded unpaired bases at the 3′ end of ssPS30. This was blocked by biotin modification at the 3′ end (lanes 2 to 5 in Fig. 3C, fig. S8, and tables S1 and S2). MIF's 3′ exonuclease activity was also supported by cleavage assays in which the 5′bRF or 5′b3E substrates were used (Fig. 3C, fig. S8, and tables S1 and S2). Moreover, we used poly(A) (PA30), which lacks secondary structure and cannot be stained by ethidium bromide (EtBr) (Fig. 3C, top). We found that MIF's 3′ exonuclease activity allowed it to cleave 5′ biotin–poly(A) (5′bPA30), but not 3′ biotin–poly(A) (3′bPA30), so that MIF's 3′ exonuclease activity can occur independently of secondary structure (Fig. 3C, bottom, and fig. S8). MIF endonuclease activity was also influenced by secondary structure, because it cleaved short unpaired bases of ssDNA at the 3′ end adjacent to the stem loop (5′bPS40, 3′bPS40, 5′b3F1, 3′b3F1, and 5′bL3), as well as 3′-OH or 3′-biotin adjacent to the stem loop (3′bSL and 3′bLF) (Fig. 3C and fig. S8). In contrast to its exonuclease activity, MIF's endonuclease activity was not blocked by biotin modification at the substrate's 3′ end (3′bSL, 3′bLF, 3′bPS40, and 3′b3F1). However, 5′bL3, a sequence not related to PS30 but with a similar stem-loop structure, was cleaved by MIF, but with less efficiency (Fig. 3C and fig. S8). These results indicate that MIF has both 3′ exonuclease and endonuclease activities and cleaves unpaired bases of stem-loop ssDNA at the 3′ end. In the presence of AIF, AIF also increased the binding of MIF to ssDNAs, including 5′bPS30, as well as 5′bSL, which has no 5′ unpaired bases (fig. S10D). Nevertheless, we found that AIF increased both exonuclease and endonuclease activities of MIF (0.5 μM) on 5′bPS30, 3′bPS30, and 3′bSL (fig. S10E). However, AIF has a rather weak effect, if any, on the nuclease activity of MIF at 4 μM (fig. S10F). At this higher concentration, MIF itself can efficiently bind and cleave ssDNAs. These data suggest that AIF may enhance MIF nuclease activity by increasing its binding to ssDNAs. To further study where MIF cleaves DNA and to avoid the potential interference of biotin labeling, we used non-labeled PS30 and 3F1, which has only one unpaired base at the 3′ end of the stem-loop structure as substrates and customized two different DNA ladders based on PS30. After incubation of these substrates with MIF (2 μM) for 2 hours, two major products of 20 and 22 nucleotides were detected (Fig. 3D). Faint bands of higher molecular mass were also observed. These bands were more obvious in the experiment in which PS30 was biotin labeled and the incubation time was 1 hour (Fig. 3D). MIF cleavage of the 3F1 substrate yielded only a 29-nucleotide (nt) band consistent with cleavage of one unpaired base at the 3′ end of the stem-loop structure (Fig. 3, D and E). These data indicate that PS30 is initially cleaved by MIF after “A23↓T24↓T25↓” (arrow indicates cleavage) by both 3′ exonuclease and endonuclease activity (Fig. 3E, left). Then the resulting product appears to form a new stem-loop structure, as predicted by the online RNA/DNA structure prediction software (http://rna.urmc.rochester.edu/RNAstructureWeb/Servers/Predict1/Predict1.html) (Fig. 3E, right). MIF then cleaves at the new unpaired bases at the 3′ end of this stem-loop structure after “G20↓G21↓G22↓”. We conclude that MIF cleaves unpaired bases at the 3′ end adjacent to the stem loop at the +1 to ~+3 positions through both 3′ exonuclease and endonuclease activities. AIF interacts with MIF and recruits MIF to the nucleus Wild-type (WT) glutathione S-transferase–tagged AIF (GST-AIF) associated with MIF, and wild-type GST-MIF associated with AIF in GST pulldown analyses from cell lysates (Fig. 4A; fig. S11, A to D; and supplementary text). We mapped the MIF-AIF binding domain. MIF bound to AIF at amino acids 567 to 592 (fig. S11, A to C, and supplementary text). Conversely, the MIF E22A mutant showed reduced binding to GST-AIF, whereas the E22D and E22Q mutants still bound to GST-AIF (Fig. 4, A and B, and fig. S11D). The other PD-D/E(X)K and C57A;C60A mutations still bound GST-AIF (fig. S11D). Thus, MIF E22 appears to be critical for AIF binding. Endogenous AIF also coimmunoprecipitated with MIF from cortical neurons treated with NMDA (500 μM) but was barely detectable in untreated cultures (Fig. 4, C and D). MIF was localized predominantly to the cytosol of both cortical neurons (Fig. 4E) and HeLa cells (fig. S12A). Both MIF and AIF translocated to the nucleus in cortical neurons treated with NMDA (Fig. 4, E and F) and HeLa cells stimulated with MNNG (fig. S12A). Depletion of AIF with shRNA led to a loss of MIF translocation to the nucleus, but depletion of MIF did not prevent translocation of AIF to the nucleus in cells exposed to NMDA (Fig. 4, E and F). Subcellular fractionation into nuclear and postnuclear fractions confirmed the translocation of MIF and AIF to the nucleus in cultured cortical neurons exposed to NMDA (Fig. 4, G to I). AIF was required for MIF translocation (Fig. 4, E to I). DPQ prevented accumulation of both MIF and AIF in the nucleus in HeLa cells treated with MNNG (fig. S12, A to C) and cortical neurons treated with NMDA (fig. S13, A to C). Consistent with the notion that NMDA excitotoxicity involves nitric oxide production, the nitric oxide synthase inhibitor nitro-arginine (N-Arg) prevented accumulation of both MIF and AIF in the nucleus (fig. S13, A to C). We transduced primary cortical cultures from WT MIF knockout mice with lentivirus carrying Flag-tagged MIF (MIF-WT-Flag) or MIF mutants (MIF-E22Q-Flag and MIF-E22A-Flag) to confirm that AIF and MIF binding is required for MIF nuclear accumulation after NMDA administration (Fig. 5, A and B). Wild-type MIF and E22Q interacted with AIF, but MIF E22A did not bind to AIF (Fig. 5B). In nontransduced MIF knockout cultures and in MIF knockout cultures transduced with MIF-WT-Flag, MIF-E22Q-Flag, and MIF-E22A-Flag, AIF translocated to the nucleus when cells were exposed to NMDA (Fig. 5, C and D). Both MIF wild-type and MIF E22Q also translocated to the nucleus; however, the MIF E22A mutant, which is deficient in AIF binding, failed to do so (Fig. 5, C and D). Separation of nuclear and post-nuclear fractions confirmed the observations made by immunofluorescence (Fig. 5, E to G). These results indicate that MIF's interaction with AIF is required for the nuclear translocation of MIF. MIF nuclease activity is required for chromatinolysis and parthanatos To determine whether MIF's nuclease activity and AIF-mediated recruitment are required for parthanatos, we transduced MIF knockout cultures with the nuclease-deficient MIF E22Q mutant and the AIF binding–deficient MIF E22A mutant. Consistent with the shRNA experiments, cortical cultures lacking MIF were resistant to NMDA excitotoxicity (Fig. 6, A and B). Transduction of cells with wild-type MIF or the tautomerase-deficient mutant MIF P2G fully restored NMDA excitotoxicity; conversely, neither MIF E22Q nor MIF E22A restored NMDA excitotoxicity (Fig. 6, A and B). By the comet assay, a method to measure DNA damage, we found that NMDA administration in wild-type cortical neurons resulted in substantial numbers of neurons with DNA damage, whereas no such damage was detected in MIF knockout neurons (Fig. 6, C to F). Transduction of knockout neurons with wild-type MIF, but not with MIF E22Q or MIF E22A, restored DNA damage in cells treated with NMDA (Fig. 6, C to F). Depletion of MIF with shRNA in HeLa cells with two different shRNAs resulted in a reduced number of cells showing damaged DNA after treatment with MNNG compared with DNA in cells treated with nontargeted shRNA (fig. S14, A to D). A pulsed-field gel electrophoresis assay of genomic DNA confirmed that NMDA administration caused large DNA fragments in wild-type cortical neurons but not in MIF knockout cortical neurons (Fig. 6G). No obvious large DNA fragments were observed in MIF knockout neurons transduced with MIF E22Q or MIF E22A (Fig. 6G). Transduction of knockout neurons with wild-type MIF or MIF P2G restored NMDA-induced formation of large DNA fragments (Fig. 6G). HeLa cells lacking MIF after we used CRISPR-Cas9 were resistant to MNNG toxicity (fig. S15, A to C). Transduction of knockout HeLa cells with wild-type MIF or MIF P2G restored MNNG-induced formation of large DNA fragments and toxicity (fig. S15). These results indicate that MIF is the major nuclease involved in large-scale DNA fragmentation during MNNG- or NMDA-induced parthanatos, which is independent from MIF's tautomerase activity. To evaluate the requirement of MIF nuclease activity and MIF binding to AIF in cell death due to parthanatos in vivo, we transduced MIF knockout mice with adeno-associated virus serotype 2 virus (AAV2) containing wild-type MIF, or the nuclease-deficient MIF E22Q mutant or the AIF-binding-deficient MIF E22A mutant by injecting the different AAV2 MIFs into the intracerebroventricular zone of newborn mice. The effectiveness of transduction was confirmed by immunostaining for MIF-Flag in the cortex, striatum, and hippocampus in adult mice (fig. S16, A and B). Two-month old male mice were then subjected to 45-min transient occlusion of the middle cerebral artery (MCAO). Despite the similar intensity of the ischemic insult (fig. S16C), infarct volume as previously reported (34) was reduced in MIF knockout mice in the cortex, striatum, and hemisphere by about 75% compared to that in their wild-type counterparts (Fig. 7, A to D). Moreover, the neuroprotection in MIF knockout mice remained for at least 7 days (Fig. 7, C and D). Expression of wild-type MIF, but not MIF E22Q or MIF E22A, in the MIF knockout mice restored infarct volume to that observed in wild-type animals (Fig. 7, A to D). We assessed behavior by spontaneous activity in the open field task on days 1, 3, and 7 after MCAO. Consistent with the infarct data, MIF knockout mice had improved behavioral scores compared to those of wild-type mice. MIF knockout mice expressing wild-type MIF had behavioral scores equivalent to those of wild-type mice whereas expression of MIF E22Q or MIF E22A had no effect (Fig. 7, E and F). Over 3 and 7 days, the behavioral scores of MIF knockout mice remained higher than those of wild-type treated mice (Fig. 7, F and G). A corner test measuring sensorimotor function showed that all mice do not show a side preference before MCAO surgery. However, wild-type mice and MIF knockout mice expressing wild-type MIF had significantly (P < 0.05 to P < 0.001, one-way analysis of variance (ANOVA)] increased turning toward the non-impaired side at days 1, 3, and 7 after MCAO (Fig. 7G), indicating these mice have more severe sensory and motor deficits. No preference was observed in MIF knockout mice and MIF knockout mice with expression of MIF E22Q or MIF E22A (Fig. 7G). Significant (P < 0.0001, one-way ANOVA) DNA damage as assessed by pulse field gel electrophoresis was observed at days 1, 3, and 7 after MCAO in wild-type mice or MIF knockout mice expressing wild-type MIF (Fig. 7, H and I). DNA damage was reduced in the MIF KO mice and MIF knockout mice expressing E22Q or E22A MIF (Fig. 7, H and I). We examined the localization of AIF and MIF by confocal microscopy in the penumbra region of the stroke (fig. S17, A and B). Consistent with the observation in cultured cortical neurons, AIF significantly (P < 0.001, one-way ANOVA) translocated to the nucleus at 1, 3, and 7 days after MCAO in wild-type animals. In MIF knockout animals as well as MIF knockout mice injected with MIF wild-type, E22Q, and E22A AIF significantly (P < 0.001, one-way ANOVA) translocated to the nucleus at 1 and 3 days after MCAO and there was reduced translocation of AIF at 7 days (fig. S17, A and B). Both MIF wild-type and MIF E22Q also significantly (P < 0.001, one-way ANOVA) translocated to the nucleus at 1 and 3 days after MCAO and there was reduced translocation at 7 days; however, the AIF binding–deficient mutant MIF E22A failed to do so (fig. S17, A and B). These data indicate that MIF is required for AIF-mediated neurotoxicity and DNA cleavage and that AIF is required for MIF translocation in vivo. Conclusion We identified MIF as a PAAN. Prior crystallization studies of MIF allowed us to show via 3-D modeling that MIF is structurally similar to PD-D/E(x)K nucleases (25, 26). The MIF monomer, which has pseudo 2-fold symmetry does not contain the core PD-D/E(X)K structure since the MIF monomer has four β strands next to the two α helices, and the orientations of the β-strands within an isolated monomer do not fit the requirement of the PD-D/E(x)K topology (23). However, our structure-activity analyses based on the MIF trimer, which has 3-fold symmetry, indicated that the interactions of the β strands of each monomer with the other monomers results in a MIF PD-D/E(x)K structure that consists of four β strands next to two α strands (23). Two of the β strands are parallel (β-4 and β-5), whereas the other two strands (β-6 and β-7) (from the adjacent monomer) are antiparallel. This topology exquisitely supports the idea that MIF's nuclease activity requires the trimer as the monomers do not support the required topology and is consistent with MIF existing as a trimer. The PD-D/E(X)K domains in MIF are highly conserved in vertebrates. The glutamic acid residue (E22) in the first α helix of MIF is critical for its nuclease activity, which is consistent with prior reports that this glutamic acid in the first α helix of many exonuclease-endonuclease-phosphatase (EEP) domain superfamily nucleases is highly conserved, and it is the active site for nuclease activity (25, 26). MIF has both 3′ exonuclease and endonuclease activity. It preferentially binds to 5′ unpaired bases of ssDNA with the stem-loop structure and cleaves its 3′ unpaired bases. AIF interacts with MIF and recruits MIF to the nucleus where MIF binds and cleaves genomic DNA into large fragments similar to the size induced by stressors that activate parthanatos. MIF binding to AIF facilitates its cleavage of double-stranded genomic DNA, and, based on the chromatin immunoprecipitation sequencing ChiP-seq data, the average distance of MIF binding is about 15 to 60 kb, which is comparable to the size of large DNA fragments caused by MIF. MIF's cleavage of genomic DNA into 20- to 50-kb fragments is likely due to its rare binding on genomic DNA. Knockout of MIF reduces DNA fragmentation induced by stimuli that activate PARP-1–dependent cell death. Mutating a key amino acid residue, glutamic acid residue (E22), in the PD-D/E(X)K motif eliminates MIF's nuclease activity and protects cells from parthanatos both in vitro and in vivo. Disruption of the AIF and MIF protein-protein interaction prevents the translocation of MIF from the cytosol to the nucleus, which also protects against PARP-1–dependent cell death both in vitro and in vivo. Neither MIF's thiol-protein oxidoreductase activity nor tautomerase activity are involved in its actions as a nuclease. Knockout of MIF, a MIF nuclease–deficient mutant and a MIF AIF binding–deficient mutant all reduce infarct volume and have long-lasting behavioral rescue in the focal ischemia model of stroke in mice. Thus, MIF is a PAAN that is important in cell death because of activation of PARP-1 and the release of AIF (2). Future studies are required to further determine whether the stem-loop–ssDNA binding activity or the 3′ exonuclease and endonuclease activities of MIF, is important for its in vivo PAAN activity. In addition, our stroke data from MIF knockout mice indicate that other nucleases other than MIF might be involved in ischemic neuronal cell death. However, how these nucleases interact with MIF and contribute to PARP-1–induced cell death requires future studies. MIF has a variety of pleiotropic actions. It is widely distributed throughout the brain (35, 36). It functions as a nonclassically secreted cytokine and may play important roles in cancer biology, immune responses, and inflammation (18, 37). MIF also has important roles in cellular stress and apoptosis (34, 38, 39). How MIF's nuclease activity relates to its role in the immune system and its other actions requires future studies. Like PARP, inhibition of MIF nuclease activity is an attractive target for acute neurologic disorders. However, it may have advantages over PARP inhibition in chronic neurodegenerative diseases where long-term inhibition of PARP could impair detection and repair of DNA damage. Inhibition of MIF's nuclease activity could bypass this potential concern and offer a therapeutic opportunity for various disorders. Materials and methods Human protein chip high-throughput screening Human protein chips (16K and 5K), which were prepared by spotting more than 16,000 or 5000 highly purified proteins onto special nitrocellulose-coated slides (16), were incubated in renaturation buffer containing 50 mM Tris-HCl, pH 8.0, 100 mM NaCl, 1 mM DTT, 0.3% Tween 20 for 1 hour at 4°C. After Blocking with 5% non-fat dry milk for 1 hour at room temperature, protein chips were incubated with purified mouse AIF protein (50 nM, NP_036149) in 1% milk for 1 hour. Protein interaction was then determined either by sequentially incubating with rabbit anti-AIF antibody (JH532, JHU) (9, 11) and Alexa Fluor 647 donkey anti-rabbit IgG, or Alexa Fluor 647 donkey anti-rabbit IgG only as negative control. Protein microarrays were scanned with GenePix 4000B Microscanner (Tecan) using the Cy5 image and the median fluorescence of each spot was calculated. We used the same procedure described previously to identify interacting proteins (16). Reverse transfection format siRNA-based screen for PARP-1–dependent cell viability On-Target plus SMARTpool siRNAs targeting AIF-interacting proteins resulting from human protein chip high throughput screening were customized in 96-well plates from Dharmacon. The plates were rehydrated using DharmaFECT 1 transfection reagent at room temperature for 30 min. HeLa cells were then seeded in the plates with the cell density at 1 × 104/well. 48 hours after transfection, cells were treated with MNNG (50 μM) or DMSO for 15 min and then incubated in normal complete medium for 24 hours. After adding Alamar Blue for 1–4 hours, cell viability was determined by fluorescence at excitation wavelength 570 nm and Emission wavelength 585 nm. PARP-1 siRNAs were used as the positive control and non-target siRNAs as the negative control. Nuclease assays Human genomic DNA (200 ng/reaction, Promega), pcDNA (200 ng/reaction) or PS30 and its related and non-related substrates (1 μM) was incubated with wild-type MIF or its variants at a final concentration of 0.25–8 μM as indicated in 10 mM Tris-HCl buffer (pH 7.0) containing 10 mM MgCl2 or specific buffer as indicated, for 1 hour (with pcDNA and small DNA substrates) or 4 hours (with human genomic DNA) at 37°C. The reaction was terminated with loading buffer containing 10 mM EDTA and incubation on ice. The human genomic DNA samples were immediately separated on a 1.2% pulse field certified agarose in 0.5 × TBE buffer with initial switch time of 1.5 s and a final switch time of 3.5 s for 12 hours at 6 V/cm. pcDNA samples were determined by 1% agarose gel. Small DNA substrates were separated on 15% or 25% TBE-urea polyacrylamide (PAGE) gel or 20% TBE PAGE gel. The gel was then stained with 0.5 μg/ml Ethidium Bromide (EtBr) followed by electrophoretic transfer to a nylon membrane. Biotin-labeled DNA was further detected by chemiluminescence using the Chemiluminescent Nucleic Acid Detection Module (Thermo Scientific). Electrophoretic mobility shift assay (EMSA) EMSA assays were performed using the Light-Shift Chemiluminescent EMSA kit (Thermo Scientific) following the manufactures instruction. Briefly, purified MIF protein (2 μM) was incubated with biotin-labeled DNA substrates (10 nM) in the binding buffer containing 10 mM MgCl2 for 30 min on ice. Then samples were separated on 6% retardation polyacrylamide followed by electrophoretic transfer to a nylon membrane. Biotin-labeled DNA was further detected by chemiluminescence using the Chemiluminescent Nucleic Acid Detection Module (Thermo Scientific). Comet assay Comet assays were conducted following protocols provided by Trevigen (Gaithersburg, MD). Briefly, HeLa cells with or without MNNG treatment and cortical neurons with or without NMDA treatment were washed with ice-cold PBS 6 hours after the treatment, harvested by centrifugation at 720 g for 10 min and re-suspended in ice-cold PBS (Ca2+ and Mg2+ free) at 1 × 105 cells/ml. Cells were then combined with 1% low melting point agarose in PBS (42°C) in a ratio of 1:10 (v/v), and 50 μl of the cell-agarose mixture was immediately pipetted onto the Comet Slide and placed flatly at 4°C in the dark for 30 min to enhance the attachment. After being lysed in lysis buffer, slides were immersed with alkaline unwinding solution (200 mM NaOH, pH >13, 1 mM EDTA) for 1 hour at RT. The comet slides were transferred and electrophoresed with 1 L of alkaline unwinding solution at 21 Volts for 30 min in a horizontal electrophoresis apparatus. After draining the excess electrophoresis buffer, slides were rinsed twice with dH2O and then fixed with 70% ethanol for 5 min and stained with SYBR Green for 5 min at 4°C. Cell images were captured using a Zeiss epifluorescent microscope (Axiovert 200M) and image analysis was performed with CASP software (version 1.2.2). The length of the “comet tail,” which is termed as the length from the edge of the nucleus to the end of the comet tail, for each sample, was measured. Protein expression and purification Human EndoG (NM_004435), cyclophilin A (NM_021130), mouse AIF (NM_012019), human MIF (NM_002415) cDNA and their variants were subcloned into glutathione S-transferase (GST)-tagged pGex-6P-1 vector (GE Healthcare) by EcoRI and XhoI restriction sites and verified by sequencing. The protein was expressed and purified from Escherichia coli by glutathione sepharose. The GST tag was subsequently proteolytically removed for the nuclease assay. MIF point mutants were constructed by polymerase chain reaction (PCR) and verified by sequencing. The purity of MIF proteins that were used in the nuclease assays was further confirmed by mass spectrometry. MIF proteins purified by FPLC were also used in the nuclease assays and no obvious difference was observed between FPLC MIF and non-FPLC MIF proteins. GST protein was used as a negative control in the nuclease assay. Middle cerebral artery occlusion (MCAO) Cerebral ischemia was induced by 45 min of reversible MCAO as previously described (40). Adult male MIF knockout (KO) mice (2 to 4 months old, 20 to 28 g) were anesthetized with isoflurane and body temperature was maintained at 36.5 ± 0.5°C by a feedback-controlled heating system. A midline ventral neck incision was made, and unilateral MCAO was performed by inserting a 7.0 nylon monofilament into the right internal carotid artery 6–8 mm from the internal carotid/pterygopalatine artery bifurcation via an external carotid artery stump. Sham-operated animals were subjected to the same surgical procedure, but the suture was not advanced into the internal carotid artery. After 1 day, 3 days or 7 days of reperfusion, mice were perfused with PBS and stained with triphenyl tetrazolium chloride (TTC). The brains were further fixed with 4% PFA and sliced for immunohistochemistry (9, 11, 41). ChIP-seq We preformed ChIP-seq as previously described (42, 43). Briefly, HeLa Cells were first treated with DMSO or MNNG (50 μM, 15 min). 5 hours after MNNG treatment, cells were cross-linked with 1% formaldehyde for 20 min at 37°C, and quenched in 0.125 M glycine. Chromatin extraction was performed before sonication. The anti-MIF antibody (ab36146, Abcam) was used and DNA was immunoprecipitated from the sonicated cell lysates. The libraries were prepared according to Illumina's instructions accompanying the DNA Sample kit and sequenced using an Illumina HiSEq. 2000 with generation of 50 bp single-end reads. Detailed procedures are as follows. HeLa cells were treated with DMSO or MNNG (50 μM) for 15 min and cultured in the fresh medium for an additional 5 hours. Cells were then cross-linked with 1% formaldehyde for 10 min at 37°C, and the reaction was quenched in 0.125 M glycine for 20 min at room temperature. Chromatin was extracted using SimpleChIP Enzymatic Chromatin IP kit from Cell Signaling Technology (Cat# 9003), and sonicated 30 s on and 30 s off for 15 cycles using a Bioruptor Twin (Diagenode). The quality and size of sheared chromatin DNA were examined on an agarose gel by DNA electrophoresis. 10% of chromatin was kept as input and the rest of the chromatin was diluted and pre-cleared using 10 μl Magnetic protein G agarose slurry for 30 min at 4°C to exclude nonspecific binding to protein G agarose beads directly. The pre-cleared chromatin was incubated overnight with an anti-MIF antibody (3 μg/ml, ab36146, Abcam) or control IgG (3 μg/ml) in the presence of Magnetic protein G agarose slurry (30 μl) at 4°C. After washing the protein G agarose beads for 3 times, half of the protein G agarose/antibody complex was subjected to immunoblot assays to check the quality of the immunoprecipitation. Another half of the protein G agarose/antibody complex was eluted in 170 μl of elution buffer containing 1% SDS, 0.1 M NaHCO3 at 65°C. The eluates as well as the chromatin input were treated with 1 mg/ml RNase A at 37°C for 30 min, and reverse-crosslinked by incubating at 65°C for 4 hours after adding 3 μl of 5 M NaCl and 1 μl of 10 mg/ml proteinase K. Finally the chromatin DNA was purified using phenol/chloroform/isoamyl alcohol and precipitated by ethanol. The ChIP and input DNA libraries were prepared using Illumina's Truseq DNA LT Sample Prep Kit according to the instructions. The final product was amplified for 15 cycles. The quality and the size of the insert was analyzed using a bioanalyzer. Sequencing was performed in the Next Generation Sequencing Center at Johns Hopkins using an Illumina HiSEq. 2000 with generation of 50 bp single-end reads. The ChIP-seq raw data have been deposited in the GEO database accession #: GSE65110. ChIP-seq data analysis Raw data from the HiSEq. 2000 was converted to FASTQ using CASAVA v1.8 and demultiplexed. Reads were mapped to the human genome (hg19) using Bowtie2 (v2.0.5) using the default parameters. Converted SAM files were passed to MACS (v1.4.1) for peak calling using the default parameters. Peaks from DMSO- and MNNG-treated libraries were reported in .bed format and are provided in GEO. Peaks differentially identified in the DMSO- and MNNG-treated groups were parsed by a custom R script. Sequence corresponding to peaks identified in only MNNG-treated, but not DMSO-treated libraries were fed into SeSiMCMC_4_36, Chipmunk v4.3+, and MEMEchip v4.9.0 for motif discovery using default parameters. Data transfer: The CASAVAv1.8 software was used to convert the raw files into FASTQ files as well demultiplex the lanes. MIF-DNA docking methods A DNA duplex structure (44) (PDB accession 1BNA) and a single-stranded DNA structure [PDB accession 2RPD (45)] were docked onto the surface of MIF [PDB accession 1FIM (24)] using Hex-8.0. protein-DNA docking program (46, 47). The HEX program uses a surface complementarity algorithm to identify contact between protein and DNA. MIF surfaces were generated using Pymol. All images were viewed and labeled with pdb viewer, Pymol. The MIF-DNA docked models are shown as obtained from the HEX program. Lentivirus, adeno-associated virus (AAV) construction and virus production Mouse MIF-WT-Flag (NM_010798), MIF-E22Q-Flag and MIF-E22A-Flag were subcloned into a lentiviral cFugw vector by AgeI and EcoRI restriction sites, and its expression was driven by the human ubiquitin C (hUBC) promoter. Human MIF and mouse MIF shRNAs were designed using the website <http://katahdin.cshl.org/siRNA/RNAi.cgi?type=shRNA>. The program gave 97 nt oligo sequences for generating shRNAmirs. Using PacI SME2 forward primer 5' CAGAAGGTTAATTAAAAGGTATATTGCTGTTGACAGTGAGCG 3' and NheI SME2 reverse primer 5' CTAAAGTAGCCCCTTGCTAGCCGAGGCAGTAGGCA 3', we then PCR amplified them to generate the second strand and added PacI and NheI restriction sites to clone the products into pSME2, a construct that inserts an empty shRNAmir expression cassette in the pSM2 vector with modified restriction sites into the cFUGw backbone. This vector expresses GFP. The lentivirus was produced by transient transfection of the recombinant cFugw vector into 293FT cells together with three packaging vectors: pLP1, pLP2, and pVSV-G (1.3:1.5:1:1.5). The viral supernatants were collected at 48 and 72 hours after transfection and concentrated by ultracentrifugation for 2 hours at 50,000 g. MIF-WT-Flag, MIF-E22Q-Flag and MIF-E22A-Flag were subcloned into a AAV-WPRE-bGH (044 a.m./CBA-pI-WPRE-bGH) vector by BamHI and EcoRI restriction sites, and its expression was driven by chicken β-actin (CBA) promoter. All AAV2 viruses were produced by the Vector BioLabs. Sequences of MIF substrates, templates, and primers Sequences of MIF substrates, templates and primers used for shRNA constructs and point mutation constructs are provided in Table S1. Cell culture, transfection, lentiviral transduction, and cytotoxicity HeLa cells were cultured in Dulbecco's modified Eagle's medium (Invitrogen) supplemented with 10% fetal bovine serum (HyClone). V5-tagged MIF was transfected with Lipofectamine Plus (Invitrogen). Primary neuronal cultures from cortex were prepared as previously described (9). Briefly, the cortex was dissected and the cells were dissociated by trituration in modified Eagle's medium (MEM), 20% horse serum, 30 mM glucose, and 2 mM l-glutamine after a 10-min digestion in 0.027% trypsin/saline solution (Gibco-BRL). The neurons were plated on 15-mm multiwell plates coated with polyornithine or on coverslips coated with polyornithine. Neurons were maintained in MEM, 10% horse serum, 30 mM glucose, and 2 mM l-glutamine in a 7% CO2 humidified 37°C incubator. The growth medium was replaced twice per week. In mature cultures, neurons represent 70 to 90% of the total number of cells. Days in vitro (DIV) 7 to 9, neurons were infected by lentivirus carrying MIF-WT-Flag, MIF-E22Q-Flag, or MIF-E22A-Flag [1 × 109 units (TU)/ml] for 72 hours. Parthanatos was induced by either MNNG (Sigma) in HeLa cells or NMDA (Sigma) in neurons. HeLa cells were exposed to MNNG (50 μM) for 15 min, and neurons (DIV 10 to 14) were washed with control salt solution [CSS, containing 120 mM NaCl, 5.4 mM KCl, 1.8 mM CaCl2, 25 mM tris-Cl, and 20 mM glucose (pH 7.4)], exposed to 500 μmM NMDA plus 10 μmM glycine in CSS for 5 min, and then exposed to MEM containing 10% horse serum, 30 mM glucose, and 2 mM l-glutamine for various times before fixation, immunocytochemical staining, and confocal laser scanning microscopy. Cell viability was determined the following day by unbiased objective computer-assisted cell counting after staining of all nuclei with 7 μM Hoechst 33342 (Invitrogen) and dead cell nuclei with 2 μM propidium iodide (Invitrogen). The numbers of total and dead cells were counted with the Axiovision 4.6 software (Carl Zeiss). At least three separate experiments using at least six separate wells were performed with a minimum of 15,000 to 20,000 neurons or cells counted per data point. For neuronal toxicity assessments, glial nuclei fluoresced at a different intensity than neuronal nuclei and were gated out. The percentage of cell death was determined as the ratio of live to dead cells compared with the percentage of cell death in control wells to account for cell death attributed to mechanical stimulation of the cultures. Pull-down, coimmunoprecipitation, and immunoblotting For the pull-down assay, GST-tagged MIF or AIF proteins immobilized glutathione Sepharose beads were incubated with 500 μg of HeLa cell lysates, washed in the lysis buffer, and eluted in the protein loading buffer. For coimmunoprecipitation, 1 mg whole-cell lysates were incubated overnight with AIF antibody (1 μg/ml) in the presence of protein A/G Sepharose (Santa Cruz Biotechnology), followed by immunoblot analysis with mouse anti-Flag antibody (Clone M1, Sigma), mouse anti-V5 (V8012, Sigma) or Goat anti-MIF (ab36146, Abcam). The proteins were separated on denaturing SDS-PAGE and transferred to a nitrocellulose membrane. The membrane was blocked and incubated overnight with primary antibody (50 ng/ml; mouse anti-Flag; rabbit anti-AIF; or goat anti-MIF) at 4°C, followed by horseradish peroxidase (HRP)–conjugated donkey anti-mouse, anti-rabbit or anti-goat for 1 hour at RT. After washing, the immune complexes were detected by the SuperSignalWest Pico Chemiluminescent Substrate (Pierce). Subcellular fractionation The nuclear extracts (N) and postnuclear cell extracts (PN), which is the fraction prepared from whole-cell lysates after removing nuclear proteins, were isolated in hypotonic buffer (9, 11). The integrity of the nuclear and postnuclear subcellular fractions was determined by monitoring histone H3 or H4 and MnSOD or mitochondria antibody (MTC02) (Mito) immunoreactivity (9, 11). Immunocytochemistry, immunohistochemistry, and confocal microscopy For immunocytochemistry, cells were fixed 4 hours after MNNG or NMDA treatment with 4% paraformaldehyde, permeabilized with 0.05% Triton X-100, and blocked with 3% BSA in PBS. AIF was visualized by Donkey anti-Rabbit Cy3 or donkey anti-rabbit 647. MIF was visualized by donkey anti-mouse cy2 (2 μg/ml), donkey anti-goat Cy2 or donkey anti-goat 647. Immunohistochemistry was performed with an antibody against Flag. Immunofluorescence analysis was carried out with an LSM710 confocal laser scanning microscope (Carl Zeiss) as described (9). Quantification of relative percentage levels of AIF and MIF in subcellular fractions The relative levels of AIF and MIF in different fractions were quantified and calculated as the percentage of their total proteins based on the intensity of protein signals relative to the protein amount prepared from the same number of cells (6 million). The detailed calculation is as follows: 1) The signal intensity of each interested band was measured and normalized to their mitochondrial and nuclear markers, with the total proteins of CSS in Fig. 4, G to I and knockout neurons treated with NMDA in Fig. 5 E to G. A volume factor was used to calculate the relative amount of total protein (T), post-nuclear protein (PN) and nuclear protein (N) prepared from the same number (6 million) of cells. As such, the relative ratio of different samples in the same fractions and the same sample in different fractions will be calculated as the relative intensity of total protein (Ti), post-nuclear fraction (PNi) and nuclear fraction (Ni). 2) A Z factor for the adjusted total proteins for each sample was determined via Z = (PNi + Ni)/Ti. 3) Relative protein levels in PN and N fractions were calculated as follows: PN% = (PNi × Z)/ Ti × 100%; N% = (Ni × Z)/Ti × 100%; T% = (Ti × Z)/ Ti × 100%. FPLC The native state and purity of the purified recombinant MIF were determined using standard calibration curve between elution volume and molecular mass (kD) of known molecular weight native marker proteins in Akta Basic FPLC (Amersham-Pharmacia Limited) using Superdex 200 10/300GL column (GE Healthcare, Life Sciences). The gel filtration column was run in standard PBS buffer at a flow rate of 0.5 ml/min. The following molecular weight standards were used: Ferritin (440 kD), aldolase (158 kD), conalbumin (75 kD), ovalbumin (43 kD), carbonic anhydrase (29 kD), and ribonuclease (13.7 kD) respectively (GE Healthcare, Life Sciences). Eluted fractions containing MIF were resolved on 12% SDS-PAGE and stained with commassie blue to check the purity of the protein. Mass spectrometry analysis for MIF protein purity MIF proteins used for nuclease assays were also examined by mass spectrometry in order to exclude any possible contamination from other known nucleases. We performed analyses using different criteria at a 95% and lower confidence levels in order to capture any remote possibility of a nuclease. Analysis and search of the NCBI database using all species reveal that no known nuclease that can digest single or double-stranded DNA was detected in the MIF protein that was used in the nuclease assays. Circular dichroism (CD) spectroscopy CD spectroscopy was performed on a AVIV 420 CD spectrometer (Biomedical Inc., Lakewood, NJ, USA). Near-UV CD spectra were recorded between 240–320 nm using a quartz cuvette of 0.5 cm path length with protein samples at a concentration of 2 mg/ml at room temperature. Far UV CD spectra were also recorded at room temperature between 190–260 nm using quartz sandwich cuvettes of 0.1 cm path length with protein samples at a concentration of 0.2 mg/ml (48). The proteins were suspended in PBS buffer with or without magnesium chloride (5.0 mM) and/or zinc chloride (0.2 mM). The CD spectra were obtained from 0.5 nm data pitch, 1 nm/3 s scan speed and 0.5 s response time were selected for the recordings. Oxido-reductase activity assay The thiol-protein oxidoreductase activity of MIF was measured using insulin as the substrate as described previously (30). Briefly, the insulin assay is based on the reduction of insulin and subsequent insolubilization of the insulin β-chain. The time-dependent increase in turbidity is then measured spectrophotometrically at 650 nm. The reaction was started by adding 5 μM MIF to WT, E22A, E22Q, C57A;C60A or and P2G mutants dissolved in 20 mM sodium phosphate buffer (pH 7.2), and 200 mM reduced glutathione (GSH) to ice-cold reaction mixture containing 1 mg/ml insulin, 100 mM sodium phosphate buffer (pH 7.2) and 2 mM EDTA. MIF insulin reduction was measured against the control solution (containing GSH) in the same experiment. Tautomerase activity assay Tautomerase activity was measured using d-dopachrome tautomerase as the substrate as described previously (49). Briefly, a fresh solution of d-dopachrome methyl ester was prepared by mixing 2 mM l-3,4 dihydroxyphenylalanine methyl ester with 4 mM sodium peroxidate for 5 min at room temperature and then placed directly on ice before use. The enzymatic reaction was initiated at 25°C by adding 20 μl of the dopachrome methyl ester substrate to 200 μl of MIF WT, E22A, E22Q, C57A;C60A (final concentration 5 μM) or and P2G mutants prepared in tautomerase assay buffer (50 mM potassium phosphate, 1 mM EDTA, pH 6.0). The activity was determined by the semi-continuous reduction of OD 475 nm using a spectrophotometer. Quantification of noncleaved genomic DNA Noncleaved genomic DNA was quantified as percentage (%) of the total genomic DNA that included both noncleaved genomic DNA and cleaved genomic DNA in each individual group. Quantification of cells with AIF and MIF nuclear translocation Nuclear translocation of AIF and MIF was calculated as the percentage of total cells in each individual immunostained image. At least 5 to 12 images were quantified for each group. 500 or more neurons were counted for each condition. White indicates the overlay of AIF (red), MIF (green) and 4′,6′-diamidino-2-phenylindole (DAPI) (blue) suggesting the nuclear translocation of both AIF and MIF. Pink indicates the overlay of AIF (red) and DAPI (blue) suggesting the nuclear translocation of AIF only. Cyan indicates the overlay of MIF (green) and DAPI (blue) suggesting the nuclear translocation of MIF. Representative immunostaining images of MIF and AIF nuclear translocation were shown in Figs. 4E and 5C and fig. S17A. Intracerebroventricular (ICV) injection Three microliters of AAV2-MIF WT, E22Q and E22A (1 × 1013 GC/ml, Vector BioLabs) were injected into both sides of intracerebroventricular of the newborn MIF KO mice (41). The expression of MIF and its variants were checked by immunohistochemistry after MCAO surgery at 8–16 weeks of age. Neurobehavioral activity Spontaneous motor activity was evaluated 1 day, 3 days and 7 days after MCAO by placing the animals in a mouse cage for 5 min. A video camera was fitted on top of the cage to record the activity of a mouse in the cage. Neurological deficits were evaluated by an observer blinded to the treatment and genotype of the animals with a scale of 0–5 (0, no neurological deficit; 5, severe neurological deficit). The following criteria were used to score deficits: 0 = mice appeared normal, explored the cage environment and moved around in the cage freely; 1 = mice hesitantly moved in cage but could occasionally touch the walls of the cage, 2 = mice showed postural and movement abnormalities, and did not approach all sides of the cage, 3 = mice showed postural and movement abnormalities and made medium size circles in the center of cage, 4 = mice with postural abnormalities and made very small circles in place, 5 = mice were unable to move in the cage and stayed at the center. Recordings were evaluated by observers blinded to the treatment and genotype of the animals. The corner test was performed 1 day, 3 days and 7 days after MCAO to assess sensory and motor deficits following both cortical and striatal injury. A video camera was fitted on top of the cage to record the activity of a mouse in the cage for 5 min. The mice were placed between two cardboards each with a dimension of 30 cm × 20 cm × 0.5 mm attached to each other from the edges with an angle of 30°. Once in the corner, the mice usually rear and then turn either left or right. Before stroke mice do not show a side preference. Mice with sensory and motor deficits following stroke will turn toward the non-impaired side (right). Percent of right turn = right turns/total turns × 100 was calculated and compared. Recordings were evaluated by observers blinded to the treatment and genotype of the animals. Animals The Johns Hopkins Medical Institutions are fully accredited by the American Association for the Accreditation of Laboratory Animal Care (AAALAC). All research procedures performed in this study were approved the Johns Hopkins Medical Institutions Institutional Animal Care and Use Committee (IACUC) in compliance with the Animal Welfare Act regulations and Public Health Service (PHS) Policy. All animal studies were performed in a blinded fashion. Mouse genotype was determined by K.N. Stroke surgery was performed by R.A. Mouse genotypes were decoded after the stroke surgery, mouse behavior tests and data analysis. Based on their genotype, mice were grouped as WT, KO, KO-WT, KO-E22Q and KO-E22A. Within each group, mice were randomly assigned to subgroups including sham, 1 day-post stroke, 3 days- or 7 days-post stroke. Statistical analysis Unless otherwise indicated, statistical evaluation was carried out by Student's t test between two groups and by one-way analysis of variance (ANOVA) followed by post hoc comparisons with the Bonferroni test using GraphPad Prism software within multiple groups. Data are shown as means ± SEM. P <0.05 is considered significant. Experiments for quantification were performed in a blinded fashion. In order to ensure adequate power to detect the effect, at least 3 independent tests were performed for all molecular biochemistry studies and at least 5 mice from 3 different litters were used for animal studies. Supplementary Material 01 ACKNOWLEDGMENTS This work was supported by NIH grant K99/R00 NS078049, American Heart Association (AHA) National Scientist Development Grant 12SDG11900071, and the University of Texas Southwestern Medical Center Department of Pathology Startup funds and UT Rising Stars to Y.W.; and grants from National Institute on Drug Abuse, NIH, DA000266; and National Institute of Neurological Disorders and Stroke, NIH, R01 NS067525, R37 NS067525, and NS38377 to T.M.D. and V.L.D. T.M.D. is the Leonard and Madlyn Abramson Professor in Neurodegenerative Diseases. The authors acknowledge the joint participation by the Adrienne Helis Malvin Medical Research Foundation through its direct engagement in the continuous active conduct of medical research in conjunction with the Johns Hopkins Hospital and the Johns Hopkins University School of Medicine and the foundation's Parkinson's Disease Program M-2016. ChIP-seq data are archived in National Center for Biotechnology Information, NIH, GEO, GSE65110. Y.W. contributed to all aspects of the project. R.A., G.K.U., K.N., H.P., B.K., L.B., M.M.H., C.C., R.C., S.M.E., M.M.H., T.-I.K., S.N., G.M., and H.S. helped with some experiments. J.S.J., S.B., and H.Z. provided protein chips. Z.X. and J.Q. helped with the bioinformatics analysis. Y.W., V.L.D., and T.M.D. designed experiments and wrote the paper. The study was conceived and scientifically directed by V.L.D. and T.M.D. Y.W., H.P., V.L.D., and T.M.D. are inventors on a patent application submitted by Johns Hopkins University that covers the use of inhibitors of MIF nuclease activity to treat stroke and other disorders. The supplementary materials contain additional data. Fig. 1 Identification of MIF as a key cell-death effector mediating PARP-1–dependent cell death (A) Strategy for identifying AIF-associated proteins involved in PARP-1–dependent cell death. (B) siRNA-based PARP-1–dependent cell viability high-throughput screening in HeLa cells 24 hours after MNNG treatment (50 μM, 15 min); n = 8. The experiments were repeated in four independent tests ***P < 0.001, one-way ANOVA. (C) Schematic representation of MIF's PD-D/E(X)K domains. Single-letter abbreviations for the amino acid residues are as follows: A, Ala; C, Cys; D, Asp; E, Glu; F, Phe; G, Gly; H, His; I, Ile; K, Lys; L, Leu; M, Met; N, Asn; P, Pro; Q, Gln; R, Arg; S, Ser; T, Thr; V, Val; W, Trp; Y, Tyr; and X, any amino acid. (D) Alignment of the nuclease domain of human MIF and other nucleases. Arrows above the sequence indicate β strands and rectangles represent α helices. Amino acid residues that were mutated are indicated with an arrow and number (see Results). Nuclease and CxxCxxHx(n)C domains are highlighted in green and pink, respectively. (E) Crystal structure of MIF trimer (pdb:1GD0) (left) and MIF PD-D/E(x)K motif in trimer (right). Fig. 2 MIF is a nuclease that cleaves genomic DNA (A) In vitro MIF (2 μM) nuclease assay with pcDNA as substrate. (B) In vitro pulsed-field gel electrophoresis MIF (4 μM) nuclease assay with human genomic DNA (hgDNA) as a substrate in buffer containing Mg2+ (10 mM) with or without EDTA (50 mM) or Ca2+ (2 mM) with or without EDTA (25 mM). (C) Pulsed-field gel electrophoresis assay of MNNG-induced DNA damage in MIF-deficient HeLa cells and wild-type (WT) HeLa cells treated with or without DPQ (30 μM) or ISO-1 (100 μM). NT/MNNG (nontargeting shRNA/MNNG). (D) Nuclease assay of MIF WT and mutants (4 μM) using human genomic DNA as the substrate. Fig. 3 MIF binds and cleaves ssDNA (A) MIF DNA binding motif determined by ChIP-seq. (B) Binding of MIF to biotin-labeled small DNA substrates with different structures or different sequences in an EMSA assay (see fig. S8 for illustrations of substrates, and tables S1 and S2 for sequences). Arrow indicates the DNA-MIF complex. Asterisk indicates nonspecific bands. PCS, positive control substrate from the LightShift Chemiluminescent EMSA Kit (Thermo Scientific) containing 60 base pairs (bp) of 5′ biotin-labeled duplex. With or without BSA, bovine serum albumin; PC, Epstein-Barr nuclear antigen extract from the LightShift Chemiluminescent EMSA Kit or MIF. (C) MIF cleavage of unpaired bases at the 3′ end of the stem loop of 5′ or 3′ biotin-labeled small DNA substrates with various structures or sequences in a nuclease assay (see fig. S8 for illustrations of substrates, and tables S1 and S2 for sequences). Experiments were replicated four times with three independent preparations of MIF protein. (D) MIF cleavage of 3′ unpaired bases from nonlabeled PS30 and 3F1 substrates. DNA ladders 1 and 2 were customized with PS30 and its cleavage products by removing its 3′ nucleotides one by one. DNA ladder 1 was prepared using PS30, PS28, PS26, PS24, PS22, and PS20. DNA ladder 2 was prepared using PS29, PS27, PS25, PS23, and PS21. (E) MIF cleavage sites on nonlabeled PS30 and 3F1 substrates. Fig. 4 Requirement of AIF for the recruitment of MIF to the nucleus in NMDA excitotoxicity (A) Binding of immobilized GST-MIF WT and GST-MIF variants to AIF. (B) Nuclease activity and AIF-binding activity of MIF WT and MIF variants. (C and D) Coimmunoprecipitation (IP) of MIF and AIF in control (CSS) and NMDA-treated cortical neurons. Asterisk indicates IgG. Ab, antibody. (D) Intensity of treated versus untreated cultures. *P < 0.05, Student's t test. (E) Images of nuclear translocation of AIF and MIF after NMDA treatment in WT, AIF knockdown, and MIF knockdown cortical neurons. AIF shRNA (AIF sh) and MIF shRNA (MIF sh) caused a 71.3 ± 5.2% and 73.3 ± 6.1% protein reduction, respectively. White color indicates the overlay of AIF, MIF, and 4′,6′-diamidino-2-phenylindole (DAPI), showing the nuclear translocation of AIF and MIF. Purple color indicates the overlay of AIF and DAPI, showing the nuclear translocation of AIF. Z stacks illustrating the x,z and y,z axis are provided to demarcate the nucleus. Arrowheads indicate cells with the high magnification. (F) Quantification of the percentage of cells with nuclear translocation of MIF and AIF after NMDA treatment in WT, AIF knockdown, and MIF knockdown cortical neurons. CSS, control salt solution. ****P < 0.0001, versus its CSS control; ####P < 0.0001, versus its WT treated with NMDA, one-way ANOVA. (G) Immunoblots of nuclear translocation of AIF and MIF after NMDA treatment in WT, AIF knockdown, and MIF knockdown cortical neurons. Compare total protein (T), post-nuclear fraction (PN), nuclear fraction (N), and Mito, mitochondrial antibody. (H and I) Relative levels of AIF and MIF in T, PN, and N. Means ± SEM. Experiments were replicated at least three times; ****P < 0.0001, versus its CSS control; ##P < 0.01, ###P < 0.0001, versus its total protein, one-way ANOVA. Fig. 5 MIF E22A mutant prevents AIF's recruitment of MIF to the nucleus in NMDA-excitotoxicity (A) Expression of MIF in WT and knockout (KO) neurons. (B) Coimmunoprecipitation of Flag-tagged MIF variants and AIF in cortical neurons after NMDA treatment. (C) Images of nuclear translocation of AIF and exogenous MIF WT and MIF variants after NMDA treatment in MIF KO cortical neurons. Scale bar, 20 μm. White color indicates the overlay of AIF, MIF, and DAPI, showing the nuclear translocation of AIF and MIF. Purple color indicates the overlay of AIF and DAPI, showing the nuclear translocation of AIF. Z stacks illustrating the x,z and y,z axis are provided to demarcate the nucleus. (D) Quantification of the percentage of cells with nuclear translocation of AIF and exogenous MIF WT and MIF variants after NMDA treatment in MIF KO cortical neurons. ****P < 0.0001, versus KO group; ###P < 0.001, versus KO-WT group, one-way ANOVA. (E) Immunoblots of nuclear translocation of AIF and exogenous MIF WT and MIF variants after NMDA treatment in MIF KO cortical neurons. H4, histone H4; mito, mitochondrial antibody. (F and G) Relative levels of AIF and MIF in total protein (T), postnuclear fraction (PN) and nuclear fraction (N). Means ± SEM. Experiments were replicated at least three times. **P <0.01, ****P < 0.001, versus KO control group; ##P < 0.01, ####P < 0.0001, versus its total protein, one-way ANOVA. Fig. 6 MIF nuclease activity is critical for DNA damage and PARP-1–dependent cell death in cortical neurons (A) Representative images and (B) quantification of NMDA-induced (500 μM for 5 min) excitotoxicity in MIF WT, KO, and KO cortical neurons expressing MIF WT, E22Q, E22A, or P2G. Scale bar, 200 μm. (C) Representative images and (D to F) quantification of NMDA-induced DNA damage 6 hours after treatment determined by the comet assay in MIF WT, KO, and KO neurons expressing MIF WT, E22Q, E22A, or P2G. Dashed lines indicate the center of the head and tail. Scale bar, 20 μm. (G) Pulsed-field gel electrophoresis assay of NMDA-induced DNA damage 6 hours after treatment in MIF WT and KO neurons and KO neurons expressing MIF WT, E22Q, E22A, or P2G. Means ± SEM are shown in (B), (D), (E), and (F). *P < 0.05, **P < 0.01, ***P < 0.001, one-way ANOVA; ns, nonsignificant. Fig. 7 MIF nuclease activity is critical for DNA damage and ischemic neuronal cell death in vivo (A) Representative images of triphenyl tetrazolium chloride staining of MIF WT, KO, and KO mice that were injected with AAV2-MIF WT, E22Q, or E22A 24 hours after 45 min of middle cerebral artery occlusion (MCAO). (B to D) Quantification of infarction volume in cortex, striatum, and hemisphere 1 day or 7 days after 45-min MCAO. WT MCAO (n = 29); KO MCAO (n = 20); KO-WT MCAO (n = 23). KO-E22Q (n = 22) and KO-E22A MCAO (n = 19). *P < 0.05, ***P < 0.001, versus KO group at the same time point; ##P < 0.01, ###P < 0.001, the same group at 7 days versus at 1 day after 45-min MCAO, one-way ANOVA. (E to G) Neurological deficit was evaluated by [(E) and (F)] open field on a scale of 0 to 5 and (G) corner test evaluated by percentage of right turns at 1 day, 3 days, or 7 days after MCAO surgery. WT MCAO (n = 16); KO MCAO (n = 12); and KO-WT MCAO (n = 16). KO-E22Q MCAO (n = 16) and KO-E22A MCAO (n = 16). Means ± SEM. *P < 0.05, ***P < 0.001, one-way ANOVA in (E) and (G). **P < 0.01, two-way ANOVA in (F), WT and KO-WT versus KO, KO-E22Q, and KO-E22A at different time points. (H) DNA fragmentation determined by pulsed-field gel electrophoresis in the penumbra 1 day, 3 days, or 7 days after 45-min MCAO surgery in MIF WT, KO, and KO mutant mice, which were injected with AAV2-MIF WT, E22Q, or E22A. WT MCAO (n = 15); KO MCAO (n = 15); and KO-WT MCAO (n = 15). KO-E22Q (n = 15) and KO-E22A MCAO (n = 15). (I) Quantification of noncleaved genomic DNA. Means ± SEM. ****P < 0.0001, versus its sham treatment group, one-way ANOVA. SUPPLEMENTARY MATERIALS www.sciencemag.org/content/354/6308/aad6872/suppl/DC1 REFERENCES AND NOTES 1 Bai P Biology of poly(ADP-ribose) polymerases: The factotums of cell maintenance Mol. Cell 2015 58 947 958 doi: 10.1016/j.molcel.2015.01.034; pmid: 26091343 26091343 2 Fatokun AA Dawson VL Dawson TM Parthanatos: Mitochondrial-linked mechanisms and therapeutic opportunities Br. J. Pharmacol 2014 171 2000 2016 doi: 10.1111/bph.12416; pmid: 24684389 24684389 3 Wang Y Dawson VL Dawson TM Poly(ADP-ribose) signals to mitochondrial AIF: A key event in parthanatos Exp. Neurol 2009 218 193 202 doi: 10.1016/j.expneurol.2009.03.020; pmid: 19332058 19332058 4 Pacher P Szabo C Role of the peroxynitrite-poly(ADP-ribose) polymerase pathway in human disease Am. J. Pathol 2008 173 2 13 doi: 10.2353/ajpath.2008.080019; pmid: 18535182 18535182 5 Szabó C Dawson VL Role of poly(ADP-ribose) synthetase in inflammation and ischaemia-reperfusion Trends Pharmacol. Sci 1998 19 287 298 doi: 10.1016/S0165-6147(98)01193-6; pmid: 9703762 9703762 6 Martire S Mosca L d'Erme M PARP-1 involvement in neurodegeneration: A focus on Alzheimer's and Parkinson's diseases Mech. Ageing Dev 2015 146–148 53 64 doi: 10.1016/j.mad.2015.04.001; pmid: 25881554 7 Virág L Robaszkiewicz A Rodriguez-Vargas JM Oliver FJ Poly(ADP-ribose) signaling in cell death Mol. Aspects Med 2013 34 1153 1167 doi: 10.1016/j.mam.2013.01.007; pmid: 23416893 23416893 8 Wang H Apoptosis-inducing factor substitutes for caspase executioners in NMDA-triggered excitotoxic neuronal death J. Neurosci 2004 24 10963 10973 doi: 10.1523/JNEUROSCI.3461-04.2004; pmid: 15574746 15574746 9 Wang Y Poly(ADP-ribose) (PAR) binding to apoptosis-inducing factor is critical for PAR polymerase-1–dependent cell death (parthanatos) Sci. Signal 2011 4 ra20 doi: 10.1126/scisignal.2000902; pmid: 21467298 21467298 10 Yu SW Apoptosis-inducing factor mediates poly(ADP-ribose) (PAR) polymer-induced cell death Proc. Natl. Acad. Sci. U.S.A 2006 103 18314 18319 doi: 10.1073/pnas.0606528103; pmid: 17116881 17116881 11 Yu SW Mediation of poly(ADP-ribose) polymerase-1–dependent cell death by apoptosis-inducing factor Science 2002 297 259 263 doi: 10.1126/science.1072221; pmid: 12114629 12114629 12 Wang X Yang C Chai J Shi Y Xue D Mechanisms of AIF-mediated apoptotic DNA degradation in Caenorhabditis elegans Science 2002 298 1587 1592 doi: 10.1126/science.1076194; pmid: 12446902 12446902 13 David KK Sasaki M Yu SW Dawson TM Dawson VL EndoG is dispensable in embryogenesis and apoptosis Cell Death Differ 2006 13 1147 1155 doi: 10.1038/sj.cdd.4401787; pmid: 16239930 16239930 14 Irvine RA Generation and characterization of endonuclease G null mice Mol. Cell. Biol 2005 25 294 302 doi: 10.1128/MCB.25.1.294-302.2005; pmid: 15601850 15601850 15 Xu Z Endonuclease G does not play an obligatory role in poly(ADP-ribose) polymerase-dependent cell death after transient focal cerebral ischemia Am. J. Physiol. Regul. Integr. Comp. Physiol 2010 299 R215 R221 doi: 10.1152/ajpregu.00747.2009; pmid: 20427721 20427721 16 Hu S Profiling the human protein-DNA interactome reveals ERK2 as a transcriptional repressor of interferon signaling Cell 2009 139 610 622 doi: 10.1016/j.cell.2009.08.037; pmid: 19879846 19879846 17 Calandra T Protection from septic shock by neutralization of macrophage migration inhibitory factor Nat. Med 2000 6 164 170 doi: 10.1038/72262; pmid: 10655104 10655104 18 Merk M Mitchell RA Endres S Bucala R d-Dopachrome tautomerase (d-DT or MIF-2): Doubling the MIF cytokine family Cytokine 2012 59 10 17 doi: 10.1016/j.cyto.2012.03.014; pmid: 22507380 22507380 19 Kosinski J Feder M Bujnicki JM The PD-(D/E)XK superfamily revisited: Identification of new members among proteins involved in DNA metabolism and functional predictions for domains of (hitherto) unknown function BMC Bioinformatics 2005 6 172 doi: 10.1186/1471-2105-6-172; pmid: 16011798 16011798 20 Kratz K Deficiency of FANCD2-associated nuclease KIAA1018/FAN1 sensitizes cells to interstrand crosslinking agents Cell 2010 142 77 88 doi: 10.1016/j.cell.2010.06.022; pmid: 20603016 20603016 21 MacKay C Identification of KIAA1018/FAN1, a DNA repair nuclease recruited to DNA damage by monoubiquitinated FANCD2 Cell 2010 142 65 76 doi: 10.1016/j.cell.2010.06.021; pmid: 20603015 20603015 22 Sugimoto H Suzuki M Nakagawa A Tanaka I Nishihira J Crystal structure of macrophage migration inhibitory factor from human lymphocyte at 2.1 A resolution FEBS Lett 1996 389 145 148 doi: 10.1016/0014-5793(96)00553-4; pmid: 8766818 8766818 23 Sun HW Bernhagen J Bucala R Lolis E Crystal structure at 2.6-A resolution of human macrophage migration inhibitory factor Proc. Natl. Acad. Sci. U.S.A 1996 93 5191 5196 doi: 10.1073/pnas.93.11.5191; pmid: 8643551 8643551 24 Suzuki M Crystal structure of the macrophage migration inhibitory factor from rat liver Nat. Struct. Biol 1996 3 259 266 doi: 10.1038/nsb0396-259; pmid: 8605628 8605628 25 Laganeckas M Margelevicius M Venclovas C Identification of new homologs of PD-(D/E)XK nucleases by support vector machines trained on data derived from profile-profile alignments Nucleic Acids Res 2011 39 1187 1196 doi: 10.1093/nar/gkq958; pmid: 20961958 20961958 26 Steczkiewicz K Muszewska A Knizewski L Rychlewski L Ginalski K Sequence, structure and functional diversity of PD-(D/E)XK phosphodiesterase superfamily Nucleic Acids Res 2012 40 7016 7045 doi: 10.1093/nar/gks382; pmid: 22638584 22638584 27 Pingoud V On the divalent metal ion dependence of DNA cleavage by restriction endonucleases of the EcoRI family J. Mol. Biol 2009 393 140 160 doi: 10.1016/j.jmb.2009.08.011; pmid: 19682999 19682999 28 Lubetsky JB The tautomerase active site of macrophage migration inhibitory factor is a potential target for discovery of novel anti-inflammatory agents J. Biol. Chem 2002 277 24976 24982 doi: 10.1074/jbc.M203220200; pmid: 11997397 11997397 29 Fingerle-Rowson G A tautomerase-null macrophage migration-inhibitory factor (MIF) gene knock-in mouse model reveals that protein interactions and not enzymatic activity mediate MIF-dependent growth regulation Mol. Cell. Biol 2009 29 1922 1932 doi: 10.1128/MCB.01907-08; pmid: 19188446 19188446 30 Kudrin A Human macrophage migration inhibitory factor: A proven immunomodulatory cytokine? J. Biol. Chem 2006 281 29641 29651 doi: 10.1074/jbc.M601103200; pmid: 16893895 16893895 31 Rosengren E The macrophage migration inhibitory factor MIF is a phenylpyruvate tautomerase FEBS Lett 1997 417 85 88 doi: 10.1016/S0014-5793(97)01261-1; pmid: 9395080 9395080 32 Machanick P Bailey TL MEME-ChIP: Motif analysis of large DNA datasets Bioinformatics 2011 27 1696 1697 doi: 10.1093/bioinformatics/btr189; pmid: 21486936 21486936 33 Zhao G Zhao B Tong Z Mu R Guan Y Effects of 2′-O-methyl nucleotide substitution on EcoRI endonuclease cleavage activities PLOS ONE 2013 8 e77111 doi: 10.1371/journal.pone.0077111; pmid: 24194862 24194862 34 Inácio AR Ruscher K Leng L Bucala R Deierborg T Macrophage migration inhibitory factor promotes cell death and aggravates neurologic deficits after experimental stroke J. Cereb. Blood Flow Metab 2011 31 1093 1106 doi: 10.1038/jcbfm.2010.194; pmid: 21063426 21063426 35 Bacher M MIF expression in the rat brain: Implications for neuronal function Mol. Med 1998 4 217 230 pmid: 9606175 9606175 36 Zhang W Expression of macrophage migration inhibitory factor in the mouse neocortex and posterior piriform cortices during postnatal development Cell. Mol. Neurobiol 2014 34 1183 1197 doi: 10.1007/s10571-014-0094-1; pmid: 25118614 25118614 37 Calandra T Roger T Macrophage migration inhibitory factor: A regulator of innate immunity Nat. Rev. Immunol 2003 3 791 800 doi: 10.1038/nri1200; pmid: 14502271 14502271 38 Hudson JD A proinflammatory cytokine inhibits p53 tumor suppressor activity J. Exp. Med 1999 190 1375 1382 doi: 10.1084/jem.190.10.1375; pmid: 10562313 10562313 39 Kleemann R Intracellular action of the cytokine MIF to modulate AP-1 activity and the cell cycle through Jab1 Nature 2000 408 211 216 doi: 10.1038/35041591; pmid: 11089976 11089976 40 Andrabi SA Iduna protects the brain from glutamate excitotoxicity and stroke by interfering with poly(ADP-ribose) polymer-induced cell death Nat. Med 2011 17 692 699 doi: 10.1038/nm.2387; pmid: 21602803 21602803 41 Glascock JJ Delivery of therapeutic agents through intracerebroventricular (ICV) and intravenous (IV) injection in mice J. Vis. Exp 2011 2001 2968 pmid: 21988897 42 Chen Y Systematic evaluation of factors influencing ChIP-seq fidelity Nat. Methods 2012 9 609 614 doi: 10.1038/nmeth.1985; pmid: 22522655 22522655 43 Feng J Liu T Qin B Zhang Y Liu XS Identifying ChIP-seq enrichment using MACS Nat. Protoc 2012 7 1728 1740 doi: 10.1038/nprot.2012.101; pmid: 22936215 22936215 44 Drew HR Structure of a B-DNA dodecamer: Conformation and dynamics Proc. Natl. Acad. Sci. U.S.A 1981 78 2179 2183 doi: 10.1073/pnas.78.4.2179; pmid: 6941276 6941276 45 Masuda T Ito Y Terada T Shibata T Mikawa T A non-canonical DNA structure enables homologous recombination in various genetic systems J. Biol. Chem 2009 284 30230 30239 doi: 10.1074/jbc.M109.043810; pmid: 19729448 19729448 46 Ghoorah AW Devignes MD Smaïl-Tabbone M Ritchie DW Protein docking using case-based reasoning Proteins 2013 81 2150 2158 doi: 10.1002/prot.24433; pmid: 24123156 24123156 47 Macindoe G Mavridis L Venkatraman V Devignes MD Ritchie DW HexServer: An FFT-based protein docking server powered by graphics processors Nucleic Acids Res 2010 38 W445 W449 Web Server doi: 10.1093/nar/gkq311; pmid: 20444869 20444869 48 Reddy PP Molecular dynamics of the neuronal EF-hand Ca2+-sensor Caldendrin PLOS ONE 2014 9 e103186 doi: 10.1371/journal.pone.0103186; pmid: 25058677 25058677 49 Bendrat K Biochemical and mutational investigations of the enzymatic activity of macrophage migration inhibitory factor Biochemistry 1997 36 15356 15362 doi: 10.1021/bi971153a; pmid: 9398265 9398265

PMC005xxxxxx/PMC5135007.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 0376422 6405 Pediatrics Pediatrics Pediatrics 0031-4005 1098-4275 24515517 5135007 10.1542/peds.2013-2570 HHSPA831836 Article Epidemiology of Tuberculosis in Young Children in the United States Pang Jenny MD, MPH a Teeter Larry D. PhD b Katz Dolly J. PhD c Davidow Amy L. PhD d Miranda Wilson MPH e Wall Kirsten MHS f Ghosh Smita MS c Stein-Hart Trudy MS g Restrepo Blanca I. PhD h Reves Randall MD, MSc f Graviss Edward A. PhD, MPH b on behalf of the Tuberculosis Epidemiologic Studies Consortium a Department of Epidemiology, University of Washington, Seattle, Washington b Department of Pathology and Genomic Medicine, Center for Molecular and Translational Human Infectious Diseases Research, Houston Methodist Research Institute, Houston, Texas c Division of TB Elimination, Centers for Disease Control and Prevention, Atlanta, Georgia d Global TB Institute and Department of Preventive Medicine and Community Health, New Jersey Medical School at Rutgers, The State University of New Jersey, Newark, New Jersey e New York State Department of Health, Albany, New York f Denver Metro Tuberculosis Control Program, Denver Public Health Department, Denver, Colorado g Tennessee Department of Health TB Elimination Program, Nashville, Tennessee h School of Public Health in Brownsville, University of Texas Health Science Center at Houston, Brownsville, Texas Address correspondence to Larry D. Teeter, PhD, Houston Methodist Research Institute, 6565 Fannin St, Mail Station: MGJ3-010, Houston, TX 77030. ldteeter@houstonmethodist.org 28 11 2016 10 2 2014 3 2014 02 12 2016 133 3 e494e504 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. OBJECTIVES To estimate tuberculosis (TB) rates among young children in the United States by children’s and parents’ birth origins and describe the epidemiology of TB among young children who are foreign-born or have at least 1 foreign-born parent. METHODS Study subjects were children <5 years old diagnosed with TB in 20 US jurisdictions during 2005–2006. TB rates were calculated from jurisdictions’ TB case counts and American Community Survey population estimates. An observational study collected demographics, immigration and travel histories, and clinical and source case details from parental interviews and health department and TB surveillance records. RESULTS Compared with TB rates among US-born children with US-born parents, rates were 32 times higher in foreign-born children and 6 times higher in US-born children with foreign-born parents. Most TB cases (53%) were among the 29% of children who were US born with foreign-born parents. In the observational study, US-born children with foreign-born parents were more likely than foreign-born children to be infants (30% vs 7%), Hispanic (73% vs 37%), diagnosed through contact tracing (40% vs 7%), and have an identified source case (61% vs 19%); two-thirds of children were exposed in the United States. CONCLUSIONS Young children who are US born of foreign-born parents have relatively high rates of TB and account for most cases in this age group. Prompt diagnosis and treatment of adult source cases, effective contact investigations prioritizing young contacts, and targeted testing and treatment of latent TB infection are necessary to reduce TB morbidity in this population. tuberculosis epidemiology young children foreign birth Tuberculosis (TB) in children <5 years old (young children) is a sentinel event because it represents recent or ongoing disease transmission rather than reactivation of remotely acquired infection; as such, it points to deficiencies in TB control.1 In the United States, an average 416 TB cases were reported annually in young children from 2007 to 2011.2 Recent exposure to an adult with TB disease (the source case) is the primary mode of transmission to young children.3,4 In the United States, although >60% of all TB cases occur among foreign-born persons, 90% of cases in young children occur among US-born persons.5–7 Approximately two-thirds of US-born young children with TB have foreign-born parents, suggesting that this is an important population for prevention efforts.8,9 However, to date, no studies have calculated TB rates in young children by country of birth and parents’ origins. We used a multistate, cross-sectional observational study to calculate TB rates in young children in 3 subgroups: (1) foreign-born children, (2) US-born children with at least 1 foreign-born parent, and (3) US-born children of US-born parents. To improve understanding of TB in children who are foreign born or have foreign-born parents, we describe the children’s epidemiologic and clinical characteristics and characteristics of source cases. METHODS This cross-sectional population-based study of TB in children is part of a general investigation of TB among foreign-born persons in the United States,10 conducted in 2005–2007 by the Centers for Disease Control and Prevention (CDC)-funded Tuberculosis Epidemiologic Studies Consortium (TBESC).11 Each of the consortium’s 20 US sites was located in or partnered with a state or municipal health department (Table 1). All sites obtained institutional review board approval; for the observational study, the child’s parent or guardian provided written informed consent. The US Department of Health and Human Services provided a Certificate of Confidentiality to each site to protect participants’ sensitive information from involuntary disclosure. Additional information on measures used to protect confidentiality and encourage participation has been published.10 Study Population Study sites collected TB case counts for all young children during the study’s recruitment period (Table 1) and queried health department staff or the child’s parents to determine parental nativity. Children were eligible for the observational study if they were (1) <5 years old at TB treatment initiation, (2) foreign born or US born with at least 1 foreign-born parent, and (3) living in a TBESC jurisdiction. Sites began and ended the study at different times, based on timing of institutional review board approvals and recruitment rates. Recruitment began at all sites in 2005 and continued a median of 18 months (range = 13–25 months; Table 1). Additional recruitment details have been published.10 CDC’s National Tuberculosis Surveillance System was queried to determine what proportion of TB cases in young children were reported from the study jurisdictions in 2005 and 2006. Data Collection Data for the observational study were derived from (1) a structured in-person interview with a child’s parent or guardian for demographic and visa information, (2) routine TB surveillance reports for clinical variables, and (3) local health department resources for information on source cases. The source case is a person with TB who transmitted the infection to a child. When a young child is diagnosed with TB, the local health department attempts to identify the source case to break the chain of transmission. All source cases identified by participating health departments were offered enrollment and interviewed with a standard questionnaire. The parent’s or guardian’s English proficiency was assessed with a US Census question that has been shown to correlate strongly with scores on standardized English proficiency tests: “I would like to know how well you speak English. Would you say you speak English very well, well, not well, or not at all?”10,12 Persons who responded that they spoke English “very well” were interviewed in English unless they requested an interpreter; all others were interviewed in their native language by using bilingual interviewers and interpreters. All questionnaires and consent documents were translated from English into 10 languages that, together with English, accommodated almost 90% of participants; for the remainder, professional translation occurred at the time of interview.10 Calculation of TB Rates To account for differences in recruitment periods, each of the site’s total number of TB cases in children <5 years old was annualized to estimate the average number of cases in a 12-month period. Population information on children <5 years old by children’s and parents’ origin (US-born children of US-born parents, US-born children with at least 1 foreign-born parent, and foreign-born children) for all counties in the study’s catchment area was obtained by special request to the US Census Bureau’s American Community Survey for 2005–2007 (see Supplemental Table 6). TB incidence was determined by dividing the annualized number of TB cases in each subgroup by the average annual population census estimates obtained over a 3-year period (2005–2007) for all study sites; 95% confidence intervals (CIs) were calculated by Poisson distribution methods.13,14 To assess the impact of excluding children with unknown parental birth country, the annualized number of such children was added to each US-born category, and TB incidence and 95% CI recalculated. Analysis of Data From Observational Study US-born and foreign-born children were compared by clinical and sociodemographic variables by using Pearson’s χ2 or Fisher’s exact tests. Analyses were conducted with Stata 10.0 (Stata Corp, College Station, TX), SAS 9.3 (SAS Institute, Inc, Cary, NC), and Microsoft Excel 2007 (Microsoft, Inc, Redmond, WA). RESULTS A total of 364 TB cases in children <5 years old were identified in the 20 US study sites (Table 1); this represented 49.6% of all TB cases in young children reported to the National Tuberculosis Surveillance System in 2005 and 2006. More than 80% (n = 303) of the children were US born, including 194 (64%) with at least 1 foreign-born parent, 76 (25%) with US-born parents, and 33 (11%) whose parents’ birth countries were unknown. Sixty-one children (17%) were foreign born. TB Rate Estimates Estimated TB rates per 100 000 population for children <5 years old were 2.57 for all children, 24.03 for foreign-born children, 4.81 for US-born children with at least 1 foreign-born parent, and 0.75 for US-born children of US-born parents. More than half of the cases (53%) occurred among US-born children with foreign-born parents (Table 2). When the 33 children with unknown parental birthplaces were included in the group with foreign-born parents, that group’s TB rate increased to 5.61; including them in the group with US-born parents increased that rate to 1.08. Observational Study Results: Demographics A total of 255 children (61 foreign born, 194 US born) were eligible for the observational study, of whom 149 (58%) were enrolled: 27 (44%) foreign born and 122 (63%) US born; US born children were more likely to be enrolled (P = .01). Reasons for nonenrollment included parental refusal, >6 months elapsed since treatment start, inability to contact the child’s parents, or the child moved; case counts by nonenrollment reason were not provided by study sites. Two-thirds of enrolled children were Hispanic; the median age was 2 years. The foreign-born children were from Mexico (9), Central America/South America/Caribbean (5), sub-Saharan Africa (5), South Asia/East Asia/Pacific (5), and Eastern Europe/Central Asia (3); 6 (22%) were adopted by US-born parents. Adoptees were from Guatemala, China, Ethiopia, and Kazakhstan, and 2 were from the Russian Federation. US-born children were more likely than foreign-born children to be Hispanic and to be diagnosed before their first birthday (Table 3). Of the 143 children with at least 1 foreign-born parent, more than half the parents (57%) were from Mexico. Almost half (48%) of the foreign-born parents reported they were undocumented at time of US entry, and 31% spoke no English. Fifteen (10%) of the 149 parents said “yes” when asked if they feared they or their children would be deported when they took the children for TB treatment. Care-Seeking Behavior US-born children were more likely to have identified source cases, have medical insurance, and be diagnosed through medical evaluation for symptoms; foreign-born children were more likely to be diagnosed through medical screenings (Table 3). Of the 70 children without identified source cases, 37 (53%) were diagnosed during medical visits for symptoms and 33 (47%) during screenings, most commonly well-child examinations (n = 18). Of the 79 children with identified source cases, 51 (65%) were diagnosed through the source case’s contact investigation (n = 47) or evaluation for a known TB exposure (n = 4), 24 (30%) were diagnosed during medical visits for symptoms, and 4 (5%) from well-child examinations. Clinical Findings Forty-seven children (32%) had some extrapulmonary TB; extrapulmonary involvement was more common in symptomatic children (40% vs 14%; P < .01). Meningeal and miliary presentations were more common in infants (24% vs 5%; P < .01; data not shown). Of the 140 children who received tuberculin skin tests, 89% were positive. All but 1 of 45 culture-confirmed TB cases occurred in US-born children (Table 3). Of the 44 with isoniazid, rifampin, and ethambutol drug susceptibility testing (DST), 1 (2.3%) had multidrug-resistant (MDR) TB (resistance to both isoniazid and rifampin) and 5 (11.4%) had isoniazid-resistant, rifampin-susceptible TB. Three children who were only clinically confirmed had source cases with MDR TB. Of the 35 isolates with pyrazinamide (PZA) DST, 5 (14.3%) were resistant. All children with PZA-resistant TB were US born with a Mexican-born parent (33% PZA resistance among the 15 children of Mexican-born parents with PZA DST results) and had extrapulmonary involvement. Those who had eaten unpasteurized dairy products were more likely to have PZA resistance: 60% vs 7% (P = .02). Ninety-nine children (66%) had TB symptoms; most commonly fever (56%) and cough (54%) (Table 4). A median 52 days elapsed between symptom onset and TB treatment initiation: 44.5 for US-born and 115 for foreign-born children (Table 4). A median 2 clinical visits occurred before treatment initiation; 15 of 99 symptomatic participants had >4 visits (maximum 7). Source of Infection Conservatively, 102 (68%) of the participants contracted TB in the United States (Table 3), 79 with known source cases and 23 with no foreign travel (4 with foreign visitors in the previous year and 19 without) (Fig 1). Thirty-six children (24%) had TB transmission risks only outside the United States (21 foreign born with no other identified risk factors, 15 with foreign travel in the 2 years before diagnosis). Eleven children (7%) had foreign visitors and foreign travel. US-born children were nearly 10 times as likely as foreign-born children to have been infected in the United States (Table 3). Source Case Characteristics Seventy-nine children (53%) in the observational study had known source cases. Fifty-one (65%) of the source cases lived in the child’s household. The source case was diagnosed first for 54 (68%) children and the child was diagnosed first for 16 (20%). The timing of diagnosis relative to the source was unknown for 9 children, including 2 in whom the epidemiologic linkage was confirmed through Mycobacterium tuberculosis genotyping. Source cases for 52 (66%) of the 79 children were enrolled in the study; 6 of the source cases infected 2 different enrolled children each. Among the 46 unique enrolled source cases (Table 5), 42 (91%) were foreign born, including at least 38% who were undocumented at diagnosis. The median time from US entry to TB treatment initiation was 5.3 years (interquartile range [IQR] 2.7–13.1). Most source cases reported cough (91%) and were acid-fast bacilli sputum smear positive (89%) or M tuberculosis sputum culture positive (98%). For 3 enrolled US-born children with only clinically confirmed TB, the 2 foreign-born source cases had MDR TB. The median time from the source case’s symptom onset to the source case’s treatment initiation was 105 days (IQR 72–190 days), to the child’s treatment initiation was 132 days (IQR 101–212 days, Table 5). DISCUSSION TB in young children is of particular clinical and programmatic concern because (1) such cases are markers for recent or ongoing transmission, (2) young children are more likely to progress from infection to disease, and (3) young children are more likely to develop severe manifestations of TB, such as meningeal or miliary disease.15,16 Foreign-born children have long been known to be at high risk of TB and have been a focus of prevention efforts.17,18 But the risk of TB among children of foreign-born parents has been harder to define because of the difficulty of obtaining parental birth information and population denominators. This is the first US study to calculate TB rates among children younger than 5 years by national origin of both the children and their parents. The study found that the TB rate for young children who are US born but have at least 1 foreign-born parent was more than 6 times that of US-born young children with US-born parents. Although foreign-born children had higher TB rates, they accounted for only 1.8% of the total population of children <5 years old and 17% of TB cases in young children. Most TB cases (53%) in young children were reported among the 29% of the US population of young children with at least 1 foreign-born parent. This pediatric population should be a focus of TB prevention and control efforts. Routine ascertainment of children’s and parents’ countries of birth can help identify children at risk for both TB and latent TB infection (LTBI). It took a median of 52 days to initiate TB therapy for the symptomatic children in this study. Unfortunately, symptoms such as fever and cough are also seen in the more common viral upper respiratory infections, making it difficult to select young children who should be further evaluated for TB. We propose that young children who are foreign born or US born with foreign-born parents be assessed for TB if they have coughs lasting for >3 weeks and are not experiencing clinical improvement.15,19–21 Even though specimen collection from young children is challenging, confirming TB diagnosis and identifying drug susceptibilities is critical. In our study, PZA-resistant M tuberculosis complex strains were associated with a parent being born in Mexico and with consumption of unpasteurized dairy products, suggesting infection with Mycobacterium bovis. Speciation to confirm M bovis was not performed, but all children with PZA-resistant mycobacterial disease had extrapulmonary involvement, an epidemiologic characteristic of M bovis disease, and many had eaten unpasteurized dairy products, a known source of M bovis infection.22 If culture cannot be obtained, these epidemiologic and clinical characteristics can help support a diagnosis of PZA-resistant TB. The importance of prioritizing young pediatric contacts for evaluation in contact investigations23 is underscored by our finding that more than one-third of the children in the observational study were diagnosed that way. Prevention of TB in young children also requires preventing TB transmission from adults to children. In the observational study, more than 60% of US-born children had an identified source case, but a substantial amount of time was required to initiate treatment of these source cases: a median of 105 days from symptom onset. Earlier identification of the adult source case might have prevented transmission to these children or allowed for preventive treatment before the children’s infections had progressed to disease. Factors associated with longer time to treatment initiation could include declining public health funding, which would reduce resources required for extensive and detailed contact tracing and source case investigations.24 Other factors might be barriers to care that are unique to foreign-born populations; many of the individuals who were identified source cases in this study did not speak English, were undocumented immigrants, and expressed fears of deportation. Although early diagnosis and prompt treatment of TB disease is a primary component of TB control, it is preferable to prevent TB disease through (1) identification, screening, and LTBI testing of children at high risk and (2) treatment of infected children. Testing of children who do not have TB risk factors is strongly discouraged because of a low yield of positives and a high yield of false-positives.17 In response to the need for easily administered screening instruments, questionnaires have been developed and validated for screening children <18 years old in clinics and office-based practices; those determined to be at high risk can be tested for LTBI.17,19,25 These questionnaires ask about presence of a household member born outside the United States, child’s birth place and foreign travel, foreign visitors to the child’s household, and other risk factors. Different questions and combinations of questions have been associated with different levels of sensitivity and specificity. The high rates of TB in young children who are foreign born or have foreign-born parents suggest that, at least for children <5 years old, all those who are born in countries with moderate to high TB rates (eg, Asia, Middle East, Africa, Latin America, and countries of the former Soviet Union), or have a parent who was born in such a country, should be considered for LTBI testing, regardless of responses to other questions.26 Of note, although 1 set of current guidelines for LTBI testing in children recommends testing of children with recent foreign travel to countries with high TB rates,23 more than two-thirds of the children in our observational study were exposed in the United States. In an effort to curtail importation of TB cases among US immigrants, the CDC’s Division of Global Migration and Quarantine in 2007 revised the prearrival requirements for immigrant children and refugees to require that all children ages 2 to 14 be tested for TB infection and, if positive, screened for TB with a chest radiograph.27,28 Although this change will help identify both children with TB and those who could benefit from LTBI treatment in the United States, it will not help those children who acquire infection after arrival. An important issue is the generalizability of the study’s calculated TB rates to the nation as a whole, as the population studied was not a national sample. However, the sample was drawn from 16 states and represented almost half of all TB cases diagnosed in young children in the United States during the study enrollment years. Moreover, the calculated overall rate of 2.57 per 100 000 young children in the study jurisdictions is similar to the national rate of 2.38 per 100 000 in 2005–2007.29 Although parental origin was unknown for 33 US-born children, a sensitivity analysis that assigned those children to US or foreign parentage changed the estimated TB rates only modestly. A limitation of the observational study is that only ~60% of eligible cases were enrolled, and US-born children were more likely to be enrolled; to the extent that children with particular characteristics (eg, age, site of disease) were differentially enrolled by nativity, this could affect study comparisons. However, the clinical characteristics of the study population were consistent with those previously reported for pediatric TB, including (1) the proportion and types of extrapulmonary TB overall and by age group, (2) the higher prevalence of PZA resistance among children of Mexican descent, and (3) the proportion of laboratory-confirmed cases (30% in the study, 25% nationally, P = .15).16,30–32 An additional limitation to the observational study is that source cases were identified through epidemiologic links; it is unclear how many were confirmed by genotyping. However, other studies have shown tight concordance (85%–91%) between genotyping results and epidemiologic data in patient-source case pairs for young children.33,34 Finally, PZA analysis was limited by missing data. CONCLUSIONS Ultimately, prevention of TB in young children will require more effective testing and treatment of adults at high risk of LTBI, so they do not develop TB and transmit it to young children.35 Foreign-born adults represent the largest high-risk population in the United States.7 All medical providers should assess patients’ country of birth routinely as part of any clinical encounter. This practice will identify persons likely to have LTBI and TB disease; treatment will prevent disease transmission. Supplementary Material Supplemental material We thank the investigators and staff at all the participating TBESC sites: James McAuley, MD, MPH, Sharon Welbel, MD, Judith Beison, BS, Chicago, IL; Frank Wilson, MD, MPH, Cheryl LeDoux, MT, MPH, Little Rock, AR; Jennifer Flood, MD, MPH, Sumi Sun, MPH, Richmond, CA; Holly Anger, MPH, Paul Colson, PhD, Yael Hirsch-Moverman, PhD, MPH, Hugo Ortega, BA, Jiehui Li, MBBS, MSc, New York, NY; Lourdes Yun, MD, MSPH, Denver, CO; Jane Tapia, BSN, Atlanta, GA; Jessie Wing, MD, MPH, Honolulu, HI; Wendy Cronin, PhD, Susan Dorman, MD, Frances Maurer, BSN, MS, Baltimore, MD; Sue Etkind, RN, MS, Sharon Sharnprapai, MS, John Bernardo, MD, C. Robert Horsburgh Jr, MD, Arnaud Barbosa, Boston, MA; Wendy Sutherland, MPH, Hodan Guled, MPH, Jane Schulz, RN, MPH, Sarah Solarz, MPH, Minneapolis, MN; John Grabau, PhD, MPH, Margaret Oxtoby, Stephen Hughes, PhD, Albany, NY; Rachel Royce, PhD, MPH, Carol Dukes-Hamilton, MD, Research Triangle Park, NC; Connie Haley, MD, MPH, Jon Warkentin, MD, MPH, Tamara Chavez-Lindell, MPH, Nashville, TN; Stephen Weis, DO, Patrick Moonan, DrPH, Guadalupe Munguia, MD, MPH, Fort Worth, TX; Ashutosh Tamhane, MD, PhD, MSPH, Birmingham, AL; Nandini Selvam, MPH, PhD, Anna Sevilla, MPH, Newark, NJ; Gary Goldbaum, MD, MPH, Masa Narita, MD, Seattle, WA; Lisa Pascopella, PhD, MPH, L. Masae Kawamura, MD, Baby Djojonegoro, MPH, San Francisco, CA; Charles Wallace, PhD, Austin, TX; Frank Valdez Jr, Yuly Orozco, Houston, TX; Izelda Zarate, Brownsville, TX; Marie McMillan, RN, Kesner Accime, Ft Lauderdale, FL. The authors also thank Andrea T Cruz, MD, MPH, for her critical review of the manuscript, as well as the CDC project coordinator Lolem Ngong, MPH. FUNDING: Funded by the Centers for Disease Control and Prevention. ABBREVIATIONS CDC Centers for Disease Control and Prevention CI confidence interval DST drug susceptibility test IQR interquartile range LTBI latent tuberculosis infection MDR multidrug-resistant PZA pyrazinamide TB tuberculosis TBESC Tuberculosis Epidemiologic Studies Consortium FIGURE 1 Sources of TB infection for 149 children <5 years old who were foreign born or had foreign-born parents and were diagnosed with TB in 20 US jurisdictions 2005–2006.a Three of 5 foreign-born children had history of previous travel or foreign-born visitors. Of the 74 US-born children, 1 had lived outside the United States, 10 had travelled outside the United States, 16 reported foreign-born household visitors, and 50 (68% of 74) had no risks identified other than known contact to active TB in the United States. TABLE 1 Reported TB Cases in Children <5 Years Old by Study Site, Nativity, and Enrollment Period, 2005–2006a Enrollment Site Jurisdictions All Children Foreign Born US Born Years Enrolling Arkansas Department of Health State of Arkansas     3   1     2 1.4 University of Alabama, Birmingham State of Alabama   12   2   10 1.4 California Department of Health Services and University of California, San Francisco Alameda, Orange, San Diego, Santa Clara, and San Francisco counties   30   3   27 1.9 Denver Public Health and Hospitals Authority Six counties in the Denver metropolitan area     4   2     2 1.5 Broward County Department of Health Broward County, Florida     4   0     4 1.5 Emory University Twenty counties in the Atlanta metropolitan area   23   4   19 1.4 Hawaii Department of Health State of Hawaii     1   0     1 1.9 American Lung Association of Metropolitan Chicago Four counties in the Chicago metropolitan area   17 14     3 2.0 Maryland Department of Health and Mental Hygiene State of Maryland   13   3   10 1.6 Massachusetts Department of Public Health State of Massachusetts   12   1   11 1.6 Minnesota Department of Health State of Minnesota   12   4     8 1.4 University of Medicine and Dentistry of New Jersey State of New Jersey   17   1   16 1.3 New York City Department of Health/Charles P. Felton National TB Center at Harlem New York City (NYC)   37   6   31 1.3 New York State Department of Health, Health Research Inc State of New York excluding New York City   15   2   13 1.7 RTI International State of North Carolina   21   3   18 1.3 Tennessee Department of Health State of Tennessee   24   4   20 1.7 Texas Department of State Health Services and University of North Texas Health Science Center State of Texas 115 10 105 2.1 Seattle-King County Department of Public Health King County, WA     4   1     3 1.4 Total 364   61 303 a Case numbers reflect total case counts for children <5 years old in the given jurisdiction during that study site’s specific 2005 and 2006 enrollment period, not the entire 2 calendar years. TABLE 2 TB Cases and Rates Among Children Who Are <5 Years Old by Birth Origin and Parents’ Birth Origin in 20 Jurisdictions in the United States Total Foreign Born US Born With at Least 1 Foreign-Born Parent US Born of US-Born parents No. of TB cases in children <5 y olda (%) 364 (100) 61 (16.7) 194 (53.3) 76 (20.9) Annualized case number 217.12 36.75 117.18 43.86 Population <5 y of age estimated from US Census ACS 2005–2007 (%) 8 450 150 (100) 152 890 (1.8) 2 433 890 (28.8) 5 863 360 (69.4) TB rateb (95% CI) 2.57 (2.23–2.91) 24.03 (16.21–31.85) 4.81 (3.94–5.68) 0.75 (0.53–0.97) ACS, American Community Survey. a The total number of 364 includes 33 (9%) children, not shown in the other columns, with unknown parental birth countries. b TB rates per 100 000 population, annualized and adjusted for study enrollment period. To estimate the impact of excluding the 33 children with unknown parental birth country on incidence, the entire annualized number (19.33) was added to each US-born category, and TB incidence and 95% CIs were calculated to be 5.61 (4.67–6.55) for US-born children with at least 1 foreign-born parent; 1.08 (0.81–1.35) for US born with US-born parents. TABLE 3 Demographic and Clinical Characteristics, Care-Seeking Behavior, and Access to Health Care Among 149 Children <5 Years Old Newly Diagnosed With TB in the United States in 2005–2006 and Enrolled in Observational Study, by the Child’s Nativity Participant Characteristics All Cases (N = 149) US Born (N = 122) Foreign Born (N = 27) Risk Ratioa 95% CI n % (n/N) n % (n/N) n % (n/N) Age  Younger than 1 y 38 26 36 30 2 7 4.28   1.06–17.22  Between 1 and 4 y 111 74 86 70 25 93 referent Gender  Male 68 46 54 44 14 52 0.78   0.39–1.54  Female 81 54 68 56 13 48 referent Hispanic  Yes 99 66 89 73 10 37 3.37   1.67–6.80  No 50 34 33 27 17 63 referent Country or world region of parent’s birthb  Mexico 81 54 72 59 9 33 2.38   1.15–4.96  Latin America/South America/Caribbean 29 19 25 20 4 15 1.39   0.52–3.71  South Asia/East Asia/Pacific 19 13 15 12 4 15 0.84   0.33–2.17  Sub-Saharan Africa 14 9 10 8 4 15 0.60   0.24–1.48  United Statesc 6 4 0 0 6 22 excluded Parent’s visa status at US entryd  Undocumented 72 48 63 52 9 33 1.82   0.84–3.96  Documented 57 38 44 36 13 48 referent  Unknown 20 13 15 12 5 19 excluded Informant speaks English “not at all”  Yes 46 31 37 30 9 33 0.89   0.43–1.84  No 103 69 85 70 18 67 referent Parent feared child’s deportatione  Yes 15 10 11 9 4 15 0.81   0.32–2.01  No 107 72 84 69 23 85 referent  Not applicable 21 14 21 17 0 0 excluded  No opinion 6 4 6 5 0 0 excluded Health insurance  Yes 113 76 100 82 13 48 3.38   1.76–6.51  No 36 24 22 18 14 52 referent Reason for initially seeking health caref  Screenings and well-baby examinations 37 25 17 14 20 74 0.12   0.05–0.25  Symptoms 61 41 56 46 5 19 3.05   1.22–7.61  Contact investigation or known TB exposure 51 34 49 40 2 7 6.51   1.60–26.38 Source case identified  Yes 79 53 74 61 5 19 4.97   1.99–12.41  No 70 47 48 39 22 81 referent Probable location of TB transmission  Inside United States 102 68 97 80 5 19 9.55   3.85–23.66  Outside United States 36 24 15 12 21 78 0.09   0.04–0.21  Exposure risks both inside and outside United States 11 7 10 8 1 4 2.07   0.31–13.86 TB site  Pulmonary only 102 68 83 68 19 70 0.91   0.43–1.94  Pulmonary and extrapulmonary 21 14 17 14 4 15 0.94   0.36–2.45  Extrapulmonary only 26 17 22 18 4 15 1.22   0.46–3.22 Abnormal chest radiograph  Yes 130 87 104 85 26 96 0.29   0.04–2.03  No 17 11 16 13 1 4 referent  Unknown/not done 2 1 2 2 0 0 excluded Specimen positive by culture  Yes 45 30 44 36 1 4 11.25   1.57–80.38  Culture negative 38 26 25 20 13 48 0.37   0.19–0.71  Not done or unknown 66 44 53 43 13 48 0.81   0.41–1.60 Tuberculin skin test positive at diagnosis  Yes 124 83 99 81 25 93 0.62   0.16–2.37  No 16 11 14 11 2 7 referent  Unknown/not done 9 6 9 7 0 0 excluded TB symptoms  Yes 99 66 83 68 16 59 1.36   0.68–2.71  No 50 34 39 32 11 41 referent a Risk ratio estimates the relative risk of US-born children compared with foreign-born children (with asymptotic 95% CIs). b At least 1 parent born in given region. When discordant and US-born for 1 parent, foreign region given (n = 22); Mexico given when Mexican and Latin American parents (n = 4). c All children were adopted (adoptee birth countries were Guatemala, China, Ethiopia, Kazakhstan, and 2 from the Russian Federation). d At least 1 parent with given visa status. When discordant, undocumented given over documented. e Interview informant was asked: “When you took your child for tuberculosis treatment, were you afraid you or your child might be sent back to the country you came from?” f Reasons for initially seeking health care among adopted children were post adoption check-ups (n = 4), well-baby examinations, and symptoms (n = 1). TABLE 4 Characteristics of 99 Symptomatic Children <5 Years Old Newly Diagnosed With TB in 20 Jurisdictions in the United States in 2005–2006 and Enrolled in Observational Study, by the Child’s Nativity Symptomatic Participant Characteristics All Cases (N = 99) US Born (N = 83) Foreign Born (N = 16) Risk Ratioa 95% CI n % (n/N) n % (n/N) n % (n/N) Cough  Yes 53 54 45 54     8 50     1.15 0.47–2.83  No 46 46 38 46     8 50 referent Fever  Yes 55 56 51 61     4 25     3.75 1.30–10.82  No 44 44 32 39   12 75 referent Night sweats  Yes 43 43 34 41     9 56     0.60 0.24–1.48  No 56 57 49 59     7 44 referent Weight loss  Yes 38 38 33 40     5 31     1.37 0.52–3.64  No 61 62 50 60   11 69 referent Lymphadenopathy  Yes 20 20 16 19     4 25     0.78 0.28–2.16  No 77 78 65 78   12 75 referent  Unknown   2   2   2   2     0   0 excluded Reason for initially seeking health careb  Screenings and well-baby examinations 18 18   8 10   10 63     0.13 0.06–0.32  Symptoms 57 58 52 63     5 31     2.99 1.12–7.95  Contact investigation or known TB exposure 24 24 23 28     1   6     4.80 0.67–34.47 Median days from symptom onset to treatment initiation (IQR) 52 (23–117) 44.5 (21–112) 115 (36–160)     0.999 0.998–1.000 Median number of physician visits from symptom onset to treatment initiationc (IQR)   2 (1–3)   2 (1–2.5)     2 (1–3)     1.023 0.951–1.100 a Risk ratio estimates the relative risk of US-born children compared with foreign-born children (with asymptotic 95% CIs). b Reasons for initially seeking health care among symptomatic adopted children were post-adoption check-ups (n = 2), well-baby examinations, and symptoms (n = 1). c Number of physician visits required for TB diagnosis for persons reporting symptoms. TABLE 5 Demographic and Clinical Characteristics of 46 Persons With TB Who Were Identified as the Source of Transmission for 52 Children With TB Enrolled in the Observational Study Source Case Characteristic n (N%) Gender  Male    18 (39)  Female    28 (61) Age at treatment initiation, median (IQR) 30.9 (25.2–39.6) Country or world region of birth  United States      4 (9)  Latin America/South America/Caribbean      9 (20)  Mexico    24 (52)  Sub-Saharan Africa      3 (6)  South Asia/East Asia/Pacific      6 (13) Speaks English “not at all”    15 (33) Sputum smear positive for acid-fast bacilli    41 (89) Sputum culture positive for TB    45 (98) Chest radiograph abnormality  Cavitary    28 (61)  Noncavitary consistent with TB    17 (37)  Unknown      1 (2) Reason for seeking care  Symptoms    37 (80)  Contact investigation      7 (15)  Other (regular check-up and pregnancy/child birth)      2 (4) Reported cough  Yes    42 (91)  No      4 (9) Years from source case’s US entry to TB treatment initiation (N = 42), median (IQR)   5.3 (2.7–13.1) Days from source case’s onset of symptoms to treatment initiation (N = 41); median (IQR)  105 (72–190) Days from source case’s onset of symptoms to child’s TB treatment initiation (N = 47),a median (IQR)  132 (101–212) Days from source case’s TB treatment initiation to child’s TB treatment initiation (N = 52),a median (IQR) 11.5 (1–33.5) Visa status at interview (N = 42)  Undocumented    16 (38)  Temporary      5 (12)  Permanent    18 (43)  Unknown      3 (7) Overseas medical examination before US entry (N = 42)  Yes    10 (24)  No    29 (69)  Unknown      3 (7) US medical examination to change visa (N = 42)  Yes      5 (12)  No    37 (88) a Multiple children infected by source case. WHAT’S KNOWN ON THIS SUBJECT More than 60% of all US tuberculosis cases occur among foreign-born persons, but ~90% of cases in young children occur among US-born; many of these children have foreign-born parents, suggesting that this is an important population for prevention. WHAT THIS STUDY ADDS This is the first study to calculate tuberculosis rates in US-born children by parental nativity. Compared with US-born children with US-born parents, rates were 32 times higher in foreign-born children and 6 times higher in US-born children with foreign-born parents. Drs Katz, Davidow, and Reves conceptualized and designed the study, and critically reviewed the manuscript; Drs Pang, Teeter, and Graviss participated in data collection and analytic planning, carried out the analyses, and drafted the initial manuscript; Mr Miranda, Ms Wall, Ms Ghosh, Ms Stein-Hart, and Dr Restrepo participated in data collection and analytic planning, and reviewed and revised the manuscript; and all authors approved the final manuscript as submitted. The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention. FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose. POTENTIAL CONFLICT OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose. 1 Lobato MN Mohle-Boetani JC Royce SE Missed opportunities for preventing tuberculosis among children younger than five years of age Pediatrics 2000 106 6 Available at: www.pediatrics.org/cgi/content/full/106/6/e75 2 Centers for Disease Control and Prevention (CDC) NCHHSTP Atlas Available at: www.cdc.gov/nchhstp/atlas/. Accessed Apr 23, 2013 3 Khan EA Starke JR Diagnosis of tuberculosis in children: increased need for better methods Emerg Infect Dis 1995 1 4 115 123 8903180 4 Smith KC Tuberculosis in children Curr Probl Pediatr 2001 31 1 1 30 11217659 5 Centers for Disease Control and Prevention Reported Tuberculosis in the United States, 2009 Atlanta, GA US Department of Health and Human Services, CDC 2010 6 Centers for Disease Control and Prevention Reported Tuberculosis in the United States, 2010 Atlanta, GA US Department of Health and Human Services, CDC 2011 7 Centers for Disease Control and Prevention Reported Tuberculosis in the United States, 2011 Atlanta, GA US Department of Health and Human Services, CDC 2012 8 Kenyon TA Driver C Haas E Valway SE Moser KS Onorato IM Immigration and tuberculosis among children on the United States-Mexico border, County of San Diego, California Pediatrics 1999 104 1 Available at: www.pediatrics.org/cgi/content/full/104/1/e8 9 Winston CA Menzies HJ Pediatric and adolescent tuberculosis in the United States, 2008–2010 Pediatrics 2012 130 6 Available at: www.pediatrics.org/cgi/content/full/130/6/e1425 10 Davidow AL Katz D Reves R Bethel J Ngong L Tuberculosis Epidemiologic Studies Consortium The challenge of multisite epidemiologic studies in diverse populations: design and implementation of a 22-site study of tuberculosis in foreign-born people Public Health Rep 2009 124 3 391 399 19445415 11 Katz D Albalak R Wing JS Combs V Tuberculosis Epidemiologic Studies Consortium Setting the agenda: a new model for collaborative tuberculosis epidemiologic research Tuberculosis (Edinb) 2007 87 1 1 6 16895763 12 Kominski R How good is “How well”? An examination of the Census English-speaking ability question Paper presented at: the American Statistical Association annual meeting Aug 6–11, 1989 Washington, DC 13 Stuart A Ord K Kendall’s Advanced Theory of Statistics, Distribution Theory 1 6th London, UK Arnold 1998 351 14 Elandt-Johnson RC Johnson NL Survival Models and Data Analysis New York, NY John Wiley & Sons 1980 69 15 Munoz FM Starke JR Tuberculosis in children Reichman LB Hershfield ES Tuberculosis: a Comprehensive International Approach 2nd New York, NY Marcel Dekker, Inc 2000 553 586 16 CDC Slide sets—Epidemiology of pediatric tuberculosis in the United States, 1993–2008 (Slide 19) Available at: www.cdc.gov/tb/publications/slidesets/pediatricTb/default.htm. Accessed March 11, 2013 17 Pediatric Tuberculosis Collaborative Group Targeted tuberculin skin testing and treatment of latent tuberculosis infection in children and adolescents Pediatrics 2004 114 supp 4 1175 1201 18 American Academy of Pediatrics Tuberculosis Pickering LK Baker CJ Kimberlin DW Long SS Red Book: 2012 Report of the Committee on Infectious Diseases Elk Grove Village, IL American Academy of Pediatrics 2012 740 19 Froehlich H Ackerson LM Morozumi PA Pediatric Tuberculosis Study Group of Kaiser Permanente, Northern California Targeted testing of children for tuberculosis: validation of a risk assessment questionnaire Pediatrics 2001 107 4 Available at: www.pediatrics.org/cgi/content/full/107/4/e54 20 Marchant JM Masters IB Taylor SM Chang AB Utility of signs and symptoms of chronic cough in predicting specific cause in children Thorax 2006 61 8 694 698 16670171 21 Marais BJ Obihara CC Gie RP The prevalence of symptoms associated with pulmonary tuberculosis in randomly selected children from a high burden community Arch Dis Child 2005 90 11 1166 1170 16243872 22 Dankner WM Davis CE Mycobacterium bovis as a significant cause of tuberculosis in children residing along the United States-Mexico border in the Baja California region Pediatrics 2000 105 6 Available at: www.pediatrics.org/cgi/content/full/105/6/e79 23 National Tuberculosis Controllers Association; Centers for Disease Control and Prevention (CDC) Guidelines for the investigation of contacts of persons with infectious tuberculosis. Recommendations from the National Tuberculosis Controllers Association and CDC MMWR Recomm Rep 2005 54 RR-15 1 47 24 American Public Health Association Budget cuts straining capacity of public health departments: services in demand Nations Health 2010 40 1 16 25 Ozuah PO Ozuah TP Stein REK Burton W Mulvihill M Evaluation of a risk assessment questionnaire used to target tuberculin skin testing in children JAMA 2001 285 4 451 453 11242430 26 American Thoracic Society Targeted tuberculin testing and treatment of latent tuberculosis infection. This official statement of the American Thoracic Society was adopted by the ATS Board of Directors, July 1999. This is a Joint Statement of the American Thoracic Society (ATS) and the Centers for Disease Control and Prevention (CDC). This statement was endorsed by the Council of the Infectious Diseases Society of America. (IDSA), September 1999, and the sections of this statement Am J Respir Crit Care Med 2000 161 4 pt 2 S221 S247 10764341 27 CDC Revised technical instructions for tuberculosis screening and treatment for panel physicians MMWR Morb Mortal Wkly Rep 2008 57 292 293 28 CDC CDC immigration requirements: technical instructions for tuberculosis screening and treatment Atlanta, GA US Department of Health and Human Services, CDC 2009 Available at: www.cdc.gov/immigrantrefugeehealth/exams/ti/panel/tuberculosis-panel-technical-instructions.html. Accessed October 17, 2013 29 CDC OTIS 2010 TB Data Request National Tuberculosis Surveillance System, United States, 1993–2010 US Department of Health and Human Services, Centers for Disease Control and Prevention, Division of TB Elimination, CDC WONDER Online Database 11 2012 Available at: http://wonder.cdc.gov/tb-v2010.html. Accessed May 22, 2013 30 Centers for Disease Control and Prevention (CDC) Human tuberculosis caused by Mycobacterium bovis—New York City, 2001–2004 MMWR Morb Mortal Wkly Rep 2005 54 24 605 608 15973241 31 Hlavsa MC Moonan PK Cowan LS Human tuberculosis due to Mycobacterium bovis in the United States, 1995–2005 Clin Infect Dis 2008 47 2 168 175 18532886 32 Rodwell TC Moore M Moser KS Brodine SK Strathdee SA Tuberculosis from Mycobacterium bovis in binational communities, United States Emerg Infect Dis 2008 14 6 909 916 18507901 33 Sun SJ Bennett DE Flood J Loeffler AM Kammerer S Ellis BA Identifying the sources of tuberculosis in young children: a multistate investigation Emerg Infect Dis 2002 8 11 1216 1223 12453345 34 Wootton SH Gonzalez BE Pawlak R Epidemiology of pediatric tuberculosis using traditional and molecular techniques: Houston, Texas Pediatrics 2005 116 5 1141 1147 16264001 35 Hill AN Becerra J Castro KG Modelling tuberculosis trends in the USA Epidemiol Infect 2012 140 10 1862 1872 22233605

PMC005xxxxxx/PMC5135018.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 0410462 6011 Nature Nature Nature 0028-0836 1476-4687 27882964 5135018 10.1038/nature20166 NIHMS829950 Article Bacteria establish an aqueous living space in plants crucial for virulence Xin Xiu-Fang 1 Nomura Kinya 1 Aung Kyaw 12 Velásquez André C. 1 Yao Jian 1* Boutrot Freddy 3 Chang Jeff H. 4 Zipfel Cyril 3 He Sheng Yang 1256† 1 Department of Energy Plant Research Laboratory, Michigan State University, East Lansing, MI 48824, USA 2 Howard Hughes Medical Institute, Gordon and Betty Moore Foundation, Michigan State University, East Lansing, MI 48824, USA 3 The Sainsbury Laboratory, Norwich Research Park, NR4 7UH Norwich, UK 4 Department of Botany and Plant Pathology and Center for Genome Research and Biocomputing, Oregon State University, Corvallis, OR 97331, USA 5 Department of Plant Biology, Michigan State University, East Lansing, MI 48824, USA 6 Plant Resilience Institute, Michigan State University, East Lansing, MI 48824, USA † Correspondence to: Sheng Yang He; hes@msu.edu * Current address: Department of Biological Sciences, Western Michigan, Kalamazoo, MI 49008, USA 16 11 2016 24 11 2016 23 5 2017 539 7630 524529 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. High humidity has a profound influence on the development of numerous phyllosphere diseases in crop fields and natural ecosystems, but the molecular basis of this humidity effect is not understood. Previous studies emphasize immune suppression as a key step in bacterial pathogenesis. Here we show that humidity-dependent, pathogen-driven establishment of an aqueous intercellular space (apoplast) is another crucial step in bacterial infection of the phyllosphere. Bacterial effectors, such as Pseudomonas syringae HopM1, induce establishment of the aqueous apoplast and are sufficient to transform non-pathogenic P. syringae strains into virulent pathogens in immune-deficient Arabidopsis under high humidity. Arabidopsis quadruple mutants simultaneously defective in a host target (MIN7) of HopM1 and in pattern-triggered immunity could not only recapitulate the basic features of bacterial infection, but also exhibit humidity-dependent dyshomeostasis of the endophytic commensal bacterial community in the phyllosphere. These results highlight a new conceptual framework for understanding diverse phyllosphere-bacterial interactions. Introduction The terrestrial phyllosphere (the above-ground parts of plants) represents one of the most important habitats on Earth for microbial colonization. Although the vast majority of phyllosphere microbes exhibit benign commensal associations and maintain only modest populations, adapted phyllosphere pathogens can multiply aggressively under favorable environmental conditions and cause devastating diseases. In crop fields, phyllosphere bacterial disease outbreaks typically occur after rainfalls and a period of high humidity1-3, consistent with the famous “disease triangle” (host-pathogen-environment) dogma formulated more than 50 years ago4. The molecular basis of the profound effect of high humidity on bacterial infection of the phyllosphere is not understood. Many plant and animal pathogenic bacteria, including the model phyllosphere bacterial pathogen Pseudomonas syringae, carry a type III secretion system (T3SS), which is used to deliver disease-promoting “effector” proteins into the host cell as a primary mechanism of pathogenesis5,6. Studies of how individual type III effectors promote bacterial disease in plants and animals show that effector-mediated suppression of host immunity is a common theme in both plant-bacterial 7-9 and animal-bacterial interactions 10,11. However, due to the apparent molecular complexities in bacterial diseases, the fundamental question as to what minimal set of host processes that must be subverted to allow basic bacterial pathogenesis to occur has not been answered in any plant or animal pathosystem. Immune-suppression and pathogenesis To test the hypothesis that host immunity may be the only process that needs to be subverted for bacterial pathogenesis in the phyllosphere, we performed infection assays in Arabidopsis polymutants severely defective in multiple immune pathways: (i) fls2/efr/cerk1 (fec), which is mutated in three major pattern recognition receptor (PRR) genes relevant to P. syringae pv. tomato (Pst) DC3000 infection12, (ii) bak1-5/bkk1-1/cerk1 (bbc; see Methods), which is compromised in immune signaling downstream of multiple PRRs13,14, and (iii) dde2/ein2/pad4/sid2 (deps), which is defective in all three major defense hormone pathways (salicylic acid, jasmonate and ethylene)15. Two nonpathogenic mutant derivatives of Pst DC3000 were used: the hrcC− mutant (defective in type III secretion)16 and the DC3000D28E mutant, in which the T3SS remains intact, but 28 of 36 type III effectors are deleted17. As shown in Fig. 1a, hrcC− and DC3000D28E mutants grew very poorly not only in wild-type Col-0, but also in immune-compromised mutants when infiltrated into the apoplast, suggesting that host immunity is unlikely to be the only process subverted by Pst DC3000 during infection. High humidity required for pathogenesis During the active pathogenesis phase, phyllosphere bacterial pathogens such as Pst DC3000 live mainly in the air-filled apoplast, which is connected directly to open air through epidermal pores called stomata. The water status inside the apoplast could therefore be influenced by air humidity during pathogen infection. In crop fields, phyllosphere bacterial disease outbreaks typically occur after rainfalls and a period of high humidity1-3,18, following the “disease triangle” dogma in plant pathology. In addition, one of the earliest and common symptoms of phyllosphere bacterial diseases is the appearance of “water soaking” in infected tissues, although whether water-soaking plays an active role in bacterial pathogenesis remains unclear. These key phenomena could be demonstrated in the laboratory. Whereas Pst DC3000 multiplied to a very high level under high humidity (~95%; mimicking high humidity after rains in crop fields), it multiplied to a much lower level under low humidity (< 60%) (Fig. 1b), as reflected also in a lower disease severity (Fig. 1c). The ability of Pst DC3000 to multiply increased as humidity rose; in contrast, the hrcC− mutant multiplied poorly under all tested humidity conditions (Fig. 1d). The most aggressive infection by Pst DC3000 was associated with the appearance, usually within one day after infection, of water soaking in the infected Arabidopsis leaves under high humidity (Fig. 1e). Water-soaked spots could also be observed in Pst DC3000-infected leaves of another host species, tomato (Fig. 1f). Real-time imaging (Supplementary Video 1) showed that the initial water-soaked spots mark the areas of later disease symptoms (necrosis and chlorosis), and revealed, interestingly, that water soaking was a transient process and it disappeared before the onset of late disease symptoms. Using a Pst DC3000 strain tagged with a luciferase reporter (DC3000-lux19), we found that water soaking areas and luciferase signals are detected nonuniformly across the leaf, but they overlap extensively (Fig. 1g, Extended Data Fig. 1), revealing that water-soaked areas are where bacteria multiply aggressively in the phyllosphere before the onset of late disease symptoms. P. syirngae water-soaking effectors The DC3000D28E mutant never caused water soaking under any condition (e.g., high humidity/inoculum). We therefore transformed each of the 28 Pst DC3000 effector genes, individually, back to the DC3000D28E mutant to identify the effector(s) that cause water soaking. Most effectors did not (see Fig. 2a for avrPto, as an example); but hopM1 and avrE (together with their respective type III secretion chaperone genes shcM and avrF) did (Fig. 2a). We found this result interesting because, although HopM1 and AvrE show no sequence similarity, they were previously shown to be functionally redundant in virulence and they are highly conserved in diverse P. syringae strains and/or other phytopathogenic bacteria20,21. Moreover, transgenic overexpression of 6xHis:HopM122 or 6xHis:AvrE23 under control of dexamesathone (DEX)-inducible promoter (10 μM DEX used) also caused water soaking under high humidity (Fig. 2b). In contrast, transgenic expression of AvrPto, like D28E (avrPto), did not. These results show that HopM1 and AvrE, either delivered by bacteria or when overexpressed transgenically inside the plant cells, are each sufficient to cause water soaking. Bacterial mutant analysis showed that HopM1 and AvrE are necessary for Pst DC3000 to cause water soaking during infection, as the avrE−/hopM1− double mutant20 could not cause water soaking, even when the inoculum of the avrE−/hopM1− mutant was adjusted to reach a similar population with Pst DC3000 when water soaking was assessed (Fig. 2c). In contrast, Pst DC3000 and the avrE− and hopM1− single mutants20 caused strong initial water soaking (Extended Data Fig. 2a) and later disease symptoms (Extended Data Fig. 2b) and multiplied aggressively in a high humidity-dependent manner, while the avrE−/hopM1− double mutant multiplied poorly regardless of the humidity setting (Fig. 2d). Transgenic expression of 6xHis:HopM1 in Arabidopsis (in these experiments 0.1 nM was used to induce low-level expression of HopM1 so that HopM1 alone does not cause extensive water soaking) restored the ability of the avrE−/hopM1− double mutant to cause water soaking and multiply highly under high humidity (Extended Data Fig. 2c, d). These results revealed that, unlike the other 34 effectors present in the avrE−/hopM1− double mutant, the virulence functions of HopM1 and AvrE are uniquely dependent on external high humidity. Why would the virulence functions of HopM1 and AvrE be dependent on the external humidity? We hypothesized that perhaps the primary function of HopM1 and AvrE is to create an aqueous apoplast per se (i.e., bacteria “prefer” to living in an aqueous environment in the apoplast), the maintenance of which requires high humidity as the leaf apoplast is directly connected to open air through stomata. If so, it may be possible to substitute the function of HopM1 and AvrE by simply providing water to the apoplast. To directly test this hypothesis, we performed transient water supplementation experiments in which Col-0 plants infiltrated with the avrE−/hopM1− mutant were kept water-soaked, transiently, for the first 12 h to 16 h to mimic the kinetics of transient water soaking normally occurring during Pst DC3000 infection (Supplementary Video 1). Remarkably, transient apoplast water supplementation was sufficient to restore the multiplication (100- to 1000-fold) of the avrE−/hopM1− mutant almost to the level of Pst DC3000 (Fig. 2e), as well as appearance of severe disease symptoms (Fig. 2f). As controls, Pst DC3000, the hrcC− mutant and CUCPB5452 (which contains avrE and hopM1 genes but has much reduced virulence due to deletion of other type III effectors24) grew only slightly better (<10 fold) with transient water-supplementation (Fig. 2e). These results demonstrate that the primary virulence function of HopM1 and AvrE can be effectively substituted by supplying water, transiently, to the apoplast. HopM1’s host target in water soaking To investigate the mechanism by which HopM1 creates aqueous apoplast, we focused on the host targets of HopM1 in Arabidopsis. We have previously shown that HopM1 is targeted to the trans-Golgi-network/early endosome (TGN/EE) in the host cell and mediates proteasome-dependent degradation of several host proteins, including MIN7 (also known as BEN1), which is a TGN/EE-localized ADP ribosylation factor-guanine nucleotide exchange factor involved in vesicle trafficking22,25,26. Although the min7 mutant plant partially allows increased bacterial multiplication22,25, the exact role of MIN7 during pathogen infection remains enigmatic. A previous study showed that HopM1’s virulence function is fundamentally different from that of canonical immune-suppressing effectors, such as AvrPto17. In light of our discovery of HopM1’s primary role in creating water-soaking in this study, we tested the intriguing possibility that MIN7 may be a key player in modulating apoplast water soaking in response to bacterial infection. Excitingly, we found that the min7 mutant plant allowed apoplast water soaking to occur in the absence of HopM1/AvrE (i.e., during infection by the avrE−/hopM1− mutant; Fig. 3a, Extended Data Fig. 3c), and allowed the avrE−/hopM1− mutant to multiply (Extended Data Fig. 3a, b). Thus, genetic removal of MIN7 is sufficient to mimic the virulence function of HopM1, albeit partially, in causing apoplast water soaking. The min7 mutant plant is defective in endocytic recycling of plasma membrane (PM) proteins and has an abnormal PM26, suggesting that HopM1 degrades MIN7 possibly to compromise host PM integrity as a mechanism to create an infection-promoting aqueous apoplast (Extended Data Fig. 4). If apoplast water soaking is an essential step of pathogenesis, we hypothesized that plants may have evolved defense mechanisms to counter it. Indeed, we found that Pst DC3000 (avrRpt2)-triggered effector-triggered immunity (ETI)27 completely blocked water-soaking, even when the inoculum of Pst DC3000 (avrRpt2) was raised to reach a population similar to Pst DC3000 when water soaking was assessed (Fig. 3b, c, Extended Data Fig. 5a-b). When transferred from high (~95%) to low (~50%) humidity, Pst DC3000 (avrRpt2)-infected leaves quickly wilted, indicating extensive ETI-associated programmed cell death. In contrast, Pst DC3000-infected, water-soaked leaves returned to pre-infection healthy appearance (Fig. 3b), indicating little host cell death during apoplast water soaking. Furthermore, Pst DC3000 (avrRpt2)-triggered ETI stabilized the MIN7 protein (Fig. 3d). These results therefore uncovered a previously unrecognized battle between bacterial virulence (creating apoplast water soaking) and host defense (preventing apoplast water soaking), in part linked to MIN7 stability, to take control of apoplast water availability. Reconstitution of P. syringae infection The discovery of apoplast water soaking as a key process of bacterial pathogenesis prompted us to investigate a new model in which PTI suppression and creation of apoplast water soaking are two principal pathogenic processes sufficient for bacterial infection of the phyllosphere. To test this hypothesis, we infected Col-0 and two PTI-compromised mutant plants (i.e., fec and bbc) with DC3000D28E, DC3000D28E (avrPto) or DC3000D28E (hopM1/shcM) and found that only DC3000D28E (hopM1/shcM), but not DC3000D28E or DC3000D28E (avrPto), caused strong water soaking, multiplied aggressively (almost to the Pst DC3000 level) and produced prominent disease symptoms in the fec and bbc mutant plants (Fig. 4a-c) in a high humidity-dependent manner (Fig. 4d). Furthermore, unlike PTI mutants, the npr1-6 mutant plant, which is defective in salicylic acid-dependent defense (Extended Data Fig. 6a-c), could not rescue the ability of DC3000D28E (hopM1/shcM) to multiply (Fig. 4a). Thus, a combination of defective PTI and presence of an aqueous-apoplast-inducing effector (HopM1) could almost fully convert a non-pathogenic mutant into a virulent pathogen in the Arabidopsis phyllosphere. If immune suppression and creation of apoplast water soaking are two principal pathogenic processes sufficient for bacterial infection of the phyllosphere, we reasoned that we might be able to construct a multi-host-target mutant that simulates the two processes. Such mutant plant might allow an otherwise nonpathogenic mutant bacterium (e.g., the hrcC− mutant) to colonize the phyllosphere, thereby reconstituting basic features of a phyllosphere bacterial infection. For this purpose, we mutated the MIN7 gene in PTI mutants (fec and bbc) and generated min7/fls2/efr/cerk1 (mfec) and min7/bak1-5/bkk1-1/cerk1 (mbbc) quadruple mutants using CRISPR technology (see Methods; Extended Data Fig. 7a). The quadruple mutant plants display a similar morphology as wild type Col-0 plants (Extended Data Fig. 7b) and have a tendency of showing some water-soaking spots, especially in mature leaves, under high humidity (Extended Data Fig. 7c, d). Excitingly, these mutants allow the nonpathogenic hrcC− mutant to multiply aggressively under high (~95%) humidity, to a final population that was ~100 fold higher than in Col-0 plants 5 days after inoculation, with the mbbc plants showing a greater susceptibility than the mfec plants (Fig. 5a). In addition, in these quadruple mutant plants, the hrcC− mutant induced prominent disease chlorosis and necrosis (Fig. 5b, Extended Data Fig. 7e), which were not observed for the hrcC− strain in Col-0, min7 or PTI mutants. Thus, a dual disruption of MIN7 and PTI signaling is sufficient to reconstitute the basic features of a model phyllosphere bacterial disease. Consistent with this conclusion, transient water supplementation to the leaf apoplast was sufficient to enhance the growth of the hrcC− mutant in the bbc triple mutant, but not in Col-0 plants (Fig. 5c). To our knowledge, this is the first infectious model disease, in plant or animal, for which basic pathogenesis has been reconstituted using biologically relevant host target mutants. Dyshomeostasis of commensal bacteria The inability of the nonpathogenic hrcC− mutant to multiply aggressively in wild-type phyllosphere resembles that of the commensal bacterial community that resides in the apoplast of healthy leaves. Consistent with this, only low levels of the endophytic phyllosphere bacterial community were detectable in wild type Col-0 plants (Fig. 5d). However, after plants were shifted from regular growth conditions (~60% relative humidity, day 0; Fig. 5d) to high humidity conditions (~95% relative humidity), the mfec and mbbc quadruple mutant plants, but not Col-0 plants, allowed excessive proliferation of the endogenous endophytic bacterial community (Fig. 5d, Extended Data Table 1), in a high humidity dependent manner (Extended Data Fig. 8a). Furthermore, the excessive proliferation of the endophytic bacterial community was associated with mild tissue chlorosis and necrosis in some leaves (Extended Data Fig. 8b). We found this result intriguing as a recent study showed that overgrowth of a beneficial root-colonizing fungus in immune-compromised (against fungal pathogens) plants also led to harmful effects in Arabidopsis28, illustrating a potentially common theme that the levels of commensal and beneficial microbiota must be strictly controlled by the host for optimal plant health. Future comprehensive in planta 16S rRNA amplicon-based analysis will be needed to determine whether there are also humidity-dependent changes in the composition of commensal bacterial communities in the Col-0, mfec and mbbc plants. Discussion Results from this study suggest a new conceptual framework for understanding phyllosphere-bacterial interactions (Fig. 5e). Specifically, we have identified PTI signaling and MIN7, presumably via vesicle trafficking, as two key components of the elusive host barrier that functions to limit excessive and potentially harmful proliferation of nonpathogenic microbes (e.g., hrcC− mutant) in the phyllosphere. Pathogenic bacteria, like Pst DC3000, have evolved T3SS effectors not only to disarm PTI signaling, but also to establish an aqueous living space in a humidity-dependent manner in order to aggressively colonize the phyllosphere. This new conceptual framework integrates host, pathogen and environmental factors, providing a critical insight into the enigmatic basis of the profound effect of humidity on the development of numerous bacterial diseases, consistent with the “disease triangle” dogma in plant pathology. Prior to this study, humidity was commonly thought to promote bacterial movements on the plant surface and invasion into plant tissues. Our study, however, revealed a striking and previously unrecognized effect of high humidity on the function of bacterial effectors inside the plant apoplast. An aqueous apoplast could potentially facilitate the flow of nutrients to bacteria, promote the spread/egression of bacteria, and/or affect apoplastic host defense responses, the latter of which may explain some of the previously observed effects of HopM1, AvrE and MIN7 on plant immunity21,23,25 and suggest a potential “cross-talk” between plant immune responses and water availability. Most of our current knowledge on plant-pathogens and plant-microbiome interactions are derived from studies under limited laboratory conditions. This study illustrates a need for future research to consider the dynamic climate conditions in which plants and microbes live in nature in order to uncover new biological phenomena involved in host-microbe interactions. Research that unravels the molecular bases of environmental influences of disease development should help us understand the severity, emergence and/or disappearance of infectious diseases in crop fields and natural ecosystems, especially in light of the dramatically changing drought/humidity patterns associated with global climate change. Methods Plant materials and bacterial strains Arabidopsis thaliana plants were grown in the “Arabidopsis Mix” soil (equal parts of SUREMIX [Michigan Grower Products Inc., Galesburg, MI], medium vermiculate and perlite; autoclaved once) or Redi-Earth soil (Sun Gro® Horticulture) in environmentally-controlled growth chambers, with relative humidity at 60%, temperature at 22 °C and 12h light/12h dark cycle. Five-week-old plants were used for bacterial inoculation and disease assays. The bak1-5/bkk1-1/cerk1mutant plant was generated by crossing the bak1-5/bkk1-1 mutant14 with the cerk1 mutant29. PCR-based genotyping was performed in F2 progeny to obtain a homozygous triple mutant. The npr1-6 (Fig. 4a) mutant was the SAIL_708_F09 line ordered from the Arabidopsis Biological Resource Center, and confirmed to be a knock-out mutant and defective in SA signaling (Extended Data Fig. 6). Bacterial disease assays Syringe-infiltration and dip-inoculation were performed. Briefly, Pst DC3000 and mutant strains were cultured in Luria-Marine (LM30) medium containing 100mg/L rifampicin (and/or other antibiotics if necessary) at 28°C to OD600 of 0.8 - 1.0. Bacteria were collected by centrifugation and re-suspended in sterile water. Cell density was adjusted to OD600 = 0.2 (~1×108 cfu/ml). For syringe-infiltration, bacterial suspension was further diluted to cell densities of 1×105 to 1×106 cfu/ml. Unless stated otherwise, infiltrated plants were first kept under ambient humidity for 1-2 h for water to evaporate, and, after the plant leaves returned to pre-infiltration appearance, plants were kept under high humidity (~95%; by covering plants with domes) or other specified humidity settings for disease to develop. For dip-inoculation, plants were dipped in the bacterial suspension of OD600 = 0.2, with 0.025% Silwet L-77 added, and then kept under high humidity (~95%) immediately for disease to develop. Different humidity settings were achieved by placing a plastic dome over a flat (in which plants are grown) with different degrees of opening. A humidity/temperature Data Logger (Lascar) was placed inside the flat to record the humidity and/or temperature over the period of disease assay. For quantification of Pst DC3000 bacterial populations, Arabidopsis leaves were surface-sterilized in 75% ethanol and rinsed in sterile water twice. Leaf disks were taken using a cork borer (9.5mm in diameter) and ground in sterile water. Colony-forming units were determined by serial dilutions and plating on LM plates containing 100mg/L rifampicin. Two leaf disks from two leaves were pooled together as one technical replicate, and 4 technical replicates are included in each biological experiment. Experiments were repeated at least three times. CRISPR-Cas9-mediated mutation of the MIN7 gene The one-plasmid CRISPR-Cas9 cloning system31 was used to mutate MIN7 in the fls2/efr/cerk1 and bak1-5/bkk1-1/cerk1 plants. MIN7-sgRNA primers containing target mutation regions were as follows, with MIN7 sequence underlined. MIN7-sgRNA-F: GATTGATCATTTGGAAGGGGATCC MIN7-sgRNA-R: AAACGGATCCCCTTCCAAATGATC The constructs containing MIN7-sgRNA and Cas9 were cloned in pCAMBIA1300, which were then mobilized into Agrobacterium tumefaciens for plant transformation. For genotyping of MIN7-mutated lines, total DNA was extracted from individual lines and the regions containing the CRISPR target sites were amplified by PCR using the following primers: MIN7-sgRNA-F2: GATGCTGCTTTGGATTGTCTTC MIN7-sgRNA-R2: AATGGCTCCCCATGCACTGCGATA For genotyping, the PCR products were digested by the BamHI restriction enzyme and plant lines showing an (partially or completely) uncut band were chosen. The PCR products of putative homozygous T2 lines, identified based on a lack of cutting by BamHI, were sequenced. The lines showing a frame-shift mutation and an absence of Cas9 gene based on PCR using the following primers were identified as homozygous lines. The T3 and T4 progeny of homozygous lines were used for disease assays. Primers for PCR-amplifying Cas9 gene: Cas9-F: CCAGCAAGAAATTCAAGGTGC Cas9-R: GCACCAGCTGGATGAACAGCTT Imaging of bacterial colonization with luciferase assay Four-week-old Arabidopsis Col-0 plants were dip-inoculated with Pst DC3000 or Pst DC3000-lux strain. The infected plants were fully covered with plastic dome to maintain high humidity. Leaves were excised from the infected plants 2 days post inoculation and the light signals were captured by a charge-coupled device (CCD) using ChemiDoc™ MP system (Bio-Rad). MIN7 protein blot Arabidopsis leaves were syringe-infiltrated with bacteria or H2O and kept under high humidity (~95%) for 24h. Leaf disks were homogenized in 2xSDS buffer, boiled for 5 min and centrifuged at 10,000 × g for 1 min. Supernatants containing the total protein extracts were subjected to separation by SDS-polyacrylamide gel electrophoresis (PAGE). A MIN7 antibody22 was used in the western blot to detect the MIN7 protein. Uncropped blot/gel images are included in Supplementary Figure 1. Bacterial community quantification Five-week old plants were sprayed with H2O and covered with a plastic dome to keep high humidity (~95%) for 5 days. To quantify the endophytic bacterial community, leaves were detached, sterilized in 75% ethanol for 1 min (Extended Data Fig. 9) and rinsed in sterile water twice. Leaves were weighed and ground in sterile water using a TissueLyser (Qiagen; at the frequency of 30 times per second for 1 min) in the presence of 3 mm Zirconium oxide grinding beads (Glen Mills; 5 beads in each tube). After serial dilutions, bacterial suspensions were plated on R2A plates, which were kept at 22°C for 4 days before colonies were counted. Colony-forming units were normalized to tissue fresh weight. 16S rRNA amplicon sequence analysis of endophytic bacterial community The Col-0, mfec and mbbc plants were sprayed with water and kept under high humidity (~95%) for 5 days. Leaves were surface-sterilized in 75% ethanol for 1 min and rinsed in sterile water twice. Leaves from four plants were randomly selected (2 leaves from each plants; 8 leaves in total) and were divided in 4 tubes (2 leaves in each tube) and ground in sterile water. Bacterial suspensions were diluted (Col-0 samples were diluted to 10−3 and mfec and mbbc samples were diluted to 10−5) and, for each genotype, 15 μl suspension from each tube of the right dilution (10−3 dilution for Col-0 and 10−5 for mfec and mbbc) were pooled together and plated on R2A plates, which were kept at 22°C for 4 days. Fifty colonies from each genotype were randomly picked and genomic DNA was extracted and PCR was performed with AccuPrime high-fidelity Taq DNA polymerase (Invitrogen) and primers 799F/1392R33 to amplify bacterial 16S rRNA gene. The PCR product was sequenced and taxonomy of each bacterium (family level) was determined by Ribosomal Database Project at Michigan State University (https://rdp.cme.msu.edu/)34. Data analysis, statistics and experimental repeats The specific statistical method used, the sample size and the results of statistical analyses are described in the relevant figure legends. Sample size was determined based on experimental trials and in consideration of previous publications on similar experiments to allow for confident statistical analyses. The Student’s two-tailed t-test was performed for comparison of means between two data points. One-way or two-way ANOVA with Tukey’s test was used for multiple comparisons within a dataset, with p value set at 0.05. ANOVA analysis was performed with the GraphPad Prism software. Data Availability The bacterial 16S rRNA sequences in Extended Data Table 1 have been deposited in the National Center for Biotechnology Information (NCBI) GenBank database under accession numbers KX959313-KX959462. Other data that support the findings of this study are available from the corresponding author upon request. Extended Data Extended Data Figure 1 Water soaking does not affect luminescence signal. Col-0 plants were dip-inoculated with bacteria at 2×108 cfu/ml, and kept under high humidity (~95%) for 2 days. Imaging was performed in the same way as in Fig. 1g. Water-soaked leaves were air-dried for about 2 h and imaged again (right panel). Images were representative of leaves from more than four plants. Extended Data Figure 2 a-b, The virulence of the avrE−/hopM1− mutant is insensitive to humidity settings. a, Col-0 plants were syringe-infiltrated with indicated bacteria at 2×105 cfu/ml. Inoculated plants were kept under high (~95%) humidity, and pictures were taken 24 h post infiltration. b, Col-0 plants were syringe-infiltrated with Pst DC3000, the avrE− mutant, the hopM1− mutant or the avrE−/hopM1− mutant at 2×105 cfu/ml. Inoculated plants were kept under high (~95%) or low (20-40%) humidity. Pictures were taken 3 days post inoculation. Images were representative of leaves from more than four plants. c-d, The 6xHis:HopM1 transgenic plants were infiltrated with 0.1 nM DEX, the avrE−/hopM1− mutant (at 1×105 cfu/ml) or both. H2O was infiltrated as control. Infiltrated plants were kept at high humidity (~95%). Leaf pictures were taken 24 h post infiltration (c) and bacterial populations were determined 3 days post infiltration (d). * indicates a significant difference, as determined by Student’s t-test; (two-tailed); ***, p=1.03×10−5. n=6 technical replicates from there independent experiments (n=2 in each experiment); error bars, mean±s.d. Extended Data Figure 3 Bacterial multiplication and water soaking in Col-0 and the min7 mutant. a, The Col-0 and min7 plants were dip-inoculated with Pst DC3000, the avrE−/hopM1− mutant or the hrcC− mutant at 1×108 cfu/ml. Bacterial populations were determined 4 days post inoculation. * indicates a significant difference between Col-0 and min7 plants, as determined by Student’s t-test (two-tailed); *, p=1.61×10−2 and 3.12×10−2 for DC3000 and hrcC−, respectively; ***, p=1.41×10−4 for avrE−/hopM1−. n=4 technical replicates; error bars, mean±s.d. Experiments were repeated three times. b-c, The Col-0 and min7 plants were syringe-infiltrated with Pst DC3000, the avrE− /hopM1− mutant or the hrcC− mutant at 1×106 cfu/ml. Bacterial populations were determined 3 days post inoculation (b) and leaf pictures were taken 38 h after infiltration to show water soaking in min7 leaves (c). * indicates a significant difference between Col-0 and min7 plants, as determined by Student’s t-test (two-tailed); **, p=1.63×10−3 for avrE−/hopM1−; ns, not significant (p=0.72 and 0.14 for DC3000 and hrcC−, respectively). n=3 technical replicates; error bars, mean±s.d. Experiments were repeated three times. Images were representative of leaves from more than four plants. Extended Data Figure 4 Pst DC3000 delivers a total of 36 effectors into the plant cell. Many effectors, including AvrPto, appear to suppress pattern-triggered immunity (PTI). AvrPto inhibits pattern recognition receptor (PRR) function8. Two conserved effectors, HopM1 and AvrE, create an aqueous apoplast in a humidity-dependent manner. AvrE is localized to the host plasma membrane (PM)23; its host target is currently unknown. HopM1 targets MIN7 (an ARF-GEF protein) in the trans-Golgi-network/early endosome (TGN/EE), which is involved in recycling of PM proteins26. Extended Data Figure 5 a, Col-0 leaves were syringe-infiltrated with Pst DC3000 (1×106 cfu/ml) or Pst DC3000 (avrRpt2) (1×107 cfu/ml). Plants were kept under high humidity (~95%) for 24 h to observe water soaking and then shifted to low humidity (~25%) for 2 h to observe ETI-associated tissue collapse. Pictures were taken before and after low humidity exposure (a) and bacterial populations were determined 24 h post infiltration to show similar population levels (b). * indicates a significant difference of bacterial population, as determined by Student’s t-test (two-tailed); *, p=0.033. n=3 technical replicates; error bars, mean±s.d. Experiments were repeated three times. This is an experimental replicate of Fig. 3b and 3c (without rps2). Extended Data Figure 6 Characterization of the npr1-6 mutant. a, A diagram showing the T-DNA insertion site in the npr1-6 mutant. Blue boxes indicate exons in the NPR1 gene. b, RT-PCR results showing that the npr1-6 line cannot produce the full-length NPR1 transcript. Primers used (NPR1 sequence is underlined): NPR1-F: agaattcATGGACACCACCATTGATGGA; NPR1-R: agtcgacCCGACGACGATGAGAGARTTTAC; UBC21-F: TCAAATGGACCGCTCTTATC; UBC21-R: TCAAATGGACCGCTCTTATC. Uncropped gel images are included in Supplementary Figure 1. c, The npr1-6 line, similar to npr1-1, is greatly compromised in benzothiadiazole (BTH)-mediated resistance to Pst DC3000 infection. The Col-0, npr1-1 and npr1-6 plants were sprayed with 100μM BTH and, 24 h later, dip-inoculated with Pst DC3000 at 1×108 cfu/ml. Bacterial populations were determined 3 days post inoculation. * indicates a significant difference between mock and BTH treatment, as determined by Student’s t-test (two-tailed); *, p=0.027; ***, p=1.6×10−4; ns, not significant (p=0.19). n=3 technical replicates; error bars, mean±s.d. Experiments were repeated three times. Extended Data Figure 7 Construction and characterization of the mfec and mbbc quadruple mutants. a, CRISPR-Cas9-mediated mutations in the 4th exon of the MIN7 gene (exons indicated by blue boxes) in the quadruple mutant lines used in this study. The underlined sequence in the wild type (WT) indicates the region targeted by sgRNA. The number “399” indicates the nucleotide position in the MIN7 coding sequence. “+1” and −1” indicate frame shifts in the mutant lines. b, Col-0 and various mutants used in this study have similar growth, development and morphology. Four-week-old plants are shown. c, The mfec and mbbc plants show a tendency of developing sporadic water soaking under high humidity. Five-week-old regularly-grown (~60% relative humidity) Col-0, mfec and mbbc plants were shifted to high humidity (~95%) for overnight and pictures of mature leaves were taken after high humidity incubation. d, Even leaves of mfec and mbbc plants that do not have sporadic water-soaking have a tendency to develop some water soaking after hrcC− inoculation. Five-week old Col-0, mfec and mbbc plants were dip-inoculated with hrcC− at 1×108 cfu/ml, and kept under high humidity (~95%). Leaf pictures were taken 2 days post inoculation. Images were representative of leaves from at least four plants. e, The non-pathogenic hrcC− mutant causes significant necrosis and chlorosis in the quadruple mutant plants. Col-0, mfec and mbbc plants were dip-inoculated with the hrcC− strain at 1×108 cfu/ml. Pictures were taken 9 days post inoculation. This is one of the four independent experimental repeats of the results presented in Fig. 5b. Extended Data Fig. 8 a, Increased endophytic bacterial community in the mfec and mbbc plants depend on high humidity. Col-0, mfec and mbbc plants were either sprayed with H2O and kept under high humidity (~95%) or kept under low humidity (~50%). On day 5, total populations of the endophytic bacterial community were quantified. Statistical analysis was performed by one-way ANOVA with Tukey’s test (p value set at 0.05). Bacterial populations indicated by different letters (i.e., a and b) are significantly different. n=4 technical replicates; error bars, mean±s.d. Experiments were repeated three times. b, Mild chlorosis and necrosis in leaves is associated with increased endophytic bacterial community level in the mfec and mbbc quadruple mutant plants. Plants were sprayed with H2O and kept under high (~95%) humidity. Pictures were taken 10 days after spray. Individual leaves are enlarged and shown in the lower panel, showing mild chlorosis and necrosis in some of the mfec and mbbc leaves. Extended Data Fig. 9 Validation of 1 min as an effective surface sterilization time. Five-week old Col-0 plants were sprayed with H2O and kept under high humidity (~95%) for 5 days. Leaves were detached, surface sterilized in 75% ethanol for 20s, 40s, 1min or 2min and then rinsed in sterile water twice. No sterilization (0s) was used as control. Leaves were ground in sterile water and bacterial numbers were determined by serial dilutions and counting of colony-forming units on R2A plates. Statistical analysis was performed by one-way ANOVA with Tukey’s test (p value set at 0.05). Bacterial populations indicated by different letters (i.e., a and b) are significantly different. n=4 technical replicates; error bars, mean±s.d. Experiments were repeated twice with similar results. Extended Data Table 1 Order/Family Col-0 mfec mbbc Bacillales  Paenibacillaceae 15 (30%) ND ND Burkholderiales  Comamonadaceae 8 (16%) 12 (24%) 9 (18%)  Burkholderiaceae 4 (8%) 1 (2%) 22 (44%)  Alcaligenaceae 3 (6%) 19 (38%) 12 (24%) Flavobacteriales  Flavobacteriaceae 6 (12%) 1 (2%) 1 (2%) Xanthomonadales  Xanthomonadaceae 4 (8%) 9 (18%) ND Sphingomonadales  Sphingomonadaceae 3 (6%) ND 1 (2%) Sphingobacteriales  Sphingobacteriaceae 3 (6%) ND ND  Chitinophagaceae 1 (2%) ND ND Rhizobiales  Rhizobiaceae 2 (4%) 5 (10%) ND Cytophagales  Cytophagaceae 1 (2%) ND ND Pseudomonadales  Pseudomonadaceae ND 1 (2%) 5 (10%) Actinomycetales  Microbacteriaceae ND 2 (4%) ND Supplementary Material Supplemental video1 Supplementary Video 1: A movie showing the process of Pst DC3000 infection of Arabidopsis plants. Five-week-old Col-0 plants were dip-inoculated with Pst DC3000 at 1×108 cfu/ml. Plants were kept under high humidity (~95%) and the disease symptoms were recorded over 4 days. The process was sped up by 8,640-fold (24 h to 10 seconds). The recording started 7 h after inoculation and the red arrow indicates one leaf, as an example, that showed the transient appearance of water soaking. 01 Supplementary Figure 1: Uncropped gel/blot images. Red boxes indicate cropped sections that are used in the main or Extended Data figures. Diagram in a indicates how the two gel blots in b and c were generated. Acknowledgements We thank He lab members for insightful discussions and constructive suggestions. We thank James Kremer for help with setting up real-time disease imaging experiments and advice on 16S rRNA amplicon sequencing, Koichi Sugimoto for providing tomato plants (cv. Castle Mart), and Caitlin Thireault for technical help. This project was supported by funding from Gordon and Betty Moore Foundation (GBMF3037), National Institutes of Health (GM109928) and the Department of Energy (the Chemical Sciences, Geosciences, and Biosciences Division, Office of Basic Energy Sciences, Office of Science; DE–FG02–91ER20021 for infrastructural support). C.Z acknowledges support from The Gatsby Charitable Foundation. Figure 1 Full-scale Pst DC3000 infection requires high humidity and is tightly associated with apoplast “water soaking”. See Methods for syringe-infiltration or dip-inoculation of plants described in all figures. a, Bacterial populations in Col-0, fls2/efr/cerk1 (fec), bak1-5/bkk1-1/cerk1 (bbc) and dde2/ein2/pad4/sid2 (deps) leaves 2 days post infiltration with bacteria at 1×106 cfu/ml. Humidity: ~95%. Two-way ANOVA with Tukey’s test (p value set at 0.05) was performed. No significant differences were found for DC3000 populations in different plant genotypes (indicated by the same letter a), whereas differences were found for hrcC− or DC3000D28E populations in different plant genotypes, as indicated by different letters of the same type (a’ vs. b’ for hrcC− and a” vs. b” for DC3000D28E). n=4 technical replicates; error bars, mean±s.d. Experiments were repeated three times with similar results. b-c, Bacterial populations (b) and disease symptoms (c) 3 days post infiltration with Pst DC3000 at 1×105 cfu/ml. * indicates a significant difference determined by Student’s t-test (two-tailed); ***, p=1.08×10−6. n=4 technical replicates; error bars, mean±s.d. Experiments were repeated four times with similar results. d, Bacterial populations in Col-0 leaves 3 days post infiltration with bacteria at 1×105 cfu/ml. Statistical analysis was the same as in a. Significant differences were found for DC3000 populations under different humidities, as indicated by different letters (a, b, c and d). No significant differences were found in hrcC− populations (indicated by the same letter a’). n=3 technical replicates; error bars, mean±s.d. Experiments were repeated three times with similar results. e, Pictures of the abaxial sides of Col-0 leaves 24 h post infiltration with Pst DC3000 at 1×106 cfu/ml. Humidity: ~95%. Dark spots on the leaf indicate water soaking spots. Red boxes indicate “zoomed-in” regions. f, Picture of a tomato leaf (cv. Castle Mart) 3 days after infiltration with Pst DC3000 at 1×104 cfu/ml. Humidity: ~95%. Yellow circles in e and f indicate infiltration sites. Images were representative of water-soaked leaves from more than four plants. g, Col-0 plants were dip-inoculated with bacteria at 2×108 cfu/ml. Humidity: ~95%. Bacterial colonies in inoculated leaves were visualized 2 days later by a charge-coupled device (upper panel) and pictures of leaves were taken to show water soaking spots (middle panel). Bottom panel shows merged images, with the artificial red color labeling Pst DC3000-lux bacteria. Experiments were repeated three times. Images were representative of leaves from more than four plants. Figure 2 Type III effectors AvrE and HopM1 are necessary and sufficient to cause water soaking. a, Pictures of Col-0 leaves 24 h post infiltration with bacteria (1-2×108 cfu/ml). Humidity: ~95%. b, Pictures of leaves of transgenic 6xHis:HopM122, 6xHis:AvrE23 or AvrPto32 plants after spray with 10μM dexamethasone (DEX; to induce effector gene expression). Humidity: ~95%. Col-0 or Col-0 gl plants were non-transgenic parental controls. Images were representative of leaves from more than four plants. c, Pictures of Col-0 leaves (left) and bacterial populations (right) 24 h post infiltration with Pst DC3000 (1×106 cfu/ml) or the avrE−/hopM1− strain (1×107 cfu/ml). Humidity: ~95%. Student’s t-test (two-tailed) was performed; ns, not significant (p=0.104). n=3 biological replicates; error bars, mean±s.d. Experiments were repeated three times. d, Bacterial populations in Col-0 plants 3 days post infiltration with bacteria at 2×105 cfu/ml. *** indicates a significant difference (p=1.07×10−6, 8.07×10−7 and 5.95×10−7 for DC3000, the avrE− mutant and the hopM1− mutant, respectively) of bacterial population between different humidities, as determined by Student’s t-test (two-tailed); ns, not significant (p=0.13). n=4 technical replicates; error bars, mean±s.d. Experiments were repeated three times. e-f, Bacterial populations (e) and leaf pictures (f) in Col-0 leaves 3 days post infiltration with bacteria at 1×105 cfu/ml. In the “− H2O” treatment, plants were air-dried normally (for ~2 h) and then kept under high humidity (~95%). In the “+H2O” treatment, plants were kept under high (80-95%) humidity after syringe-infiltration to allow slow evaporation of water (for ~16 h, until no visible apoplast water can be seen). ** (p=8.29×10−3 and 1.14×10−3 for DC3000 and hrcC−, respectively) and *** (p=7.61×10−7 and 9.82×10−4 for avrE−/hopM1− and CUCPB5452, respectively) indicate significant differences between “− H2O” and “+H2O” treatments as determined by Student’s t-test (two-tailed). n=3 technical replicates; error bars, mean±s.d. Experiments were repeated three times. Figure 3 Effects of MIN7 and effector-triggered immunity on water soaking. a, The min7 leaves, but not Col-0 leaves, showed partial water soaking 48 h after dip-inoculation with the avrE−/hopM1− mutant at 1×108 cfu/ml. Humidity: ~95%. Water soaking disappeared after transition to low humidity (~25%) to allow evaporation of apoplast water. Images were representative of leaves from more than four plants. b-c, ETI blocks apoplast water soaking. Col-0 and rps2 leaves were infiltrated with Pst DC3000 (1×106 cfu/ml) or Pst DC3000 (avrRpt2) (1×107 cfu/ml for Col-0 and 1×106 cfu/ml for rps2 plants). Plants were kept under high humidity (~95%) for 24 h to observe water soaking and then shifted to low humidity (~50%) for 4 h to observe ETI-associated tissue collapse. Pictures were taken before and after low humidity exposure (b) and bacterial populations were determined 24 h post infiltration to show similar population levels (c). Statistical analysis of data in c was performed by one-way ANOVA with Tukey’s test (p value set at 0.05), and no significant difference was detected. n=3 technical replicates; error bars, mean±s.d. Experiments were repeated three times. d, MIN7 protein is stabilized during ETI revealed by immunoblot. Col-0 or min7 leaves were infiltrated with bacteria (1×107 cfu/ml25) or H2O and kept under high humidity (~95%) for 24 h before protein extraction. Asterisk indicates a non-specific band. Coomassie blue staining shows equal loading. See Supplementary Figure 1 for cropping. Figure 4 hopM1/shcM transform the non-pathogenic DC3000D28E mutant into a highly virulent pathogen in PTI-deficient mutant plants in a humidity-dependent manner. a-c, Bacterial populations (a) and disease symptoms (b) 3 days post infiltration with bacteria indicated at 1×106 cfu/ml. Humidity: ~95%. Statistical analysis was performed by one-way ANOVA with Tukey’s test (p value set at 0.05). Bacterial populations indicated by different letters (i.e., a, b and c) are significantly different (ab is not significantly different from a or b). n=4 technical replicates; error bars, mean±s.d. Experiments were repeated three times. Water-soaking symptom was recorded 24 h post inoculation (c). d, Bacterial populations 3 days post infiltration with DC3000D28E (hopM1/shcM) at 1×106 cfu/ml under indicated humidities. Statistical analysis was the same as in (a). Bacterial populations indicated by different letters (i.e., a, b and c) are significantly different. n=4 technical replicates; error bars, mean±s.d. Experiments were repeated three times. Images were representative of leaves from at least four plants. Figure 5 Disease reconstitution experiments. a-b, The hrcC− bacterial populations 5 days (a) and disease symptoms 10 days post dip-inoculation (b) in Col-0, fec, bbc, min7, min7/fls2/efr/cerk1 (mfec) and min7/bak1-5/bkk1-1/cerk1 (mbbc) plants. Humidity: ~95%. Statistical analysis was performed by one-way ANOVA with Tukey’s test (p value set at 0.05). Bacterial populations indicated by different letters (i.e., a, b, c and d) are significantly different (ad is not significantly different from a or d). n=4 technical replicates; error bars, mean±s.d. Experiments were repeated four times. c, The hrcC− bacterial populations in Col-0 and bbc leaves 3 days post infiltration with bacteria at 1×106 cfu/ml. The “− H2O” and “+ H2O” conditions are the same as in Fig. 2e. Statistical analysis was performed by one-way ANOVA with Tukey’s test (p value set at 0.05). Bacterial populations indicated by different letters (i.e., a, b and c) are significantly different (ab is not significantly different from a or b). n=3 technical replicates; error bars, mean±s.d. Experiments were repeated three times. d, The Col-0, fec, bbc, min7, mfec and mbbc plants were mock-sprayed with H2O and kept under high humidity (~95%). On day 0 (before water spray) and day 5, total populations of the endophytic bacterial community were quantified by counting colony-forming units on R2A plates, after surface sterilization of leaves with 75% ethanol, leaf homogenization and serial dilutions. Statistical analysis is the same as in (a). Bacterial populations indicated by different letters (i.e., a and b) are significantly different. n=4 technical replicates; error bars, mean±s.d. Experiments were repeated three times. e, A new model for Pst DC3000 pathogenesis in Arabidopsis. Dashed arrows indicate a possible interplay, at spatial and temporal scales, between “immune suppression” and “wet apoplast” during pathogenesis. Supplementary Information is linked to the online version of the paper at www.nature.com/nature. Author Contributions X-F.X, K.N, and S.Y.H designed the experiments. K.A performed the Pst DC3000-lux imaging experiment. A.C.V performed biological repeats of bacterial infection experiments shown in Fig. 1a. J.Y characterized an unpublished plant mutant line. X-F.X and K.N performed all other experiments, including bacterial infections, protein blotting and generation of Arabidopsis mfec and mbbc mutant lines. F.B and C.Z contributed unpublished plant mutant materials. J.H.C contributed unpublished Pst DC3000 effector constructs. X-F.X and S.Y.H wrote the manuscript with input from all co-authors. Reprints and permissions information is available at www.nature.com/reprints. The authors declare no competing financial interests. Readers are welcome to comment on the online version of the paper. References 1 Miller S Rowe R Riedel R Bacterial spot, speck, and canker of Tomatoes 1996 Ohio State University Extension Fact Sheet HYG-3120-96 2 Pernezny K Zhang S Bacterial speck of tomato 2005 University of Florida IFAS Extension PP-10 3 Schwartz HF Bacterial diseases of beans 2011 Colorado State University Extension Fact Sheet No: 2.913 4 Stevens RB Plant Pathology, an Advanced Treatise 1960 3 Academic Press New York 5 Buttner D He SY Type III protein secretion in plant pathogenic bacteria Plant Physiol 2009 150 1656 1664 19458111 6 Galán J Collmer A Type III secretion machines: bacterial devices for protein delivery into host cells Science 1999 284 1322 1328 10334981 7 Asai S Shirasu K Plant cells under siege: plant immune system versus pathogen effectors Curr. Opin. Plant Biol 2015 28 1 8 26343014 8 Dou D Zhou JM Phytopathogen effectors subverting host immunity: different foes, similar battleground Cell Host Microbe 2012 12 484 495 23084917 9 Macho AP Zipfel C Targeting of plant pattern recognition receptor-triggered immunity by bacterial type-III secretion system effectors Curr. Opin. Microbiol 2015 23 14 22 25461568 10 Asrat S Davis KM Isberg RR Modulation of the host innate immune and inflammatory response by translocated bacterial proteins Cell Microbiol 2015 17 785 795 25850689 11 Sperandio B Fischer N Sansonetti PJ Mucosal physical and chemical innate barriers: Lessons from microbial evasion strategies Sem. Immunol 2015 27 111 118 12 Gimenez-Ibanez S Ntoukakis V Rathjen JP The LysM receptor kinase CERK1 mediates bacterial perception in Arabidopsis Plant Signal. Behav 2009 4 539 541 19816132 13 Macho AP Zipfel C Plant PRRs and the activation of innate immune signaling Mol. Cell 2014 54 263 272 24766890 14 Schwessinger B Phosphorylation-dependent differential regulation of plant growth, cell death, and innate immunity by the regulatory receptor-like kinase BAK1 PLoS Genet 2011 7 e1002046 21593986 15 Tsuda K Sato M Stoddard T Glazebrook J Katagiri F Network properties of robust immunity in plants PLoS Genet 2009 5 e1000772 20011122 16 Yuan J He SY The Pseudomonas syringae Hrp regulation and secretion system controls the production and secretion of multiple extracellular proteins J. Bacteriol 1996 178 6399 6402 8892851 17 Cunnac S Genetic disassembly and combinatorial reassembly identify a minimal functional repertoire of type III effectors in Pseudomonas syringae Proc. Natl. Acad. Sci. USA 2011 108 2975 2980 21282655 18 Hirano SS Upper CD Population biology and epidemiology of Pseudomonas syringae Annu. Rev. Phytopathol 1990 28 155 177 19 Fan J Crooks C Lamb C High-throughput quantitative luminescence assay of the growth in planta of Pseudomonas syringae chromosomally tagged with Photorhabdus luminescens luxCDABE Plant J 2008 53 393 399 17971037 20 Badel JL Shimizu R Oh HS Collmer A A Pseudomonas syringae pv. tomato avrE1/hopM1 mutant is severely reduced in growth and lesion formation in tomato Mol. Plant Microbe Interact 2006 19 99 111 16529372 21 DebRoy S Thilmony R Kwack YB Nomura K He SY A family of conserved bacterial effectors inhibits salicylic acid-mediated basal immunity and promotes disease necrosis in plants Proc. Natl. Acad. Sci. USA 2004 101 9927 9932 15210989 22 Nomura K A bacterial virulence protein suppresses host innate immunity to cause plant disease Science 2006 313 220 223 16840699 23 Xin XF Pseudomonas syringae effector Avirulence protein E localizes to the host plasma membrane and down-regulates the expression of the NONRACE-SPECIFIC DISEASE RESISTANCE1/HARPIN-INDUCED1-LIKE13 gene required for antibacterial immunity in Arabidopsis Plant Physiol 2015 169 793 802 26206852 24 Wei CF A Pseudomonas syringae pv. tomato DC3000 mutant lacking the type III effector HopQ1-1 is able to cause disease in the model plant Nicotiana benthamiana Plant J 2007 51 32 46 17559511 25 Nomura K Effector-triggered immunity blocks pathogen degradation of an immunity-associated vesicle traffic regulator in Arabidopsis Proc. Natl. Acad. Sci. USA 2011 108 10774 10779 21670267 26 Tanaka H Kitakura S De Rycke R De Groodt R Friml J Fluorescence imaging-based screen identifies ARF GEF component of early endosomal trafficking Curr. Biol 2009 19 391 397 19230664 27 Kim MG Two Pseudomonas syringae type III effectors inhibit RIN4-regulated basal defense in Arabidopsis Cell 2005 121 749 759 15935761 28 Hiruma K Root Endophyte Colletotrichum tofieldiae Confers Plant Fitness Benefits that Are Phosphate Status Dependent Cell 2016 165 464 474 26997485 29 Miya A CERK1, a LysM receptor kinase, is essential for chitin elicitor signaling in Arabidopsis Proc. Natl. Acad. Sci. USA 2007 104 19613 19618 18042724 30 Preston G Deng WL Huang HC Collmer A Negative regulation of hrp genes in Pseudomonas syringae by HrpV J. Bacteriol 1998 180 4532 4537 9721292 31 Feng Z Multigeneration analysis reveals the inheritance, specificity, and patterns of CRISPR/Cas-induced gene modifications in Arabidopsis Proc. Natl. Acad. Sci. USA 2014 111 4632 4637 24550464 32 Hauck P Thilmony R He SY A Pseudomonas syringae type III effector suppresses cell wall-based extracellular defense in susceptible Arabidopsis plants Proc. Natl. Acad. Sci. USA 2003 100 8577 8582 12817082 33 Bai Y Functional overlap of the Arabidopsis leaf and root microbiota Nature 2015 528 364 369 26633631 34 Wang Q Garrity GM Tiedje JM Cole JR Naïve Bayesian Classifier for Rapid Assignment of rRNA Sequences into the New Bacterial Taxonomy Appl Environ Microbiol 2007 73 5261 5267 17586664

PMC005xxxxxx/PMC5135085.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 100941354 21750 Nat Immunol Nat. Immunol. Nature immunology 1529-2908 1529-2916 26901152 5135085 10.1038/ni.3396 NIHMS832065 Article Stromal cells control the epithelial residence of DCs and memory T cells by regulated activation of TGF-β Mohammed Javed 1 Beura Lalit K 2 Bobr Aleh 3 Astry Brian 1 Chicoine Brian 1 Kashem Sakeen W 1 Welty Nathan E 1 Igyártó Botond Z 1 Wijeyesinghe Sathi 1 Thompson Emily A 2 Matte Catherine 45 Bartholin Laurent 6 Kaplan Alesia 7 Sheppard Dean 8 Bridges Alina G 3 Shlomchik Warren D 910 Masopust David 2 Kaplan Daniel H 11112 1 Department of Dermatology, Center for Immunology, University of Minnesota, Minneapolis, Minnesota USA 2 Department of Microbiology and Immunology, Center for Immunology, University of Minnesota, Minneapolis, Minnesota USA 3 Department of Laboratory Medicine and Pathology, Mayo Clinic, Rochester, Minnesota, USA 4 Department of Medicine, Yale University School of Medicine, New Haven, Connecticut, USA 5 Department of Immunobiology, Yale University School of Medicine, New Haven, Connecticut, USA 6 Centre de Recherche en Cancérologie de Lyon, INSERM U1052, CNRS UMR5286, Lyon, France 7 Department of Laboratory Medicine and Pathology, University of Minnesota, Minneapolis, Minnesota, USA 8 Department of Medicine, University of California, San Francisco, San Francisco, California, USA 9 Department of Medicine, University of Pittsburgh Cancer Center Institute, Pittsburgh, Pennsylvania, USA 10 Department of Immunology, University of Pittsburgh Cancer Center Institute, Pittsburgh, Pennsylvania, USA 11 Department of Dermatology, University of Pittsburgh, Pennsylvania, USA 12 Department of Immunology, University of Pittsburgh, Pennsylvania, USA Correspondence should be addressed to: D.H.K. (dankaplan@pitt.edu) 24 11 2016 22 2 2016 4 2016 01 4 2017 17 4 414421 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. Cells of the immune system that reside in barrier epithelia provide a first line of defense against pathogens. Langerhans cells (LCs) and CD8+ tissue-resident memory T cells (TRM cells) require active transforming growth factor-β1 (TGF-β) for epidermal residence. Here we found that integrins αvβ6 and αvβ8 were expressed in non-overlapping patterns by keratinocytes (KCs) and maintained the epidermal residence of LCs and TRM cells by activating latent TGF-β. Similarly, the residence of dendritic cells and TRM cells in the small intestine epithelium also required αvβ6. Treatment of the skin with ultraviolet irradiation decreased integrin expression on KCs and reduced the availability of active TGF-β, which resulted in LC migration. Our data demonstrated that regulated activation of TGF-β by stromal cells was able to directly control epithelial residence of cells of the immune system through a novel mechanism of intercellular communication. The skin is exposed to a wide variety of potential pathogens and commensal microorganisms. Keratinocytes (KCs) that form the stratified squamous epithelium of the epidermis create a physical barrier and also a niche for cells of the immune system that provide an active immunological barrier, notably Langerhans cells (LCs) and CD8+ tissue-resident memory T cells (TRM cells). LCs are a radioresistant, self-renewing subset of dendritic cell (DCs) that reside exclusively in the epidermis1. LCs migrate from the epidermis to the skin-draining lymph nodes (LNs), where they present antigen acquired in peripheral tissue to naive and central memory T cells1. Migration occurs both homeostatically and in response to microbial or inflammatory cues, including exposure to hapten, ultraviolet (UV) light and skin infection2,3. LCs are required for the induction of responses of the TH17 subset of helper T cells to specific cutaneous infections and also suppress skin immune responses in a variety of contexts4–9. TRM cells are a subset of memory T cells that maintain long-term residence in barrier tissues10. In the skin, CD8+ TRM cells reside in the epidermis and provide protective memory responses to infection with herpes simplex virus or vaccinia virus11–13. They are also thought to mediate autoimmune diseases such as vitiligo and alopecia areata14,15. Transforming growth factor-β1 (TGF-β) is a pleotropic cytokine that has been long considered an essential growth factor for LCs16. However, TGF-β signaling is also required for LCs to maintain their epidermal residence17–19. LCs successfully populate the epidermis in mice with LC-specific genetic ablation of TGF-β receptors (TGF-βRI or TGF-βRII) but spontaneously migrate to skin-draining LNs17,18. Notably, LC-specific ablation of TGF-β induces LC migration, which indicates that autocrine TGF-β is required for the epidermal residence of LCs17. The homeostasis of TRM cells also depends on TGF-β. CD8+ TRM cells unable to signal through TGF-β receptors fail to express integrin αEβ7 (CD103) and do not maintain residence in barrier epithelia20. TGF-β is secreted as a biologically inactive complex non-covalently bound to latency-associated peptide (LAP)21. Dissociation of LAP from TGF-β can be mediated by the integrins αvβ6 and αvβ8, which bind to an RGD (Arg-Gly-Asp) sequence in LAP; this allows TGF-β to become biologically active. In vivo, the activation of TGF-β by αvβ6 and αvβ8 in the epidermis is nonredundant22,23. Here we found that αvβ6 or αvβ8 expressed by KCs was required for the maintenance of epidermal LCs. αvβ6 and αvβ8 were expressed in non-overlapping patterns throughout the epidermis, and ablation or inhibition of αvβ6 or αvβ8 in KCs reduced the availability of active TGF-β, which resulted in loss of LCs from the epidermal region where that integrin was normally expressed. Similarly, residence of DCs in the epithelium of the small intestine and residence of CD8+ TRM cells in both intestinal epithelium and epidermis required αvβ6. In addition, UV irradiation reduced integrin expression by KCs and diminished the amount of activated TGF-β, which resulted in LC migration that was overcome by constitutive TGF-β signaling in LCs. Thus, regulated transactivation of LAP–TGF-β by KCs determined the epithelial residence of both DCs and TRM cells. RESULTS Epidermal maintenance of LCs in human skin requires TGF-β To determine whether the epidermal residence of LCs requires TGF-β signaling in human skin, we compared the number of epidermal LCs in skin samples obtained from untreated (control) patients with that in patients treated with losartan. Losartan is an antagonist to the angiotensin II type I receptor commonly used to treat hypertension but also attenuates TGF-β-mediated pathologies by inhibiting TGF-β signaling through an unknown mechanism24,25. The average number of LCs per linear mm of skin was reduced from 21.2 in age- and sex-matched control patients (Supplementary Table 1) to 14.3 in losartan-treated patients (Fig. 1a and Supplementary Fig. 1). Thus, the maintenance of LCs in human epidermis seemed to require TGF-β. αvβ6 inhibits homeostatic LC migration by activating TGF-β On the basis of the findings reported above, we hypothesized that homeostatic LC migration would require a loss of TGF-β signaling. To test our hypothesis, we bred huLangerin-CreERT2 mice (which have tamoxifen-inducible expression of Cre recombinase from the LC-specific gene encoding human langerin (huLangerin)) with TGF-βRI–CA mice (which express a loxP-STOP cassette followed by a mutant, constitutively active (CA), hemagglutinin-tagged form of TGF-βRI that signals independently of ligand) to generate TGF-βRI–CALC mice (with inducible expression of constitutively active TGF-βRI in LCs)26. Treatment of TGF-βRI–CALC mice with tamoxifen efficiently induced the expression of TGF-βRI–CA in LCs without affecting their viability (Supplementary Fig. 2a,b). Notably, ligand-independent TGF-β signaling resulted in a normal number of LCs in the epidermis but significantly fewer LCs in the LNs of TGF-βRI–CALC mice (Fig. 1b,c), indicative of a failure of homeostatic migration. As a comparison, ablation of TGF-β in LCs produced the opposite phenotype: enhanced migration (Fig. 1b,c). Since TGF-β is secreted as a latent complex, we next investigated whether its activation was required for the epidermal maintenance of LCs. Consistent with reports of reduced numbers of LCs in integrin β6–deficient (Itgb6−/−) mice22,23, we found spontaneous migration of LCs in Itgb6−/− mice (Fig. 1b,c). Similarly, in vivo blockade of αvβ6 activity in wild-type mice by intradermal injection of a neutralizing antibody resulted in local reduction in the number of LCs but not that of dermal DCs (Fig. 1d,e and Supplementary Fig. 2c). Notably, neutralization of αvβ6 in tamoxifen-treated TGF-βRI–CALC mice did not induce LC migration (Fig. 1d,e). Similarly, tamoxifen treatment largely restored the epidermal LC network in Itgb6−/− TGF-βRI–CALC mice (Supplementary Fig. 2d,e). Thus, αvβ6-mediated activation of TGF-β was needed to maintain LC residence. αvβ6 on interfollicular KCs controls epidermal LC residence Since αvβ6 is expressed by several epithelial tissues, we next investigated whether it was expressed by KCs in the epidermis. We noticed that KCs in wild-type mice expressed αvβ6, as assessed by immunofluorescence of epidermal mounts, but its expression was not uniform throughout the epidermis (Fig. 2a). We defined KCs on the basis of their location and relationship to the hair follicle, as interfollicular (IFE), isthmus (IM) and bulge KCs (Supplementary Fig. 3a). Precursors of LCs are recruited into the epidermis first through the IM before they populate the IFE region but are actively excluded from the bulge area27. Using surface markers to identify individual KC subsets27 (Supplementary Fig. 3b), we noted that αvβ6 expression was evident in the IFE and bulge regions but was absent from the IM (Fig. 2a,b,e). The residual LCs in Itgb6−/− mice seemed to ‘preferentially’ be located in the IM (Fig. 2a). To confirm that observation, we quantified LCs in the IFE region or IM in wild-type and Itgb6−/− mice. The number of LCs was significantly lower in the IFE region of Itgb6−/− mice than in that of wild-type mice but remained unaffected in the IM of Itgb6−/− mice (Fig. 2c). In addition, the few LCs that remained in the epidermis of Itgb6−/− mice were located in close proximity to the hair follicle (Fig. 2d). In the epidermis, Itgb6 mRNA expression was higher in IFE KCs than in IM KCs, but Itgb6 mRNA was detectable in LCs (Fig. 2e). To determine whether maintenance of epidermal LCs relied on expression of αvβ6 by KCs or LCs, we generated bone marrow (BM) chimeras by reconstituting irradiated wild-type(CD45.2+) recipient mice or Itgb6−/− (CD45.2+) recipient mice with congenic wild-type (CD45.1+) BM cells (wild-type→wild-type mice or wild-type→Itgb6−/− mice, respectively). A hematopoietic chimerism of >95% was achieved after 6 weeks in blood and LNs (Supplementary Fig. 4a,b). As expected, since LC are radio-resistant, they remained of host origin in wild-type mice (Fig. 2f). In contrast, donor-derived LCs were recruited into the epidermis of wild-type→Itgb6−/− mice, which resulted in 50% LC chimerism (Fig. 2f and Supplementary Fig. 4c). The frequency of host Itgb6−/− LCs and donor wild-type LCs in the epidermis were equivalent in wild-type→Itgb6−/− mice (Fig. 2f and Supplementary Fig. 4c), which indicated that expression of αvβ6 by LCs was not required for their retention in the epidermis. Instead, the absence of αvβ6 on KCs of wild-type→Itgb6−/− mice led to a failure of LC precursors to spread beyond the hair follicle, which resulted in the absence of LCs in the IFE region (Fig. 2g), similar to that seen in Itgb6−/− mice. Thus, expression of αvβ6 on IFE KCs controlled the epidermal residence of LC. Non-overlapping expression of αvβ6 and αvβ8 by KC subsets In addition to αvβ6, αvβ8 also activates LAP–TGF-β21. αvβ8 is expressed by DCs and has a non-redundant role in TGF-β-mediated development of TH17 cells and regulatory T cells28,29. Expression of αvβ8 by epithelial cells, however, has not been reported, to our knowledge. To investigate whether cells in the epidermis expressed αvβ8, we compared the expression of αvβ8-encoding mRNA on sorted populations of epidermal cells. Unexpectedly, we detected minimal expression of Itgb8 (which encodes β8) in LCs (Fig. 3a). Instead, we found that IM KCs had the highest expression of Itgb8, while IFE KCs had much lower expression of Itgb8 (Fig. 3a), an expression pattern reciprocal to that seen for Itgb6 (Fig. 2e). This raised the possibility that the presence of LCs in the IM of Itgb6−/− mice might have resulted from αvβ8-mediated activation of TGF-β. To test that hypothesis, we bred Itgb8loxP mice (which have loxP-flanked Itgb8 alleles) with huLangerin-Cre mice (which express Cre recombinase from the LC-specific gene encoding human langerin) to generate mice with LC-specific ablation of Itgb8 (Itgb8ΔLC), or with K14-Cre mice (which express Cre recombinase from the KC-specific gene encoding keratin 14) to generate mice with KC-specific ablation of Itgb8 (Itgb8ΔKC). The number of LCs in the epidermis and LNs was unaltered in Itgb8ΔLC mice relative to that in wild-type mice (Supplementary Fig. 5a–c). In contrast, the number of LCs in skin-draining LNs was elevated and the number of LCs was decreased in IM of the epidermis of Itgb8ΔKC mice, relative to that in wild-type mice (Fig. 3b,c). Notably, the number of LCs was reduced in the IM of the hair follicle but remained unaltered in the IFE region in Itgb8ΔKC mice, relative to that in wild-type mice, as assessed by immunofluorescence imaging of whole tail mounts and transverse sections of back skin (Fig. 3d,e). To exclude the possibility that TGF-β might be activated though mechanisms other than those involving αvβ6 or αvβ8, we crossed Itgb8ΔKC mice with Itgb6−/− mice. The resulting Itgb8ΔKCItgb6−/− mice lacked expression of both αvβ6 and αvβ8 (data not shown) and were devoid of epidermal LCs (Fig. 3f,g), which confirmed the absence of compensatory mechanisms for activation of TGF-β in the epidermis. Finally, we directly investigated the necessity for αvβ6 and αvβ8 in the KC-mediated activation of latent TGF-β using an in vitro assay for activated TGF-β30. Primary mouse KCs were co-cultured with mink lung reporter cells that had been transfected with a TGF-β-responsive promoter that drives the expression of luciferase. The ability of KCs to activate latent TGF-β was significantly reduced by antibody blockade of αvβ6 in both wild-type cells and Itgb8ΔKC cells, relative to its activation in their counterparts treated with control antibody (Fig. 3h). Similarly, the amount of active TGF-β was lower in cultures of KCs derived from Itgb8ΔKC mice than in cultures of KCs derived from wild-type mice and was further reduced by blockade of αvβ6 (Fig. 3h). Together these data supported a model in which inactive LAP–TGF-β secreted by LCs is converted into an active form through the action of either αvβ6 or αvβ8 expressed by spatially distinct subsets of KCs, which results in the maintenance of LCs in the epidermis. Intraepithelial DCs expressing integrin αEβ7 (CD103) in the small intestine have been reported31. Unlike the skin, the epithelium of the small intestine expressed αvβ6 but not αvβ8 (Supplementary Fig. 6a). Blockade of αvβ6 via antibody administered weekly for 4 weeks resulted in a reduction in the number of intraepithelial CD103+ DCs, as assessed by immunofluorescence imaging, but did not alter the number of DCs in the lamina propria (Supplementary Fig. 6b–d). Thus, residence of DCs in both the skin and intestinal barrier surfaces required epithelial αvβ6. UV irradiation diminishes TGF-β activation by KCs In addition to their homeostatic migration, LCs migrate in response to exogenous stimuli such as exposure to UV irradiation2. We next sought to determine whether reduced release of TGF-β from LAP by KCs expressing αvβ6 or αvβ8 controlled UV-induced migration. We first exposed the shaved skin of tamoxifen-treated wild-type and TGF-βRI–CALC mice to UVB at a dose of 100 mJ/cm2. After 4 d, we observed a nearly complete absence of LCs, as assessed on the basis of expression of major histocompatibility complex (MHC) class II and langerin, in the epidermis of UVB-treated wild-type mice (Fig. 4a,b). The few remaining MHCII+ cells had very low expression of langerin protein (data not shown), which identified them as short-term monocyte-derived cells and not true, long-term LCs32. In contrast, many langerin-positive LCs remained in the epidermis of UVB-treated TGF-βRI–CALC mice (Fig. 4a,b). A similar result was obtained by epicutaneous application of the fluorescent hapten TRITC (Supplementary Fig. 7). Notably, LCs in UVB-treated TGF-βRI–CALC mice had higher expression of the costimulatory molecule CD86 than that of LCs from their counterparts not treated with UVB, indicative of some degree of activation, but failed to increase their expression of CCR7, the chemokine receptor required for the migration of DCs into draining LNs (Fig. 4c). Ex vivo KCs from the UVB-irradiated skin of wild-type mice had lower expression of Itgb6 in IFE KCs and of Itgb8 in IM KCs than that of their counterparts from wild-type mice not treated with UVB (Fig. 4d). Similarly, UVB irradiation reduced the expression of both Itgb6 and Itgb8 in primary cultured wild-type KCs and also reduced the activation of TGF-β by UV irradiation–exposed primary KCs, as measured with luciferase reporter cells (Fig. 4e–g). Thus, in the setting of UV irradiation, KCs reduced their expression of αvβ6 and αvβ8, which resulted in smaller amounts of active TGF-β. That in turn led to a failure of LCs to maintain epidermal residence that was overcome by constitutive TGF-β signaling. Epidermal CD8+ TRM cell residence requires αvβ6 and αvβ8 Studies have demonstrated that local TGF-β is essential for the persistence of TRM cell precursors in the epidermis20. CD103 (encoded by Itgae) pairs with integrin β7 to bind the ligand E-cadherin in various epithelial tissues. TGF-β induces CD103 expression on CD8+ T cells in the small intestine, where CD103 is required for the retention of CD8+ TRM cells in the epithelium but not in the underlying lamina propria33–35. On the basis of those findings and our own observations, we hypothesized that in addition to maintaining the epidermal residence of LCs, the integrin-mediated activation of TGF-β might also be required for the retention of epidermal TRM cells. To test our hypothesis, we adoptively transferred Thy-1.1+ lymphocytic choriomeningitis virus (LCMV)-specific P14 CD8+ T cells into wild-type, Itgb6−/−, Itgb6−/−Itgb8ΔKC or LC-deficient mice, infected the host mice with LCMV (Armstrong strain) and applied the hapten DNFB (an established technique for seeding the epidermis with P14 T cells) topically onto the skin of the host mice20. At day 42 after infection, Thy-1.1+ P14 TRM cells were readily detectable in wild-type host mice by immunofluorescence microscopy of epidermal whole mounts (Fig. 5a,b). In contrast, TRM cells were much less frequent in the epidermis of Itgb6−/− mice and were completely absent from that of Itgb6−/−Itgb8ΔKC mice, despite a normal number of these cells in both blood and LNs of Itgb6−/− and Itgb6−/−Itgb8ΔKC mice (Fig. 5a,b and Supplementary Fig. 8a). Host mice that lacked LCs had a normal number of TRM cells (Fig. 5a,b and Supplementary Fig. 8a), which indicated that the retention of TRM cells in the epidermis was independent of LCs, consistent with published reports36. To exclude the possibility that T cells failed to enter the epidermis of Itgb6−/− and Itgb6−/−Itgb8ΔKC mice, rather than persisting there, we first assessed the number of P14 cells using a strategy similar to that described above, except we analyzed mice at day 7 after infection, which coincided with the effector phase of the T cell response. We observed an equivalent number of P14 cells in the skin, blood and LNs of wild-type, Itgb6−/− and Itgb6−/−Itgb8ΔKC host mice (Fig. 5c,d and Supplementary Fig. 8b). Although P14 cells entered the epidermis in similar numbers in all mice, the frequency of such cells that expressed CD103 was significantly lower in Itgb6−/− and Itgb6−/−Itgb8ΔKC mice than in wild-type mice (Fig. 5e), consistent with a requirement for active TGF-β for the expression of CD103 by TRM cells20,37. In addition, we treated wild-type recipient mice weekly with neutralizing antibody to αvβ6 for 4 weeks starting 56 d after infection, once TRM cell residence had been established. Neutralization of αvβ6 in wild-type mice efficiently reduced the number of P14 cells in the epidermis but not in the LNs or blood (Fig. 5f,g and Supplementary Fig. 8c), which demonstrated that the epithelial residence of TRM cells required αvβ6. Thus, αvβ6 and αvβ8 were required for the development and maintenance of CD8+ TRM cells in epidermis. Residence of intestine epithelial CD8+ TRM cells requires αvβ6 Since the epithelium of the small intestine expressed αvβ6 and blocking αvβ6 function in vivo led to a reduced number of intraepithelial DCs, we next investigated whether the residence of TRM cells in intestinal epithelium also required αvβ6. Using an experimental model similar to that described above to generate LCMV-specific TRM cells, we analyzed wild-type and Itgb6−/− mice host mice at both a memory time point (day 42 after infection) and an effector time points (day 7 after infection) to quantify P14 cells in the epithelium and lamina propria of the small intestine. Consistent with our findings obtained for the epidermis, there was a significantly reduced number of TRM cells, with no difference in the entry of P14 cells, during the effector phase in the intestinal epithelium of Itgb6−/− mice, compared with that of wild-type mice (Fig. 6a,b,d,e). The number of P14 cells at the memory and effector phases in intestinal lamina propria was similar in Itgb6−/− mice and wild-type mice (Fig. 6c,f). Neutralization of αvβ6 in wild-type mice after TRM cells had been established (day 56 after infection) efficiently reduced the number of P14 cells in the intestinal epithelium but not in the lamina propria (Fig. 6g–i), which demonstrated that αvβ6 was required for the maintenance of TRM cells in the epithelium of the small intestine even after TRM cells had been established. DISCUSSION Here we have demonstrated that regulated activation of LC-derived LAP–TGF-β by KCs determined whether LCs migrated from the epidermis. We found that KCs expressed αvβ6 and αvβ8 in non- overlapping patterns and that deletion or blockade of integrins reduced the availability of active TGF-β, which resulted in loss of LCs from the epidermal region in which that integrin was normally expressed. The migration of LCs in response to UV irradiation required loss of TGF-β signaling in LCs and was accompanied by reduced expression of integrins by KCs. We also found that expression of αvβ6 was required for the persistence of TRM cells in the epidermis and the epithelium of the small intestine. Together these data demonstrate a key role for stromal cells in determining the residence of cells of the immune system within barrier epithelia. LC migration is thought to result from a combination of LC-intrinsic activation through engagement of pattern-recognition receptors and inflammatory cytokines such as IL-1β and TNF38. Our data obtained with TGF-βRI–CALC mice demonstrated that loss of TGF-β signaling in LCs was required for homeostatic migration and for both UVB-induced migration and hapten-induced migration. Loss of integrin expression by KCs was also sufficient to induce LC migration. Notably, homeostatic and inflammation-induced LC migration is unaffected by the absence of the adaptor MyD88 in LCs3,39,40. Thus, we speculate that inflammatory mediators known to drive LC migration might act indirectly through KCs and that reduced activation of TGF-β by KCs is a general mechanism that prompts LC migration. Whether KCs modulate integrin expression in the steady state or in response to inflammatory mediators other than those we have tested, however, remains to be determined. TRM cell differentiation is induced by environmental factors, of which TGF-β remains the best characterized. As with LCs, TRM cells also depended on the transactivation of TGF-β by KCs for epidermal residence. The number of TRM cells was reduced in Itgb6−/− and Itgb6−/−Itgb8ΔKC mice, and antibody blockade of αvβ6 reduced the number of epidermal TRM cells. The recruitment of effector cells into the epidermis was unaffected by the absence of these integrins, but the frequency of cells expressing CD103 was reduced. These results are consistent with a requirement for TGF-β during TRM cell development but also support a model in which tonic exposure to TGF-β, regulated by local TGF-β-activating integrin expression, is constitutively required for their long-term epidermal retention. Notably, the number of TRM cells was unaffected by the absence of LCs, which indicated that the LCs were not the source of TGF-β that maintained the TRM cells. Dendritic epidermal T cells produced copious amounts of TGF-β (data not shown) but have been reported to remain spatially distinct from TRM cells, so they are unlikely to be critical for the maintenance of TRM cells in the epidermis36. KCs are also a major source of TGF-β and probably support epidermal TRM cell residence, although the possibility that TGF-β from TRM cells themselves is involved cannot be excluded. It remains unclear why LCs depend on autocrine TGF-β and do not have access to TGF-β from these other sources. Epithelial TRM cells in the small intestine require TGF-β for CD103 expression and development33,34,41. Unlike integrin expression in the epidermis, αvβ6 but not αvβ8 was expressed by epithelial cells in the small intestine, consistent with published findings42. The absence of TRM cells in Itgb6−/− mice and after neutralization of αvβ6 demonstrated that, as in the skin, epithelial TRM cells in the intestine also required αvβ6 for epithelial retention and probably also for their development. Since αvβ6 is widely expressed by many barrier epithelia, TRM cell residence in other tissues might have a similar requirement for this integrin. αvβ8 has been best studied in DCs, in which it provides non-redundant activation of TGF-β and promotes the development of TH17 cells and regulatory T cells28,29. The unexpected expression of αvβ8 by IM KCs raised the possibility that differential regulation of Itgb6 and Itgb8 by distinct subsets of KCs might allow regional variation in the leukocyte occupancy of the epidermal niche. A published report has shown that CD103+ DCs are sparsely distributed in intestinal epithelium during steady state and are recruited from the lamina propria in response to infection31. It remains unclear whether steady-state CD103+ DCs represent a true population of intraepithelial DCs or a population that continually migrates between the epithelium and lamina propria. We observed fewer epithelial CD103+ DCs than lamina propria CD103+ DCs after inhibition of αvβ6, which suggested that these cells also depended on continuous TGF-β signaling to maintain CD103 expression, similar to TRM cells; thus, these findings potentially extend our findings to epithelial DCs beyond the epidermis. In summary, we have defined a novel mechanism of retention of DCs and TRM cells in two barrier epithelia through the regulated activation of TGF-β. Compounds that block TGF-β and inhibit αvβ6 are in clinical trials as cancer therapeutics and to treat fibrosis during lung, liver and kidney diseases43. These compounds might have utility in treating diseases in which TRM cells or LCs are pathogenic. In particular, clearing an established population of TRM cells could be especially effective, since they do not circulate and are unlikely to be replenished without the recruitment of additional CD8+ effector cells. In addition, regulated expression of integrins and activation of TGF-β by stromal cells might participate in numerous other TGF-β-mediated processes, including immunological, carcinogenic, tissue-repair, aging and developmental processes. Whether the regulated transactivation of TGF-β we observed in the epidermis and intestinal epithelium occurs in other barrier and non-barrier tissues as well as other contexts remains to be investigated. ONLINE METHODS Mice huLangerin-Cre, huLangerin-CreERT2, TGF-βloxP and TGF-βRI–CA and huLangerin-DTA mice have been previously described17,26,44. huLangerin-CreERT2 mice were crossed to the TGF-βRI–CA line to generate TGF-βRI–CALC mice. Itgb6−/− and Itgb8loxP mice were provided by D. Sheppard (University of California, San Francisco). C57BL/6 (wild-type) and K14-Cre mice were purchased from Jackson Laboratories. K14-Cre and huLangerin-Cre mice were crossed with Itgb8loxP mice to obtain Itgb8ΔKC mice and Itgb8ΔLC mice, respectively. We used age-matched female mice that were between 4 and 12 weeks of age in all our experiments. All mice were housed and bred in microisolator cages and were fed irradiated food and acidified water. The University of Minnesota Institutional Care and Use Committee approved all the experimental procedures on mice. Antibodies Antibodies 6.3g9 (used for detection of αvβ6 by flow cytometry and neutralization of αvβ6)45 and ch2A1 (used for detection of αvβ6 in mouse epidermis by immunofluorescence) were provided by S. Violette (Biogen Idec); 6.3g9 was also obtained from American Type Culture Collection. Mouse anti-hemagglutinin (6E2) was purchased from Cell Signaling Technologies. Zenon mouse IgG1 labeling kit (Life Technologies) was used to label 6E2 and 6.3G9 with biotin. Streptavidin–phycoerythrin (PE) and streptavidin–PE–indotricarbocyanine (Cy7) (Life Technologies) were used to detect biotin-labeled antibodies. Anti–collagen IV (AB769) was purchased from EMD Millipore. Fluorochrome-conjugated antibodies to mouse CD3e (145-2c11), CD11b (M1/70), CD11c (N418), CD45.1 (A20), CD45.2 (104), CD80 (16-10A1), CD86 (GL1), CD103 (2E7), Epcam (G8.8), MHCII (M5/114.15.2), Sca-1 (D7) and Thy-1.1 (OX-7) and unconjugated anti–human CD1a (HI149) were purchased from BioLegend. Antibodies to mouse langerin (L31), CCR7 (4B12) and CD49f (eBioGoH3) were purchased from eBioscience. Anti–mouse CD34 (RAM34) was purchased from BD Biosciences. Anti–human langerin (12D6) was purchased from Abcam. Fluorescein-conjugated goat antibody to green fluorescent protein (800-102-215) was purchased from Rockland Immunochemicals. All antibodies and streptavidin reagents were used at a dilution of 1:200, except for anti-CCR7 (1:20), anti-langerin (L31) (1:400) and anti-Thy-1.1 (OX-7) (1:1,000). All antibodies and their dilutions were previously validated in the lab or as recommended by the manufacturer. Tamoxifen treatment Tamoxifen (T5648; Sigma-Aldrich) was dissolved in 1/10th volume of 200 proof ethanol following incubation at 37 °C for 15–30 min with occasional vortexing. Corn oil (Sigma-Aldrich) was added for a final concentration of 10 mg/ml and was administered to mice for 5 consecutive days by intraperitoneal injection at 0.05 mg per g of mouse weight. Histology of human and mice skin Approval was obtained from Mayo Clinic Institutional Review Board (IRB#13-005614) to search the Mayo Clinic electronic records for patients who were treated with the angiotensin receptor blocker losartan and had undergone excisional biopsy performed for a variety of dermatologic problems. The control group was selected as patients who underwent excisional biopsy and did not receive losartan. Each excisional biopsy had tips of the excision that usually represented unaffected normal tissue, which was confirmed by hematoxylin-and-eosin staining. Slides with pathology and excisions of soles and palms in both groups were excluded. Archived formalin-fixed, paraffin-embedded human skin sections were deparaffinized and subjected to an antigen-retrieval procedure by microwave treatment in citrate buffer (pH 6). After sections cooled to 25 °C, they were blocked with PBS containing 3% BSA and 5% goat serum for 60 min at room temperature. Sections were incubated overnight at 4 °C with anti-CD1a or anti–human langerin (both identified above) and washed, and bound antibodies were detected with horseradish peroxidase–conjugated anti-mouse IgG (115-036-062; Jackson ImmunoResearch). Bound secondary antibodies were detected using a Diaminobenzidine Peroxidase Substrate Kit (Vector) following manufacturer’s instructions and were counterstained using Gills hematoxylin (Vector). Epidermal LCs were counted and are presented as linear density number per mm. Mouse skin was embedded in OCT compound and frozen in liquid nitrogen. Transverse mouse skin sections that were 7-μm thick were prepared from frozen blocks using a cryostat and were fixed for 5 min in chilled acetone. Sections were blocked with PBS containing 0.1% tween, 3% BSA and 2% rat normal serum for 60 min before staining overnight with Alexa Fluor 488–conjugated anti–mouse MHC class II and Alexa Fluor 647– conjugated anti-langerin (both identified above) in PBS containing 0.1% tween and 3% BSA. DAPI (4,6-diamidino-2-phenylindole) was used for nuclear counterstaining. Images were captured using Leica DM5500 epifluorescent microscope with digital system and LAS AF software (version 1.5.1). Immunofluorescence of epidermis Epidermal whole mounts for immunofluorescence were prepared as previously described8. Epidermal sheets were prepared from ear, back and tail skin by affixing of the epidermis side to slides with double-sided adhesive (3M). Slides were incubated in 10 mM EDTA in PBS for 2 h at 37 °C, followed by physical removal of the dermis. Epidermal mounts were fixed in chilled acetone for 5 min. They were subsequently stained and images captured similar to transverse skin sections mentioned above. LCs and TRM cells in epidermis were counted either by ImageJ software or manually in a blinded fashion. Freezing, immunofluorescence and microscopy of small intestine Harvested mouse small intestines were frozen as described46. Specimens were embedded in the tissue-freezing medium OCT and were snap-frozen in an isopentane liquid bath. Frozen blocks were cut to prepare 7-μm-thick sections using a Leica cryostat. Immunofluorescence microscopy was performed using a Leica DM5500 B microscope six-color fluorescent system with motorized z-focus stage for fully automated image stitching. Separate images taken with a 20× objective were collected for each channel and overlaid to obtain a multicolor image. Image processing and enumeration was done using ImageJ64 and Adobe Photoshop (version 6) as described47. Enumeration of MHCII+ DCs and P14 CD8+ T cells was done manually in Adobe Photoshop, and ImageJ64 software was used to enumerate nuclei in each image. All software-based counts were periodically manually validated. Flow cytometry Epidermal and LN single-cell suspensions were prepared for flow cytometry as previously described8. Single-cell suspensions of epidermal cells were obtained from trunk or ear skin and were incubated for 2 h at 37 °C in 0.3% trypsin (Sigma-Aldrich) in 150 mM NaCl, 0.5 mM KCl and 0.5 mM glucose. The epidermis was physically separated from the dermis and disrupted by mincing and vigorous pipetting. The resulting cells were filtered through a 40-μm filter. LNs (axillary, brachial and inguinal) were incubated in 400 U/ml collagenase D (Roche Applied Science) for 30 min before filtration through a 40-μm filter. Single-cell suspensions were pretreated for 10 min with blocking antibody to CD16/CD32 (2.4G2; American Type Culture Collection) and were stained with antibodies to extracellular markers (all identified above) at 4 °C. For staining of CCR7, epidermal single cells were incubated at room temperature for 60 min. The fixable viability dye eFluor 780 (eBioscience) was used for live-dead discrimination. Intracellular staining of langerin, αvβ6 and green fluorescent protein was performed with a BD Bioscience Cytofix/Cytoperm kit (BD Biosciences) in accordance with the manufacturer’s instructions. Anti-αvβ6 and anti-hemagglutinin (identified above) were labeled with biotin using the zenon mouse IgG1 labeling kit followed by incubation with streptavidin-PE or streptavidin-PE-Cy7. Samples were analyzed on LSRII flow cytometers (BD Biosciences). Data were analyzed with FlowJo software (TreeStar, Ashland, OR). Generation of chimeric mice 6-week-old Itgb6−/− and wild type C57BL/6 CD45.2 mice were lethally irradiated using a X-ray irradiator and received two split doses of 500 cGy each. The following day, BM cells were prepared following erythrocyte lysis using ACK lysis buffer (Biowhittaker) from congenically marked C57BL/6 CD45.1 mice, and 5 × 106 BM cells injected intravenously into irradiated mice. Mice were rested for at least 6 weeks before experiments. The efficiency of chimerism was determined by flow cytometry of congenic markers on peripheral blood mononuclear cells, spleen, LN and epidermis (antibodies identified above). Neutralization of αvβ6 activity in vivo TGF-βRI–CALC and wild-type control mice were treated with tamoxifen as described above and injected intradermally with 20 μl in ear of 1 mg/ml anti-αvβ6 (6.3g9) or isotype-matched control antibody (ADWA-21; provided by D. Sheppard) under ketamine and xylazine sedation (100 mg and 10 mg/kg body weight, respectively) on days 2 and 5 of tamoxifen treatment. Epidermis was analyzed 3 d following the second antibody injection. In some experiments, 3g9 was administered intra-peritoneally at a dose of 10 mg/kg weekly for up to 4 weeks. UVB irradiation of mice TGF-βRI–CALC and wild-type control mice were treated with tamoxifen as described above. Mice were sedated and backs shaved using hair clippers on third day of tamoxifen treatment. Shaved backs were exposed to 100 mJ/cm2 UV-B radiation using two TL 20W/12RS bulbs (Philips). Tissues were harvested 4 d after UV irradiation. Epidermal cell sorting Epidermal single cells were prepared as described above. Dead cells were gated out using fixable viability dye efluor 780 (eBioscience). Different epidermal cell types were identified and sorted on FACSAria cell sorter using the following markers (antibodies identified above): dendritic epidermal T cells, CD45+CD3e+; LCs, CD45+MHCII+; IFE KCs, CD45−Sca1+CD49f+CD34−; IM KCs, CD45−Sca1−CD34−Epcam+; bulge KCs, CD45−Sca1−CD34+. Purity of the sorted cells was determined by post-sort analysis and consistently exceeded 95%. Cell culture Mink lung epithelial reporter cells stably transfected with a plasmid containing luciferase-encoding cDNA downstream of a TGF-β-sensitive portion of the plasminogen activator inhibitor 1 promoter (tMLEC) were cultured as previously described30. Primary KCs from newborn mice were prepared and cultured according to established method48. Skin was floated overnight with dermis side down at 4 °C in 0.25% trypsin (25 050-Cl; Corning Cellgro). The following day, epidermis was separated from dermis, and single cells were prepared by mincing and vigorous pipetting. After filtration using 100-μm filters, epidermal cells were plated overnight in Eagle’s MEM (06-174G, Lonza) supplemented with 1.4 mmol/L CaCl2 and 8% chelexed FCS. After 16–20 h of culture, the dishes were washed twice with PBS and Eagle’s MEM with 8% chelexed serum and 0.05 mmol/L CaCl2 (KC medium) was added that was replaced every 2 d. Primary KCs were cultured to at least 90% confluence before being treated with 50 mJ/cm2 UVB radiation. RNA isolation and quantitative PCR Total RNA from flow cytometry–sorted epidermal cells and primary KC cultures was extracted with the RNeasy Mini extraction kit (Qiagen) following the manufacturer’s instructions and was quantified using Nanodrop (NanoDrop). cDNA was generated using a High Capacity cDNA Reverse Transcription Kit (Applied Biosystems) and was subjected to quantitative PCR using TaqMan Gene Expression Master Mix and TaqMan Gene Expression Assays for Hprt, Itgb6, Itgb8, Itgav and Fermt1 (Kindlin-1). ABI Prism 7900HT (Applied Biosystems) was used to complete the quantitative PCR. All protocols while using the kits were completed according the manufacturer’s instructions. All cycling threshold (Ct) values were normalized to Hprt expression. Latent TGF-β-activation assay Latent TGF-β activation by KCs was determined by coculture with thymic mink lung epithelial cells (tMLECs) as described49. tMLEC were resuspended in DMEM containing 10% FCS and were plated at 1.6 × 104 cells/well of a 96-well cell culture treated plate for 3 h at 37 °C and 5% CO2. Control primary KCs and UVB-exposed KCs were harvested from the culture dish by gentle trypsinization using 0.25% trypsin and 2.21 mM EDTA (25 053-Cl; Corning Cellgro). KCs were resuspended at 2 × 105 cells/ml in KC medium with 1% chelexed serum, and 100 μl was added to tMLEC cultures after replacement of the medium. Separate groups of KCs were pre-incubated with 10 μg/ml anti-αvβ6 (6.3g9) for 60–90 min before addition to tMLEC cultures. The cells were cultured for 16–20 h, following which the reporter cells were lysed and assayed for luciferase activity using Bright Glo Luciferase Assay System (Promega). Generation of P14 immune chimeras and treatments 5 × 104 naive Thy-1.1+ P14 T cells (CD8+ T cells with transgenic expression of a T cell antigen receptor specific to the LCMV epitope of glycoprotein amino acids 33–41) were transferred intravenously into wild-type, Itgb6−/− or Itgb6−/−Itgb8ΔKC mice on day -1, followed by infection with LCMV Armstrong strain (2 × 105 plaque-forming units per mouse) on day 0. The contact-sensitizing agent 2,4-dinitrofluorobenzene (DNFB) was applied on skin at day 3 to pull effector P14 cells to the site of sensitization. For this, a ~1-cm2 area on mice skin was shaved, and 20 μL of 0.2% DNFB in acetone–olive oil (4:1) was applied. Tissues were harvested either on day 7 or day 42 for analysis. Separate cohorts of wild-type mice were injected intraperitoneally weekly for 4 weeks with 10 mg/kg isotype-matched control antibody (ADWA-21; provided by D. Sheppard) or anti-αvβ6 (6.3g9) starting at day 56 after infection. Separate cohorts of wild-type mice were also injected intraperitoneally with isotype-matched control antibody (ADWA-21; provided by D. Sheppard) or anti-αvβ6 (6.3g9) on days 2 and 5 after infection and were analyzed 3 d later (day 8 after infection). Statistical analysis All the statistical analysis on the data was done using GraphPad Prism software version 6.0. Statistical significance of differences between two groups with normally distributed data was determined using Student’s unpaired two-tailed t-test. For data that did not distribute normally between two groups, a two-tailed Mann-Whitney test was done. For groups of more than two, a one-way analysis of variance was used along with Tukey’s post-analysis tests. Sample sizes were chosen based on prior experiences with experimental mouse models. Mice of a particular genotype were randomly assigned to various treatment groups within a study. Blinding was used to count positive cells in immunofluorescent images in Figures 5 and 6. We thank D. Sheppard (University of California, San Francisco) for Itgb6−/− and Itgb8loxP mice and the isotype-matched control antibody ADWA-21; A. Glick, N. Blazanin and A. Ravindran for technical assistance with mouse KC culture; S. Violette (Biogen Idec) for antibody clones 6.3g9 and ch2A1 (both specific for αvβ6); J. Mitchell for technical assistance with confocal and epifluorescence microscopy; T. Martin, J. Motl and P. Champoux for technical assistance with flow cytometry and cell sorting; D. Rifkin and M. Vassallo for providing detailed protocols of an in vitro latent TGF-β-activation assay; the Research Animal Resources staff at the University of Minnesota for animal care; the Mayo Clinic division of Biostatistics and Bioinformatics for assistance in searching control and losartan-treated skin samples; and M. Jenkins for critical reading of the manuscript. Supported by the US National Institutes of Health (AR060744 to D.H.K., AI084913 to D.M.), the Dermatology Foundation (J.M.) and the American Skin Association (J.M.). Figure 1 Activation of latent TGF-β by αvβ6 inhibits homeostatic LC migration. (a) Quantification of LCs in archival skin specimens from untreated (control) and losartan-treated patients, detected by immunohistochemistry with antibody to CD1a (anti-CD1a) or anti-langerin. Each symbol represents an individual patient; horizontal lines indicate the average. (b) Microscopy of whole mounts of back epidermis from cohorts (n = 10 mice in each) of wild-type mice (WT), TGF-βRI–CALC mice (with inducible expression of constitutively active TGF-βRI in LCs) and TGF-βLC mice (with inducible ablation of TGF-β in LCs) 9 d after the start of tamoxifen treatment, as well as epidermis from adult Itgb6−/− mice, all stained for MHC class II (green). (c) Ratio of the number of LCs in the epidermis (Epi) (obtained by counting of the MHCII+ cells in b) or skin draining LNs (LN) (obtained by flow cytometry) of Itgb6−/−, TGF-βRI–CALC and TGF-βLC mice to that in their wild-type counterparts. Each symbol represents an individual mouse (with results from the same mouse joined by a solid line); dashed horizontal lines indicate a ratio of 1.0. (d) Microscopy of whole mounts of ear epidermis from tamoxifen-treated wild-type and TGF-βRI–CALC mice given intradermal injection of anti-αvβ6 or isotype-matched control antibody (Isotype) (n = 4 mice per genotype per group), stained for MHC class II (green). (e) Quantification of LCs per high-power field (HPF) in the mice in d. Each symbol represents LCs per HPF; small horizontal lines indicate the average. Scale bars (b,d), 100 μm. NS, not significant (P > 0.05); *P < 0.01 and **P < 0.0001 (two-tailed unpaired Student’s t test (a,c) or Tukey’s multiple comparisons test (e)). Data are representative of experiments with n = 42 total donors (a) or are representative of (b,d) or pooled from (c,e) three independent experiments. Figure 2 Epidermal residence of LCs requires expression of αvβ6 by IFE epidermal KCs. (a) Microscopy of whole mounts of back epidermis from wild-type and Itgb6−/− mice, stained for MHC class II (red) and αvβ6 (green); arrowheads indicate bulge region of telogen hair; asterisks indicate IFE region. (b) Flow cytometric analysis of the expression of αvβ6 by IM, IFE and bulge KCs (identified as in Supplementary Fig. 3b) from the back skin of wild-type and Itgb6−/− mice (key). (c) Quantification of LCs in transverse sections of back skin from wild-type and Itgb6−/− mice, identified by colocalization of langerin and MHC class II within the IFE region or IM. (d) Distance of IFE LCs from the hair follicle (HF) in sections as in c. (e) Quantitative RT-PCR analysis of Itgb6 mRNA in sorted epidermal populations of dendritic epidermal T cells (DETC), LCs, and IFE and IM KCs in wild-type mice; results are presented relative to those of the control gene Hprt. (f) Frequency of host- and donor-derived epidermal LCs in lethally irradiated wild-type→wild-type (WT→WT) or wild-type→Itgb6−/− (WT→Itgb6−/−) (CD45.2+) chimeras 6–8 weeks after transfer of BM from wild-type (CD45.1+) donors, analyzed by flow cytometry. (g) Microscopy of whole mounts of back epidermis from chimeras as in f, stained for MHC class II (green); autofluorescence shows hair shafts. Scale bars (a,g), 100 μm. ND, not detected. Each symbol (c,d,f) represents LCs per HPF (c), distance from HF (d) or an individual mouse (f); small horizontal lines indicate the mean. *P < 0.01 and **P < 0.0001 (two-tailed unpaired Student’s t test). Data are representative of three independent experiments with eight mice per genotype (a,b,g) or six independent experiments (e; mean + s.e.m.) or are pooled from two independent experiments with two mice per group (c,d) or three independent experiments (f). Figure 3 LC residence is controlled by spatially distinct expression of αvβ6 and αvβ6 on KC subsets through activation of latent-TGF-β. (a) Quantitative RT-PCR analysis of Itgb8 mRNA in sorted epidermal populations from wild-type mice (as in Fig. 2e). (b) Quantification of LCs in skin-draining LNs of wild-type and Itgb8ΔKC mice. (c) Quantification of LCs (per HPF) in IM of wild-type and Itgb8ΔKC mice (as in Fig. 2c). (d) Microscopy of whole mounts of tail epidermis (top) and transverse sections of back skin (bottom) from wild-type and Itgb8ΔKC mice, stained for langerin (red) and with the DNA-binding dye DAPI (blue); dashed lines, dermal-epidermal junction; SG, sebaceous gland; asterisks (bottom), IM. (e) Quantification of LCs in the IFE region and IM of the stained transverse skin sections in d. (f) Microscopy of whole mounts of back epidermis from wild-type, Itgb6−/− and Itgb6−/−Itgb8ΔKC mice, stained for MHC class II (green); autofluorescence shows hair shafts. (g) Quantification of LCs in f. (h) Luciferase activity in mink lung TGF-β reporter cells cultured together with wild-type or Itgb8ΔKC primary KCs pre-incubated with anti-αvβ6 or isotype-matched control antibody; results are presented as relative light units (RLU), normalized to those of wild-type control cells. Each symbol (b,c,e,g,h) represents an individual mouse (b,c,g), LCs per HPF (e) or RLU values (f); small horizontal lines indicate the mean (±s.e.m. in h). Scale bars, 50 mu;m (d) or 100 mu;m (f). *P < 0.05 and **P < 0.0001 (two-tailed unpaired Student’s t test (a–c,e) or Tukey’s multiple-comparisons test (g,h)). Data are representative of six independent experiments (a; mean + s.e.m) or three independent experiments (d,f), or are pooled from three independent experiments (b,c,e,g) with n = 5 mice per group (d,e) or three to six independent experiments (h). Figure 4 UV irradiation promotes LC migration through diminished integrin expression and TGF-β activation. (a) Microscopy of whole mounts of back epidermis from tamoxifen-treated wild-type and TGF-βRI–CALC mice (n = 5 per group) 4 d after sham treatment (-UV) or UVB irradiation (+UV), stained for MHC class II (green) and langerin (red). Scale bars, 100 mu;m. (b) Quantification of LCs (per HPF) in a. (c) Flow cytometry analyzing the expression of CD86 and CCR7 by epidermal LCs from mice as in a, gated as CD45+MHCII+CD11bint langerin-positive cells. Numbers in quadrants indicate percent cells in each. (d) Quantitative RT-PCR analysis of Itgb6 mRNA (left vertical axis) and Itgb8 mRNA (right vertical axis) in flow cytometry–sorted populations of IFE and IM KCs from wild-type mice 18 h after sham treatment or UVB irradiation (presented as in Fig. 2e). (e,f) Quantitative RT-PCR analysis of Itgb6 mRNA (e) and Itgb8 mRNA (f) in mouse primary KC cultures 24 h after sham treatment or UVB irradiation (presented as in Fig. 2e). (g) Luciferase activity of mink lung TGF-β reporter cells cultured together with primary keratinocytes given sham treatment or UVB irradiation; results are normalized to those of sham-treated cells. Each symbol (b,d–g) represents LCs per HPF (b), an independent experiment (d–f) or RLU values (g); small horizontal lines indicate the mean (±s.e.m. in d–g). *P < 0.01, **P < 0.001 and ***P < 0.0001 (two-tailed Mann-Whitney test (b,f) or two-tailed unpaired t test (d,e,g)). Data are representative of three independent experiments (a,c) or are pooled from three independent experiments with five mice per group (b) or six to seven independent experiments (d–g). Figure 5 αvβ6 and αvβ8 are required for the residence of CD8+ TRM cells in epidermis. (a) Microscopy of whole mounts of back epidermis from wild-type, Itgb6−/−, Itgb6−/−Itgb8ΔKC and LC-deficient (ΔLC) mice given injection of Thy-1.1+ P14 cells on day -1 and infected with LCMV on day 0, followed by topical treatment with DNFB on day 3 (for epidermal seeding of P14 T cells), stained for Thy-1.1 (red) and MHC class II (green) on day 42 after infection. (b) Quantification of Thy-1.1+ cells in a. (c) Microscopy of whole mounts of back epidermis from wild-type, Itgb6−/− and Itgb6−/−Itgb8ΔKC treated as in a, stained for Thy-1.1 (red) and MHC class II (green) on day 7 after infection. (d) Quantification of Thy-1.1+ cells in c. (e) Frequency of CD103+Thy-1.1+ cells in epidermis of the mice in c, assessed by flow cytometry on day 7 after infection. (f) Microscopy of whole mounts of back epidermis from wild-type mice treated intraperitoneally with anti-αvβ6 or isotype-matched control antibody once a week for 4 weeks starting at 56 d after infection with LCMV and epidermal seeding of TRM cells as in a, stained for Thy-1.1 (red) and MHC class II (green). (g) Quantification of Thy-1.1+ cells in f. Autofluorescent hair follicles are visible in a,c,f. Scale bars (a,c,f), 100 μm. *P < 0.05, **P < 0.01 and ***P < 0.0001 (Tukey’s multiple-comparisons test (b,e) or two-tailed Mann-Whitney test (g)). Data are representative of two to three independent experiments with n = 4 mice per group (a,b; mean + s.e.m. in b) or two independent experiments with n = 4–5 mice per group (c–g; mean + s.e.m. in d,e,g). Figure 6 αvβ6 is required for residence of TRM cells in intestinal epithelium. (a) Microscopy of small intestine from wild-type and Itgb6−/− mice 42 d after infection with LCMV, stained for Thy-1.1 (red) and collagen IV (ColIV) (green) and with DAPI (blue); arrowheads indicate TRM cells in the epithelial layer. (b,c) Quantification of Thy-1.1+ cells (per 106 nucleated cells) in the epithelial layer (above collagen IV (arrowheads) in a) (b) and lamina propria (LP) (below collagen IV in a) (c) of the small intestine (SI) of mice as in a. (d) Microscopy of small intestine from wild-type and Itgb6−/− mice 7 d after infection with LCMV, stained as in a. (e,f) Quantification of Thy-1.1+ cells in the epithelial layer (as in b) (e) and lamina propria (as in c) (f) of the small intestine of mice as in d. (g) Microscopy of small intestine from wild-type mice infected with LCMV and then treated intraperitoneally with anti-αvβ6 or isotype-matched control antibody starting at 56 d after infection, stained as in a. (h,i) Quantification of Thy-1.1+ cells in the epithelial layer (as in b) (h) and lamina propria (as in c) (i) of the small intestine of mice as in g. Scale bars (a,d,g), 50 μm. *P < 0.05 (two-tailed Mann-Whitney test). Data representative of three independent experiments with n = 4 mice per group (a–c; mean + s.e.m. in b,c) or two independent experiments with n = 4–5 mice per group (d–i; mean + s.e.m. in e,f,h,i). AUTHOR CONTRIBUTIONS J.M., D.M. and D.H.K. designed and interpreted experiments; J.M. performed most experiments; L.K.B., B.A., B.C., S.W.K., N.E.W., B.Z.I., S.W. and E.A.T. performed experiments and provided technical assistance; C.M. and W.D.S. provided technical and conceptual assistance; L.B. and D.S. provided reagents and technical assistance; A.B., A.K. and A.G.B. collected and analyzed data from control and losartan treated patients; and J.M. and D.H.K. wrote the manuscript and all authors edited it. COMPETING FINANCIAL INTERESTS The authors declare no competing financial interests. Note: Any Supplementary Information and Source Data files are available in the online version of the paper. 1 Merad M Ginhoux F Collin M Origin, homeostasis and function of Langerhans cells and other langerin-expressing dendritic cells Nat Rev Immunol 8 935 947 2008 19029989 2 Merad M Langerhans cells renew in the skin throughout life under steady-state conditions Nat Immunol 3 1135 1141 2002 12415265 3 Haley K Langerhans cells require MyD88-dependent signals for Candida albicans response but not for contact hypersensitivity or migration J Immunol 188 4334 4339 2012 22442445 4 Igyártó BZ Skin-resident murine dendritic cell subsets promote distinct and opposing antigen-specific T helper cell responses Immunity 35 260 272 2011 21782478 5 Kobayashi T Dysbiosis and staphylococcus aureus colonization drives inflammation in atopic dermatitis Immunity 42 756 766 2015 25902485 6 King JK Langerhans cells maintain local tissue tolerance in a model of systemic autoimmune disease J Immunol 195 464 476 2015 26071559 7 Kautz-Neu K Langerhans cells are negative regulators of the anti-Leishmania response J Exp Med 208 885 891 2011 21536741 8 Kaplan DH Jenison MC Saeland S Shlomchik WD Shlomchik MJ Epidermal langerhans cell-deficient mice develop enhanced contact hypersensitivity Immunity 23 611 620 2005 16356859 9 Obhrai JS Langerhans cells are not required for efficient skin graft rejection J Invest Dermatol 128 1950 1955 2008 18337832 10 Schenkel JM Masopust D Tissue-resident memory T cells Immunity 41 886 897 2014 25526304 11 Gebhardt T Memory T cells in nonlymphoid tissue that provide enhanced local immunity during infection with herpes simplex virus Nat Immunol 10 524 530 2009 19305395 12 Jiang X Skin infection generates non-migratory memory CD8+ TRM cells providing global skin immunity Nature 483 227 231 2012 22388819 13 Ariotti S T cell memory. Skin-resident memory CD8+ T cells trigger a state of tissue-wide pathogen alert Science 346 101 105 2014 25278612 14 Rashighi M CXCL10 is critical for the progression and maintenance of depigmentation in a mouse model of vitiligo Sci Transl Med 6 223ra23 2014 15 Bertolini M Uchida Y Paus R Toward the clonotype analysis of alopecia areata-specific, intralesional human CD8+ T lymphocytes J Investig Dermatol Symp Proc 17 9 12 2015 16 Li MO Flavell RA TGF-β: a master of all T cell trades Cell 134 392 404 2008 18692464 17 Bobr A Autocrine/paracrine TGF-β1 inhibits Langerhans cell migration Proc Natl Acad Sci USA 109 10492 10497 2012 22689996 18 Kel JM Girard-Madoux MJH Reizis B Clausen BE TGF-β is required to maintain the pool of immature Langerhans cells in the epidermis J Immunol 185 3248 3255 2010 20713882 19 Yasmin N Identification of bone morphogenetic protein 7 (BMP7) as an instructive factor for human epidermal Langerhans cell differentiation J Exp Med 210 2597 2610 2013 24190429 20 Mackay LK The developmental pathway for CD103+CD8+ tissue-resident memory T cells of skin Nat Immunol 14 1294 1301 2013 24162776 21 Travis MA Sheppard D TGF-β activation and function in immunity Annu Rev Immunol 32 51 82 2014 24313777 22 Yang Z Absence of integrin-mediated TGFβ1 activation in vivo recapitulates the phenotype of TGFβ1-null mice J Cell Biol 176 787 793 2007 17353357 23 Aluwihare P Mice that lack activity of αvβ6- and αvβ8-integrins reproduce the abnormalities of Tgfb1- and Tgfb3-null mice J Cell Sci 122 227 232 2009 19118215 24 Cohn RD Angiotensin II type 1 receptor blockade attenuates TGF-β-induced failure of muscle regeneration in multiple myopathic states Nat Med 13 204 210 2007 17237794 25 Lanz TV Angiotensin II sustains brain inflammation in mice via TGF-β J Clin Invest 120 2782 2794 2010 20628203 26 Bartholin L Generation of mice with conditionally activated transforming growth factor β signaling through the TGFβ1/ALK5 receptor Genesis 46 724 731 2008 18821589 27 Nagao K Stress-induced production of chemokines by hair follicles regulates the trafficking of dendritic cells in skin Nat Immunol 13 744 752 2012 22729248 28 Melton AC Expression of αvβ8 integrin on dendritic cells regulates Th17 cell development and experimental autoimmune encephalomyelitis in mice J Clin Invest 120 4436 4444 2010 21099117 29 Travis MA Loss of integrin αvβ8 on dendritic cells causes autoimmunity and colitis in mice Nature 449 361 365 2007 17694047 30 Abe M An assay for transforming growth factor-β using cells transfected with a plasminogen activator inhibitor-1 promoter-luciferase construct Anal Biochem 216 276 284 1994 8179182 31 Farache J Luminal bacteria recruit CD103+ dendritic cells into the intestinal epithelium to sample bacterial antigens for presentation Immunity 38 581 595 2013 23395676 32 Seré K Two distinct types of Langerhans cells populate the skin during steady state and inflammation Immunity 37 905 916 2012 23159228 33 El-Asady R TGF-β-dependent CD103 expression by CD8+ T cells promotes selective destruction of the host intestinal epithelium during graft-versus-host disease J Exp Med 201 1647 1657 2005 15897278 34 Casey KA Antigen-independent differentiation and maintenance of effector-like resident memory T cells in tissues J Immunol 188 4866 4875 2012 22504644 35 Masopust D Dynamic T cell migration program provides resident memory within intestinal epithelium J Exp Med 207 553 564 2010 20156972 36 Zaid A Persistence of skin-resident memory T cells within an epidermal niche Proc Natl Acad Sci USA 111 5307 5312 2014 24706879 37 Park CO Kupper TS The emerging role of resident memory T cells in protective immunity and inflammatory disease Nat Med 21 688 697 2015 26121195 38 Merad M Sathe P Helft J Miller J Mortha A The dendritic cell lineage: ontogeny and function of dendritic cells and their subsets in the steady state and the inflamed setting Annu Rev Immunol 31 563 604 2013 23516985 39 Harberts E Zhou H Fishelevich R Liu J Gaspari AA Ultraviolet radiation signaling through TLR4/MyD88 constrains DNA repair and plays a role in cutaneous immunosuppression J Immunol 194 3127 3135 2015 25716994 40 Wilson NS Normal proportion and expression of maturation markers in migratory dendritic cells in the absence of germs or Toll-like receptor signaling Immunol Cell Biol 86 200 205 2008 18026177 41 Zhang N Bevan MJ Transforming growth factor-β signaling controls the formation and maintenance of gut-resident memory T cells by regulating migration and retention Immunity 39 687 696 2013 24076049 42 Knight PA Enteric expression of the integrin αvβ6 is essential for nematode-induced mucosal mast cell hyperplasia and expression of the granule chymase, mouse mast cell protease-1 Am J Pathol 161 771 779 2002 12213704 43 Akhurst RJ Hata A Targeting the TGFβ signalling pathway in disease Nat Rev Drug Discov 11 790 811 2012 23000686 44 Kaplan DH Autocrine/paracrine TGFβ1 is required for the development of epidermal Langerhans cells J Exp Med 204 2545 2552 2007 17938236 45 Weinreb PH Function-blocking integrin αvβ6 monoclonal antibodies: distinct ligand-mimetic and nonligand-mimetic classes J Biol Chem 279 17875 17887 2004 14960589 46 Beura LK Lymphocytic choriomeningitis virus persistence promotes effector-like memory differentiation and enhances mucosal T cell distribution J Leukoc Biol 97 217 225 2015 25395301 47 Steinert EM Quantifying memory CD8 T cells reveals regionalization of immunosurveillance Cell 161 737 749 2015 25957682 48 Dlugosz AA Glick AB Tennenbaum T Weinberg WC Yuspa SH Isolation and utilization of epidermal keratinocytes for oncogene research Methods Enzymol 254 3 20 1995 8531694 49 Annes JP Chen Y Munger JS Rifkin DB Integrin αvβ6-mediated activation of latent TGF-β requires the latent TGF-β binding protein-1 J Cell Biol 165 723 734 2004 15184403

PMC005xxxxxx/PMC5135094.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 100883691 21720 Cell Microbiol Cell. Microbiol. Cellular microbiology 1462-5814 1462-5822 26639759 5135094 10.1111/cmi.12554 NIHMS832626 Article Porphyromonas gingivalis Initiates a Mesenchymal-like Transition through ZEB1 in Gingival Epithelial Cells Sztukowska Maryta N. 1 Ojo Akintunde 1 Ahmed Saira 1 Carenbauer Anne L. 1 Wang Qian 1 Shumway Brain 2 Jenkinson Howard F. 3 Wang Huizhi 1 Darling Douglas S. 1 Lamont Richard J. 1* 1 Department of Oral Immunology and Infectious Diseases, University of Louisville School of Dentistry, Louisville, Kentucky, United States of America 2 Department of Surgical and Hospital Dentistry, University of Louisville School of Dentistry, Louisville, Kentucky, United States of America 3 School of Oral and Dental Sciences, University of Bristol, Bristol, United Kingdom * Corresponding author: 570 South Preston Street, University of Louisville, Louisville, KY, 40202, Phone: 502-852-2112, rich.lamont@louisville.edu 30 11 2016 13 1 2016 6 2016 01 6 2017 18 6 844858 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. Summary The oral anaerobe Porphyromonas gingivalis is associated with the development of cancers including oral squamous cell carcinoma (OSCC). Here we show that infection of gingival epithelial cells with P. gingivalis induces expression and nuclear localization of the ZEB1 transcription factor which controls epithelial-mesenchymal transition (EMT). P. gingivalis also caused an increase in ZEB1 expression as a dual species community with Fusobacterium nucleatum or Streptococcus gordonii. Increased ZEB1 expression was associated with elevated ZEB1 promoter activity and did not require suppression of the miR-200 family of micro RNAs. P. gingivalis strains lacking the FimA fimbrial protein were attenuated in their ability to induce ZEB1 expression. ZEB1 levels correlated with an increase in expression of mesenchymal markers, including vimentin and MMP-9, and with enhanced migration of epithelial cells into matrigel. Knockdown of ZEB1 with siRNA prevented the P. gingivalis-induced increase in mesenchymal markers and epithelial cell migration. Oral infection of mice by P. gingivalis increased ZEB1 levels in gingival tissues, and intracellular P. gingivalis were detected by antibody staining in biopsy samples from OSCC. These findings indicate that FimA-driven ZEB1 expression could provide a mechanistic basis for a P. gingivalis contribution to OSCC. Introduction Once considered implausible, the concept that bacteria can be associated with cancer development is now well established. Indeed, a causal relationship between Helicobacter pylori and gastric cancer has been demonstrated (Kim et al., 2011), and a growing body of evidence supports the relationship between specific bacteria and various types of cancer (Garrett, 2015, Sahingur and Yeudall, 2015). For example, Fusobacterium nucleatum, a common inhabitant of the oral cavity, is over-represented in colorectal carcinoma (Castellarin et al., 2012, Kostic et al., 2012) and can induce colorectal carcinogenesis by activating E-cadherin/β-catenin signaling (Rubinstein et al., 2013). F. nucleatum can also inhibit natural killer (NK) cell cytotoxicity and killing of various tumors (Gur et al., 2015). High levels of antibodies to Porphyromonas gingivalis, a keystone pathogen in periodontal diseases, correlate with a greater than 2-fold increased risk of pancreatic cancer (Michaud, 2013). P. gingivalis is also associated with oral squamous cell carcinoma (OSCC). The surfaces of OSCCs harbor higher levels of Porphyromonas compared to contiguous healthy mucosa (Nagy et al., 1998), and P. gingivalis can be detected within gingival carcinomas by immunohistochemistry (Katz et al., 2011). Moreover, recent studies have established that combined infection with P. gingivalis and F. nucleatum promotes tumor progression in an oral-specific chemical carcinogenesis mouse model (Gallimidi et al., 2015). P. gingivalis and oral epithelial cells engage in an intricate molecular dialog, one consequence of which is entry of bacterial cells into the cytoplasm of the host cell (Lamont and Hajishengallis, 2015, Lamont et al., 1995). Primary cultures of epithelial cells containing P. gingivalis do not undergo apoptotic cell death and indeed P. gingivalis can suppress several proapoptotic pathways. In response to P. gingivalis infection Jak1/Akt/Stat3 signaling is activated with resultant increase in Bcl2 and inhibition of intrinsic mitochondrial apoptotic pathways (Yilmaz et al., 2004, Mao et al., 2007). By an independent mechanism P. gingivalis upregulates the level of miR-203 which suppresses expression of SOCS3, consequently impeding apoptosis (Moffatt and Lamont, 2011). In tandem with suppression of apoptosis, P. gingivalis promotes acceleration of primary epithelial cells through the S-phase of the cell cycle by impacting cyclin/CDK activities and reducing the amount of p53 (Kuboniwa et al., 2008). The process is dependent on the major fimbriae of P. gingivalis as a mutant deficient in FimA, the structural fimbrial subunit protein, does not induce increased cell proliferation. While inhibition of apoptosis and enhanced replication of cells can contribute directly to tumor development, it is unknown if P. gingivalis is capable of initiating the malignant transformation or oncogenic progression of epithelial cells. The epithelial-mesenchymal transition (EMT) is a process by which epithelial cells change shape and acquire a motile phenotype (Lamouille et al., 2014). The EMT is required for normal development and wound healing; however, it is also associated with the generation of self-renewing tumor-initiating cells, and in a malignant tumor it gives rise to a population of migratory and invasive cancer cells (Lamouille et al., 2014). This switch in cell differentiation and behavior is controlled by a group of transcription factors including the zinc-finger E-box-binding homeobox 1 and 2 proteins (ZEB1/2), SNAIL and TWIST (Vandewalle et al., 2009, Scanlon et al., 2013). The ZEB1 (δEF1, Zfhx1a, Zfhep) and ZEB2 (SIP1) transcription factors are critical EMT activators that bind to 5′-CACCTG sequences and repress transcription of epithelial specific genes such as E-cadherin (cdh1) (Vandewalle et al., 2009). ZEB can also positively regulate genes associated with the mesenchymal phenotype such as those encoding vimentin and matrix-metalloproteinases (Vandewalle et al., 2009, Lamouille et al., 2014). ZEB1/2 are in the TGFβ signaling pathway, binding SMADs and having essential effects on embryonic development (Gheldof et al., 2012). ZEB1 has been implicated in activating EMT and metastasis in several type of cancers (Sanchez-Tillo et al., 2012, Jia et al., 2012). Moreover, ZEB1/2 are linked to the miR-200 family in a reciprocal negative feedback loop whereby each regulates the expression of the other (Brabletz and Brabletz, 2010). H. pylori has been shown to upregulate expression of ZEB1 which can initiate an EMT and cancer stem-cell properties in infected gastric epithelial cells (Baud et al., 2013, Bessede et al., 2014). In this study we show that P. gingivalis can increase ZEB1 levels in gingival epithelial cells in a fimbriae dependent manner. Upregulation of ZEB1 was dependent on increased promoter activity. Elevated expression of ZEB1 was associated with a partial mesenchymal phenotype in P. gingivalis-infected gingival epithelial cells, including increased migration. We also detected P. gingivalis antigens in oral carcinoma in situ and poorly differentiated cancer, and mice orally infected with P. gingivalis had an increase in ZEB1 mRNA expression in gingival tissues. The results suggest a novel mechanism by which oral bacteria such as P. gingivalis can contribute to a mesenchymal phenotype, and potentially drive the progression of cancer. Results P. gingivalis upregulates ZEB1 in gingival epithelial eells We investigated the impact of P. gingivalis on ZEB1 expression in TIGK cells using qRT-PCR and immunoblotting. As shown in Figure 1A, P. gingivalis increased ZEB1 mRNA levels in a time and dose dependent manner, with maximal induction occurring after 24 h infection with a MOI of 100. An increase in the amount of ZEB1 protein was also observed at 24 h following P. gingivalis infection at both MOI 50 and 100 (Fig 1B). As P. gingivalis infections of oral tissue are chronic conditions, we further examined ZEB1 activity 72 h after P. gingivalis infection. MOIs of 1, 10 and 50 were used as at MOI 100 the proteases of P. gingivalis can cause detachment of cells from the substratum. While an MOI 1 did not affect ZEB1 expression, mRNA levels were increased by P. gingivalis at MOI 10 and 50 (Figure 1C). The ability of P. gingivalis at MOI 10 to increase ZEB1 expression after 72 h, but not earlier, indicates that infection of epithelial cells with low numbers of the organism has the potential to elevate ZEB1 over extended times, possibly as a result of intracellular P. gingivalis replication and cell to cell spread (Lamont et al., 1995, Yilmaz et al., 2006). To corroborate the nuclear location of the ZEB1 transcription factor following P. gingivalis infections, TIGKs were examined by CLSM with quantitative image analysis (Figures 1D and E). After P. gingivalis infection there was increased expression of ZEB1 protein in the nucleus where it is functionally active. ZEB1 responses to P. gingivalis are strain and fimbriae dependent P. gingivalis is a host adapted organism with a nonclonal population structure, and isolates from different individuals often vary considerably (Tribble et al., 2013). Hence, we next examined the ability of different strains of P. gingivalis to enhance ZEB1 mRNA levels. As shown in Figure 1F, an additional ATCC strain (49417) and two low passage clinical isolates (11029 and 10512) induced ZEB1 expression to a similar degree as the type strain 33277. In contrast, the commonly used laboratory strain W83 did not significantly increase ZEB1 expression. One of the major differences among P. gingivalis strains is in the expression of fimbriae (Nadkarni et al., 2014). The two ATCC strains, along with the two low passage isolates, all expressed FimA, the structural subunit protein of the major fimbriae (Supporting Information Figure S1). Strain W83 does not expresses FimA (Nishikawa and Duncan, 2010), which prompted us to speculate that FimA may be an effector protein for ZEB1 induction. This concept was corroborated by the failure of an isogenic fimA mutant of 33277 to increase ZEB1 mRNA levels (Figure 1G). We also found that induction of ZEB1 expression required direct contact between P. gingivalis and epithelial cells (Figure 1G), consistent with a role for the FimA adhesin. Fimbriated P. gingivalis activate JNK signaling (Watanabe et al., 2001) which has been reported to increase transcription of ZEB1 (Zhang et al., 2012). However, siRNA knockdown of JNK in TIGK cells did not impede the ability of P. gingivalis to upregulate ZEB1 (Supporting Information Figure S2). In addition, pharmacological inhibition of Akt with LY294002 also failed to reduce P. gingivalis-mediated ZEB1 induction (Supporting Information Figure S3). Hence the signaling pathways activated by P. gingivalis fimbriae that converge on Zeb1 remain to be determined, and this topic is under active investigation in our laboratory. The FimA fimbriae are required for maximal invasion of P. gingivalis into gingival epithelial cells (Lamont and Jenkinson, 1998). However, invasion per se was not required for ZEB1 induction as a mutant of P. gingivalis that is invasion-deficient, due to disruption of the gene encoding the serine phosphatase SerB (Takeuchi et al., 2013), retained the ability to upregulate ZEB1 (Supporting Information Figure S4). Nonetheless, the spatial definition of P. gingivalis initiation of ZEB1 activation requires further study. Moreover, when P. gingivalis was separated from the epithelial cells by a semi-permeable membrane, ZEB1 levels were lower than the control uninfected condition, indicating that there may be components secreted by P. gingivalis that can antagonize ZEB1 regulation in the absence of FimA mediated contact. Thus, multiple effectors of P. gingivalis may be capable of impacting ZEB1 expression, with the effect of whole cells representing the collective output of several distinct interactions and signaling pathways. P. gingivalis communities regulate ZEB1 expression On mucosal surfaces bacteria rarely exist as monospecies accumulations but rather as complex multispecies communities. P. gingivalis engages in synergistic community formation with S. gordonii and F. nucleatum, common inhabitants of the oral microbiota, and in vivo these organisms can be found in close association (Benitez-Paez et al., 2014, Valm et al., 2011, Wright et al., 2014, Hendrickson et al., 2014). Individually, neither S. gordonii nor F. nucleatum were capable of regulating ZEB1 expression, indicating that of these three widespread oral species, P. gingivalis has the most potential to effect an EMT through ZEB1 (Figure 2). Importantly, P. gingivalis remained effective at elevating ZEB1 mRNA in the context of a community with either S. gordonii or F. nucleatum, consistent with recent reports demonstrating that a community of P. gingivalis and F. nucleatum can promote tumor progression in animal models (Gallimidi et al., 2015). Thus, the tumorigenic properties of P. gingivalis can prevail in the presence of co-colonizing organisms, an important principle for in vivo relevance. P. gingivalis upregulates ZEB1 promoter activity and downregulates miR200 An increase in the amount of steady state mRNA levels can result from an increase in transcription or a decrease in degradation. To begin to distinguish between these possibilities, we first examined transcriptional activity from the ZEB1 promoter using a series of ZEB1 upstream regulatory regions promoting transcription of the luc gene. These human ZEB1 promoter constructs contain sequences important for regulation of ZEB1 expression in several cell types (Manavella et al., 2007, Liu et al., 2007). P. gingivalis stimulated the activity of all of these promoter constructs (Figure 3A), indicating that the increase of ZEB1 mRNA induced by P. gingivalis can occur through an elevated transcription rate. These data also localize the response element(s) within the first 400 bp of the promoter. An additional mechanism by which ZEB1 is controlled posttranscriptionally is through the action of the miR-200 family of microRNAs (Brabletz and Brabletz, 2010). miR-200 family members target conserved recognition sites on the 3′ UTR of ZEB1 mRNA (Brabletz and Brabletz, 2010), and thus a decrease in miR-200 leads to higher levels of ZEB1 mRNA. However, we did not observe a reduction in the amount of miR-200b, miR-200c or miR-205 in cells at 6 h after P. gingivalis infection when levels of ZEB1 mRNA begin to rise (Figure 3B–D). Indeed there was a slight increase in miR-200 family expression, indicating that the increase in ZEB1 levels at 6 h is not the consequence of decreased miR-200 expression. The action of ZEB1, in turn, represses the transcription of the miR-200 family, and consistent with this at 24 and 48 h after P. gingivalis infection, miR-200b, miR-200c and miR-205 levels were reduced. A control microRNA, miR-21, which is not involved in ZEB1 feedback regulation, did not show a significant decrease in expression (Supporting Information Figure S5). Hence, the pattern of miRNA expression is consistent with the results from the promoter-reporter constructs in pointing toward increased mRNA synthesis as the initial cause of the elevated levels of steady state mRNA for ZEB1. Changes in epithelial and mesenchymal marker expression upon P. gingivalis infection The expression pattern of ZEB1 targets in TIGKs infected with P. gingivalis was assessed by qRT-PCR (Table 1). Mesenchymal markers N-cadherin, vimentin and matrix metalloproteinase (MMP)-9 were upregulated at 24 h after infection by P. gingivalis at MOI 50 and 100, while fibronectin levels were increased by P. gingivalis at MOI 100. An increase in vimentin protein expression was confirmed by immunoblotting (Supporting Information Figure S6), and elevated MMP-9 amounts following infection with fimbriated P. gingivalis was corroborated by zymography (Figure 4). While both pro and active forms of MMP-9 were increased by P. gingivalis wild type, there was no difference in the ratio of active MMP-9 to total (cleaved and pro-MMP9) between parental and fimbrial deficient mutant strains. Under these infection conditions, therefore, fimbriated P. gingivalis elevate the amount of MMP-9 produced by TIGK cells but do not modulate MMP-9 activation. In contrast expression of MMP-2, which may be more predominantly regulated by Twist (Yang et al., 2013), was not impacted by P. gingivalis infection. Of the epithelial markers tested, collagen 1 (COL1A1) and cytokeratin 19 (KRT19) were suppressed by P. gingivalis. Collectively, these results support the concept that infection by P. gingivalis can contribute to the process of transition toward a mesenchymal phenotype. Although the mesenchymal marker integrin α5 (ITGA5) and the epithelial marker cytokeratin 7 (KRT7) were unaffected by P. gingivalis, variability in expression of important cell proteins is not unexpected as control of expression by ZEB1 is cell and context dependent (Lamouille et al., 2014). One of the major targets of ZEB1 is E-cadherin, and ZEB1 mediated repression of E-cadherin, with associated disruption of E-cadherin dependent junctions, is an important marker of EMT. We did not observe differential regulation of E-cadherin following P. gingivalis infection (not shown). However, the gingival epithelium is highly porous with only sparse interconnections (Bosshardt and Lang, 2005) and expression of E-cadherin is very low (Heymann et al., 2001). Thus a reduction of E-cadherin may not be as important for the EMT of gingival epithelial cells as in other cell types. To corroborate the role ZEB1 in the differential regulation of mesenchymal markers, siRNA mediated knockdown was performed. Reduction of ZEB1 mRNA and protein following siRNA transfection was confirmed by qRT-PCR and immunoblotting, respectively (Figure 5A,B). TIGKs with diminished ZEB1 expression were then infected with P. gingivalis MOI 50 or 100 over 24 h. As shown in Figure 5C and D, ZEB1 deficiency prevented P. gingivalis induced modulation of expression of vimentin and MMP-9. P. gingivalis promotes migration of epithelial cells Cells that acquire an EMT phenotype display an invasive behavior in vitro, and thus we tested the ability of P. gingivalis to increase the migration of TIGK cells into matrigel. Figure 6 shows that P. gingivalis infection resulted in a greater than 2-fold increase in TIGK cell invasion into the gel compared to control cells. Knockdown of ZEB1 prevented P. gingivalis-induced TIGK cell migration, verifying the importance of ZEB1 in this aspect of the P. gingivalis-dependent partial mesenchymal phenotype. P. gingivalis elevates ZEB1 levels in vivo To determine whether P. gingivalis’ ability to increase ZEB1 levels also occurs in vivo, mice were orally infected with P. gingivalis, and gingival tissue recovered 1, 3 and 8 days following the final inoculation. Levels of P. gingivalis on the gingival tissues were determined by qPCR (Supporting Information Figure S7) and remained constant over the 8-day period. As shown in Figure 7, colonization with P. gingivalis induced an increase in gingival tissue expression of ZEB1 mRNA over 8 days compared to sham infected animals. Thus, P. gingivalis has the potential to stimulate ZEB1 and contribute to an EMT in an animal model. Presence of P. gingivalis in human OSCC P. gingivalis could exacerbate carcinogenesis at several stages through its ability to increase ZEB1 expression, but only if the bacteria are physically associated with the developing cancer. We investigated whether P. gingivalis bacteria are present within oral squamous cell carcinoma biopsy samples. Immunofluorescence microscopy with a specific polyclonal P. gingivalis antiserum labeled discrete speckles in the cells of a poorly differentiated OSCC sample and a carcinoma in situ, whereas antibodies to S. gordonii showed little or no labeling (Figure 8A–B). Similar results were seen with two carcinoma in situ cases and two poorly differentiated carcinomas. Confocal microscopy more clearly detected the particles which were fluorescently labeled by the P. gingivalis antibodies. Serial optical sections were taken at 0.4 microns, and individual particles were found to persist in 5 to 7 adjacent optical slices (Supporting Information Figure S8). This estimates the fluorescent particles to be 2.0 to 2.8 microns in size, consistent with intact P. gingivalis. As described in primary gingival epithelial cells, P. gingivalis was observed in both the cytoplasm and in the nuclei (Belton et al., 1999). Discussion Typically, the gingival epithelium provides a major physical barrier to oral pathogens. Disruption of the gingival barrier by inducing an EMT may enhance the ability of P. gingivalis to invade the tissue and enhance access to nutrients derived from inflammatory tissue breakdown (Hajishengallis, 2014). Hence, up-regulation of ZEB1 by P. gingivalis can be seen as providing an evolutionary advantage to the organism. Beyond this, the ability of P. gingivalis to stimulate ZEB1 expression could have several distinct clinically relevant effects. ZEB1 influences multiple stages of carcinogenesis, including the initial transformation, progression, EMT leading to metastasis, and resistance to therapy (Sanchez-Tillo et al., 2012, Zhang et al., 2014). Therefore, the presence of P. gingivalis, interacting with other environmental effectors, may enhance the initiation of oral cancer within the pre-cancerous field, or increase carcinogenic progression. The ability to manipulate ZEB1 location and function constitutes an important attribute of bacteria with a potential role in carcinogenesis. P. gingivalis is a keystone member of dysbiotic oral communities, which in combination with its ability to spread systemically, and enhance cell survival and proliferation, supports epidemiological evidence of an association with cancers such as OSCC (Whitmore and Lamont, 2014, Lamont and Hajishengallis, 2015). Moreover, in established invasive OSCC lines, P. gingivalis activates the ERK1/2-Ets1, p38/HSP27, and PAR2/NF-κB pathways to promote cellular invasion (Inaba et al., 2014). The results of the present study indicate the P. gingivalis may induce nontransformed gingival epithelial cells to undergo a partial EMT through the upregulation of ZEB1. We show that P. gingivalis increases the transcriptional activity of the ZEB1 gene and increases ZEB1 protein levels in the nucleus. Infection of epithelial cells with P. gingivalis upregulated expression of genes associated with the mesenchymal phenotype and knockdown of ZEB1 attenuated this effect. P. gingivalis also induced a migratory phenotype in epithelial cells which was ZEB1-dependent. Oral infection of mice with P. gingivalis stimulated ZEB1 expression in the gingival tissues and biopsy tissue from human OSCC carcinoma in situ and poorly differentiated cancer showed the presence of P. gingivalis. On the hard and soft tissues of the oral cavity P. gingivalis is an inhabitant of multispecies communities. Organisms such as S. gordonii and F. nucleatum provide mutual physiological support, and P. gingivalis in the context of a community is phenotypically distinct from single species accumulations (Wright et al., 2013). In addition, infection of epithelial cells with the early colonizing streptococci can reprogram specific signaling pathways such that they do not respond to the later colonizing P. gingivalis (Handfield et al., 2005). We found here that while neither S. gordonii nor F. nucleatum modulated ZEB1 mRNA levels, combinations of P. gingivalis with either species retained the capacity to upregulate ZEB1. As porphyromonads, fusobacteria and streptococci are all found in higher numbers on the surfaces of OSCC compared to contiguous healthy mucosa (Nagy et al., 1998), it is likely therefore that these microbial communities contribute to the EMT. P. gingivalis strains exhibit extensive genetic variation as a result of genomic rearrangements (Naito et al., 2008), and horizontal gene transfer is considered an adaptive strategy for long term survival in the oral environment (Nadkarni et al., 2014, Tribble et al., 2007). Most strains of P. gingivalis expresses fimbriae comprised of FimA major fimbrial subunit proteins, although in some strains, such as the commonly used lab strain W83, FimA production is very low due to a mutation in the FimS histidine kinase component of the FimS/FimR TCS that controls transcription of the fimA operon (Nishikawa and Duncan, 2010). Results with a variety of strains and isogenic mutants of P. gingivalis indicate that FimA is the effector of P. gingivalis responsible for upregulation of ZEB1. FimA is a major antigen on the P. gingivalis surface and can also incite the production of proinflammatory cytokines (Lamont and Jenkinson, 1998, Bostanci and Belibasakis, 2012). FimA is capable of manipulating a number of signal transduction pathways and transcription factors in different cell types (Zhou and Amar, 2007, Hajishengallis et al., 2012), and the processes that lead to increased ZEB1 promoter activity require further study. The potential importance of FimA expressing P. gingivalis lineages in the events that can lead to tumor development is corroborated by the role of this protein in the acceleration of the epithelial cell cycle (Kuboniwa et al., 2008). FimA fimbriae, which do not share significant homology to other fimbrial proteins (Enersen et al., 2013), may thus constitute an attractive target for novel biomarkers or therapeutics. ZEB1 can be regulated at the transcriptional level and posttranscriptionally regulated by the miR-200 family through a double negative feedback loop (Brabletz and Brabletz, 2010). In epithelial cells miR-200s inhibit ZEB expression and maintain the epithelial phenotype. By contrast, in mesenchymal cells elevated ZEB activity suppresses expression of the miR-200s. Our results show that the increased levels of ZEB1 were associated with a reduction in the amounts of the miR200 family, potentially facilitating a stable transition to the partial mesenchymal phenotype. The initial upregulation of ZEB1 in epithelial cells, however, was not associated with a decrease in the amounts of the miR200 family, but with an increase in ZEB1 promoter activity. NF-κB activation also promotes ZEB1 transcription (Vandewalle et al., 2009); however, in epithelial cells P. gingivalis suppresses the activation of NF-κB by dephosphorylation of the p65 subunit at the S536 residue (Takeuchi et al., 2013). It is unlikely, therefore, that NF-κB is involved in P. gingivalis-induced ZEB1 upregulation. P. gingivalis induced expression of the mesenchymal markers and decreased expression of the epithelial markers. siRNA knockdown of ZEB1 abrogated the ability of P. gingivalis to regulate epithelial and mesenchymal gene expression, establishing ZEB1 as a major transcriptional effector of the P. gingivalis-induced partial EMT. Epithelial markers down regulated by P. gingivalis included cytokeratin 19 which is characteristically expressed in cells of the junctional epithelium within the gingival tissues, and has also been reported to be suppressed in OSCC (Khanom et al., 2012). The mesenchymal relevant genes induced by P. gingivalis included N-cadherin, vimentin, fibronectin and MMP-9. N-cadherin is a calcium dependent cell-cell adhesion glycoprotein, which is upregulated in EMT and some studies have found associated with OSCC (Zhao et al., 2012). Vimentin is a cytoskeletal intermediate filament protein involved in maintaining cell shape and stabilizing cytoskeletal interactions. Expression of vimentin is associated with OSCC tumorigenesis (Lee et al., 2015), and vimentin has been proposed as a predictor of the malignant potential of high risk oral lesions (Sawant et al., 2014). Fibronectin is a component of the cell matrix involved in cell migration processes including metastasis, and expression of alternatively spliced segments of fibronectin is related to OSCC tumorigenesis (Kamarajan et al., 2010). MMP-9 is secreted as inactive proproteins which are activated by proteolytic cleavage. As a gelatinase, MMP-9 can degrade collagen IV in the basement membrane and extracellular matrix facilitating tumor growth, invasion, metastasis, and angiogenesis (Westermarck and Kahari, 1999). MMP-9 plays a crucial role in the development of several human malignancies, including OSCC (Kruger et al., 2005, Bedal et al., 2014). Moreover, epithelial cells infected with P. gingivalis showed ZEB1-dependent increased migration into matrigel, a phenotype consistent with increased MMP-9 activity and with an overall partial mesenchymal phenotype. To begin to translate our results from reductionist in vitro models to the in vivo situation, we orally infected mice with P. gingivalis and examined ZEB1 expression in gingival tissues. Although P. gingivalis is not a normal member of the mouse oral microbiota, it does colonize transiently and causes alveolar bone loss (Hajishengallis et al., 2015). Our results show for the first time that P. gingivalis colonization of the gingival tissues in vivo leads to upregulation of ZEB1. Further in vivo evidence for a role of P. gingivalis in oral tumor development was provided by IF analysis of OSCC biopsy samples. Antigenic based detection of P. gingivalis within biopsy samples from OSCC poorly differentiated cancer and carcinoma in situ corroborates a similar study in which P. gingivalis antigens were detected in ten gingival squamous cell carcinomas of differing degrees of differentiation (Katz et al., 2011). The use of optical sectioning in the current study established that the size of particles detected by P. gingivalis antibodies was in the 2–3 μm range, consistent with whole organisms, rather than shed antigens or outer membrane vesicles. Intimate association of P. gingivalis with OSCC lesions shows that the organism has the opportunity as well as the capability, to contribute to EMT in vivo. Oral cancers are among the most prevalent (Jemal et al., 2008), and despite considerable advances in diagnosis and therapeutic options, the 5-year survival rate has remained stable at approximately 50% among all tumor stages during the past decades (Wikner et al., 2014). The early phase of OSCC is often asymptomatic, therefore the identification of both novel biomarkers and contributing etiological agents is important for improving survival rates. The results of the current study suggest that infection with FimA-positive P. gingivalis can induce a ZEB1 dependent partial EMT. The detection of P. gingivalis, or of FimA, in early erythroplakia or leukoplakia lesions, therefore, may have utility for the early detection of lesion likely to progress to malignant status. Materials and Methods Bacterial strains, eukaryotic cells, and growth and infection conditions Porphyromonas gingivalis strain ATCC 33277 and its isogenic mutant ΔfimA, strains ATCC 49417, W83, and low passage clinical isolates 11029, 10512 (laboratory strains), were cultured in trypticase soy broth (TSB) supplemented with yeast extract (1 mg/ml), hemin (5 μg/ml) and menadione (1 μg/ml). Tetracycline (1 μg/ml) was incorporated into the medium for the growth of ΔfimA. Fusobacterium nucleatum ATCC 25586 was cultured in brain heart infusion (BHI) broth supplemented with hemin (5 μg/ml) and menadione (1 μg/ml). Streptococcus gordonii DL1 was grown in BHI supplemented with yeast extract (5 μg/ml). All bacterial strains were cultured anaerobically at 37°C. Human telomerase immortalized keratinocytes (TIGKs) derived from gingival epithelium were maintained at 37°C and 5% CO2 in Dermalife-K serum free culture medium (Lifeline Cell Technology, Carlsbad, CA) as described (Moffatt-Jauregui et al., 2013). TIGKs between passages 10 and 20 at 70% confluence were stimulated with bacteria as described for individual experiments. For transwell (Corning, Corning NY) assays, TIGKs were cultured in the lower compartment and P. gingivalis added to the upper chamber. Immunoblotting TIGK cells were solubilized in cold cell lysis reagent (Cell Signaling, Danvers, MA) containing Protease Inhibitor and PhosSTOP Phosphatase Inhibitor (Roche, Indianapolis, IN). Proteins (40 ng) were separated by 10% SDS-polyacrylamide gel electrophoresis, blotted onto a PVDF membrane and blocked in 5% BSA in TBS with 0.1% Tween20. Blots were probed at 4°C overnight with primary antibodies followed by 1 h with HRP-conjugated secondary antibody at room temperature. Antigen-antibody binding were detected using ECL Substrate (Thermoscientific, Hudson NH). Primary antibodies targeted ZEB1 (Novus, Littleton, CO) or vimentin (Cell Signaling). Duplicate blots were probed with GAPDH antibodies (Cell Signaling) as a loading control RNA extraction and quantitative reverse transcription-PCR (qRT-PCR) Total RNA from TIGK cells and from homogenized gingival tissue was isolated and purified with PerfectPure RNA kit (5Prime, Gaithersburg, MD). miRNA was isolated and purified from TIGKs with PureLink miRNA isolation kit (Invitrogen, Carlsbad, CA). RNA concentrations were determined by spectrophotometry (NanoDrop Technology, Wilmington, DE). cDNA from total RNA and miRNA was synthesized (2 μg RNA/reaction volume) using a High Capacity cDNA Reverse Transcription kit and a TaqMan MicroRNA Reverse Transcription kit (Applied Biosystems, Grand Island, NY), respectively. Real time RT-PCRs used TaqMan Fast universal PCR mastermix and gene expression assays for Zeb1, vimentin, MMP-9, ITGA5, fibronectin, KRT7, COL-1A1 and GAPDH. Negative RT reactions were included in each assay. TaqMan microRNA assays were used to quantify the mature miRNA-200b, mi-RNA-200c, miRNA-205, miRNA-21 and RNU48. Real time qPCR was performed on an Applied Biosystems StepOne Plus cycler with StepOne software V2.2.2 and autocalculated threshold cycle selected. The cycle threshold (Ct) values were determined, mRNA and miRNA expression levels were normalized to GAPDH and RNU48, respectively, and expressed relative to controls following the 2−ΔΔCT method. Luciferase reporter assay TIGK cells were transfected with ZEB1 promoter constructs: Z1p 1000-Luc (−912 bp to +43 bp of the ZEB1 gene), Z1p.400-Luc (−367 bp to +43 bp) and Z1p.196-Luc (−212 bp to +43 bp); at 2 μg/105 cells using FuGENE6 Transfection Reagent (Promega, Madison, WI). Following 48 h in transfection media, cells were returned to TIGK medium for further 24 h, prior to the stimulation with P. gingivalis. Cells were lysed and the reporter activity was determined with the Dual-Glo Luciferase Assay System (Promega). Firefly luciferase activity was normalized on the basis of Renilla luciferase activity in the same samples. Zymography The activities of MMP2 and MMP9 in culture supernatant collected from control uninfected and P. gingivalis-infected TIGK cells, were determined using gelatin zymography as described (Inaba et al., 2014). Samples were mixed with SDS sample buffer without reducing reagents, then separated on 10% SDS-polyacrylamide gels containing 0.1% gelatin. The gels were incubated at 37°C with 2.5% Triton X-100 for 1 h, and then in 20 mM Tris-HCl (pH 7.5) containing 200 mM NaCl and 5 mM Ca Cl2 for 48 h. After staining with 5% Coomassie Brilliant Blue R-250, gelatinolytic activities were visualized as clear bands against a blue background and quantified using ImageJ. RNA interference TIGKs were transfected with predesigned 30 nM siRNA (siGENOME SMARTpool siRNA) targeting ZEB1 or control siRNA (GE Healthcare Dharmacon, Lafayette, CO) for 24 h using LipoJet (SignaGen, Gaithersburg, MD) transfection reagent. At 48 h after transfection, the medium was replaced and cells were incubated for a further 24 h prior to infection. Immunofluorescence and confocal laser scanning microscopy TIGK cells were grown on glass coverslips, washed twice in phosphate-buffered saline (PBS) and fixed with 4% paraformaldehyde for 10 min. Permeabilization was with 0.2% TritonX-100 for 10 min at room temperature prior to blocking in 10% goat serum for 20 min. ZEB1 was detected by reacting with primary antibodies at 1:100 overnight at 4°C, followed by Alexa Fluor 488-conjugated anti-rabbit secondary antibodies (Life Technologies) at 1:200 for 3 h in the dark. Following a 20 min blocking in 0.1% goat serum, actin was labeled with Texas Red-phalloidin (Life Technologies) for 2 h at room temperature in the dark. Coverslips were mounted with on glass slides using ProLong Gold with DAPI (4′6-diamidino-2-phenylindole) mounting medium (Invitrogen) prior to imaging with a Leica SP8 confocal inverted fluorescence microscope. Images were analyzed using Volocity 6.3 Software (PerkinElmer, Waltham, MA). Matrigel invasion assay Cell motility was measured by assessment of the migration rate of TIGKs using a BD BioCoat Matrigel Invasion Chamber (BD Biosciences, San Jose, CA). Cells (2 × 105) were plated on transwell filters coated with matrigel. The lower compartment of the invasion chambers contained cell culture medium. After 18 h incubation at 37°C, cells remaining on the upper surface of the filter were removed, and the cells that migrated through the filter were fixed with 1% methanol and stained with toluidine blue. Cells were enumerated from three random 20× fields for each filter using a Nikon Eclipse TS100 microscope. Animal infection BALB/c mice were orally infected with 107 cfu P. gingivalis 33277 five times at 2-day intervals as described previously (Daep et al., 2011) and approved by the University of Louisville Institutional Animal Care and Use Committee. The levels of P. gingivalis colonization were determined by qPCR with the P. gingivalis 16SrRNA gene as described (Daep et al., 2011). On days 1, 3 and 8 after the last infection, mice were euthanized and the upper and lower jaw with gingival tissue were recovered. After RNA extraction, the ratio of ZEB1 mRNA to GAPDH mRNA for each mouse was determined by qRT-PCR using the 2− ΔCT method. Human oral tissue immunofluorescent and immunohistochemical staining Paraffin embedded human tongue biopsy samples were sectioned at 4 μm, dewaxed and rehydrated. The slides were blocked with 10% goat serum for 1 h, and reacted with P. gingivalis 33277 or S. gordonii antibodies 1:1000 for 2 h at room temperature. Secondary Alexa Fluor 488 conjugated anti-rabbit antibody (1:500) was for 1 h, following which slides were treated with DAPI (1:2000). Slides were mounted with VectaShield (Vector Labs, Burlingame, CA) and imaged with a Zeiss Axiocam MRc5 fluorescence microscope. Procedures were approved by the University of Louisville Institutional Review Board. Statistical analyses Statistical analyses were conducted using the GraphPad Prism software. Data were evaluated by one-way ANOVA with Tukey’s multiple comparison test. Experimental data presented are representative of at least three biological replicates. Supplementary Material Supplementary Material Figure S1. Immunoblot of whole cell lysates of P. gingivalis strains probed with polyclonal antibodies to the FimA protein of strain 33277. Figure S2. JNK knockdown does not affect regulation of Zeb1 by P. gingivalis. A. TIGK cells were transiently transfected with siRNA to JNK1/2 (si Jnk, 100 nM, Sigma) or scrambled siRNA (si ctr). Control (ctr) cells were nontransfected. JNK mRNA levels in transfected cells were measured by qRT-PCR. Data were normalized to GAPDH mRNA and are expressed relative to ctr. Results are means ± SD, n = 3; *** P < 0.001. B. Transfected TIGKs cells were infected with P. gingivalis 33277 for 24 h at MOI 100. ZEB1 mRNA was measured by qRT-PCR, the data were normalized to GAPDH mRNA and are expressed relative to the noninfected (NI) control. Results are means ± SD, n = 3; *** P < 0.001 compared to NI; NS: not significant. Figure S3. Pharmacological inhibition of Akt does not affect regulation of Zeb1 by P. gingivalis. TIGK cells were preincubated with 10 μM LY294002 or vehicle (DMSO) only for 1 h and infected with P. gingivalis 33277 MOI 50 or100 for 6 h. ZEB1 mRNA levels were measured by qRT-PCR, normalized to GAPDH mRNA and are expressed relative to noninfected (NI) controls. Results are means ± SD, n = 3; * P < 0.05; *** P < 0.001; NS: not significant. Figure S4. A non-invasive mutant of P. gingivalis can induce ZEB1 expression. qRT-PCR of ZEB1 mRNA expression in TIGK cells infected with P. gingivalis 33277 (Pg WT) or a ΔserB mutant. Data were normalized to GAPDH mRNA and are expressed relative to noninfected (NI) controls. Results are means ± SD, n = 3; *** P < 0.001 compared to NI; NS: not significant. Figure S5. Expression of miRNA-21 is not down-regulated by P. gingivalis. TIGK cells were infected with P. gingivalis 33277 (Pg) at MOI 100 for the time indicated. miRNA levels were measured by qRT-PCR, normalized to RNU48 miRNA, and expressed relative to noninfected (NI) controls. Results are means ± SD, n = 3; *** P < 0.001 compared to NI. Figure S6. P. gingivalis increases expression of vimentin. Immunoblot of lysates of TIGK cells infected with P. gingivalis 33277 for 24 h at the MOI indicated. Control cells were uninfected (NI). Duplicate blots were probed with antibodies to vimentin or GAPDH (loading control). Figure S7. Colonization of mice. Mice were orally infected with 107 cfu P. gingivalis five times at 2-days intervals. Bacterial samples were collected along the gingiva of the upper molars. Samples were lysed, DNA extracted and qPCR performed with primers specific for P. gingivalis 16S DNA. For enumeration, genomic DNA was isolated from laboratory cultures of P. gingivalis 33277 (numbers determined by viable counting) and a series of dilutions prepared. The number of gene copies in the oral samples was determined by comparison with the standard curve. In the sham infected animals, 2 of 5 mice were colonized with low levels of organisms sufficient similar to P. gingivalis to give a positive result. P. gingivalis levels from day 1, 3 and 8 were statistically greater than sham infected (P < 0.0001) but were not statistically different from each other. Figure S8. Fluorescent confocal microscopy of a carcinoma in situ biopsy sample probed with P. gingivalis antibodies (green) and stained with DAPI (blue). Cells were imaged at magnification ×63. Red arrows point to a discrete fluorescent spot, yellow arrows indicate the same position where that spot is absent. Numbers are the slice number in an optical stack of 40 slices at 0.4 μm. Fluorescent spots are present in typically 5 to 7 adjacent optical slices (0.4 μm slices), indicating that the fluorescent particles are about 2.0 to 2.8 μm in size, consistent with the size of P. gingivalis. Grant Sponsors: We thank the NIH/NIDCR for support through DE011111, DE017921, DE016690 and DE023193. Figure 1 P. gingivalis up-regulates ZEB1 expression in TIGK cells in a FimA-dependent manner A. TIGKs were infected with P. gingivalis 33277 at the MOI and time indicated. ZEB1 mRNA levels were measured by qRT-PCR. Data were normalized to GAPDH mRNA and are expressed relative to noninfected (NI) controls. Results are means ± SD; n = 3; * P < 0.05; *** P < 0.001. B. Immunoblot of lysates of TIGK cells infected with P. gingivalis 33277 for 24 h at the MOI indicated. Control cells were uninfected (NI). Duplicate blots were probed with antibodies to ZEB1 or GAPDH (loading control). C. ZEB1 mRNA levels in TIGKs after 72 h infection with P. gingivalis 33277 at MOI indicated. qRT-PCR data were normalized to GAPDH mRNA and are expressed relative to noninfected (NI) controls. Results are means ± SD, n = 3; ** P < 0.01; *** P < 0.001. D. Fluorescent confocal microscopy of TIGK cells infected with P. gingivalis 33277 (Pg) at MOI 50 or MOI 100 for 24 h. Control cells were noninfected (NI). Cells were fixed and probed with ZEB1 antibodies (green). Actin (red) was stained with Texas Red-phalloidin, and nuclei (blue) stained with DAPI. Cells were imaged at magnification ×63, and shown are representative merged images of projections of z-stacks obtained with Volocity software. Bar = 10 μm. E. Nuclear localization of ZEB1 calculated by Pearson’s correlation coefficient from images in C (n=100 cells) using Volocity software. Results are mean ± SD; ** P < 0.01; *** P < 0.001. F. qRT-PCR of ZEB1 mRNA expression in TIGK cells infected with P. gingivalis (Pg) strains at MOI 100 for 24 h. qRT-PCR data were normalized to GAPDH mRNA and are expressed relative to noninfected (NI) controls. Results are means ± SD, n = 3; * P < 0.05; *** P < 0.001. G. qRT-PCR of ZEB1 mRNA expression in TIGK cells infected with P. gingivalis 33277 (Pg WT), ΔfimA mutant, or W83, or Pg WT in the presence of membrane insert. qRT-PCR data were normalized to GAPDH mRNA and are expressed relative to noninfected (NI) controls. Results are means ± SD, n = 3; *** P < 0.001 compared to NI; ### P < 0.001 compared to Pg WT. Figure 2 P. gingivalis communities regulate ZEB1 expression A. qRT-PCR of ZEB1 mRNA expression in TIGK cells infected with P. gingivalis 33277 (Pg), S. gordonii (Sg) or a combination of Pg and Sg at MOI 100 for 24 h. B. qRT-PCR of ZEB1 mRNA expression in TIGK cells infected with P. gingivalis 33277 (Pg), F. nucleatum (Fn) or a combination of Pg and Fn at MOI 100 for 24 h. Data were normalized to GAPDH mRNA and are expressed relative to noninfected (NI) controls. Results are means ± SD; n = 3; ** P < 0.01; *** P < 0.001. Figure 3 P. gingivalis regulates ZEB1 promoter activity and increases in ZEB1 levels are not dependent on the miRNA 200 family A. TIGK cells were transiently transfected with ZEB1 promoter-luciferase plasmids: Z1p 1000-Luc (−912 bp to +43 bp), Z1p.400-Luc (−367 bp to +43 bp) or Z1p.196-Luc (−212 bp to +43 bp); or with a constitutively-expressing Renilla luciferase reporter. Transfected cells were infected with P. gingivalis 33277 (Pg) at MOI 100 for 24 h. Control cells were noninfected (Ctr). Luciferase activity was normalized to the level of Renilla luciferase. Results are mean ± SD, n = 3; * P < 0.01; *** P < 0.001. B–D. Expression of mature miRNA-200b (B), miRNA-200c (C), or miRNA-205 (D) in TIGK cells infected with P. gingivalis 33277 MOI 100 for the times indicated. miRNA levels were measured by qRT-PCR, normalized to RNU48 miRNA, and expressed relative to noninfected (NI) controls. Results are means ± SD, n = 3; * P < 0.05; ** P < 0.01; *** P < 0.001. Figure 4 P. gingivalis induces MMP9 expression in a FimA-dependent manner TIGKs were infected with P. gingivalis 33277 (WT) or ΔfimA mutant at MOI 10 for 24 h, or left uninfected. A. Culture supernatants were analyzed for MMP9 and MMP2 by gelatin zymography. B. Quantitative analysis of 4 independent zymograms using ImageJ. * P < 0.05; *** P < 0.001. Figure 5 ZEB1 knockdown suppresses TIGK responses to P. gingivalis A. TIGK cells were transiently transfected with siRNA to ZEB1 (si Zeb1) or scrambled siRNA (si ctr). Control (ctr) cells were nontransfected. ZEB1 mRNA levels in transfected cells were measured by qRT-PCR. Data were normalized to GAPDH mRNA and are expressed relative to ctr. Results are means ± SD, n = 3; *** P < 0.001. B. Immunoblot of lysates of TIGK cells transfected (as in A) and probed with antibodies to ZEB1 or GAPDH (loading control). C–D. Transfected TIGK cells (as in A) were infected with P. gingivalis 33277 for 24 h at the MOI indicated. The Effect of ZEB1 knockdown on expression of vimentin (B) and MMP9 (C) was measured by qRT-PCR. Data were normalized to GAPDH mRNA and are expressed relative to the noninfected (NI) control. Results are means ± SD, n = 3; * P < 0.05; ** P < 0.01; *** P < 0.001 compared to NI. ### P < 0.001 compared to si ctr. Figure 6 P. gingivalis increases TIGK migration in a ZEB1-dependent manner Quantitative analysis of TIGK migration through matrigel-coated transwells. TIGK cells were transiently transfected with siRNA to ZEB1 (si Zeb1) or scrambled siRNA (si ctr). Control cells were nontransfected. TIGKs were infected with P. gingivalis 33277 for 24 h at the MOI indicated. Data are presented as the average number of cells invading through inserts coated with matrigel. Results are means ± SD, n = 3; *** P < 0.001 compared to NI. ### P < 0.001 compared to si ctr. Figure 7 P. gingivalis induces ZEB1 expression in vivo A. qRT-PCR of ZEB1 mRNA expression relative to GAPGH control in gingival tissues from mice infected with P. gingivalis or sham infected (NI). Tissue samples were collected at days 1, 3 and 8 after infection. Each point represents the determination from a single animal. * P < 0.05; ** P < 0.01. Figure 8 P. gingivalis antigens are present in OSCC Tissue biopsy of (A) poorly differentiated carcinoma, and (B) a tongue carcinoma in situ. Biospsy sections were stained with H&E, P. gingivalis 33277 antibodies (Anti-Pg 1:1000) or S. gordonii antibodies (Anti-Sg 1:1000) in the presence or absence of DAPI. Controls had no primary antibody. Red blood cells in the connective tissue are autofluorescent. Samples were imaged with a Zeiss Axiocam MRc5 fluorescence microscope at the magnification indicated. Table 1 Changes in expression of mesenchymal and epithelial markers in TIGK cells infected with P. gingivalis 33277. Fold change induced by P. gingivalis 33277 MOI 50 MOI 100 Mesenchymal markers Vimentin 1.72 ± 0.14 ** 3.4 ± 0.41 ** ITGA5 1.06 ± 0.12 1.29 ± 0.23 MMP-9 8.943 ± 0.38 *** 12.77 ± 0.75 *** Fibronectin 1.16 ± 0.05 3.13 ± 0.16 *** N-cadherin 2.55 ± 0.34** 2.4 ± 0.17** Epithelial markers KRT7 0.97 ± 0.09 1.29 ± 0.09 KRT19 1.2 ± 0.07 0.63 ± 0.003 *** COL-1A1 0.63 ± 0.06 ** 0.67 ± 0.035 *** Data represent qRT-PCR results of the individual genes normalized to that of GAPDH mRNA and expressed relative to noninfected cells. Results are means ± SD, n = 6; ** P < 0.01; *** P < 0.001 Conflict of Interest The authors have no conflict of interest to declare. Baud J Varon C Chabas S Chambonnier L Darfeuille F Staedel C 2013 Helicobacter pylori initiates a mesenchymal transition through ZEB1 in gastric epithelial cells PLoS One 8 e60315 23565224 Bedal KB Grassel S Oefner PJ Reinders J Reichert TE Bauer R 2014 Collagen XVI induces expression of MMP9 via modulation of AP-1 transcription factors and facilitates invasion of oral squamous cell carcinoma PLoS One 9 e86777 24466237 Belton CM Izutsu KT Goodwin PC Park Y Lamont RJ 1999 Fluorescence image analysis of the association between Porphyromonas gingivalis and gingival epithelial cells Cell Microbiol 1 215 223 11207554 Benitez-Paez A Belda-Ferre P Simon-Soro A Mira A 2014 Microbiota diversity and gene expression dynamics in human oral biofilms BMC Genomics 15 311 24767457 Bessede E Staedel C Acuna Amador LA Nguyen PH Chambonnier L Hatakeyama M 2014 Helicobacter pylori generates cells with cancer stem cell properties via epithelial-mesenchymal transition-like changes Oncogene 33 4123 4131 24096479 Bosshardt DD Lang NP 2005 The junctional epithelium: from health to disease J Dent Res 84 9 20 15615869 Bostanci N Belibasakis GN 2012 Porphyromonas gingivalis: an invasive and evasive opportunistic oral pathogen FEMS Microbiol Lett 333 1 9 22530835 Brabletz S Brabletz T 2010 The ZEB/miR-200 feedback loop--a motor of cellular plasticity in development and cancer? EMBO Rep 11 670 677 20706219 Castellarin M Warren RL Freeman JD Dreolini L Krzywinski M Strauss J 2012 Fusobacterium nucleatum infection is prevalent in human colorectal carcinoma Genome Res 22 299 306 22009989 Daep CA Novak EA Lamont RJ Demuth DR 2011 Structural dissection and in vivo effectiveness of a peptide inhibitor of Porphyromonas gingivalis adherence to Streptococcus gordonii Infect Immun 79 67 74 21041492 Enersen M Nakano K Amano A 2013 Porphyromonas gingivalis fimbriae J Oral Microbiol 5 doi:10.3402 Gallimidi AB Fischman S Revach B Bulvik R Maliutina A Rubinstein AM 2015 Periodontal pathogens Porphyromonas gingivalis and Fusobacterium nucleatum promote tumor progression in an oral-specific chemical carcinogenesis model Oncotarget 6 22613 22623 26158901 Garrett WS 2015 Cancer and the microbiota Science 348 80 86 25838377 Gheldof A Hulpiau P van Roy F De Craene B Berx G 2012 Evolutionary functional analysis and molecular regulation of the ZEB transcription factors Cell Mol Life Sci 69 2527 2541 22349261 Gur C Ibrahim Y Isaacson B Yamin R Abed J Gamliel M 2015 Binding of the Fap2 protein of Fusobacterium nucleatum to human inhibitory receptor TIGIT protects tumors from immune cell attack Immunity 42 344 355 25680274 Hajishengallis G 2014 The inflammophilic character of the periodontitis-associated microbiota Mol Oral Microbiol 29 248 257 24976068 Hajishengallis G Krauss JL Liang S McIntosh ML Lambris JD 2012 Pathogenic microbes and community service through manipulation of innate immunity Adv Exp Med Biol 946 69 85 21948363 Hajishengallis G Lamont RJ Graves DT 2015 The enduring importance of animal models in understanding periodontal disease Virulence 6 229 235 25574929 Handfield M Mans JJ Zheng G Lopez MC Mao S Progulske-Fox A 2005 Distinct transcriptional profiles characterize oral epithelium-microbiota interactions Cell Microbiol 7 811 823 15888084 Hendrickson EL Wang T Beck DA Dickinson BC Wright CJ Lamont RJ Hackett M 2014 Proteomics of Fusobacterium nucleatum within a model developing oral microbial community Microbiologyopen 3 729 751 25155235 Heymann R Wroblewski J Terling C Midtvedt T Obrink B 2001 The characteristic cellular organization and CEACAM1 expression in the junctional epithelium of rats and mice are genetically programmed and not influenced by the bacterial microflora J Periodontol 72 454 460 11338297 Inaba H Sugita H Kuboniwa M Iwai S Hamada M Noda T 2014 Porphyromonas gingivalis promotes invasion of oral squamous cell carcinoma through induction of proMMP9 and its activation Cell Microbiol 16 131 145 23991831 Jemal A Siegel R Ward E Hao Y Xu J Murray T Thun MJ 2008 Cancer statistics, 2008 CA Cancer J Clin 58 71 96 18287387 Jia B Liu H Kong Q Li B 2012 Overexpression of ZEB1 associated with metastasis and invasion in patients with gastric carcinoma Mol Cell Biochem 366 223 229 22466758 Kamarajan P Garcia-Pardo A D’Silva NJ Kapila YL 2010 The CS1 segment of fibronectin is involved in human OSCC pathogenesis by mediating OSCC cell spreading, migration, and invasion BMC Cancer 10 330 20579373 Katz J Onate MD Pauley KM Bhattacharyya I Cha S 2011 Presence of Porphyromonas gingivalis in gingival squamous cell carcinoma Int J Oral Sci 3 209 215 22010579 Khanom R Sakamoto K Pal SK Shimada Y Morita K Omura K 2012 Expression of basal cell keratin 15 and keratin 19 in oral squamous neoplasms represents diverse pathophysiologies Histol Histopathol 27 949 959 22648550 Kim SS Ruiz VE Carroll JD Moss SF 2011 Helicobacter pylori in the pathogenesis of gastric cancer and gastric lymphoma Cancer Lett 305 228 238 20692762 Kostic AD Gevers D Pedamallu CS Michaud M Duke F Earl AM 2012 Genomic analysis identifies association of Fusobacterium with colorectal carcinoma Genome Res 22 292 298 22009990 Kruger A Arlt MJ Gerg M Kopitz C Bernardo MM Chang M 2005 Antimetastatic activity of a novel mechanism-based gelatinase inhibitor Cancer Res 65 3523 3526 15867341 Kuboniwa M Hasegawa Y Mao S Shizukuishi S Amano A Lamont RJ Yilmaz O 2008 P. gingivalis accelerates gingival epithelial cell progression through the cell cycle Microbes Infect 10 122 128 18280195 Lamont RJ Chan A Belton CM Izutsu KT Vasel D Weinberg A 1995 Porphyromonas gingivalis invasion of gingival epithelial cells Infect Immun 63 3878 3885 7558295 Lamont RJ Hajishengallis G 2015 Polymicrobial synergy and dysbiosis in inflammatory disease Trends Mol Med 21 172 183 25498392 Lamont RJ Jenkinson HF 1998 Life below the gum line: pathogenic mechanisms of Porphyromonas gingivalis Microbiol Mol Biol Rev 62 1244 1263 9841671 Lamouille S Xu J Derynck R 2014 Molecular mechanisms of epithelial-mesenchymal transition Nat Rev Mol Cell Biol 15 178 196 24556840 Lee CH Chang JS Syu SH Wong TS Chan JY Tang YC 2015 IL-1beta promotes malignant transformation and tumor aggressiveness in oral cancer J Cell Physiol 230 875 884 25204733 Liu Y Costantino ME Montoya-Durango D Higashi Y Darling DS Dean DC 2007 The zinc finger transcription factor ZFHX1A is linked to cell proliferation by Rb-E2F1 Biochem J 408 79 85 17655524 Manavella PA Roqueiro G Darling DS Cabanillas AM 2007 The ZFHX1A gene is differentially autoregulated by its isoforms Biochem Biophys Res Commun 360 621 626 17610840 Mao S Park Y Hasegawa Y Tribble GD James CE Handfield M 2007 Intrinsic apoptotic pathways of gingival epithelial cells modulated by Porphyromonas gingivalis Cell Microbiol 9 1997 2007 17419719 Michaud DS 2013 Role of bacterial infections in pancreatic cancer Carcinogenesis 34 2193 2197 23843038 Moffatt-Jauregui CE Robinson B de Moya AV Brockman RD Roman AV Cash MN 2013 Establishment and characterization of a telomerase immortalized human gingival epithelial cell line J Periodontal Res 48 713 721 23441958 Moffatt CE Lamont RJ 2011 Porphyromonas gingivalis induction of microRNA-203 expression controls suppressor of cytokine signaling 3 in gingival epithelial cells Infect Immun 79 2632 2637 21536793 Nadkarni MA Chhour KL Chapple CC Nguyen KA Hunter N 2014 The profile of Porphyromonas gingivalis kgp biotype and fimA genotype mosaic in subgingival plaque samples FEMS Microbiol Lett 361 190 194 25353706 Nagy KN Sonkodi I Szoke I Nagy E Newman HN 1998 The microflora associated with human oral carcinomas Oral Oncol 34 304 308 9813727 Naito M Hirakawa H Yamashita A Ohara N Shoji M Yukitake H 2008 Determination of the genome sequence of Porphyromonas gingivalis strain ATCC 33277 and genomic comparison with strain W83 revealed extensive genome rearrangements in P. gingivalis DNA Res 15 215 225 18524787 Nishikawa K Duncan MJ 2010 Histidine kinase-mediated production and autoassembly of Porphyromonas gingivalis fimbriae J Bacteriol 192 1975 1987 20118268 Rubinstein MR Wang X Liu W Hao Y Cai G Han YW 2013 Fusobacterium nucleatum promotes colorectal carcinogenesis by modulating E-cadherin/beta-catenin signaling via its FadA adhesin Cell Host Microbe 14 195 206 23954158 Sahingur SE Yeudall WA 2015 Chemokine function in periodontal disease and oral cavity cancer Front Immunol 6 214 25999952 Sanchez-Tillo E Liu Y de Barrios O Siles L Fanlo L Cuatrecasas M 2012 EMT-activating transcription factors in cancer: beyond EMT and tumor invasiveness Cell Mol Life Sci 69 3429 3456 22945800 Sawant SS Vaidya M Chaukar DA Alam H Dmello C Gangadaran P 2014 Clinical significance of aberrant vimentin expression in oral premalignant lesions and carcinomas Oral Dis 20 453 465 23865921 Scanlon CS Van Tubergen EA Inglehart RC D’Silva NJ 2013 Biomarkers of epithelial-mesenchymal transition in squamous cell carcinoma J Dent Res 92 114 121 23128109 Takeuchi H Hirano T Whitmore SE Morisaki I Amano A Lamont RJ 2013 The serine phosphatase SerB of Porphyromonas gingivalis suppresses IL-8 production by dephosphorylation of NF-kappaB RelA/p65 PLoS Pathog 9 e1003326 23637609 Tribble GD Kerr JE Wang BY 2013 Genetic diversity in the oral pathogen Porphyromonas gingivalis: molecular mechanisms and biological consequences Future Microbiol 8 607 620 23642116 Tribble GD Lamont GJ Progulske-Fox A Lamont RJ 2007 Conjugal transfer of chromosomal DNA contributes to genetic variation in the oral pathogen Porphyromonas gingivalis J Bacteriol 189 6382 6388 17573478 Valm AM Mark Welch JL Rieken CW Hasegawa Y Sogin ML Oldenbourg R 2011 Systems-level analysis of microbial community organization through combinatorial labeling and spectral imaging Proc Natl Acad Sci U S A 108 4152 4157 21325608 Vandewalle C Van Roy F Berx G 2009 The role of the ZEB family of transcription factors in development and disease Cell Mol Life Sci 66 773 787 19011757 Watanabe K Yilmaz O Nakhjiri SF Belton CM Lamont RJ 2001 Association of mitogen-activated protein kinase pathways with gingival epithelial cell responses to Porphyromonas gingivalis infection Infect Immun 69 6731 6737 11598045 Westermarck J Kahari VM 1999 Regulation of matrix metalloproteinase expression in tumor invasion FASEB J 13 781 792 10224222 Whitmore SE Lamont RJ 2014 Oral bacteria and cancer PLoS Pathog 10 e1003933 24676390 Wikner J Grobe A Pantel K Riethdorf S 2014 Squamous cell carcinoma of the oral cavity and circulating tumour cells World J Clin Oncol 5 114 124 24829858 Wright CJ Burns LH Jack AA Back CR Dutton LC Nobbs AH 2013 Microbial interactions in building of communities Mol Oral Microbiol 28 83 101 23253299 Wright CJ Xue P Hirano T Liu C Whitmore SE Hackett M Lamont RJ 2014 Characterization of a bacterial tyrosine kinase in Porphyromonas gingivalis involved in polymicrobial synergy Microbiologyopen 3 383 394 24811194 Yang G Yuan J Li K 2013 EMT transcription factors: implication in osteosarcoma Med Oncol 30 697 23975634 Yilmaz O Jungas T Verbeke P Ojcius DM 2004 Activation of the phosphatidylinositol 3-kinase/Akt pathway contributes to survival of primary epithelial cells infected with the periodontal pathogen Porphyromonas gingivalis Infect Immun 72 3743 3751 15213114 Yilmaz O Verbeke P Lamont RJ Ojcius DM 2006 Intercellular spreading of Porphyromonas gingivalis infection in primary gingival epithelial cells Infect Immun 74 703 710 16369027 Zhang P Cai Y Soofi A Dressler GR 2012 Activation of Wnt11 by transforming growth factor-beta drives mesenchymal gene expression through non-canonical Wnt protein signaling in renal epithelial cells J Biol Chem 287 21290 21302 22556418 Zhang P Wei Y Wang L Debeb BG Yuan Y Zhang J 2014 ATM-mediated stabilization of ZEB1 promotes DNA damage response and radioresistance through CHK1 Nat Cell Biol 16 864 875 25086746 Zhao D Tang XF Yang K Liu JY Ma XR 2012 Over-expression of integrin-linked kinase correlates with aberrant expression of Snail, E-cadherin and N-cadherin in oral squamous cell carcinoma: implications in tumor progression and metastasis Clin Exp Metastasis 29 957 969 22638656 Zhou Q Amar S 2007 Identification of signaling pathways in macrophage exposed to Porphyromonas gingivalis or to its purified cell wall components J Immunol 179 7777 7790 18025224

PMC005xxxxxx/PMC5135582.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 9502500 8794 Clin Cancer Res Clin. Cancer Res. Clinical cancer research : an official journal of the American Association for Cancer Research 1078-0432 27354472 5135582 10.1158/1078-0432.CCR-15-2190 NIHMS800337 Article Hsp90 Inhibitor Ganetespib Sensitizes Non-Small Cell Lung Cancer to Radiation but Has Variable Effects with Chemoradiation Wang Yifan 1210 Liu Hui 1810 Diao Lixia 3 Potter Adam 4 Zhang Jianhu 1 Qiao Yawei 1 Wang Jing 3 Proia David A. 59 Tailor Ramesh C. 6 Komaki Ritsuko 7 Lin Steven H. 17 1 Department of Experimental Radiation Oncology, the University of Texas M.D. Anderson Cancer Center, Houston, TX 77030, USA 2 The University of Texas Graduate School of Biomedical Sciences, Houston, TX 77030, USA 3 Department of Bioinformatics and Computational Biology, the University of Texas M.D. Anderson Cancer Center, Houston, TX 77030, USA 4 Texas A&M School of Medicine, College Station, TX 77843, USA 5 Synta Pharmaceuticals Corp, Lexington, MA 02421, USA 6 Department of Radiation Physics, the University of Texas M.D. Anderson Cancer Center, Houston, TX 77030, USA 7 Department of Radiation Oncology, the University of Texas M.D. Anderson Cancer Center, Houston, TX 77030, USA Corresponding Author: Steven H. Lin, MD, PhD, Department of Radiation Oncology, the University of Texas MD Anderson Cancer Center, 1515 Holcombe Blvd., Unit 097, Houston, TX 77030 USA. shlin@mdanderson.org; Phone: 713-563-8490 8 Current Address: Department of Investigational Cancer Therapeutics, the University of Texas M.D. Anderson Cancer Center, Houston, TX 77030, USA 9 Current Address: C4 Therapeutics Inc., Boston, MA 02142, USA 10 These two authors contributed equally to this work 9 7 2016 28 6 2016 1 12 2016 01 6 2017 22 23 58765886 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. Purpose HSP90 inhibition is well known to sensitize cancer cells to radiation. However, it is currently unknown if additional radiosensitization could occur in the more clinically relevant setting of chemoradiation (CRT). We used the potent HSP90 inhibitor ganetespib to determine if it can enhance CRT effects in NSCLC. Experimental Design We first performed in vitro experiments in various NSCLC cell lines combining radiation with or without ganetespib. Some of these experiments included clonogenic survival assay, DNA damage repair and cell cycle analysis, and Reverse Phase Protein Array. We then determined if chemotherapy affected ganetespib radiosensitization by adding carboplatin-paclitaxel to some of the in vitro and in vivo xenograft experiments. Results Ganetespib significantly reduced radiation clonogenic survival in a number of lung cancer cell lines, and attenuated DNA damage repair with irradiation. Radiation caused G2/M arrest that was greatly accentuated by ganetespib. Ganetespib with radiation also dose-dependently up-regulated p21 and down-regulated pRb levels that were not apparent with either drug or radiation alone. However, when carboplatin-paclitaxel was added, ganetepsib was only able to radiosensitize some cell lines but not to others. This variable in vitro CRT effect was confirmed in vivo using xenograft models. Conclusions Ganetespib was able to potently sensitize a number of NSCLC cell lines to radiation but has variable effects when added to platinum-based doublet CRT. For optimal clinical translation, our data emphasizes the importance of preclinical testing of drugs in the context of clinically-relevant therapy combinations. Hsp90 inhibitor non-small cell lung cancer chemoradiotherapy Introduction Lung cancer is the leading cause of cancer death in the United States and has a 5-year relative survival rate of only 16% (1). Non-small cell lung cancer (NSCLC) accounts for about 85% of all lung cancer cases and the lack of significant treatment advance is related to the highly resistant nature of this disease. While chemotherapy provides only useful palliation for stage IV NSCLC, the treatment of locally advanced, unresectable NSCLC is with curative intent using concomitant chemotherapy and radiotherapy (chemoradiotherapy, CRT), which produces longer overall survival than sequential chemotherapy and radiation therapy (2–6), but the outcomes remain poor. The median survival ranges from 17–28 months, despite significantly increased toxicity of the combination therapy. There is a strong need to improve therapy efficacy in NSCLC without substantially increasing normal tissue toxicity. Indeed, recent clinical trials have investigated the combination of CRT with molecularly targeted agents, either with angiogenesis inhibitors or with EGFR targeting agents. Unfortunately, either due to intolerable toxicities (7) or to lack of efficacy (8), these trials have not advanced the management of this disease. Hsp90 is a molecular chaperone protein ubiquitously present in cells; however its function is critically important for the maintenance of cancer cell growth (9, 10). Inhibiting its function has been extensively studied for its potent antitumor effect (9, 11, 12). An attractive feature of targeting Hsp90 is that the cytotoxicity of Hsp90 inhibitors is tumor selective (13). Hsp90 inhibition has also been known to be radiation sensitizing on tumor cells (14–25). Some of these studies indicated that the radiation sensitization is also tumor selective since normal cells are not affected (14, 15, 17). However, the clinical development of Hsp90 inhibitors has been hampered by the severe toxicities of first generation inhibitors, including severe ocular and hepatic toxicities (26, 27). Ganetespib, a second generation Hsp90 inhibitor with little to no ocular or hepatic toxicities, has been safely used in thousands of patients in over 60 clinical trials internationally. A completed phase II randomized trial in stage IV NSCLC combining ganetespib with docetaxel compared to docetaxel alone has demonstrated efficacy signal in a subgroup of patients (28), and therefore it was further tested in a phase III randomized trial (GALAXY II) (NCT01798485). Despite the safety and promising efficacy of this drug in advanced NSCLC, the experience of combining ganetespib with CRT is limited. While radiation sensitizing effect is well known for this class of inhibitors, one simply cannot assume that synergy could be seen with CRT. This was once assumed for EGFR inhibitors when preclinical studies demonstrated synergy with radiotherapy alone (29, 30), but ultimately failed in a number of phase III clinical trials when EGFR targeting agents were combined with CRT in oropharyngeal cancer, esophageal cancer, and NSCLC(8, 31, 32). While this manuscript was under preparation, Gomez-Casal et al. reported the HSP90 inhibitor ganetespib radiosensitizes human lung adenocarcinoma cells(33), however whether ganetespib enhanced therapeutic effects of chemoradiotherapy was not demonstrated. The purpose of the current study is to evaluate the cytotoxic action of the combination of radiation with ganetespib and test its potential to synergize with CRT for the treatment of NSCLC. Materials and Methods Cell culture, reagents and irradiator The human non-small lung cancer cell lines H460, A549, H1299, H1650, H358 and H2087 cells were all obtained from the American Type Culture Collection (ATCC) and routinely maintained in RPMI-1640 medium supplemented with 10% FBS, and 10,000 U/mL of penicillin-streptomycin. Cell lines were authenticated at Characterized Cell Line Core Facility at MD Anderson Cancer Center using the Short Tandem Repeat method every 6 months of use in the lab. Ganetespib [3-(2,4-dihydroxy-5-isopropylphenyl)-4-(1-methyl-1H-1,2,4-triazol-5(4H)-1] was provided by Synta Pharmaceuticals Corp. Cells and animals were irradiated with a JL Shepherd Mark I-68A 137Cs irradiator with 137Cs sources at the doses from 0–6 Gy. The Cesium-Irradiator output (cGy/min) was measured in-air using an ADCL (Accredited-Dosimetry-Calibration-Laboratory) calibrated ion-chamber. Dosimetry in Simulated-Irradiation-Geometry was performed employing Gafchromic-Film “EBT3”. For Film-Dosimetry, mouse was simulated by dosimetrically-equivalent Gel “SuperFlab”. EBT3 response in Simulated-Geometry versus in-Air Reference-Calibration-Geometry provided dose-rate in mouse. The treatment set-up employed table, cerrobend-block and mouse-restrainer on top. Cesium beam pointed up. All these pieces were provided with mutually inter-locking pins to ensure set-up reproducibility. For reduced penumbra, the block was provided with appropriate divergence. Clonogenic survival assay (CSA) to determine ganetespib radiation sensitization effect The effectiveness of the combination of ganetespib and ionizing radiation was assessed by CSA. H460, A549, H1299, and H1650 cells were seeded (100–2000 cells/well) in duplicate in 6-well plates. The medium was changed 16 hours after plating and the cells were treated with either vehicle (DMSO) or ganetespib (30 nM). Five hours following ganetespib treatment, the cells were subjected to irradiation at doses from 0–6 Gy. Twenty-four hours after ganetespib treatment, media was changed and the cells were maintained in normal culture conditions. On about 12th–20th day, the medium was removed and cell colonies were stained with crystal violet (0.1% in 20% methanol) (Sigma-Aldrich, St. Louis, MO, USA). Colony numbers were assessed visually and colonies containing > 50 normal-appearing cells were counted. The surviving fraction was calculated using SigmaPlot 10.0 (San Jose, CA, USA). Clonogenic survival assay for CRT with ganetespib The combination effect of chemo radiation and ganetespib was assayed in vitro using clonogenic survival assay. Cells were seeded in 6-well plates 16 hours prior to treatment. The cells were treated by combination chemotherapy (Paclitaxel 3.51 nM with Carboplatin 24.23 nM) and/or ganetespib (30 nM) followed by radiation (2 Gy) after 4–5 hours. Drugs were washed out 24 hours after the treatment and the cells were maintained in normal cultural conditions for 12–15 days. The colonies were stained by crystal violet (0.1% in 20% methanol) (Sigma-Aldrich, St. Louis, MO, USA). Colony numbers were assessed visually and colonies containing > 50 normal-appearing cells were counted. The surviving fraction was calculated using GraphPad Prism 6 (La Jolla, CA, USA). DNA repair foci formation assay H460 and A549 cells were grown as monolayers on chamber slides with plastic bottom (Nunc Lab-Tek, Roskilde, Denmark) and were treated with DMSO or ganetespib (25 nM, 50 nM) 24 hours after seeding into culture chambers. Five hours after ganetespib treatment, the cells were subjected to irradiation at dose of 2 Gy. Thirty minutes, 4 hours, 24 hours and 48 hours after irradiation, cells were fixed in 4% paraformaldehyde in PBS for 15 min at room temperature and washed in PBS. The cells were then permeabilized in 0.5% Triton X-100 for 10 min, and blocked in PBS with 3% BSA (Bovine serum albumin) for 20 min at room temperature. The cells were sequentially incubated with anti-53BP1 antibody (Cell Signaling Technology, Danvers, MA, USA) at 1:100 dilution overnight at 4 °C and Alexa Fluor® 488 Conjugate secondary antibody (Cell Signaling Technology, Danvers, MA, USA) at a 1:1000 dilution for 1 hour at 37°C in PBS with 1.5% BSA and washed three times in PBS. Nuclei were counterstained with 1:500 4-diamidino-2-phenylindole dihydrochloride (DAPI) in PBS. Cover glasses were mounted in Vectashield (Vector Laboratories). Fluorescence images were captured with Leica fluorescence microscope equipped with a CCD camera and images were imported into Advanced Spot Image analysis software. DNA repair foci were quantified using ImageJ software (NIH public domain). Flow cytometry cell cycle analysis H460 and A549 cells were treated with vehicle (DMSO) or ganetespib (50 nM) with or without irradiation as described above and were harvested 4 hours, 16 hours, 24 hours, 48 hours, 72 hours and 96 hours after irradiation and fixed in 70% ethanol for 1 hour at 4°C. The cells were stained with propidium iodide (Sigma-Aldrich, St. Louis, MO, USA) for 15 min in the dark and FACS analysis was performed using a Becton-Dickinson FACS Calibur flow cytometer (BD Biosciences, Heidelberg, Germany) per the manufacturer’s instructions. Assays were performed in triplicates. The percentage of cells in each phase of the cell cycle (sub-G1, G1, S and G2/M) was determined with Flow-Jo (TreeStar Inc., Ashland, OR). Flow cytometry apoptosis assay Cells were treated with vehicle (DMSO) or ganetespib (50 nM) followed by irradiation (2 Gy) after 4 hours. Cells were trypsinized 24 hours post irradiation and stained by AnnexinV-FITC and propidium iodide using the Abcam apoptosis detection kit (ab14085, Abcam, Cambridge, United Kingdom) according to manufacturer’s instructions. The flow analysis was performed on Becton-Dickinson FACS Calibur flow cytometer (BD Biosciences, Heidelberg, Germany) and analyzed with Flow-Jo (TreeStar Inc., Ashland, OR). The assays were done in triplicate. Reverse phase protein array (RPPA) H460 and A549 cell lines were treated with vehicle (DMSO) or ganetespib (10 nM, 25 nM, 50 nM and 100 nM) with or without irradiation as described above and were harvested 24 hours after ganetespib treatment. Cell lysates were prepared using lysis buffer containing 1% Triton X-100, 50mM HEPES, pH 7.4, 150mM NaCl, 1.5mM MgCl2, 1mM EGTA, 100mM NaF, 10mM Na pyrophosphate, 1mM Na3VO4, 10% glycerol, containing freshly added protease and phosphatase inhibitors and 4× SDS sample buffer containing 40% Glycerol, 8% SDS, 0.25M Tris-HCL, pH 6.8. and 10% volume of 2-mercaptoethanol, arrayed on nitrocellulose-coated FAST slides (Whatman), and probed for a standard list of antibodies (34). Immunoblot analysis H460 cell lines were treated with vehicle (DMSO) or ganetespib (25 nM, 50 nM) with or without irradiation (2 Gy, 4 Gy) as described above and were harvested 24 hours after genetespib treatment. After washing with ice-cold PBS, cell lysates were collected in RIPA Buffer (Life Technologies, Grand Island, NY, USA) with protease inhibitor cocktail (Roche, Eugene, OR USA). Protein concentrations were measured by protein assay kit (Life Technologies, Grand Island, NY, USA). Total cellular lysates were loaded onto 4–12% Bis-Tris gel (Bio-Rad, Hercules, CA, USA), separated by electrophoresis and electro-transferred onto nitrocellulose membranes. Membranes were blocked for 1 h at room temperature in 5% nonfat dry milk in Tris-Buffer saline containing 0.1% Tween20 (TBST). The membrane was incubated with primary antibodies (p21, phospho-CDC2, CDC2, phospho-p90RSK, p90RSK, phospho-AKT, AKT, phospho-mTOR, mTOR, phospho-EGFR, EGFR, pS6, cleaved caspase 9 and caspase 7, phospho-histone γH2AX, GAPDH, β-actin antibodies are all from Cell Signaling Technology, Danvers, MA, USA) at 4 °C overnight and subsequently incubated with a secondary antibody for 1 hour at room temperature. After washing the membrane with TBST, HRP activity was detected using Clarity Western ECL substrate (Bio-Rad, Hercules, CA, USA) and analyzed by ChemiDoc imager (Bio-Rad, Hercules, CA, USA). In vivo NSCLC xenograft model Female NCr nu/nu mice (5 weeks old, from Taconic Biosciences, Germantown, NY, USA) were used for tumor studies. Animals were maintained in an Association for the Assessment and Accreditation of Laboratory Animal Care approved facility in accordance with current regulations of the U.S. Department of Agriculture and Department of Health and Human Services. Experimental methods were approved by and in accordance with institutional guidelines established by the Institutional Animal Care and Use Committee. For H460 xenografts, mice were subcutaneously inoculated with a total of 1 × 106 H460 cells in 20 ul PBS into their right hind legs to prepare the tumor model. For A549 xenografts, mice were subcutaneously inoculated with a total of 5 × 106 cells in 50 ul PBS into the right hind legs. When the tumor reached 100 mm3 (H460) or 150 mm3 (A549), mice were randomized into different treatment groups (4–10 animals per group). Mice were treated with Carboplatin (30mg/kg)-Paclitaxel (10mg/kg) intraperitoneally once weekly for two weeks and/or intravenously via the tail vein with 100 mg/kg ganetespib dissolved in 200 ul 10/18 DRD (10% DMSO, 18% Cremophore RH 40, 3.8% dextrose) once weekly for two weeks. Mice not included in the ganetespib treatment groups received the same volume of solvent. Five hours after drug treatment, irradiation was applied locally to the tumor-bearing legs of unanesthetized mice at 2 Gy once daily for 5 days. Tumor volumes (V) were calculated by caliper measurements of the width (W), and length (L) of each tumor using the formula: V=0.5 × (L × W2). The tumor growth curve was calculated using GraphPad Prism 6 (La Jolla, CA, USA). Immunohistochemistry Tumors were fixed in 1:10 diluted formalin (Fisher scientific, Cat# 245-684, Waltham, MA USA) overnight and embedded in paraffin. Tissue sections (2 mm) were deparaffinized in 100% xylene and rehydrated through incubation in descending ethanol dilutions (100–60%) followed by boiling at 125 °C for 2 min in Citrate buffers (10 mM Sodium Citrate pH 6). To reduce the endogenous peroxidase activity, slides were treated with 3% H2O2 for 10 min and subsequently probed with the primary antibody anti-cleaved caspase 7 (Cell Signaling Technology, Danvers, MA, USA) at 1:1000 dilution and secondary antibody SignalStain® Boost IHC Detection Reagent (HRP, Rabbit) (Cell Signaling Technology, Danvers, MA, USA). DAB (SignalStain® DAB Substrate Kit, Cell Signaling Technology, Danvers, MA, USA) was used as chromogen. Tissue sections were imaged using a PerkinElmer Vectra 2 microscope (PerkinElmer, Waltham, MA, USA) and analyzed with PerkinElmer Inform 2.1 analysis software (PerkinElmer, Waltham, MA, USA) at MD Anderson Cancer Center Flow Cytometry & Cellular Imaging Core Facility. Statistical analysis Statistical significance was assessed by Student’s t-test (2 sample assuming unequal variances) and expressed as standard mean error. A difference was considered significant if p < 0.05. RPPA statistical analysis used two-way analysis of variance (ANOVA) performed on a marker-by-marker base to test for the interaction between radiation and drug treatment and one-way ANOVA for drug only effect. Pairwise comparisons between different drug doses were done using Tukey's honest significance (HSD) Test with 95% family-wise confidence level. To control for multiple testing, the resulting p-values, computed from test statistics applied, were modeled using a beta-uniform mixture (BUM) model in order to select p-value cutoffs according to pre-defined false discovery rates (FDRs). All statistical analyses of RPPA data were performed using R (version R3.1.0) and Bioconductor packages (http://www.r-project.org/). Results Ganetespib sensitizes NSCLC cells to radiation We assessed the ability of ganetespib to radiosensitize human NSCLC cells of varying genetic backgrounds using clonogenic survival curve assays. We evaluated the Kras mutant/p53 wild type cells H460 and A549, Kras wild type/p53 mutant cells H1650 cells, and the Kras wild type/p53 null H1299 cell lines. Cells were exposed to ganetespib at 30 nM for 5 hours, subsequently irradiated with gamma-rays and incubated for a further 19 hours in the presence of ganetespib. Irradiation in combination with ganetespib had a strong radiosensitizing effect on H460 and H1299 cells. Moderate sensitizing effects of ganetespib were observed for the A549 and H1650 cells (Figure 1 A–D). The radiosensitivity enhancement ratios at a survival rate of 50% were 1.87 in the H460 cell, 1.67 in the A549, 1.65 in H1650 and 2.4 in H1299. The results indicated that ganetespib can potentiate the radiation effect in different NSCLC cells. Ganetespib inhibits radiation-induced DNA damage repair foci in NSCLC cancer cells Double-strand DNA breaks induced by radiation, if unrepaired, can lead to genomic instability and cell death. To determine if radiation induced DNA damage repair can be hampered by ganetespib, we measured the DNA damage repair response by evaluating 53BP1 foci formation (Figure 2A–C). In both the H460 and A549 cells, a substantial rise in 53BP1 foci formation is seen after 30 minutes of 2 Gy radiation for both the IR alone group and the IR+ganetespib group, but not apparent with ganetespib alone. While this effect is sustained at 4 hours, by 24 hours, the number of foci was reduced in the IR alone group but not to the same level as the ganetespib alone group. However, in both cell lines, the number of foci in the cells exposed to both ganetespib and radiation was sustained significantly longer compared with radiation alone at both the 24 and 48 hour time points. This suggests that ganetespib impairs the repair of radiation-induced double-strand DNA breaks. G2/M arrest induced by irradiation is further intensified by ganetespib pretreatment As Hsp90 inhibitors are known to affect the cell cycle checkpoint (19, 21–24, 35, 36), as does irradiation, which causes a G1/S arrest in P53 WT cells (37), we next determined the interaction of radiation with ganetespib on the cell cycle effects in both H460 and A549 cells (Figure 3A–D). G2/M arrest was induced by ganestepib, radiation, and the combination, at 4 hours postirradiation. For both H460 and A549, the irradiation effect normalized by 16 hours, but for ganetespib alone the effect continued for 16 hours, but is mostly restored to baseline by 48 hours. The addition of ganetespib appeared to further augment the radiation-induced G2/M arrest in both H460 and A549 cells at 16–24 hour, an effect that is not fully restored to baseline in H460 until nearly 96 hours post-irradiation (Figure 3B). Ganetespib pretreatment accentuates the radiation effect by altering expression level of checkpoint proteins and kinase activities We next sought to define the mechanisms by which ganetespib reduced clonogenic survival and sensitized NSCLC cells to radiation. Cell lysate was prepared from H460 cells and A549 cells treated with ganetespib at 10 nM, 25 nM, 50 nM and 100 nM, with or without 2 and 4 Gy radiation for 24 hours, and RPPA was performed. As shown in Supplemental Figures 1 and 2, many oncogenic proteins were downregulated by treatment of ganetespib alone including phospho-Akt, phospho-mTOR, phospho-p90RSK, and EGFR. Although radiation can induce the activation of Akt and mTOR at the doses of 2–4 Gy, as previously reported(38, 39), the combination of ganetespib suppressed this activation along with other phosphokinases in general (Supplemental Figure 3A). In addition to oncogenic proteins, ganetespib pretreatment followed by radiation also altered the expression levels of checkpoint proteins (Figure 4A–D, Supplemental Figure 3B). Ganetespib dose-dependently upregulated pCDC2 (pCDK1), but in combination with radiation there was instead a downregulation of pCDC2 at higher ganetespib levels. CDC2 is known to be involved in the G2/M checkpoint by regulating the activity of cyclin-dependent kinase inhibitor p21 which regulates G2/M transition. Ganetespib combined with radiation reduced the activity of phospho-p90RSK and phospho-S6, which is known as a regulator of CDC2 (Figure 4A–D, Supplemental Figure 3B) at the dose of radiation (2–6 Gy) and ganetespib (25–100 nM). In A549 cells, cell cycle checkpoint proteins p21, p27, Chk1, Rb and phospho-S6 are also altered after the combined treatment (Figure 4C–D). Ganetespib combined with radiation also increased cleaved caspase 7 and caspase 9 and DNA damage repair protein phospho-γH2AX that were not apparent with radiation alone (Supplemental Figure 3B). Apoptosis was also determined by Annexin V-PI staining and flow cytometry analysis (Supplemental Figure 4 A–B). There was a significant increase in the percentage of apoptotic cells with ganetespib treatment alone, an effect that was not enhanced with radiation. Ganetespib demonstrated variable sensitizing effects with CRT in vitro and in vivo Since CRT is the standard treatment for unresectable locally advanced NSCLC, adding a radiation sensitizer like ganetespib may improve clinical outcomes. In order to assay whether ganetespib enhances the treatment efficacy of CRT, we performed clonogenic survival assay using A549 and H460 (Figure 5A–B). Robust clonogenic survival reduction was again seen in both cell lines when treated with ganetespib and IR. When chemotherapy was added, ganetespib was able to further enhance treatment efficacy for both chemotherapy and CRT in H460 cells but not in A549 cells. In A549, there was even a paradoxical increase in clonogenic survival when ganetespib was added to chemotherapy or CRT. Since A549 and H460 are both p53 wild-type cells, we tested two cell lines that harbored p53 mutations, H2087 and H358 (Figure 5 C–D). While ganetespib was again able to sensitize both of the cell lines to radiation, this was no longer apparent in the H2087 cells when co-treated with chemotherapy. Next, we assessed the effect of ganetespib on anti-tumor treatment sensitivity in vivo compared to CRT alone. For this, we established a treatment protocol that reflected a shortened version of what’s done clinically for unresectable NSCLC using fractionated daily 2 Gy radiation and concurrent chemotherapy with carboplatin and paclitaxel for 5 days, followed by consolidation chemotherapy alone for the second week. We compared this “standard therapy” to chemotherapy alone or radiation alone, with or without two doses of once-weekly ganetespib. As shown in Figure 6A, in H460 xenografts, chemotherapy alone showed only minimal tumor growth inhibition, but when combined with ganetespib, the anti-tumor effect of chemotherapy was significantly increased. As expected, combined CRT had much stronger anti-tumor effect than either chemotherapy or radiation alone. However, ganetespib added to CRT produced the greatest tumor growth delay (Figure 6A). The addition of ganetespib to CRT delayed tumor growth by 7 days compared with CRT and by 17 days compared with non-treated control. To determine if the DNA damage and cytotoxic effects of treatment in vivo could be compared to what we saw in vitro, we performed immunohistochemical staining for 53BP1 foci and cleaved caspase 7 on tumors removed at the end of the first 5 days of treatment. Compared to all the treatment groups, the tumors treated with the combination of ganetespib with CRT had significantly higher levels of P53BP1 foci and cleaved caspase 7 levels (Supplemental Figure 5). These results demonstrate that the enhanced DNA damage and cytotoxic effect of ganetespib with radiation seen in vitro could be further enhanced with concurrent CRT. Given the paradoxical in vitro clonogenic survival assay results of the A549 cell line, we determined if such results could be recapitulated in vivo (Figure 6B). Ganetespib alone has antitumor effects by causing tumor growth delay for nearly a week. CRT had an even stronger effect on preventing tumor regrowth, but when ganetespib was added to CRT it caused rapid tumor progression after an initial period of tumor control. Discussion In clinical radiotherapy, tumor radioresistance is one of the causes of local failure after radiotherapy. The development of drugs that can enhance the sensitivity of tumor cells to radiation is of great importance to improve the outcomes of lung cancer therapy. Although there are many studies that have focused on the development of radiosensitizers, the targeted agents that have been tested clinically, namely vascular targeting drugs and EGFR inhibitors, have so far not been clinically useful when combined with CRT for NSCLC. Unfortunately most of rationale for the combination has been based on preclinical studies using radiation alone. The best example was the EGFR targeting drugs. It has been well established in preclinical studies that EGFR inhibition sensitizes radiation for multiple tumor types, including head and neck cancer and NSCLC(29, 40). The approach seemed promising when a survival benefit was demonstrated for cetuximab when it was combined with radiotherapy compared to radiotherapy alone in head and neck cancer in a phase III randomized trial(32). Unfortunately, the benefit of cetuximab disappeared when it was combined with CRT, as shown in two large randomized trials for head and neck cancer (RTOG 0522 (41)) and unresectable NSCLC (RTOG 0617 (8)). This experience highlights the importance of critically evaluating radiosensitizers in preclinical models that at least modestly reflect the treatment regime used in the clinical setting. Several studies have shown that Hsp90 inhibitors can enhance radiation sensitivity of human cancer cell lines of different origin(14–25). These sensitizing effects are the result of the Hsp90 inhibitor-mediated abrogation of the G2 checkpoint, apoptosis and the inhibition of DNA repair. Studies have shown that one of the causes of sensitization could be inhibition of DNA double strand break (DSB) repair (17–21, 36). Checkpoint arrest mainly at G2/M phase has also been suggested as a cause of radiosensitization with Hsp90 inhibitors (19–24, 36). Radiosensitization effect in vivo by Hsp90 inhibitors has also been demonstrated (14, 21, 23). These data strongly suggest that targeting Hsp90 with its inhibitors represents a promising strategy for enhancing the sensitivity of cancer cells to radiation (19–24). Ganetespib is an investigational small molecule inhibitor of Hsp90 with favorable pharmacologic properties that distinguishes it from other first- and second-generation Hsp90 inhibitors in terms of potency, safety, and tolerability (27, 42). Ganetespib has also been shown to possess robust antitumor activity against a variety of cancer types in preclinical studies, including lung, breast, and prostate (38, 43–47). In addition, it has been shown in NSCLC cell lines that a synergistic combinatorial benefit was seen with the taxanes such as paclitaxel or docetaxel (43). As expected, we demonstrated that ganetespib significantly reduced clonogenic survival of various lung cancer cell lines, attenuated DNA damage repair, induced cell cycle arrest by the enhanced upregulation of negative cell cycle regulators (p21 and/or p27) and/or downregulation of positive regulators (cyclins D1 and E, CDK1, CDK2 and CDK4) that was greater than either treatment alone. Since CRT with carboplatin and paclitaxel is the standard treatment regime for unresectable NSCLC, adding ganetespib to standard CRT may have synergistic effects that need preclinical validation. Importantly, we have demonstrated that ganetespib has variable effects when combined with CRT in vitro and in two xenograft NSCLC models in vivo. Our in vitro clonogenic survival assay results showed that ganetespib alone enhances radiation effects in a panel of cell lines, but it demonstrated variable effects when combined with chemo or CRT. The xenograft models of H460 and A549 confirmed what we observed in vitro. H460 xenograft tumors gained additional benefits by adding ganetespib when treated with CRT, showing delayed tumor growth and increased DNA damage and apoptosis. On the other hand, A549 xenograft tumors was very well controlled by CRT alone but progressed rapidly when ganetespib was added. The xenograft results are consistent with what we saw in vitro, where there was also an increase in clonogenic cell survival when ganetespib was added to either chemotherapy or CRT. On the surface, the cell lines that do not respond to ganetespib when combined to CRT do not appear to be related to Kras or p53 mutation statuses. The mechanism for the variable effects of ganetespib in the various cell lines needs further investigation. Our findings imply that only a subgroup of lung cancer patients may benefit from HSP90 inhibition when receiving CRT. This may warrant the need for a predictable biomarker that could potentially identify patients to receive or avoid HSP90 inhibitors in combination with chemotherapy or CRT. It is important to emphasize the fact that drugs which sensitize cancer cells to radiation may not have the same effect when added with chemotherapy or CRT. This may explain the failure of previous trials using EGFR inhibitors in combination with CRT as well as the recent futility closure of the GALAXY-II docetaxel-ganetespib trial in lung cancer. Future development of the optimal drug and radiation combination needs to be tested in rigorous preclinical models within the context of clinically-relevant therapy combinations. Supplementary Material 1 2 Research Support: Funding was provided in part by The Mabuchi Program in Targeted Radiotherapy, United Against Lung Cancer, and The University of Texas MD Anderson Cancer Center and by the National Cancer Institute Cancer Center Support Grant CA016672. We thank Dr. Jared Burks, director of MD Anderson Cancer Center Flow Cytometry & Cellular Imaging Core Facility, for expert assistance with quantification of immunohistochemistry staining. SHL receive research funding support from STCube Pharmaceuticals, Roche/Genentech, and is the principle investigator of a clinical trial combining Ganetespib with chemoradiotherapy in esophageal cancer (NCT02389751) but does not receive financial support from Synta Pharmaceuticals, the manufacturer of Ganetespib. DAP was an employee of Synta Pharmaceuticals. S. H. Lin is funded in part through a research contract with STCube Pharmaceuticals, which does not manufacture or market any of the drugs discussed. D.A. Proia was director at Synta Pharmaceuticals Corp. Figure 1 Ganetespib is a potent radiation sensitizer in non-small cell lung cancer cell lines Clonogenic survival curves for H460 and A549 (Kras mutant, p53 WT) (A–B), H1650 (Kras WT, p53 mutant) (C) and H1299 (Kras WT and p53 null) (D), cells treated with or without 30 nM ganetespib for 5 hours prior to irradiation followed by an additional 19 hours of post-irradiation incubation in ganetespib containing medium. Figure 2 Ganetespib attenuate DNA damage repair in H460 and A549 cells assessed by p53BP1 foci formation Detection of p53BP1 foci formation was performed 30 mins, 4 hrs, 24 hrs and 48 hrs after irradiation (2 Gy) with or without 50 nM ganetespib pre-treatment prior to irradiation (A). Statistical data are presented as average foci number of 12 high power (20×) fields (B-C). Asterisks indicate p < 0.05. Figure 3 Ganetespib increases radiation induced G2M arrest H460 (A–B) and A549 (C–D) cells were treated with or without ganetespib (50 nM) for 5 hrs and then irradiated with 2 Gy. Cells were collected at each time point thereafter and analyzed by flow cytometry for the percentage of cells in G2/M. Drug treatment was continued after irradiation in the ganetespib-treated groups. Figure 4 Ganetespib altered expression level and activity of client growth factors and cell cycle progression proteins H460 and A549 cells were treated with ganetespib at 0 nM, 10 nM, 25 nM, 50 nM and 100 nM 5 hrs prior to irradiation (0 Gy, 2 Gy and 4 Gy) followed by an additional 19 hours of post-irradiation incubation in ganetespib containing medium. H460 and A549 cell lysates were collected for RPPA analysis (A–D). Heat map of RPPA results indicated altered activities or expression level of proteins (B and D). Figure 5 Ganetespib demonstrates variable effects with CRT in vitro Clonogenic survival assay results of H460, A549 (A–B) and H2087, H358 (C–D) tested with radiation alone, chemotherapy alone, or in combination with ganetespib. Figure 6 Ganetespib shows differential effects with CRT in H460 and A549 xenografts Ncr nude mice were injected with 1 × 106 H460 cells or 5 × 106 A549 subcutaneously in the right leg. When tumor volume reached around 100 mm3 for H460 or 150 mm3 for A549, mice bearing established xenograft (n=4–10/group) were exposed to radiation at 2 Gy daily for 5 days, with or without i.p. 30 mg/kg Carboplatin + 10 mg/kg Paclitaxel 1×/week for 2 weeks and combined with i.v. 100 mg/kg ganetespib 1×/week for 2 weeks, either alone or in combination. Tumor volumes are indicated as average of each treatment group and the error bars are the SEM. Translational Relevance The current study shows that ganetespib, a selective and highly potent Hsp90 inhibitor, inhibits clonogenic survival in NSCLC cells, and synergizes the efficacy of irradiation. This provided the motivation to explore the combination of ganetespib with CRT to improve the clinical outcomes of unresectable NSCLC. When it was tested with CRT, ganetespib was able to sensitize some cell lines, but in others it had either no radiation sensitization effect or even abrogated the CRT effect. Our data cautions the assumption that drugs that sensitize to radiation would also sensitize to CRT. Rigorous and systematic preclinical testing of drugs must be done in the clinical context in which the disease is being treated. Identifying ways to circumvent this variable effect or predictive biomarkers will be needed for optimal translation of radiation sensitizing drugs to the clinical setting. Disclosure of Potential Conflicts of Interest No potential conflicts of interest were disclosed by the other authors. Authors' Contributions Conception and design: S.H. Lin, Y. Wang, H. Liu Development of methodology: H. Liu, Y. Wang, A. Potter, J. Zhang, Y. Qiao Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): H. Liu, Y. Wang, A. Potter, J. Zhang, Y. Qiao Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): H. Liu, L. Diao, J. Wang Writing, review, and/or revision of the manuscript: S.H. Lin, Y. Wang, H. Liu, D.A. Proia, Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): S.H. Lin, H. Liu, Y. Wang, J. Zhang, Y. Qiao, D.A. Proia Study supervision: S.H. Lin References 1 Jemal A Siegel R Ward E Hao Y Xu J Thun MJ Cancer statistics, 2009 CA Cancer J Clin 2009 59 225 249 19474385 2 Chemotherapy in non-small cell lung cancer: a meta-analysis using updated data on individual patients from 52 randomised clinical trials. Non-small Cell Lung Cancer Collaborative Group BMJ 1995 311 899 909 7580546 3 Auperin A Le Pechoux C Rolland E Curran WJ Furuse K Fournel P Meta-analysis of concomitant versus sequential radiochemotherapy in locally advanced non-small-cell lung cancer J Clin Oncol 2010 28 2181 2190 20351327 4 Curran WJ Jr Paulus R Langer CJ Komaki R Lee JS Hauser S Sequential vs. concurrent chemoradiation for stage III non-small cell lung cancer: randomized phase III trial RTOG 9410 J Natl Cancer Inst 2011 103 1452 1460 21903745 5 Komaki R Scott CB Sause WT Johnson DH Taylor SGt Lee JS Induction cisplatin/vinblastine and irradiation vs. irradiation in unresectable squamous cell lung cancer: failure patterns by cell type in RTOG 88-08/ECOG 4588. Radiation Therapy Oncology Group. Eastern Cooperative Oncology Group Int J Radiat Oncol Biol Phys 1997 39 537 544 9336129 6 Rowell NP O'Rourke NP Concurrent chemoradiotherapy in non-small cell lung cancer Cochrane Database Syst Rev 2004 CD002140 7 Sandler A Gray R Perry MC Brahmer J Schiller JH Dowlati A Paclitaxel-carboplatin alone or with bevacizumab for non-small-cell lung cancer N Engl J Med 2006 355 2542 2550 17167137 8 Bradley JD Paulus R Komaki R Masters G Blumenschein G Schild S Standard-dose versus high-dose conformal radiotherapy with concurrent and consolidation carboplatin plus paclitaxel with or without cetuximab for patients with stage IIIA or IIIB non-small-cell lung cancer (RTOG 0617): a randomised, two-by-two factorial phase 3 study The Lancet Oncol 2015 16 187 199 25601342 9 Garcia-Carbonero R Carnero A Paz-Ares L Inhibition of HSP90 molecular chaperones: moving into the clinic The Lancet 2013 14 e358 e369 23896275 10 Trepel J Mollapour M Giaccone G Neckers L Targeting the dynamic HSP90 complex in cancer Nat Rev Cancer 2010 10 537 549 20651736 11 Patel HJ Modi S Chiosis G Taldone T Advances in the discovery and development of heat-shock protein 90 inhibitors for cancer treatment Expert Opin Drug Discov 2011 6 559 587 22400044 12 Neckers L Workman P Hsp90 molecular chaperone inhibitors: are we there yet? Clin Cancer Res 2012 18 64 76 22215907 13 Kamal A Thao L Sensintaffar J Zhang L Boehm MF Fritz LC A high-affinity conformation of Hsp90 confers tumour selectivity on Hsp90 inhibitors Nature 2003 425 407 410 14508491 14 Bisht KS Bradbury CM Mattson D Kaushal A Sowers A Markovina S Geldanamycin and 17-allylamino-17-demethoxygeldanamycin potentiate the in vitro and in vivo radiation response of cervical tumor cells via the heat shock protein 90-mediated intracellular signaling and cytotoxicity Cancer Res 2003 63 8984 8995 14695217 15 Russell JS Burgan W Oswald KA Camphausen K Tofilon PJ Enhanced cell killing induced by the combination of radiation and the heat shock protein 90 inhibitor 17-allylamino-17-demethoxygeldanamycin: a multitarget approach to radiosensitization Clin Cancer Res 2003 9 3749 3755 14506167 16 Machida H Nakajima S Shikano N Nishio J Okada S Asayama M Heat shock protein 90 inhibitor 17-allylamino-17-demethoxygeldanamycin potentiates the radiation response of tumor cells grown as monolayer cultures and spheroids by inducing apoptosis Cancer Sci 2005 96 911 917 16367912 17 Noguchi M Yu D Hirayama R Ninomiya Y Sekine E Kubota N Inhibition of homologous recombination repair in irradiated tumor cells pretreated with Hsp90 inhibitor 17-allylamino-17-demethoxygeldanamycin Biochem Biophys Res Commun 2006 351 658 663 17083915 18 Koll TT Feis SS Wright MH Teniola MM Richardson MM Robles AI HSP90 inhibitor, DMAG, synergizes with radiation of lung cancer cells by interfering with base excision and ATM-mediated DNA repair Mol Cancer Ther 2008 7 1985 1992 18645008 19 Stingl L Stuhmer T Chatterjee M Jensen MR Flentje M Djuzenova CS Novel HSP90 inhibitors, NVP-AUY922 and NVP-BEP800, radiosensitise tumour cells through cell-cycle impairment, increased DNA damage and repair protraction Br J Cancer 2010 102 1578 1591 20502461 20 Ha K Fiskus W Rao R Balusu R Venkannagari S Nalabothula NR Hsp90 inhibitor-mediated disruption of chaperone association of ATR with hsp90 sensitizes cancer cells to DNA damage Mol Cancer Ther 2011 10 1194 1206 21566061 21 Zaidi S McLaughlin M Bhide SA Eccles SA Workman P Nutting CM The HSP90 inhibitor NVP-AUY922 radiosensitizes by abrogation of homologous recombination resulting in mitotic entry with unresolved DNA damage PLoS One 2012 7 e35436 22523597 22 Djuzenova CS Blassl C Roloff K Kuger S Katzer A Niewidok N Hsp90 inhibitor NVP-AUY922 enhances radiation sensitivity of tumor cell lines under hypoxia Cancer Biol Ther 2012 13 425 434 22286776 23 Gandhi N Wild AT Chettiar ST Aziz K Kato Y Gajula RP Novel Hsp90 inhibitor NVP-AUY922 radiosensitizes prostate cancer cells Cancer Biol Ther 2013 14 347 356 23358469 24 Milanovic D Firat E Grosu AL Niedermann G Increased radiosensitivity and radiothermosensitivity of human pancreatic MIA PaCa-2 and U251 glioblastoma cell lines treated with the novel Hsp90 inhibitor NVP-HSP990 Radiat Oncol 2013 8 42 23448094 25 Enomoto A Fukasawa T Takamatsu N Ito M Morita A Hosoi Y The HSP90 inhibitor 17-allylamino-17-demethoxygeldanamycin modulates radiosensitivity by downregulating serine/threonine kinase 38 via Sp1 inhibition Eur J Cancer 2013 49 3547 3558 23886587 26 Supko JG Hickman RL Grever MR Malspeis L Preclinical pharmacologic evaluation of geldanamycin as an antitumor agent Cancer Chemother Pharmacol 1995 36 305 315 7628050 27 Zhou D Liu Y Ye J Ying W Ogawa LS Inoue T A rat retinal damage model predicts for potential clinical visual disturbances induced by Hsp90 inhibitors Toxicol Appl Pharmacol 2013 273 401 409 24090817 28 Ramalingam S Goss G Rosell R Schmid-Bindert G Zaric B Andric Z A randomized phase II study of ganetespib, a heat shock protein 90 inhibitor, in combination with docetaxel in second-line therapy of advanced non-small cell lung cancer (GALAXY-1) Ann Oncol 2015 26 1741 1748 25997818 29 Milas L Fan Z Andratschke NH Ang KK Epidermal growth factor receptor and tumor response to radiation: in vivo preclinical studies Int J Radiat Oncol Biol Phys 2004 58 966 971 14967457 30 Nasu S Ang KK Fan Z Milas L C225 antiepidermal growth factor receptor antibody enhances tumor radiocurability Int J Radiat Oncol Biol Phys 2001 51 474 477 11567823 31 Crosby T Hurt CN Falk S Gollins S Mukherjee S Staffurth J Chemoradiotherapy with or without cetuximab in patients with oesophageal cancer (SCOPE1): a multicentre, phase 2/3 randomised trial The Lancet Oncol 2013 14 627 637 23623280 32 Ang KK Zhang Q Rosenthal DI Nguyen-Tan PF Sherman EJ Weber RS Randomized phase III trial of concurrent accelerated radiation plus cisplatin with or without cetuximab for stage III to IV head and neck carcinoma: RTOG 0522 J Clin Oncol 2014 32 2940 2950 25154822 33 Gomez-Casal R Bhattacharya C Epperly MW Basse PH Wang H Wang X The HSP90 Inhibitor Ganetespib Radiosensitizes Human Lung Adenocarcinoma Cells Cancers 2015 7 876 907 26010604 34 Tibes R Qiu Y Lu Y Hennessy B Andreeff M Mills GB Reverse phase protein array: validation of a novel proteomic technology and utility for analysis of primary leukemia specimens and hematopoietic stem cells Mol Cancer Ther 2006 5 2512 2521 17041095 35 Begg AC Stewart FA Vens C Strategies to improve radiotherapy with targeted drugs Nature reviews Cancer 2011 11 239 253 21430696 36 Niewidok N Wack LJ Schiessl S Stingl L Katzer A Polat B Hsp90 Inhibitors NVP-AUY922 and NVP-BEP800 May Exert a Significant Radiosensitization on Tumor Cells along with a Cell Type-Specific Cytotoxicity Transl Oncol 2012 5 356 369 23066444 37 Kuerbitz SJ Plunkett BS Walsh WV Kastan MB Wild-type p53 is a cell cycle checkpoint determinant following irradiation Proc Natl Acad Sci U S A 1992 89 7491 7495 1323840 38 He S Zhang C Shafi AA Sequeira M Acquaviva J Friedland JC Potent activity of the Hsp90 inhibitor ganetespib in prostate cancer cells irrespective of androgen receptor status or variant receptor expression Int J Oncol 2013 42 35 43 23152004 39 Zhan M Han ZC Phosphatidylinositide 3-kinase/AKT in radiation responses Histol Histopathol 2004 19 915 923 15168354 40 Raben D Helfrich B Bunn PA Jr Targeted therapies for non-small-cell lung cancer: biology, rationale, and preclinical results from a radiation oncology perspective Int J Radiat Oncol Biol Phys 2004 59 27 38 15142632 41 Adkins D Ley J Wildes TM Michel L RTOG 0522: huge Investment in patients and resources and no benefit with addition of cetuximab to radiotherapy--why did this occur? J Clin Oncol 2015 33 1223 1224 25713440 42 Ying W Du Z Sun L Foley KP Proia DA Blackman RK Ganetespib, a unique triazolone-containing Hsp90 inhibitor, exhibits potent antitumor activity and a superior safety profile for cancer therapy Mol Cancer Ther 2012 11 475 484 22144665 43 Proia DA Sang J He S Smith DL Sequeira M Zhang C Synergistic activity of the Hsp90 inhibitor ganetespib with taxanes in non-small cell lung cancer models Invest New Drugs 2012 30 2201 2209 22227828 44 Acquaviva J Smith DL Sang J Friedland JC He S Sequeira M Targeting KRAS-mutant non-small cell lung cancer with the Hsp90 inhibitor ganetespib Mol Cancer Ther 2012 11 2633 2643 23012248 45 Shimamura T Perera SA Foley KP Sang J Rodig SJ Inoue T Ganetespib (STA-9090), a nongeldanamycin HSP90 inhibitor, has potent antitumor activity in in vitro and in vivo models of non-small cell lung cancer Clin Cancer Res 2012 18 4973 4985 22806877 46 Friedland JC Smith DL Sang J Acquaviva J He S Zhang C Targeted inhibition of Hsp90 by ganetespib is effective across a broad spectrum of breast cancer subtypes Invest New Drugs 2014 32 14 24 23686707 47 Proia DA Zhang C Sequeira M Jimenez JP He S Spector N Preclinical activity profile and therapeutic efficacy of the HSP90 inhibitor ganetespib in triple-negative breast cancer Clin Cancer Res 2014 20 413 424 24173541

PMC005xxxxxx/PMC5135586.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 0413066 2830 Cell Cell Cell 0092-8674 1097-4172 27839864 5135586 10.1016/j.cell.2016.10.023 NIHMS824113 Article FMN2 makes perinuclear actin to protect nuclei during confined migration and promote metastasis Skau Colleen T. 1 Fischer Robert S. 1 Gurel Pinar 1 Racine-Thiam Hawa 12 Tubbs Anthony 3 Baird Michelle A. 14 Davidson Michael W. 4 Piel Matthieu 2 Alushin Gregory M. 1 Nussenzweig Andre 3 Steeg Patricia S. 5 Waterman Clare M. 1* 1 Cell Biology and Physiology Center, National Heart Lung and Blood Institute, National Institutes of Health, Bethesda, MD 20892, USA 2 Institut Curie, CNRS UMR 144, 26 rue d’Ulm, 75005 Paris, France 3 Laboratory of Genome Integrity, National Cancer Institute, National Institutes of Health, Bethesda, MD 20892, USA 4 Magnet Lab, Florida State University, Tallahassee FL, USA 5 Women’s Malignancies Branch, Center for Cancer Research, National Cancer Institute, National Institutes of Health, Bethesda, MD 20892, USA * Lead Contact and Correspondence to: Clare M. Waterman Ph.D., Cell Biology and Physiology Center, National Heart, Lung and Blood Institute, National Institutes of Health, Building 50 South Drive, Room 4537 MSC 8019, Bethesda Maryland 20892-8019, T: (301)-435-2949, F: (301)-480-6012, watermancm@nhlbi.nih.gov Admin: Schwanna Thacker schwanna.thacker@nih.gov 20 10 2016 10 11 2016 01 12 2016 01 12 2017 167 6 15711585.e18 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. Summary Cell migration in confined 3D tissue microenvironments is critical for both normal physiological functions and dissemination of tumor cells. We discovered a cytoskeletal structure that prevents damage to the nucleus during migration in confined microenvironments. The formin-family actin filament nucleator FMN2 associates with and generates a perinuclear actin/focal adhesion (FA) system that is distinct from previously characterized actin/FA structures. This system controls nuclear shape and positioning in cells migrating on 2D surfaces. In confined 3D microenvironments, FMN2 promotes cell survival by limiting nuclear envelope damage and DNA double-strand breaks. We found that FMN2 is upregulated in human melanomas, and show that disruption of FMN2 in mouse melanoma cells inhibits their extravasation and metastasis to the lung. Our results indicate a critical role for FMN2 in generating a perinuclear actin/FA system that protects the nucleus and DNA from damage to promote cell survival during confined migration, and thus promote cancer metastasis. eTOC A Perinuclear actin-based armor protects the nucleus and its contents from damage when cells need to migrate through tiny spaces. Introduction Cell migration is critical for angiogenesis, leukocyte extravasation and tissue surveillance, connective tissue maintenance by fibroblasts, as well as pathological conditions such as metastatic cancer. In these contexts, cells must migrate through complex and tightly confined 3D microenvironments within tissues. To invade dense tissues, cells can remodel their microenvironment via proteinases that degrade the extracellular matrix (ECM)(Chang & Werb 2001), but can also squeeze through constrictions much smaller than their diameter without degrading the ECM(Wolf et al. 2007). Spatial regulation of actin filament assembly can mediate leading edge protrusion through small ECM pores(Ridley et al. 2003); however squeezing the large and stiff nucleus through constrictions requires extensive nuclear deformation(Ridley et al. 2003; Friedl et al. 2011; Thiam et al. 2016). The role of nuclear mechanics, shape, and integrity in confined cell migration and how the cytoskeleton participates in these processes has begun to be elucidated. The nucleus and DNA can be damaged by mechanical stress(Kumar et al. 2014; Hatch et al. 2013; Zhang et al. 2015; Raab et al. 2016), suggesting that cells have nucleo-protective mechanisms that operate during confined migration. The lamin intermediate filament network that underlies the nuclear envelope (NE) provides mechanical stability to the nucleus (Broers 2004). In immune and cancer cells migrating in tight confines, the NE experiences transient tears at sites of discontinuity in the lamina and is associated with DNA damage(Raab et al. 2016; Denais et al. 2016). Lamins connect to the cytoplasmic cytoskeleton via “LINC complexes” that traverse the NE and interact with the actomyosin, microtubule, or intermediate filament cytoskeletons(Lombardi & Lammerding 2011). These cytoskeletal-nuclear connections mediate nuclear positioning during migration and force transmission from the outside of the cell to the nucleus (Maniotis et al. 1997; Lombardi et al. 2011). The cytoskeleton plays other important roles in nuclear function(Belin et al. 2015; Baarlink et al. 2013; Lottersberger et al. 2015; Schramek et al. 2014). Actin polymerization on the cytoplasmic face of the NE drives nuclear deformation to help squeeze the nucleus through pores in the microenvironment(Thiam et al. 2016). Actin filaments inside the nucleus and cytoplasmic microtubules are both involved in DNA double-strand break (DSB) repair(Belin et al. 2015; Baarlink et al. 2013; Lottersberger et al. 2015), while myosin 2 regulates nuclear retention of p53(Schramek et al. 2014). Furthermore, altered actin structures around the nucleus have been described in cancer cells(Revach et al. 2015; Koshkina et al. 2013) and changes in nuclear deformability are linked to cancer progression(Davidson et al. 2014). Thus, cytoskeletal/nuclear interactions are critical to nuclear function, implicating nuclear mechanics and nucleus-cytoskeleton connections in disease. Cell migration is mediated by a diverse array of actin-based structures that generate protrusive lamellipodia and filopodia as well as contracting networks and bundles that maintain cell shape, drive FA turnover, and promote cell body and nuclear movement. Different actin structures are built by different actin nucleation factors that are regulated in time and space(Skau & Waterman 2015). The Arp2/3 complex, which generates short networks of branched filaments builds lamellipodia at the cell leading edge(Svitkina 1999) and drives nuclear deformation in confined migration(Thiam et al. 2016). Formins are a family of actin nucleating proteins encoded by fifteen different genes in humans, and which generate unbranched actin filaments(Schonichen & Geyer 2010). Specific formins mediate formation of contractile stress fibers (SF), transmembrane actin-associated (TAN) lines on the dorsal cell surface that position the nucleus, filopodia at the leading edge, isotropic cortical actin networks, and actin important for mitochondrial fission(Skau & Waterman 2015; Campellone & Welch 2010). However, the cellular functions of many formins remain poorly characterized. Here, we explored the role of a focal adhesion (FA)-associated formin(Kuo et al. 2011), FMN2, in FA function and cell migration in 2D and 3D. FMN2 is highly expressed in oocytes and the nervous system, and in human and mouse, mutations and deletions are associated with infertility and intellectual disability(Leader & Leder 2000; Leader et al. 2002; Law et al. 2014; Almuqbil et al. 2013), as well as several cancers(Leader & Leder 2000; Charfi et al. 2011; Liu et al. 2012; Araujo et al. 2014; Gruel et al. 2014; Lynch et al. 2013). In oocytes, FMN2 mediates formation of an actin mesh that positions the spindle during oogenesis(Leader et al. 2002; Montaville et al. 2014) and generates actin filaments in the nucleus(Belin et al. 2015). However, nothing is known about the role of FMN2 in FAs or cell migration. We find that FMN2 generates a perinuclear actin/FA system that regulates nuclear position and shape during migration on planar substrates, and in confined 3D migration is required for cell survival. This is due to a unique role of FMN2 in preventing NE rupture and DNA double-strand breaks (DSB) that occur during confined migration. Furthermore, we find a striking role for FMN2 in promoting metastasis of melanoma cells to the lung. Results FMN2 associates with a perinuclear membrane, actin and FA system that moves with the nucleus during cell migration We sought to characterize the role of the formin FMN2 in FA function and cell migration. Immunolocalization analysis of primary mouse embryonic fibroblasts (MEF) plated on fibronectin (FN)-coated coverslips and spinning disk confocal (SDC) microscopy showed FMN2 localization to fibrils underneath the nucleus that partially co-localized with a subset of perinuclear actin bundles and FA (Figure 1A). 3D reconstructions of SDC Z-stacks showed that FMN2 localized to the ventral cell surface and foci in the nucleus(Belin et al. 2015), but was absent from the dorsal cell surface (Figure 1B). Expression of C-terminally tagged FMN2-GFP (Figures 1C–G, S1) and imaging with total internal reflection fluorescence (TIRF) microscopy (Figure 1C) confirmed that FMN2 was concentrated at the ventral cell surface in MEFs, and was localized similarly in human endothelial and intestinal adenocarcinoma cells, canine kidney epithelial cells and mouse melanoma cells (Figures S1, 7). Immunostaining of FMN2 and microtubules, vimentin intermediate filaments, or the FA formin INF2 revealed no substantial co-localization (Figure S1). Furthermore, inhibition of myosin-2 or integrin engagement failed to disrupt perinuclear localization of FMN2, although it became less fibrillar and more punctate (Figures 1D, S1). Depolymerization of actin caused FMN2 to form rings, which together with the fact that FMN2 contains a consensus myristoylation sequence, suggested that FMN2 may associate with membrane vesicles (Figure 1E). Indeed, detergent extraction induced loss of FMN2-GFP (Figure 1F). Thus, FMN2 localizes to a membrane system that is associated with perinuclear actin bundles and FA at the dorsal cell surface. We then examined the composition and dynamics of the perinuclear actin/FA system (Figure S2). Immunolocalization showed that similar to leading edge SF and FA that mediate cell migration and ECM remodeling(Hotulainen & Lappalainen 2006), perinuclear actin bundles and FAs contained zyxin, tropomyosin, paxillin, Hic5, vinculin, ILK, FAK, talin, phospho-p130Cas, phospho-Src and phospho-FAK (Figure S2G–H). However, in contrast with leading edge SF and FA, perinuclear SF and FA co-localized with non-muscle myosin 2B (M2B), lacked α-actinin and VASP (Figure S2A) and were incapable of remodeling fluorescent FN (Figure S2D). Furthermore, while leading edge SF terminated in FA at one end, perinuclear actin bundles had FA at both ends (Figure S2B). Although the nuclear envelope (NE) protein SUN2 that associates with TAN lines on the dorsal cell surface associated with ventral perinuclear actin bundles (Figure S1F), over-expression of a dominant-negative KASH construct(Lombardi et al. 2011) that disrupts TAN lines did not affect perinuclear actin bundles (not shown). Thus, FMN2 associates with a novel perinuclear actin/FA system that is compositionally distinct from other cellular actin/FA systems. We then characterized the dynamics of FMN2 and the perinuclear actin/FA system. Time-lapse SDC of MEF co-expressing FMN2-GFP, mCherry-α-actinin (FA and lamellipodia marker) and mCardinal-H2B (nuclear marker) showed that FMN2 fibrils localized to and moved with the trailing half of the nucleus during cell migration (Figure 1G; Movie S1). Imaging of myosin 2 light chain-GFP and the nucleus showed that perinuclear actin bundles impinged on and deformed the nucleus (Figure 1H; Movie S2). Photobleaching GFP-actin and kymograph analysis revealed that bleached marks near FA-anchored ends of a perinuclear actin bundle moved toward the bundle center (Figure S2C), indicating polarized assembly of actin at FA and bundle contraction. TIRF microscopy of paxillin-GFP (FA marker) and the nucleus showed that perinuclear FA appeared in the cell center and elongated in the direction of nuclear movement (Figure S2F; Movie S3). Quantitative analysis showed that compared to leading edge FA, a greater fraction of perinuclear FA elongated and they persisted longer (Figure S2E). Thus, FMN2 localizes to membranes associated with a compositionally and dynamically unique contractile perinuclear actin-FA system that assembles and moves in a coordinated fashion with the nucleus during cell migration. FMN2 is required for the formation and maintenance of the perinuclear actin/FA system to control nuclear shape and position in MEF migrating on planar ECMs We then probed the role of FMN2 in the perinuclear actin/FA system and cell function by siRNA-mediated knockdown (FMN2-KD, 70% reduction by siRNA pool or 3′ untranslated region targeting). This resulted in near absence of FMN2 immunofluorescence signal in individual transfected cells (Figure S3A). FMN2-KD cells exhibited a striking loss of the perinuclear actin/FA system and concomitant redistribution of M2B to the cell periphery (Figure 2A–C). However FMN2-KD had no effect on cell area, leading edge SF and FA, fibronectin bundles under the cell, SUN2-marked TAN lines, or microtubules (Figure 2A–D; S3; Movie S4). Re-expression of FMN2-GFP in FMN2-KD cells rescued perinuclear actin bundles and FA and M2B localization (Figure 2A–C). Therefore, FMN2 is specifically required for generating and maintaining the perinuclear actin-FA system. We next examined the role of FMN2 and the perinuclear actin-FA system in nucleus morphometry. Analysis of cell and nuclear shape showed that the aspect ratios of the nucleus and the cell were approximately equal in mock-transfected controls(Versaevel et al. 2012), such that degree of cell extension was mirrored by the ellipticity of the nucleus (Figure 3A–B). In contrast, in FMN2-KD cells this correlation decreased, and there was an increase in nuclear lobularity (Figure 3B). Immunostaining lamin A/C or lamin B or live imaging of mCherry-lamin B (Figure 3D, E; S4A) in FMN2-KD cells showed that in spite of the nuclear lobularity, nuclear lamina remained intact. FMN2-KD had no effect compared to control on the number of bi-nucleate cells, cell migration velocity or wound closure (not shown; Figure S4). However, in control cells, the nucleus centroid tracked with the cell centroid during random migration and was located to the rear of the centrosome during directional wound-healing(Gomes et al. 2005), while in FMN2-KD cells, the nucleus drifted from the cell center over time and the centrosome was positioned randomly (Figures 3C, S4). Alternating between wide-field epi-fluorescence (Epi) and TIRF imaging of the nucleus showed that control cells maintained closer apposition of the nucleus with the substrate (as indicated by the Epi:TIRF intensity ratio) than did FMN2-KD cells (Figure 3F–G). All defects in nuclear parameters were rescued by re-expression of FMN2-GFP in FMN2-KD cells (Figure 3). Thus, FMN2 maintains nuclear shape and position during cell migration on planar substrates. FMN2 is required for the perinuclear actin FA system in cells in 3D collagen ECMs We then sought to determine the role of FMN2 and the perinuclear actin-FA system in cells migrating in a 3D microenvironment. 3D immunolocalization analysis of MEF cultured in collagen gels showed that FMN2 localized to the cortex around the perimeter of one half of the nucleus, forming a cup-like structure that extended into one pole of the cell (Figure 4A). Time-lapse imaging of FMN2-GFP, actin and nuclei during 3D migration showed that the cup of FMN2-GFP fibrils moved with the rear of the nucleus as the cell crawled and changed direction (Figure 4B; Movie S5). We then assessed the effects of loss of FMN2 on MEF in 3D ECMs. While control cells were spindle-shaped and exhibited actin bundles with associated FA running across the nucleus along the long axis of the cell cortex, FMN2-KD MEF were less polarized with projections distributed around the cell periphery and displayed an isotropic actin mesh that lacked FAs in the perinuclear region (Figure 4C–D; S5). Re-expression of FMN2-GFP rescued these effects (Figure 4D, S5). Thus, FMN2 generates a perinuclear actin/FA system in both 2D and 3D microenvironments. FMN2 protects against NE damage to promote cell survival during migration in confined microenvironments In the course of our investigation of the effects of FMN2-KD on actin and FA organization in MEFs in 3D microenvironments, we noted that depletion of FMN2 reduced survival of cells migrating in collagen gels in a concentration dependent manner, but had no effect on their viability on 2D ECMs (Figure 5A, S6A,B). This suggested that their susceptibility to death may be due to migration through physically confining microenvironments. To test this, we employed transwell assays in which cells chemotax through an ECM-coated membrane containing 8 or 12 μM pores (Figure 5B). Collecting, plating and imaging cells after they passed through the membrane showed that compared to GFP-H2B-expressing controls, fewer FMN2-KD cells transited through the 12 μM pore-size membrane, and this difference was enhanced by reducing pore size to 8 μM (Figure 5B). The invasion defect induced by depletion of FMN2 was rescued by re-expression of FMN2-GFP. Together, these results suggest that FMN2 promotes cell survival during migration in confining 3D microenvironments. We hypothesized that the perinuclear actin and FA system generated by FMN2 could provide protection to the nucleus during confined migration. To test this, we imaged mEmerald tagged to a nuclear localization sequence (NLS-mEmerald) as a marker of NE integrity in cells migrating through PDMS micro-channels containing narrow constrictions (Figure 5C). Control cells entered channels and traversed multiple constrictions, exhibiting only transient leaks of NLS-mEmerald into the cytoplasm(Denais et al. 2016; Raab et al. 2016) (Figure 5C; Movie S6). Conversely, most FMN2-KD MEF exhibited catastrophic and irreversible release of NLS-mEmerald in the wide portion of the channel, with only 1.4% surviving traversal of a constriction (Figure 5C). In contrast, in cells cultured on coverslips, there was no difference in the nuclear:cytoplasmic ratio of NLS-mEmerald between control and FMN2-KD (Figure S6C–D). These data indicate that FMN2 protects the NE from damage to promote cell survival during migration and invasion through tightly confined microenvironments. FMN2 protects cells migrating in confined microenvironments from DNA double-strand breaks (DSBs) that are detected in an ATM-dependent manner We next examined the effects of FMN2-KD on DNA damage after migration in confinement(Raab et al. 2016; Denais et al. 2016) (Figure 6). We immunostained cells plated on 2D ECMs or collected after migration through transwells for phosphorylated histone H2AX (γH2AX )(Rogakou et al. 1998) and 53BP1 as focal markers of DNA DSB repair(Schultz et al. 2000; Lukas et al. 2011) (Figure 6A–B). This showed that compared to control, FMN2-KD cells had a striking increase in DNA repair sites (Figure 6B), with smaller pores in the transwells eliciting greater DNA damage (Figure 6A–B). We also varied the porosity of collagen gels by changing the temperature of collagen polymerization (higher temperature = smaller pores)(Raub et al. 2007). After cells penetrated the collagen for 20h there was a porosity-dependent increase in DNA damage sites in FMN2-KD cells, but not for controls (Figure 6C–D). To determine the signaling pathway responsible for detecting and repairing the DNA damage(Awasthi et al. 2015), we treated cells with ATM and ATR-specific kinase inhibitors during invasion through 8μm transwells. This showed that when ATM, but not ATR was inhibited, γH2AX foci were strongly reduced in both control and FMN2-KD cells (Figure 6I). The increases in γH2AX and 53BP1 staining induced by loss of FMN2 and invasion through transwells or migration in collagen ECMs were rescued by re-expression of full-length FMN2-GFP (Figure 6A–D). Importantly, no DNA repair sites were detected in either control or FMN2-KD cells cultured on 2D ECMs, and there was no difference in the time course of resolution of the repair sites after exposure to 7500 μJ/cm2 of radiation (Figure S6D,E). These results show that FMN2 protects cells migrating in confined microenvironments from DNA DSBs that are detected in an ATM-dependent manner. FMN2-mediated actin polymerization is required for protection against DNA damage We hypothesized that the perinuclear actin-FA system generated by FMN2 may protect the nucleus from damage during migration in confinement. We first tested this by reducing actin bundles by low-dose treatment with the myosin 2 inhibitor blebbistatin (20μM). This abolished most SFs, including the perinuclear actin system (Figure 6F), yet had no effect on the localization of FMN2 (Figure 1D), peripheral actin arcs (Figure 6F), or migration in transwells (not shown). Blebbistatin treatment significantly increased the proportion of mock-transfected cells with γH2AX foci after passage through an 8μm pore transwell, although not to the same extent as that induced by FMN2-KD (Figure 6F–G). In contrast, blebbistatin did not enhance the effects of FMN2 knockdown on the level of γH2AX foci (Figure 6G). This shows that full actomyosin contractility, and possibly the SFs that it generates, are required to protect cells from DNA damage during confined migration. We next used an FMN2 point mutant deficient in actin polymerization to directly test the requirement for FMN2-generated actin filaments in protection from DNA damage during confined migration. We mutated a conserved isoleucine in the FH2 domain of FMN2 to an alanine (FMN2-I1226A), which inhibits actin polymerization activity in other formin family proteins(Ramabhadran, Pinar S Gurel, et al. 2012). We verified that compared to the wild-type FH2 domain of FMN2, the mutant was deficient in enhancement of actin filament assembly in vitro (Figure 6H). In FMN2-KD cells in both 2D and 3D ECMs, FMN2-I1226A-GFP localized to the perinuclear region, but formed punctae rather than extended fibrils (Figure 6I; Movie S7), and failed to rescue both the loss of the perinuclear actin/FA system in 2D culture and the increase in γH2AX foci induced by migration through 8μm transwells or in 3D collagen ECMs (Figure 6I). Therefore, the actin assembly activity of FMN2 and the perinuclear actin/FA system it creates are critical for protection of cells from DNA damage during migration through confined microenvironments. FMN2 promotes metastasis of murine melanoma Given the requirement for FMN2 for cell survival during confined migration, we hypothesized that FMN2 may be critical for invasive migration of cancer cells in metastasis. We found in public databases (https://www.oncomine.org/) that FMN2 is highly upregulated in melanoma, an extremely invasive cancer that generally metastasizes to lung and/or brain. We thus examined the requirement for FMN2 in metastasis of B16-F10 melanoma cells, (Fidler 1973) which are highly invasive and show high FMN2 expression compared to primary mouse melanocytes (Figure S7A). We generated B16-F10 cell lines with CRISPR-mediated deletion of FMN2 (FMN2-KO), with or without FMN2-GFP rescue (Figure S7B), and validated that FMN2 localization and its role in forming the perinuclear actin/FA system and protection from DNA damage and promotion of survival in a transwell invasion assay was similar in B16-F10 cells (Figures 7A,B; S4C; S7A,E,F) as that documented in MEFs (Figures 1A,C; 2A; 5B; 6A; S4C). We tested the requirement for FMN2 in melanoma metastasis in mice in vivo using a tail-vein injection model to determine if cells were capable of surviving the circulatory system, extravasating, and forming metastases in the lung. Compared to wild-type B16-F10 cells (WT), cells with stable overexpression of FMN2 (FMN2-OE) induced a 1.6-fold increase in lung surface metastases (Figures 7C, D, S7F), and metastases on the heart, diaphragm, thymus and lining of the thoracic cavity (not shown), and increased lung weight by two-fold. Strikingly, compared to WT, FMN2-KO cells exhibited a 91% reduction in in lung surface metastases (Figures 7C, D, S7F) and a complete absence of metastases in other organs as well as a corresponding decrease in lung weight (not shown). Histological analysis showed both circumscribed and invasive lesions in lungs bearing WT and FMN2-OE cells, with only infrequent micrometastases induced by FMN2-KO cells (Figure 7E, F, S7G). However, these micrometastases had a more invasive (less compact) morphology compared large metastases induced by WT cells (Figure S7H). Immunostaining of lung sections for γH2AX revealed a significant decrease in cell nuclei with DNA DSB sites within metastases induced by WT compared to FMN2-OE cells (Figure 7G, H). However, significantly more nuclei displayed γH2AX foci in the few metastases induced by FMN2-KO cells as compared to either WT or FMN2-OE (Figure 7G, H) suggesting cells lacking FMN2 acquire other mechanisms for surviving high levels of DNA damage. Together, these results indicate that FMN2 promotes formation of a perinuclear actin/FA system and protects melanoma cells from DNA damage during confined migration in vitro, and is critical to their ability to survive the circulatory system, extravasate, and form lung metastases with limited DNA damage in vivo. Discussion We show for the first time the role of a formin actin nucleation factor, FMN2, in generating a novel contractile perinuclear actin/FA system that protects the nucleus and DNA from mechanically induced damage during confined migration in 3D microenvironments, and that promotes melanoma metastasis. We find that FMN2 is essential for nuclear shape and position maintenance in fibroblasts migrating on 2D ECMs. In contrast, in 3D ECM microenvironments, FMN2 and its actin nucleating activity are required for cell survival. In the absence of FMN2, fibroblasts migrating through confined 3D microenvironments experience NE rupture and DNA DSBs that are sensed in an ATM kinase-dependent manner. We demonstrate that the DNA damage induced by loss of FMN2 are the consequence of physical constraints, as tighter confinement induces worse damage. The rapid death of cells lacking FMN2 during confined migration suggests FMN2 as a potential target for inhibiting migration of cells in dense and stiff tissues such as tumors(Butcher et al. 2009) or through small pores such as those formed during intravasation and extravasation from the bloodstream. Indeed, we find that FMN2 generates a similar actin/FA system in melanoma cells, and promotes their survival and protects them from DNA damage during migration through confined 3D microenvironments. Furthermore, we demonstrate that melanoma cells lacking FMN2 are largely blocked from surviving the circulatory system, extravasating to the lung, and forming metastases. We thus suggest that the high level of FMN2 seen in several types of cancers may endow them with a robust perinuclear actin/FA system is advantageous to their survival during the physical challenges of metastasis. Collectively, our results highlight the importance of the FMN2-generated perinuclear actin/FA system in preventing mechanical damage to the nucleus and genetic material during confined migration, and support FMN2 as a potential target for inhibiting metastasis of cancer cells. Our data provide critical new insight in the growing literature connecting nuclear function and DNA damage with the cytoskeleton and its role in physically regulating these processes(Maniotis et al. 1997; Kumar et al. 2014; Lottersberger et al. 2015; Yamada et al. 2013; Belin et al. 2015; Harada et al. 2014). It has been known for decades that force from the outside of the cell can be transmitted to the nucleus via the actin cytoskeleton(Maniotis et al. 1997). More recent work shows that force induces reorganization of nuclear proteins and DNA damage signaling (Kumar et al. 2014; Bekker-Jensen et al. 2006). The cytoskeleton may also have multiple roles in DNA repair, including increasing mobility of broken DNA and clearing DSBs(Belin et al. 2015),(Lottersberger et al. 2015). Our data support yet another role for the cytoskeleton in preventing NE rupture and DNA damage during cell migration in confined environments. Recent evidence shows that ruptures of the NE during confined migration of immune or cancer cells leads to aberrant nucleo-cytoplasmic mixing and DNA DSBs, which are both resolved upon ESCRT III- and lamin-mediated NE repair(Denais et al. 2016; Raab et al. 2016). In addition, NE collapse in micronuclei triggers massive DNA DSB and chromosome rearrangements(Hatch et al. 2013; Maciejowski et al. 2015). Together with our results, this suggests that FMN2 and the perinuclear actin-FA system it generates may prevent NE rupture or facilitate NE repair by either ESCRT III and lamin-dependent or -independent pathways. How loss of nuclear integrity and induction of DNA damage during confined migration leads to cell death is unclear(Denais et al. 2016; Raab et al. 2016). But because γ-H2AX staining in apoptotic cells is either diffuse or forms a ring at the nuclear periphery (Solier & Pommier 2014), while in cells lacking FMN2 γ-H2AX forms discrete foci, we suspect that they may die by a pathway distinct from classical apoptosis. Indeed, it is widely acknowledged that cell death in response to DNA damage is variable and can be mediated through several disparate pathways(Borges et al. 2008). Different cell types utilize different cytoskeletal mechanisms to mediate mesenchymal, ameboid, or bleb-based motility (Clark & Vignjevic 2015), suggesting the likelihood that different cell types employ different cytoskeleton-based mechanisms of nucleo-protection. Although mice lacking FMN2 exhibit female infertility due to defects in meiosis(Leader et al. 2002; Yi et al. 2013), they develop and live normally(Leader & Leder 2000), suggesting that cells undergoing the invasive migration programs that occur during normal development, wound healing, and immune responses utilize other nucleo-protective mechanisms. Indeed, dendritic cells of the immune system utilize Arp2/3-mediated actin polymerization to facilitate nuclear deformation and movement during tightly confined migration(Thiam et al. 2016), while immune cells of various types express little-to-no FMN2 mRNA (http://refdic.rcai.riken.jp). Thus, although the phenotype of nuclear damage upon confined migration observed in immune cells(Raab et al. 2016; Denais et al. 2016) is similar to what we document in mesenchymal cells in the absence of FMN2, these two cell types clearly utilize different mechanisms for protecting their nuclei from mechanical damage and DNA breakage. Thus, there are likely other cell-type specific mechanisms for cytoskeleton-based nucleprotection during cell migration in different developmental and disease contexts. The functions identified for FMN2 in primary fibroblasts were confirmed in metastatic melanoma cells. Tumor metastasis is a major contributor to cancer patients’ deaths, due to the deleterious effects of the lesions or complications of treatment. While FMN2 levels had no impact on 2D cell motility, invasion in vitro and metastasis in vivo were significantly reduced by its knockout. These data extend previous findings that 3D assays provide novel information(Bissell & Hines 2011) and highlight another actin regulatory network as important in metastasis. Invasion through confined spaces could occur as tumor cells extravasate the circulatory system and push their way into a tissue as an expanding metastasis. The metastatic process is complex and thought to be driven by phenotypic plasticity(Turajlic & Swanton 2016; Celia-Terrassa & Kang 2016). Given the link of FMN2 and DNA damage, it is possible that FMN2 levels regulate the threshold of accumulated mutations that can provide the multiple required phenotypes without producing lethal damage. The lack of FMN2 expression in immune cells strengthens the case for FMN2 as an anti-metastasis target whose inhibition would not interfere with immune function or immunotherapy. In addition, it will be interesting to determine the effect of FMN2 inhibition on the chemotherapeutic efficacy of DNA repair inhibitors. Individual members of the formin family have been shown to mediate formation of functionally distinct actin structures in cells, including SFs, filopodia, isotropic cortical actin networks, and mitochondria-associated actin(Skau & Waterman 2015; Campellone & Welch 2010). We find that FMN2 specifically mediates the formation of a previously uncharacterized, compositionally and dynamically unique perinuclear actin/FA system. The regulation and localization of actin nucleation factors are likely critical for specifying the function of actin structures in cells(Skau & Waterman 2015). The N-terminal myristoylation site may promote FMN2 localization to the unidentified perinuclear membrane compartment to facilitate local formation of nucleus-associated actin bundles. We speculate that these membranes could be extensions of the outer nuclear membrane, thus providing a direct association between polymerizing actin and the NE. Previous reports have demonstrated interactions between FMN2 and the actin nucleation factor, Spire,(Vizcarra et al. 2011) as well as with the cell cycle regulator p21(Yamada et al. 2013). We cannot rule out a role of Spire or p21 in localizing FMN2 to the perinuclear region or mediating its function in nucleo-protection. Future studies using point mutants and isolated domains of FMN2 will provide insight into how localization and activation of FMN2 are controlled in nucleo-protection and promotion of metastasis. Our data illuminate a critical aspect of cell migration in confined 3D environments. Beyond the ability of the cell to protrude its edge and retract its tail, a cell must also maintain the integrity of its DNA and protect it from external forces that might induce damage. It now seems that multiple cytoskeletal elements are involved in positioning the nucleus and regulating nuclear function. Here we show how one system, the perinuclear actin/FA system generated by the formin FMN2, has an essential function in this process during mesenchymal cell migration in confined microenvironments, thereby promoting metastasis of melanoma cells. CONTACT FOR REAGENT AND RESOURCE SHARING Clare M. Waterman is the Lead Contact for reagent and resource sharing. All published reagents can be shared on an unrestricted basis; reagent requests should be directed to and will be fulfilled by the lead author. EXPERIMENTAL MODEL AND SUBJECT DETAILS Animal Models All animal experiments were performed in accordance with approved protocols from the Institutional Animal Care and Use Committee of the NHLBI. Six-week old female C57BL/6J mice (8 mice per condition for a total of 64 mice) were obtained from Jackson labs and housed four or five to a cage as per institutional regulations. Animals were monitored daily for signs of distress or inflammation at the injection site. Animals were euthanized as per NIH guidelines with carbon dioxide Cell Lines Primary murine embryonic fibroblasts were obtained by the investigators as previously described (Thievessen et al. 2013; Skau et al. 2015). Fibroblasts were isolated from mice that were maintained according to guidelines approved by the National Heart, Lung and Blood Institute Animal Care and Use Committee. Four to six month old pregnant mice from C57J/BL6 X C57J/BL6 timed matings were obtained from Jackson Labs (Bar Harbor, Maine, USA). At embryonic day 13.5, mice were euthanized as per NIH guidelines with carbon dioxide followed by secondary termination via cervical dislocation. The uterus was removed, then individual embryos were removed from the uterus and placenta, decapitated and their internal organs removed. Tissue was cut into pieces and incubated at 37 degrees for 30 minutes in 0.25 mg/ml Trypsin-EDTA (Life Technologies, Grand Island, NY, USA) with mild vortexing every 10 minutes. After digestion, cell suspension was passed through 100 μm nylon-mesh cell strainer, then cells were collected by centrifugation at 1,600 rpm for 8 minutes. Media containing trypsin, lipids, etc was removed and single mouse embryonic fibroblasts (MEF) were re-suspended and transferred to sterile tissue culture dishes in DMEM/20% FBS (Gibco, Grand Island, NY, USA). Non-adherent cells were removed after 4 hours. Adherent cells were allowed to attach overnight then sub-cultured for transfection, or frozen in liquid nitrogen in FBS/10% DMSO without further passaging. After thawing from liquid nitrogen, MEF were passaged no more than five times. Human foreskin fibroblasts and human umbilical vein endothelial cells were obtained from American Type Culture Collection (Manassas, VA, USA) and maintained at 37 degrees in DMEM/10% FBS at 5% CO2. Caco2 and MDCK cells were a gift of Dr James Anderson (NHLBI) and maintained according to standard protocols. B16-F10 melanoma cells of C57BL/6J origin were obtained from ATCC (Manassas, VA, USA) and were cultured according to supplier’s recommendations in DMEM+10% FBS. Before injection into mice, all cell lines (CRISPR, wild type overexpression, vector, etc) were subject to IDEXX IMPACT I testing for common infectious agents and were negative for all agents tested including Mycoplasma. METHOD DETAILS Reagents and Transfection (See also KEY REAGENTS TABLE) Mammalian expression vectors containing cDNAs encoding mApple- or EGFP-actin, EGFP-paxillin, mCardinal-, GFP- or mCherry-H2B, mCherry-α-actinin, mCherry-Lamin B, NLS-mEmerald, or myosin light chain-GFP were purchased from Addgene (Cambridge MA). FMN2-GFP expression vector was purchased from GeneCopoeia (Rockville, MD, USA). MEF were transiently transfected using Amaxa Kit V solution (Amaxa Nucleofector, Lonza, Walkersville MD), program “MEF alternate program T-020” and 1 ug DNA, then recovered in DMEM/20% FBS overnight before plating on coverslips coated with 5 ug/mL human plasma fibronectin (EMD Millipore, Billerica, MA, USA) or 0.01% poly-l-lysine (Catalog #P4707, Sigma-Aldrich, St. Louis, MO, USA) (see below for imaging conditions). B16-F10 cells were transfected with 700ng-1 ug DNA using Amaxa Kit V solution (Amaxa Nucleofector, Lonza, Walkersville, MD, USA), program “B16-F10 ATCC P020”. ON-TARGETplus SMARTpool siRNA and 3′UTR siRNA against FMN2 was purchased from GE Healthcare Dharmacon (Lafayette, CO, USA). The myosin II inhibitor blebbistatin was purchased from Toronto Research Chemicals (USA) and used at 20–50 μM. Latrunculin A (used at 500 nM) and Triton X-100 were purchased from Sigma-Aldrich (St. Louis, MO, USA). Alexa-488 and -560 phalloidin were obtained from Invitrogen (Carlsbad CA, USA). The nuclear stain 4′,6-Diamidino-2-Phenylindole (DAPI) was obtained from Thermo Scientific, Grand Island, NY, USA. Alexa-647-labeled fibronectin was a gift of Dr. K. Yamada (NIH, Bethesda, MD, USA) and was used at 2.5 ug/mL with 2.5 ug/mL unlabeled human plasma fibronectin added simultaneously. ATR inhibitor AZ-20 was obtained from Tocris (Minneapolis, MN, USA) and ATM inhibitor KU-55933 from SelleckChem (Houston, TX, USA). The following primary antibodies were used for indirect immunofluorescence(See also KEY REAGENTS TABLE): anti-FMN2 clone Y-15 (goat; Santa Cruz Biotechnology, Dallas, TX), anti-FMN2 (rabbit, catalog # ab72052 Abcam, Cambridge UK), anti-INF2 (rabbit; ProteinTech, Chicago, IL, USA); anti-paxillin clone 349, anti-Hic5 clone 34/Hic5 (both mouse; BD Biosciences, San Jose, CA, USA); anti-SUN2 (rabbit, catalog #ab124916, Abcam, Cambridge, UK); anti-phospho-paxillin catalog #44-720G, anti-phospho-SRC catalog #44-660G (both rabbit; ThermoFischer, Grand Island, NY, USA); anti-phospho-FAK clone 31H5L17 (rabbit; Invitrogen/Life Technologies, Grand Island, NY, USA); anti-VASP clone 9A2 (catalog #3132S), anti-ILK (catalog #3862), anti-myosin-IIA (catalog # 3403P), anti-myosin-light-chain II (catalog #3672), anti-phospho-myosin-light-chain II T18/S19 (catalog #3674P), anti-phospho P130Cas (catalog #4011), anti-phospho-histone H2A.X clone 20E3 (catalog #9718), and anti 53BP1 catalog #4937 (all rabbit; Cell Signaling Technologies, Danvers, MA, USA); anti-α-actinin clone BM-75.2, anti-tropomyosin clone TM311, anti-talin clone 8D4, anti-vinculin clone VIN-11-5, anti-gamma-tubulin clone GTU-88, anti-lamin A/C clone 4C11, anti-vimentin clone V9 (all mouse; Sigma-Aldrich, St. Louis, MO, USA);anti-fibronectin (rabbit, catalog #F3648; Sigma-Aldrich, St. Louis, MO, USA); ; anti-tubulin clone DM1A (mouse, catalog #ab7291, Abcam, USA); anti-zyxin (rabbit; gift of Dr. M Beckerle, University of Utah, Salt Lake City, UT, USA); anti-myosin-IIB (rabbit, catalog # PRB-445P; BioLegend/Covance, Dedham, MA, USA). The following secondary antibodies were used for indirect immunofluorescence: donkey anti-rabbit Alexa 647, Cy2 and Cy3 conjugate, donkey anti-mouse Alexa 647, Cy2 and Cy3 conjugates, (Jackson ImmunoResearch Laboratories, West Grove, PA, USA). Collagen Gel Polymerization Rat tail collagen type I was purchased from Corning (Bedford, MA, USA) and diluted to a final concentration of 2 or 3 mg/mL in MEM (Gibco) with 5 ug/mL fibronectin and DMEM/0% FBS and the pH adjusted with a 7.5% sodium bicarbonate solution. For experiments in which collagen porosity was manipulated, collagen was polymerized at 4, 25, or 37 degrees in 35 mm glass bottom dishes (No 1.5, MatTek, Ashland, MA, USA) for 36 hours, then cells were seeded on top of the gel and allowed to penetrate for 20 hours. For all other experiments involving 3D collagen gels, cells were trypsinized and mixed with liquid collagen, and this mixture was added to uncoated 35 mm glass bottom dishes and allowed to polymerize at 37 degrees overnight. For live imaging, collagen gels were covered the following day with DMEM/5% FBS without phenol red. For fixation, collagen gels were fixed without adding media. For 2D collagen-coating of coverslips, collagen was diluted to 50 ug/mL in 0.02 M acetic acid and added to cleaned coverslips at a final concentration of 5 ug/cm2, incubated for 1 hour at 37 degrees, then washed with 1X phosphate buffered saline before plating cells. Transwell Migration and Invasion Assays Twenty-four well QCM Chemotaxis Cell Migration Assay (“Boyden chamber”) kits with 8 μm pores were obtained from EMD Millipore (Billerica, MA, USA) or 12 μm pores were obtained from Chemicon International (Temecula, California, USA) and coated with 10 ug/mL fibronectin (EMD Millipore). Serum-starved cells were seeded in the upper chamber and allowed to migrate for 8 hours into DMEM/10%FBS in the lower chamber. Inserts were removed to a new well with 225 uL cell detachment solution (provided) and incubated at 37 degrees for 30 minutes, pelleted briefly, re-suspended in a small volume and plated on glass coverslips coated with 5 ug/mL fibronectin for 25–30 minutes at 37 degrees, then fixed and stained as described below. Invasiveness of all B16-F10 cells lines was assayed by migration through a commercially available 24-well invasion assay (Corning BioCoat Matrigel Invasion Chamber, Corning, Bedford, MA, USA). As per the manufacturer’s instructions, Matrigel inserts on an 8um pore-size filter were rehydrated with warm DMEM for 2 hours, then 5 × 104 serum-starved cells were added to the upper chamber and allowed to migrate for 24 hours into DMEM containing 10% FBS. Cells that passaged through were detached from the membrane and quantified using the fluorometric reagent supplied in the QCM Chemotaxis kit in a SpectraMax plate reader (Molecular Devices, Sunnyvale, CA, USA) with 480/520 filter set. UV Damage UV damage was induced in fibroblasts plated on fibronectin-coated coverslips by placing them uncovered in a CL-1000 254 nm Ultraviolet Crosslinker (UVP, Upland, CA, USA). For time course of recovery from UV irradiation, cells were exposed to 7500 uJ/cm2, then recovered in fresh DMEM/20% FBS for 0.5, 7, 24, or 48 hours. FMN2 Mutagenesis and Cloning Plasmid containing the cDNA encoding murine FMN2-GFP was obtained from GeneCopoeia. The conserved isoleucine in the FH2 domain of FMN2 was identified based on(Ramabhadran, Pinar S. Gurel, et al. 2012) and the following primers were used to make the I 1225 to A mutation at nucleotide 1226: Forward: 5′-aatgcagactagacattagagctcctactgcttgtgaccttttgttg-3′ Reverse: 5′-caacaaaaggtcacaagcagtaggagctctaatgtctagtctgcatt-3′ using the QuikChange II Site Directed Mutagenesis Kit (Agilent, Santa Cruz, CA, USA) to generate FMN2 I1226A GFP. From this plasmid, the following primers were used to amplify the FH2 domain (amino acids 1139–1529) of FMN2: Forward: 5′ TACTTCCAATCCAATGCAgctaggaagcagctgatcgagcc 3′ Reverse: 5′ TTATCCACTTCCAATGTTATTAttatttaaagtcagagctgaa 3′ where the sequence in all capital letters represents the LIC cloning site for insertion of the fragment into the pET His6 StrepII TEV LIC cloning vector (2HR-T) obtained from Addgene (catalog # 29718; Cambridge, MA, USA). The FH2 domain was amplified with Phusion polymerase (New England Biolabs, Ipswich MA) and constructs were checked by sequencing. PDMS Microchannel Generation Micro-channels were prepared as previously described(Heuze et al. 2011; Faure-Andre et al. 2008). Briefly, polydimethylsiloxane (PDMS) (SYLGARD® 184, DOW CORNING) was used to prepare 7μm X 5μm micro-channels with 3.5 um X 4.2 um constrictions from a custom-made mold. The PDMS chamber and a 35 mm glass bottom dish (FD35–100, WPI) were plasma activated before being bound to each other. Binding was strengthened in a 65°C oven for 1h. After strengthening, microchannels were plasma cleaned then incubated with 10 μg/mL of fibronectin at RT for 1 h then washed with PBS before being incubated with DMEM/20% FBS (Gibco, Grand Island, NY, USA) for at least 1 h at 37 °C and 5% CO2 prior to cell loading. Fixation and Immunofluorescence For fixation of cells on coverslips, indirect immunofluorescence was performed as described (Pasapera et al. 2010) Briefly, cells plated on 5 ug/mL fibronectin overnight were fixed for 20 minutes with 4% paraformaldehyde (Electron Microscopy Science, Hatfield, PA, USA), permeabilized with 0.5% triton-X100, then washed with 0.1 M glycine followed by 3 washes with TBS. Fixed cells were blocked for at least 1 hour in 2% BSA/0.1% Tween-20 with Alexa-488 phalloidin (1:400 dilution, Invitrogen)in TBS and incubated for at least 2 hours in primary antibody, washed, and incubated at least 1 hour in fluorescently-conjugated secondary antibody. Coverslips were then mounted on glass slides with DAKO fluorescent mounting media (Agilent Technologies, Carpinteria, CA, USA). For fixation of cells in collagen gels, indirect immunofluorescence was performed as described(Fischer et al. 2009) with minor modifications. Briefly, cells in collagen gels were fixed for 30 minutes with 4% paraformaldehyde (Electron Microscopy Science, Hatfield, PA, USA), washed once with TBS, permeabilized with 0.5% triton-X100 for 35 minutes, then washed with 0.1 M glycine for 30 minutes followed by 3 washes with 1X TBS+0.1% Triton. Fixed cells were blocked for at least 6 hours to overnight in 2% BSA/0.1% Tween-20/2% normal goat serum with Alexa-560 phalloidin (1:400 dilution, Invitrogen) and incubated overnight in primary antibody, washed with 1X TBS+0.1% Triton, and incubated at least 5 hours in fluorescently-conjugated secondary antibody with 4′,6-Diamidino-2-Phenylindole (DAPI; Thermo Scientific, Grand Island, NY, USA) followed by 3 washes with 1X TBS+0.1% Triton, then 3 washes with 1X TBS without Triton. Collagen gels were then mounted in glass-bottom MatTek dishes (see above) with DAKO fluorescent mounting media (Agilent Technologies, Carpinteria, CA, USA) and covered with glass coverslips. Imaging Spinning disk confocal imaging of both fixed and live cells was performed with a Plan Apo 60×1.40NA Ph oil immersion objective lens on an inverted Eclipse Ti microscope system (Nikon Instruments, Melville, NY, USA) equipped with the Nikon PerfectFocus™ system, a servo-motor controlled X-Y stage and a PZ-2000 Piezo Z stage (Applied Scientific Instrumentation, Eugene, OR, USA), and a spinning disk confocal scan head (CSU-X; Yokogawa, Tokyo, JP). Laser illumination was provided by 405 nm, 488 nm, 561 nm and/or 655 nm solid state lasers fitted in a custom laser combiner module (Spectral Applied Research, Richmond Hill, ON, CN) and delivered to the confocal scan head (or the Nikon TIRF illuminator, see below) via a single mode optical fiber (Oz Optics, Carp, ON, CN). Photobleaching was accomplished using a FRAPPA dual galvanometer scan head (Andor Technology, South Windsor, CT, USA). Appropriate multi-bandpass dichromatic mirror and single bandpass emission filters (Semrock In., Rochester NY, USA) were used to select emission wavelength. Pairs or sets of 4 images were captured in immediate succession with one of two cooled CCD cameras (CoolSNAP HQ2 or CoolSNAP MYO; Photometrics, Tuscon, AZ, USA) operated in the 14-bit mode. For live cell imaging, temperature was controlled with either an Air Stream Stage Incubator (Nevtek, Williamsville, VA, USA) or a LiveCell Incubation Chamber (Pathology Devices, Westminster, MD, USA) which also controlled humidity. Dual-color time-lapse TIRF microscopy of EGFP and mApple- or mCherry-tagged proteins in living cells was performed at 37°C using an Apo TIRF 100×1.49 NA oil immersion objective lens (Nikon Instruments, Melville, NY, USA) on an inverted Eclipse Ti microscope system (Nikon Instruments, Melville, NY, USA) with an evanescent field depth of ~100 nm. Pairs of EGFP (using 488 nm laser illumination) and mCherry (using 561 nm laser illumination) images were captured in rapid succession at 5s intervals using either an EMCCD (Cascade II:1024; Photometrics, Tuscon, AZ, USA) or cooled CCD (CoolSNAP HQ2, Photometrics, Tuscon, AZ, USA) camera. Alternation between TIRF imaging of fluorescence and widefield epifluorescence imaging was performed by automated switching between the two illumination pathways. Widefield illumination was provided by an LED system (Lumencor, Beaverton OR, USA). Long-term phase-contrast imaging of individual cells migrating randomly or in monolayer wound healing assays on coverslips coated with 5 ug/mL human plasma fibronectin or in 3D collagen ECMs was performed on an inverted Eclipse Ti microscope system (Nikon Instruments, Melville, NY, USA) with a Plan Fluor Ph 10X/0.30 NA dry objective lens (Nikon Instruments, Melville, NY, USA). Illumination provided by a quartz-halogen bulb was attenuated with 546nm, heat-cut, and neutral density filters. Pairs of phase contrast and EGFP spinning-disk confocal (using 488 nm laser illumination) images were captured by automated switching between the two light paths at 5 minute intervals at multiple stage positions for 16 hours. All live cell experiments were performed in Phenol red–free DMEM containing 5% FBS, 20 mM Hepes, and 10 U/ml oxyrase (Oxyrase, Mansfield, OH, USA) as imaging medium. All electronic functions on the Eclipse Ti microscope systems were controlled by MetaMorph imaging software (Molecular Devices, Sunnyvale, CA, USA). Protein Preparation and Purification cDNAs encoding mouse FMN2 FH2 domain, corresponding to amino acids 1137–1529 (with or without a point mutation that resulted in a I1225 to A amino acid substitution) were expressed in BL21(DE3)-Rosetta2 E. coli cells (Millipore, DarmStadt, FGR) as His6-fusion proteins, following procedures used previously (Kim et al. 2015). Briefly, expression was induced in log phase cultures with 0.5 mM IPTG at 16°C. After expression, cells were harvested at 5000 RCF, then lysed by sonic disruption in lysis buffer. Lysates were clarified by centrifugation at 20,000 RCF, and then extracts were passed over Ni-NTA agarose (Qiagen, Hilden, FGR) equilibrated in wash buffer, and eluted off with elution buffer. The His6 tag was cleaved with tobacco etch virus protease (1:50 ratio) during overnight dialysis in wash buffer. The protein was then passed over Ni-NTA agarose again to remove any uncleaved product, and then further purified by gel filtration with Supderdex200 (GE Healthcare Life Sciences, Pittsburgh PA, USA) in gel filtration buffer. Peak fractions were pooled, dialyzed into storage buffer, then flash-frozen in liquid nitrogen and stored at −80C. The wild type and I1225A variant of the FMN2 FH1-FH2 domain (amino acids 1139–1529) were also purified, but very low expression and poor solubility prevented extensive biochemical characterization. The following buffers were used for protein purification: Lysis buffer (50mM Tris-Cl pH 8.0, 200mM NaCl, 10mM Imidazole, 1mM DTT, and Complete ULTRA protease inhibitor tablets from Roche Life Sciences, Indianapolis, IN, USA), Wash buffer (50mM Tris-Cl pH 8.0, 200mM NaCl, 10mM Imidazole, 1mM DTT), Elution buffer (50mM Tris-Cl pH 8.0), 200mM NaCl, 250mM Imidazole, 1mM DTT), Gel Filtration buffer (10mM Tris-Cl pH 8.0, 50mM KCl, 1mM DTT), Storage buffer (Gel Filtration buffer in 50% glycerol). The following buffers were used for polymerization assays: G-buffer (2 mM Tris, pH 8, 0.5 mM DTT, 0.2 mM ATP, 0.1 mM CaCl2, and 0.01% NaN3), G-Mg buffer (same as G-buffer but with 0.1 mM MgCl2 instead of CaCl2), 10× K50MEI (500 mM KCl, 10 mM MgCl2, 10 mM EGTA, and 100 mM imidazole, pH 7.0), and polymerization buffer (G-Mg buffer plus 1× K50MEI) Pyrene-actin filament assembly assays To monitor the polymerization of actin filaments, 40μL of 6μM actin (10% pyrene labeled, catalog # 8102-01 Hypermol, Bielefeild, FGR)) was mixed with 80μL of formin (150–750nM) in 1.5x polymerization buffer. Pyrene fluorescence (365/407nm) was monitored with a PTI Technologies (Ovnard CA, USA) QuantaMaster fluorimeter within 30s-1min of mixing actin and formins. Construction of genetically modified B16-F10 cell lines Two independent oligonucleotide sets for production of sgRNAs for targeting genomic FMN2 for excision by CRSPR-Cas-9 technology were designed using an online tool (crispr.mit.edu): Oligo Set 1 5′ CACCGTTTTGTGCGTAGATCCTCGA 3′ 5′ AAACTCGAGGATCTACGCACAAAAC 3′ Oligo Set 2 5′ CACCGGCAACTGTAATTCAGCAAC 3′ 5′ AAACGTTGCTGAATTACAGTTGCC 3′ Oligos were cloned into CRISPR/Cas9 vectors vector (pSpCas9(BB)-2A-PuroV2.0, AddGene, Cambridge MA) using established protocols, transformed into DH5α bacteria, mini-prepped and validated by sequencing. B16-F10 cells were then transfected with 700 ng of CRISPR vector with listed oligos or empty CRISPR vector (vector only control) containing a puromycin selectable marker. 48 hours after transfection, cells were subjected to selection by 1 ng/ul puromycin then diluted down to isolate single cells which were grown in DMEM+10% FBS+puromycin. At least four clones of each CRISPR oligo set were isolated and two clones of each CRISPR oligo set were used for further analysis and tail-vein injection (see below). B16-F10 cells were similarly transfected with the FMN2-GFP expression vector (which contained a geneticin selectable marker) and clones were selected in 800 ng/mL G418 to obtain FMN2 over-expressing lines. Similarly, the process was repeated with both CRISPR oligo sets and simultaneously the FMN2-GFP vector, or with empty CRISPR vector and simultaneously with mEGFP-N1 expression vector (vector+GFP control, AddGene, Cambridge MA). Cells were selected with both G418 and puromycin to obtain lines with FMN2 deleted and re-expressing full-length FMN2 to serve as rescue control lines and vector+GFP control lines (see Supplemental Figure 7D). At least two clones of each cell line were obtained and the expected level of FMN2 protein expression validated by western blot. All B16 cells lines were subjected to analysis by growth curve as well as measurement of random migration velocity and migration through a commercially available invasion assay (see above). In Vivo Metastasis Experiments All animal experiments were performed in accordance with approved protocols from the Institutional Animal Care and Use Committee of the NHLBI. Six-week old female C57BL/6J mice were injected with 2.0 × 105 cells in 100 uL of PBS+10% OptiMem into the tail vein on day 0. Four mice were injected with each cell line, for a total of 8 mice per genotype (i.e. four mice with clone 1 of FMN2 overexpressing cells, four mice with clone 2 of FMN2 overexpressing cells for a total of 8 mice with FMN2 overexpressing cells, etc). Animals were monitored daily for signs of distress or inflammation at the injection site. On day 15, all animals were euthanized with CO2 and lungs inflated with ~1.5 mL 10% neutral-buffered formalin (Sigma-Aldrich, St Louis MO, USA) using a 19 gauge needle through the trachea. Lungs (with heart, thymus and trachea attached) were then removed and fixed in 20 mL 10% neutral-buffered formalin overnight. After fixation for 16 hours, surface metastases were counted and tissue was imaged with the camera of an iPhone 5S (Apple, Inc, Cupertino, CA, USA). Tissue Histology Lungs were transferred to 70% ethanol and processed for histological analysis with a Leica ASP 300 tissue processor. Briefly, lungs were dehydrated through gradients of alcohols and xylenes and embedded in paraffin. Two series of paraffin sections were cut. The first set of sections was cut at the lung surface and the second set was cut 200 micrometers deeper, closer to the center of lung. The right lung was cut longitudinally and the left lung was cut along the short axis. Every other slide was stained with H&E using a Leica Autostainer (Leica Biosystems, Buffalo Grove IL, USA). After mounting, H&E stained sections were imaged on a Hamamatsu NanoZoomer 2.0-RS in 40X mode and images processed using NDP.view 2 (Hamamatsu Photonics, Hamamatsu JP). Tissue Immunohistochemistry Unstained lung sections were mounted on slides then treated for immunohistochemistry: deparaffinization in three changes of 100% xylene, rehydration and washing with 100%, 95%, then 70% ethanol, then washing extensively with PBS. Quenching of autofluorescence was accomplished by two ten-minute incubations with 1 mg/mL sodium borohydride (Sigma Aldrich, St. Louis, MO, USA). Antigen retrieval was performed by washing slides with PBS +0.05% Tween-20 (Thermo Fisher Scientific, Waltham, MA, USA), then boiling slides in PBS+0.05%Tween-20+10mM sodium citrate (Sigma Aldrich, St. Louis, MO, USA) five times for two and half minutes each time in a microwave oven, cooling two minutes in between each heating. Slides were then cooled to room temperature for 30 minutes, then washed once with PBS +0.05% Tween-20. Immunostaining of γH2AX was done by first blocking tissue with 5% bovine serum albumin in PBS+0.5% Tween-20+0.1% Triton-X-100 for 4 hours, then incubating in primary antibody overnight in the presence of 1% bovine serum albumin in PBS+0.5% Tween-20+0.1% Triton-X-100 + 1 ug/mL wheat-germ agglutinin Alex Fluor 488 conjugate (Thermo Fisher Scientific, Waltham, MA, USA) +1:300 dilution of anti-phospho-histone H2A.X clone 20E3 at 4 degrees. Following two PBS+0.05% Tween-20 washes, Alexa Fluor 647 donkey anti-rabbit secondary antibody (Jackson ImmunoResearch Laboratories, West Grove, PA, USA) was applied in 1% bovine serum albumin in PBS+0.5% Tween-20+0.1% Triton-X-together 100 with 14.3 nM DAPI (Thermo Scientific) at fluorescent mounting media (Dako, Carpinteria CA, USA)and imaged by spinning disk confocal microscopy as described above. QUANTIFICATION AND STATISTICAL ANALYSIS Image Analysis Number and length of stress fibers were determined by hand-tracing the relevant features on Alexa-488 phalloidin-stained images of control and FMN2 KD MEF. Adhesion morphometry measurements were determined by hand-tracing adhesions on paxillin immunofluorescence images and quantifying with ImageJ (NIH, Bethesda, MD, USA). FA lifetime and assembly rate were measured from TIRF image sequences taken every 5 s of EGFP-paxillin in control and FMN2 KD cells. FA nucleation rate was measured by overlaying two TIRF images of EGFP-paxillin taken 100 seconds apart and counting newly generated adhesions, then tiling this procedure every 10 seconds for 20 minutes. Correlation coefficient of cell and nucleus was obtained by hand-tracing in ImageJ and measuring Shape Descriptors to obtain aspect ratio of the cell and nucleus respectively. Aspect ratios were then correlated to each other using the Correlation Coefficient function in Microsoft Excel. Nuclear lobularity was obtained by hand-tracing nuclei in ImageJ and measuring Perimeter and Area of traced objects. Analysis of the fraction of the nucleus in each quadrant of the cell was performed by hand-tracing the cell from phase images, then using ImageJ to find the centroid of the traced area. The perimeter of the nuclei was then traced, and the fraction of total perimeter located in each quadrant obtained and graphed over time. Similarly, perimeter of nuclei and cells were traced in ImageJ then the centroid of traced objects found and tracked over time to obtain centroid position over time. For epi/TIRF ratio images, background corrected images were ratioed using the Ratio Plus plugin for ImageJ and displayed using the lookup table Green-Fire-Blue. The value of ratioed images was then quantified using ImageJ integrated density measurement function. To quantify cell migration, the Manual Tracking plugin for ImageJ was used to follow movement of the nucleolus of cells in phase-contrast images of cells migrating on FN-coated coverslips for 16 hours at 20X magnification. Percent of wound closure was assessed by creating a wound in a monolayer of cells and measuring the width of the wound at time 0, the time of wounding. Twelve and 24 hours later, the width of the wound was assessed again by imaging with transmitted light at 4X with an Evos XL Cell Imaging System (Thermo Fisher Scientific); five images along the length of the wound were examined and reported as fraction of original wound width. For cells migrating through collagen gels, death was quantified by observing cell morphology and whether or not cells moved from movies taken over 12 hours. Similarly, cell death was assayed by morphology of cells as visualized at 10X with transmitted light with an Evos XL Cell Imaging System after cells had been in collagen gels in DMEM+5% FBS without phenol red for 48 hours at 37 degrees, 5% CO2. Anisotropy of the actin around the nucleus was quantified using FibrilTool(Boudaoud et al. 2014) according to the described protocol, in phalloidin-stained cells in the area of the cell that correlated with the DAPI staining of the nucleus. Aspect ratio of cells in 3D was assessed by drawing a bounding ellipse surrounding a 3D maximum intensity projection of the cell (created from spinning disk confocal Z-stacks of the phalloidin-stained actin in the cell) and measuring the aspect ratio of the ellipse using ImageJ Shape Descriptors function. To determine the fraction of cells that migrated through a Boyden chamber or invasion assay, cells were harvested using the provided cell detachment solution and counted using a Countess Automated Cell Counter (Thermo Fisher Scientific) and compared to the starting number of cells. Harvested cells were also imaged on an Evos XL Cell Imaging System with transmitted light and GFP fluorescence at 4X magnification, then the number of cells positive for H2B-GFP fluorescence was counted by hand. To determine the percent survival of cells migrating through PDMS microchannels, cells were scored by hand in phase images to determine whether the cell was in a channel or in a constriction, and to determine if the plasma membrane was intact, and in fluorescence images to determine if the nucleus was intact. To assess NLS distribution in cells migrating on FN-coated glass coverslips, perimeter of the cells and nuclei were traced by hand in ImageJ. Nuclear area was then subtracted from total cell area and integrated density of NLS-mEmerald fluorescence for nuclear area and cell area without nucleus were calculated using ImageJ Set Measurements, then expressed as a ratio of fluorescence in the nucleus to fluorescence in the cell without the nucleus. This was repeated for a single cell every 50 minutes over the duration of the movie as well as for 10 individual cells at a single time point. Cells with γH2AX foci were scored by hand from immunofluorescence images for the presence or absence of greater than 2 γH2AX foci in the nucleus in the region defined by the area stained with DAPI. To score surface metastases, lungs were dabbed dry and placed in a petri dish then were examined by eye using forceps to expose all surfaces. To quantify lung metastases in hematoxylin and eosin stained lung sections, color images were acquired with a Hamamatsu NanoZoomer as described above. Relative areas of tissue occupied by melanoma metastases was quantified by measuring the area and perimeter of both metastases and whole lung cross-sections using hand selected regions using FIJI/ImageJ. “Metastasis compactness” was calculated as the ratio of the area:perimeter ratio of the metastasis to the area:perimeter ratio of a circle of similar area, which was found using the average Feret’s diameter of the metastasis within the section. Thus, if the metastasis cross section was perfectly circular, it would have a metastasis compactness ratio of 1.0, while lower numbers indicate a more tortuous periphery and less compact metastasis. Only non-surface metastases completely contained within the image were used to measure metastasis compactness. Statistical Analysis For Figures 2B, 3B, 5A and B, 7D, F and H, and Supplemental Figures 2E, 4C and D, 5A and B, 6B, and 7F and G, mean values and standard deviations were calculated using Microsoft Excel (Redmond, WA, USA). Two-tailed type 2 (sample equal variance; homoscedastic) student’s T tests were performed on entire data sets using Excel. p values of <0.05 were reported as moderately significant (*), p values of <0.01 were reported as highly significant (**) (see also Figure Legends). For Figures 6B, D, E, G, and I, Fisher’s exact test was performed using MatLab (Mathworks, Natick, MA) using the syntax [h,p,stats] = fishertest(x) to test the null hypothesis that there are no nonrandom associations between the two categorical variables in x, against the alternative that there is a nonrandom association. The result h is 1 if the test rejects the null hypothesis at the 5% significance level, or 0 otherwise; p is the significance level of the test. Supplementary Material 1 Movie S1: Dynamics of FMN2-GFP in migrating MEF MEF expressing FMN2-GFP, mCherry-α-actinin, and mCardinal-H2B were cultured on fibronectin-coated coverslips and imaged by SDC microscopy. Time indicates minutes:seconds, scale bar equals 10 μm. Related to Figure 1. 10 Supplemental Figure 1: FMN2 and INF2 Localization Related to Figure 1. (A) Primary mouse embryo fibroblasts (MEF) transfected with FMN2-GFP (green) were plated on fibronectin-coated coverslips, fixed, and immuno-labeled with antibodies against FMN2 (red) then imaged by spinning disk confocal microscopy (SDC) Bottom panel shows recognition of FMN2-GFP by FMN2 antibody via co-localization (color overlay). Bar=20 um. (B) MEF plated on fibronectin-coated coverslips were fixed and immuno-labeled with antibodies against INF2 (blue) and paxillin (green) and actin stained with fluorescent phalloidin (actin, red) and cells were imaged by spinning disk confocal microscopy (SDC, ventral confocal section). Boxed region in upper row (Bar= 10 μm) shown zoomed below (Bar= 10 μm), white arrowhead highlights FMN2 at an FA at the terminus of a perinuclear actin bundle, red arrow highlights lack of INF2 at this location. (C) Human umbilical vein endothelial cells (HUVECS, top row), Madin-Darby canine kidney cells (MDCK, middle row), or human colorectal adenocarcinoma cells (Caco-2, bottom row) plated on fibronectin-coated coverslips were fixed and immuno-labeled with antibodies against FMN2 (green), paxillin (red) or actin was stained with fluorescent phalloidin (actin, red). DNA was stained with DAPI (blue) and cells were imaged by SDC (ventral confocal section). Boxed region in fourth column (Bar= 10 μm) shown zoomed at right (Bar= 10 μm). (D) MEF plated on coverslips coated with 0.01% poly-l-lysine were fixed and immuno-labeled with antibodies against FMN2 (green), actin was stained with fluorescent phalloidin (red) and DNA was stained with DAPI (blue) and cells were imaged by spinning disk confocal microscopy (SDC). Boxed region in fourth column (Bar= 10 μm) shown zoomed at right (Bar= 10 μm) (E) MEF plated on fibronectin-coated coverslips were fixed and immuno-labeled with antibodies against FMN2 (green) and tubulin (red, above) or vimentin (red, below) and DNA stained with DAPI (blue) then imaged by SDC. Bar= 10um. Boxed region in fourth column (Bar= 10 μm) shown zoomed at right (Bar= 10 μm). 11 Supplemental Figure 2: Perinuclear actin bundles and FA are compositionally and dynamically distinct from leading edge actin bundles and FA Related to Figure 1. (A) Primary mouse embryo fibroblasts (MEF) plated on fibronectin-coated coverslips were (A) fixed and immuno-labeled with antibodies against α-actinin (green, Atn) and myosin IIB (red, MIIB), actin was stained with fluorescent phalloidin and cells were imaged by spinning disk confocal microscopy (SDC). Boxes in perinuclear (blue) and leading edge (yellow) regions on left are shown zoomed at right. White arrowheads highlight sites of co-localization, red arrowheads highlight lack of co-localization. Bar= 10 um. (B) SDC image of a live MEF co-expressing mApple actin (red) and paxillin-GFP (green). Boxed region on left (Bar = 20 um) shown zoomed at right (Bar=10 um), white arrowheads highlight FA at both ends of a perinuclear actin bundle. (C) SDC images of a live MEF expressing GFP-actin. Left: image just after laser photobleaching of three bars across two different sets of perinuclear actin bundles (Bar= 20 um). Position of the nucleus highlighted by a red dashed line, time-lapse of the white boxed region is shown zoomed in the series at center (Bar=1um). Center: Time shown in seconds relative to the time of photobleaching, red lines highlight the position of the photobleach marks at time=0s. Right: Kymograph along the actin bundle shown at left (D= distance= 20um, T= time= 200s). (D) Left panels: SDC images of live MEFs expressing GFP-paxillin plated on fibronectin-coated coverslips with Alexa-647 conjugated fibronectin in the imaging media. White box indicates area zoomed in time series at right. Bar=10um. Right panels: Time lapse image series, elapsed time shown in minutes, all panels are oriented with the cell leading edge toward the top. Left panels: leading edge FA with associated fibronectin; right panels: perinuclear FA lacking fibronectin. (E) Quantitative analysis of FA dynamics from total internal reflection fluorescence (TIRF) images for perinuclear FA (PNA) and leading edge FA (LEA). LEA are FA within 20 um of a protruding lamella region, PNA are FA within the region of the nucleus defined by mCherry histone H2B fluorescence. All graphs show mean +/− SD, N.S.: non=significant, **:p<0.01, Student’s T-test. Top: Percentage of newly formed FA that did not disassemble soon after forming but went on to elongate (Maturation Fraction; n=250 FAs per condition). Second line left: Rate of FA elongation as measured from kymographs of individual FA (Assembly Rate; n=50 FAs per condition). Second line right: Time in min from first FA formation to final disappearance (Adhesion lifetime; n=50 FAs per condition). Third line left: aspect ratio of FA at their largest as measured by hand-tracing (n=50 FAs per condition). Third line right: average number of FAs in the cell at a single time point (Adhesion Number; n=25 cells per condition). Bottom line left: average area of individual FA at a single time point (Adhesion Area; n=50 FAs per condition). Bottom line right: average of the maximum size FAs reach at any point in their life (Maximum Adhesion Area; n=50 adhesions per condition). (F) TIRF images (evanescent field depth = 100nm) of a live MEF co-expressing GFP-paxillin (green) together with mCherry histone H2B (H2B, red), Bar= 10um. Left: Regions of the yellow (LEA and blue (PNA) boxes are shown zoomed at right. White arrow highlights PNA, white arrowhead highlights the edge of the protruding lamellipodium. Right: Time lapse image series, elapsed time shown in minutes, all panels are oriented with the cell leading edge towards the top (Bar= 2um). Upper panel, the arrowhead highlights a LEA that forms and elongates towards the cell center. Center panel, the arrowhead highlights a PNA that forms and slides and elongates towards the leading edge. (G) Primary MEF plated on fibronectin-coated coverslips were fixed and stained with fluorescent phalloidin (Act) and/or immuno-labeled with antibodies against (left) actin-associated proteins, (center), FA-associated proteins or (right) phospho-proteins at FA. Tropomyosin (TM), myosin IIA (MIIA), myosin II regulatory light chain (MLC), phosphorylated myosin II regulatory light chain (pMLC), integrin linked kinase (ILK), phosphorylated Src (pSRC), phosphorylated p130Cas (pP130Cas), focal adhesion kinase (FAK), phosphorylated focal adhesion kinase (pFAK), paxillin (Pxn). Bar= 10 um. 12 Supplemental Figure 3: Localization of proteins in cells depleted of FMN2 Related to Figure 2. (A) Primary mouse embryo fibroblasts (MEF) plated on fibronectin-coated coverslips were mock-transfected (Control, top panels), transfected with a pool of siRNAs targeting FMN2 open reading frame (FMN2 ORF siRNA, middle panels) or an siRNA targeting the 3′untranslated region of FMN2 (FMN2 3′ UTR siRNA). MEF were fixed and immuno-labeled with antibodies against paxillin (blue) and FMN2 (red); actin was stained with fluorescent phalloidin (green) and the nucleus was stained with DAPI (blue) and cells were imaged by spinning disk confocal microscopy (SDC). Boxed regions above (Bar= 20 μm) are shown zoomed below (Bar= 5 μm). (B) SDC images of control or FMN2 ORF siRNA MEFs that were immuno-labeled with antibodies against paxillin (red, upper panels) or α-actinin (red, lower panels) and actin was stained with fluorescent phalloidin (green) and the nucleus stained with DAPI (blue). Boxes in perinuclear (blue) and leading edge (yellow) regions in upper panels (Bar= 20 μm) are shown zoomed below (Bar= 5 μm). White arrowheads highlight sites of co-localization, red arrowheads, lack of co-localization. (C) SDC images of control or FMN2 3′ UTR siRNA MEFs that were immuno-labeled with antibodies against paxillin (red) and fibronectin (blue) and actin was stained with fluorescent phalloidin (green). The position of the nucleus outlined with a white dashed line, boxed regions above (Bar = 20 μm) are shown zoomed below (Bar= 5 μm). White arrowheads highlight sites of co-localization, red arrowheads, lack of co-localization. (D) SDC images of control or FMN2 ORF siRNA MEFs that were immuno-labeled with antibodies against tubulin (green), actin was stained with fluorescent phalloidin (red) and DNA in the nucleus was stained with DAPI (blue). Bar = 20 μm. 13 Supplemental Figure 4: FMN2 controls nuclear position but not lamin integrity or cell velocity Related to Figure 3. (A) Primary mouse embryo fibroblasts (MEF) plated on fibronectin-coated coverslips were co-transfected with nuclear localization signal (NLS)-mEmerald and mCherry-lamin B, together with (FMN2 KD) or without (Control) siRNA targeting the 3′untranslated region of FMN2 and imaged by spinning disk confocal microscopy (SDC). Bar=10 μm. Elapsed time in minutes is shown. (B) MEF transfected with (FMN2 KD) or without (Control) an siRNA targeting the 3′untranslated region of FMN2 were plated as a confluent monolayer on fibronectin-coated coverslips and fixed 6 hours after induction of a scratch wound and immuno-labeled with an antibody against γ-tubulin (red) as a marker of the microtubule organizing center; actin was stained with phalloidin (green) and the DNA in the nucleus was stained with DAPI (blue). Cells were imaged by SDC. Bar= 10 μm. White arrowheads indicate MTOC positioned in front of the nucleus, red arrowheads indicate MTOC positioned behind or to the side of nuclei. (C) Quantitative analysis of random migration of MEF over 16 hours on fibronectin-coated coverslips from phase imaging. n=10 cells per condition; mean +/− SD, N.S.: not significant, Student’s T-test. (D) Quantitative analysis of closure of a scratch wound by MEF co-transfected with H2B-GFP together with (FMN2 KD) or without (Control) siRNA targeting the 3′untranslated region of FMN2 (FMN2 KD) and MEF transfected with H2B-GFP, siRNA targeting the 3′untranslated region of FMN2, and re-expressing full length FMN2-GFP (Rescue). Wound closure was measured by transmitted light microscopy 24 hours after wounding and expressed as a fraction of the original wound width. n=5 images per condition; mean +/− SD, N.S.: not significant, Student’s T-test. 14 Supplemental Figure 5: FMN2 generates anisotropic actin bundles around the nucleus in 3D and promotes polarized cell shape Related to Figure 4. (A) Quantitative analysis of the anisotropy of phalloidin-labeled actin bundles (using FibrilTool8) from spinning-dick confocal images of MEF cultured in 3D collagen gels (3 mg/ml) that had been transfected with H2B-GFP together with (FMN2 KD) or without (Control) siRNA targeting the 3′untranslated region of FMN2, and MEF transfected with H2B-GFP, siRNA targeting the 3′untranslated region of FMN2, and re-expressing full length FMN2-GFP (Rescue). n=10 cells per condition; mean +/− SD, **:p<0.01; *:p<0.05; N.S.: not significant, Student’s T-test. (B) Quantitative analysis of cell polarization as measured by aspect ratio of the cell in control, FMN2-KD, and Rescue MEF cultured in 3D collagen gels (3 mg/ml). n=10 cells per condition; mean +/− SD, **:p<0.01, N.S.: not significant, Student’s T-test. 2 Movie S2: Dynamics of actin at stress fibers and paxillin at FA in MEF MEF expressing mApple-Actin (left panel) and paxillin-EGFP (center panel) with overlay (right panel) were cultured on fibronectin-coated coverslips and imaged by SDC microscopy. Time indicates minutes:seconds, scale bar equals 10 μm. Related to Figure 1. 3 Movie S3: Dynamics of paxillin at PNA in relation to the nucleus in MEF MEF expressing paxillin-EGFP (left panel) and mCherry-H2B (center panel) with overlay (right panel) were cultured on fibronectin-coated coverslips and imaged by TIRF microscopy. Time indicates minutes:seconds, scale bar equals 10 μm. Related to Figure 1. 4 Movie S4: Dynamics of actin and paxillin in control and FMN2 KD MEF cultured on glass coverslips Mock-transfected MEF (Control; left panel) and MEF transfected with siRNA targeting FMN2 (FMN2 KD; right panel) expressing paxillin-EGFP and mApple-actin were cultured on fibronectin-coated coverslips and imaged by spinning disc confocal microscopy. Time indicates minutes:seconds, scale bar equals 20 μm. Related to Figure 2. 5 Movie S5: Dynamics of actin and FMN2 in MEF cultured in 3D ECM MEF expressing mApple-Actin (left panel) and FMN2-EGFP (center panel) with overlay (right panel) were cultured in 3 mg/mL and imaged by SDC microscopy. Single confocal slice is shown. Time indicates minutes:seconds, scale bar equals 10 μm. Related to Figure 4. 6 Movie S6: Movement of the nucleus in MEF migrating in confined PDMS channels MEF transfected with mEmerald-NLS (control; top), or mEmerald-NLS and siRNA targeting FMN2 (FMN2 KD; bottom) migrating into PDMS microchannels with 7 μM channel opening and 3.5 μM constrictions. Time indicates minutes:seconds, scale bar equals 10 μm. Related to Figure 5. 7 Movie S7: Dynamics of FMN2 I1226A mutant in FMN2 KD MEF MEF transfected with siRNA targeting FMN2 (FMN2 KD), mApple-Actin (left panel), FMN2 I1226A GFP (center panel) and mCardinal-H2B with overlay (right panel) were cultured on fibronectin-coated coverslips and imaged by SDC microscopy. Time indicates minutes:seconds, scale bar equals 10 μm. Related to Figure 6. 8 Supplemental Figure 6: FMN2 promotes cell survival in 3D collagen gels but, does not affect baseline DNA damage or DNA repair after UV exposure in cells cultured on 2D ECMs Related to Figure 6. (A) Primary mouse embryo fibroblasts (MEF) were cultured in 3D collagen gels (3 mg/ml) and subjected to time-lapse phase-contrast imaging for 12hr. A representative phase contrast image of MEF transfected with GFP-histone H2B (Control) or co-transfected with GFP-histone H2B and siRNAs targeting the 3′untranslated region of FMN2 alone (FMN2 KD, KD) or co-transfected with siRNAs targeting the 3′untranslated region of FMN2 together with FMN2-GFP (Rescue) is shown at the start of imaging and 12 hours later. Bar= 20 μm. (B) Quantitative analysis of cell death in MEF transfected with GFP-histone H2B (Control) or co-transfected with GFP-histone H2B and siRNAs targeting the 3′untranslated region of FMN2 alone (FMN2 KD, KD) after culturing for 48 hours in a 3D collagen gel. n=25 cells per condition; mean +/− SD, **:p<0.01, Student’s T-test. (C) Quantitation of the nuclear to cytoplasmic fluorescence ratio of NLS-mEmerald in control and FMN2 KD MEF randomly migrating on fibronectin-coated glass coverslips in (left) fifteen individual cells and (right) a single cells over 7 and a half hours. Black dots on left graph represent individual cells; bar is the average. (D) MEF were mock-transfected (Control, top panels), or transfected with a pool of siRNAs targeting FMN2 open reading frame (FMN2 ORF siRNA, bottom panels) and plated on fibronectin-coated coverslips for 25 minutes then fixed and immuno-labeled with antibodies against phospho-γH2AX (red); actin was stained with fluorescent phalloidin (green) and the nucleus was stained with DAPI (blue) and cells were imaged by spinning disk confocal microscopy (SDC). Bar=10 um. (E) Quantitation of DNA damage as measured by induction of phospho-γH2AX foci after MEF were exposed to 7500 uJ/cm2 of UV light and allowed to recover for the time indicated. n=25 cells per condition; mean +/− SD, **:p<0.01, Student’s T-test. 9 Supplemental Figure 7: Characterization of B16-F10 melanoma cells lines with modified levels of FMN2 and their behavior in a model of murine metastasis Related to Figure 7. (A-H) Analysis of B16-F10 cell lines: B16-F10 melanoma cells (B16-F10) or B16-F10 melanoma cells lines stably expressing Cas9 vector (Vector), co-expressing Cas9 vector and GFP vector (Vector + GFP), with CRISPR-mediated deletion of FMN2 by two distinct targeting sequences, either alone (CRISPR1, CRISPR 2) or together with stable expression of FMN2-GFP (C1+WT, C2+WT), or stably over-expressing FMN2-GFP (WT-OE). (A) Western blot showing FMN2 and GAPDH protein levels in mouse primary melanocytes and B16-F10 melanoma cells. (B)Western blots showing FMN2 and tubulin protein levels in B16-F10 melanoma cells and B16-F10 melanoma cells lines. (C) Spinning disk confocal (SDC) image series of live B16-F10 cells expressing FMN2-GFP (lower left, green) with mApple-actin (upper left, red) and mCardinal-H2B (upper right, blue). Time in minutes is shown. Bar=20 um. (D) Table summarizing B16-F10 clonal cell lines generated and isolated for characterization in in vitro assays and use in murine experimental metastasis tail-vein injection model in vivo, figure(s) in which each cell line appears are noted in the right column. (E) Table summarizing the similarities and differences between the behavior of B16-F10 cell lines versus MEF control cells or MEF treated with siRNA targeting FMN2 in the various assays performed in this study. For MEFs, +/− FMN2 represents with (-) or without (+) siRNA-mediated knockdown of FMN2, for B16-F10 cells, +/− FMN2 represents with (-) or without (+) CRSPR-mediated deletion of FMN2. (F) Quantification of random cell migration speed for the B16-F10cell lines after plating on fibronectin-coated coverslips (left) or relative passage through a matrigel-coated filter (8um pore size) in a transwell invasion assay (left-center). n=25 cells per condition, for data on migration speed, only comparisons that are statistically significantly different (*: p<0.05, student’s T test) are shown, all other comparisons are not significantly different. Number of surface metastases appearing on the lung of mice 15 days post tail vein injection with the B16-F10 cell lines (right-center, n=4 mice per pooled genotype) Mean +/− SD is shown, *:p<0.05, **:p<0.01, Student’s T-test, dashed bars indicate that only the one clone was significantly different than B16-F10, the solid bars indicates that both clones were significantly different than B16-F10. Quantification of average intensity of γH2AX fluorescence in the nucleus of the B16-F10 cell lines (right, n=at least 10 cells per condition, mean +/−SD is shown, all values are not significant as compared to B16 cells except *, p<0.05, **, p<0.01, student’s T test. (G) Single slices of paraffin-embedded hematoxylin and eosin (H&E) stained lung slices 15 days post tail vein injection with the B16-F10 cell lines. Images were obtained on NanoZoomer and magnified to 20X with NDPview software. Bar= 200 um. (H) Quantitation of metastasis compactness calculated as the ratio of the area: perimeter of the metastasis in H&E sections like those shown in Figure 7E and in (G) above. Mean +/−SEM is shown, **: p<0.01, *: p<0.05, N.S.=not significant, Student’s T-test; n = 10 tissue sections with at least 8 metastases measured per condition (n of metastases: B16: n=15, CRISPR: n= 8, Rescue: n=26, WT OE: n=52). The authors would like to thank Dr. Mary Beckerle (Huntsman Cancer Institute), Dr James Anderson (NHLBI), and Dr Ken Yamada (NICDR) for reagents; Drs. Margot Quinlan and Alex M.L. Wu (NCI) for technical advice; William Shin for work on the Waterman Lab microscopes and Schwanna Thacker for administrative assistance. This work is supported by the Intramural Programs of the National Heart, Lung, and Blood Institute, (CTS, HR-T, MB, GMA and CMW) and the National Cancer Institute (AT, AN and PSS), NIH, Bethesda, MD; an NIH Director’s Early Independence Award (1DP5OD17885-1) (PG and GMA); and ANR “Netoshape” and “Association pour la Recherche contre le Cancer” (n SFI20101201669; MP). Figure 1 FMN2 associates with a perinuclear membrane, actin and FA system that moves with the nucleus during cell migration (A–H) Primary mouse embryo fibroblasts (MEF) plated on fibronectin-coated coverslips were (A–B) fixed and immuno-labeled with antibodies against FMN2 and paxillin, actin was stained with fluorescent phalloidin, and the nucleus stained with DAPI. Cells were imaged by spinning disk confocal microscopy (SDC). (A) Ventral confocal sections, arrowhead: FMN2 (red) at FA (actin, green; paxillin, blue). (B) Maximal intensity projection (upper) and X-Z 3-D reconstruction (bottom). White line: position of the X-Z image. (C–G) SDC (D–G) or Total internal reflection fluorescence microscopy (C, TIRF, evanescent field depth = 100nm) images of MEF expressing FMN2-GFP (green) with mCardinal histone H2B (blue, D, F, G; red, C, H) and mApple actin (red,D,E,F), mCherry α-actinin (red,G). (C) Boxed region of color overlay (left) is zoomed at right. (D,E) Cells treated with (D) 50μM blebbistatin or (E) 500nM latrunculin-A for 1 hour prior to imaging. Boxed regions are zoomed on the right (D) or bottom (E), arrowheads in (D): perinuclear FMN2 fibrils associated with actin clusters. (F) Cells were imaged before (Pre-Treat) and immediately after extraction with 1% Triton-x-100. (G) Left: Single time-point of color-encoded overlay, Right: time lapse image series of the boxed region from the left, time in minutes shown. Bars in A–G (left) = 10μM; C and G (right) and E (bottom)= 5um. (H) SDC image of a live MEF expressing myosin 2 regulatory light chain-GFP (MLC, green) and mCherry histone H2B (H2B, red). Boxed region in top image (Bar = 10um) zoomed at below (Bar=10um), arrowheads: myosin 2-containing bundle deforming nucleus. See also Figures S1–2 and movies S1–3. Figure 2 FMN2 is required for the formation and maintenance of the perinuclear actin/FA system (A–C) MEF plated on FN-coated coverslips were mock-transfected (Control) or transfected with a single siRNA targeting the 3′untranslated region of FMN2 alone (FMN2 KD) or together with FMN2-GFP (Rescue). (A, C) MEF were fixed and immuno-labeled with antibodies against paxillin (A, red) or myosin IIB (C, red), actin was stained with fluorescent phalloidin (green) and the nucleus with DAPI (blue). Cells imaged by SDC. Boxes in perinuclear (blue) and leading edge (yellow) regions on left (Bar= 10um): zoomed at right (Bar= 5um). White arrowheads in blue boxes of Control and Rescue panels: perinuclear actin bundle and FA; red arrowheads in FMN2 KD: lack thereof. In the Rescue panel, FMN2-GFP (red arrow far right) is in blue in the color overlay. (B) Quantitative analysis of morphometry in MEF from images like those shown in (A) for Control (Con, n= 48 perinuclear adhesions (PNA), 70 sub-nuclear stress fibers (SNF), and 186 leading edge adhesions (LEA) from 10 cells), FMN2 KD (n= 4 PNA, 13 SNF, and 170 LEA from 10 cells), and Rescue (n= 52 PNA, 63 SNF, and 170 LEA from 10 cells) conditions. LEAs are FA within 20 um from a protruding region; PNAs or SNFs are FA and actin bundles, respectively, within the region of the nucleus defined by DAPI. Mean +/− SD, N.S.: non-significant, **:p<0.01, Student’s T-test. (C) Average area of fifteen cells is given in um2 Bar= 10 um. (D) SDC images of a dorsal confocal section of control or FMN2 KD MEF immuno-labeled with antibodies against SUN2 (red); actin stained with fluorescent phalloidin (green) and the nucleus with DAPI (blue). Bar = 20 μm. Boxed region in SUN2 panel: zoomed at right; green arrowheads: enrichment of SUN2 along SNF. See also Figure S3 and movie S4. Figure 3 FMN2 and the perinuclear actin-FA system control nuclear shape and position in MEF (A–C) MEF plated on fibronectin-coated coverslips were transfected with GFP-histone H2B (H2B, green (Control, Con)) or with GFP-histone H2B and an siRNA targeting the 3′untranslated region of FMN2 alone (FMN2 KD, KD) or together with FMN2-GFP (Rescue). (A–B) MEF were imaged by time lapse phase-contrast (Phase, gray-scale) and SDC microscopy (A) and analyzed (B,C). (A) Boxed regions at left (Bar= 25um) zoomed and magnified at right (Bar= 15 um). Right: Outline of the nucleus from the H2B channel at 20 min intervals; time encoded by the color scale bar shown. (B) Upper panel: Correlation coefficient between the aspect ratio of the nucleus and the aspect ratio of the cell. Lower panel: Nuclear lobularity (area to perimeter ratio of the nucleus). n= 25 cells per condition Mean +/− SD, N.S.: non=significant, **:p<0.01, Student’s T-test. (C) Upper left: Cartoon example of the division of the cell into four quadrants (Q1–Q4) based on the position of its area centroid. Upper right: Fraction of the nucleus in each cell quadrant over time. Bottom: plot of the position of the cell centroid (Cell) and the position of the nuclear centroid (Nuc) for one control (Con) and FMN2-KD (KD) MEF. (D) MEF were fixed and immuno-labeled with antibodies against lamins A and C (red, Lam, Lamin A/C), actin stained with fluorescent phalloidin (green) and the nucleus with DAPI (blue); cells were imaged by SDC. Boxed regions at left (Bar= 10 um) zoomed and magnified at right (Bar= 5 um). (E) MEF were fixed and immuno-labeled with antibodies that recognize lamin B (red), actin stained with fluorescent phalloidin (green) and the nucleus with DAPI (blue); cells were imaged by SDC (Bar= 10 um) (F) MEF co-expressing mApple-actin (blue) and histone H2B-GFP were imaged by time-lapse wide-field epifluorescence (green) and TIRF (red) microscopy. Boxed regions at left (Bar = 10um) zoomed and magnified at right; ratio of epi H2B-GFP signal to TIRF H2B-GFP signal shown with higher intensity coded by warm color. Bar= 5 um, time in minutes. (G) Quantitative analysis of epi:TIRF fluorescence ratio of the nucleus of a single control, FMN2 KD and Rescue cell over time. See also Figure S4 Figure 4 FMN2 localizes in the perinuclear region in fibroblasts migrating through 3D collagen ECMs (A–D) MEFs were cultured in 3D collagen gels (3 mg/ml) for 16–18 hr. (A) MEF were fixed and immuno-labeled with antibodies against FMN2 (Endogenous FMN2, green), actin stained with fluorescent phalloidin (actin, red), and the nucleus with DAPI (blue). Cells were imaged by SDC. Left 3 panels: maximal intensity projections of z-stacks, white arrowhead: cortical FMN2. Boxed regions in upper panels (Bar = 20um): zoomed and magnified in lower panels (Bar= 5 um). Right panel: X/Z section (at the white line in left panels) from 3D reconstruction. (B) Time-lapse imaging of a single confocal slice of a MEF co-expressing FMN2-GFP (green), mCardinal histone H2B (blue) and mApple actin (red). Bar= 20um time in minutes. (C,D) Maximal intensity projections of SDC Z-stacks of MEF mock-transfected (Control) or transfected with an siRNAs targeting the 3′untranslated region of FMN2 alone (FMN2 KD) or together with FMN2-GFP (Rescue) fixed and stained with fluorescent phalloidin (actin, green), and the nucleus stained with DAPI (C, red, D, blue) or immuno-labeled with antibodies against vinculin (D, red). In Rescue panels, FMN2-GFP is in blue in the color overlay. Boxed regions in upper panels (Bar= 10 um) shown zoomed below (Bar= 5um). In (C), green arrowheads: perinuclear actin bundles, red arrowhead: lack thereof. Bottom panels: X/Z section (at the white line in upper panels) from a 3D reconstruction. In (D), regions of the yellow (leading edge adhesions) and blue (perinuclear adhesions) boxes zoomed and shown below, green arrowheads: FAs. See also Figure S5 and movie S5 Figure 5 FMN2 protects against NE damage to promote cell survival during migration in confined microenvironments (A,B) MEFs were transfected with GFP-histone H2B (H2B, green (Control, Con) or co-transfected with GFP-histone H2B and an siRNA targeting the 3′untranslated region of FMN2 alone (FMN2 KD, KD) or together with FMN2-GFP (Rescue). Mean +/− SD is shown, N.S. Not significant, **:p<0.01, *:p<0.05, student’s T test. (A) MEF cultured on collagen-coated coverslips or in 3D collagen gels (2 mg/ml or 3 mg/ml) were subjected to time-lapse phase-contrast imaging for 12hr; fraction of dead cells was determined visually (n=30 cells per condition) (B) MEF collected and plated after passage through 8 or 12um pore-size transwells. Left panels: Phase contrast and fluorescence images of GFP-H2B in Control and FMN2 KD panels or FMN2-GFP in the Rescue panel, Bar= 100 um. Upper right panel: Quantitation of percent of cells added to the transwell chambers that were retrieved after passage through the filter. Lower right panel: Quantitation of fraction of cells that passed through the filter that were expressing GFP-H2B (Controls and FMN2-KD) or FMN2-GFP (Rescue). (C) MEF were transfected with mEmerald with a nuclear localization signal sequence (NLS-mEmerald (Control, Con)) or co-transfected with NLSmEmerald and an siRNA targeting the 3′untranslated region of FMN2 (FMN2 KD, KD) and migrated through micro-fabricated channels. Upper left: Cartoon depiction of the dimensions of micro-channels, showing three time-points as a cell enters and proceeds through the constricted portion. Upper right: phase contract image of a micro-channel array. Cells added on the side with pillars; red boxed area: single micro-channel, white arrowhead; cell entering a micro-channel. Middle: Time-lapse SDC images of NLS-mEmerald in control and FMN2-KD cells in micro-channels, time in min, Bar= 20um. Bottom: Quantitation of fraction of cells that survive or die within the wide (Channel) and constricted (Constriction) portions of the micro-channel. See also Figure S6 and movie S6 Figure 6 FMN2-mediated actin polymerization is required for protection against double-stranded DNA breaks detected by ATM during migration in confined microenvironments (A–G) MEF were transfected with GFP-histone H2B (H2B, green (Control, Con)) or co-transfected with GFP-histone H2B and an siRNA targeting the 3′untranslated region of FMN2 alone (FMN2 KD, KD) or together with FMN2-GFP (Rescue). (A,C,F) SDC images of MEF fixed and immunostained for γH2AX or 53BP1 (red), actin stained with fluorescent phalloidin (green) and the nucleus with DAPI (blue). In the Rescue rows actin is red, FMN2-GFP is green, and γH2AX or 53BP1 is blue. Bar= 10um. (B,D,E,G,I bottom left) Quantitation of fraction of transfected cells with γH2AX foci. Mean is shown, N.S. Not significant, **:p<0.01, *:p<0.05, Fisher’s exact test. (B) Quantitation of MEF plated after passage through 8um or 12um pore-size transwells and MEF plated without passage through the transwell (Glass), n=50 cells per condition. (C) Maximal intensity projections of SDC z-stacks (C) and quantitation (D, n=10 cells per condition) for MEF cultured in 3D collagen gels (3 mg/ml) for 16–18 hr. (E) MEF were treated with 30μM AZ-20 (ATRi) or 250nM KU-55933 (ATMi) for 24hr prior to and during passage through an 8um pore-size transwell and SDC images of cells were quantitated. (F–G) MEF were treated with 20μM blebbistatin for 2 hr prior to and during passage through an 8um pore-size transwell and SDC images of cells (F, Bar= 10um) were quantitated (G) (H) Actin polymerization (2 μM total, 10% pyrene-labeled) was monitored by increase in pyrene fluorescence over time for actin alone, actin plus WT or I1226A FMN2 FH2 domain (amino acids 1139–1529 of full-length FMN2). (I) MEF co-transfected with mApple-Actin (red), mCardinal-histone H2B (blue), an siRNA targeting the 3′untranslated region of FMN2, and FMN2-I1226A-GFP (green) were imaged by SDC on FN-coated coverslips (upper panel) or in 3D collagen gels (3mg/ml, lower right panels). Bars= 10 um. Lower left: Quantitation after passage of MEF through an 8um pore-size transwell. See also Figure S6 and movie S7. Figure 7 FMN2 promotes metastasis of murine melanoma cells (A–H) Analysis of B16-F10 cell lines: B16-F10 melanoma cells (B16-F10) or B16-F10 melanoma cell lines stably expressing Cas9 vector (Vector), co-expressing Cas9 vector and GFP vector (Vector + GFP), with CRISPR-mediated deletion of FMN2 by two distinct targeting sequences, either alone (CRISPR1, CRISPR 2) or with stable expression of FMN2-GFP (C1+WT, C2+WT), or stably over-expressing FMN2-GFP (WT-OE). (A) Ventral SDC images of B16-F10 cells plated on FN-coated coverslips, fixed and immuno-labeled with antibodies against FMN2 (red). Actin stained with fluorescent phalloidin (green) and the nucleus with DAPI. Scale bar: 10 μM (B) Ventral SDC images of CRISPR FMN2-KO cells plated on FN-coated coverslips, fixed and immuno-labeled with antibodies against paxillin (red). Actin stained with fluorescent phalloidin (green) and the nucleus with DAPI (blue). Scale bar: 10 μM (C) Representative images of murine lungs fixed with 10% neutral-buffered formalin on day 15 after tail vein injections. Black spots: metastases of melanoma cells. Bar=1 cm (D) Quantification of number of metastases visible on the surface of the lungs. Two independent clones for each genotype were examined and data pooled (n=8 mice per pooled genotype). Individual clone data is in Figure S7. Mean +/− SD is shown. Values are not significantly different from B16-F10 except those marked with *, p<0.05 and **, p<0.01, student’s T test. (E) Images of paraffin-embedded hematoxylin and eosin (H&E) stained lungs were obtained on NanoZoomer and magnified to 2.5X with NDPview software. Black arrowheads: metastases. Bar=200 um. (F) Quantification of percent of tissue area occupied by metastases (Met) in 5 representative sections for each genotype. Mean +/− SD is shown. All values are not significantly different from B16-F10 except **, p<0.01, student’s T test. (G) SDC images of paraffin-embedded lung tissue stained for γH2AX (red) with wheat-germ agglutinin (WGA, green) for cell membrane and the DAPI for the nuclei (blue). Scale bar: 10 um. (H) Quantification of fraction of cells within a tumor (as identified by lack of WGA staining and dark area in phase images) with γH2AX foci; n=at least 70 cells for each of three representative areas. Mean +/− SD is shown, **: p<0.01, student’s T test. See also Figure S7. KEY RESOURCES TABLE REAGENT or RESOURCE SOURCE IDENTIFIER Antibodies anti-FMN2 clone Y15; discontinued Santa Cruz Biotechnology Dallas TX Cat # sc-22729 anti-FMN2 Abcam Cambridge, UK Cat # ab72052 anti-INF2 ProteinTech, Chicago, IL Cat # 20466-1-AP anti-paxillin clone 349 BD Biosciences San Jose CA Cat # 610052 anti-Hic5 clone 34/Hic5 BD Biosciences San Jose CA Cat # 611164 anti-SUN2 Abcam, Cambridge, UK Cat #ab124916 anti-phospho-paxillin ThermoFischer, Grand Island, NY Cat #44-720G anti-phospho-SRC ThermoFischer, Grand Island, NY Cat #44-660G anti-phospho-FAK clone 31H5L17 Invitrogen/Life Technologies, Grand Island, NY Cat #700255 anti-VASP clone 9A2 Cell Signaling Technologies Danvers MA Cat #3132S anti-ILK Cell Signaling Technologies Danvers MA Cat #3862 anti-myosin-IIA Cell Signaling Technologies Danvers MA Cat # 3403P anti-myosin-light-chain II Cell Signaling Technologies Danvers MA Cat #3672 anti-phospho-myosin-light-chain II T18/S19 Cell Signaling Technologies Danvers MA Cat #3674P anti-phospho P130Cas Cell Signaling Technologies Danvers MA Cat #4011 anti-phospho-histone H2A.X clone 20E3 Cell Signaling Technologies Danvers MA Cat #9718 anti 53BP1 Cell Signaling Technologies Danvers MA Cat #4937 anti-α-actinin clone BM-75.2 Sigma-Aldrich St. Louis MO Cat #A5044 anti-tropomyosin clone TM311 Sigma-Aldrich St. Louis MO Cat #T2780 anti-talin clone 8D4 Sigma-Aldrich St. Louis MO Cat #T3287 anti-vinculin clone VIN-11-5 Sigma-Aldrich St. Louis MO Cat #V4505 anti-gamma-tubulin clone GTU-88 Sigma-Aldrich St. Louis MO Cat #T5326 anti-lamin A/C clone 4C11 Sigma-Aldrich St. Louis MO Cat #SAB4200236 anti-vimentin clone V9 Sigma-Aldrich St. Louis MO Cat # V6389 anti-fibronectin Sigma-Aldrich St. Louis MO Cat #F3648 anti-tubulin clone DM1A Abcam Cambridge, UK Cat #ab7291 anti-zyxin gift of Dr. M Beckerle, Huntsman Cancer Institute, University of Utah, Salt Lake City, UT anti-myosin-IIB BioLegend/Covance Dedham MA Cat # PRB-445P donkey anti-rabbit Alexa 647 Jackson ImmunoResearch Laboratories West Grove PA Cat # 711-605-152 donkey anti-rabbit Cy2 Jackson ImmunoResearch Laboratories West Grove PA Cat # 711-225-152 donkey anti-rabbit Cy3 Jackson ImmunoResearch Laboratories West Grove PA Cat # 711-165-152 donkey anti-mouse Alexa 647 Jackson ImmunoResearch Laboratories West Grove PA Cat # 715-605-150 donkey anti-mouse Cy2 Jackson ImmunoResearch Laboratories West Grove PA Cat # 715-225-150 donkey anti-mouse Cy3 Jackson ImmunoResearch Laboratories West Grove PA Cat # 715-165-150 Chemicals, Peptides, and Recombinant Proteins Human Plasma Fibronectin EMD Millipore Billerica MA Cat# FC010 Blebbistatin Toronto Research Chemicals Cat# B592500 Latrunculin A Sigma-Aldrich St. Louis MO Cat# L5163-100UG Alexa-488 phalloidin Invitrogen Carlsbad CA Cat # A12379 Alexa 560 phalloidin Invitrogen Carlsbad CA Cat# A12380 Alexa-647-labeled fibronectin Lab of K Yamada, NICDR, NIH, Bethesda MD. Pyrene-labeled actin Hypermol, Bielefeld, Germany Cat# 8102-01 ATR inhibitor AZ-20 Tocris Minneapolis MN Cat# 5198 ATM inhibitor KU-55933 SelleckChem Houston, TX Cat# S1092 BL21(DE3)-Rosetta2 E. coli cells EMD Millipore Billerica MA Cat# 70954-3 Rat tail collagen type I Corning Bedford MA Cat# 354236 4′,6-Diamidino-2-Phenylindole (DAPI) Thermo Scientific Grand Island NY Cat# D1306 Polydimethylsiloxane (PDMS) #184 Krayden Inc Westminster CO Cat# DC4019862 DAKO fluorescent mounting media Agilent Technologies Carpinteria CA Cat# S302380-2 Oxyrase Oxyrase Mansfield OH Cat# EC-0500 Wheat-germ agglutinin Alex Fluor 488 conjugate Thermo Fisher Scientific Waltham MA Cat# W11261 Phusion High Fidelity DNA polymerase New England Biolabs, Ipswitch MA Cat# M0530 Complete Ultra protease inhibitor Roche Life Sciences, Indianapolis, IN Cat# 58927991001 Critical Commercial Assays Twenty-four well QCM Chemotaxis Cell Migration Assay, 8 um pores EMD Millipore Billerica MA Cat# ECM509 Twenty-four well Chemotaxis Cell Migration Assay, 12 um pores Cell BioLabs San Diego CA Cat# CBA-108 BioCoat Matrigel Invasion Chamber Corning Bedford MA Cat# 354480 Kit V transfection solution Amaxa Nucleofector, Lonza, Walkersville MD Cat# VCA-1003 QuikChange XL-II site-ditected mutagenesis kit Agilent Technologies, Santa Cruz, CA Cat# 200521 Deposited Data Experimental Models: Cell Lines Human foreskin fibroblasts ATCC Manassas VA ATCC® SCRC-1041 Human umbilical vein endothelial cells ATCC Manassas VA ATCC® PCS-100-010 Caco2 cells Gift of J. Anderson ATCC® HTB-37 MDCK cells Gift of J. Anderson ATCC® CRL-2936 B16-F10 Melanoma cells ATCC Manassas VA ATCC® CRL-6475 Primary mouse embryonic fibroblasts This paper Experimental Models: Organisms/Strains C57J/BL6 mice Jackson Labs Recombinant DNA pET His6 StrepII TEV LIC cloning vector (2HR-T) Addgene Cambridge MA Cat # 29718 FMN2-GFP GeneCopoeia Rockville MD EX-Mm31870-M98 mEGFP-N1 Addgene Cambridge MA Plasmid #54767 mApple-Actin Addgene Cambridge MA Plasmid #54862 EGFP-Actin Addgene Cambridge MA Plasmid #56421 EGFP-Paxillin Addgene Cambridge MA Plasmid #15233 mCardinal-H2B Addgene Cambridge MA Plasmid #56162 EGFP-H2B Addgene Cambridge MA Plasmid #56436 mCherry-H2B Addgene Cambridge MA Plasmid #55056 mCherry-α-actinin Addgene Cambridge MA Plasmid #54975 NLS-mEmerald Addgene Cambridge MA Plasmid #54206 mCherry-Lamin B Addgene Cambridge MA Plasmid #55069 Myosin light chain-GFP Addgene Cambridge MA Plasmid #56282 pSpCas9(BB)-2A-PuroV2.0 Addgene Cambridge MA Plasmid#62988 Sequence-Based Reagents Primer for FMN2 I→A Mutant Forward 5′-aatgcagactagacattagagctcctactgcttgtgaccttttgttg-3′ This paper; synthesized by Eurofins Operon Primer for FMN2 I→A Mutant Reverse 5′-caacaaaaggtcacaagcagtaggagctctaatgtctagtctgcatt-3′ This paper; synthesized by Eurofins Operon ON-TARGETplus SMARTpool siRNA against FMN2 GE Healthcare Dharmacon Lafayette, CO Cat # L-048687-01-0005 3′UTR siRNA against FMN2 GE Healthcare Dharmacon Lafayette, CO Cat# J-048687-12-0002 Primer for FMN2 FH2 domain amplification Forward 5′ TACTTCCAATCCAATGCAgctaggaagcagctgatcgagcc 3′ This paper; synthesized by Eurofins Operon Primer for FMN2 FH2 domain amplification Reverse5′ TTATCCACTTCCAATGTTATTAttatttaaagtcagagctgaa This paper; synthesized by Eurofins Operon Oligo Set 1 for targeting genomic FMN2 5′ CACCGTTTTGTGCGTAGATCCTCGA 3′ 5′ AAACTCGAGGATCTACGCACAAAAC 3′ This paper; synthesized by Eurofins Operon Oligo Set 2 targeting genomic FMN2 5′ CACCGGCAACTGTAATTCAGCAAC 3′ 5′ AAACGTTGCTGAATTACAGTTGCC 3′ This paper; synthesized by Eurofins Operon Software and Algorithms MetaMorph imaging software Molecular Devices Sunnyvale CA ImageJ NIH Bethesda MD FibrilTool Hamant Lab (82) NDP.View2 Hamamatsu Bridgewater NJ MatLab MathWorks, Natick, MA Other Highlights The formin FMN2 generates a perinuclear actin/adhesion system FMN2 regulates nuclear position during 2D migration FMN2 protects against DNA damage and cell death during invasive 3D migration FMN2 promotes metastasis of melanoma cells to the lung Author Contributions CTS designed and performed experiments, generated reagents, analyzed data and wrote the paper. RSF designed and performed experiments. PG, HRT, and MAB generated reagents and performed experiments. AT performed experiments and provided technical advice. MWD generated reagents. MP, GMA, AN, and PSS provided technical expertise and interpretation. CMW designed experiments, interpreted data and wrote the paper. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. Almuqbil M 2013 De novo deletion of FMN2 in a girl with mild non-syndromic intellectual disability Araujo AN 2014 Genome-wide copy number analysis in a family with p.G533C RET mutation and medullary thyroid carcinoma identified regions potentially associated with a higher predisposition to lymph node metastasis The Journal of clinical endocrinology and metabolism 99 6 E1104 12 24601688 Awasthi P Foiani M Kumar A 2015 ATM and ATR signaling at a glance Journal of cell science 128 23 4255 62 26567218 Baarlink C Wang H Grosse R 2013 Nuclear actin network assembly by formins regulates the SRF coactivator MAL Science (New York, NY) 340 6134 864 7 Bekker-Jensen S 2006 Spatial organization of the mammalian genome surveillance machinery in response to DNA strand breaks The Journal of cell biology 173 2 195 206 16618811 Belin BJ Lee T Mullins RD 2015 DNA damage induces nuclear actin filament assembly by formin-2 and Spire-1/2 that promotes efficient DNA repair eLife 4 Bissell MJ Hines WC 2011 Why don’t we get more cancer? A proposed role of the microenvironment in restraining cancer progression Nature medicine 17 3 320 9 Borges HL Linden R Wang JYJ 2008 DNA damage-induced cell death: lessons from the central nervous system Cell research 18 1 17 26 18087290 Boudaoud A 2014 FibrilTool, an ImageJ plug-in to quantify fibrillar structures in raw microscopy images Nature Protocols 9 2 457 463 24481272 Broers JLV 2004 Decreased mechanical stiffness in LMNA−/− cells is caused by defective nucleo-cytoskeletal integrity: implications for the development of laminopathies Human Molecular Genetics 13 21 2567 2580 15367494 Butcher DT Alliston T Weaver VM 2009 A tense situation: forcing tumour progression Nature reviews. Cancer 9 2 108 22 19165226 Campellone KG Welch MD 2010 A nucleator arms race: cellular control of actin assembly Nature reviews. Molecular cell biology 11 4 237 51 20237478 Celià-Terrassa T Kang Y 2016 Distinctive properties of metastasis-initiating cells Genes & development 30 8 892 908 27083997 Chang C Werb Z 2001 The many faces of metalloproteases: cell growth, invasion, angiogenesis and metastasis Trends in Cell Biology 11 11 S37 S43 11684441 Charfi C 2011 Gene profiling of Graffi murine leukemia virus-induced lymphoid leukemias: identification of leukemia markers and Fmn2 as a potential oncogene Blood 117 6 1899 910 21135260 Clark AG Vignjevic DM 2015 Modes of cancer cell invasion and the role of the microenvironment Current Opinion in Cell Biology 36 13 22 26183445 Davidson PM 2014 Nuclear deformability constitutes a rate-limiting step during cell migration in 3-D environments Cellular and molecular bioengineering 7 3 293 306 25436017 Denais CM 2016 Nuclear envelope rupture and repair during cancer cell migration Science (New York, NY) 352 6283 353 8 Faure-André G 2008 Regulation of dendritic cell migration by CD74, the MHC class II-associated invariant chain Science (New York, NY) 322 5908 1705 10 Fidler IJ 1973 Selection of successive tumour lines for metastasis Nature: New biology 242 118 148 9 4512654 Fischer RS 2009 Local cortical tension by myosin II guides 3D endothelial cell branching Current biology : CB 19 3 260 5 19185493 Friedl P Wolf K Lammerding J 2011 Nuclear mechanics during cell migration Current opinion in cell biology 23 1 55 64 21109415 Gomes ER Jani S Gundersen GG 2005 Nuclear movement regulated by Cdc42, MRCK, myosin, and actin flow establishes MTOC polarization in migrating cells Cell 121 3 451 63 15882626 Gruel N 2014 Polarity gene alterations in pure invasive micropapillary carcinomas of the breast Breast cancer research : BCR 16 3 R46 24887297 Harada T 2014 Nuclear lamin stiffness is a barrier to 3D migration, but softness can limit survival The Journal of cell biology 204 5 669 82 24567359 Hatch EM 2013 Catastrophic nuclear envelope collapse in cancer cell micronuclei Cell 154 1 47 60 23827674 Heuzé ML 2011 Cell migration in confinement: a micro-channel-based assay Methods in molecular biology (Clifton, NJ) 769 415 34 Hotulainen P Lappalainen P 2006 Stress fibers are generated by two distinct actin assembly mechanisms in motile cells The Journal of cell biology 173 3 383 94 16651381 Kim LY 2015 The structural basis of actin organization by vinculin and metavinculin Journal of molecular biology Koshkina NV Yang G Kleinerman ES 2013 Inhibition of Cdc42-interacting protein 4 (CIP4) impairs osteosarcoma tumor progression Current cancer drug targets 13 1 48 56 22920438 Kumar A 2014 ATR mediates a checkpoint at the nuclear envelope in response to mechanical stress Cell 158 3 633 46 25083873 Kuo J-C 2011 Analysis of the myosin-II-responsive focal adhesion proteome reveals a role for β-Pix in negative regulation of focal adhesion maturation Nature cell biology 13 4 383 93 21423176 Law R 2014 Biallelic Truncating Mutations in FMN2, Encoding the Actin-Regulatory Protein Formin 2, Cause Nonsyndromic Autosomal-Recessive Intellectual Disability Leader B 2002 Formin-2, polyploidy, hypofertility and positioning of the meiotic spindle in mouse oocytes Nature cell biology 4 12 921 8 12447394 Leader B Leder P 2000 Formin-2, a novel formin homology protein of the cappuccino subfamily, is highly expressed in the developing and adult central nervous system Mechanisms of development 93 1–2 221 31 10781961 Liu P 2012 Genome-wide association and fine mapping of genetic loci predisposing to colon carcinogenesis in mice Molecular cancer research : MCR 10 1 66 74 22127497 Lombardi ML 2011 The interaction between nesprins and sun proteins at the nuclear envelope is critical for force transmission between the nucleus and cytoskeleton The Journal of biological chemistry 286 30 26743 53 21652697 Lombardi ML Lammerding J 2011 Keeping the LINC: the importance of nucleocytoskeletal coupling in intracellular force transmission and cellular function Biochemical Society transactions 39 6 1729 34 22103516 Lottersberger F 2015 53BP1 and the LINC Complex Promote Microtubule-Dependent DSB Mobility and DNA Repair Cell 163 4 880 893 26544937 Lukas J Lukas C Bartek J 2011 More than just a focus: The chromatin response to DNA damage and its role in genome integrity maintenance Nature cell biology 13 10 1161 9 21968989 Lynch J 2013 Metastasis suppressor microRNA-335 targets the formin family of actin nucleators PloS one 8 11 e78428 24223803 Maciejowski J 2015 Chromothripsis and Kataegis Induced by Telomere Crisis Cell 163 7 1641 1654 26687355 Maniotis AJ Chen CS Ingber DE 1997 Demonstration of mechanical connections between integrins, cytoskeletal filaments, and nucleoplasm that stabilize nuclear structure Proceedings of the National Academy of Sciences of the United States of America 94 3 849 54 9023345 Montaville P 2014 Spire and Formin 2 synergize and antagonize in regulating actin assembly in meiosis by a ping-pong mechanism PLoS biology 12 2 e1001795 24586110 Pasapera AM 2010 Myosin II activity regulates vinculin recruitment to focal adhesions through FAK-mediated paxillin phosphorylation The Journal of cell biology 188 6 877 90 20308429 Raab M 2016 ESCRT III repairs nuclear envelope ruptures during cell migration to limit DNA damage and cell death Science 352 6283 e1002904 e1002904 Ramabhadran V Gurel PS Higgs HN 2012 Mutations to the Formin Homology 2 Domain of INF2 Protein Have Unexpected Effects on Actin Polymerization and Severing Journal of Biological Chemistry 287 41 34234 34245 22879592 Ramabhadran V Gurel PS Higgs HN 2012 Mutations to the formin homology 2 domain of INF2 protein have unexpected effects on actin polymerization and severing The Journal of biological chemistry 287 41 34234 45 22879592 Raub CB 2007 Noninvasive assessment of collagen gel microstructure and mechanics using multiphoton microscopy Biophysical journal 92 6 2212 22 17172303 Revach O-Y 2015 Mechanical interplay between invadopodia and the nucleus in cultured cancer cells Scientific reports 5 9466 25820462 Ridley AJ 2003 Cell migration: integrating signals from front to back Science (New York, NY) 302 5651 1704 9 Rogakou EP 1998 DNA Double-stranded Breaks Induce Histone H2AX Phosphorylation on Serine 139 Journal of Biological Chemistry 273 10 5858 5868 9488723 Schönichen A Geyer M 2010 Fifteen formins for an actin filament: a molecular view on the regulation of human formins Biochimica et biophysica acta 1803 2 152 63 20102729 Schramek D 2014 Direct in vivo RNAi screen unveils myosin IIa as a tumor suppressor of squamous cell carcinomas Science (New York, NY) 343 6168 309 13 Schultz LB 2000 p53 binding protein 1 (53BP1) is an early participant in the cellular response to DNA double-strand breaks The Journal of cell biology 151 7 1381 90 11134068 Skau CT 2015 Inverted formin 2 in focal adhesions promotes dorsal stress fiber and fibrillar adhesion formation to drive extracellular matrix assembly Proceedings of the National Academy of Sciences of the United States of America Skau CT Waterman CM 2015 Specification of Architecture and Function of Actin Structures by Actin Nucleation Factors Annual review of biophysics 44 285 310 Solier S Pommier Y 2014 The nuclear γ-H2AX apoptotic ring: implications for cancers and autoimmune diseases Cellular and molecular life sciences : CMLS 71 12 2289 97 24448903 Svitkina TM 1999 Arp2/3 Complex and Actin Depolymerizing Factor/Cofilin in Dendritic Organization and Treadmilling of Actin Filament Array in Lamellipodia The Journal of Cell Biology 145 5 1009 1026 10352018 Szczepanowska J Korn ED Brzeska H 2006 Activation of myosin in HeLa cells causes redistribution of focal adhesions and F-actin from cell center to cell periphery Cell motility and the cytoskeleton 63 6 356 74 16607629 Thiam H-R 2016 Perinuclear Arp2/3-driven actin polymerization enables nuclear deformation to facilitate cell migration through complex environments Nature communications 7 10997 Thievessen I 2013 Vinculin-actin interaction couples actin retrograde flow to focal adhesions, but is dispensable for focal adhesion growth The Journal of cell biology 202 1 163 77 23836933 Turajlic S Swanton C 2016 Metastasis as an evolutionary process Science (New York, NY) 352 6282 169 75 Versaevel M Grevesse T Gabriele S 2012 Spatial coordination between cell and nuclear shape within micropatterned endothelial cells Nature communications 3 671 Vizcarra CL 2011 Structure and function of the interacting domains of Spire and Fmn-family formins Proceedings of the National Academy of Sciences of the United States of America 108 29 11884 9 21730168 Wolf K 2007 Multi-step pericellular proteolysis controls the transition from individual to collective cancer cell invasion Nature Cell Biology 9 8 893 904 17618273 Yamada K 2013 Identification and functional characterization of FMN2, a regulator of the cyclin-dependent kinase inhibitor p21 Molecular cell 49 5 922 33 23375502 Yi K Rubinstein B Li R 2013 Symmetry breaking and polarity establishment during mouse oocyte maturation Philosophical transactions of the Royal Society of London. Series B, Biological sciences 368 1629 20130002 24062576 Zhang C-Z 2015 Chromothripsis from DNA damage in micronuclei Nature 522 7555 179 84 26017310

PMC005xxxxxx/PMC5135595.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 9200608 2299 Cancer Epidemiol Biomarkers Prev Cancer Epidemiol. Biomarkers Prev. Cancer epidemiology, biomarkers & prevention : a publication of the American Association for Cancer Research, cosponsored by the American Society of Preventive Oncology 1055-9965 1538-7755 27550750 5135595 10.1158/1055-9965.EPI-16-0203 NIHMS811391 Article Traditional breast cancer risk factors in Filipina Americans compared to Chinese and Japanese Americans in Los Angeles County Wu Anna H. 1 Vigen Cheryl 1 Lee Eunjung 1 Tseng Chiu-Chen 1 Butler Lesley M. 23 1 Keck School of Medicine, University of Southern California, Los Angeles, CA 90089 2 Cancer Control and Population Sciences, University of Pittsburgh Cancer Institute, Pittsburgh, PA, USA 3 Department of Epidemiology, University of Pittsburgh Graduate School of Public Health, Pittsburgh, PA, USA Corresponding Author: Anna H. Wu, Department of Preventive Medicine, University of Southern California Keck School of Medicine, 1441 Eastlake Avenue, Rm 4443, Los Angeles CA 90089 (anna.wu@med.usc.edu) phone: 323 865-0484; fax: 323-865-0139 25 8 2016 22 8 2016 12 2016 01 6 2017 25 12 15721586 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. Background Filipina Americans have one of the highest breast cancer incidence rates among Asian Americans for poorly understood reasons. Methods Breast cancer risk factors were investigated in a population-based study of Filipina (790 cases, 587 controls), Japanese (543 cases, 510 controls), and Chinese (913 cases, 904 controls) Americans; cases were identified by the Los Angeles County Cancer Surveillance Program and controls were matched to cases on age, ethnicity, and neighborhood. Multivariable conditional logistic regression was performed by Asian ethnicity. Results In Filipina, Chinese, and Japanese Americans, breast cancer risk decreased significantly with increasing parity (all P trends <0.0001). Breast cancer risk increased with increasing quartiles of cumulative menstrual months in premenopausal (P trend =0.019) and postmenopausal Filipina (P trend=0.008), in premenopausal (P trend=0.0003) but not postmenopausal Chinese (P trend=0.79), and in neither premenopausal (P trend=0.092) nor postmenopausal (P trend=0.75) Japanese Americans. For postmenopausal Filipina and Japanese, greater weight gain since age 18 (P trend =0.019 and 0.053 respectively), high current body mass index (both P trend<0.01), and greater waist circumferences (both P trend <0.04) were statistically significant; these associations were weaker for postmenopausal Chinese women. Conclusions Cumulative menstrual months and body size factors were statistically significant risk factors for Filipina. Total menstrual months were associated with breast cancer among Chinese but not for Japanese, while body size factors were significantly associated with risk among Japanese but not among Chinese. Impact Characterization of breast cancer risk factors in Filipina will help to generate hypotheses for their high breast cancer incidence. Filipina Japanese Chinese body size menstrual and reproductive risk factors cumulative menstrual months Introduction Historically, Asian women in Asia and in the U.S. have among the lowest incidence of breast cancer worldwide. Exceptions are the relatively high breast cancer incidence rates in the Philippines and among Filipina Americans. In 2013, age-adjusted breast cancer incidence rates (per 100,000) were higher in the Philippines (87.5) than in Japan (55.0) or China (45.4) (1). Among Asians in the U.S., the highest breast cancer incidence rates during the period 2004 to 2008 were for Japanese Americans (104.9) and Filipina Americans (103.7) (2). However, compared with incidence during the period 1990 to 1994, there has been a 21% increase in Filipina women compared with a 6.2% increase in Japanese women. It is of note that most Filipina American women are recent immigrants, compared with the mostly US-born, and second or later generation Japanese American women. In one of the few analytic epidemiological studies of breast cancer among Filipina women in Manila, Philippines, the ‘classical’ risk factors (e.g., excess body weight) did not fully explain the high breast cancer incidence in Manila, especially when compared to other urban Asian populations. However, this previous study had few cases (n=123) and limited information on lifestyle factors (3). To date, there are no epidemiologic studies of breast cancer focusing on Filipina-American women. We initiated a population-based case-control study among Filipina, Japanese and Chinese American women in Los Angeles (LA) County between 1995 and 2001. Since then, in order to investigate Asian ethnicity-specific associations with breast cancer, we have expanded the database to include cases diagnosed between 2003 and 2006, and additional controls. To better understand the high breast cancer rates in Filipina Americans, we investigated traditional breast cancer risk factors in pre- and postmenopausal Filipina, Japanese, and Chinese Americans in LA County. We also calculated lifetime cumulative number of menstrual months (4, 5), as an index of total exposure to endogenous estrogens. Materials and methods Study design and population The study population and methods used in this population-based case-control study have been described previously (6, 7). In brief, breast cancer patients were identified by the LA County Cancer Surveillance Program, the population-based cancer registry covering LA County, a member of the National Cancer Institute’s Surveillance, Epidemiology, and End Results (SEER) program, and the statewide California Cancer Registry. Patients included in this analysis were women who were identified as Chinese, Japanese, or Filipina between the ages of 25 and 74 inclusive at the time of diagnosis of an incident breast cancer. Case patients were diagnosed between 1995 and 2001 or between 2003 and 2006. In total, we identified 3,797 eligible case patients (1,496 Chinese, 865 Japanese, 1,436 Filipina) and interviewed 2,303 cases (929 Chinese, 547 Japanese, and 827 Filipina) (response rate of 61%). Among those who did not participate, 869 declined to be interviewed (375 Chinese, 222 Japanese, 272 Filipina), 77 had died (17 Chinese, 24 Japanese, 36 Filipina), and 548 could not be located (175 Chinese,72 Japanese, 301 Filipina). The 2,035 control subjects (923 Chinese, 518 Japanese, and 594 Filipina) were selected from the neighborhoods where the case patients resided at the time of diagnosis. A well-established algorithm was used to identify neighborhood controls for population-based case-control studies conducted in LA County as this provides a mechanism of matching on socioeconomic status which is likely to influence various lifestyle habits (8, 9). We initially defined a specified sequence of houses to be visited in the neighborhoods where index cases lived at the time of diagnosis. We then sought to interview the first eligible resident in the sequence. If the first eligible control subject refused to participate, the second eligible one in the sequence was asked, and so on. Letters were left when no one was home, and follow-up was by mail and telephone. Controls were sought to frequency-match to the cases on specific Asian ethnicities and 5-year age groups. On average, a suitable control was identified after visiting a mean of 60 households (48.1 for Chinese, 58.0 for Japanese, and 74.6 for Filipina). Of the controls interviewed, 64% were the first identified eligible control (range was 64% for Filipina and 67% for Chinese), 18% were the second-identified eligible control, and 18% were the third or later eligible control. Data collection Cases and controls were interviewed using a standardized, structured questionnaire. Filipina and Japanese American women were interviewed in English as almost all were English-speaking whereas a Chinese-translated questionnaire was used for subjects who were not English-speaking. Interviews were conducted in Mandarin or Cantonese for 36.9% (n=337) of Chinese cases and 33.3% (n=301) of Chinese controls. To the extent possible, each case-control pair was interviewed by the same interviewer. The questionnaire covered demographic characteristics and migration history, menstrual and reproductive history, body size, physical activity, family history of breast cancer, and diet history (10). To assess menstrual history of subjects, we asked the age when they had their first menstrual period, age when their menstrual periods became established at regular intervals (i.e. there was a predictable amount of time between menstrual periods). Subjects were asked the total number of pregnancies they had. For each pregnancy, the outcome of the pregnancy (i.e., livebirth, stillbirth, induced abortion, spontaneous abortion, and tubal or ectopic pregnancy), the length of the pregnancy, the month/year when the pregnancy ended, whether the baby was breastfed, and the duration of breastfeeding of each birth were asked. Calendars were used to chart major life events, reproductive histories and hormone use. Lifetime history of hormone use (oral contraceptives (OCs) and menopausal hormones) was obtained with the aid of an album with color photographs of all preparations used in the U.S. For each episode of exogenous hormone use, age at starting and age at stopping use were asked (10). Subjects were asked about their height and usual weight history at age 18 years, at age 30 years and each decade thereafter when they were not pregnant. Relative body weight was evaluated by body mass index (BMI), calculated as the weight in kilograms divided by the square of height in meters (kg/m2). The change in weight from age 18 to current weight (i.e., before diagnosis or at interview) was calculated. We examined BMI using the recommended 5 category cut points (<22.9, 23 –24.9, 25–27.4, 27.5–29.9, ≥30 kg/m2) for studies in Asian Americans (11), which incorporated the standard WHO definition (<25, 25–29.9, and ≥30 kg/m2) of normal, overweight, and obese as well as the corresponding WHO Asian BMI definition (<22.9, 23–27.5, and ≥27.5 kg/m2). We calculated total months of menstruation as age at menarche subtracted from age at menopause for postmenopausal women and from age at interview or cancer diagnosis for premenopausal women, and then subtracted anovulatory periods due to months of complete and incomplete pregnancies, lactation, OC use, or missing periods (if 3 or more periods were missed in a row). We excluded women who started hormone therapy before menopause (112 cases, 56 controls) or had hysterectomy only (153 cases, 126 controls) because age at menopause was not known, women who did not have regular cycles (132 cases, 136 controls), and other reasons (5 cases, 7 controls). Statistical Analysis The results presented below are based on 2,190 cases (393 in situ, 1151 localized, 634 advanced stage, 12 stage unknown) and 1,983 controls for whom we have information on body size, menstrual and reproductive factors as well as the covariates included for adjustment. We excluded 56 cases (22 Chinese, 24 Japanese, 10 Filipina) and 18 control women (8 Chinese, 5 Japanese, 5 Filipina) from the analysis because they were found to have had a previous cancer (mostly cervical, uterine, or colorectal cancer) diagnosis at the time of interview. We calculated Asian ethnic-specific odds ratios (ORs; relative risk estimates) and their corresponding 95% confidence intervals (CIs) and P values by conditional logistic regression methods, with matched sets defined jointly by Asian ethnicity (Filipina, Japanese, Chinese), and reference age (<39, 40–44, 45–49, 50–54, 55–59, 60–64, 65–69, 70+ years) (10). All regression models included the following covariates: years of residence in the US (US born, >20 years, 11–20 years, ≤10 years), education (less than high school, high school, some college, college graduate), income, interviewer, family history of breast cancer (no, yes any first degree relative with breast cancer), and history of benign breast disease (no, yes). The menstrual and reproductive variables we considered included age at menarche, pregnancy (never, only incomplete pregnancy [i.e., did not result in live birth, or stillbirths at less than 28 weeks of pregnancy], complete pregnancy [i.e., live births and stillbirths at 28 weeks of pregnancy or more]), parity (no full-term pregnancy, full-term pregnancies), number of births (live births and stillbirths), age at birth (live births and stillbirths), breast feeding, duration of breast feeding, use of OCs, duration of OC use, type of menopause and age at menopause, and lifetime cumulative number of menstrual months. Menstrual and reproductive factors were mutually adjusted for. Age at menarche (≤11, 12, 13, 14, 15+ years) and parity (0, 1, 2, 3, 4+ births) were also included in analyses to examine the effects of body size factors. The body size variables we considered included height, current (i.e., before diagnosis or interview) BMI, BMI at ages 18, 30 and 40, weight change (current weight-weight at age 18), waist and hip circumferences and waist to hip ratio (WHR). For anthropometric exposures, test for trend (p values) were performed by coding each variable as a grouped (quartile) linear variable (i.e., as 1, 2, 3, and 4). In addition, we present risk associations with exposure variables of interest as a continuous variable. P values less than 5% were considered statistically significant and all P values quoted are 2-sided. All analyses were performed by using EPILOG Windows (version 1.01s) statistical software system (Pasadena, CA) and the SAS statistical software system (version 9.3; SAS Institute, Cary, NC). Results Filipina control women were more educated, were more likely to be foreign born, had spent fewer years in the U.S., and had the oldest age at immigration, compared with Chinese and Japanese control women (Table 1). Filipina controls had more births and were younger at first birth compared with the other Asian women. Filipina and Japanese women displayed similar mean current BMI, which was higher than in Chinese women. Mean age at natural menopause and ovulatory months did not differ between the three Asian American groups. These patterns were similarly observed when Filipina breast cancer cases were compared to Chinese and Japanese cases. Table 1 also shows that breast cancer cases tended to display a higher risk profile compared to control women of the same Asian race/ethnicity for almost all the variables we studied. Menstrual and reproductive factors and breast cancer risk associations did not differ by menopausal status among Asian ethnic subgroups, with the exception for age at menarche (see below), and thus ORs are presented in all women (Table 2). Women who were ever pregnant, including those who only had an incomplete pregnancy had lower risk than women who were never pregnant. Breast cancer risk decreased per increasing number of pregnancies (P values <0.001 for all three Asian groups). Having a greater number of births was inversely associated with breast cancer (OR per birth) similarly among Filipina (0.78, 95% CI 0.73–0.84), Chinese (0.82, 95% 0.76–0.89) and Japanese women (0.78, 95% CI 0.70–0.88). Age at first birth, age at last birth, and breastfeeding were not associated with risk in Filipina women or in the other Asian women. Use of OC was inversely associated with risk in all three Asian groups combined (P trend=0.040), but the trend was borderline statistically significant only in Chinese American women (P trend=0.056) and not in Japanese and Filipina women. Overall, age at menarche was not associated with risk in Filipina, Chinese or Japanese women. However, the effect of age at menarche differed between premenopausal and postmenopausal women. In premenopausal women, risk was lower with each year that menarche was delayed in Filipina (OR 0.92, 95% CI 0.83–1.03), Chinese (0.89, 95% CI 0.81–0.98) and Japanese (0.88, 95% CI 0.75–1.03) but risk increased with later age at menarche in postmenopausal women (respective P for interaction was 0.02, 0.04, and 0.08). Breast cancer risk was non-significantly higher with later age at natural menopause (OR for >54 vs ≤49) in Filipina (OR=1.71) and Japanese (OR=1.33) American women, but not in Chinese women (OR=0.87) (Table 2). To investigate whether the timing of births in relation to migration to the US might explain the unexpected null associations with age at first birth and breastfeeding in this study, we examined risk associations separately for women who had all their births prior to migration, women who had all their births after migration, and women who had one or more birth before migration as well as after migration to the US (Supplemental Table 1). The risk associations with births, age at first birth, and breastfeeding, were generally the same irrespective whether participants completed births before or after migration to the US. We also compared the results in migrants to those of US born Asian Americans (314 cases [238 Japanese, 59 Chinese, and 17 Filipina] and 413 controls [290 Japanese, 102 Chinese, and 21 Filipina]). Although the results in US born Asian Americans were not statistically significant, compared to migrants the risk association per birth was similar (0.87, 95% CI 0.74–1.03), while the increased risk per 5 years delay in age at first birth was higher (1.05, 95% CI=0.86–1.29), and the risk reduction per 6 months of breast-feeding was stronger (0.95, 95% CI=0.87–1.05). Breast cancer risk increased with increasing number of menstrual months in Filipina women (P trend=0.0008); women in the highest quartile (>399 months) had more than a 2-fold increased risk (OR=2.55, 95% CI 1.50–4.33) than those in the lowest quartile (≤283 months); this was observed in both premenopausal (P trend=0.019), postmenopausal (P trend=0.008), and parous (P trend=0.014) but not nonparous (P trend=0.24) Filipina women (Table 3). Risk patterns were similar in Chinese (P trend=0.003), statistically significant in premenopausal (P trend =0.0003) and parous women (P trend=0.017) but not in postmenopausal (P trend =0.79) or nonparous Chinese women (P trend=0.92). Results were weaker in Japanese (P trend=0.16) but there was a borderline statistically significant association in premenopausal women (P trend= 0.092). In the three Asian groups combined, there was a significant trend of increasing risk with increasing number of menstrual months in premenopausal (P trend <0.0001) but not in postmenopausal women (P trend=0.12). Risk associations were statistically significant in ever parous women (P trend=0.002) and borderline statistically significant in never parous women (P trend=0.08). The risk associations were similar when we fully adjusted for diabetes, physical activity, and dietary patterns (Table 3, footnote 4). The patterns of associations were also similar by tumor stage at diagnosis (Supplemental Table 2) and by hormone receptor status of breast cancer (Supplemental Table 3). High BMI at age 18 was inversely associated with risk of breast cancer in premenopausal Filipina women (P trend=0.03) but risk was not significantly influenced by BMI at ages 30 and 40, current BMI, weight gain since age 18, waist and hip circumferences or WHR. There were no significant associations between these body size factors and risk in premenopausal Chinese and Japanese women (Table 4). Breast cancer risk of postmenopausal Asian women was unrelated to BMI at age 18 (P trend=0.93) but was positively associated with BMI at ages 30 (P trend=0.026) and 40 (P trend=0.002); results for BMI at ages 30 and 40 were statistically significant in Filipina and Japanese women, respectively. Current BMI and weight gain since 18 were positively associated with breast cancer risk, regardless of Asian ethnicity (Table 5). High current BMI was positively associated with risk; the OR per 5 unit increase was 1.40 (95% CI 1.11–1.78) in Filipina, 1.29 (95% CI 1.02–1.62) in Japanese, and 1.20 (95% CI 0.96–1.50) in Chinese women. Weight gain since age 18 was also positively associated with risk; the OR per 5 kg increase was 1.18 (95% CI 1.07–1.30) in Filipina, 1.10 (95% CI 1.00–1.21) in Japanese, and 1.09 (95% CI 0.99–1.21) in Chinese women. Breast cancer risk in postmenopausal Filipina women increased with increasing waist circumference (P trend=0.04) but risk was unrelated to hip circumference or WHR. Both waist (P trend=0.013) and hip (P trend=0.01) circumferences, but not WHR were positively associated with risk in postmenopausal Japanese women. Risk in postmenopausal Chinese women was not associated with waist and hip circumferences or WHR (Table 5). The patterns of associations with current BMI and weight gain and risk in postmenopausal women were also similar by tumor stage at diagnosis (Supplemental Table 2). Discussion Numerous observational studies have investigated reproductive factors (e.g., age at menarche, parity, breastfeeding) and breast cancer risk although few of these studies were conducted in Asian American women (10, 12), and even fewer had adequate sample size to examine risk patterns separately in Filipina, Chinese and Japanese women. To explore reasons for the higher breast cancer rates among Filipina Americans (2), we examined the role of body size and menstrual and reproductive factors in Filipina Americans, comparing their risk profiles to those of Chinese and Japanese Americans. In addition, we investigated the influence of lifetime menstrual months as a marker of cumulative exposure to endogenous hormones. It is of interest that both total menstrual months, and body size factors (age 30 BMI, current BMI, weight gain since age 18 and waist circumferences) were statistically significant breast cancer risk factors for postmenopausal Filipina women. In contrast, cumulative number of menstrual months was a significant risk factor for Chinese, particularly premenopausal Chinese women, but not for Japanese women, while body size factors were significantly associated with risk among postmenopausal Japanese but not among postmenopausal Chinese women. Breast cancer risk was lower with later age at menarche among premenopausal Filipina, Chinese, and Japanese Americans; the ORs ranged from 0.88 to 0.92 per year delay. However this inverse relationship was not observed in postmenopausal Asian Americans. Our findings in premeno-pausal women are consistent with results from a large international pooled analysis (13), but the positive association in postmenopausal women was unexpected. There was internal consistency in our data as average age at menarche was slightly older in postmenopausal then premenopausal Filipina (13.1 vs 12.7), Chinese (13.4 vs 13.1) and Japanese (13.1 vs 12.4) women. Early menarche (i.e., ≤12 years) was also more common for Filipina (42%), Japanese (56%) and Chinese (35%) in LA County than their Asia counterparts in the Philippines (26%)(3), Japan (7% –22%) (14, 15) and China (9%) (16). Although studies in China (16, 17) and a meta-analysis of case-control studies from Japan conducted prior to 1995 have found a significant risk reduction with later age at menarche (18), recent cohort studies from Japan have reported mixed results. No significant associations were found in either the Japan Collaborative Cohort Study (14) or the Miyagi (19) cohorts whereas in the Japan Public Health Center cohort, a significant protective association of late menarche was found in premenopausal women but not in postmenopausal women (20). Secular changes in age at menarche in the U.S., and in Asia, and the influence of increasing body size in U.S. adolescent girls, as well as small sample sizes, and recall biases may have contributed to some of the inconsistencies in the different studies. Risk reduction with increasing parity was comparable for Filipina, Chinese, and Japanese women in LA County (OR per birth ranged from 0.78 to 0.82). Nulliparity was most prevalent in Japanese (24.5%), intermediate in Filipina (15.2%), and lowest for Chinese (13.5%) in LA County. Interestingly, the prevalence of nulliparity among U.S. Filipina control women is similar to that in the Philippines (13.5%) (3), whereas the prevalence of nulliparity was higher for Chinese and Japanese in LA County than their respective counterparts in China (3–5%) (16, 21) or Japan (2–10%) (15, 19, 20). Compared with age at first birth ≤24 years, older age was associated with an increased risk in Filipina, Chinese, and Japanese women combined (ORs were 1.17, 1.13, and 1.11, P trend=0.36 with age at first birth 25–29, 30–35, and >35, respectively). These findings on age at first birth in Asian Americans are weaker than results from Asia (3, 14, 16, 20) and western populations (22). This discrepancy may be due to the few (9.7% Filipina, 3.9% Japanese, 3.2% Chinese) Asian American control women with young age at first birth (i.e., <20 years) (data not shown) that typically comprise the reference group, compared to data in the Philippines (22%) (3), Japan (8%)(23), and China (5%) (16). Breastfeeding was not associated with breast cancer risk in the three Asian American groups. This is in contrast to an international pooled analysis which found a statistically significant 4.3% reduction in breast cancer risk for every 12 months of breastfeeding (24). Less than half of parous Filipina (48%), Chinese (48%), and Japanese (48%) in LA County reported breastfeeding and the average duration of breastfeeding was less than one year. Breastfeeding was not included in the previous case-control study from the Philippines (3) but was reported by approximately 80% of parous women in Japan and China (16, 19–21). In studies conducted in the 1970s and 1980s, long duration of breastfeeding (i.e., >3 years) had a significant protective effect of breast cancer risk in urban Shanghai and Tianjin, China (17, 25, 26). There was no significant association between risk and duration of breastfeeding in more recent studies conducted in Shanghai, China (16, 21). The highest category of breastfeeding (≥24 months) was reported by 17% of control women in Shanghai in the 1990s (16) compared to 41% who reported breastfeeding 3 or more years in Shanghai in the mid-1980s (17). Results on breastfeeding and breast cancer risk in studies from Japan are mixed (27); no significant association was found in recent cohort studies but information on duration of breastfeeding was not provided (19, 20). Our results showed that the risk associations with births, age at first birth, and breastfeeding, were generally the same irrespective whether participants completed births before or after migration to the US (Supplemental Table 1). Much larger studies will be needed to better understand the complex interplay between age at migration and its corresponding influences on number and timing of pregnancy as well as breastfeeding practices in different Asian groups. We explored whether risk of breast cancer in Asian Americans was associated with cumulative menstrual months, a surrogate of total endogenous exposure to estrogen (4, 5). Our data provided clear evidence that greater duration of menstrual months was positively associated with risk of breast cancer; this was particularly strong in all three groups of premenopausal Asian women (OR per 50 months was 1.18, 95% CI=1.09–1.29). In pre- and postmenopausal women combined, the association was strongest in Filipina (OR per 50 months=1.22, 95% CI 1.09–1.36), intermediate in Chinese (OR per 50 months=1.16, 95% CI 1.05–1.28), and weakest in Japanese (OR=1.07, 95% CI 0.96–1.19) women. These novel findings in premenopausal women and in Filipina are supportive of the findings on lifetime cumulative number of menstrual cycles and risk in western populations (4, 5). Duration of menstruation was associated with breast cancer in Shanghai, China (21) and among postmenopausal women in Japan (19), although it is not clear whether duration of breastfeeding, pregnancy and months of OC use were excluded from these analyses. To estimate the number of menstrual months, we used self-reported information on duration of each pregnancy, months of breast feeding, months of OC use, and age at menopause (for postmenopausal women only). However, we did not calculate number of ovulatory cycles because this would require information on duration of menstrual cycles and assumptions about absence of cycles during lactation, other outcomes of pregnancy, and OC use. Although OC use is typically associated with a small increased risk of breast cancer in studies conducted in western populations (28), in this and a previous case-control study of Asian Americans we conducted in the 1980s (29), use of OCs was inversely associated with risk in combined analyses of Asian American women, but did not achieve statistical significance in Asian ethnic specific analysis. No significant association between OC use and breast cancer risk was reported in recent studies in China, Japan and the Philippines (3, 19, 30). The null findings in Asian populations may be due in part to the low prevalence of OC use and the relatively short duration of use in these study populations, compared with Asian Americans and western populations. We reported few significant associations between body size factors and risk in premenopausal Filipina, Chinese, and Japanese women. This is in contrast to the positive association between current BMI and risk in premenopausal women in Shanghai, China (31, 32), Japan (23), and the Philippines (3) and the significant inverse association between recent BMI and risk in premenopausal Caucasian U.S. women (33, 34). The inconsistent direction in the association between current BMI and risk in premenopausal Asian vs. Caucasian women is not well-understood (34). Anovulation as a result of obesity (i.e., BMI ≥30 kg/m2) has been hypothesized to explain the inverse association in Caucasian women (35). The extent to which this applies to Asian populations is not known but the prevalence of obesity (i.e., BMI ≥30 kg/m2) is still low in Asian Americans; reported by 1.3% of Chinese, 4.7% of Filipina, and 6.8% of Japanese control women in LA County, compared to about 15% of premenopausal Caucasian women in studies that reported inverse associations (36, 37). In addition, BMI may not be the most appropriate measure of obesity in Asian Americans (38, 39). Previous studies on weight at age 20 and weight changes in relation to risk in premenopausal Asians are sparse and conflicting (32, 40). The body size effects on risk among postmenopausal Filipina Americans in LA County (Table 4) were as strong as the risk associations observed among postmenopausal Japanese Americans in the Multiethnic Cohort (MEC) which showed HRs of 1.07 (95% CI 1.03–1.12) per 5 kg increase of weight gain, and 1.15 (95% CI 1.04–1.27) per 5 kg/m2 increase in current BMI (41). Although BMI was not a risk factor for postmenopausal women in the Philippines, these results were based on 36 postmenopausal cases and 274 control women and the analysis was limited to dichotomizing BMI as above or below 25 kg/m2 (3). Current BMI and weight gain since age 20 were significant risk factors for postmenopausal women in China (32) and Japan (23, 42). While high BMI (>30 kg/m2) was more prevalent in LA County Asian Americans (7.0% of Japanese, 4.5% of Filipina, and 4.0% of Chinese women) than in Japan (2.8%) (23), it is still much lower than in Whites (16%) (37). Weight gain of 9 kg or more since age 18 was reported by 65% of Filipina, 49% of Japanese, and 44% of Chinese women in LA County, comparable to results in China (51%) (32) but higher than in Japan (33%) (43). Our findings of high weight gain/obesity in Filipina are consistent with previous studies (44), showing that Filipina tend to be at higher risk of obesity than other Asian American groups. The high prevalence of obesity in Japanese American women is also consistent with the trends of increasing obesity with succeeding generations of Asians in the US (45). Some strengths and limitations of our study should be noted. The overall participation rate was modest (61% among cases and 64% among controls). Although our participation rate is not unlike that reported in other population-based studies conducted during this time period, 14% of the identified cases had moved outside of LA County (21% for Filipina, 12% for Chinese, and 8% for Japanese). The non-interviewed breast cancer patients tended to be older than those who were interviewed, but the interviewed and non-interviewed group did not differ in terms of neighborhood socioeconomic status or tumor stage at diagnosis (Supplemental Table 4). While we cannot rule out the possibility of recall bias in this case-control study, there was internal consistency in our data such as younger average age at menarche among US born than non-US born Asian Americans. In addition, we had the advantage of using a lifetime calendar approach which allowed us to collect detailed information on relevant menstrual and reproductive events that were used to calculate the cumulative index of menstrual months (see Methods). However, we did not have information on cycle length and assumed absence of ovulation during the duration of breastfeeding. Because a common questionnaire and study protocol was used in the three Asian American ethnic groups, we were able to compare risk estimates in Filipina, Chinese, and Japanese Americans as well as prevalence of risk factors. We recognize the challenges of comparing risk factor patterns between different studies conducted in Asia to those of Filipina, Japanese, and Chinese Americans because of cohort effects as well as economic and westernization effects on body size, and menstrual and reproductive factors. To the extent possible, we selected as comparison larger population-based studies in Asia that were conducted during comparable time frames as our study in LA County. In summary, the present analysis represents a first attempt to examine reasons for the high breast cancer incidence in Filipina women compared to other Asian groups. Both cumulative menstrual months and body size factors were important risk factors for LA County Filipina women, while only menstrual months was a risk factor for Chinese and only body size factors were related to risk for Japanese women. The prevalence of weight gain since age 18 (i.e., at least 9 kg) was also higher in postmenopausal Filipina than postmenopausal Chinese and Japanese women. Thus, these factors may explain, in part, the higher incidence of breast cancer among Filipina women. However, this analysis only covered the well-established breast cancer risk factors related to menstrual and reproductive events and body size characteristics. Future analysis will need to examine the role of diet, comorbidities, physical activity, and other lifestyle factors to obtain a more complete understanding of the reasons for the high breast cancer incidence among Filipina women compared to other women in Asia as well as other Asian Americans in the US. Supplementary Material 1 We are grateful to all the study participants for their contributions and support. We thank the entire data collection team, especially Annie Fung and June Yashiki. Incident breast cancer cases for this study were collected by the USC Cancer Surveillance Program (CSP), which is supported under subcontract by the California Cancer Surveillance Program (CSP), which is supported under subcontract by the California Department of Health. The CSP is also part of the National Cancer Institute’s Division of Cancer Prevention and Control Surveillance, Epidemiology, and End Results Program, under contract number N01CN25403. The ideas and opinions expressed herein are those of the authors, and endorsement by the State of California, the California Department of Health Services, or the National Cancer Institute is not intended nor should be inferred. Grant Support: The Asian American Breast Cancer Study was supported by the California Breast Research Program grants 1RB-0287, 3PB-0120, 5PB-0018, 10PB-0098 (Wu AH, Tseng CC). This work is also supported by USC Norris Comprehensive Cancer Center Core Support grant (P30 CA14089) (Wu) and a NIEHS grant (2P30 ES007048021) (Wu). Table 1 Demographic, and menstrual and reproductive factors of Filipina, Chinese, and Japanese control women in Los Angeles County Controls P value1 Cases P value1 Cases vs controls Chinese Japanese Filipina Chinese Japanese Filipina Ch Jp Fi Fi vs. Ch Fi vs. Jp Ch vs Jp Fi vs. Ch Fi vs. Jp Ch vs Jp N 896 505 582 891 519 780 Mean age (SD) 49.3 (9.6) 52.8 (11.3) 52.0 (10.4) <0.01 0.19 <0.01 52.2 (9.9) 55.8 (11.1) 53.0 (10.0) 0.12 <0.01 <0.01 <0.01 <0.01 0.07 Education level, %   Less than high   school 23.2 13.9 9.3 28.7 17.3 8.6   High school/some   college 19.4 34.1 17.0 18.2 34.9 14.2   College graduate 37.1 35.0 59.3 <0.01 <0.01 <0.01 35.0 32.4 62.8 <0.01 <0.01 <0.01 0.06 0.39 0.47   Graduate 20.3 17.0 14.4 18.1 15.4 14.4 Immigration history   U.S. born, % 14.1 73.9 4.6 <0.01 <0.01 <0.01 9.8 67.2 2.8 <0.01 <0.01 <0.01 <0.01 0.02 0.10   Mean age at   migration (SD) 30.9 (12.6) 24.1 (8.4) 33.6 (13.1) <0.01 <0.01 <0.01 32.0 (13.0) 24.1 (9.2) 33.5 (12.1) 0.02 <0.01 <0.01 0.08 0.95 0.89   Mean years in U.S. 18.4 (11.4) 25.2 (12.8) 18.5 (10.6) 0.90 <0.01 <0.01 20.0 (11.1) 28.4 (12.3) 19.5 (10.4) 0.34 <0.01 <0.01 <0.01 0.03 0.09 Mean current BMI (SD) 22.1 (3.1) 23.3 (4.0) 23.4 (3.4) <0.01 0.51 <0.01 22.4 (3.3) 23.4 (4.0) 23.9 (3.7) <0.01 0.01 <0.01 0.04 0.83 0.02 Mean age at menarche (SD) 13.2 (1.6) 12.5 (1.6) 12.9 (1.7) <0.01 <0.01 <0.01 13.2 (1.6) 12.6 (1.7) 13.0 (1.8) 0.03 <0.01 <0.01 0.81 0.09 0.34 Nulliparous, % 13.6 24.4 15.1 0.46 <0.01 <0.01 19.4 31.0 23.2 0.07 <0.01 <0.01 <0.01 0.02 <0.01 Mean # of births2 (SD) 2.2 (1.1) 2.3 (1.1) 3.1 (1.9) <0.01 <0.01 0.20 2.2 (1.1) 2.2 (0.9) 2.6 (1.6) <0.01 <0.01 0.35 0.87 0.05 <0.01 Mean age at first birth2 (SD) 27.1 (4.8) 27.4 (5.1) 26.4 (5.5) <0.01 <0.01 0.41 27.0 (4.8) 27.3 (4.9) 27.3 (5.0) 0.36 0.96 0.44 0.69 0.77 <0.01 Mean months of breastfeed (SD) 5.9 (10.6) 7.7 (13.1) 8.9 (19.8) <0.01 0.31 <0.01 6.1 (11.9) 5.3 (9.9) 6.7 (14.0) 0.39 0.10 0.29 0.69 <0.01 0.03 Postmenopausal,% 40.3 53.1 52.9 <0.01 0.99 <0.01 51.2 62.6 57.6 0.01 0.08 <0.01 <0.01 <0.01 0.10 Mean age at natural menopause (SD) 49.6 (4.2) 49.9 (4.2) 49.2 (4.0) 0.24 0.06 0.42 50.0 (3.7) 50.7 (4.1) 49.6 (4.0) 0.19 <0.01 0.04 0.24 0.08 0.22 Mean menstrual months3 (SD) 339.6 (82.7) 331.3 (105.0) 342.4 (83.2) 0.56 0.08 0.14 362.9 (76.3) 361.0 (98.0) 365.4 (79.5) 0.55 0.42 0.71 <0.01 <0.01 <0.01 Family history breast cancer,% 7.1 8.9 9.3 0.17 0.92 0.28 13.7 15.6 12.8 0.65 0.18 0.36 <0.01 <0.01 0.05 Had breast cyst,% 12.8 15.8 15.5 0.18 0.93 0.14 20.3 16.0 23.5 0.13 <0.01 0.05 <0.01 0.98 <0.01 1 P values in controls: Fi vs Ch (Filipina vs Chinese); Fi vs Jp (Filipina vs Japanese), and Ch vs Jp (Chinese vs Japanese); 2 Births included live births and stillbirths 3 Based on 757 Chinese, 410 Japanese, and 489 Filipina controls and 746 Chinese, 403 Japanese, and 641 Filipina cases Table 2 Menstrual and reproductive factors and risk of breast cancer in Filipina, Chinese and Japanese Americans in Los Angeles County Chinese Japanese Filipina All Ca Co OR1 (95%CI) Ca Co OR1 (95%CI) Ca Co OR1 (95%CI) OR2 (95%CI) Pregnancy Never 143 87 1.0 113 82 1.0 159 66 1.00 1.00 Incomplete only 30 35 0.60 (0.33–1.08) 48 41 0.89 (0.52–1.53) 22 22 0.45 (0.22–0.90) 0.69 (0.49– 0.96) 1+ livebirth 718 774 0.49 (0.36–0.66) 358 382 0.57 (0.40–0.82) 599 494 0.49 (0.35–0.69) 0.54 (0.45– 0.64) Pregnancy   0 143 87 1.00 113 82 1.00 159 66 1.00 1.00   1 86 98 0.55 (0.37–0.83) 78 64 0.76 (0.48–1.21) 92 64 0.62 (0.39–0.99) 0.67 (0.52– 0.86)   2 207 220 0.54 (0.38–0.76) 128 141 0.58 (0.39–0.87) 167 106 0.69 (0.46–1.04) 0.60 (0.48– 0.74)   3 195 204 0.50 (0.35–0.71) 106 115 0.55 (0.36–0.84) 145 118 0.48 (0.32–0.72) 0.53 (0.43– 0.66)   ≥4 260 287 0.43 (0.30–0.60) 94 103 0.59 (0.38–0.92) 217 228 0.37 (0.25–0.53) 0.45 (0.37– 0.56) P trend <0.0001 0.008 <0.0001 <0.0001 Per pregnancy 0.84 (0.78–0.91) 0.89 (0.82–0.97) 0.84 (0.79–0.89) 0.87 (0.84– 0.9) P value <0.0001 0.008 <0.0001 <0.0001 Births 3   0 173 122 1.00 161 123 1.00 181 88 1.00 1.00   1 171 185 0.62 (0.44–0.86) 87 71 0.79 (0.52–1.20) 133 76 0.89 (0.59–1.34) 0.74 (0.60– 0.91)   2 336 365 0.57 (0.42–0.76) 154 181 0.57 (0.40–0.81) 222 139 0.80 (0.56–1.16) 0.64 (0.53– 0.77)   3 137 141 0.48 (0.33–0.69) 92 90 0.56 (0.37–0.86) 116 133 0.40 (0.27–0.59) 0.49 (0.39– 0.61)   ≥4 74 83 0.35 (0.22–0.55) 25 40 0.29 (0.15–0.53) 128 146 0.32 (0.22–0.48) 0.34 (0.26– 0.44) P trend <0.0001 <0.0001 <0.0001 <0.0001 Per birth 0.82 (0.76–0.89) 0.78 (0.70–0.88) 0.78 (0.73–0.84) 0.81 (0.77– 0.84) P value <0.0001 <0.0001 <0.0001 <0.0001 Age at first birth among parous women   ≤24 197 227 1.00 103 106 1.00 194 193 1.00 1.00   25–29 329 333 1.29 (0.98–1.71) 152 156 1.04 (0.69–1.56) 225 171 1.09 (0.78–1.52) 1.17 (0.98– 1.41)   30–35 155 166 1.42 (1.00–2.00) 76 92 1.01 (0.61–1.68) 123 101 0.88 (0.59–1.32) 1.13 (0.91– 1.42)   >35 37 48 0.94 (0.55–1.61) 27 28 0.93 (0.45–1.96) 57 29 1.08 (0.60–1.96) 1.11 (0.79– 1.55) P trend 0.31 0.96 0.80 0.36 Per 5 years 1.06 (0.92–1.21) 1.01 (0.83–1.22) 1.00 (0.86–1.15) 1.04 (0.95– 1.13) P value 0.43 0.96 0.96 0.42 Age at last birth among parous women   ≤25 84 76 1.00 42 39 1.00 58 42 1.00 1.00   26–30 262 271 0.87 (0.59–1.26) 122 106 1.15 (0.66–2.20) 150 124 0.98 (0.59–1.63) 0.97 (0.75– 1.26)   30–35 249 275 0.94 (0.64–1.39) 122 156 0.87 (0.49–1.54) 208 172 1.10 (0.66–1.82) 0.98 (0.75– 1.27)   >35 123 152 0.81 (0.53–1.24) 71 81 1.09 (0.58–2.05) 183 155 1.05 (0.62–1.76) 0.97 (0.73– 1.29) P trend 0.52 0.82 0.70 0.89 Per 5 years 0.97 (0.86–1.10) 0.95 (0.79–1.14) 0.97 (0.85–1.12) 0.98 (0.90– 1.06) P value 0.43 0.59 0.71 0.56 Breastfeeding among parous women   0 months 382 370 1.00 185 181 1.00 284 236 1.00 1.00   >0 to <0.5 yr 136 178 0.74 (0.55–0.99) 78 63 1.28 (0.81–2.01) 157 110 1.18 (0.83–1.67) 0.96 (0.79– 1.16)   0.5 to ≤1yr 77 103 0.82 (0.57–1.16) 47 55 1.17 (0.70–1.94) 61 40 1.55 (0.95–2.53) 1.01 (0.79– 1.28)   >1 to <2 yr 64 78 0.78 (0.53–1.16) 28 54 0.82 (0.45–1.49) 40 55 0.68 (0.41–1.11) 0.76 (0.58– 0.99)    ≥2 yr 57 44 1.49 (0.93–2.39) 19 29 1.04 (0.51–2.10) 56 53 1.10 (0.67–1.82) 1.18 (0.88– 1.59) P trend 0.99 0.86 0.84 0.78 Per 6 months 1.03 (0.97–1.10) 0.97 (0.89–1.07) 0.96 (0.91–1.01) 0.99 (0.95– 1.02) P value 0.36 0.57 0.14 0.39 Oral contraceptives   Never used 484 444 1.00 242 191 1.00 475 309 1.00 1.00   >0 to <1 yr 186 187 1.03 (0.79–1.33) 76 67 1.24 (0.80–1.90) 136 119 0.95 (0.69–1.30) 1.03 (0.87– 1.23)   1 to 5 142 147 0.96 (0.71–1.28) 98 117 0.87 (0.59–1.27) 111 88 1.07 (0.75–1.53) 0.97 (0.80– 1.16)   >5 to ≤10 46 67 0.75 (0.49–1.15) 46 62 0.80 (0.49–1.31) 38 44 0.70 (0.42–1.16) 0.75 (0.57– 0.97)   >10yr 32 51 0.61 (0.37–1.00) 57 68 0.84 (0.52–1.37) 20 22 0.84 (0.40–1.65) 0.80 (0.60– 1.07)   P trend 0.056 0.24 0.40 0.040 Per 5 years of use 0.84 (0.72, 0.97) 0.94 (0.82, 1.07) 0.89 (0.73, 1.09) 0.91 (0.84– 0.99) P value 0.0154 0.326 0.37 0.024 Age at menarche ≤11 102 110 1.00 117 126 1.00 146 97 1.00 1.00   12 213 206 1.00 (0.71–1.43) 149 157 0.98 (0.68–1.41) 182 150 0.82 (0.57–1.19) 0.95 (0.77, 1.16)   13 220 229 0.89 (0.63–1.27) 134 125 0.99 (0.68–1.44) 171 152 0.72 (0.50–1.03) 0.87 (0.71, 1.07)   14 184 177 0.96 (0.67–1.39) 56 55 0.83 (0.51–1.36) 122 86 1.04 (0.69–1.57) 0.98 (0.79, 1.23)   ≥15 172 174 0.80 (0.55–1.19) 63 42 1.13 (0.67–1.91) 159 97 0.99 (0.67–1.48) 0.96 (0.76, 1.21) P trend 0.23 1.00 0.64 0.85 Per 1 year delay 0.96 (0.90–1.03) 0.99 (0.91–1.08) 1.01(0.95–1.08) 0.99 (0.95– 1.03) P value 0.23 0.79 0.69 0.72 Chinese Japanese Filipina All Age at menarche– Premenopausal ≤11 62 69 1.00 55 56 1.00 84 52 1.00 1.00   12 126 133 0.94 (0.60–1.48) 58 77 0.84 (0.48–1.48) 89 75 0.78 (0.47–1.30) 0.85 (0.64– 1.12)   13 112 138 0.80 (0.51–1.27) 50 67 0.61 (0.34–1.10) 69 80 0.54 (0.32–0.90) 0.66 (0.50– 0.88)   14 81 107 0.73 (0.45–1.18) 19 22 0.63 (0.28–1.43) 41 33 0.84 (0.45–1.58) 0.73 (0.53– 1.01)   ≥15 54 88 0.57 (0.34–0.97) 12 15 0.70 (0.27–1.83) 48 34 0.73 (0.39–1.36) 0.62 (0.44– 0.88) P value 0.014 0.13 0.26 0.003 Per 1 year delay 0.89 (0.81–0.98) 0.88 (0.75–1.03) 0.92(0.83–1.03) 0.91 (0.85– 0.96) P 0.016 0.11 0.14 0.0016 Age at menarche– Postmenopausal ≤11 40 41 1.00 62 70 1.00 62 45 1.00 1.00   12 87 73 1.19 (0.67–2.14) 91 80 1.26 (0.77–2.06) 93 75 0.94 (0.55–1.61) 1.12(0.83–1.51)   13 108 91 1.09 (0.62–1.92) 84 58 1.57 (0.93–2.63) 102 72 0.97 (0.57–1.66) 1.18 (0.88–1.59)   14 103 70 1.41 (0.80–2.52) 37 33 1.11 (0.59–2.11) 81 53 1.30 (0.73–2.30) 1.34 (0.97–1.85)   ≥15 118 86 1.25 (0.70–2.26) 51 27 1.67 (0.85–3.30) 111 63 1.30 (0.75–2.24) 1.43 (1.04–1.96) P value 0.37 0.17 0.14 0.015 Per 1 year delay 1.04 (0.94–1.14) 1.06 (0.95–1.18) 1.08 (0.99–1.18) 1.06 (1.01– 1.12) P 0.46 0.34 0.10 0.026 Age at natural menopause   ≤49 117 109 1.00 69 60 1.00 133 105 1.00 1.00   50–54 161 112 1.37 (0.91–2.05) 103 88 0.95 (0.56–1.60) 127 91 1.09 (0.71–1.67) 1.17 (0.91– 1.49)   >54 29 26 0.87 (0.44–1.71) 35 21 1.33 (0.62–2.85) 36 18 1.71 (0.83–3.51) 1.23 (0.83– 1.83) P trend 0.65 0.59 0.21 0.19 Per 5 year delay 1.02 (0.99–1.06) 1.01 (0.96–1.05) 1.02 (0.98–1.06) 1.02 (0.99–1.04) P 0.25 0.84 0.42 0.17 Age at menopause (natural + complete hysterectomy)   ≤49 161 144 1.00 96 93 1.00 181 141 1.00 1.00   50–54 172 117 1.36 (0.94–1.98) 114 95 1.09 (0.69–1.72) 137 96 1.07 (0.73–1.58) 1.22 (0.97– 1.52)   >54 30 26 0.85 (0.45–1.61) 36 22 1.35 (0.67–2.00) 38 18 1.60 (0.80–3.20) 1.23 (0.85– 1.79) P trend 0.61 0.42 0.26 0.098 Per 5 year delay 1.02 (0.99–1.06) 1.02 (0.97–1.06) 1.01 (0.97–1.04) 1.02 (1.00–1.04) P value 0.22 0.47 0.65 0.064 1 Adjusted for age, education, income, years of residence in the US among non-US born, interviewer, family history of breast cancer, benign breast diseases. All the other variables (except the ovulatory months) were mutually adjusted for each other. 2 As above and also adjusted for Asian ethnicity. 3 Included stillbirths reported in 28 cases and 25 controls. Table 3 Cumulative menstrual months and risk of breast cancer in Filipina, Chinese and Japanese Americans in Los Angeles County Chinese Japanese Filipina All Ca Co OR1 (95%CI) Ca Co OR1 (95%CI) Ca Co OR1 (95%CI) OR2 (95%CI) Menstrual months3,4   ≤283 102 168 1.00 79 118 1.00 90 106 1.00 1.00   284 to ≤345 173 214 1.16 (0.80–1.69) 70 81 1.06 (0.65–1.73) 124 123 1.11 (0.72–1.71) 1.07 (0.84– 1.35)   346 to ≤399 220 192 1.57 (1.04–2.36) 84 87 1.05 (0.63–1.75) 192 139 1.61 (1.01–2.56) 1.36 (1.06– 1.76)   >399 251 183 1.91 (1.20–3.03) 170 124 1.53 (0.88–2.68) 235 121 2.55 (1.50–4.33) 1.83 (1.38– 2.44) P trend 0.003 0.16 0.0002 <0.0001 Per 50 months 1.16 (1.05–1.28) 1.07 (0.96–1.19) 1.22 (1.09–1.36) 1.13 (1.06– 1.19) P value 0.0037 0.23 0.008 0.0001 By menopausal status Premenopausal women   ≤283 70 145 1.00 57 88 1.00 62 69 1.00 1.00   284 to ≤345 112 165 1.26 (0.81–1.98) 42 59 0.99 (0.54–1.83) 66 72 0.96 (0.53–1.74) 1.07 (0.80– 1.44)   346 to ≤399 126 117 2.30 (1.34–3.93) 39 41 1.18 (0.58–2.39) 97 70 1.44 (0.75–2.78) 1.65 (1.17– 2.33)   >399 95 70 2.99 (1.55–5.78) 42 28 2.32 (0.99–547) 82 41 2.59 (1.15–5.84) 2.67 (1.75– 4.08) P trend 0.0003 0.092 0.019 <0.0001 Per 50 months 1.23 (1.07–1.42) 1.11 (0.96–1.28) 1.26 (1.06–1.49) 1.18 (1.09– 1.29) P value 0.0038 0.17 0.008 0.0001 Postmenopausal women   ≤283 32 23 1.00 22 30 1.0 28 37 1.00 1.00   284 to ≤345 61 49 1.00 (0.48–2.06) 28 22 1.36 (0.56–3.30) 58 51 1.32 (0.68–2.58) 1.08 (0.72– 1.63)   346 to ≤399 94 75 0.93 (0.45–1.91) 45 46 0.88 (0.38–1.99) 95 69 1.69 (0.86–3.33) 1.08 (0.72– 1.62)   >399 156 113 1.09 (0.51–2.31) 128 96 1.23 (0.54–2.80) 153 80 2.61 (1.24–5.52) 1.36 (0.89– 2.08) P trend 0.79 0.75 0.008 0.12 Per 50 months 1.08 (0.93–1.26) 1.02 (0.87–1.20) 1.18 (1.01–1.39) 1.07 (0.98– 1.17) P value 0.32 0.79 0.042 0.13 By pregnancy history No pregnancy   ≤283 15 16 1.0 19 27 1.00 10 9 1.00 1.00   284 to ≤345 20 11 2.39 (0.53–0.79) 18 10 1.87 (0.53–6.67) 16 12 0.83 (0.15–4.47) 1.64 (0.82– 3.27)   346 to ≤399 33 19 2.80 (0.47–6.67) 20 13 1.02 (0.30–3.51) 34 16 0.97 (0.13–7.40) 1.64 (0.77– 3.49)   >399 50 29 1.35 (0.22–8.23) 36 15 2.66 (0.75–9.46) 71 19 3.02 (0.29–1.32) 2.25 (0.96– 5.30) P trend 0.92 0.21 0.24 0.08 Per 50 months 1.11 (0.79–1.56) 1.19 (0.98–1.46) 1.21 (0.83–1.76) 1.18 (1.01– 1.37) P 0.55 0.087 0.32 0.03 Yes pregnancy   ≤283 87 152 1.00 60 91 1.0 80 97 1.00 1.00   284 to ≤345 153 203 1.07 (0.72–1.59) 52 71 0.85 (0.49–1.47) 108 111 1.11 (0.70–1.75) 1.00 (0.77– 1.29)   346 to ≤399 187 173 1.41 (0.91–2.18) 64 74 0.92 (0.51–1.67) 158 123 1.47 (0.90–2.39) 1.25 (0.95– 1.65)   >399 201 154 1.73 (1.05–2.84) 134 109 1.22 (0.65–2.29) 164 102 1.97 (1.11–3.49) 1.56 (1.14– 2.13) P trend 0.017 0.50 0.014 0.0022 Per 50 months 1.13 (1.01–1.26) 1.01 (0.89–1.14) 1.14 (1.01–1.30) 1.08 (1.01– 1.15) P 0.028 0.93 0.037 0.0023 1 Adjusted for age, education, income, years of residence in the US among non-US born, interviewer, family history of breast cancer, benign breast diseases. All the other variables (except the ovulatory months) were mutually adjusted for each other. 2 As above and also adjusted for Asian ethnicity 3 Total menstrual months were calculated as age at menarche subtracted from age at menopause for postmenopausal women and from age at interview or cancer diagnosis for premenopausal women, and then subtracting anovulatory periods due to complete and incomplete pregnancies, lactation, OC use, or missing periods (if 3 or more periods were missing in a row). We excluded women who started hormone therapy before menopause (112 cases, 63 controls) or had hysterectomy only (153 cases, 126 controls) because age at menopause was not known, women who did not have regular cycles (122 cases, 136 controls), and other reasons (3 cases, 7 controls). 4 The ORs per 50 menstrual months when additionally adjusted for history of diabetes, years of physical activity, and dietary pattern (vegetable/soy pattern, and ethnic meats and carbohydrates); the fully adjusted RR per 50 menstrual months were: Chinese: 1.18 (95% CI 1.07–1.31), P=0.001; Japanese: 1.07 (95% CI 0.96–1.20), P=0.20; and Filipina: 1.23 (95% CI 1.10–1.39), p=0.0005 Table 4 Body size and risk of breast cancer in premenopausal Filipina, Chinese and Japanese women in Los Angeles County Chinese Japanese Filipina All Ca Co OR1 (95%CI) Ca Co OR1 (95%CI) Ca Co OR1 (95%CI) OR2 (95%CI) Height (cm) ≤155 116 129 1.00 74 76 1.00 126 116 1.00 1.00 156–160 122 159 0.93 (0.65–1.34) 62 83 0.76 (0.46–1.28) 113 78 1.58 (1.02–2.45) 1.02 (0.81– 1.29) 161–165 112 129 1.00 (0.68–1.46) 35 52 0.73 (0.40–1.33) 54 50 1.20 (0.71–2.02) 0.99 (0.76– 1.29) >165 85 118 0.82 (0.55–1.24) 23 26 1.12 (0.54–2.31) 38 30 1.28 (0.70–2.34) 1.01 (0.75– 1.34) P trend 0.45 0.80 0.35 0.99 Per 10 cm 1.02 (0.80–1.30) 0.91 (0.61–1.37) 1.12 (0.81–1.55) 1.04 (0.88– 1.23) P 0.89 0.65 0.49 0.65 BMI at age 18 (kg/m2)   <20 306 378 1.00 98 107 1.00 261 193 1.00 1.00   20 – ≤22.9 87 114 1.01 (0.71–1.42) 74 101 0.86 (0.55–1.37) 53 62 0.58 (0.36–0.92) 0.83 (0.66– 1.04)   >22.9 21 29 0.84 (0.46–1.55) 22 27 1.03 (0.51–2.06) 15 14 0.64 (0.28–1.49) 0.89 (0.60– 1.32) P trend 0.71 0.80 0.03 0.18 Per 5 kg/m2 0.99 (0.74–1.33) 1.04 (0.68–1.59) 0.70 (0.48–1.01) 0.94 (0.78– 1.14) P 0.96 0.85 0.059 0.54 BMI at age 30 (kg/m2)   <20 212 234 1.00 71 70 1.00 118 100 1.00 1.00   20 – ≤22.9 161 222 0.81 (0.60–1.09) 80 99 0.86 (0.52–1.41) 137 118 1.20 (0.79–1.82) 0.91 (0.74– 1.12)   >22.9–≤24.9 34 41 0.98 (0.57–1.67) 23 36 0.79 (0.39–1.57) 40 30 1.15 (0.63–2.13) 0.91 (0.65– 1.27)   >24.9 17 27 0.61(0.30– 1.23) 19 27 0.72 (0.34–1.53) 29 20 2.02 (0.96–4.24) 1.00 (0.67– 1.49) P trend 0.17 0.34 0.10 0.69 Per 5 kg/m2 0.88 (0.66–1.71) 0.83 (0.58–1.17) 1.31 (0.93–1.84) 1.00 (0.83– 1.19) P value 0.39 0.29 0.12 0.95 BMI at age 40 (kg/m2)   <20 99 99 1.00 41 29 1.00 47 33 1.00 1.00   20 – ≤22.9 156 183 0.93 (0.64–1.35) 62 72 0.80 (0.41–1.58) 102 80 1.01 (0.55–1.86) 0.91 (0.69– 1.19)   >22.9–≤24.9 67 70 1.08 (0.68–1.72) 33 39 0.61 (0.28–1.34) 60 53 0.86 (0.44–1.65) 0.86 (0.62– 1.20)   >24.9 47 62 0.74 (0.44–1.24) 25 30 0.78 (0.34–1.78) 61 47 0.97 (0.49–1.92) 0.84 (0.60– 1.20) P trend 0.45 0.40 0.93 0.31 Per 5 kg/m2 0.98 (0.75–1.28) 0.80 (0.55–1.17) 1.10 (0.79–1.53) 0.97 (0.82, 1.16) P value 0.87 0.25 0.58 0.76 Current BMI (kg/m2)   ≤22.9 304 381 1.00 127 144 1.00 182 150 1.00 1.00   >22.9–≤24.9 73 78 1.19 (0.81–1.74) 32 45 0.73 (0.41–1.32) 67 60 0.92 (0.57–1.46) 0.95 (0.74– 1.22)   >24.9 – ≤27.5 41 59 0.76 (0.48–1.20) 23 23 1.02 (0.49–2.13) 46 43 0.85 (0.49–1.47) 0.85 (0.62– 1.15)   >27.5 –≤29.9 6 10 0.73 (0.24–2.23) 5 9 0.78 (0.23–2.69) 20 8 2.02 (0.79–5.17) 1.16 (0.66– 2.05)   ≥30 13 7 1.83 (0.67–4.97) 7 16 0.46 (0.16–1.30) 16 13 0.92 (0.38–2.24) 1.06 (0.63– 1.77) P trend 0.96 0.19 0.81 0.82 Per 5 kg/m2 1.02 (0.80–1.29) 0.77 (0.71–1.07) 1.14 (0.85–1.51) 0.99 (0.85– 1.15) P value 0.91 0.099 0.38 0.82 Weight gain since age 18 (kg) ≤3.6 127 171 0.88 (0.63–1.24) 85 89 1.37 (0.83–2.26) 49 56 0.65 (0.38–1.12) 0.93 (0.73– 1.18) >3.6–9.1 144 184 1.00 69 88 1.00 98 88 1.00 1.00 >9.1– ≤14.1 85 83 1.10 (0.73–1.64) 20 28 0.83 (0.40–1.73) 85 47 1.61 (0.97–2.70) 1.22 (0.92– 1.61) >14.1 to ≤22.7 44 66 0.75 (0.47–1.20) 18 23 0.98 (0.44–2.18) 71 59 1.03 (0.61–1.73) 0.87 (0.64– 1.17) >22.7 14 17 0.88 (0.39–1.96) 2 7 0.37 (0.07–1.99) 26 19 1.24 (0.58–2.65) 0.94 (0.58– 1.53) P trend 0.49 0.49 0.37 0.80 Per 5 kg 1.02 (0.92–1.12) 0.82 (0.69–0.97)3 1.15 (1.03–1.30)3 1.02 (0.95– 1.08) P value 0.71 0.017 0.015 0.63 Waist(cm) ≤71.1 104 128 1.00 69 71 1.00 38 46 1.00 1.00 >71.1– ≤75.5 161 198 1.01 (0.70–1.44) 48 57 0.74 (0.41–1.33) 100 76 1.84 (1.02–3.34) 1.07 (0.82– 1.38) >75.5 – ≤85.1 100 127 0.94 (0.63–1.42) 39 58 0.64 (0.35–1.19) 99 72 1.83 (0.99–3.38) 1.03 (0.78– 1.36) >85.1 57 66 1.01 (0.62–1.65) 34 46 0.76 (0.40–1.47) 87 77 1.67 (0.90–3.10) 1.03 (0.76– 1.40) P trend 0.89 0.29 0.27 0.94 Per 5 cm 1.00 (0.92–1.09) 0.95 (0.86–1.05) 1.06 (0.96–1.17) 1.01 (0.96– 1.06) P value 0.89 0.32 0.23 0.69 Hip (cm) ≤92.7 136 170 1.00 70 69 1.00 87 66 1.00 1.00 >92.7 – ≤96.5 88 117 0.91 (0.62–1.33) 39 50 0.84 (0.46–1.51) 62 53 0.87 (0.50–1.52) 0.88 (0.68– 1.15) >96.5 – ≤102.9 138 155 1.08 (0.76–1.52) 51 67 0.84 (0.48–1.47) 105 87 0.96 (0.59–1.56) 1.03 (0.80– 1.31) >102.9 60 77 0.91 (0.58–1.43) 29 46 0.77 (0.40–1.50) 70 65 0.82 (0.48–1.40) 0.87 (0.65– 1.16) P trend 0.99 0.44 0.57 0.62 Per 5 cm 1.00 (0.76–1.29) 0.92 (0.75–1.13) 1.04 (0.93, 1.16) 1.01 (0.95– 1.07) P value 0.96 0.42 0.51 0.80 Waist hip ratio ≤0.76 122 150 1.00 58 63 1.00 33 38 1.00 1.00 >0.76 – ≤0.80 132 150 1.15 (0.80–1.65) 45 59 0.77 (0.42–1.39) 70 56 1.79 (0.92–3.48) 1.08 (0.82– 1.41) >80.0 – ≤0.845 93 131 0.95 (0.64–1.41) 49 58 0.78 (0.42–1.42) 102 89 1.41 (0.73–2.65) 0.92 (0.70– 1.21) >0.845 75 88 1.07 (0.69–1.67) 37 52 0.62 (0.33–1.20) 119 88 1.84 (0.97–3.49) 1.08 (0.80– 1.45) P trend 0.98 0.18 0.18 0.62 Per 0.01 0.99 (0.76–1.29) 0.94 (0.66–1.34) 1.28 (0.90–1.81) 1.01 (0.92– 1.10) P value 0.94 0.73 0.17 0.92 1 Adjusted for age, education, income, years of residence in the US among non-US born, interviewer, age at menarche, parity, family history of breast cancer, and benign breast diseases. 2 As above but also adjusted for Asian ethnicity 3 RR per 5 kg weight gain differed significantly between Japanese and Filipina American women (P interaction =0.003) Table 5 Body size and risk of breast cancer in postmenopausal Filipina, Japanese and Chinese women in Los Angeles County Chinese Japanese Filipina All Ca Co OR1 (95%CI) Ca Co OR1 (95%CI) Ca Co OR1 (95%CI) OR2 (95%CI) Height (cm) ≤155 163 128 1.00 136 140 1.00 229 166 1.00 1.00 156–160 136 115 0.94 (0.66, 1.36) 117 70 1.75 (1.14, 2.67) 125 85 1.07 (0.73, 1.56) 1.18 (0.95, 1.46) 161–165 104 75 1.10 (0.73, 1.65) 50 37 1.56 (0.9, 2.69) 63 40 1.20 (0.74–1.94) 1.23 (0.95, 1.59) 165 53 43 1.09 (0.66, 1.81) 22 21 1.27 (0.63, 2.56) 32 17 1.43 (0.72, 2.84) 1.15 (0.83, 1.61) P trend 0.61 0.10 0.30 0.13 Per 10 cm 1.08 (0.81, 1.45) 1.28 (0.91, 1.82) 1.12 (0.83, 1.53) 1.12 (0.94, 1.33) P value 0.60 0.16 0.46 0.21 BMI 18 (kg/m2) <20 295 231 1.00 166 134 1.00 319 224 1.00 1.00   20 – ≤22.9 64 71 0.76 (0.50–1.14) 117 107 1.07 (0.72–1.60) 97 49 1.43 (0.94–2.19) 1.03 (0.82, 1.29) >22.9 25 16 1.20 (0.60–2.39) 35 26 1.41 (0.76–2.61) 15 19 0.42 (0.19–0.93) 0.94 (0.64, 1.38) P trend 0.71 0.32 0.67 0.93 Per 5 kg/m2 1.01 (0.72–1.41) 1.10 (0.76–1.61) 0.89 (0.63–1.27) 0.99 (0.85, 1.16) P value 0.97 0.61 0.53 0.93 BMI at age 30 (kg/m2)   <20 208 165 1.00 112 83 1.00 150 117 1.00 1.00   20 – ≤22.9 150 124 1.03 (0.73–1.45) 137 118 0.92 (0.61–1.39) 192 122 1.36 (0.96–1.96) 1.11 (0.90, 1.36)   >22.9–≤24.9 38 33 1.03 (0.58–1.80) 45 45 0.99 (0.56–1.74) 63 40 1.31 (0.78–2.19) 1.12 (0.83, 1.51)   >24.9 18 12 1.41(0.64– 3.12) 30 22 1.39 (0.70–2.75) 31 14 2.48 (1.16–5.32) 1.71 (1.14, 2.58) P trend 0.53 0.50 0.021 0.026 per 5 kg/m2 1.11 (0.81–1.53) 1.15 (0.82–1.62) 1.36 (0.99–1.86) 1.21 (1.01, 1.45) P value 0.52 0.43 0.06 0.041 BMI at age 40 (kg/m2)   <20 126 119 1.00 76 59 1.00 78 60 1.00 1.00   20 – ≤22.9 180 143 1.34 (0.93–1.93) 126 114 0.81 (0.51–1.29) 166 119 1.28 (0.81–2.02) 1.21 (1.01, 1.45)   >22.9–≤24.9 72 44 1.72 (1.05–2.82) 60 55 0.99 (0.57–1.73) 107 53 1.61 (0.96–2.71) 1.45 (1.09, 1.92)   >24.9 42 38 1.33 (0.77–2.30) 63 39 1.77 (0.98–3.21) 91 63 1.40 (0.82–2.37) 1.50 (1.11, 2.03) P trend 0.086 0.049 0.16 0.002 per 5 kg/m2 1.21 (0.98–1.37) 1.36 (1.03–1.81) 1.31 (0.99–1.74) 1.29 (1.11, 1.51) P value 0.17 0.032 0.06 0.001 Current BMI (kg/m2) ≤22.9 244 217 1.00 146 137 1.00 163 127 1.00 1.00 >22.9–≤24.9 106 69 1.44 (0.98–2.12) 72 61 1.19 (0.74–1.88) 116 75 1.43 (0.94–2.16) 1.30 (1.03– 1.64) >24.9 – ≤27.5 65 47 1.30 (0.83–2.03) 58 31 2.56 (1.46–4.47) 94 71 1.14 (0.74–1.74) 1.39 (1.07– 1.80) >27.5 – ≤29.9 24 14 1.60 (0.77–3.35) 25 21 1.66 (0.82–3.34) 39 21 1.65 (0.87–3.13) 1.60 (1.09– 2.33) ≤30 17 14 1.26 (0.58–2.74) 24 18 1.69 (0.80–3.57) 37 14 2.79 (1.35 –5.77) 1.83 (1.21– 2.76) P trend 0.11 0.008 0.012 0.0001 Per 5 kg/m2 1.20 (0.96–1.50)3 1.29 (1.02–1.62)3 1.40 (1.11–1.78)3 1.29 (1.13– 1.46)3 P value 0.12 0.032 0.005 0.0001 Weight gain ≤3.64 81 70 1.01 (0.65–1.58) 91 91 1.07 (0.68–1.73) 40 39 0.83 (0.45–1.52) 0.95 (0.73– 1.26) >3.64–9.09 117 109 1.00 80 74 1.00 84 65 1.00 1.00 >9.09– ≤14.1 77 63 1.14 (0.72–1.81) 60 49 1.23 (0.72–2.10) 103 72 1.34 (0.82–2.17) 1.22 (0.93– 1.60) >14.1 to ≤22 .7 76 64 1.10 (0.69–1.75) 67 40 1.95 (1.11–3.42) 152 85 1.54 (0.97–2.44) 1.45 (1.11– 1.90) >22.7 33 12 2.76 (1.30–5.85) 20 18 1.25 (0.57–2.76) 52 31 1.61 (0.88–2.96) 1.74 (1.19– 2.56) P trend 0.064 0.053 0.019 0.0002 Per 5 kg 1.09 (0.99–1.21) 1.10 (1.00–1.21) 1.18 (1.07–1.30) 1.13 (1.07– 1.19) P value 0.082 0.058 0.001 <0.0001 Waist(cm) ≤71.1 80 55 1.00 50 46 1.00 26 22 1.00 1.00 >71.1– ≤75.5 134 119 0.74 (0.47–1.16) 81 90 0.85 (0.48–1.48) 74 63 0.98 (0.47–2.05) 0.86 (0.63– 1.16) >75.5 – ≤85.1 122 101 0.87 (0.55–1.42) 77 64 1.08 (0.60–1.92) 139 98 1.16 (0.57–2.33) 1.04 (0.76– 1.42) >85.1 106 80 0.95 (0.58–1.56) 105 64 1.74 (0.97–3.12) 194 122 1.52 (0.77–3.02) 1.34 (0.98– 1.83) P trend 0.79 0.0125 0.04 0.004 Per 5cm 1.00 (0.92–1.09) 1.12 (1.03–1.22) 1.05 (0.97–1.14) 1.06 (1.01– 1.11) P value 0.99 0.012 0.23 0.02 Hip(cm) ≤92.7 119 91 1.00 80 76 1.00 94 59 1.00 1.00 >92.7 – ≤96.5 82 71 0.97 (0.65–1.57) 69 57 1.31 (0.78–2.22) 53 55 0.58 (0.33–1.00) 0.92 (0.69– 1.21) >96.5 – ≤102.9 158 119 1.12 (0.79–1.72) 84 64 1.17 (0.76–1.90) 151 92 1.15 (0.73–1.81) 1.17 (0.92– 1.50) >102.9 83 74 0.94 (0.60–1.48) 80 64 2.08 (1.21–3.56) 135 99 1.06 (0.67–1.68) 1.22 (0.94– 1.60) P trend 0.95| 0.01 0.44 0.05 Per 5 cm 1.02 (0.92–1.13) 1.12 (1.01, 1.24) 1.03 (0.80–1.32) 1.06 (1.01– 1.12) P value 0.70 0.038 0.81 0.03 Waist/Hip Ratio (WHR) ≤0.76 92 66 1.00 43 57 1.00 31 21 1.00 1.00 >0.76 – ≤0.80 117 93 0.92 (0.58–1.44) 75 66 1.17 (0.66–2.07) 55 64 0.54 (0.26–1.13) 0.88 (0.65– 1.20) >80.0 – ≤0.845 115 96 0.85 (0.54–1.34) 80 66 1.27 (0.72–2.25) 118 79 0.96 (0.48–1.94) 1.05 (0.78– 1.43) >0.845 118 100 0.86 (0.54–1.37) 115 75 1.53 (0.87–2.68) 229 141 0.98 (0.67–1.91) 1.13 (0.84– 1.52) P trend 0.50 0.12 0.16 0.14 Per 0.1 0.93 (0.72–1.20) 1.31 (0.98–1.76) 1.03 (0.80–1.43) 1.08 (0.93– 1.25) P value 0.56 0.021 0.81 0.30 1 Adjusted for age, education, income, years of residence in the US among non-US born, interviewer, age at menarche, parity, family history of breast cancer, benign breast disease, and type of menopause status and age at menopause. 2 As above but also adjusted for Asian ethnicity 3 Full adjustment included history of diabetes, years of physical activity, and dietary pattern (vegetable/soy pattern, and ethnic meats and carbohydrates); the fully adjusted RR per 5 kg/m2 current BMI were: Chinese 1.14 (95% CI 0.90–1.44), p=0.27;, Japanese 1.23 (95% CI 0.97–1.55), p=0.092; Filipina 1.31 (95% CI 1.02–1.67), p=0.036; and all subjects combined 1.22 (95% CI 1.07–1.39), p=0.003. 4 Full adjustment included history of diabetes, years of physical activity, and dietary pattern (vegetable/soy pattern, and ethnic meats and carbohydrates); the fully adjusted RR per 5 kg weight gain were: Chinese 1.07 (95% CI 0.96–1.18), p=0.22; Japanese 1.08 (95% CI 0.98–1.20), p=0.13; Filipina 1.16 (95% CI 1.05–1.28), p=0.005; and all subjects combined 1.11 (95% CI 1.05–1.17), p=0.0005. Disclosure of Potential Conflicts of interest: None References 1 Global Burden of Disease Cancer C Fitzmaurice C Dicker D Pain A Hamavid H Moradi-Lakeh M The Global Burden of Cancer 2013 JAMA Oncol 2015 1 505 527 26181261 2 Gomez SL Noone AM Lichtensztajn DY Scoppa S Gibson JT Liu L Cancer incidence trends among Asian American populations in the United States, 1990–2008 J Natl Cancer Inst 2013 105 1096 1110 23878350 3 Gibson LJ Hery C Mitton N Gines-Bautista A Parkin DM Ngelangel C Risk factors for breast cancer among Filipino women in Manila Int J Cancer 2010 126 515 521 19626603 4 Clavel-Chapelon F Cumulative number of menstrual cycles and breast cancer risk: results from the E3N cohort study of French women Cancer Causes Control 2002 13 831 838 12462548 5 Chavez-MacGregor M Elias SG Onland-Moret NC van der Schouw YT Van Gils CH Monninkhof E Postmenopausal breast cancer risk and cumulative number of menstrual cycles Cancer Epidemiol Biomarkers Prev 2005 14 799 804 15824146 6 Wu AH McKean-Cowdin R Tseng CC Birth weight and other prenatal factors and risk of breast cancer in Asian-Americans Breast Cancer Res Treat 2011 130 917 925 21710135 7 Wu AH Vigen C Razavi P Tseng CC Stancyzk FZ Alcohol and breast cancer risk among Asian-American women in Los Angeles County Breast Cancer Res 2012 14 R151 23185976 8 Pike MC Peters RK Cozen W Probst-Hensch NM Felix JC Wan PC Estrogen-progestin replacement therapy and endometrial cancer J Natl Cancer Inst 1997 89 1110 1116 9262248 9 Ross RK Paganini-Hill A Wan PC Pike MC Effect of hormone replacement therapy on breast cancer risk: estrogen versus estrogen plus progestin J Natl Cancer Inst 2000 92 328 332 10675382 10 Wu AH Yu MC Tseng CC Pike MC Body size, hormone therapy and risk of breast cancer in Asian-American women Int J Cancer 2007 120 844 852 17131315 11 Jih J Mukherjea A Vittinghoff E Nguyen TT Tsoh JY Fukuoka Y Using appropriate body mass index cut points for overweight and obesity among Asian Americans Prev Med 2014 65 1 6 24736092 12 Wu AH Ziegler RG Pike MC Nomura AM West DW Kolonel LN Menstrual and reproductive factors and risk of breast cancer in Asian-Americans Br J Cancer 1996 73 680 686 8605107 13 Collaborative Group on Hormonal Factors in Breast C Menarche, menopause, and breast cancer risk: individual participant meta-analysis, including 118 964 women with breast cancer from 117 epidemiological studies Lancet Oncol 2012 13 1141 1151 23084519 14 Tamakoshi K Yatsuya H Wakai K Suzuki S Nishio K Lin Y Impact of menstrual and reproductive factors on breast cancer risk in Japan: results of the JACC study Cancer Sci 2005 96 57 62 15649257 15 Hirose K Tajima K Hamajima N Takezaki T Inoue M Kuroishi T Effect of body size on breast-cancer risk among Japanese women Int J Cancer 1999 80 349 355 9935173 16 Gao YT Shu XO Dai Q Potter JD Brinton LA Wen W Association of menstrual and reproductive factors with breast cancer risk: results from the Shanghai Breast Cancer Study Int J Cancer 2000 87 295 300 10861490 17 Yuan JM Yu MC Ross RK Gao YT Henderson BE Risk factors for breast cancer in Chinese women in Shanghai Cancer Res 1988 48 1949 1953 3349468 18 Nagata C Hu YH Shimizu H Effects of menstrual and reproductive factors on the risk of breast cancer: meta-analysis of the case-control studies in Japan Jpn J Cancer Res 1995 86 910 915 7493908 19 Kawai M Minami Y Kuriyama S Kakizaki M Kakugawa Y Nishino Y Reproductive factors, exogenous female hormone use and breast cancer risk in Japanese: the Miyagi Cohort Study Cancer Causes Control 2010 21 135 145 19816778 20 Iwasaki M Otani T Inoue M Sasazuki S Tsugane S Japan Public Health Center-based Prospective Study G Role and impact of menstrual and reproductive factors on breast cancer risk in Japan Eur J Cancer Prev 2007 16 116 123 17297387 21 Bao PP Shu XO Gao YT Zheng Y Cai H Deming SL Association of hormone-related characteristics and breast cancer risk by estrogen receptor/progesterone receptor status in the shanghai breast cancer study Am J Epidemiol 2011 174 661 671 21768404 22 Ursin G Bernstein L Wang Y Lord SJ Deapen D Liff JM Reproductive factors and risk of breast carcinoma in a study of white and African-American women Cancer 2004 101 353 362 15241834 23 Wada K Nagata C Tamakoshi A Matsuo K Oze I Wakai K Body mass index and breast cancer risk in Japan: a pooled analysis of eight population-based cohort studies Ann Oncol 2014 25 519 524 24412821 24 Collaborative Group on Hormonal Factors in Breast C Breast cancer and breastfeeding: collaborative reanalysis of individual data from 47 epidemiological studies in 30 countries, including 50302 women with breast cancer and 96973 women without the disease Lancet 2002 360 187 195 12133652 25 Tao SC Yu MC Ross RK Xiu KW Risk factors for breast cancer in Chinese women of Beijing Int J Cancer 1988 42 495 498 3170023 26 Wang QS Ross RK Yu MC Ning JP Henderson BE Kimm HT A case-control study of breast cancer in Tianjin, China Cancer Epidemiol Biomarkers Prev 1992 1 435 439 1302554 27 Nagata C Mizoue T Tanaka K Tsuji I Tamakoshi A Wakai K Breastfeeding and breast cancer risk: an evaluation based on a systematic review of epidemiologic evidence among the Japanese population Jpn J Clin Oncol 2012 42 124 130 22210922 28 Collaborative Group on Hormonal Factors in Breast C Breast cancer and hormonal contraceptives: collaborative reanalysis of individual data on 53 297 women with breast cancer and 100 239 women without breast cancer from 54 epidemiological studies Lancet 1996 347 1713 1727 8656904 29 Ursin G Wu AH Hoover RN West DW Nomura AM Kolonel LN Breast cancer and oral contraceptive use in Asian-American women Am J Epidemiol 1999 150 561 567 10489994 30 Xu WH Shu XO Long J Lu W Cai Q Zheng Y Relation of FGFR2 genetic polymorphisms to the association between oral contraceptive use and the risk of breast cancer in Chinese women Am J Epidemiol 2011 173 923 931 21382839 31 Shu XO Jin F Dai Q Shi JR Potter JD Brinton LA Association of body size and fat distribution with risk of breast cancer among Chinese women Int J Cancer 2001 94 449 455 11745429 32 Shin A Matthews CE Shu XO Gao YT Lu W Gu K Joint effects of body size, energy intake, and physical activity on breast cancer risk Breast Cancer Res Treat 2009 113 153 161 18228135 33 Van Den Brandt PA Spiegelman D Yaun SS Adami HO Beeson L Folsom AR Pooled analysis of prospective cohort studies on height, weight, and breast cancer risk Am J Epidemiol 2000 152 514 527 10997541 34 Amadou A Ferrari P Muwonge R Moskal A Biessy C Romieu I Overweight, obesity and risk of premenopausal breast cancer according to ethnicity: a systematic review and dose-response meta-analysis Obes Rev 2013 14 665 678 23615120 35 Wei S Schmidt MD Dwyer T Norman RJ Venn AJ Obesity and menstrual irregularity: associations with SHBG, testosterone, and insulin Obesity (Silver Spring) 2009 17 1070 1076 19180069 36 John EM Sangaramoorthy M Phipps AI Koo J Horn-Ross PL Adult body size, hormone receptor status, and premenopausal breast cancer risk in a multiethnic population: the San Francisco Bay Area breast cancer study Am J Epidemiol 2011 173 201 216 21084558 37 Berstad P Coates RJ Bernstein L Folger SG Malone KE Marchbanks PA A case-control study of body mass index and breast cancer risk in white and African-American women Cancer Epidemiol Biomarkers Prev 2010 19 1532 1544 20501755 38 Low S Chin MC Ma S Heng D Deurenberg-Yap M Rationale for redefining obesity in Asians Ann Acad Med Singapore 2009 38 66 69 19221673 39 Goh LG Dhaliwal SS Welborn TA Lee AH Della PR Ethnicity and the association between anthropometric indices of obesity and cardiovascular risk in women: a cross-sectional study BMJ Open 2014 4 e004702 40 Suzuki R Iwasaki M Inoue M Sasazuki S Sawada N Yamaji T Body weight at age 20 years, subsequent weight change and breast cancer risk defined by estrogen and progesterone receptor status--the Japan public health center-based prospective study Int J Cancer 2011 129 1214 1224 21064092 41 White KK Park SY Kolonel LN Henderson BE Wilkens LR Body size and breast cancer risk: the Multiethnic Cohort Int J Cancer 2012 131 E705 E716 22120517 42 Suzuki S Kojima M Tokudome S Mori M Sakauchi F Wakai K Obesity/weight gain and breast cancer risk: findings from the Japan collaborative cohort study for the evaluation of cancer risk J Epidemiol 2013 23 139 145 23438695 43 Kawai M Minami Y Kuriyama S Kakizaki M Kakugawa Y Nishino Y Adiposity, adult weight change and breast cancer risk in postmenopausal Japanese women: the Miyagi Cohort Study Br J Cancer 2010 103 1443 1447 20842123 44 Maxwell AE Crespi CM Alano RE Sudan M Bastani R Health risk behaviors among five Asian American subgroups in California: identifying intervention priorities J Immigr Minor Health 2012 14 890 894 22089979 45 Bates LM Acevedo-Garcia D Alegria M Krieger N Immigration and generational trends in body mass index and obesity in the United States: results of the National Latino and Asian American Survey, 2002–2003 Am J Public Health 2008 98 70 77 18048787

PMC005xxxxxx/PMC5135607.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 100966035 22040 Trends Mol Med Trends Mol Med Trends in molecular medicine 1471-4914 1471-499X 27887808 5135607 10.1016/j.molmed.2016.10.003 NIHMS823889 Article Norovirus Regulation by Host and Microbe Baldridge Megan T. 1* Turula Holly 2 Wobus Christiane E. 2 1 Department of Medicine, Division of Infectious Diseases, Washington University School of Medicine, St. Louis, MO, USA 2 Department of Microbiology and Immunology, University of Michigan Medical School, Ann Arbor, MI, USA * Correspondence: mbaldridge@wustl.edu, M.T. Baldridge 26 10 2016 22 11 2016 12 2016 01 12 2017 22 12 10471059 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. Norovirus (NoV) infection is the leading cause of epidemic gastroenteritis globally, and can lead to detrimental chronic infection in immunocompromised hosts. Despite its prevalence as a cause of diarrheal illness, the study of human NoVs (HNoVs) has historically been limited by a paucity of models. The use of murine NoV (MNoV) to interrogate mechanisms of host control of viral infection has facilitated the exploration of different genetic mouse models, revealing roles for both innate and adaptive immunity in viral regulation. MNoV studies have also recently identified important interactions between the commensal micro-biota and NoV with clear extensions to HNoVs. In this review, we discuss the most current understanding of how the host, the microbiome, and their interactions regulate NoV infections. Noroviruses: Tenacious Pathogens In the United States alone, human noroviruses (HNoVs) are responsible for approximately 20 million cases of acute gastroenteritis annually, leading to over 70 000 hospitalizations and nearly 800 deaths [1]. HNoV infections are also a global problem, causing approximately US$60 billion in societal costs every year [2]. HNoVs cause a species-specific infection, but recent developments are overcoming the historical lack of cell culture and small animal models [3–5]. Nevertheless, the direct study of factors regulating HNoV pathogenesis in the natural host will always be limited. To counter this limitation, HNoV infections are studied in non-human hosts or related NoVs are investigated in their natural hosts as detailed in a recent review [6]. Among the available models, murine NoV (MNoV), first described in 2003, provides the most widely used, readily tractable model system to explore viral and host factors regulating NoV infection [7]. MNoV infection is studied in vitro macrophages, dendritic cells (DCs), and B cells [4,8], as well as in vivo in mice [9]. Together, these studies have revealed novel host pathways critical to the regulation of NoV infection, and facilitated the exploration of NoV interactions with the commensal microbiome, a critically important player in mucosal infection. In this review, first, we briefly summarize parameters of NoV infections including transmission, symptoms, and viral tropism. Second, we explore the known mechanisms of host regulation of NoVs, with a focus on innate and adaptive immune regulators. Lastly, we detail recent work exploring the interactions of NoVs with the microbiota, describing the coordinate effects of host and microbial control of NoVs, and providing a comprehensive examination of the complex interactions between NoV, host, and bacteria. Future studies of the multifaceted regulation of NoV infection using existing and newly developed models will undoubtedly yield new scientific insights that may ultimately reduce the global burden of disease. NoV Infection and Disease NoV is a genus in the Caliciviridae (see Glossary) family. These non-enveloped icosahedral viruses have a single-stranded, positive-sense RNA genome, and are classified into at least six genogroups on the basis of their nucleotide sequence [10]. Genogroup I (GI), GII, and GIV viruses infect humans, with GII being the most prevalent, while GV viruses infect rodents (Table 1) [11]. The NoV genome contains three to four open reading frames (ORFs). ORF1 encodes nonstructural proteins including viral protein, genome-linked (VPg) and the RNA-dependent RNA polymerase (RdRp). ORF2 and ORF3 encode structural capsid proteins VP1 and VP2, respectively [10]. ORF4 is only found in MNoVs and encodes virulence factor VF1 [12]. NoV transmission typically occurs by the fecal–oral route from contaminated surfaces, food or water, and by person-to-person spread [10] but transmission via droplets, through aerosolization of HNoV-containing vomitus, can also occur [13,14]. Outbreaks occur in places where people gather (e.g., cruise ships, day-care centers, hospitals). They are facilitated by the low numbers of virions able to cause infections (i.e., low infectious dose) [15,16], high amounts of viral shedding [17], high environmental stability of HNoV [18], and a relative viral resistance to disinfectants [19]. After an average 1.2-day incubation period, HNoV infection induces symptoms including abdominal pain, nausea, vomiting, and diarrhea, which typically resolve within 1–4 days [17,20,21]. However, viral shedding may occur for weeks to months in asymptomatic healthy hosts [22], and years in immunocompromised patients [23]. The latter have been postulated to serve as a reservoir for future outbreaks [24]. There is no significant correlation between presentation of symptoms and viral burden, duration, or magnitude of NoV shedding, but enhanced cytokine responses correlate with HNoV symptoms and suggest immune mediation [17]. Complications can occur following acute infection and include postinfectious irritable bowel syndrome [25,26], life-threatening dehydration [27], necrotizing enterocolitis [28], and exacerbation of Crohn’s disease [29]. NoVs cause species-specific infections. Thus, HNoVs can only infect animal models with reduced immune responses [3]. Alternatively, models relying on the natural infection of surrogate viruses can be used (recently reviewed, [6,30]). MNoV, a natural mouse pathogen endemic to animal facilities throughout the world [31–33], has been the most widely used surrogate model. MNoV cultivation in multiple cell types in vitro, the ability to genetically manipulate both virus and host, and the use of acute [murine norovirus 1 (MNV-1)] and chronic (e.g., MNV.CR6, MNV-3) MNoV strains add to the strengths of this model system [24,30]. The cellular and tissue tropism is a critical determinant of pathogenesis and an active area of investigation in the NoV field (Table 1) [34]. Recently, a model was proposed based on experimental evidence, whereby MNoVs use microfold (M) cells to overcome the epithelial barrier in order to infect B cells, macrophages, and DCs in the intestine, before being trafficked to local lymph nodes and distal sites by DCs [35–38]. B cells are also targets for HNoVs [4], but other targets exist, since humans deficient in B cells are still susceptible to HNoV infection [39]. Recent immunofluorescence analysis of small intestinal biopsy samples from HNoV-infected immunocompromised patients revealed the presence of HNoV infection in intestinal epithelial cells, CD68+ or DC-SIGN+ phagocytes (e.g., macrophages, DCs), and CD3+ cells (T cells or intraepithelial lymphocytes) [40]. A tropism of HNoV for enterocytes was subsequently confirmed by cultivating HNoV in human intestinal enteroid monolayer cultures [5]. MNoV antigen is also observed in small intestinal epithelial cells of immunodeficient signal transducer and activator of transcription 1 (Stat1)- and recombination activating gene (Rag1)/Stat1-deficient mice [41,42]. Taken together, the data indicate that both MNoV and HNoVs share a tropism for intestinal immune and epithelial cells. However, whether all the same cell types are infected in immunocompetent hosts remains to be confirmed. Cellular tropism of NoVs is determined at the level of virus entry [43]. This was confirmed recently following the identification of CD300LF and CD300LD as functional receptors for MNoV [44,45]. Expression of murine CD300LF and CD300LD in multiple nonsusceptible cells, including HeLa or HEK293T cells from nonmurine hosts, supported MNoV infection, while infection could be reduced by competition with soluble protein or antibody [44,45]. Expression of human CD300F was unable to substitute for murine CD300LF, nor was antihuman CD300F able to block infection, indicating that restriction of NoVs may be due to species-specific variation in these molecules, rendering them determinants of species specificity. CD300LF and CD300LD belong to a family of type I transmembrane proteins with an immunoglobulin-like extracellular domain that can bind lipids in the plasma membrane [46]. Both proteins are expressed in myeloid cells, which are known MNoV target cells [47,48]. These findings raise questions regarding their physiological role during MNoV infection in vivo. Preincubation of MNV-1 with soluble CD300LF prevents mortality of Stat1-deficient mice, and Cd300lf−/− mice are resistant to viral shedding following oral infection with MNV.CR6 [44]. Whether MNoV establishes tissue infection in Cd300lf−/− or CD300ld−/− mice, however, has not been reported. These data pave the way for future investigations into the molecular details governing NoV entry. Another key player during NoV infection is fucosyltransferase 2, or FUT2, whose protein expression is critical for susceptibility to most (but not all) HNoV strains [49,50]. FUT2 activity is required during the synthesis of ABO histo-blood group antigens (HBGAs), which are attachment factors or receptors that facilitate binding of some caliciviruses to cells [51,52]. Some individuals harbor an inactivating mutation in FUT2, which leads to the absence of ABO antigens, and these individuals (nonsecretors) have lower susceptibility to symptomatic infection [53]. This is likely due to the inability of certain HNoV strains to infect enterocytes, since intestinal organoids from nonsecretors are resistant to infection by some HNoVs [5]. Antibodies that prevent the interaction between the virus and HBGAs correlate with development of protective immune responses to HNoV infection [54,55], and current vaccine and immunoprophylaxis strategies under development are aimed at eliciting antibodies that block HBGA binding [56,57]. Future studies will be required to understand the precise function(s) of different susceptibility factors during NoV infection, which in turn may influence vaccine design. Host Immune Regulators of NoV Host genes broadly important for immunity are critical in the control of HNoV infection, as evidenced by primary or secondary immunodeficiencies, which have been associated with chronic NoV infection [58]. The immune response and cytokines in particular may also control symptomatology in HNoV infection [17]. Serum T-helper type 1 (Th1), Th2 cytokines, and interleukin-8 (IL-8) were more elevated in symptomatic versus asymptomatic patients, despite equivalent viral titers [17]. However, the underlying genetic causes for these specific immune responses were not explored. Future work to identify additional human genes affecting HNoV infection and symptomatology will thus be important in order to aid in vaccine development and treatment. Recent NoV vaccination efforts (reviewed in [59]) have relied upon HNoV-recombinant viruslike particles, produced by spontaneous self-assembly of the capsid protein VP1, to provide insight into immune correlates of protection [54,56]. The advent of novel in vitro culture methods for HNoV [4,5] now makes development of live-attenuated vaccines possible, and may facilitate defining functional host immune responses to HNoV. Genetic knockout mouse models have proved extremely tractable for interrogating and identifying specific host genes regulating NoV pathogenesis (Table 2). Indeed, MNoV was first identified in immunodeficient mice highlighting the importance of the type I interferon (IFN) pathway, including the type I IFN receptor, Ifnar1, and the downstream transcription factor Stat1 in MNoV control [9,60–63]. The production of, and response to, IFNs has been consistently shown to be important for MNoV regulation. Pattern recognition receptors for viral RNA including melanoma differentiation-associated protein 5 (MDA5), a retinoic acid-inducible gene (RIG-I)-like receptor, Toll-like-receptor 3 (TLR3), and nucleotide-binding oligomerization domain (NOD)-like receptor family member NLRP6 also limit viral replication in vivo [64,65]. Pattern recognition receptors stimulate transcription of IFNs by activating the IFN regulatory transcription factor (IRF) family members IRF3 and IRF7, which additionally control MNoV viral levels [60]. Ifnar1−/− mice, which lack the ability to respond to type I IFNs, die upon MNoV challenge [8,9,60–63]. This lethality phenotype is exacerbated by an additional mutation in Ifngr1, the type II IFN receptor, suggesting combinatorial antiviral effects of type I and II IFNs [61,66]. Interestingly, for persistent strains of MNoV, Ifnar1 does not control viral shedding or levels in the gut, but rather, regulates the ability of virus to spread to extraintestinal sites such as spleen [67,68]. Fecal shedding of these strains is instead limited by type III IFN signaling through Ifnlr1 [67,68]. Intraperitoneal administration of type III IFN cures persistent MNoV infection, even in the absence of adaptive immunity, suggesting this pathway as a potential therapeutic target for humans [67]. IFN-λ administration is an especially attractive approach, as it is a safe treatment for chronic hepatitis C virus [69]. Work in HEK293FT cells demonstrated that replication of transfected HNoV RNA is also sensitive to pretreatment with type I and III IFN, but unlike MNoV, blocking such responses does not affect viral replication [70]. This suggests that there may be alternate ways in which HNoV and MNoV engage IFN pathways [70]. After engagement of IFNs with their receptors, downstream transcription factors drive the production of a variety of antiviral factors. STAT1, which mediates IFN-stimulated gene transcription in type I, II, and III IFN signaling pathways, is critical for controlling viral loads, shedding, and lethality, as observed by infection of Stat1−/− mice with acute and persistent MNoV strains [8,9,66–68]. Transcription factor IRF1 is important for type II IFN-mediated control of MNoV in vitro in macrophages, and in vivo Irf1−/− mice exhibit enhanced lethality to MNoV challenge in the context of antibody-mediated IFNAR1 depletion [66]. Other factors downstream of IFN signaling implicated in MNoV control include proteins involved in macroautophagy. Mice lacking autophagy genes Atg5, Atg12, Atg7, or Atg16L1 in macrophages fail to control MNoV in vivo on an Ifnar1-deficient background, and autophagy-deficient macrophages fail to respond to type II IFN-directed control of acute MNoV replication in vitro [61]. However, the specific mechanisms by which autophagy may control MNoV remain to be elucidated. Interestingly, ATG16L1 is also important for preventing intestinal pathology induced by persistent MNoV infection [71]. This effect does not appear to occur through direct control of MNoV levels in the gut, as mice hypomorphic for Atg16L1 exhibit an inflammatory bowel disease (IBD)-like phenotype with viral infection [71]. Another study has shown that mice deficient in Il10, an anti-inflammatory cytokine, also exhibit intestinal pathology with persistent MNoV infection; in Il10−/− mice, altered MNoV levels may drive inflammation as MNoV is detectable in epithelial and lamina propria cells in the first two days after infection [72]. These findings suggest that although inflammation is important for controlling MNoV, unchecked inflammation induced by the virus may have pathologic consequences. In addition to innate immunity, the adaptive immune system also controls MNoV (Table 2). For example, MNV-1 is not effectively cleared in mice lacking either set of, or both sets of, lymphocytes (e.g., B and T cells), that is, Rag1-, immunoglobulin heavy constant mu (Ighm/muMt)-, and MHC Class II-deficient mice [63,73–75]. Antibody responses are an important component of MNoV control and protection against repeat challenge, as Rag1−/− and muMt−/− mice fail to clear acute MNoV infection [73,74]. CD4+ and CD8+ T cells also play combinatorial roles in regulating MNoV; mice lacking CD4+ T cells present higher tissue titers but normal viral clearance, while mice lacking CD8+ T cells exhibit delayed viral clearance [63,74–76]. Infection with different MNoV strains elicits differential innate and adaptive immune responses. While adaptive immune responses are critical for the control of acute MNoV strains, persistent strains such as MNV-3 and MNV.CR6 fail to generate effective T-cell responses, which may contribute to viral persistence [75,76]. Acute MNoV strain MNV-1 stimulates greater type I and III IFN responses than persistent strain MNV.CR6, which may contribute to the development of enhanced adaptive immune responses and viral clearance [67]. However, a robust adaptive immune response alone is not sufficient for clearance of acute MNoV, as MNV-1 persists in mice lacking type I IFN signaling in DCs [62]. Extrapolating from MNoV findings, it is reasonable to assume that targeted IFN-related therapies might be potentially effective in HNoV clearance; however, the importance of differential adaptive immune responses to genetically diverse NoV strains may represent a challenge and will require further investigation. Interactions of NoV with the Microbiota The intestinal lumen houses a large number and variety of commensal microbes (bacteria, viruses, fungi, protozoa, and single-celled archaea species), collectively referred to as the microbiome. Pathogens infecting the intestine encounter these microbes, and NoV infections are influenced by those encounters with harmful or beneficial outcomes to the host. To date, a direct interaction of HNoV capsids with members of the microbiota is known for a few commensal (e.g., Enterobacter cloacae) and pathogenic (e.g., Clostridium difficile) bacterial species [77,78]. Interactions are mediated via HBGA-like carbohydrates expressed on the surface of these bacteria [78]. Both HNoVs and MNoVs have been reported to bind additional carbohydrate moieties [79–83], including sialic acid residues (Table 1), which are widely expressed on bacteria [84]. Thus, MNoVs likely interact with members of the microbiome as well, although specific interactions remain to be demonstrated. Although not universally seen, both HNoV and MNoV can alter host gut microbial communities. This dysbiosis typically results in an enhanced bacterial Firmicutes-to-Bacteroidetes (F/B) ratio, an alteration also linked to liver fibrosis and obesity [85,86]. A subset (20%) of HNoV-infected individuals exhibited enhanced F/B ratios [87], while MNV-1 enhanced F/B ratios in C57BL/6 mice at Day 5 postinfection [88]. However, longitudinal analyses of both C57BL/6 (MNV-1, MNV.CR6) and Swiss Webster (MNV-1, MNV-4) mice failed to detect MNoV-induced dysbiosis of gastrointestinal bacteria at the phylum level [89]. A reduction in intestinal Lactobacillus strains occurred in MNV1-infected ICR mice, with a concurrent increase in Proteobacteria delta, mirroring findings from HNoV [89]. On the flip side, experimentally induced bacterial dysbiosis also alters MNoV infection. For instance, pretreatment of mice with an antibiotic cocktail including vancomycin, ampicillin, metronidazole, and neomycin (VNAM) reduces acute MNV-1 titers [4] and reduces or prevents infection by persistent MNoV strains MNV-3 and MNV.CR6, respectively [4,68]. Importantly, MNV.CR6 viral loads are restored with a fecal transplant from untreated to antibiotic-treated animals [68]. Although antibiotic treatment reduces intestinal MNoV titers, it is not an advisable treatment option for HNoV infections, given the overall beneficial impact of the microbiota on human health [90]. From the perspective of the cellular level, co-infection of macrophages with MNoV and various bacteria has been shown to reduce viral titers, likely stemming from bacterial triggering of innate immune responses [91–93]. However, by contrast, MNoV-induced inhibition of inflammatory cytokines appears to foster bacterial growth. These findings highlight the need to further study complex, trans-kingdom interactions and their effects on the host. The Bermuda Triangle: Host, Microbiota, and NoV Recent work has provided much insight into different host regulators of NoVs and clearly identified an important role for commensal microbes in NoV infection. However, these findings only begin to outline a still poorly defined triangle of interactions between the host, bacteria, and the virus. As outlined earlier, commensal bacteria in vivo serve a proviral function by facilitating successful viral infection [4,68]. This proviral function is countered by host antiviral factors in the IFN-λ pathway, including IRF3, IFNLR1, and STAT1, such that in the absence of these host molecules, persistent MNoV is less dependent on bacteria for infection (Figure 1A, Key Figure). The mechanism by which these host molecules alter the infectious potential of MNoV and its interaction with bacteria remains unclear. Consequently, these interactions will require a greater depth of understanding of all three arms of contact: MNoV interactions with commensal bacteria, interactions of bacteria with type III IFN signaling pathways, and the specific antiviral effects of type III IFN signaling on MNoV. In another interesting example of other enteric kingdoms of life interacting with MNoV in a proviral fashion, helminth infection with Trichinella spiralis leads to elevated intestinal MNV-1 loads due to STAT6-dependent changes in macrophage function [94]. These examples indicate that NoVs can benefit from the presence of other kingdoms of life normally present in the intestine, but that these beneficial effects may be mediated by the host. In other cases, the combination of MNoV and commensal bacteria may induce pathologic consequences to a susceptible host, such as persistent MNoV infection potentiating intestinal inflammation in particular genetic contexts. For example, Atg16L1−/− or Il10−/− mice, which serve as models for IBD, develop intestinal inflammation following MNoV infection [71,72]. Treatment of mice with antibiotics following MNoV viral infection, or the infection of germ-free mice, prevents injury-induced pathology. Consequently, in these IBD models, MNoV and commensal bacteria work synergistically to induce inflammation via inflammatory host factors [tumor necrosis factor-α (TNF-α) and IFN-γ], and in opposition to other protective host factors (ATL16L1 and IL-10; Figure 1B). MNoV can also act in concert with pathogenic bacteria to drive inflammation. In a mouse model of Helicobacter bilis-induced IBD, mice deficient in ATP-binding cassette transporter multidrug resistance gene 1a (Mdr1a) develop more severe colitis during co-infection with Gram-negative pathogen H. bilis and persistent MNoV, than with H. bilis infection alone [95] (Figure 1C). Host factor MDR1A, linked to human IBD and potentially important in intestinal barrier function, thus appears to suppress this co-infection phenotype [95]. The pathology-enhancing effect of MNoV in the context of IBD risk factor genes, however, was not observed in mice deficient for mothers against decapentaplegic homolog 3 (Smad3) and Il10 [96,97]. Acute MNoV infection also enhances lethality by systemic Escherichia coli, via type I IFNs and TNF-α, as well as via NOD1/NOD2- and RIP2-enhanced bacterial recognition on cell surfaces [98]. Thus, in addition to promoting potentially deleterious inflammatory effects from commensal bacteria under certain genetic contexts in mice, MNoV can also enhance inflammation, which is driven by known pathogenic bacteria such as E. coli. The host inflammatory response, while fighting these pathogens, is subverted by such multiple stimuli, turning on itself, and causing severe intestinal tissue damage to the host. Extrapolating to humans, the wide genetic diversity, differential co-infections, and unique commensal microbes harbored by each person may thus differentially predispose individuals to a range of symptoms and potential long-term effects from NoV-induced inflammation. The effects of inherent human variability during NoV infection thus represent an unexplored area of investigation. In other contexts, MNoV infection can be beneficial for the host. Interestingly, in the absence of the microbiota, MNoV acts via type IIFN to restore abnormalities in intestinal tissues and immune function [99]. Wild-type mice treated with a VNAM antibiotics cocktail exhibited enhanced dextran sodium sulfate-induced intestinal injury and morbidity, which could be rescued by MNoV infection, but not in an Ifnar1−/− genetic background. In addition, MNoV infection of VNAM-treated mice significantly reduced weight loss, diarrhea, and intestinal histopathology following infection with Citrobacter rodentium, a model for enteropathogenic E. coli [99]. Thus, MNoV contributed in protecting mice against injury, presumably by replacing IFN signaling functions of commensal bacteria [99] (Figure 1D). Another study found that in ampicillin-treated mice, MNoV infection partially restored antibacterial defenses against vancomycin-resistant Enterococcus faecium, again supporting a role for MNoV in replacing the commensal micro-biota’s role in immune signaling [100]. While these studies were undertaken in the absence of the microbiota, MNoV infection of conventionally housed mice has also been found to increase mouse survival following intranasal bacterial infection with Pseudomonas aeruginosa [101]. Indeed, MNoV infection was associated with reduced P. aeruginosa-induced neutrophil infiltration in the lung and decreased proinflammatory cytokine (IL-6 and TNF-α) production, thus resulting in protected alveolar capillary barrier permeability and reduced acute lung injury [101] (Figure 1E). MNoV can thus confer protection against a variety of different pathogenic bacteria by stimulating effective immunological responses; whether these findings apply broadly to all mucosal pathogens or are specific to particular pathogen subsets remains to be determined. The intestinal environment with its rich mixture of different commensal and pathogenic microbes, host cells, microbial products, and cytokines is a complex locale which we are only beginning to understand. Systematic interrogation of the individual connections between these different components will start to shed light on the context-dependent effects of these factors on each other. Ultimately, studies will need to consider the translatability of these findings to HNoV infection as appropriate models are being developed. Concluding Remarks NoV infections cause a significant public health burden worldwide, but no directed treatment or prevention strategies are currently approved. Historically, this has been in large part due to the many technical hurdles that scientists have faced working with NoV. However, over the last two decades, NoV researchers have made significant leaps forward, increasing our general understanding of NoV biology (Box 1). These developments were aided in part by the utility of the MNoV model. Recent advances have included the identification of previously unknown host regulators, and the ascription of a novel function for the microbiota in modulating NoV infection at both cellular and host levels. Such information is invaluable for the development of vaccines and antivirals. Furthermore, NoV-infected cell types are being identified, which shifts the paradigm away from an exclusively enterotropic virus to that of a virus exhibiting a broader tropism also encompassing myeloid cell lineages [8,40]. The implications of this expanded cellular tropism for pathogenesis and vaccine responses are only just beginning to be explored. In addition, determining the translatability of findings from MNoV to HNoV will require gaining a better appreciation of the aspects by which HNoV infection and disease are best modeled with current experimental strategies. In addition, the development of improved models will undoubtedly be a focus of investigation over the decades to come (see Outstanding Questions). Thus, as future discoveries reveal further insights into NoVs, these are expected to have a significant positive impact on public health, hopefully reducing the associated personal and economic burden of this infectious disease. Box 1 The Clinician’s Corner Norovirus is the leading cause of epidemic gastroenteritis, and can persist indefinitely in immunocompromised hosts. This may be the case for patients with genetic immunodeficiencies, or those at risk of infection post-transplantation. In addition to causing acute gastroenteritis, norovirus infection has been suggested as a possible trigger for other enteric diseases, including postinfectious irritable bowel syndrome, necrotizing enterocolitis in infants, and exacerbation of inflammatory bowel disease. Additional evidence is needed to bolster these connections, and the mechanisms by which norovirus might cause long-term disease remain unclear. The commensal bacteria of the gut have been shown to help facilitate norovirus infection. Because different individuals harbor different commensal bacterial communities, it is possible that interindividual variation in gut bacteria could contribute to differing severity of norovirus infection. There is no evidence, however, that antibiotic treatment would be either helpful or therapeutic for an ongoing norovirus infection. Both the innate and adaptive immune systems have been shown to be important for norovirus regulation in mouse models, suggesting that future targeting of the innate immune system may be important therapeutically for patients with chronic norovirus infection. Outstanding Questions Which are the major cell types infected during norovirus (NoV) infection in immunocompetent hosts? How does the tropism for immune cells influence immunity against NoVs? Which aspects of human NoV biology are best modeled by murine NoV? Which aspects require other existing or yet to be developed models? What are the physiological functions of histo-blood group antigens, CD300LF, and CD300LD during NoV infection? How does the wide genetic diversity of commensal microbes in individuals, in addition to differential co-infections with different pathogens, influence NoV disease outcomes? How will microbial/NoV interactions affect vaccine development? Can interferon-λ or other host regulators be contemplated as a putative treatment for persistent NoV infections? If so, how? M.T.B. was supported by National Institutes of Health (NIH) training grant T32CA009547. Work in the laboratory of C.E.W. was supported by NIH grants AI102106, AI080611, and AI103961. H.T. was supported in part by a University of Michigan Rackham Merit fellowship and NIH training grants T32AI007413 and T32DK094775. Glossary Caliciviridae (caliciviruses) family of positive-sense, nonsegmented, single-stranded RNA viruses including norovirus. CD300LF and CD300LD inhibitory receptors of the immunoglobulin superfamily, known to be expressed on myeloid cells, and recently identified as viral receptors for murine norovirus. CD68+ or DC-SIGN+ phagocytes CD68 is a surface marker for macrophages, and DC-SIGN (or CD209) is a marker for dendritic cells, both found in the lamina propria of the human intestine. Crohn’s disease a type of chronic inflammatory bowel disease that may affect any part of the gastrointestinal tract, characterized by symptoms including abdominal pain, diarrhea, fever, and weight loss. Dextran sodium sulfate a compound used to induce colitis in mouse models; it interferes with intestinal barrier function and stimulates inflammation, thereby mimicking clinical features of inflammatory bowel diseases. Fucosyltransferase 2 enzyme that transfers fucose sugars, and governs the molecular basis of the histo-blood group antigens, which are surface markers for red blood cells. Macroautophagy a process by which cells degrade and recycle intracellular contents. Microfold (M) cells a subtype of intestinal epithelial cell, found in gut- and mucosa-associated lymphoid tissue, which facilitates transport of microbes from the gut lumen to the lamina propria. Necrotizing enterocolitis a medical condition primarily affecting premature infants, often leading to death, where portions of the bowel undergo necrosis. Pattern recognition receptors receptors that recognize conserved features of pathogens and initiate signaling cascades resulting in innate immune responses. They are divided into four families, including Toll-like receptors, nucleotide-binding oligomerization domain-like receptors, and RIG-I-like receptors. T-helper type 1 (Th1) and Th2 T helper (Th) cells are T cells that facilitate B-cell antibody class switching and activation of cytotoxic T cells. Th1 cells are important for control of intracellular bacteria and protozoa and classically produce effector cytokine interferon-γ (IFN-γ type II IFN). Th2 cells are important for the control of extracellular parasites such as helminths, and are associated with cytokines interleukin-4 (IL-4), IL-5, and IL-13. Type I, II, and III interferons (IFNs) signaling cytokines produced by host cells in response to exposure to pathogens. Type I (including IFN-α and IFN-β) and type III IFNs (such as IFN-λ) are classically triggered with viral infection, with type III IFN being more associated with mucosal infections. Type II IFN (or IFN-γ) is usually associated with bacterial infections. Figure 1 (A) Commensal bacteria facilitate persistent MNoV infection in a manner counteracted by innate immune factors IRF3, IFNLR1, and STAT1. (B) Commensal bacteria and MNoV are both needed to drive TNF-α and IFN-γ-mediated intestinal inflammation in mice lacking Atg16l1 or Il10. (C) MNoV exacerbates intestinal inflammation and lethality during pathogenic bacterial infection by potentiating inflammatory signals, including TNF-α and type I IFN, and host responses to bacteria. (D) In the absence of commensal bacteria, MNoV may replace protective functions of enteric microbes via type I IFN signaling. (E) In other contexts, MNoV can serve a protective role by decreasing the host inflammatory response to bacterial pathogens. IFN, interferon; IFNLR1, interferon λ receptor 1; IL, interleukin; IRF3, interferon regulatory transcription factor 3; NOD, nucleotide-binding oligomerization domain; STAT1, signal transducer and activator of transcription 1; TNF, tumor necrosis factor. Table 1 Major Mechanisms of Infection in HNoV and MNoV. Human and murine norovirus (HNoV and MNoV) infections exhibit distinct characteristics, such as symptomatology and known attachment factors/receptors, but share overlap in the carbohydrate nature of their attachment factors, in cellular tropism, as well as in harboring a potential for persistent viral shedding. Human norovirus Murine norovirus Symptoms Abdominal pain, nausea, vomiting, and diarrhea [17,20,21] Asymptomatic in wild-type mice; potential for lethality in immunocompromised mice (Table 2) Duration of infection Acute symptomatic phase (1–4 days), which may be followed by viral shedding for weeks to months [22–24] Acute strains cleared in 7–10 days; persistent strains are shed for many months/lifetime of animal [24,30] In vitro tropism B cells and enterocytes [4,5] Macrophages, dendritic cells, microglial cells, and B cells [4,8,44] In vivo tropism Intestinal epithelial cells, myeloid cells, and lymphoid cells in immunocompromised patients [40] Intestinal epithelial cells, macrophages, dendritic cells, B cells, and Kupffer cells (stellate macrophages in the liver) in immunocompromised mice [4,41,42,72] Known attachment factors and receptors Histo-blood group antigens are attachment factors conferring susceptibility to most HNoV strains [51,52]. Some strains bind heparan sulfate, sialic acid, and β-galactosylceramide [81–83]. No proteinaceous receptors are known Strain-dependent attachment factors include terminal sialic acid residues on gangliosides and glycoproteins, and glycans on N-linked proteins [79,80]; CD300LF and CD300LD are proteinaceous viral receptors [44,45] Table 2 Host Genes Involved in MNoV Regulation. Mouse genes implicated in viral infection, innate immune control, and adaptive immune control of murine norovirus are presented. MNV-1 and MNV-1.CW3 (here CW3) are acute strains of MNoV that infect systemic organs and are cleared by wild-type (WT) mice, while CR3, CR6, MNV/Hannover, and MNV-3 are persistent strains of MNoV that infect the intestine, are shed into the feces, and are not cleared in WT mice. Viral infection Gene Role MNoV strain(s) In vitro effect In vivo effect Refs Cd300lf and CD300ld Viral receptors CW3/CR6 Myeloid cell lines and primary macrophages lacking Cd300lf or CD300ld are not infected Oral CR6 infection is prevented in Cd300lf−/− mice [44,45] Rag2 and Il2rg Development of Peyer’s patches and M cells CW3/CR3 MNoV is transcytosed by M-cell-like murine intestinal epithelial mlCcl2 cells Oral MNoV infection is reduced in mice lacking M cells [37,38] Innate immunity Gene Role MNoV strain(s) In vitro effect In vivo effect Refs Mda5 Viral recognition MNV-1 Higher viral replication and defective cytokine response in Mda5−/− dendritic cells Higher viral replication in Mda5−/− mice [64] Tlr3 Viral recognition MNV-1 No effect observed Higher viral replication in Tlr3−/− mice [64] Nlrp6 Viral recognition MNV-1 N/A Higher viral replication and levels of fecal shedding in Nlrp6−/− mice, and virus persists in Nlrp6−/− mice [65] Irf3 Induction of IFN CW3 Irf3−/−Irf7−/− dendritic cells and macrophages show higher viral replication, and macrophages show impaired IFN production Higher viral replication in Irf3−/− mice [60] CR6 N/A Irf3−/− mice are resistant to the effects of antibiotics in preventing infection [68] Irf7 Induction of IFN CW3 Irf3−/−Irf7−/− dendritic cells and macrophages show higher viral replication, and macrophages show impaired IFN production Higher viral replication in Irf7−/− mice [60] Ifnar1 Type I IFN response MNV-1/CW3 Ifnar1−/− dendritic cells and macrophages show higher viral replication Ifnar1−/− mice succumb to lethal infection, virus replicates to higher level, and virus persists in conditional knockout mice lacking Ifnar1 in dendritic cells or macrophages [8,9,60–63] CR6 N/A Enteric virus infects systemically in Ifnar1−/− mice [68,69] Ifngr1 Type II IFN response CW3 N/A Addition of Ifngr1 deficiency to type I IFN receptor deficiency causes increased viral replication and mortality [61,66] Ifnlr1 Type III IFN response CR6 N/A Higher viral replication and levels of fecal shedding in Ifnlr1−/− mice, and Ifnlr1−/− mice are resistant to the effects of antibiotics in preventing infection [67,68] Stat1 Type I, II, and III IFN response MNV-1/CW3 Stat1−/− dendritic cells and macrophages show higher viral replication, and Stat1 is required for IFN-γ-mediated inhibition of MNoV replication in macrophages Stat1−/− mice succumb to lethal infection [8,9,66] CR6 N/A Higher viral replication and levels of fecal shedding in Stat1−/− mice, and Stat1−/− mice are resistant to the effects of antibiotics in preventing infection [67,68] Irf1 Type II IFN Response CW3 Irf1 is required for IFN-γ-mediated inhibition of MNoV replication in macrophages Addition of Irf1 deficiency to type I IFN receptor deficiency causes increased viral replication and mortality [66] Atg5-Atg12, Atg7, and Atg16L1 Type II IFN response CW3 Atg5–Atg12, Atg7, and Atg16L1 are required for IFN-γ-mediated inhibition of MNoV replication in macrophages Addition of macrophage-specific Atg5-deficiency to type I IFN receptor deficiency causes increased viral replication and mortality [61] CR6 N/A Persistent viral infection induces intestinal pathology in Atg16L1-hypomorphic mice, dependent upon TNF-α and IFN-γ, though viral levels are similar to wild-type mice [71] Il10 Anti-inflammatory cytokine MNV/Hannover1 N/A Persistent viral infection induces intestinal pathology in Il10-deficient mice; though overall viral levels are similar to wild-type mice, MNoV localization is altered at early time points in knockout mice [72] Adaptive immunity Gene Role MNoV strain(s) In vitro effect In vivo effect Refs Rag1 B- and T-lymphocyte development CW3/MNV-3 N/A MNoV persists in Rag1−/− mice, and adoptively transferred immune splenocytes can clear infection; adoptively transferred MNV-specific CD8+ T cells can decrease viral loads [63,73–75] Ighm/muMt B-lymphocyte development CW3/MNV-3 N/A MNoV persists in muMt−/− mice; adoptively transferred immune splenocytes derived from B-cell-deficient mice or antibody production-deficient mice are unable to efficiently clear persistent infection in Rag1−/− mice; B cells are critical for protective immunity [63,73,74,76] MHC II CD4+ T-cell development CW3/MNV-3 N/A MHC II −/− mice have higher viral titers in the ileum than wild-type mice, but clear infection normally; CD4+ T cells are critical for protective immunity [63,74,76] MHC I and β2M CD8+ T-cell development CW3 N/A MNoV persists in MHC I × β2M−/− mice for longer periods than in wild-type mice, but is eventually cleared; acute control of virus is also regulated by CD8+ T cells [74,76] Trends Murine norovirus provides a practical model to study of norovirus infections and pathogenesis. Murine and human noroviruses share a tropism for immune cells and intestinal epithelial cells. Norovirus infection is regulated by multiple host genes, including viral entry factors, innate immune mediators, and components of the adaptive immune system. The commensal bacteria of the gastrointestinal microbiome can modulate norovirus infection. Recent findings indicate that norovirus and commensal bacteria stimulate overlapping host immune pathways, and furthermore, that certain bacteria are required to mediate some of the pathologic effects of norovirus infection. Host and microbial effects on norovirus may function in complex combinations to regulate viral infection and pathogenesis. 1 Hall AJ 2013 Norovirus disease in the United States Emerg Infect Dis 19 1198 1205 23876403 2 Bartsch SM 2016 Global economic burden of norovirus gastroenteritis PLoS ONE 11 e0151219 27115736 3 Taube S 2013 A mouse model for human norovirus mBio 4 e00450 e513 23860770 4 Jones MK 2014 Enteric bacteria promote human and mouse norovirus infection of B cells Science 346 755 759 25378626 5 Ettayebi K 2016 Replication of human noroviruses in stem cell-derived human enteroids Science 353 1387 1393 27562956 6 Wobus CE 2016 Animal models of norovirus infection Viral Gastroenteritis: Molecular Epidemiology and Pathogenesis 1 Svensson L 397 422 Academic Press 7 Wobus CE 2006 Murine norovirus: a model system to study norovirus biology and pathogenesis J Virol 80 5104 5112 16698991 8 Wobus CE 2004 Replication of norovirus in cell culture reveals a tropism for dendritic cells and macrophages PLoS Biol 2 e432 15562321 9 Karst SM 2003 STAT1-dependent innate immunity to a Norwalk-like virus Science 299 1575 1578 12624267 10 Green KY 2013 Caliciviridae: the noroviruses Fields Virology 6 Knipe DM 582 608 Lippincott Williams & Wilkins 11 Vinje J 2015 Advances in laboratory methods for detection and typing of norovirus J Clin Microbiol 53 373 381 24989606 12 McFadden N 2011 Norovirus regulation of the innate immune response and apoptosis occurs via the product of the alternative open reading frame 4 PLoS Pathog 7 e1002413 22174679 13 Kirby AE 2016 Vomiting as a symptom and transmission risk in norovirus illness: evidence from human challenge studies PLoS ONE 11 e0143759 27116105 14 Jones RM Brosseau LM 2015 Aerosol transmission of infectious disease J Occup Environ Med 57 501 508 25816216 15 Teunis PF 2008 Norwalk virus: how infectious is it? J Med Virol 80 1468 1476 18551613 16 Atmar RL 2014 Determination of the 50% human infectious dose for Norwalk virus J Infect Dis 209 1016 1022 24253285 17 Newman KL 2016 Norovirus in symptomatic and asymptomatic individuals: cytokines and viral shedding Clin Exp Immunol 184 347 357 26822517 18 Colas de la Noue A 2014 Absolute humidity influences the seasonal persistence and infectivity of human norovirus Appl Environ Microbiol 80 7196 7205 25217015 19 Tung G 2013 Efficacy of commonly used disinfectants for inactivation of human noroviruses and their surrogates J Food Prot 76 1210 1217 23834796 20 Kaufman SS 2014 Treatment of norovirus infections: moving antivirals from the bench to the bedside Antiviral Res 105 80 91 24583027 21 Lee RM 2013 Incubation periods of viral gastroenteritis: a systematic review BMC Infect Dis 13 446 24066865 22 Teunis PF 2015 Shedding of norovirus in symptomatic and asymptomatic infections Epidemiol Infect 143 1710 1717 25336060 23 Echenique IA 2016 Prolonged norovirus infection after pancreas transplantation: a case report and review of chronic norovirus Transpl Infect Dis 18 98 104 26460906 24 Karst SM 2015 The molecular pathology of noroviruses J Pathol 235 206 216 25312350 25 Bonani M 2016 Chronic norovirus infection as a risk factor for secondary lactose maldigestion in renal transplant recipients: a prospective parallel cohort pilot study Transplantation Published online August 1, 2016. http://dx.doi.org/10.1097/TP.0000000000001376 26 Futagami S 2015 Systematic review with meta-analysis: post-infectious functional dyspepsia Aliment Pharmacol Ther 41 177 188 25348873 27 Glass RI 2009 Norovirus gastroenteritis N Engl J Med 361 1776 1785 19864676 28 Pelizzo G 2013 Isolated colon ischemia with norovirus infection in preterm babies: a case series J Med Case Rep 7 108 23594884 29 Hubbard VM Cadwell K 2011 Viruses, autophagy genes, and Crohn’s disease Viruses 3 1281 1311 21994779 30 Karst SM 2014 Advances in norovirus biology Cell Host Microbe 15 668 680 24922570 31 Hsu CC 2005 Development of a microsphere-based serologic multiplexed fluorescent immunoassay and a reverse transcriptase PCR assay to detect murine norovirus 1 infection in mice Clin Diagn Lab Immunol 12 1145 1151 16210475 32 Kim JR 2011 Prevalence of murine norovirus infection in Korean laboratory animal facilities J Vet Med Sci 73 687 691 21187680 33 McInnes EF 2011 Prevalence of viral, bacterial and parasitological diseases in rats and mice used in research environments in Australasia over a 5-y period Lab Anim (NY) 40 341 350 22012194 34 Karst SM Tibbetts SA 2016 Recent advances in understanding norovirus pathogenesis J Med Virol 88 1837 1843 27110852 35 Elftman MD 2013 Multiple effects of dendritic cell depletion on murine norovirus infection J Gen Virol 94 1761 1768 23636823 36 Karst SM Wobus CE 2015 A working model of how noroviruses infect the intestine PLoS Pathog 11 e1004626 25723501 37 Gonzalez-Hernandez MB 2014 Efficient norovirus and reovirus replication in the mouse intestine requires microfold (M) cells J Virol 88 6934 6943 24696493 38 Gonzalez-Hernandez MB 2013 Murine norovirus transcytosis across an in vitro polarized murine intestinal epithelial monolayer is mediated by M-like cells J Virol 87 12685 12693 24049163 39 Brown JR 2016 Norovirus infections occur in B-cell-deficient patients Clin Infect Dis 62 1136 1138 26908782 40 Karandikar UC 2016 Detection of human norovirus in intestinal biopsies from immunocompromised transplant patients J Gen Virol 97 2291 2300 27412790 41 Mumphrey SM 2007 Murine norovirus 1 infection is associated with histopathological changes in immunocompetent hosts, but clinical disease is prevented by STAT1-dependent interferon responses J Virol 81 3251 3263 17229692 42 Ward JM 2006 Pathology of immunodeficient mice with naturally occurring murine norovirus infection Toxicol Pathol 34 708 715 17074739 43 Guix S 2007 Norwalk virus RNA is infectious in mammalian cells J Virol 81 12238 12248 17855551 44 Orchard RC 2016 Discovery of a proteinaceous cellular receptor for a norovirus Science 353 933 936 27540007 45 Haga K 2016 Functional receptor molecules CD300lf and CD300ld within the CD300 family enable murine noroviruses to infect cells Proc Natl Acad Sci USA 113 E6248 E6255 27681626 46 Borrego F 2013 The CD300 molecules: an emerging family of regulators of the immune system Blood 121 1951 1960 23293083 47 Tian L 2016 Enhanced efferocytosis by dendritic cells underlies memory T-cell expansion and susceptibility to autoimmune disease in CD300f-deficient mice Cell Death Differ 23 1086 1096 26768664 48 Albino D 2012 ESE3/EHF controls epithelial cell differentiation and its loss leads to prostate tumors with mesenchymal and stem-like features Cancer Res 72 2889 2900 22505649 49 Currier RL 2015 Innate susceptibility to norovirus infections influenced by FUT2 genotype in a United States pediatric population Clin Infect Dis 60 1631 1638 25744498 50 Lindesmith L 2003 Human susceptibility and resistance to Norwalk virus infection Nat Med 9 548 553 12692541 51 Tan M Jiang X 2014 Histo-blood group antigens: a common niche for norovirus and rotavirus Expert Rev Mol Med 16 e5 24606759 52 Sestak K 2014 Role of histo-blood group antigens in primate enteric calicivirus infections World J Virol 3 18 21 25392814 53 Kambhampati A 2016 Host genetic susceptibility to enteric viruses: a systematic review and metaanalysis Clin Infect Dis 62 11 18 26508510 54 Atmar RL 2011 Norovirus vaccine against experimental human Norwalk virus illness N Engl J Med 365 2178 2187 22150036 55 Ramani S 2015 Mucosal and cellular immune responses to Norwalk virus J Infect Dis 212 397 405 25635121 56 Lindesmith LC 2015 Broad blockade antibody responses in human volunteers after immunization with a multivalent norovirus VLP candidate vaccine: immunological analyses from a phase I clinical trial PLoS Med 12 e1001807 25803642 57 Chen Z 2013 Development of Norwalk virus-specific monoclonal antibodies with therapeutic potential for the treatment of Norwalk virus gastroenteritis J Virol 87 9547 9557 23785216 58 Green KY 2014 Norovirus infection in immunocompromised hosts Clin Microbiol Infect 20 717 723 25040790 59 Riddle MS Walker RI 2016 Status of vaccine research and development for norovirus Vaccine 34 2895 2899 27036510 60 Thackray LB 2012 Critical role for interferon regulatory factor 3 (IRF-3) and IRF-7 in type I interferon-mediated control of murine norovirus replication J Virol 86 13515 13523 23035219 61 Hwang S 2012 Nondegradative role of Atg5-Atg12/Atg16L1 autophagy protein complex in antiviral activity of interferon gamma Cell Host Microbe 11 397 409 22520467 62 Nice TJ 2016 Type I interferon receptor deficiency in dendritic cells facilitates systemic murine norovirus persistence despite enhanced adaptive immunity PLoS Pathog 12 e1005684 27327515 63 Zhu S 2013 Identification of immune and viral correlates of norovirus protective immunity through comparative study of intra-cluster norovirus strains PLoS Pathog 9 e1003592 24039576 64 McCartney SA 2008 MDA-5 recognition of a murine norovirus PLoS Pathog 4 e1000108 18636103 65 Wang P 2015 Nlrp6 regulates intestinal antiviral innate immunity Science 350 826 830 26494172 66 Maloney NS 2012 Essential cell-autonomous role for interferon (FN) regulatory factor 1 in IFN-gamma-mediated inhibition of norovirus replication in macrophages J Virol 86 12655 12664 22973039 67 Nice TJ 2015 Interferon-lambda cures persistent murine norovirus infection in the absence of adaptive immunity Science 347 269 273 25431489 68 Baldridge MT 2015 Commensal microbes and interferon-lambda determine persistence of enteric murine norovirus infection Science 347 266 269 25431490 69 Muir AJ 2014 A randomized phase 2b study of peginterferon lambda-1a for the treatment of chronic HCV infection J Hepatol 61 1238 1246 25064437 70 Qu L 2016 Replication of human norovirus RNA in mammalian cells reveals lack of interferon response J Virol 90 8906 8923 27466422 71 Cadwell K 2010 Virus-plus-susceptibility gene interaction determines Crohn’s disease gene Atg16L1 phenotypes in intestine Cell 141 1135 1145 20602997 72 Basic M 2014 Norovirus triggered microbiota-driven mucosal inflammation in interleukin 10-deficient mice Inflamm Bowel Dis 20 431 443 24487272 73 Chachu KA 2008 Antibody is critical for the clearance of murine norovirus infection J Virol 82 6610 6617 18417579 74 Chachu KA 2008 Immune mechanisms responsible for vaccination against and clearance of mucosal and lymphatic norovirus infection PLoS Pathog 4 e1000236 19079577 75 Tomov VT 2013 Persistent enteric murine norovirus infection is associated with functionally suboptimal virus-specific CD8 T cell responses J Virol 87 7015 7031 23596300 76 Zhu S 2016 Norovirus antagonism of B-cell antigen presentation results in impaired control of acute infection Mucosal Immunol 9 1559 1570 27007673 77 Li D 2015 Binding to histo-blood group antigen-expressing bacteria protects human norovirus from acute heat stress Front Microbiol 6 659 26191052 78 Miura T 2013 Histo-blood group antigen-like substances of human enteric bacteria as specific adsorbents for human noroviruses J Virol 87 9441 9451 23804639 79 Taube S 2009 Ganglioside-linked terminal sialic acid moieties on murine macrophages function as attachment receptors for murine noroviruses J Virol 83 4092 4101 19244326 80 Taube S 2012 Murine noroviruses bind glycolipid and glycoprotein attachment receptors in a strain-dependent manner J Virol 86 5584 5593 22438544 81 Bally M 2012 Norovirus GII. 4 virus-like particles recognize galactosylceramides in domains of planar supported lipid bilayers Angew Chem Int Ed Engl 51 12020 12024 23097253 82 Rydell GE 2009 Human noroviruses recognize sialyl Lewis x neoglycoprotein Glycobiology 19 309 320 19054801 83 Tamura M 2004 Genogroup II noroviruses efficiently bind to heparan sulfate proteoglycan associated with the cellular membrane J Virol 78 3817 3826 15047797 84 Almagro-Moreno S Boyd EF 2010 Bacterial catabolism of nonulosonic (sialic) acid and fitness in the gut Gut Microbes 1 45 50 21327116 85 DeMinicis S 2014 Dysbiosis contributes to fibrogenesis in the course of chronic liver injury in mice Hepatology 59 1738 1749 23959503 86 Compare D 2016 The gut bacteria-driven obesity development Dig Dis 34 221 229 27028448 87 Nelson AM 2012 Disruption of the human gut microbiota following norovirus infection PLoS ONE 7 e48224 23118957 88 Hickman D 2014 The effect of malnutrition on norovirus infection mBio 5 e01032 e1113 24595373 89 Nelson AM 2013 Murine norovirus infection does not cause major disruptions in the murine intestinal microbiota Microbiome 1 7 24451302 90 Pettigrew MM 2016 The human microbiota: novel targets for hospital-acquired infections and antibiotic resistance Ann Epidemiol 26 342 347 26994507 91 Agnihothram SS 2015 Infection of murine macrophages by Salmonella enterica serovar Heidelberg blocks murine norovirus infectivity and virus-induced apoptosis PLoS ONE 10 e0144911 26658916 92 Li D 2016 Anti-viral effect of Bifidobacterium adolescentis against noroviruses Front Microbiol 7 864 27375585 93 Lee H Ko G 2016 Antiviral effect of vitamin A on norovirus infection via modulation of the gut microbiome Sci Rep 6 25835 27180604 94 Osborne LC 2014 Coinfection. Virus-helminth coinfection reveals a microbiota-independent mechanism of immuno-modulation Science 345 578 582 25082704 95 Lencioni KC 2008 Murine norovirus: an intercurrent variable in a mouse model of bacteria-induced inflammatory bowel disease Comp Med 58 522 533 19149409 96 Lencioni KC 2011 Lack of effect of murine norovirus infection on a mouse model of bacteria-induced colon cancer Comp Med 61 219 226 21819691 97 Hsu CC 2014 Infection with murine norovirus 4 does not alter Helicobacter-induced inflammatory bowel disease in Il10−/−mice Comp Med 64 256 263 25296012 98 Kim YG 2011 Viral infection augments Nod1/2 signaling to potentiate lethality associated with secondary bacterial infections Cell Host Microbe 9 496 507 21669398 99 Kernbauer E 2014 An enteric virus can replace the beneficial function of commensal bacteria Nature 516 94 98 25409145 100 Abt MC 2016 TLR-7 activation enhances IL-22-mediated colonization resistance against vancomycin-resistant enterococcus Sci Transl Med 8 327ra25 101 Thepaut M 2015 Protective role of murine norovirus against Pseudomonas aeruginosa acute pneumonia Vet Res 46 91 26338794

PMC005xxxxxx/PMC5135625.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 7612148 6787 Psychoneuroendocrinology Psychoneuroendocrinology Psychoneuroendocrinology 0306-4530 1873-3360 27810703 5135625 10.1016/j.psyneuen.2016.10.022 NIHMS826480 Article Histone deacetylase and acetyltransferase inhibitors modulate behavioral responses to social stress McCann Katharine E. * Rosenhauer Anna M. Jones Genna M.F. Norvelle Alisa Choi Dennis C. Huhman Kim L. Neuroscience Institute, Georgia State University, 161 Jesse Hill Jr. Drive, Atlanta, GA 30303, USA arosenhauer1@student.gsu.edu (A.M. Rosenhauer), gjones26@student.gsu.edu (G.M.F. Jones), anorvelle@gsu.edu (A. Norvelle), dchoi15@gsu.edu (D.C. Choi), khuhman@gsu.edu (K.L. Huhman). * Corresponding author. kmccann3@gsu.edu (K.E. McCann) 6 11 2016 26 10 2016 1 2017 01 1 2018 75 100109 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. Histone acetylation has emerged as a critical factor regulating learning and memory both during and after exposure to stressful stimuli. There are drugs that we now know affect histone acetylation that are already in use in clinical populations. The current study uses these drugs to examine the consequences of acutely increasing or decreasing histone acetylation during exposure to social stress. Using an acute model of social defeat in Syrian hamsters, we systemically and site-specifically administered drugs that alter histone acetylation and measured subsequent behavior and immediate-early gene activity. We found that systemic administration of a histone deacetylase inhibitor enhances social stress-induced behavioral responses in males and females. We also found that systemic administration completely blocks defeat-induced neuronal activation, as measured by Fos-immunoreactivity, in the infralimbic cortex, but not in the amygdala, after a mild social defeat stressor. Lastly, we demonstrated that site-specific administration of histone deacetylase inhibitors in the infralimbic region of the prefrontal cortex, but not in the basolateral amygdala, mimics the systemic effect. Conversely, decreasing acetylation by inhibiting histone acetyltransferases in the infralimbic cortex reduces behavioral responses to defeat. This is the first demonstration that acute pharmacological manipulation of histone acetylation during social defeat alters subsequent behavioral responses in both males and females. These results reveal that even systemic administration of drugs that alter histone acetylation can significantly alter behavioral responses to social stress and highlight the importance of the infralimbic cortex in mediating this effect. Valproic acid Social defeat Epigenetics Medial prefrontal cortex 1. Introduction DNA transcription is necessary for development and maintenance of experience-dependent, long-term memories that elicit subsequent changes in behavior. The removal or addition of acetyl groups to the histones around which DNA is wrapped by histone deacetylases (HDACs) or histone acetyltransferases (HATs) alters the likelihood of gene transcription. Inhibition of Class I HDACs enhances long-term memory at each stage of memory processing (e.g., acquisition, consolidation, reconsolidation, extinction), while HAT inhibition impairs memory (Kilgore et al., 2010; Reolon et al., 2011). For example, acquisition of conditioned fear is enhanced following the administration of a Class I HDAC inhibitor, as is reconsolidation of that memory (Bredy and Barad, 2008), while inhibition of HATs during fear conditioning blocks acquisition and consolidation of that fear memory (Maddox et al., 2013; Monsey et al., 2015). HDAC inhibitors, including valproic acid (VPA), are already being used clinically to treat a variety of illnesses such as epilepsy and bipolar disorder, but their effects on learning suggest that they may also be useful in a range of neuropsychiatric illnesses, such as posttraumatic stress disorder (PTSD) or specific phobia, wherein fear learning is potentially aberrant (Bredy and Barad, 2008; Parsons and Ressler, 2013). Further investigation into how these drugs impact long-term behavioral and physiological reactions to stress may lead us to the development of more targeted treatments and interventions, many of which could be immediately available for clinical populations. While the initial data are encouraging, most studies completed to date have used physical stressors (e.g., foot/tail shock) and only a few studies have examined the role of histone acetylation in more ethologically relevant models of stress-induced behavioral change (Covington et al., 2015; Espallergues et al., 2012; Hollis et al., 2011). Social defeat models have strong face and construct validity for human anxiety and depressive behavior (Hollis and Kabbaj, 2014; Huhman, 2006; Toth and Neumann, 2013), but the majority of these models use relatively severe, repeated exposure to social defeat in male mice. While the study of chronic social stress is important, not all social stressors that humans experience are chronic in nature. Acute social stress or trauma can also lead to sudden and discernable changes in behavior, sometimes leading to psychopathology (e.g., PTSD). Furthermore, using an acute model of social stress allows a much more precise determination of when acquisition and consolidation of stress-related learning are occurring, therefore we can test hypotheses about these processes in a way that is not possible in chronic models. Our laboratory studies acute social defeat stress in Syrian hamsters. Hamsters provide a unique social stress model because both males and females are highly territorial, and these animals do not require complex housing conditions to elicit conspecific aggression or reliable behavioral responses to defeat in the laboratory. Home cage animals of both sexes will readily attack an intruding conspecific. However, after losing one agonistic encounter hamsters abandon all territorial aggression and, instead, become highly submissive and socially avoidant (Huhman, 2006; McCann et al., 2014; McCann and Huhman, 2012). This behavioral change has been termed conditioned defeat and lasts for at least one month in the majority of hamsters (Huhman et al., 2003). The conditioned defeat model is unique among social defeat models for several reasons. First, the agonistic interactions in hamsters are highly ritualized so that they rarely result in physical injury; thus, it is possible to examine the behavioral and physiological effects of social stress in the absence of physical injury or trauma and the concomitant inflammatory response. In addition, striking behavioral and physiological changes, including social avoidance and elevated cortisol, are observed after even a single, relatively mild defeat (Huhman et al., 1991, 2003; McCann and Huhman, 2012). Finally, unlike models using rats or mice, conditioned defeat in hamsters allows examination of defeat-induced behavior in both sexes. Thus, our model of acute social stress provides an excellent opportunity to study the behavioral and physiological responses in both males and females with much more precise temporal specificity compared with chronic defeat models that use only males, that test only extended periods of social stress, or that use species wherein wounding is common during social interactions. We have made significant progress in delineating the neural circuitry mediating conditioned defeat. It is well established that the amygdala is a crucial site of plasticity necessary for processing and responding to emotional and fearful stimuli (Davis, 1992; Fanselow and Gale, 2003; McGaugh, 2004), and we have demonstrated that the basolateral amygdala (BLA) as well as the medial prefrontal cortex (mPFC) are critical components of the neural circuit mediating conditioned defeat (Jasnow and Huhman, 2001; Jasnow et al., 2005; Markham et al., 2012, 2010). The persistence of the behavioral changes observed after a single social defeat in hamsters suggests that these behavioral changes might be mediated by epigenetic mechanisms. A better understanding of the molecular mechanisms subserving conditioned defeat may lead us to a clearer understanding of how even brief exposure to social stress impacts future social behavior. The purpose of the present study was to test the hypothesis that epigenetic changes within the neural circuit that mediates conditioned defeat contribute to the observed behavioral changes after acute social stress. 2. Material and methods 2.1. Animals Adult male and female Syrian hamsters (Mesocricetus auratus) were obtained from Charles River Laboratories (Wilmington, MA) or bred in-house from animals obtained from Charles River. Subjects (approximately 12 weeks, 120–130 g) were individually housed in a polycarbonate cage (23 × 43 × 20 cm) and were handled daily for at least one week before any behavioral manipulations began. The colony room was temperature-controlled, and animals were kept on a 14:10 light/dark cycle. All cages contained corncob bedding and cotton nesting material, and food and water were available ad libitum. Same sex resident aggressors (RAs) were used for social defeat training and for social avoidance testing. RAs are larger, individually housed hamsters that readily attack an intruder placed in their home cage. Female subjects were paired with ovariectomized female RAs because aggression in intact females varies over the estrous cycle and aggression in ovariectomized females is more reliable. Behavioral manipulations were done in a dedicated testing suite within the vivarium during the first 3 h of the dark phase of the daily light/dark cycle. All procedures and protocols were approved by the Georgia State University Institutional Animal Care and Use Committee and are in accordance with the standards outlined in the National Institutes of Health Guide for Care and Use of Laboratory Animals. 2.2. Social defeat training For social defeat training, subjects were placed into the home cage of a same-sex RA as described previously (McCann et al., 2014; McCann and Huhman, 2012). Estrous cycles of female subjects were monitored via vaginal swabs for at least two cycles before the experiment, and females were defeated on Diestrus 1 (D1) and tested on Diestrus 2 (D2) because we have previously shown this results in the most pronounced avoidance after social defeat (unpublished observations). A clear cage top was placed on top of the RA's cage to prevent either animal from escaping the cage during a 5 min or 15 min defeat session. The two different training durations were chosen to avoid ceiling and floor effects, respectively, and the choice of which to use was based on a priori hypotheses of the directionality of the expected behavioral effect. Shorter defeat was used when submission and avoidance was expected to increase and longer defeat was used when submission and avoidance was expected to decrease in drug-treated animals as compared with vehicle controls. The holding box used for social avoidance testing, described below, was placed in the RA's cage during training. At the end of the defeat, subjects were returned to their home cages. Animals were monitored during defeat to ensure that no injury occurred to either animal. No-defeat controls were placed in a novel cage with soiled RA bedding and a holding box for the same amount of time as the defeat group and were subsequently returned to their home cage until social avoidance testing. Behavior emitted by RAs and by subjects during defeat training was recorded and scored by trained observers that were blind to experimental condition to ensure that pre-training drug infusions did not alter either the amount of aggression displayed by the RAs toward the subjects or the amount of submission shown by the subjects during defeat training. 2.3. Social avoidance testing Social avoidance testing was conducted as described previously (McCann et al., 2014; McCann and Huhman, 2012) and was recorded for later analysis. In brief, 24 h after social defeat training, subjects were placed in a clean, novel testing arena (23 × 40 × 20 cm) with an unfamiliar RA placed inside a smaller holding box on one end of the arena. The holding box for the unfamiliar RA was constructed of perforated plastic that allowed the subject to see, hear, and smell the unfamiliar stimulus animal but not to come into direct contact with it. For scoring purposes, the testing arena was divided into eight sections (Fig. 1). Time spent in the far half of the testing arena (operationally defined as avoidance) as well as total number of line crosses (a measure of locomotor behavior) were scored. A line cross was counted when the subject's head and both front paws crossed over a line. Frequencies of specific behaviors (e.g., flees, risk assessments, flank marks), as defined previously (McCann and Huhman, 2012; Song et al., 2014), were also counted. 2.4. Cannulation and microinjections For site-specific injections, males were implanted with bilateral cannulae targeting the BLA or with a unilateral cannula targeting the mPFC. Coordinates for guide cannulae used to target the BLA and mPFC were measured from bregma and were as follows for BLA: +0.0AP, ±4.0ML, −3.0DV from dura perpendicular, and for mPFC: +3.0AP, ±1.6ML, −3.2DV from dura at a 20° angle toward the midline to avoid the central sinus. Anesthesia was induced with 5% isoflurane, and animals were maintained at 3–5% isoflurane in a stereotaxic apparatus for the entire surgical procedure. Animals were handled for 1 week after surgery before any experimental manipulations. The compounds and concentrations listed below were injected directly into the site of interest using an infusion pump (Harvard Apparatus) and a Hamilton syringe connected to an injection needle by 50-gauge polyethylene tubing. In order to minimize damage to the area being injected, a shorter guide cannula (26-gauge) was used, and the final depth was reached with a smaller (33-gauge) injection needle that projected from the guide cannula (BLA: 3.3 mm below the guide; mPFC: 1.2 mm below the guide). The injection needle was left in the cannula guide for 1 min post-injection to ensure diffusion of the pharmacological agent from the needle tip. Successful injections were inferred if solution flowed easily from the needle before and after injection and a small air bubble placed between the drug and the saline solution in the tubing moved during microinjection. 2.5. Pharmacological agents VPA (Sigma-Aldrich, St. Louis, MO) was dissolved in physiological saline. Intraperitoneal (IP; 100 mg/kg, 200 mg/kg, 300 mg/kg) as well as site-specific (100 μg/0.2 μl) injections of VPA were given (Bredy and Barad, 2008; Heinrichs et al., 2013; Kilgore et al., 2010; Kim et al., 2008; Nau and Loscher, 1982). IP injections were administered 2 h before defeat training because it has been clearly established that peak brain histone acetylation occurs 2 h after peripheral administration (Arent et al., 2011; Bredy and Barad, 2008; Bredy et al., 2007; Ploense et al., 2013; Tremolizzo et al., 2002). To test the temporal specificity of peripherally administered VPA in our model, we also completed two control experiments, one in which we administered VPA 1 h before defeat training and another in which we gave VPA 2 h before avoidance testing. These are both time frames in which other pharmacodynamic mechanisms of VPA would be active and are thus good controls to determine if observed effects are likely to be due to changes in histone acetylation versus other actions of VPA. Sodium butyrate (NAB; Alfa Aesar, Ward Hill, MA), another HDAC inhibitor, was used in order to further validate that the observed behavioral effects were indeed the result of HDAC inhibition and not another, more rapid effect of VPA administration such as an enhancement in GABAergic signaling. NAB was given IP (600 mg/kg, 1200 mg/kg in physiological saline) to a small subset of animals, but it induced a temporary and extreme ataxia, so systemic use was discontinued, and it was only tested site-specifically (1.32 μg/0.2 μl) (Blank et al., 2014; Heinrichs et al., 2013; Kilgore et al., 2010; Lattal et al., 2007; Mahan et al., 2012; Simon-O'Brien et al., 2015). Finally, Curcumin (Epigentek, Farmingdale, NY, 1.1 μg/0.2 μl) was dissolved in 55% DMSO. This drug was chosen because it appears to be one of the few, if not only, HAT inhibitors that is currently commercially available that does not have to be dissolved in 100% DMSO. All site-specific injections were given 30 min before social defeat (Simon-O'Brien et al., 2015; Xing et al., 2011) at a total volume of 0.2 μl to limit the spread of the injection. 2.6. Histology After social avoidance testing, cannulated animals were given an overdose of sodium pentobarbital, and 0.2 μl of ink, to match the volume of drug administration, was injected through the guide cannulae for the purpose of site verification. Brains were sectioned on a cryostat and stained with neutral red for microscopic analysis of cannula placement. Placements more than 300 μm from the target nucleus were used as anatomical, or “miss”, controls to assess site specificity of the drug effects. 2.7. Immunohistochemistry for immediate-early gene c-fos Animals were given IP injections of either saline or VPA (200 mg/kg) 2 h before a 5 min defeat and were perfused 1 h after the defeat. Postfixed brains were sectioned on a cryostat into cryoprotectant and were stored at −20°C until processing. On Day 1, sections were washed 3 × 5 min with potassium phosphate buffered saline (KPBS) and incubated in 0.3% hydrogen peroxide in KPBS for 30 min. Sections were washed again 3 × 5 min in KPBS and incubated with primary c-fos antibody (rabbit polyclonal IgG, 1:5000, Santa Cruz Biotechnology, Dallas, TX) in KPBS with 1% TritonX-100 and 1% normal goat serum overnight at room temperature. On Day 2, sections were washed 3 × 5 min with KPBS and incubated with 0.4% secondary (biotin-SP-conjugated AffiniPure goat anti-rabbit IgG, Jackson ImmunoResearch, West Grove, PA) in KPBS-T for 90 min at room temperature. Sections were again washed 3 × 5 min in KPBS and then incubated in pre-prepared avidin/biotin blocking solution (Vector Laboratories, Burlingame, CA) at room temperature for 1 h. After incubation, sections were washed 3 × 5 min with KPBS and then incubated in 3,3-diaminobenzidine (Vector Laboratories, Burlingame, CA) for 2–5 min. Sections were rinsed 2 × 5 min in KPBS, mounted using 0.15% gelatin in dH2O and allowed to dry overnight. Sections were then dehydrated for 2 min each in EtOH 50%, 70%, 95%, and 10 min in 100% EtOH, followed by 30 min in Citrosolv and then coverslipped with DPX. For analysis, a template was created for each region of interest and immunoreactive-positive cells within this area were counted using NIH ImageJ software. Bilateral counts from two or three sections per animal were averaged for each brain area. 2.8. Statistical analysis Statistics for group comparisons were completed using SPSS for Windows (PASW Statistics 22.0). Student's t-tests or ANOVA with LSD post-hoc analysis were used for all analyses. All significant results reported here had a p-value of less than 0.05. Following statistical analysis, all avoidance data were graphed as percent of control for each experiment because baseline avoidance among the experiments was somewhat variable. This variability among experiments is to be expected, particularly given that some experiments involved a 5 min and others a 15 min defeat stressor. 3. Results 3.1. Systemic administration of an HDAC inhibitor before social stress enhances conditioned defeat learning in males VPA or saline was administered IP 2 h before defeat training, and 24 h later we measured social avoidance and submission in response to a caged stimulus animal. Following a 15 min defeat, there was no difference in social avoidance during testing among animals given VPA (regardless of dose) and those given saline (F(3,33) = 0.527, p = 0.667; Fig. 2a); however, animals receiving 200 mg/kg of VPA displayed a significant increase in the number of risk assessments (F(3,33) = 2.883, p = 0.05; Fig. 2b). VPA did not alter seconds of avoidance (vehicle: 72.67 ± 7.71, 100 mg/kg: 75.63 ± 11.06, 200 mg/kg: 60 ± 5.06, 300 mg/kg: 95.29 ± 19.08; p = 0.517) or number of risk assessments (vehicle: 0 ± 0, 100 mg/kg: 0.5 ± 0.38, 200 mg/kg: 0 ± 0, 300 mg/kg: 0.14 ± 0.14; p = 0.264) in no-defeat controls, suggesting that the increase in risk assessments observed in defeated animals given VPA was not a non-specific effect of the drug on agonistic or anxiety-like behavior. In the first experiment, all defeated animals, regardless of group, exhibited social avoidance when compared with no-defeat controls, suggesting a potential ceiling effect on avoidance following a 15 min defeat. To test this possibility, animals were given 200 mg/kg VPA (the dosage shown to increase risk assessment in the first experiment) or saline IP 2 h before a 5 min defeat. Animals given VPA before this shorter defeat experience exhibited both increased social avoidance (t(17) = −2.569, p = 0.02; Fig. 2c) and increased risk assessments (t(17) = −3.882, p = 0.001; Fig. 2d) during testing compared with animals given saline. Again, there was no effect of VPA on behavior of no-defeat controls during testing (seconds of avoidance, vehicle: 71.4 ± 8.68, VPA: 78.2 ± 1.95; p = 0.482; no risk assessments observed). To test the temporal specificity of VPA, animals were given VPA 1 h before social defeat training. Animals given drug did not differ in social avoidance (t(23) = 1.593, p = 0.125; Fig. 2e) or risk assessment during testing compared with animals given saline. To further determine if VPA-enhanced conditioned defeat was specific to the acquisition of the memory of defeat, we tested defeat-induced social avoidance in animals given VPA 2 h before social avoidance testing to examine whether increased histone acetylation in the brain also had an effect on the expression of conditioned defeat. There was no difference in avoidance displayed by animals given VPA or saline (t(10) = 0.15, p = 0.883; Fig. 2f). 3.2. Systemic administration of VPA also enhances conditioned defeat learning in females Subjects in the above experiments were males, and the purpose of the next experiment was to test if systemic VPA administration also enhances the acquisition of conditioned defeat in females. Like males, females given VPA (200 mg/kg) 2 h before a 5 min defeat displayed increased social avoidance (t(11) = −2.609, p = 0.02) and risk assessments (t(11) = −2.972, p = 0.01) compared with females given saline (Fig. 3). VPA also significantly decreased flank marking exhibited by defeated females (t(11) = 2.328, p = 0.04). One animal receiving vehicle was removed from analysis because its avoidance score during testing was an outlier (z-score = 2.24). Again, there was no effect on behavior of no-defeat controls during testing (seconds of avoidance, vehicle: 88.75 ± 24.04, VPA: 93.25 ± 16.73; p = 0.883; no risk assessments observed), indicating that the behavioral effects of systemic HDAC inhibition were specific to defeated females. 3.3. Systemic administration of VPA decreases defeat-induced immediate-early gene activation in the mPFC We next used immunohistochemistry for c-fos to localize where systemically administered VPA might be acting within the neural circuit mediating conditioned defeat to enhance behavioral responses to a 5 min defeat. Fos-immunoreactive cells were counted in male hamsters in several nuclei that are crucial to conditioned defeat learning, including regions of the amygdala (basolateral, central, medial) (Fig. 4a) and mPFC (prelimbic, infralimbic) (Fig. 4b). Surprisingly, no differences from control were observed in the number of fos-positive cells in amygdala following defeat with or without HDAC inhibition (BLA: F(1,20) = 0.946, p = 0.342; central amygdala: F(1,20) = 0.556, p = 0.465; medial amygdala (F(1,20) = 0.154, p = 0.669; Fig. 4c). VPA blocked defeat-induced neuronal activation, however, in the mPFC of defeated animals that received systemic VPA (Fig. 4c), as evidenced by a decrease in the number of Fos-positive cells. There was a main effect of HDAC inhibition in the infralimbic cortex (IL) (F(1,20) = 4.897, p = 0.039) and a trend for a 5 min defeat, alone, to increase Fos activation (F(1,20) = 4.27, p = 0.052). While trending in the same direction as the IL, no main effects were observed in the prelimbic cortex (PL) (HDAC inhibition: F(1,20) = 3.075, p = 0.095; defeat: F(1,20) = 0.882, p = 0.359). 3.4. Site-specific HDAC inhibition in the mPFC, but not in the BLA, alters behavioral responses to social defeat To test if HDAC inhibition in specific regions of the conditioned defeat circuit enhances conditioned defeat learning, we next administered an HDAC inhibitor (either VPA or NAB) directly into specific nuclei. Not surprisingly given our lack of cellular activation in the amygdala after systemic HDAC inhibition, animals given drug in the BLA before defeat exhibited the same amount of avoidance (F(2,21) = 0.095; p = 0.91; Fig. 5a) as did animals given saline, further suggesting the role of the BLA in the acquisition of conditioned defeat may be independent of HDAC activity. Conversely, the activation changes observed in mPFC after systemic administration of VPA were paralleled by a behavioral effect of the drug in the mPFC. There was a main effect of HDAC inhibition in the mPFC on seconds of social avoidance exhibited during testing (F(2,16) = 4.897, p = 0.022; Fig. 5b). Animals given VPA displayed significantly more avoidance than did animals given saline (p = 0.006). Animals given NAB exhibited a strong trend towards increased avoidance over those given saline (p = 0.063) and did not differ from those given VPA (p = 0.218). There was no effect of central HDAC inhibition on seconds of avoidance exhibited by no-defeat controls (BLA, vehicle: 114.25 ± 20.14, VPA: 97.33 ± 9.71, NAB: 115.67 ± 9.82; p = 0.341; mPFC, vehicle: 85.8 ± 11.5, VPA: 92.75 ± 3.57, NAB: 95.75 ± 11.48; p = 0.768). Furthermore, HDAC inhibition in the anatomical (“miss”) controls (n = 3) for mPFC, which were located in the cingulate cortex more than 300 μm from the target nucleus, did not cause significant increases in social avoidance compared with controls (t(5) = −0.810, p = 0.455), supporting anatomical specificity of the drug effect. VPA given in the IL appeared to have a more robust effect on social avoidance (220.2s ± 22.28s, n = 5) than did VPA given in the PL (159s ± 30.57s, n = 3), but because this was not statistically significant (p = 0.151) these groups were collapsed for analysis. 3.5. HAT inhibition in the mPFC blocks the acquisition of conditioned defeat To test whether histone acetylation in the mPFC is necessary for behavioral responses to social defeat, we administered the HAT inhibitor Curcumin (Balasubramanyam et al., 2004; Kang et al., 2005) to determine if this treatment would decrease the acquisition of conditioned defeat (i.e., have the opposite effect of HDAC inhibition). Curcumin administration resulted in decreased avoidance when compared with vehicle (t(10) = 2.328, p = 0.042; Fig. 5c). For this experiment, we localized injections to the IL because HDAC inhibition had a more robust effect in this region, both systemically and site-specifically. Injections in the PL or the cingulate cortex (“miss” controls, n = 6) did not cause a significant decrease in avoidance when compared with animals receiving vehicle (t(8) = 1.795, p = 0.11), demonstrating anatomical specificity of the drug effect. 3.6. Nonspecific behavioral effects of HDAC and HAT inhibition To test if drug administration before social defeat altered the initial defeat experience and thereby caused the subsequent behavioral changes, we scored the behavior of both the aggressor and the subject during defeat training. Pharmacological manipulation of histone acetylation did not affect the amount of aggression shown by RAs during training nor the amount of submission shown by the subjects in any of the experiments. Furthermore, with the exception of animals given the highest dose of VPA in Experiment 1, drug manipulations did not affect locomotor activity during testing, as measured by number of line crosses (see Fig. 1). These data are shown in Table 1. 4. Discussion In summary, the data presented here suggest that manipulation of histone acetylation, even via systemically administered drugs, may offer a novel and effective way to alter behavioral responses to acute social stress in both males and females. The data further suggest that these treatments act, at least in part, via their action in the IL and emphasize the importance of prefrontal epigenetic regulation in mediating behavioral changes observed after exposure to acute social stress. Systemic administration of VPA before a single social defeat experience intensified subsequent behavioral responses to defeat. Our customary defeat procedure uses a 15 min, inescapable defeat. This is a relatively mild social stressor, but it is sufficient to lead to robust and quantifiable behavioral changes observed during subsequent testing (Gray et al., 2015; Jasnow and Huhman, 2001; McCann and Huhman, 2012). In our original experiment, we did not observe a change in social avoidance in animals given VPA, but this could be due to a ceiling effect. We did, however, observe a significant increase in risk assessment, which is a defensive/submissive behavior in which subjects cautiously stretch forward to investigate a potential threat. This increase in risk assessments suggests that there indeed was an increase in submission after systemic HDAC inhibition that was not captured by measuring seconds of avoidance. To test if systemic HDAC inhibition was in fact enhancing conditioned defeat learning, we next tested subjects using a shorter 5 min defeat. A 5 min defeat is not usually sufficient to elicit pronounced long-term behavioral changes; therefore, we reasoned that this defeat might provide a better starting point with which to discern possible enhancement. Using a 5 min defeat, we demonstrated that hamsters given systemic VPA did exhibit significant increases in social avoidance and risk assessment. Overall, these data demonstrate that a systemically administered HDAC inhibitor can enhance behavioral responses to acute social stress. Peripheral VPA crosses the blood brain barrier quickly, with peak concentrations of the drug found in the brain 15 min after administration, dropping to non-detectable levels at 8 h post-administration (Nau and Loscher, 1982). VPA is known to be a potent and reliable HDAC inhibitor (Gottlicher et al., 2001; Phiel et al., 2001) with peak acetylation occurring in brain 2 h after systemic administration (Tremolizzo et al., 2002). This timeframe has been established as a reliable timeframe for observing behavioral effects of brain HDAC inhibition as well (Bredy and Barad, 2008) and coincides with our main behavioral effect. VPA did not affect behavior when given 1 h before defeat, a time when the drug has already crossed the blood brain barrier but before peak brain acetylation occurs, nor when given before avoidance testing, supporting the role of histone acetylation in mediating the observed behavioral changes. There was also no effect of the drugs on no-defeat controls or on the behavior observed during defeat training when the drug is known to be in the brain and pharmacologically active. Together, these findings indicate that systemic VPA specifically enhances the acquisition of the memory of a mild social defeat stressor and that this effect coincides with the time of known peak brain acetylation. Our site-specific microinjections offer further support for a role of histone acetylation in the behavioral changes observed in response to acute social defeat. It has previously been demonstrated that ventricular and intra-mPFC administration of VPA or NAB decreases HDAC activity in the mPFC (Arent et al., 2011). Here, we demonstrate that infusion of HDAC inhibitors in the mPFC enhances conditioned defeat learning while administration of a HAT inhibitor impairs it. The opposing behavioral effects of HDAC and HAT inhibition in the mPFC strongly support a role of histone acetylation in this nucleus in mediating the behavioral responses to acute social stress. Recent evidence has demonstrated that chronic administration of an HDAC inhibitor into the mPFC after repeated defeat also decreases social avoidance (Covington et al., 2015). Together with our current data, these findings highlight an important role for epigenetic regulation in the mPFC, both during and after defeat, in modifying behavioral responses to social stress. We have previously demonstrated that the BLA is critical for acquisition and expression of conditioned defeat (Jasnow and Huhman, 2001; Markham et al., 2010). Temporary inactivation of this nucleus with a GABA-A receptor agonist blocks the acquisition and expression of defeat-induced behavioral changes (Jasnow and Huhman, 2001; Markham et al., 2010) as does an NMDA glutamate receptor antagonist (Jasnow et al., 2004), and de novo protein synthesis in this nucleus is necessary for the behavioral changes characterizing conditioned defeat (Markham and Huhman, 2008). We were thus surprised to find that acute HDAC inhibition within the BLA did not affect the acquisition of conditioned defeat. There is evidence, however, that ventricular administration of VPA or NAB and intra-amygdalar administration of VPA does not decrease HDAC activity in the BLA (Arent et al., 2011). Thus, it is entirely possible that our treatment with these two known HDAC inhibitors did not, in fact, alter histone acetylation in the BLA. Another prominent use for VPA is as an anticonvulsant or a mood stabilizer because of the drug's pharmacodynamic effect of increasing GABAergic neurotransmission (Nau and Loscher, 1982; Tunnicliff, 1999). While it might seem possible that some of the observed behavioral effects in this study result from an increase in GABA signaling, it is important to note that the enhanced avoidance and submission observed after acute systemic HDAC inhibition is specific to the time of peak brain histone acetylation. Histone acetylation (specifically at H3) reaches a peak 2 h after systemic administration, corresponding with our main behavioral effect, whereas increased GABA signaling in the brain is observed within 15 min after systemic VPA and remains elevated for up to 8 h (Nau and Loscher, 1982). We demonstrated that there was no effect of VPA on behavior when the drug was given 1 h before social defeat, a time when GABA signaling in the brain is enhanced, nor when it was given before avoidance testing, a time when GABAergic receptor agonists potently inhibit the expression of conditioned defeat (Jasnow and Huhman, 2001). Similarly, if VPA were acting primarily via a GABAergic mechanism, then we would certainly expect to see a decrease in the acquisition of conditioned defeat when VPA was given in the BLA as is seen when a GABA-A agonist is administered there (Jasnow and Huhman, 2001). Finally, NAB administration, which does not directly affect GABA signaling, caused a similar enhancement of defeat-induced behavior to VPA, while HAT inhibition in the IL, which reduces histone acetylation, reduced the acquisition of conditioned defeat. The primary shared mechanism of action for these drugs is their direct effect on the enzymes that regulate histone acetylation. Together, these data strongly dispute the argument that the observed behavioral changes resulted from an effect of VPA on GABAergic signaling and, instead strongly support the hypothesis that histone acetylation is an epigenetic mechanism contributing to the observed behavioral changes after exposure to acute social stress. Furthermore, we observed less defeat-induced cellular activation, as measured by Fos-immunoreactivity, in the IL after systemic VPA administration compared with saline. No other brain region analyzed exhibited differential Fos-immunoreactivity after HDAC inhibition. Interestingly, the activation patterns in the PL were similar to that observed in the IL, although they did not reach statistical significance. This is perhaps surprising given the literature indicating that these two regions of the mPFC have opposing roles in the expression of conditioned fear (Sierra-Mercado et al., 2011). Our results, using drugs that target histone acetylation, however, suggest that these nuclei may have similar roles in mediating the acquisition of behavioral responses to acute social stress. We have shown previously that Fos-immunoreactivity increases in the BLA after a 15 min social defeat (Markham et al., 2010); here, we show that a 5 min defeat is not sufficient to increase immediate-early gene activation in the amygdala. It is notable that there was a trend for defeat to increase Fos activation in the IL, suggesting that the IL is sensitive even to an extremely mild, 5 min social defeat stressor. We were also at first surprised to find that systemic administration of an HDAC inhibitor decreased fos-immunoreactivity in the IL. HDAC inhibition is associated with an increase in acetylation and thus an increase in gene transcription. It is important to note, however, that fos-immunoreactivity is not necessarily indicative of activation of excitatory neurons. The IL has strong inhibitory connections to the BLA and, although we did not see a corresponding increase in Fos-immunoreactivity in the BLA, it is possible that the decrease in defeat-induced neural activity in IL following HDAC inhibition leads to disinhibition of a subset of BLA neurons and that this is the mechanism by which the acquisition of conditioned defeat is enhanced after systemic or intra-mPFC HDAC inhibition. Finally, one of the unique benefits of using hamsters to model social stress is that females also display agonistic behavior to same-sex conspecifics. Females are often overlooked in other translational models of social stress because of the difficulty in eliciting spontaneous female aggression in rats and mice. Female hamsters typically exhibit more aggression during agonistic encounters than do males, and their expression of conditioned defeat after losing a fight may be less marked than that observed in males when tested with a non-aggressive opponent (Huhman et al., 2003). Defeated females, however, respond to a caged opponent with similar social avoidance to that observed in males. Here, we demonstrate that systemic VPA also has the same social stress-promoting effect in females as it does in males. Interestingly, VPA also reduced the number of flank marks in defeated females. Flank marking is a mode of social communication in which a hamster rubs its flank glands along the wall of the cage. This behavior is produced more often by dominant animals and is thought to communicate information about social status (Albers and Prishkolnik, 1992; Ferris et al., 1987). There are also significant sex differences in flank marking, with females flank marking more often than do males. In the current study, males exhibited few, if any, flank marks (mean of less than 1 flank mark per animal during a 5 min test), while most females marked during testing. The decrease in flank marking observed in the defeated females given VPA is an additional measure of submission or loss of territoriality. Together, these data are the first to show that systemic HDAC inhibition in both males and females enhances the acquisition of stress-induced behavioral changes following acute social defeat. 5. Conclusions The current study focused on the effect of acute HDAC or HAT inhibition during the experience of a mild social stressor. Social stress is particularly relevant in that it is argued to be the most common stressor experienced by humans (Bjorkqvist, 2001), and perceptions of social defeat are strongly associated with depression, anxiety disorders, social withdrawal, and submissiveness (Agid et al., 2000; Heim and Nemeroff, 2001; Nemeroff, 1998). Understanding the role that histone acetylation plays in the acquisition of socially relevant fear memories could be an important step in elucidating the molecular mechanisms underlying stress-related neuropsychiatric diseases including mood and anxiety disorders and in potentially developing better treatments to alter maladaptive behavioral responses to stressful events. Increased focus on interventions involving HAT inhibition may hold particular translational relevance because this treatment reduced behavioral responses to acute social stress. Furthermore, while systemic administration lacks anatomical resolution, it has valuable translational implications for the potential usefulness of the drugs for clinical populations, particularly when there are drugs that are already FDA-approved. The data presented here demonstrate for the first time that altering the enzymes that regulate histone acetylation, even systemically, can alter behavioral responses to acute social stress in both males and females. Acknowledgements The authors would like to acknowledge AD Guzman Bambaren, BM Thompson, KA Partrick, and TM Kahl for their assistance with this project. Role of funding source Research reported here was supported by the National Institute of Mental Health of the National Institutes of Health under Award Number R01MH062044 awarded to KLH and a Brains and Behavior Fellowship, a Honeycutt Fellowship, and a Dissertation Grant from Georgia State University awarded to KEM. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health or Georgia State University. Fig. 1 Schematic of testing arena. Dotted lines represent line markers for scoring subjects’ movements during the 5 min testing period. Time spent in the far half of the testing arena (designated by the black dotted line) was operationally defined as avoidance. Fig. 2 Systemic administration of VPA enhances conditioned defeat learning in males. Systemic VPA did not increase (A) social avoidance when given before a 15 min defeat regardless of drug dose (0 mg/kg, 100 mg/kg, 200 mg/kg, 300 mg/kg); however, animals given 200 mg/kg VPA exhibited an increase during testing in the number of (B) risk assessments (p = 0.041 compared with saline). When given 2 h before a 5 min defeat training session, systemic VPA increased both (C) social avoidance and (D) number of risk assessments observed during testing 24 h later. VPA (200 mg/kg) did not alter social avoidance when given either (E) 1 h before defeat or when given (F) 2 h before testing. *p < 0.05 compared with vehicle. Fig. 3 Systemic administration of VPA also enhances acquisition of conditioned defeat in females. VPA (200 mg/kg) increased defeat-induced social avoidance and risk assessments and decreased flank marking in females compared with their controls given saline. *p < 0.05. Fig. 4 Systemic HDAC inhibition modulates neural activity in the infralimbic cortex. Animals (n = 6 per group) were given VPA (200 mg/kg) or saline 2hr before a 5 min defeat and were sacrificed 1 h after defeat. Photomicrographs illustrate the sub-regions wherein Fos-positive cells were quantified: (A) amygdala (BLA: basolateral, CEA: central, MEA: medial) and (B) mPFC (PL: prelimbic, IL: infralimbic). (C) No significant differences were found in the amygdala (BLA, p = 0.342; CEA, p = 0.465.; MEA, p = 0.699) or PL (p = 0.095), but animals given vehicle before defeat had significantly higher fos-immunoreactivity in the IL than did all other groups. Animals given VPA exhibited fos-immunoreactivity that was comparable to no-defeat controls. *p < 0.05. Fig. 5 HDAC and HAT inhibition in the mPFC, but not the BLA, modulate behavioral responses to social defeat. HDAC inhibition (valproic acid (VPA) or sodium butyrate (NAB)) in the (A) BLA before social defeat training did not alter social avoidance during testing 24 h later. HDAC inhibition in the (B) mPFC during social defeat training significantly increased social avoidance during testing, while (C) HAT inhibition (curcumin (CUR)) specifically in the IL decreased social avoidance. *p < 0.05, +p = 0.06 compared with vehicle. Table 1 Nonspecific behavioral effects of HDAC and HAT inhibition. Aggression by RA (s) Submission by Subject (s) # Line Crosses Experiment 1: Systemic VPA Aggression: p=0.172 Submission: p=0.446 Line crosses: p=0.004 Vehicle 304.45 ± 53.75 513.73 ± 58.41 88.55 ± 6.28 100mg/kg 155.86 ± 26.43 448.57 ± 60.78 96.86 ± 7.95 200mg/kg 190.64 ± 43.1 381.36 ± 56.27 87.45 ± 5.04 300mg/kg 259.13 ± 61.58 421.25 ± 77.57 63.25 ± 2.05* Experiment 2: Systemic VPA (5min defeat) Aggression: p=0.641 Submission: p=0.873 Line crosses: p=0.332 Vehicle 72.78 ± 15.02 120.33 ± 26.64 86.89 ± 3.9 VPA 64.80 ± 7.75 125.40 ± 17.27 94.2 ± 5.98 Experiment 3: Systemic VPA in females Aggression: p=0.612 Submission: p=0.394 Line crosses: p=0.12 Vehicle 85 ± 20.54 87 ± 24.92 79 ± 6 VPA 70.86 ± 17.91 122 ± 29.57 67.57 ± 3.62 Experiment 4: Control systemic VPA Aggression: p=0.194 Submission: p=0.753 Line crosses (1hr): p=0.082 Line crosses (expression): p=0.459 Vehicle 173 ± 33.66 323.08 ± 57.36 91.92 ± 4.98 81.83 ± 9.05 VPA 241.46 ± 37.45 354.31 ± 77.76 79.69 ± 4.55 71.83 ± 9.31 Experiment 5: Intra-BLA HDAC inhibition Aggression: p=0.369 Submission: p=0.899 Line crosses: p=0.092 Vehicle 104 ± 22.57 161 ± 23.03 63.71 ± 5.22 VPA 136.36 ± 22.31 152.27 ± 18.93 76.73 ± 4.78 NAB 93 ± 19.69 144.33 ± 29.96 80.67 ± 4.52 Experiment 6: Intra-PFC HDAC inhibition Aggression: p=0.317 Submission: p=0.113 Line crosses: p=0.694 Vehicle 289.75 ± 101.50 519.25 ± 83.53 65.25 ± 14.03 VPA 224.71 ± 35.31 561.71 ± 105.4 79.43 ± 13.01 NAB 165.67 ± 33.16 318.17 ± 21.44 67.71 ± 10.3 Experiment 7: Intra-PFC HAT inhibition Aggression: p=0.108 Submission: p=0.403 Line crosses: p=0.475 Vehicle 253.25 ± 27.2 386.25 ± 56.73 93.25 ± 7.35 Curcumin 153 ± 38.9 302 ± 64.18 80.5 ± 11.35 No differences in seconds of aggression produced by the RA or seconds of submission exhibited by the subject were observed between groups in any experiment. Animals exhibited no difference in locomotor activity, as measured by the number of line crosses, during social avoidance testing with the exception of animals given the highest dose of VPA in Experiment 1. While there were no obvious signs of ataxia, animals given 300 mg/kg VPA exhibited significantly fewer line crosses than all other groups in that experiment (*p ± 0.05). All data are shown as mean ± standard error of the mean. Contributors KEM and KLH designed all experiments; KEM conducted all experiments; AMR, GMFJ, AN, and DCC contributed to design and implementation of individual experiments; KEM and KLH prepared the manuscript; all authors edited and approved the manuscript. Conflicts of interest None References Agid O Kohn Y Lerer B Environmental stress and psychiatric illness. Biomed. Pharmacother 2000 54 135 141 10840590 Albers HE Prishkolnik J Sex differences in odor-stimulated flank marking in the golden hamster (Mesocricetus auratus). Horm. Behav 1992 26 229 239 1612567 Arent CO Valvassori SS Fries GR Stertz L Ferreira CL Lopes-Borges J Mariot E Varela RB Ornell F Kapczinski F Andersen ML Quevedo J Neuroanatomical profile of antimaniac effects of histone deacetylases inhibitors. Mol. Neurobiol 2011 43 207 214 21424678 Balasubramanyam K Varier RA Altaf M Swaminathan V Siddappa NB Ranga U Kundu TK Curcumin, a novel p300/CREB-binding protein-specific inhibitor of acetyltransferase, represses the acetylation of histone/nonhistone proteins and histone acetyltransferase-dependent chromatin transcription. J. Biol. Chem 2004 279 51163 51171 15383533 Bjorkqvist K Social defeat as a stressor in humans. Physiol. Behav 2001 73 435 442 11438372 Blank M Dornelles AS Werenicz A Velho LA Pinto DF Fedi AC Schroder N Roesler R Basolateral amygdala activity is required for enhancement of memory consolidation produced by histone deacetylase inhibition in the hippocampus. Neurobiol. Learn. Mem 2014 111 1 8 24583371 Bredy TW Barad M The histone deacetylase inhibitor valproic acid enhances acquisition, extinction, and reconsolidation of conditioned fear. Learn. Memory 2008 15 39 45 Bredy TW Wu H Crego C Zellhoefer J Sun YE Barad M Histone modifications around individual BDNF gene promoters in prefrontal cortex are associated with extinction of conditioned fear. Learn. Memory 2007 14 268 276 Covington HE 3rd Maze I Vialou V Nestler EJ Antidepressant action of HDAC inhibition in the prefrontal cortex. Neuroscience 2015 298 329 335 25907440 Davis M The role of the amygdala in fear and anxiety. Annu. Rev. Neurosci 1992 15 353 375 1575447 Espallergues J Teegarden SL Veerakumar A Boulden J Challis C Jochems J Chan M Petersen T Deneris E Matthias P Hahn CG Lucki I Beck SG Berton O HDAC6 regulates glucocorticoid receptor signaling in serotonin pathways with critical impact on stress resilience. J. Neurosci 2012 32 4400 4416 22457490 Fanselow MS Gale GD The amygdala, fear, and memory. Ann. N. Y. Acad. Sci 2003 985 125 134 12724154 Ferris CF Axelson JF Shinto LH Albers HE Scent marking and the maintenance of dominant/subordinate status in male golden hamsters. Physiol. Behav 1987 40 661 664 3671532 Gottlicher M Minucci S Zhu P Kramer OH Schimpf A Giavara S Sleeman JP Lo Coco F Nervi C Pelicci PG Heinzel T Valproic acid defines a novel class of HDAC inhibitors inducing differentiation of transformed cells. EMBO J 2001 20 6969 6978 11742974 Gray CL Norvelle A Larkin T Huhman KL Dopamine in the nucleus accumbens modulates the memory of social defeat in Syrian hamsters (Mesocricetus auratus). Behav. Brain Res 2015 286 22 28 25721736 Heim C Nemeroff CB The role of childhood trauma in the neurobiology of mood and anxiety disorders: preclinical and clinical studies. Biol. Psychiatry 2001 49 1023 1039 11430844 Heinrichs SC Leite-Morris KA Rasmusson AM Kaplan GB Repeated valproate treatment facilitates fear extinction under specific stimulus conditions. Neurosci. Lett 2013 552 108 113 23916657 Hollis F Kabbaj M Social defeat as an animal model for depression. ILAR J 2014 55 221 232 Institute of Laboratory Animal Resources 25225302 Hollis F Duclot F Gunjan A Kabbaj M Individual differences in the effect of social defeat on anhedonia and histone acetylation in the rat hippocampus. Horm. Behav 2011 59 331 337 20851702 Huhman KL Hebert MA Meyerhoff JL Bunnell BN Plasma cyclic AMP increases in hamsters following exposure to a graded footshock stressor. Psychoneuroendocrinology 1991 16 559 563 1667337 Huhman KL Solomon MB Janicki M Harmon AC Lin SM Israel JE Jasnow AM Conditioned defeat in male and female Syrian hamsters. Horm. Behav 2003 44 293 299 14609551 Huhman KL Social conflict models: can they inform us about human psychopathology? Horm. Behav 2006 50 640 646 16870189 Jasnow AM Huhman KL Activation of GABA(A) receptors in the amygdala blocks the acquisition and expression of conditioned defeat in Syrian hamsters. Brain Res 2001 920 142 150 11716820 Jasnow AM Cooper MA Huhman KL N-methyl-d-aspartate receptors in the amygdala are necessary for the acquisition and expression of conditioned defeat. Neuroscience 2004 123 625 634 14706775 Jasnow AM Shi C Israel JE Davis M Huhman KL Memory of social defeat is facilitated by cAMP response element-binding protein overexpression in the amygdala. Behav. Neurosci 2005 119 1125 1130 16187840 Kang J Chen J Shi Y Jia J Zhang Y Curcumin-induced histone hypoacetylation: the role of reactive oxygen species. Biochem. Pharmacol 2005 69 1205 1213 15794941 Kilgore M Miller CA Fass DM Hennig KM Haggarty SJ Sweatt JD Rumbaugh G Inhibitors of class 1 histone deacetylases reverse contextual memory deficits in a mouse model of Alzheimer's disease. Neuropsychopharmacology 2010 35 870 880 20010553 Kim WY Kim S Kim JH Chronic microinjection of valproic acid into the nucleus accumbens attenuates amphetamine-induced locomotor activity. Neurosci. Lett 2008 432 54 57 18164815 Lattal KM Barrett RM Wood MA Systemic or intrahippocampal delivery of histone deacetylase inhibitors facilitates fear extinction. Behav. Neurosci 2007 121 1125 1131 17907845 Maddox SA Watts CS Doyere V Schafe GE A naturally-occurring histone acetyltransferase inhibitor derived from Garcinia indica impairs newly acquired and reactivated fear memories. PLoS One 2013 8 e54463 23349897 Mahan AL Mou L Shah N Hu JH Worley PF Ressler KJ Epigenetic modulation of Homer1a transcription regulation in amygdala and hippocampus with pavlovian fear conditioning. J. Neurosci 2012 32 4651 4659 22457511 Markham CM Huhman KL Is the medial amygdala part of the neural circuit modulating conditioned defeat in Syrian hamsters? Learn. Memory 2008 15 6 12 Markham CM Taylor SL Huhman KL Role of amygdala and hippocampus in the neural circuit subserving conditioned defeat in Syrian hamsters. Learn. Memory 2010 17 109 116 Markham CM Luckett CA Huhman KL The medial prefrontal cortex is both necessary and sufficient for the acquisition of conditioned defeat. Neuropharmacology 2012 62 933 939 22001285 McCann KE Huhman KL The effect of escapable versus inescapable social defeat on conditioned defeat and social recognition in Syrian hamsters. Physiol. Behav 2012 105 493 497 21945371 McCann KE Bicknese CN Norvelle A Huhman KL Effects of inescapable versus escapable social stress in Syrian hamsters: the importance of stressor duration versus escapability. Physiol. Behav 2014 129 25 29 24582674 McGaugh JL The amygdala modulates the consolidation of memories of emotionally arousing experiences. Annu. Rev. Neurosci 2004 27 1 28 15217324 Monsey MS Gerhard DM Boyle LM Briones MA Seligsohn M Schafe GE A diet enriched with curcumin impairs newly acquired and reactivated fear memories. Neuropsychopharmac 2015 40 1278 1288 Nau H Loscher W Valproic acid: brain and plasma levels of the drug and its metabolites, anticonvulsant effects and gamma-aminobutyric acid (GABA) metabolism in the mouse. J. Pharmacol. Exp. Ther 1982 220 654 659 6801254 Nemeroff CB The neurobiology of depression. Sci. Am 1998 278 42 49 Parsons RG Ressler KJ Implications of memory modulation for post-traumatic stress and fear disorders. Nat. Neurosci 2013 16 146 153 23354388 Phiel CJ Zhang F Huang EY Guenther MG Lazar MA Klein PS Histone deacetylase is a direct target of valproic acid, a potent anticonvulsant, mood stabilizer, and teratogen. J. Biol. Chem 2001 276 36734 36741 11473107 Ploense KL Kerstetter KA Wade MA Woodward NC Maliniak D Reyes M Uchizono RS Bredy TW Kippin TE Exposure to histone deacetylase inhibitors during Pavlovian conditioning enhances subsequent cue-induced reinstatement of operant behavior. Behav. Pharmacol 2013 24 164 171 23604166 Reolon GK Maurmann N Werenicz A Garcia VA Schroder N Wood MA Roesler R Posttraining systemic administration of the histone deacetylase inhibitor sodium butyrate ameliorates aging-related memory decline in rats. Behav. Brain Res 2011 221 329 332 21421011 Sierra-Mercado D Padilla-Coreano N Quirk GJ Dissociable roles of prelimbic and infralimbic cortices ventral hippocampus, and basolateral amygdala in the expression and extinction of conditioned fear. Neuropsychopharmacology 2011 36 529 538 20962768 Simon-O'Brien E Alaux-Cantin S Warnault V Buttolo R Naassila M Vilpoux C The histone deacetylase inhibitor sodium butyrate decreases excessive ethanol intake in dependent animals. Addict. Biol 2015 20 676 689 25041570 Song Z McCann KE McNeill J.K. t. Larkin TE 2nd Huhman KL Albers HE Oxytocin induces social communication by activating arginine-vasopressin V1a receptors and not oxytocin receptors. Psychoneuroendocrinology 2014 50 14 19 25173438 Toth I Neumann ID Animal models of social avoidance and social fear. Cell Tissue Res 2013 354 107 118 23760888 Tremolizzo L Carboni G Ruzicka WB Mitchell CP Sugaya I Tueting P Sharma R Grayson DR Costa E Guidotti A An epigenetic mouse model for molecular and behavioral neuropathologies related to schizophrenia vulnerability. Proc. Natl. Acad. Sci. U. S. A 2002 99 17095 17100 12481028 Tunnicliff G Actions of sodium valproate on the central nervous system. J. Physiol. Pharmacol 1999 50 347 365 10574466 Xing B Zhao Y Zhang H Dang Y Chen T Huang J Luo Q Microinjection of valproic acid into the ventrolateral orbital cortex exerts an antidepressant-like effect in the rat forced swim test. Brain Res. Bull 2011 85 153 157 21419831

PMC005xxxxxx/PMC5135628.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 101282001 33781 Brain Struct Funct Brain Struct Funct Brain structure & function 1863-2653 1863-2661 27255751 5135628 10.1007/s00429-016-1242-9 NIHMS792735 Article Sub-synaptic localization of Cav3.1 T-Type calcium channels in the thalamus of normal and parkinsonian monkeys Chen Erdong 12 Paré Jean-Francois 12 Wichmann Thomas 123 Smith Yoland 123 1 Yerkes National Primate Research Center, Emory University, Atlanta, Georgia 30322 2 Udall Center of Excellence for Parkinson’s Disease Research, Emory University, Atlanta, Georgia 30322 3 Dept. of Neurology, Emory University, Atlanta, Georgia 30322 Corresponding author: Yoland Smith, PhD, Yerkes National Primate Research Center, Emory University, 954 Gatewood Road NE, Atlanta GA 30329, Phone: (404) 727 7519, Fax: (404) 727 1266, ysmit01@emory.edu 9 6 2016 02 6 2016 3 2017 01 3 2018 222 2 735748 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. T-type calcium channels (Cav3) are key mediators of thalamic bursting activity, but also regulate single cells excitability, dendritic integration, synaptic strength and transmitter release. These functions are strongly influenced by the subcellular and subsynaptic localization of Cav3 channels along the somatodendritic domain of thalamic cells. In Parkinson’s disease, T-type calcium channels dysfunction in the basal ganglia-receiving thalamic nuclei likely contributes to pathological thalamic bursting activity. In this study, we analyzed the cellular, subcellular, and subsynaptic localization of the Cav3.1 channel in the ventral anterior (VA) and centromedian/parafascicular (CM/Pf) thalamic nuclei, the main thalamic targets of basal ganglia output, in normal and parkinsonian monkeys. All thalamic nuclei displayed strong Cav3.1 neuropil immunoreactivity, although the intensity of immunolabeling in CM/Pf was significantly lower than in VA. Ultrastructurally, 70–80% of the Cav3.1-immunoreactive structures were dendritic shafts. Using immunogold labeling, Cav3.1 was commonly found perisynaptic to asymmetric and symmetric axo-dendritic synapses, suggesting a role of Cav3.1 in regulating excitatory and inhibitory neurotransmission. Significant labeling was also found at non-synaptic sites along the plasma membrane of thalamic neurons. There was no difference in the overall pattern and intensity of immunostaining between normal and parkinsonian monkeys, suggesting that the increased rebound bursting in the parkinsonian state is not driven by changes in Cav3.1 expression. Thus, T-type calcium channels are located to subserve neuronal bursting, but also regulate glutamatergic and non-glutamatergic transmission along the whole somatodendritic domain of basal ganglia-receiving neurons of the primate thalamus. centromedian parafascicular immunogold primate Parkinson’s disease MPTP Introduction In the thalamus, low-threshold calcium spikes (LTS) and burst firing are mediated by low voltage activated (LVA) T-type calcium channels (Steriade and Llinás, 1988; Huguenard, 1996) which play key roles in sleep, arousal, and sensory gating (McCormick and Bal, 1997). Changes in T-type calcium channel function have been implicated in the pathophysiology of various neurological disorders including Parkinson’s disease (PD), absence epilepsy, neuropathic pain, and various neuropsychiatric disorders (Buzsaki et al., 1990; Jeanmonod et al., 2001; Llinás et al., 2001; Nelson et al., 2006; Belardetti and Zamponi, 2012; Cain and Snutch, 2013; Bladen et al., 2015). In addition to their long-expected role in the regulation of thalamic burst firing, several new functions have recently been assigned to these channels, such as the regulation of single cells excitability, dendritic integration, transmitter release and synaptic strengthening (Lambert et al., 2014). The onset and progression of PD motor symptoms have been closely linked to the loss of striatal dopamine and the resulting aberrant activity within basal ganglia-thalamocortical circuits (Albin et al., 1989; Delong, 1990; Bergman et al., 1990), including increased thalamic bursting (Zirh et al. 1998; Magnin et al., 2000; Pessiglione et al., 2005; Devergnas et al., 2015). A characteristic type of bursts in the mammalian thalamus is termed “rebound bursts” (Llinas and Jahnsen, 1982), which are characterized by low threshold calcium spike bursts (LTS) that result from hyperpolarization-induced de-inactivation of T-type calcium channels (Perez-Reyes, 2003). Alterations in burst kinetics of T-type calcium channels in the motor thalamus induce changes in rebound LTS bursting that may underlie motor dysfunction in PD (Pare et al., 1990; Yang et al., 2014). Increased rebound bursting has, indeed, been documented in the ventral motor thalamus of parkinsonian monkeys (Zirh et al. 1998; Magnin et al., 2000; Pessiglione et al., 2005; Devergnas et al., 2015). Based on differences in their pore-forming α1-subunit, T-type calcium channels are classified as Cav3.1, Cav3.2, and Cav3.3 (Perez-Reyes, 2003, 2006; Perez-Reyes and Lory, 2006). Although these subtypes of T-type calcium channels have similar functional characteristics (Perez-Reyes 2003; Catterall et al. 2005), they differ in their distribution across brain regions and in their activation or inactivation kinetics. In rodents, Cav3.1 is the most strongly expressed T-type calcium channel subtype in thalamocortical projection neurons, while Cav3.3 is abundantly expressed in reticular thalamic neurons (Talley et al. 1999). Although the cellular expression of these channels has been studied throughout the brain, very little is known about their spatial localization at the single cell level, and about potential changes in the localization of the channels in the parkinsonian state. Thus, using high resolution electron microscopy procedures, we studied in detail the subcellular and subsynaptic localization of the Cav3.1 channel in the basal ganglia-receiving thalamic nuclei, the parvocellular and magnocellular ventral anterior nucleus (VApc/VAmc) and centromedian/parafascicular (CM/Pf) nuclei, in normal and parkinsonian monkeys. Material and Methods Animals Six female adult (4–8 years old) Rhesus monkeys (Macaca mulatta, 5–10 kg) from the Yerkes primate center breeding colony were used in this study. Three of these animals were used as control and three were rendered parkinsonian by weekly injections of 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP). All animals were pair-housed, and had ad libitum access to food and water. All experiments were performed in accordance with the United States Public Health Service Policy on the humane care and use of laboratory animals, including the provisions of the “Guide for the Care and Use of Laboratory Animals” (Garber et al., 2011). All studies were approved by the Biosafety and Animal Care and Use Committee of Emory University. MPTP treatment Animals were treated with MPTP (0.2–0.6mg/kg i.m.; Sigma, St. Louis, MO; cumulative doses: 2.8 – 8.8 mg/kg) once per week, until they reached a state of stable moderate parkinsonism. The degree and stability of the MPTP-induced motor disability was assessed, as described in many previous studies from our laboratory (Galvan et al., 2011; 2014; Masilamoni et al., 2011; Bogenpohl et al., 2013; Villalba et al., 2014; Devergnas et al., 2014, Mathai et al., 2015). Briefly, animals were evaluated weekly in an observation cage that was equipped with infrared beams, allowing us to measure their body movements as infrared beam break events. A parkinsonism rating scale was also used to quantify impairments in ten aspects of motor function (bradykinesia, freezing, extremity posture, trunk posture, action tremor, the frequency of arm and leg movements, finger dexterity, home cage activity, and balance), each scored on a 0 to 3 scale (maximal score 30). For the three MPTP-treated monkeys used in this study, the final parkinsonism rating score ranged between 13 to 20, corresponding to moderately severe parkinsonism. The severity of the parkinsonian motor signs had to be stable for a period of at least 6 weeks after the last MPTP injection before the decision was made to sacrifice the animal. Perfusion of Animals and sectioning of tissue At the end of the experiment, the animals were sacrificed with an overdose of sodium pentobarbital (100 mg/kg, i.v.) and transcardially perfused with cold oxygenated Ringer's solution, followed by a fixative containing 4% paraformaldehyde and 0.1% glutaraldehyde in a phosphate buffer (PB) solution. After perfusion, the brains were removed from the skull, cut coronally into 10 mm thick blocks, and post-fixed overnight in 4% paraformaldehyde. The blocks were then cut into 60µm-thick coronal sections using a vibrating microtome and stored at −20 °C in an anti-freeze solution, containing 30% ethylene glycol and 30% glycerol in PB, until the time of the immunohistochemistry studies. Immunohistochemistry Antibodies used For the localization of Cav3.1, we used a highly specific monoclonal antibody (NINDS/NIMH NeuroMab, Davis, CA). According to the supplier, this antibody reacts with the >250kDa molecular weight protein associated with Cav3.1, does not cross-react with other subtypes of Cav3 channels, and does not result in any significant immunostaining or Western blot band labeling in tissue from Cav3.1 knockout mice (NeuroMab; Accession # Q9WUT2). The quality and specificity of this antibody for its use in the present study was demonstrated via Western blot analysis performed on fresh monkey brain tissue (Devergnas et al., 2016). Overall, the pattern of immunostaining throughout the monkey brain with this antibody was consistent with findings from previous anatomical and electrophysiological studies suggesting the functional expression or lack of T-type calcium channels (Cavelier and Bossu, 2003; McKay et al., 2006; Hildebrand et al., 2009; Isope et al., 2012). Light microscopic pre-embedding immunoperoxidase procedures Sections of thalamic tissue containing the VA and CM/Pf thalamic nuclei were chosen based on maps from the rhesus monkey stereotaxic brain atlas (Paxinos et al., 1999). Sections containing the basal ganglia-recipient regions of VA corresponded approximately to sections at the interaural 11.55 plane, while CM/Pf sections were at the interaural plane 7.95 in the stereotaxic atlas. Sections of tissue to be processed were removed from the anti-freeze solution and then placed in phosphate-buffered saline (PBS, 0.01M, pH 7.4). They were immersed in sodium borohydride (1% in PBS) for 20 minutes and incubated for 1 hour in PBS containing 1% normal horse serum (NHS), 1% bovine serum albumin (BSA), and 0.3% Triton X-100, followed by incubation in the primary antibody (anti-Cav3.1-NeuroMab; Accession # Q9WUT2) solution containing 1% NHS, 1% BSA, and 0.3% Triton X-100 in PBS for 24 hours at 4°C. Sections were then rinsed three times in PBS and subsequently incubated in the secondary antibody solution containing 1% NHS, 1% BSA, 0.3% Triton X-100, and biotinylated horse anti-mouse IgGs (Vector Laboratories, Burlingame, CA; used at 1:200 dilution) for 90 minutes at room temperature. After three rinses in PBS, the sections were incubated for 90 minutes in avidin-biotin peroxidase complex (ABC) solution (Vectastain standard ABC kit, Vector Laboratories; used at 1:100 dilution) including 1% BSA. To reveal the antigenic sites, the sections were first rinsed with PBS and Tris buffer (50 mM; pH 7.6), and then incubated in a solution containing 0.025% 3, 3’-diaminobenzidine tetrahydrochloride (DAB; Sigma), 10 mM imidazole, and 0.005% hydrogen peroxide in Tris buffer for 10 min. The sections were subsequently washed several times in PBS, mounted on gelatin-coated glass slides, dehydrated, and coverslipped with Cytoseal XYL™ (Richard-Allan Scientific). The slides were scanned at 20X using a ScanScope CS scanning light microscope system (Aperio Technologies, Vista, CA). Digital representations of the slides were saved and analyzed using the ImageScope software (Aperio). To help with the delineation of the nuclear borders between the basal ganglia- and cerebellar-receiving territories of the ventral motor nuclei, some adjacent sections were immunostained for the vesicular glutamate transporter 2 (vGluT2) (Kuramoto et al., 2011), a reliable marker of cerebellar glutamatergic terminals, using highly specific antibodies (Mab Technologies, Atlanta, GA, USA) and immunohistochemical procedures described in detail in our previous studies (Villalba et al., 2006; Raju et al., 2008). Additional adjacent sections from representative control and MPTP-treated monkeys were immunostained for tyrosine hydroxylase (TH) using specific monoclonal antibodies (Catalog nos. MAB 318; Millipore, Billerica, MA) and immunohistochemical procedures detailed in our previous studies (Mazloom and Smith, 2006; Masilamoni et al., 2011; Mathai et al., 2015) to demonstrate the substantial reduction in dopamine conferred by MPTP treatment of the monkeys used in this study. Electron microscopic pre-embedding immunoperoxidase protocol Sections were immersed in sodium borohydride (1% in PBS) for 20 minutes, rinsed in PBS, and placed in a cryoprotectant solution (0.05M phosphate buffer [PB], pH 7.4, 25% sucrose and 10% glycerol) for 20 minutes prior to being frozen at −80°C for 20 minutes and thawed to permeabilize cell membranes. Then, sections were put through a graded series of cryoprotectant solution (100%, 70%, 50%, and 30% in PBS), and finally washed in PBS. The subsequent tissue processing was identical to that used for light microscopy, up to the point of the use of DAB, with two important differences: Triton X-100 was omitted from all solutions, and sections were incubated in the primary antibody solution for 48 hours at 4°C. After DAB exposure, the tissue was rinsed in PB (0.1 M, pH 7.4) and treated with 1% osmium tetroxide for 20 minutes. The sections were then dehydrated through an increasing gradient of ethanol (50%, 70%, 90% and 100%) with the initial incubation done in 70% solution containing 1% uranyl acetate for 35 minutes to increase contrast under the EM. The sections were then placed in propylene oxide and subsequently embedded in epoxy resin (Durcupan, ACM; Fluka, Buchs, Switzerland) for 24 hours. Then, the sections were mounted on microscope slides, dabbed with epoxy resin, coverslipped with mineral oil-coated coverslips, and put in the oven at 60°C for 48 hours to cure the resin. After polymerization, the coverslips were taken off, and small blocks of tissue from the basal ganglia-receiving area of the VA and CM/Pf nuclei were cut from the slides and glued onto resin blocks with cyanoacrylate glue. The blocks were cut into 60nm sections using a diamond knife ultramicrotome (Ultracut T2; Leica, Nussloch, Germany) and collected on single-slot Pioloform-coated copper grids. The sections were then stained with lead citrate for 5 minutes, rinsed in distilled water, and viewed under a transmission EM (JEM-1011; JEOL USA Inc., Peabody, MA). Electron microscopic pre-embedding immunogold protocol The tissue was prepared as described above, except that sections were pre-incubated in PBS containing 5% milk, rinsed in Tris Buffer Saline-Gelatin (TBS-Gelatin). On the next day, sections were first incubated for 90 minutes with secondary goat anti-mouse Fab’ fragments conjugated to 1.4-nm gold particles (1:100; Nanoprobes, Yaphank, NY) and 1% dry milk in TBS-gelatin. Sections underwent incubation for approximately 10 minutes in the dark with aHQ Silver Kit (Nanoprobes) to increase gold particle sizes to 30–50-nm through silver intensification, as described in our previous studies (Kuwajima et al., 2007; Mitrano et al., 2010; Gonzales et al., 2013). Densitometric Analysis of Light Microscopy Material Using Imagescope image viewing software (Aperio) and the NIH Image J software (Schneider et al., 2012), we quantified the intensity of Cav3.1 immunostaining from digital 0.4x magnification images of immunostained tissue slides containing the motor thalamus in a manner similar to that reported in our previous studies (Galvan et al., 2011). The images were converted into 16-bit grayscale format and inverted. The intensity of Cav3.1 immunoreactivity was determined by measuring the optical density of Cav3.1 immunoperoxidase labeling in the basal ganglia-receiving territory of the VA and CM/Pf complex of control and MPTP-treated monkeys. Adjacent vGluT2-immunostained sections were used to help separate the cerebellar- from the basal ganglia-receiving regions of the ventral motor nuclei (Kuramoto et al. 2011; Devergnas et al., 2016). In each selected nuclear region of the three control and three MPTP-treated monkeys, measurements of optical density were taken from three immediately adjacent 2.08 mm2 areas of Cav3.1-immunostained tissue and averaged, yielding three values for each nucleus per treatment condition. The average optical density measurements in VA, CM and Pf for a single animal were calculated by taking the mean of these measurements from the three sampled areas/nucleus. Additional optical density measurements were taken in the internal capsule immediately adjacent to thalamic borders in each section in the same manner and averaged to reflect background labeling. The average values for background labeling were subtracted from measurements made in each thalamic nucleus. Student’st test was used to compare the background-corrected values between control and MPTP-treated monkeys, and between VA, CM and Pf. Analysis of Electron Microscopy Material At the electron microscopic level, immunoperoxidase labeling could be identified as a dark, amorphous deposit within neuronal elements, while immunogold labeling appeared as small dark, round particles within neuronal elements. To assess the localization of Cav3.1 labeling in normal and parkinsonian animals, 50 digital micrographs of randomly encountered Cav3.1-labeled neuronal elements were captured in each animal at 40,000X magnification and saved with a CCD camera (Orius 78; Gatan, Inc., Pleasanton, CA) that was controlled by DigitalMicrograph software (Gatan Microscopy Suite), yielding 2214 µm2 of tissue analyzed per nucleus per animal. Immunoperoxidase Material Elements labeled with the peroxidase deposit were classified, based on ultrastructural features (Peters et al., 1991), and the relative distribution of Cav3.1 immunoreactivity among neuronal elements was compared between normal and MPTP-treated animals. Labeled dendrites were categorized into small (≤0.5µm), medium (0.5µm–1µm), or large (≥1µm) profiles, based on their cross-sectional diameter. The same immunoperoxidase-stained sections were used to assess the relative prevalence of Cav3.1-labeled dendrites over the total population of dendritic profiles in the VA, CM, and Pf. Values obtained for all measures were compared between control and MPTP-treated animals using Student’s or Welche’s t-test. Immunogold Material The immunogold-stained sections were used to elucidate the specific localization of Cav3.1 labeling in relation to synaptic and non-synaptic sites along the plasma membrane of VA and CM/Pf neurons. Asymmetric (putative excitatory) and symmetric (putative inhibitory) synapses were differentiated based on the presence or absence of thick post-synaptic densities, respectively (Gray, 1959; Uchizono, 1965; Guillery, 2000). The gold particles that were directly on or within 20 nm of the plasma membrane were identified as “plasma membrane-associated”, and categorized as either perisynaptic or extrasynaptic, based on their localization relative to postsynaptic specializations. In this manuscript, the term “perisynaptic” refers to gold particles found within 20 nm of the edges of postsynaptic specializations. All other plasma membrane-bound gold particles were categorized as “extrasynaptic”. The 20-nm cut-off point was chosen based on the assertion that the distance between the immunoreactive substrate and gold particle, bridged by the primary and secondary antibodies, can be approximate to 20 nm (Blackstad et al., 1990). Statistical analyses were performed using the same procedures described for the immunoperoxidase data. Results Light microscopic (LM) immunohistochemical staining for Cav3.1 in the monkey thalamus At the LM level, a high degree of Cav3.1 immunoreactivity was found throughout the neuropil of the monkey thalamus, except in the reticular nucleus, which was completely devoid of labeling (Fig. 1). Although the VA, CM and Pf nuclei displayed Cav3.1 immunoreactivity, the intensity of immunostaining was significantly lower in the CM and Pf than in the VA of normal and MPTP-treated monkeys (p < 0.001 control; p < 0.05 post-MPTP; t-test) (Fig. 1E). Overall, there was no significant difference in the intensity of immunostaining between the normal and parkinsonian animals for each thalamic nucleus. CM and Pf displayed comparable level of immunoreactivity in either state (p = 0.39 control; p = 0.40 post-MPTP; t-test) (Fig. 1E). At high magnification, immunoreactivity was apparent throughout the neuropil, with clear localization in perikarya and dendritic processes (Fig. 1A’,C’,C”) EM immunoperoxidase localization of Cav3.1 in the VA, CM and Pf At the EM level, ~80–90% of Cav3.1 channel-immunoreactive elements comprise postsynaptic dendrites of various sizes in all three thalamic nuclei examined in both, the normal and parkinsonian states. Immunoreactivity was rarely encountered in unmyelinated axons (~3–9%), glial processes, or axon terminals (<4%) (Fig. 2). MPTP treatment did not significantly alter the relative prevalence of Cav3.1-immunoreactive elements in the VA, CM, and Pf (Fig. 2C,F,I). The only statistically significant difference was a reduction in the proportion of labeled glial processes (from 3.4% to 1.92%) in the CM of MPTP-treated animals compared with controls (p = 0.014; t-test) (Fig. 2 H). The proportion of large, medium and small immunoreactive dendrites in VA, CM and Pf was also not changed in MPTP-treated animals (Fig. 3D). None of the labeled dendritic profiles had the ultrastructural features of putative GABAergic interneurons (ie vesicle-filled and pre-synaptic to axo-dendritic synapses; Ralston 1971; Hamos et al., 1985; Ohara et al., 1989; Jones, 2007). Although the peroxidase deposit in most labeled elements was diffusely distributed without any clear interactions with specific subcellular organelles (Fig. 2A,B,D,E,G,H), there were some dendrites in which aggregates of labeling were confined to the post-synaptic densities of putative asymmetric synapses (Fig. 5A). EM immunogold localization of Cav3.1 in the VA and CM/Pf Because of the diffuse and amorphous nature of the peroxidase deposit, the immunoperoxidase method is not a reliable approach to assess the subcellular and subsynaptic localization of Cav3.1. We used the pre-embedding immunogold method to further address these issues in normal and parkinsonian monkeys. The general pattern of immunoreactivity amongst neuronal elements was similar to that found in the immunoperoxidase-stained sections of VA and CM/Pf, ie gold labeling was localized predominantly in dendrites of various sizes, with additional sparse labeling in axons, terminals and glia. In immunoreactive dendrites, 75–87% of the Cav3.1 immunogold labeling was found in direct association with plasma membranes (Fig. 4 A–E). This pattern was the same in all nuclei examined, and was not significantly different in MPTP-treated animals (Fig. 4E). Of the plasma membrane-bound gold particles in dendrites of VA, CM and Pf neurons, 12–17% were located within 20nm of the edge (ie perisynaptic) of symmetric and asymmetric synapses (Fig. 5 A–D), while the remainder was expressed extrasynaptically (Fig 6 A,B). Perisynaptic labeling was found in dendrites of all sizes (Fig. 5B–D). When normalized to the portion of membrane occupied by the individual subsynaptic domains (perisynaptic vs extrasynaptic) in each of the nuclei, the relative density of perisynaptic Cav3.1 immunogold labeling was significantly higher than would be expected from a random distribution across all three nuclei (Fig. 6 C). Discussion The results of this study demonstrate that the Cav3.1 subtype of T-type calcium channel immunoreactivity is strongly expressed in neurons of the VA, CM, and Pf thalamic nuclei, albeit to a higher level in VA than in CM/Pf, in primates. At the subcellular and subsynaptic level, Cav3.1 channels are expressed along the whole somatodendritic domain of thalamic neurons, being preferentially aggregated extrasynaptically or perisynaptically to putative glutamatergic and non-glutamatergic synapses (Fig. 7). Overall, the pattern of cellular and subcellular localization of the channel was not affected in MPTP-treated parkinsonian animals. Together, these findings indicate that Cav3.1 channels are located to subserve widespread influences over burst firing activity, but also modulation of specific glutamatergic and non-glutamatergic synapses, in basal ganglia-receiving nuclei of the primate thalamus in normal and parkinsonian state. Uniform dendritic expression of Cav3.1 immunoreactivity in the monkey thalamus The strong level of Cav3.1 immunoreactivity throughout the whole monkey thalamus is consistent with results of previous in situ hybridization and light microscopic immunohistochemical studies in rodents and cats (Talley et al., 1999; McKay et al., 2006; Kovacs et al., 2010; Parajuli et al., 2010; Liu et al., 2011). Similarly, the predominant dendritic localization of the Cav3.1 channels found in this study is consistent with previous data from the cat reticular nucleus (Kovacs et al., 2010) or different brain regions including the hippocampus, cerebellum (Christie et al., 1995; Kavalali et al., 1997; Gauck et al., 2001) and subthalamic nucleus (Song et al., 2000). However, their relative distribution along the somatodendritic axis of thalamic neurons has not yet been fully elucidated and remains a matter of debate. Although previous electrophysiological, pharmacological and neuronal modeling studies have suggested different patterns of T-type channels expression in various populations of thalamic neurons (Munsch et al., 1997; Zhou et al, 1997; Destexhe et al., 1998; Williams and Stuart, 2000; Rhodes and Llinas, 2005), these findings must be interpreted with caution because of the lack of selective T-type calcium channel blockers, the limited ability to detect calcium signaling in small-sized distal dendrites and the equivocal criteria used to define T-type currents. Our ultrastructural findings indicate that T-type calcium channels are expressed along both the proximal and distal parts of the dendritic tree of thalamic cells in VA, CM and Pf neurons in normal and parkinsonian monkeys. A similarly uniform dendritic expression of Cav3.1 immunoreactivity has also been reported in thalamocortical neurons of the lateral geniculate nucleus in mice (Parajuli et al., 2010). However, another study that used immunofluorescence as labeling method suggested that Cav3.1 immunoreactivity is confined to the cell bodies and proximal dendrites of thalamocortical neurons in the rat lateroposterior and ventral posterolateral nuclei (McKay et al., 2006). The sensitivity of different Cav3.1 antibodies combined with species and nuclear specificity in the distribution of channels along the dendritic tree of thalamic cells may explain these differences. The recent development of highly specific T-type calcium channel blockers (Shipe et al., 2008; Yang et al., 2008; Dreyfus et al., 2010; Ardashov et al., 2011; Xiang et al., 2011) will help elucidate the specific roles of T-type channels on the different parts of the dendritic tree of these thalamic neurons. In that regard, electrophysiological studies have demonstrated that an important role of spatially and temporally regulated T-type calcium channel-mediated dendritic Ca(2+) signaling properties throughout the whole dendritic tree of rat thalamocortical neurons is to mediate regenerative propagation of low threshold spikes in a behavioral state-dependent manner (Errington et al., 2010). It has also been shown that T-type calcium channel-mediated dendritic properties are highly conserved between population of thalamocortical and reticular low-threshold spiking neurons, and that these properties underlie a whole-cell somato-dendritic spike generation mechanism that makes low threshold spikes a unique global electrical and biochemical signal in neurons (Connelly et al., 2015). Differential expression level of Cav3.1 immunoreactivity in CM/Pf and VA The intensity of labeling was significantly lower in CM/Pf than in VA in both normal and parkinsonian monkeys. This difference in staining intensity may be due to striking morphological differences between projection neurons in the CM/Pf and other thalamic nuclei (Lacey et al., 2007; Jones 2007). In contrast to most thalamic neurons, including those in VA, which display a profuse “bushy” dendritic arbor, CM/Pf neurons harbor a much less arborized “reticular-like” dendritic tree (Lacey et al., 2007; Jones, 2007). In rats, the total dendritic length of single “bushy” neurons of the centrolateral nucleus can be as much as two times larger than neurons of the Pf (Lacey et al., 2007). The differential level of T-type channel expression between the Pf and other thalamic nuclei may have functional consequences. It is, indeed, noteworthy that rodent Pf neurons rarely discharge low-threshold calcium spikes during cortical slow-wave activity in vivo, compared with neurons in the centrolateral nucleus (Lacey et al., 2007). These observations suggest that the lower expression of T-type calcium channels may confer CM/Pf neurons specific physiological properties that allow them to respond differently to different states of vigilance or sleep compared with other thalamic nuclei. Other factors, such as the level of membrane depolarization of CM/Pf neurons compared with VA may also account for their differential propensity in developing LTS. If such is the case, these observations add further evidence to the concept put forward in our previous studies that CM/Pf neurons are a unique subpopulation of thalamic neurons endowed with anatomical and functional properties distinct from most other thalamic neurons (Galvan et al., 2011; Smith et al., 2004, 2014). Cav3.1 at glutamatergic and non-glutamatergic synapses Our immunogold data revealed two major structural features related to the subsynaptic localization of Cav3.1 in the monkey thalamus. They showed that most gold particles were bound to the plasma membrane of small-, medium- and large-sized dendrites in VA and CM/Pf. This predominant membrane localization of Cav3.1 can be interpreted as channel insertion in the plasma membrane, while the low level of intracellular immunoreactivity likely represents newly synthesized or internalized proteins being trafficked to intracellular organelles or different parts of the dendritic tree. Thus, these observations provide further support that “functional” membrane-bound T-type calcium channels are expressed along the entire somatodendritic domain of thalamic neurons in VA and CM/Pf. These observations are consistent with those reported in the rodent lateral geniculate nucleus (Parajuli et al., 2010). However, because these immunohistochemical data are not reliable indicators of the relative amount and physiological responsiveness of the channels at different dendritic sites, future in vitro slice studies are needed to further characterize the properties of T-type channels along the somatodendritic domain of these neurons. Our immunogold data also showed that a subset of Cav3.1 channels selectively aggregates perisynaptically at putative glutamatergic and non-glutamatergic synapses in the VA, CM and Pf of rhesus monkeys. To our knowledge, this is the first direct evidence for perisynaptic expression of T-type calcium channels in the mammalian thalamus. Although the source(s) of terminals associated with these channels was not determined, their ultrastructural features suggest that those forming asymmetric synapses likely originate from the cerebral cortex, while those forming symmetric synapses may come from GABAergic reticular neurons or interneurons (Liu et al., 1995; Ilinsky et al., 1999; Jones, 2007). However, we cannot rule out that some of these inputs may also arise from monoaminergic or cholinergic ascending brainstem afferents. Because pre-embedding immunostaining may lead to false negative results due to limited access of antibodies to their antigenic sites, our data do not allow to quantify the exact percentage of total symmetric and asymmetric synapses that express perisynaptic Cav3.1 labeling. However, based on the high prevalence of labeled synapses encountered during our electron microscopy analyses, we suggest that T-type calcium channels are frequently associated with synaptic junctions in the basal ganglia-receiving nuclei of the monkey thalamus. While the physiological role of T-type calcium channels at these synapses was not studied, it is worth discussing hypotheses regarding the function of these channels, based on rodent data from other brain regions. One possibility is that they allow calcium influx to boost depolarization induced by specific cortico-thalamic glutamatergic synapses, a mechanism for signal amplification that has been documented for other calcium channels at axo-spinous glutamatergic synapses on cortical and hippocampal pyramidal neurons (Markram and Sakmann, 1994; Magee et al., 1995; Isope et al., 2012; Lambert et al., 2013). Another possibility is that the Cav3.1 channels functionally interact and couple with group I metabotropic glutamate receptors to regulate fast calcium signaling pathway at specific cortico-thalamic synapses. A similar phenomenon has recently been described for parallel fiber synapses on Purkinje cells in the rodent cerebellar cortex (Isope et al., 2012, Hildebrand et al., 2009). This hypothesis is further supported by the fact that group I mGluRs display a perisynaptic pattern of localization similar to that described here for Cav3.1 channels in many thalamic nuclei and other brain regions (Baude et al., 1993; Liu et al., 1998; Hanson and Smith, 1999; Paquet and Smith, 2003). Because on its perisynaptic localization (Villalba et al., 2006), GABAB receptors may be another target of relevance to consider as possible functional partners with T-type channels at glutamatergic and non-glutamatergic synapses in the monkey thalamus, although functional evidence for such interaction in the thalamus and other regions remain weak and poorly understood (Guyon and Leresche, 1995; Huang et al., 2015). Thalamic expression of Cav3.1 immunoreactivity in normal and parkinsonian states In line with our recent findings (Devergnas et al., 2015), results of the present study did not reveal any major changes in the overall pattern of Cav3.1 expression in the VA and CM/Pf of MPTP-treated parkinsonian monkeys. Thus, despite clear evidence for increased thalamic rebound bursting in the parkinsonian state (Zirh et al., 1998; Magnin et al., 2000; Pessiglione et al., 2005; Devergnas et al., 2016), these functional changes are unlikely to result from abnormal cellular, subcellular and subsynaptic expression of Cav3.1 channels. However, we cannot rule out that functional properties of T-type channels might be affected in parkinsonism. On the other hand, the fact that intracerebral injections of a specific T-type calcium channel blocker in the basal ganglia-receiving region of the thalamus did not significantly affect the rate and pattern of neuronal activity of BGMT neurons in our previous study (Devergnas et al., 2016) suggests that altered T-type channel activity may not be the sole source of alterations in thalamic firing in the parkinsonian state. A thorough assessment of the functional properties of T-type calcium channels in thalamic slices from normal and dopamine-depleted animals is warranted to further address this issue. This project was supported through grants from the NIH/NINDS (R01 NS054976 [TW/YS] and P50 NS071669 [Udall Center grant, TW/YS]), a grant from the NIH BP ENDURE (SP00010548 [EC/YS]) and a grant from the NIH/ORIP to the Yerkes Center (P51 OD011132). We thank Susan Jenkins for expert technical assistance. Figure 1 (A–D) Cav3.1-immunostained coronal brain sections at different rostrocaudal levels of the thalamus in control (A,C) and MPTP-treated (B,D) monkeys. (A’,C‘,C”) show high-power micrographs of Cav3.1 immunoreactivity in the VA, CM, and Pf of a control animal. The delineation of thalamic nuclear boundaries was based on The Rhesus Monkey Brain in Stereotactic Coordinates (Paxinos et al., 2000) and a coronal atlas of the macaque brain (Lanciego and Vazquez, 2012). (E) Measurements of optical intensity taken in the VA, CM and Pf (values are mean ± SEM). Significance was assessed with the Mann-Whitney Rank Sum Test and MPTP-treatment related changes were found to be insignificant at the α = 0.05 significance level. Both the CM and Pf expressed a significantly lower optical density than VA in normal and MPTP-treated monkeys (asterisks). Scale bars: A=2mm (valid for B, C, D). A’=50µm (valid for C’, C”). Abbreviations: CM – centromedian thalamic nucleus, CN – caudate nucleus, GPe – external segment of the globus pallidus, GPi – internal segment of the globus pallidus, IC – internal capsule, Pf – parafascicular thalamic nucleus, PUT – putamen, Rt – reticular nucleus, VAmc- ventroanterior nucleus, magnocellular division; VApc- ventroanterior nucleus, parvocellular division; VL- ventrolateral nucleus; VLo – ventrolateral thalamic nucleus, pars oralis; VPl- ventroposterior nucleus, pars lateralis. Figure 2 Immunoperoxidase localization of Cav3.1 in the thalamus of control and MPTP-treated monkeys. (VA-A,B; CM-D,E; Pf-G,H) Representative electron micrographs of Cav3.1 immunoreactivity in thalamic nuclei of control (A, D, G) and MPTP-treated (B, E, H) monkeys. Immunoreactive dendritic and unmyelinated axonal processes are labeled. (C, F, I) Distribution of Cav3.1-immunoreactive elements in the VA, CM and Pf of control and MPTP-treated animals. Values are mean ± SEM and comparison between control and MPTP-treated monkeys tested with either student’s t-test or Welche’s t-test, depending on the variance of the groups as determined by the F-test. In E, note a statistically significant difference (*) between treatment groups in the density of immunoreactive glia in the CM (p = 0.014; t-test). Values are mean ± SEM. Abbreviations: U.Ax – unmyelinated axon, Den – dendrite. Scale bar in A= 0.5µm (valid for all micrographs). Figure 3 EM immunoperoxidase localization of Cav3.1 labeling in dendrites of thalamic neurons in control and MPTP-treated monkeys. (A–C) Representative examples of Cav3.1 immunoreactivity in small (≤0.5µM), medium (0.5µM–1µM), and large (≥1µM) dendrites of thalamic nuclei of control and MPTP-treated monkeys. (D) Percent of small, medium, and large dendrites out of all Cav3.1-immunoreactive dendrites. The MPTP treatment did not alter the general distribution of Cav3.1 immunoreactivity among dendritic profiles of different sizes in the VA, CM, and Pf (p > 0.05; t-test). Values are mean ± SEM. Abbreviations: Den – dendrite. Scale bar in A= 0.5µm (valid for B,C). Figure 4 Pre-embedding immunogold labeling for Cav3.1 in dendrites of VA, CM, and Pf nuclei in control and MPTP-treated monkeys. (A–D) Immunogold labeling along the plasma membrane of dendrites of various sizes in all thalamic nuclei. (E) Relative percentage of plasma membrane-bound gold labeling in dendrites of VA, CM and Pf neurons in control and MPTP-treated monkeys. In all nuclei, MPTP treatment did not significantly affect the plasma membrane localization of Cav3.1 immunoreactivity. Values are mean ± SEM. Abbreviations: Den – dendrite, Te – terminal. Scale bars: A= 0.5µm (valid for B); C= 0.1µm (valid for D). Figure 5 EM immunoperoxidase and immunogold localization of Cav3.1 at asymmetric and symmetric synapses in dendrites of thalamic neurons in control and MPTP-treated monkeys. (AD) High powered immunoperoxidase (A) and immunogold (B–D) Cav3.1 dendritic labeling. Note in (A) the dense immunoperoxidase deposit at two post-synaptic densities of putative asymmetric axo-dendritic glutamatergic synapses (*) compared with the lack of labeling at another asymmetric synapse on the same dendrite (X). (B-D) show high powered micrographs of perisynaptic Cav3.1 immunogold labeling at asymmetric (arrows) or symmetric (+) axo-dendritic synapses. Scale bar: A=0.5µm (valid for B-D). Figure 6 Quantification of Cav3.1 channel immunogold labeling associated with different sectors of the plasma membrane of thalamic neurons in VA, CM and Pf. (A-B) Comparison between the percentages of Cav3.1 gold particle labeling at extrasynaptic and perisynaptic sites (black bars) and the proportion of total plasma membrane occupied by these two subsynaptic domains based on random observations (white bars). Data are expressed as means ± SE of measurements in 6 animals. Asterisks in B indicate that the percent plasma membrane-bound gold particles perisynaptic labeling is significantly higher than that expected from a random distribution of labeling in this domain along the plasma membrane (Mann–Whitney U test; p < 0.05). (C) Relative density of Cav3.1 immunogold labeling on different portions of the plasma membrane. Normalized values were calculated as the percentage of labeling for Cav3.1 in a specific subsynaptic domain divided by the percentage of dendritic membrane that contributes to that domain along the plasma membrane. Note that the relative density of labeling for Cav3.1 is higher at perisynaptic sites than on the extrasynaptic plasma membrane. Figure 7 Summary diagram of the subsynaptic localization of Cav3.1 labeling in thalamic neurons of VA, CM and Pf. Perisynaptic labeling at asymmetric (putatively glutamatergic, green terminals with thick post synaptic density) and symmetric (unknown neurotransmitter, blue terminal with thin post synaptic density) synapses, as well as extrasynaptic and intracellular labeling are depicted (black circle). Modified from: Ladyof-tats (2007), Diagram of neurons (nerve cells), Wilkimedia Commons. References Albin RL Young AB Penney JB The functional anatomy of basal ganglia disorders Trends Neurosci 1989 12 366 375 2479133 Alexander GE Crutcher MD DeLong MR Basal ganglia-thalamocortical circuits: parallel substrates for motor, oculomotor, "prefrontal" and "limbic" functions Prog Brain Res 1990 85 119 146 2094891 Alvarez L Macias R Lopez G Alvarez E Pavon N Rodriguez-Oroz MC Juncos JL Maragoto C Guridi J Litvan I Tolosa ES Koller W Vitek J DeLong MR Obeso JA Bilateral subthalamotomy in Parkinson's disease: initial and long-term response Brain 2005 128 570 583 15689366 Alvarez L Macias R Pavon N Lopez G Rodriguez-Oroz MC Rodriguez R Alvarez M Pedroso I Teijeiro J Fernandez R Casabona E Salazar S Maragoto C Carballo M Garcia I Guridi J Juncos JL DeLong MR Obeso JA Therapeutic efficacy of unilateral subthalamotomy in Parkinson's disease: results in 89 patients followed for up to 36 months J Neurol Neurosurg Psychiatry 2009 80 979 985 19204026 Ardashov OV Pavlova AV Il'ina IV Morozova EA Korchagina DV Karpova EV Volcho KP Tolstikova TG Salakhutdinov NF Highly potent activity of (1R,2R,6S)-3-methyl-6-(prop-1-en-2-yl)cyclohex-3-ene-1,2-diol in animal models of Parkinson's disease J Med Chem 2011 54 3866 3874 21534547 Baude A Nusser Z Roberts JD Mulvihill E McIlhinney RA Somogyi P The metabotropic glutamate receptor (mGluR1 alpha) is concentrated at perisynaptic membrane of neuronal subpopulations as detected by immunogold reaction Neuron 1993 11 771 787 8104433 Belardetti F Zamponi GW Calcium channels as therapeutic targets Wiley Interdisciplinary Reviews: Membrane Transport and Signaling 2012 1 433 451 Bergman H Wichmann T Karmon B DeLong MR The primate subthalamic nucleus. II. Neuronal activity in the MPTP model of parkinsonism J Neurophysiol 1994 72 507 520 7983515 Blackstad T Karagülle T Ottersen O MORFOREL, a computer program for two-dimensional analysis of micrographs of biological specimens, with emphasis on immunogold preparations Computers in biology and medicine 1990 20 15 34 2328575 Bladen C McDaniel SW Gadotti VM Petrov RR Berger ND Diaz P Zamponi GW Characterization of novel cannabinoid based T-type calcium channel blockers with analgesic effects ACS Chem Neurosci 2015 6 277 287 25314588 Bogenpohl J Galvan A Hu X Wichmann T Smith Y Metabotropic glutamate receptor 4 in the basal ganglia of parkinsonian monkeys: ultrastructural localization and electrophysiological effects of activation in the striatopallidal complex Neuropharmacology 2013 66 242 252 22634360 Byrd CL Schwartz SJ Gonzales M Byrd CB Ciraldo RJ Sivina M Yahr WZ Greenberg JJ DDD pacemakers maximize hemodynamic benefits and minimize complications for most patients Pacing Clin Electrophysiol 1988 11 1911 1916 2463566 Buzsaki G Smith A Berger S Fisher LJ Gage FH Petit mal epilepsy and parkinsonian tremor: hypothesis of a common pacemaker Neuroscience 1990 36 1 14 2120612 Byrd CL Schwartz SJ Gonzales M Byrd CB Ciraldo RJ Sivina M Yahr WZ Greenberg JJ DDD pacemakers maximize hemodynamic benefits and minimize complications for most patients Pacing Clin Electrophysiol 1988 11 1911 1916 2463566 Carpenter CL Marks SS Watson DL Greenberg DA Dextromethorphan and dextrorphan as calcium channel antagonists Brain Res 1988 439 372 375 2451974 DeLong MR Primate models of movement disorders of basal ganglia origin Trends Neurosci 1990 13 281 285 1695404 Devergnas A Chen E Ma Y Hamada I Pittard D Kammermeier S Mullin AP Faundez V Lindsley CW Jones C Smith Y Wichmann T Anatomical localization of Cav3.1 calcium channels and electrophysiological effects of T-type calcium channel blockade in the motor thalamus of MPTP-treated monkeys J Neurophysiol 2016 115 470 485 26538609 Cain SM Snutch TP T-type calcium channels in burst-firing, network synchrony, and epilepsy Biochimica et Biophysica Acta (BBA) - Biomembranes 2013 1828 1572 1578 22885138 Catterall WA Structure and regulation of voltage-gated Ca2+ channels Annu Rev Cell Dev Biol 2000 16 521 555 11031246 Cavelier P Bossu JL Dendritic low-threshold Ca2+ channels in rat cerebellar Purkinje cells: possible physiological implications Cerebellum 2003 2 196 205 14509569 Choe W Messinger RB Leach E Eckle VS Obradovic A Salajegheh R Jevtovic-Todorovic V Todorovic SM TTA-P2 is a potent and selective blocker of T-type calcium channels in rat sensory neurons and a novel antinociceptive agent Mol Pharmacol 2011 80 900 910 21821734 Christie BR Eliot LS Ito K Miyakawa H Johnston D Different Ca2+ channels in soma and dendrites of hippocampal pyramidal neurons mediate spike-induced Ca2+ influx J Neurophysiol 1995 73 2553 2557 7666160 Connelly WM Crunelli V Errington AC The global spike: Conserved dendritic properties enable unique Ca2+ spike generation in low threshold spiking neurons J Neurosci 2015 35 15505 15522 26609149 DeLong MR Primate models of movement disorders of basal ganglia origin Trends Neurosci 1990 13 281 285 1695404 Destexhe A Neubig M Ulrich D Huguenard J Dendritic Low-Threshold Calcium Currents in Thalamic Relay Cells The Journal of Neuroscience 1998 18 3574 3588 9570789 Devergnas A Pittard D Bliwise D Wichmann T Relationship between oscillatory activity in the cortico-basal ganglia network and parkinsonism in MPTP-treated monkeys Neurobiol Dis 2014 68 156 166 24768805 Devergnas A Chen E Ma Y Hamada I Pittard D Kammermeier S Mullin AP Faundez V Lindsley CW Jones C Smith Y Wichmann T Anatomical localization of Cav3.1 calcium channels and electrophysiological effects of T-type calcium channel blockade in the motor thalamus of MPTP-treated monkeys J Neurophysiol 2016 115 470 485 26538609 Dreyfus FM Tscherter A Errington AC Renger JJ Shin HS Uebele VN Crunelli V Lambert RC Leresche N Selective T-type calcium channel block in thalamic neurons reveals channel redundancy and physiological impact of I(T)window J Neurosci 2010 30 99 109 20053892 Errington AC Renger JJ Uebele VN Crunelli V State-dependent firing determines Intrinsic dendritic Ca2+ signaling in thalamocortical neurons J Neurosci 2010 30 14843 14853 21048143 Fosset M De Weille JR Green RD Schmid-Antomarchi H Lazdunski M Antidiabetic sulfonylureas control action potential properties in heart cells via high affinity receptors that are linked to ATP-dependent K+ channels J Biol Chem 1988 263 7933 7936 2453509 Frishman WH Stroh JA Greenberg S Suarez T Karp A Peled H Calcium channel blockers in systemic hypertension Med Clin North Am 1988 72 449 499 3279287 Galvan A Hu X Smith Y Wichmann T Localization and pharmacological modulation of GABA-B receptors in the globus pallidus of parkinsonian monkeys Experimental neurology 2011 229 429 439 21419765 Galvan A Smith Y The primate thalamostriatal systems: Anatomical organization, functional roles and possible involvement in Parkinson's disease Basal Ganglia 2011 1 179 189 22773963 Galvan A Wichmann T Pathophysiology of parkinsonism Clin Neurophysiol 2008 119 1459 1474 18467168 Garber J Barbee R Bielitzki J Clayton L Donovan J Hendriksen C Kohn D Lipman N Locke P Melcher J Quimby F Turner P Wood G Wurbel H Guide for the Care and Use of Laboratory Animals 2011 8th Washington (DC) National Academies Press (US) Gauck V Thomann M Jaeger D Borst A Spatial distribution of low- and high-voltage-activated calcium currents in neurons of the deep cerebellar nuclei J Neurosci 2001 21 Rc158 11459879 Gonzales KK Pare JF Wichmann T Smith Y GABAergic inputs from direct and indirect striatal projection neurons onto cholinergic interneurons in the primate putamen J Comp Neurol 2013 521 2502 2522 23296794 Gray EG Axosomatic and axodendritic synapses in the cerebral cortex J Anat 1959 93 420 433 13829103 Gray MA Greenwell JR Argent BE Secretin-regulated chloride channel on the apical plasma membrane of pancreatic duct cells J Membr Biol 1988 105 131 142 2464063 Greenberg DA Marks SS Cation interactions with putative NMDA receptor-gated channels labeled by [3H]MK-801 in rat cerebral cortex Neurosci Lett 1988 95 236 240 2852319 Greenfield DT Greenfield LJ Hess ML Enhancement of crystalloid cardioplegic protection against global normothermic ischemia by superoxide dismutase plus catalase but not diltiazem in the isolated, working rat heart J Thorac Cardiovasc Surg 1988 95 799 813 3361928 Greenlee MW Magnussen S Spatial waveform discrimination following higher-harmonic adaptation J Opt Soc Am A 1988 5 1744 1748 3204437 Guillery RW Early electron microscopic observations of synaptic structures in the cerebral cortex: a view of the contributions made by George Gray (1924–1999) Trends Neurosci 2000 23 594 598 11137148 Guridi J Herrero MT Luquin MR Guillen J Ruberg M Laguna J Vila M Javoy-Agid F Agid Y Hirsch E Obeso JA Subthalamotomy in parkinsonian monkeys. Behavioural and biochemical analysis Brain 1996 119 Pt 5 1717 1727 8931592 Guyon A Leresche N Modulation by different GABAB receptor types of voltage-activated calcium currents in rat thalamocortical neurons J Physiol 1995 485 29 42 7658381 Hamani C Neimat JS Lozano AM Deep brain stimulation and chemical neuromodulation: current use and perspectives for the future Acta Neurochir Suppl 2007 97 127 133 17691298 Hamos JE Van Horn SC Raczkowski D Uhlrich DJ Sherman SM Synaptic connectivity of a local circuit neurone in lateral genicu- late nucleus of the cat Nature 1985 317 618 621 4058571 Hanson JE Smith Y Group I metabotropic glutamate receptors at GABAergic synapses in monkeys J Neurosci 1999 19 6488 6496 10414977 Hildebrand ME Isope P Miyazaki T Nakaya T Garcia E Feltz A Schneider T Hescheler J Kano M Sakimura K Watanabe M Dieudonne S Snutch TP Functional coupling between mGluR1 and Cav3.1 T-type calcium channels contributes to parallel fiber-induced fast calcium signaling within Purkinje cell dendritic spines J Neurosci 2009 29 9668 9682 19657020 Hopfield JF Tank DW Greengard P Huganir RL Functional modulation of the nicotinic acetylcholine receptor by tyrosine phosphorylation Nature 1988 336 677 680 3200319 Huang D Huang S Peers C Du X Zhang H Gamper N GABAB receptors inhibit low-voltage activated and high-voltage activated Ca(2+) channels in sensory neurons via distinct mechanisms Biochem Biophys Res Commun 2015 465 188 193 26239659 Huguenard JR Low-threshold calcium currents in central nervous system neurons Annual review of physiology 1996 58 329 348 Huguenard JR Gutnick MJ Prince DA Transient Ca2+ currents in neurons isolated from rat lateral habenula J Neurophysiol 1993 70 158 166 8395572 Hutchison WD Lozano AM Davis KD Saint-Cyr JA Lang AE Dostrovsky JO Differential neuronal activity in segments of globus pallidus in Parkinson's disease patients Neuroreport 1994 5 1533 1537 7948856 Ilinsky IA Ambardekar AV Kultas-Ilinsky K Organization of projections from the anterior pole of the nucleus reticularis thalami (NRT) to subdivisions of the motor thalamus: light and electron microscopic studies in the rhesus monkey J Comp Neurol 1999 409 369 384 10379824 Isope P Hildebrand ME Snutch TP Contributions of T-type voltage-gated calcium channels to postsynaptic calcium signaling within Purkinje neurons Cerebellum 2012 11 651 665 20734177 Jeanmonod D Magnin M Morel A Siegemund M Cancro A Lanz M Llinás R Ribary U Kronberg E Schulman J Zonenshayn M Thalamocortical dysrhythmia II. Clinical and surgical aspects Thalamus & Related Systems 2001 1 245 254 Jones EG The thalamus 2007 second New York, NY Cambridge University Press Kampa BM Letzkus JJ Stuart GJ Requirement of dendritic calcium spikes for induction of spike-timing-dependent synaptic plasticity The Journal of Physiology 2006 574 283 290 16675489 Kavalali ET Zhuo M Bito H Tsien RW Dendritic Ca2+ Channels Characterized by Recordings from Isolated Hippocampal Dendritic Segments Neuron 1997 18 651 663 9136773 Kopecky BJ Liang R Bao J T-type calcium channel blockers as neuroprotective agents Pflugers Arch 2014 466 757 765 24563219 Kovacs K Sik A Ricketts C Timofeev I Subcellular distribution of low-voltage activated T-type Ca2+ channel subunits (Ca(v)3.1 and Ca(v)3.3) in reticular thalamic neurons of the cat J Neurosci Res 2010 88 448 460 19774668 Krack P Limousin P Benabid AL Pollak P Chronic stimulation of subthalamic nucleus improves levodopa-induced dyskinesias in Parkinson's disease The Lancet 350 1676 Kraus RL Li Y Gregan Y Gotter AL Uebele VN Fox SV Doran SM Barrow JC Yang Z-Q Reger TS Koblan KS Renger JJ In Vitro Characterization of T-Type Calcium Channel Antagonist TTA-A2 and In Vivo Effects on Arousal in Mice Journal of Pharmacology and Experimental Therapeutics 2010 335 409 417 20682849 Kuramoto E Fujiyama F Nakamura KC Tanaka Y Hioki H Kaneko T Complementary distribution of glutamatergic cerebellar and GABAergic basal ganglia afferents to the rat motor thalamic nuclei European Journal of Neuroscience 2011 33 95 109 21073550 Kuwajima M Dehoff MH Furuichi T Worley PF Hall RA Smith Y Localization and expression of group I metabotropic glutamate receptors in the mouse striatum, globus pallidus, and subthalamic nucleus: regulatory effects of MPTP treatment and constitutive Homer deletion J Neurosci 2007 27 6249 6260 17553998 Lacey CJ Bolam JP Magill PJ Novel and distinct operational principles of intralaminar thalamic neurons and their striatal projections J Neurosci 2007 27 4374 4384 17442822 Lambert RC Bessaih T Crunelli V Leresche N The many faces of T-type calcium channels Pflugers Arch 2014 466 415 423 24043572 Leresche N Hering J Lambert RC Paradoxical potentiation of neuronal T-type Ca2+ current by ATP at resting membrane potential J Neurosci 2004 24 5592 5602 15201332 Li M Hansen JB Huang L Keyser BM Taylor JT Towards selective antagonists of T-type calcium channels: design, characterization and potential applications of NNC 55-0396 Cardiovasc Drug Rev 2005 23 173 196 16007233 Li M McCann JD Liedtke CM Nairn AC Greengard P Welsh MJ Cyclic AMP-dependent protein kinase opens chloride channels in normal but not cystic fibrosis airway epithelium Nature 1988 331 358 360 2448645 Limousin P Pollak P Benazzouz A Hoffmann D Le Bas JF Broussolle E Perret JE Benabid AL Effect of parkinsonian signs and symptoms of bilateral subthalamic nucleus stimulation Lancet 1995 345 91 95 7815888 Liu MG Lu D Wang Y Chen XF Li Z Xu Y Jin JH Wang RR Chen J Counteracting roles of metabotropic glutamate receptor subtypes 1 and 5 in regulation of pain-related spatial and temporal synaptic plasticity in rat entorhinal-hippocampal pathways Neurosci Lett 2012 507 38 42 22172929 Liu XB Munoz A Jones EG Changes in subcellular localization of metabotropic glutamate receptor subtypes during postnatal development of mouse thalamus J Comp Neurol 1998 395 450 465 9619499 Llinas R Jahnsen H Electrophysiology of mammalian thalamic neurones in vitro Nature 1982 297 406 408 7078650 Llinás R Ribary U Jeanmonod D Cancro R Kronberg E Schulman J Zonenshayn M Magnin M Morel A Siegmund M Thalamocortical dysrhythmia I. Functional and imaging aspects Thalamus & Related Systems 2001 1 237 244 Lory P Chemin J Towards the discovery of novel T-type calcium channel blockers Expert Opin Ther Targets 2007 11 717 722 17465728 Magee JC Johnston D Synaptic activation of voltage-gated channels in the dendrites of hippocampal pyramidal neurons Science 1995 268 301 304 7716525 Magnin M Morel A Jeanmonod D Single-unit analysis of the pallidum, thalamus and subthalamic nucleus in parkinsonian patients Neuroscience 2000 96 549 564 10717435 Markram H Sakmann B Calcium transients in dendrites of neocortical neurons evoked by single subthreshold excitatory postsynaptic potentials via low-voltage-activated calcium channels Proceedings of the National Academy of Sciences 1994 91 5207 5211 Masilamoni GJ Bogenpohl JW Alagille D Delevich K Tamagnan G Votaw JR Wichmann T Smith Y Metabotropic glutamate receptor 5 antagonist protects dopaminergic and noradrenergic neurons from degeneration in MPTP-treated monkeys Brain 2011 134 2057 2073 21705423 Masugi-Tokita M Shigemoto R High-resolution quantitative visualization of glutamate and GABA receptors at central synapses Curr Opin Neurobiol 2007 17 387 393 17499496 Mathai A Ma Y Paré J-F Villalba RM Wichmann T Smith Y Reduced cortical innervation of the subthalamic nucleus in MPTP-treated parkinsonian monkeys 2015 Mazloom M Smith Y Synaptic microcircuitry of tyrosine hydroxylase-containing neurons and terminals in the striatum of MPTP-treated monkeys The Journal of comparative neurology 2006 495 453 469 16485290 Mazzone P Stocchi F Galati S Insola A Altibrandi MG Modugno N Tropepi D Brusa L Stefani A Bilateral Implantation of Centromedian-Parafascicularis Complex and GPi: A New Combination of Unconventional Targets for Deep Brain Stimulation in Severe Parkinson Disease Neuromodulation 2006 9 221 228 22151710 McCormick DA Bal T Sleep and arousal: thalamocortical mechanisms Annu Rev Neurosci 1997 20 185 215 9056712 McKay BE McRory JE Molineux ML Hamid J Snutch TP Zamponi GW Turner RW CaV3 T-type calcium channel isoforms differentially distribute to somatic and dendritic compartments in rat central neurons European Journal of Neuroscience 2006 24 2581 2594 17100846 Miller LG Greenblatt DJ Roy RB Summer WR Shader RI Chronic benzodiazepine administration. II. Discontinuation syndrome is associated with upregulation of gamma-aminobutyric acidA receptor complex binding and function J Pharmacol Exp Ther 1988 246 177 182 2455789 Mitrano DA Pare JF Smith Y Ultrastructural relationships between cortical, thalamic, and amygdala glutamatergic inputs and group I metabotropic glutamate receptors in the rat accumbens J Comp Neurol 2010 518 1315 1329 20151362 Munsch T Budde T Pape HC Voltage-activated intracellular calcium transients in thalamic relay cells and interneurons Neuroreport 1997 8 2411 2418 9261800 Nedergaard S Bolam JP Greenfield SA Facilitation of a dendritic calcium conductance by 5-hydroxytryptamine in the substantia nigra Nature 1988 333 174 177 2897080 Nelson MT Todorovic SM Perez-Reyes E The role of T-type calcium channels in epilepsy and pain Curr Pharm Des 2006 12 2189 2197 16787249 Nyman G Greenlee MW Laurinen P Sampling irregularity perturbs visual reconstruction J Opt Soc Am A 1988 5 628 635 3404313 Ohara PT Chazal G Ralston HJ III Ultrastructural analysis of GABA-immunoreactive elements in the monkey thalamic ventrobasal complex J Comp Neurol 1989 283 541 558 2745753 Paquet M Smith Y Group I metabotropic glutamate receptors in the monkey striatum: subsynaptic association with glutamatergic and dopaminergic afferents J Neurosci 2003 23 7659 7669 12930805 Parajuli LK Fukazawa Y Watanabe M Shigemoto R Subcellular distribution of alpha1G subunit of T-type calcium channel in the mouse dorsal lateral geniculate nucleus J Comp Neurol 2010 518 4362 4374 20853512 Pare D Curro'Dossi R Steriade M Neuronal basis of the parkinsonian resting tremor: a hypothesis and its implications for treatment Neuroscience 1990 35 217 226 2199839 Parent A Hazrati LN Functional anatomy of the basal ganglia. II. The place of subthalamic nucleus and external pallidum in basal ganglia circuitry Brain Res Brain Res Rev 1995 20 128 154 7711765 Pasternak B Svanstrom H Nielsen NM Fugger L Melbye M Hviid A Use of calcium channel blockers and Parkinson's disease Am J Epidemiol 2012 175 627 635 22387374 Paxinos G Huang X-F Tonga AW The Rhesus Monkey Brain in Stereotaxic Coordinates 1999 1st Academic Press Peppe A Gasbarra A Stefani A Chiavalon C Pierantozzi M Fermi E Stanzione P Caltagirone C Mazzone P Deep brain stimulation of CM/PF of thalamus could be the new elective target for tremor in advanced Parkinson's Disease? Parkinsonism Relat Disord 2008 14 501 504 18337153 Perez-Reyes E Molecular Physiology of Low-Voltage-Activated T-type Calcium Channels: Physiological Reviews 2003 83 117 161 Perez-Reyes E Molecular characterization of T-type calcium channels Cell Calcium 2006 40 89 96 16759699 Perez-Reyes E Lory P Molecular biology of T-type calcium channels CNS Neurol Disord Drug Targets 2006 5 605 609 17168745 Pessiglione M Guehl D Rolland AS Francois C Hirsch EC Feger J Tremblay L Thalamic neuronal activity in dopamine-depleted primates: evidence for a loss of functional segregation within basal ganglia circuits J Neurosci 2005 25 1523 1531 15703406 Peters A Palay SL Webster HD The Fine Structure of the Nervous System: Neurons and Their Supporting Cells 1991 New York Oxford University Press Powell KL Cain SM Snutch TP O'Brien TJ Low threshold T-type calcium channels as targets for novel epilepsy treatments Br J Clin Pharmacol 2014 77 729 739 23834404 Raju DV Ahern TH Shah DJ Wright TM Standaert DG Hall RA Smith Y Differential synaptic plasticity of the corticostriatal and thalamostriatal systems in an MPTP-treated monkey model of parkinsonism European Journal of Neuroscience 2008 27 1647 1658 18380666 Ralston HJ Evidence for presynaptic dendrites and a proposal for their mechanism of action Nature 1971 230 585 587 4101752 Rhodes PA Llinas R A model of thalamocortical relay cells J Physiol 2005 565 765 781 15613378 Schneider CA Rasband WS Eliceiri KW NIH Image to ImageJ: 25 years of image analysis Nat Methods 2012 9 671 675 22930834 Shin HS Cheong EJ Choi S Lee J Na HS T-type Ca2+ channels as therapeutic targets in the nervous system Curr Opin Pharmacol 2008 8 33 41 18203662 Shipe WD Barrow JC Yang ZQ Lindsley CW Yang FV Schlegel KA Shu Y Rittle KE Bock MG Hartman GD Tang C Ballard JE Kuo Y Adarayan ED Prueksaritanont T Zrada MM Uebele VN Nuss CE Connolly TM Doran SM Fox SV Kraus RL Marino MJ Graufelds VK Vargas HM Bunting PB Hasbun-Manning M Evans RM Koblan KS Renger JJ Design, synthesis, and evaluation of a novel 4-aminomethyl-4-fluoropiperidine as a T-type Ca2+ channel antagonist J Med Chem 2008 51 3692 3695 18540666 Smith Y Galvan A Ellender TJ Doig N Villalba RM Huerta-Ocampo I Wichmann T Bolam JP The thalamostriatal system in normal and diseased states Front Syst Neurosci 2014 8 5 24523677 Smith Y Raju DV Pare JF Sidibe M The thalamostriatal system: a highly specific network of the basal ganglia circuitry Trends Neurosci 2004 27 520 527 15331233 Smith Y Villalba R Striatal and extrastriatal dopamine in the basal ganglia: An overview of its anatomical organization in normal and Parkinsonian brains Movement Disorders 2008 23 S534 S547 18781680 Song WJ Baba Y Otsuka T Murakami F Characterization of Ca(2+) channels in rat subthalamic nucleus neurons J Neurophysiol 2000 84 2630 2637 11068004 Steriade M Llinás RR The functional states of the thalamus and the associated neuronal interplay Physiol Rev 1988 68 649 742 2839857 Su PC Tseng HM Liu HM Yen RF Liou HH Subthalamotomy for advanced Parkinson disease J Neurosurg 2002 97 598 606 12296644 Talley EM Cribbs LL Lee JH Daud A Perez-Reyes E Bayliss DA Differential distribution of three members of a gene family encoding low voltage-activated (T-type) calcium channels J Neurosci 1999 19 1895 1911 10066243 Tringham E Powell KL Cain SM Kuplast K Mezeyova J Weerapura M Eduljee C Jiang X Smith P Morrison J-L Jones NC Braine E Rind G Fee-Maki M Parker D Pajouhesh H Parmar M O’Brien TJ Snutch TP T-Type Calcium Channel Blockers That Attenuate Thalamic Burst Firing and Suppress Absence Seizures Science Translational Medicine 2012 4 121ra119 121ra119 Uchizono K Characteristics of excitatory and inhibitory synapses in the central nervous system of the cat Nature 1965 207 642 643 5883646 Vila M Perier C Feger J Yelnik J Faucheux B Ruberg M Raisman-Vozari R Agid Y Hirsch EC Evolution of changes in neuronal activity in the subthalamic nucleus of rats with unilateral lesion of the substantia nigra assessed by metabolic and electrophysiological measurements Eur J Neurosci 2000 12 337 344 10651888 Villalba RM Raju DV Hall RA Smith Y GABAB receptors in the centromedian/parafascicular thalamic nuclear complex: An ultrastructural analysis of GABABR1 and GABABR2 in the monkey thalamus Journal of Comparative Neurology 2006 496 269 287 16538684 Villalba RM Wichmann T Smith Y Neuronal loss in the caudal intralaminar thalamic nuclei in a primate model of Parkinson's disease Brain Struct Funct 2014 219 381 394 23508713 Wali FA Greenidge E Tugwell AC Effect of verapamil on the contractions produced by high potassium and by noradrenaline in human isolated saphenous vein Acta Physiol Hung 1988 72 115 121 3421124 Wasserman DH Williams PE Lacy DB Green DR Cherrington AD Importance of intrahepatic mechanisms to gluconeogenesis from alanine during exercise and recovery Am J Physiol 1988 254 E518 E525 3281473 Watson DL Carpenter CL Marks SS Greenberg DA Striatal calcium channel antagonist receptors in Huntington's disease and Parkinson's disease Ann Neurol 1988 23 303 305 2967667 Williams SR Stuart GJ Action potential backpropagation and somato-dendritic distribution of ion channels in thalamocortical neurons J Neurosci 2000 20 1307 1317 10662820 Xiang Z Thompson AD Brogan JT Schulte ML Melancon BJ Mi D Lewis LM Zou B Yang L Morrison R Santomango T Byers F Brewer K Aldrich JS Yu H Dawson ES Li M McManus O Jones CK Daniels JS Hopkins CR Xie XS Conn PJ Weaver CD Lindsley CW The Discovery and Characterization of ML218: A Novel, Centrally Active T-Type Calcium Channel Inhibitor with Robust Effects in STN Neurons and in a Rodent Model of Parkinson's Disease ACS Chem Neurosci 2011 2 730 742 22368764 Yang ZQ Barrow JC Shipe WD Schlegel KA Shu Y Yang FV Lindsley CW Rittle KE Bock MG Hartman GD Uebele VN Nuss CE Fox SV Kraus RL Doran SM Connolly TM Tang C Ballard JE Kuo Y Adarayan ED Prueksaritanont T Zrada MM Marino MJ Graufelds VK DiLella AG Reynolds IJ Vargas HM Bunting PB Woltmann RF Magee MM Koblan KS Renger JJ Discovery of 1,4-substituted piperidines as potent and selective inhibitors of T-type calcium channels J Med Chem 2008 51 6471 6477 18817368 Yang YC Tai CH Pan MK Kuo CC The T-type calcium channel as a new therapeutic target for Parkinson's disease Pflugers Arch 2014 466 747 755 24531801 Yunker AMR Sharp AH Sundarraj S Ranganathan V Copeland TD McEnery MW Immunological characterization of T-type voltage-dependent calcium channel CaV3.1 (alpha1G) and CaV3.3 (alpha1I) isoforms reveal differences in their localization, expression, and neural development Neuroscience 2003 117 321 335 12614673 Zhou Q Godwin DW O'Malley DM Adams PR Visualization of calcium influx through channels that shape the burst and tonic firing modes of thalamic relaycells J Neurophysiol 1997 77 2816 2825 9163395 Zirh TA Lenz FA Reich SG Dougherty PM Patterns of bursting occurring in thalamic cells during parkinsonian tremor Neuroscience 1998 83 107 121 9466402

PMC005xxxxxx/PMC5135631.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 8902615 1463 J Chem Neuroanat J. Chem. Neuroanat. Journal of chemical neuroanatomy 0891-0618 1873-6300 27562515 5135631 10.1016/j.jchemneu.2016.08.005 NIHMS815010 Article Neurochemical Compartmentalization within the Pigeon Basal Ganglia Bruce Laura L. a Erichsen Jonathan T. b Reiner Anton c a Department of Biomedical Sciences, Creighton University, Omaha NE, USA b School of Optometry and Vision Sciences, Cardiff University, Cardiff, UK c Department of Anatomy and Neurobiology, The University of Tennessee Health Science Center, Memphis, TN, USA Corresponding author: Laura L. Bruce, Department of Biomedical Sciences, Creighton University School of Medicine, Omaha, NE 68178, lbruce@creighton.edu 8 9 2016 22 8 2016 12 2016 01 12 2017 78 6586 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. The goals of this study were to use multiple informative markers to define and characterize the neurochemically distinct compartments of the pigeon basal ganglia, especially striatum and accumbens. To this end, we used antibodies against 12 different neuropeptides, calcium-binding proteins or neurotransmitter-related enzymes that are enriched in the basal ganglia. Our results clarify boundaries between previously described basal ganglia subdivisions in birds, and reveal considerable novel heterogeneity within these previously described subdivisions. Sixteen regions were identified that each displayed a unique neurochemical organization. Four compartments were identified within the dorsal striatal region. The neurochemical characteristics support previous comparisons to part of the central extended amygdala, somatomotor striatum, and associational striatum of mammals, respectively. The medialmost part of the medial striatum, however, has several unique features, including prominent pallidal-like woolly fibers and thus may be a region unique to birds. Four neurochemically distinct regions were identified within the pigeon ventral striatum: the accumbens, paratubercular striatum, ventrocaudal striatum, and the ventral area of the lateral part of the medial striatum that is located adjacent to these regions. The pigeon accumbens is neurochemically similar to the mammalian rostral accumbens. The pigeon paratubercular and ventrocaudal striatal regions are similar to the mammalian accumbens shell. The ventral portions of the medial and lateral parts of the medial striatum, which are located adjacent to accumbens shell-like areas, have neurochemical characteristics as well as previously reported limbic connections that are comparable to the accumbens core. Comparisons to neurochemically identified compartments in reptiles, mammals, and amphibians indicate that, although most of the basic compartments of the basal ganglia were highly conserved during tetrapod evolution, uniquely avian compartments may exist as well. striatum accumbens globus pallidus ventral pallidum bed nucleus of stria terminalis avian 1. Introduction The basal ganglia play a critical role in modulating motor functions. Similar types of neurons and fibers have been identified in the basal ganglia of birds, reptiles, amphibians and mammals, which co-express similar neuropeptides and appear to have similar functions (Anderson and Reiner, 1990a; Reiner and Anderson, 1990). Homologues of the main components of the basal ganglia have been identified in mammals, birds, and reptiles, including striatum, nucleus accumbens, bed nucleus of stria terminalis, globus pallidus (or dorsal pallidum), and ventral pallidum (Medina and Reiner, 1997; Smeets et al., 2000; Roberts et al., 2002; Reiner et al., 2004b; Balint and Csillag, 2007; Kuenzel et al., 2011). The boundaries of these major subdivisions have, however, not necessarily been clearly defined in all cases. The main components of the avian striatum are the medial striatum, lateral striatum, and accumbens, each of which has a distinct neurochemical expression pattern (Reiner et al., 1994, 2004b). The medial striatum (previously named the lobus parolfactorius) has not been associated with a specific homologous field with the mammalian striatum, although it has some markers in common with the nucleus accumbens, and others in common with striatum proper (Reiner et al., 2004b). Similarly, the lateral striatum (previously the paleostriatum augmentatum) in birds is clearly striatal in nature but has not been definitively related to any one specific part of mammalian striatum. Neurochemical and connectional studies have suggested that the avian nucleus accumbens consists of regions corresponding to the mammalian accumbens core, shell, and rostrum (Balint and Csillag, 2007), but the boundaries of these regions have been elusive. Thus, in spite of numerous studies on avian basal ganglia, boundaries between and within its major subdivisions remain poorly documented, particularly in the ventral striatum, in part because only a limited number of neurochemical markers have been used for identifying different compartments. Moreover, prior studies show regional neurochemical heterogeneity throughout avian basal ganglia that needs to be better defined for a clearer understanding of avian basal ganglia organization and to facilitate its comparison to the basal ganglia in other vertebrate groups. The goals of this study were to use multiple informative markers to define and characterize the neurochemically distinct compartments of the pigeon basal ganglia, especially those of the striatum and accumbens. To this end, we used antibodies against 10 different neuropeptides, calcium-binding proteins or neurotransmitter-related enzymes known to be enriched in the basal ganglia: (1) calbindin (CALB), (2) cholecystokinin (CCK), (3) choline acetyl transferase (ChAT), (4) glutamic acid decarboxylase (GAD), (5) leucine-enkephalin-(ENK), (6) neuropeptide Y (NPY), (7) parvalbumin (PARV), (8) substance P (SP), (9) tyrosine hydroxylase (TH), and (10) vasoactive intestinal polypeptide (VIP). In addition, antibodies against calretinin (CR) and the neuropeptide cocaine- and amphetamine-regulated transcript (CART) were used to discriminate striatal accumbens territories. Our results clarify boundaries between previously described basal ganglia subdivisions in birds, and reveal considerable novel heterogeneity within these previously described subdivisions. Comparisons to neurochemically identified compartments in reptiles, mammals, and amphibians indicate that although most of the basic compartments of the basal ganglia were highly conserved during tetrapod evolution, as previously noted, uniquely avian compartments may exist as well. 2. Materials and methods The brains of sixteen adult homing pigeons (Columba livia), which were immunostained for CCK, ChAT, CR, GAD, ENK, NPY, TH, SP, VIP, CALB or PARV, were used in these studies. Sections from one of these were stained sequentially with the first 8 antibodies and provide the images in figs. 2–8. All procedures employed in this study were approved by the Animal Care Committees at State University of New York at Stony Brook and the University of Tennessee. The pigeons were deeply anesthetized with sodium pentobarbital (70 mg/kg) and perfused through the left ventricle with in 0.1M phosphate buffer (PB; pH 7.4) with 0.75% sodium chloride followed by 4% paraformaldehyde in PB. Brains were removed from the skulls, placed in fixative at 4°C for 3–4 hrs, and then transferred to a cryoprotective solution of 30% sucrose and 0.1% sodium azide in PB for 3–4 days. Brains were sectioned at 30–40 μm on a sliding, freezing microtome in the transverse plane used in the Karten and Hodos (1967) brain atlas. For each antibody, sections were washed in PB, immersed in 0.3% H2O2 in PB for 10 min, and rinsed again in PB prior to antibody incubation. Some of the tissue used in this study was used for analyses of other brain systems in earlier reports (Erichsen et al., 1991; Krebs et al., 1991; Riters et al. 1999), and the immunohistochemical procedures we employed are described there in greater detail. For the immunohistochemical staining employed to generate the tissue illustrated in figures 2–8, the following standard procedure was used (see Table 1 for optimal primary antibody concentrations and antibody sources). All sections were incubated in a 1.5% solution of normal serum in 0.1 M PB with 0.3% Triton X-100 (Sigma Chemical, St. Louis, MO) (PBX) appropriate for the animal source of the primary antibody. Following a brief wash in PB as noted above, each series of sections was then incubated in a primary antibody diluted in PBX containing 1.5% normal serum for 65–89 hours at 4°C. Several washes preceded a 1 hour incubation in biotinylated IgG (Vector Labs, Burlingame, CA) (diluted 1:200 in PBX) directed against the species originating each of the primary antibodies used in the study. After another wash in PB, the sections were preincubated in ABC solution (1:50 dilution; Vector Labs) for one hour before being placed in a solution of DAB with H2O2 for an additional 15 min. Sections were then washed, mounted onto slides, osmicated and coverslipped for subsequent examination. For immunolabeling with CALB and PARV, the method described in Laverghetta et al. (2006) was used. For SP, ENK and VIP, colchicine-treated material that had been prepared for use in prior studies was also available (Anderson and Reiner, 1990a, b, 1991). The distribution patterns of ChAT, VIP, ENK, TH, CCK, NPY, GAD, SP, CALB, PARV, CR, and CART within the lateral wall of the pigeon subpallium were analyzed in a parallel series of transverse sections through the forebrain (Figs. 2–10). Sections were imaged using a Nikon USB 5 megapixel CCD camera attached to a Nikon Optiphot microscope and connected to a High-End Desktop M55 computer. The gray scale and image contrast were adjusted, and figures labeled and formatted, using CorelDraw and Corel PhotoPaint 12. Six transverse planes through the striatum rostral to the anterior commissure were selected to illustrate the expression patterns within each neurochemical compartment using 8 stains (Fig. 1). The regional intensity of immunolabeling was densitometrically measured to quantitatively distinguish compartments (Table 2). The types of perikarya, and the dendrites and terminals that contained the markers were qualitatively analyzed to define the limits of the known subregions within the striatal and pallidal part of the subpallium, as well as to identify novel subdivisions. Immunolabeling of neuropil intensity was quantitatively measured in images prior to any contrast adjustments by recording intensity levels at three points within each compartment and at each level using the Corel Photo-Paint eyedropper tool (Table 2). Measurements were ranked at 4 levels: very intense or abundant immunolabeling, intense or abundant immunolabeling, moderately abundant or intense immunolabeling, and background or near background immunolabeling). Abundance of labeled perikarya was qualitatively rated separately (Table 3). The revised avian nomenclature of Reiner et al. (2004b) is used, but expanded upon where labeling patterns suggest there are previously unrecognized compartments. 2.1 Antibody characterization and specificity A list of all antibodies used in the present study is shown in Table 1, including the immunogens and dilutions used and their sources. Each antibody produced characteristic patterns of immunostaining that were expected based on previous reports in mammals or birds. Western blot studies of the monoclonal anti-calbindin-D-28K by the manufacturer showed a single band at 28 kD, and also showed that it does not react with other members of the EF-hand family. It recognizes calbindin, but not PARV or CR (Conde et al., 1994). Our data are consistent with a study that described the distribution pattern of CALB in the bed nucleus of stria terminalis and accumbens in the pigeon using another CALB antibody (Husband and Shimizu, 2012). In mammals calbindin immunostaining demarcates rostral associative and paralimbic subdivisions of the striatum (Morel, 2002), which it also does in the present study. The polyclonal anti-calretinin antibody recognizes a 29 kDa protein in amphibians, lizards, and chicks (Hack et al., 2000; Morona and González, 2008; Yan et al., 2010) and has been widely used in studies of diverse vertebrate species. Within the rodent basal ganglia this antibody is a marker for terminals in the olfactory tubercle, medial parts of nucleus accumbens, as well as numerous small interneurons throughout the basal ganglia (Bubser et al., 2000). The present study focused on the pattern of CR terminal labeling in the basal ganglia, which we found is consistent with results in rodents. The immunogen for the polyclonal anti-cholecystokinin octapeptide antibody is sulfated CCK-8 (26–33) coupled to bovine thyroglobulin (BTg) with glutaraldehyde. Immunostaining with this antibody is abolished by preadsorption with CCK-8. The antibody has been used studies of numerous vertebrate and invertebrate species. Although glutaraldehyde was used to produce the immunogen, this antibody works very well with paraformaldehyde-fixed tissues, as do other similarly produced antibodies against immunogens (e.g., Veenman and Reiner, 1996). This antibody is a marker for terminals in the ventral accumbens and the olfactory tubercle of rodents (Zaborszky et al., 1985; Zahm and Heimer, 1988), which is consistent with the labeling pattern seen in pigeons in the present study. The monoclonal cocaine- and amphetamine-regulated transcript peptide antibody was generated against a rat CART (54–102) fragment (Thim et al., 1998). The specificity of this antibody has been demonstrated by omission of primary antibodies and by immunoblot analysis showed a single precipitin band that migrates at about 14 kD (Subhedar et al., 2011; Singru et al., 2007). Within the basal ganglia, CART is a marker for neurons in the medial part of the accumbens shell in diverse vertebrate species, (Smith et al., 1999; Lázár et al., 2004; Barsagade et al., 2011; Subhedar et al., 2011), which is consistent with the labeling pattern observed in the present study. The anti-glutamic acid decarboxylase polyclonal antibody preferentially recognizes GAD65 and also binds to GAD67 (Oertel et al., 1981; Kaufman et al., 1991). It has been characterized by Western blot analyses using zebra finch cerebellum and forebrain, revealing bands at 61 and 59kD (Spiro et al., 1995). The rodent and pigeon basal ganglia contain numerous GAD-expressing neurons throughout, and fibers and terminals in pallidal areas stain particularly heavily with anti-GAD (Veenman et al., 1995; Sun et al. 2005), which is consistent with the labeling pattern in the present study. The mouse anti-leu-enkephalin monoclonal antibody to leu-enkephalin has been previously characterized for specificity by immunodot-blotting and by specific adsorption (Cuello et al., 1984; Milner et al., 1989), and has been used extensively in a variety of vertebrate species. In the rodent and pigeon basal ganglia anti-leu-encephalin antibodies have been used as a marker of neurons located throughout the striatal part of the basal ganglia, as well as for terminals in pallidal territories, including woolly fibers in the external segment of globus pallidus and in the ventral pallidum (Reiner et al., 1984a; Anderson and Reiner, 1990a). The present study is confirms these labeling patterns. The pattern of neuronal labeling is also consistent with prior studies using in situ hybridization histochemistry (Molnar et al., 1994). The rabbit polyclonal anti-neuropeptide Y antiserum was raised against a 36 amino acid sequence that cross reacts with human, rat, and porcine NPY, but not other closely related peptides (Kienzler et al., 2009). Within the mammalian basal ganglia, NPY is a marker for identifying the accumbens shell and ventral pallidum (Bálint and Csillag, 2007; Brauer et al., 2000), which we found useful in the present study. The mouse parvalbumin monoclonal antibody reacts with parvalbumin (12 kDa) from a variety of mammals as well as frog and fish but does not react with closely related peptides of the EF-hand family such as calmodulin and intestinal calcium-binding protein. The specificity has been examined by the manufacturer and researchers (Heizmann and Celio, 1987; Sigma-Aldrich). It has been used previously in pigeons to describe the distribution of PARV neurons in the basal ganglia (Reiner and Anderson, 1993; Lavergetta et al., 2006), and the present results are consistent with their findings. Our data are qualitatively similar to the staining patterns shown in a previous study in in the pigeon using another PARV antibody (Husband and Shimizu, 2012). In mammals, parvalbumin neuropil immunostainings demarcates a caudolateral somatomotor subdivision of the striatum (Morel, 2002), which it also does in the present study. lmmunohistochemical localization of substance P utilized a monoclonal substance P antibody (supplier: Sera-Lab, Crawley, England) raised in tissue culture from rat spleen hybridoma. The details of the production of this antibody and the specificity have been described elsewhere (Cuello et al., 1979). The specificity of the immunoreactivity for SP was assessed by the use of a blocked control in which the primary antibody was preabsorbed with synthetic SP (Reiner et al., 1983). In prior studies in mammals and birds, the striatum was found to contain numerous SP+ neurons and varying intensities of neuropil stain depending on location, whereas the neuropil of the globus pallidus and ventral pallidum are characterized by heavy SP staining but few SP+ neurons, consistent with the present results (Cuello et al., 1979; Reiner et al., 1983). The specificity of the immunostaining produced with the rabbit tyrosine hydroxylase polyclonal antiserum is well established (Pickel et al., 1975; Hervonen et al., 1980; Armstrong et al., 1981). Tyrosine hydroxylase was partially purified from bovine adrenal medulla by precipitation with ammonium sulfate and column chromatography. The antibody specificity was based on immunoelectrophoresis of the antibody run against either partially purified tyrosine hydroxylase from bovine or rat adrenal medulla which yielded a single precipitin band. No precipitin bands formed when the antibody was run against other catecholamine-synthesizing enzymes, including dopa-decarboxylase, dopamine beta-hydroxylase, or phenylethanolamine N-methyltransferase (Pickel et al., 1975). TH is a marker for dopaminergic and adrenergic neurons and fibers. For example, TH immunostaining detects dopaminergic neurons of the substantia nigra pars compacta and their terminals in the striatum. In the present study, TH immunostaining of dopaminergic terminals was used to delineate pigeon striatal subterritories. The rabbit polyclonal antiserum to vasoactive intestinal polypeptide was generously provided by Dr. J. Walsh (UCLA) and has been widely used in numerous species. The VIP antiserum is specific for the carboxyl terminal 18–28 region of VIP. Tests for specificity showed negligible reactivity with glucagon, secretin, and gastric inhibitory polypeptide (Furness et al., 1981). VIP is widespread in the brain, and we used it as a marker to delineate striatal and pallidal sub-territories. 3. Results The avian subpallium is traditionally divided into dorsal and ventral subdivisions, each of which contains striatal and pallidal territories (Reiner et al., 1994). The dorsal and ventral subpallia are associated with somatic and limbic functions of the basal ganglia, respectively. The divide between dorsal and ventral subpallial regions can be established most notably by the rich dopamine beta-hydroxylase (DBH) neuropil in the ventral subpallial striatal areas, whereas the dorsal subpallium is DBH poor (Reiner et al., 1994). In addition, we have found that a subset of regions within the DBH-poor dorsal subpallium is rich in parvalbumin, whereas the DBH-rich ventral subpallial regions, including the bed nucleus of the stria terminalis, are the opposite. This rationale for the dorsal-ventral somatic/limbic distinction in birds does not appear to apply for distinguishing mammalian dorsal and ventral subpallial regions. The differential expression patterns of Islet1, cPax6, cLmo4, and cLmo3 are consistent with this dorsal-ventral somatic/limbic distinction in birds (Abellán and Medina, 2009), and will be considered further in the Discussion. Note that we focus on subpallial territories of the lateral telencephalic wall anterior to the anterior commissure, and do not include amygdalar or septal subpallial territories. 3.1. Dorsal Subpallium Based on their neurochemistry, the medial and lateral striatum are each further divided into medial and lateral zones. The nucleus intrapeduncularis is included here as a striatal part of the dorsal subpallium, although this assignment is not unambiguous. Finally, we also recognize distinct striatal cellular islets in the boundary territory between pallium and subpallium. Only one large pallidal territory is recognized in the dorsal subpallium, namely the globus pallidus. 3.1.1. Medial part of the medial striatum Neurons As is true of dorsal striatum in general, the medial MSt contains medium spiny neurons expressing ENK, SP, and low levels of GAD immunolabeling (Fig. 11B). The sparse GAD perikaryal immunolabeling is, however, deceptive, because in situ hybridization for GAD65 and immunolabeling for GABA shows this region to be rich in GAD+/GABA+ perikarya (Veenman et al., 1995; Sun et al. 2005). Scattered ChAT+ neurons are present, although less frequent than in the more lateral parts of MSt. NPY+ and CART+ neurons are sparsely scattered throughout the medial MSt (Fig. 11A). Some VIP+ perikarya are observed in colchicine-treated material. This area has fewer CALB+ neurons than the adjacent lateral part of MSt. PARV+ neurons are also sparse in the medial MSt. Neuropil The neuropil of the medialmost part of the MSt expresses very high levels of SP, high levels of GAD and ENK, moderate levels of TH, CCK, NPY, CALB and PARV, and low levels of ChAT and VIP (Figs. 2–8). The medialmost MSt contains subtle heterogeneously distributed areas of slightly lighter and darker PARV expression, particularly near the transition with the lateralmost MSt (Fig. 9A′–D′). The low TH and CALB and higher PARV expression distinguishes the medialmost MSt, and show that it extends caudally to the level at which the tractus septopallio-mesencephalicus (TSM) reaches the basal telencephalon. Within the neuropil of medialmost MSt are “woolly fibers”, which are tight rows of terminals presumed to be from the medium spiny neurons in the striatum that are rich in SP, ENK, NPY and GAD, and that synapse along unlabeled dendrites (Fig. 11A, B). Based on prior findings for the globus pallidus (Reiner and Caraway, 1987), these woolly fiber terminals are likely to end on long, smooth dendrites of neurons that lightly express LANT6 and/or PARV. Whether these striatal neurons are interneurons or projection neurons is unknown, although PARV+/LANT6+ expressing pallidal projection neurons, but not interneurons, receive woolly fiber inputs (Reiner et al., 2004a, b). 3.1.2. Lateral part of the medial striatum Neurons Neurons containing ChAT are more plentiful in the lateral MSt than in more medial or lateral striatal compartments. The lateral MSt, like the medial MSt, is rich in medium-sized spiny projection neurons containing either SP or ENK, as evident in colchicine-treated material. Scattered NPY+ neurons were also present throughout lateral MSt. Neurons immunolabeled for GAD are scarce in the lateral MSt, although GABA immunolabeling and GAD65 in situ hybridization studies have shown that GAD-synthesizing perikarya are in fact abundant throughout both medial and lateral MSt (Veenman and Reiner, 1994; Sun et al., 2005). A few neurons immunolabeled for VIP are observed in colchicine-treated cases. Numerous PARV+ and CALB+ neurons are also present, in contrast with the medial MSt. Neuropil The striatal neuropil contains high levels of TH and GAD, moderate levels of ChAT, ENK, CCK, NPY, SP, and CALB, and low levels of VIP and PARV. The neuropil of the lateralmost MSt is distinguished from the more medial MSt by its higher expression of TH, ChAT and CALB and lower expression of PARV. The lateralmost MSt extends medially to the ventricle at its more caudal levels, that is from the level at which the TSM reaches the basal telencephalon to the anterior commissure (Figs. 8, 9F, G, G′). The lateralmost medial striatum contains some woolly fibers, which are strongly labeled for SP, NPY, and ENK, and which emanate as finger-like extensions from globus pallidus into the lateral MSt. Woolly fibers are, nonetheless, less abundant in the lateral than medial MSt. VIP and CCK are sparse in the neuropil of both the medial and lateral MSt. 3.1.3. Medial part of lateral striatum Neurons Like other striatal territories, the medial LSt is rich in medium-sized spiny projection neurons that contain either SP or ENK. The medial LSt is very poor in both ChAT and NPY neurons, in contrast to the MSt. VIP+ neurons were rarely observed, even in colchicine-treated tissue. The medial LSt contains numerous PARV+ neurons but few CALB+ neurons. Large GAD+ neurons resembling the PARV+ neurons in size and frequency are present, consistent with previous studies showing that GAD and PARV are co-expressed in the same neurons of LSt (Reiner and Anderson, 1993). Neuropil The medial part of the LSt contains high levels of TH, GAD, and PARV, moderate levels of ChAT, VIP, ENK, CCK, NPY, and SP, and low levels of CALB and (Figs. 2–8; Table 2). The medial part of the LSt extends medially to the ventricle at levels near the anterior commissure. Woolly fibers are rare, except along the border with the globus pallidus (GP), where they extend from the GP into the striatum. Except when present in woolly fibers, NPY, GAD, CCK, and VIP are expressed fairly evenly throughout the LSt neuropil. TH is present in a graded pattern that is lowest in the medial most MSt, higher in the lateralmost MSt, slightly lower in the medialmost LSt, and highest in the lateralmost LSt (Figs. 2–8). CALB is found in higher levels in the lateral part of the medial striatum than in adjacent striatal territories, whereas PARV and CART are less abundant in the lateral part of the medial striatum than in adjacent striatal territories (Figs. 9, 10). 3.1.4. Lateral part of lateral striatum Neurons The lateral LSt is very similar to the medial LSt. It is rich in SP+ and ENK+ medium-sized spiny projection neurons, but is very poor in ChAT+, NPY+, VIP+, and CALB+ neurons. There are few CALB+ neurons. Numerous large PARV+ and GAD+ neurons likely represent the same neuronal population (Reiner and Anderson, 1993). Neuropil The neuropil in the lateralmost part of the LSt contains very high levels of TH, high levels of VIP, CCK, and GAD, moderate levels of ChAT, ENK, NPY, and SP, and low levels of CALB and PARV. The lateralmost part of the LSt along the lateral margins of the globus pallidus is a neurochemically unique area, distinguished by its much higher expressions of VIP, TH, and CCK in its neuropil (Figs. 6–7), suggesting that this area may have a specialized function and thus differ from the remainder of lateral LSt. Woolly fibers are rare, except along the border with the globus pallidus (GP), where they extend from the GP into the striatum. Except when present in woolly fibers, NPY, GAD, CCK, and VIP are expressed fairly evenly throughout the lateral LSt neuropil. The lateral LSt can be readily distinguished from the medial LSt by the paucity of PARV in its neuropil. Of the markers examined in this study, the localization of TH and the calcium binding proteins provides the best means to distinguish the different striatal areas (Figs. 2–8). In the case of the lateral striatum, CALB is found in higher levels in the neuropil of the lateral striatum than in the medial striatum, whereas PARV is less abundant in the neuropil of the lateral part of the lateral striatum than in the medial parts of the lateral striatum (Fig. 9). 3.1.4. Intrapeduncular Nucleus (INP) Neurons The INP is located along the ventromedial border of globus pallidus and lateral to the MSt (Figs. 7–8). The neuronal profile of INP is largely striatal. For example, many medium-sized SP+ and ENK+ neurons are present in colchicine-treated material in INP, resembling those seen in MSt and LSt. Prior in situ hybridization studies confirm that the INP contains many SP+ neurons (Abellán and Medina, 2009) and scattered ENK+ neurons (Molnar et al., 1994), as well as medium-sized GAD+ neurons (Sun et al., 2005). NPY+ and CALB+ neurons are sparse in the INP, as they are in most other striatal areas. On the other hand, ChAT+ neurons are much more plentiful in the INP than in the medial or lateral striatum, whereas PARV+ neurons are scarcer. Neuropil The neuropil of INP contains moderate levels of ChAT, TH, GAD, and PARV, and low levels of VIP, ENK, CCK, NPY, SP, and CALB. The neuropil of the INP can be distinguished from that of the globus pallidus by the absence of woolly fibers that contain GAD, ENK and/or SP, and a more intensely PARV+ neuropil, whereas the GP neuropil is rich in woolly fibers that contain GAD, ENK and/or SP. The neuropil of INP, in general, resembles that of the striatum. The INP contains many ChAT+ fibers, likely reflecting its enrichment with cholinergic neurons. 3.1.5. Cell islands within the lamina pallio-subpallialis Neurons The dorsal and lateral edge of the striatum is formed by a band of fibers, the lamina pallio-subpallialis. Tightly clustered neurons form compact islands within it, which are particularly large and prominent rostrally (Figs. 2–3), but smaller more caudally (Figs. 4–8). Abellán and Medina have identified these islands in embryonic chick and termed them the striatal capsule. Neurons within these islands appear to contain medium-sized striatal projection neuron markers, notably SP+ and ENK+. According to Abellán and Medina (2009), these cells also express the striatal gene cLmo4 and GAD67 (their fig. 13D–F). These neurons are enriched in DARPP32+ as well (Reiner et al., 1998b). Neuropil The neuropil of the islands of the striatal capsule contain high levels of ChAT, moderate levels of ENK, TH, CALB and PARV, and low levels of VIP, GAD, CCK, NPY, and SP. The islands are most easily identified by their ChAT immunostaining because the adjacent striatum stains less intensely, whereas immunostaining for other markers is either absent or similar to that in the adjacent striatum. 3.1.6 Globus Pallidus Neurons The globus pallidus appears rostrally at approximately A11.75 as small finger-like extensions into rostral striatum (Figs. 4–5) arising from the main body of the more caudal pallidal zone (Figs. 6–8). The large GABAergic pallidal neurons are enriched in PARV, and so PARV+ neurons are plentiful in globus pallidus (Laverghetta et al., 2006). ChAT-immunolabeled neurons are relatively plentiful within the globus pallidus as well, although less numerous than in the INP. Scattered spiny SP+ and ENK+ neurons are also seen in globus pallidus (Fig. 11C), as noted previously (Reiner et al., 1983; Molnar et al., 1994), but NPY+, CCK+, and CALB+ neurons are largely absent. Neuropil Woolly fibers dominate the neuropil of the globus pallidus and contain very high levels of GAD, high levels of ENK and SP, and moderate levels of VIP (Fig. 11C). The woolly fibers arise from medium spiny neurons in the LSt, which are likewise rich in GAD, SP, and ENK, and the terminals of LSt neurons form rows of synapses along the sides of the dendrites of globus pallidus neurons. Since the globus pallidus also contains some VIP+ woolly fibers, it seems likely that some LSt medium spiny neurons do synthesize VIP, although VIP+ neurons were not detected in it by immunolabeling. The globus pallidus contains few fibers labeled for ChAT, TH, CCK, NPY, CALB or PARV, with the paucity of TH typical of a pallidal zone. 3.2. Ventral Subpallium The ventral subpallium is divided into striatal, pallidal, and mixed populations as determined by their neurochemical patterns (Abellán and Medina, 2009). Based on our findings and prior studies, we identify the following four striatal regions within the ventral subpallium: (1) nucleus accumbens, defined by Reiner et al. (2004b) as a ventral part of rostral striatum, nearest the ventricle; (2) an area lateral to nucleus accumbens that we here refer to as reticular ventral striatum; (3) a region deep to the olfactory tubercle but external to the reticular ventral striatum, which we term the paratubercular striatum; (4) a region that extends dorsal and caudal to the reticular striatum and ventral pallidum that we call the ventrocaudal striatum (Stvc). Our paratubercular striatum and Stvc corresponds to the “accumbens shell” of chicken embryos identified by Abellán and Medina (2009). A single pallidal territory of the ventral subpallium, the ventral pallidum, is considered pallidal based on the presence of woolly fibers, namely closely spaced parallel rows of synaptic terminals that immunostain strongly for DARPP32, SP, ENK, CALB or neurotensin (Reiner and Carraway, 1987; Reiner et al., 1998b). Finally, several subpallial regions have a mixture of striatal and pallidal characteristics: (1) the lateral bed nucleus of stria terminals (BStL), which is adjacent to the tip of the lateral ventricle and extends as a thin sheet rostrally along the ventricle at the medial edge of nucleus accumbens; (2) the juxtacapsular (Bjx) subdivision, which forms the lateral border of the BStL (N.B. earlier avian studies often called the BStL and Bjx the ‘nucleus accumbens’) (Reiner et al., 2004b); and (3) the olfactory tubercle, which occupies the superficial surface of the lateral subpallium and receives olfactory input (Abellán and Medina, 2009). 3.2.1. Nucleus accumbens Neurons Nucleus accumbens, as defined in Reiner et al. (2004b) is rich in SP+ and ENK+ neurons, as well as GABAergic neurons (Veenman and Reiner, 1994; Sun et al., 2005), as typical of striatal domains. ChAT+, VIP+, NPY+, CALB+, and CART+ neurons also are scattered throughout. Neuropil The neuropil of nucleus accumbens is heterogeneous, and largely characterized by very high levels of VIP, ENK, NPY, and SP, high levels of CCK and GAD, moderate levels of ChAT, TH, and CALB, and low levels of PARV (Figs. 2–5). The neurochemical characteristics of its neuropil are particularly useful for defining nucleus accumbens. For example, the rich levels of SP immunoreactivity distinguish it from adjacent dorsal striatal regions, except the medialmost MSt, which also has high SP neuropil levels. In contrast, TH is less abundant in the accumbens compared to the dorsal striatal regions, except for the medialmost MSt. The accumbens neuropil is richer in VIP, ENK, CCK, and NPY than other parts of the striatum. The VIP fibers, in particular, differentially delineate accumbens from striatal territory above and lateral to it. The NPY and ENK fibers tend to be distributed in patches of high and low abundance, not seen with other markers (Figs. 5 C, G). Caudally, the accumbens is contiguous with the rostral BStL (Fig. 5), but can be easily distinguished by the much higher levels of SP in accumbens compared to the BStL (Figs. 2–5). In addition, nucleus accumbens is one of the few striatal areas with high levels of CR and CART in the neuropil (Fig. 10). 3.2.2. Paratubercular striatum Overview The paratubercular striatum (Stp) is located superficial (i.e. ventral or external) to the accumbens and reticular striatum, but deep (i.e. ventral or internal) to the thin, rostral extension of the ventral pallidum and to the olfactory tubercle, which itself is the thin olfactory-recipient zone along the ventral surface of the subpallium. Its neuropil exhibits a unique pattern because it immunostains intensely for most markers in contrast to other striatal regions, as described below, which is the major reason for recognizing it as a distinct territory (Figs. 2–6). However, it is difficult to distinguish it from the olfactory tubercle, as both have similar staining characteristics and some neurons of Stp may receive olfactory bulb input, like the olfactory tubercle, via dendrites that extend into the olfactory terminal zone of olfactory tubercle. Nonetheless, olfactory terminals do not end in the Stp (Reiner and Karten, 1985; Atoji and Wild, 2014), and for this reason, we do not include it as part of the olfactory tubercle. The Stp expands dorsolaterally towards the MSt at the level of the reticular striatum (see below) just anterior to the rostral pole of the ventral pallidum (Fig. 5). Around the lateral aspect of VP, it blends with the Stvc (Fig. 6). Neurons The Stp contains SP+ and ENK+ neurons, but only scattered ChAT+, NPY+, and PARV+ neurons. VIP+ neurons were rarely seen. CART neurons are also present within the Stp (Fig. 10), which differentiates the Stp from most other striatal areas. Neuropil The Stp neuropil contains very high levels of VIP, ENK, TH, CCK, GAD, and SP, high levels of ChAT and NPY, moderate levels of CALB, and low levels of PARV (Figs. 2–9). Woolly fibers rich in VIP, ENK, GAD, and SP, and non-woolly fibers rich in TH and CCK are more abundant rostromedially than caudolaterally. The medial Stp neuropil also displays high levels of CART (Fig. 10H–I). 3.2.3. Reticular ventral striatum Overview The reticular part of the ventral striatum (StR) lies lateral to nucleus accumbens of Reiner et al. (2004b) and anterior to the rostral pole of the ventral pallidum (Fig. 5). It is better distinguished by its neuropil traits than by its constituent neuronal populations, as further detailed below. In particular, it is characterized by and named for the reticulated appearance imparted by the numerous fiber bundles that traverse this region, as most evident in the TH and CCK immunolabeled tissue. Neurons As a striatal region, it contains neurons rich in ENK or SP, which probably co-contain GABA (Reiner and Anderson, 1993). ChAT+ neurons are scattered throughout, in a somewhat greater abundance than in nucleus accumbens. Scattered NPY+ and CALB+ neurons are also present. Neuropil The reticular part of the ventral striatum contains very high levels of VIP and SP, high levels of ChAT, ENK, TH, CCK, NPY and GAD, but moderate levels of CALB and low levels of PARV. The high SP and VIP combined with low GAD and NPY distributions especially define this region. It exhibits a unique combination of features that distinguish it from the accumbens and the overlying MSt. First, compared to the adjacent accumbens the StR neuropil is richer in ChAT and TH, but poorer in ENK and NPY. Second, compared to the medialmost MSt, the neuropil is much richer in VIP, TH, CCK, and GAD, but poorer in PARV. Third, it is poor in ENK and NPY woolly fibers, and the more numerous ENK and NPY woolly fibers of the MSt form a clear border with the StR. Finally, the abundance of neuropil zones with lower levels of ENK, TH, CCK, NPY, GAD, and PARV, interspersed with zones displaying higher levels, distinguish it from the olfactory tubercle and paratubercular striatum below (Figs. 2–6). The pattern of TH and CCK immunolabeled fibers also characterizes this reticular region, with small zones of richly labeled neuropil interdigitating with small, poorly labeled zones, in contrast to the uniform neuropil labeling for TH and CCK throughout the dorsal striatum. 3.2.4. Ventrocaudal part of the striatum Overview The ventrocaudal part of the striatum (Stvc) appears rostrally just above the reticular striatum and extends caudally above the ventral pallidum and medial to the globus pallidus, forming a cup-shaped zone at the caudoventral pole of the lateral MSt (Figs. 7–8). It is usually included in the lateral part of the MSt because of its location just below it. Abellán and Medina (2009), however, considered this region as the caudal continuation of the paratubercular striatum, and the similar immunohistochemical labeling patterns we observed are consistent with their grouping. Although the ventrocaudal region and the MSt have some neuronal and neuropil characteristics in common, many other features readily distinguish the two regions, suggesting it is distinct from the MSt. Neurons The ventrocaudal part of MSt, like the other parts of MSt, is rich in medium-sized spiny projection neurons containing either SP or ENK, as evident in colchicine-treated material. Scattered NPY+ neurons are present, but there are few PARV+ and CALB+ neurons, similar to the medial MSt but in contrast to the lateral MSt. Neurons immunolabeled for GAD are scarce, but GABA immunolabeling and GAD65 in situ hybridization studies have shown that GAD-synthesizing perikarya are abundant in this area (Veenman and Reiner, 1994; Sun et al., 2005). Rare neurons immunolabeled for VIP are observed in colchicine-treated cases. The Stvc has scattered ChAT+ neurons, which are more abundant than in the adjacent lateral part of MSt, but far less abundant than in the adjacent globus pallidus or INP. Neuropil The neuropil of the ventrocaudal part of the striatum contains very high levels of TH and CCK, high levels of ChAT, VIP, ENK, NPY and GAD, moderate levels of CALB, and low levels of PARV. It is highly enriched in SP and ChAT. A number of these features distinguish the ventrocaudal striatum from the MSt. Notably, woolly fibers enriched with SP, ENK and/or VIP are more numerous in the Stvc. GAD woolly fibers, however, are less intensely labeled in the Stvc than in the MSt. The non-woolly part of the neuropil contains more ChAT, VIP, TH, and NPY compared to the MSt. Unlike the medial part of the MSt, but like the lateral part of the MSt, the neuropil of the ventrocaudal striatal region is rich in TH but poor in PARV. At its more rostral levels (Figs. 6, 7), it has especially high levels of TH+ and CCK+ fibers and terminals. The enrichment in CALB is similar to the lateral MSt. The similar distributions of TH and CCK in terminals in the paratubercular, reticulated, and ventrocaudal striatal regions are consistent with the possibility of their co-localization in the terminals of dopaminergic midbrain neurons in these regions. CCK has been colocalized to dopaminergic terminals in mammals (Seroogy et al., 1988). 3.2.5. Ventral Pallidum Neurons The ventral pallidum (VP) appears rostrally as a very thin band between the striatum and olfactory tubercle, and is best distinguished by its very low TH expression compared to these adjacent regions (Figs. 5D–7D). At approximately A10.75 in the Karten and Hodos atlas (1967), the VP expands dorsally, and occupies an oval zone ventrolateral to the bed nucleus of the stria terminalis (Fig. 6). It is rich in large aspiny GABAergic and PARV+ neurons, which are likely to be co-labeled projection neurons. Very few neurons with medium spiny projection neuron markers such as SP, ENK, and CALB are present, although ChAT+ neurons are scattered throughout the VP. Neuropil The neuropil of the VP contains very high levels of ENK, NPY, GAD, and SP, high levels of VIP, moderate levels of ChAT, CCK, and CALB, and low levels of PARV. The SP, ENK, GAD and NPY immunolabeling take the form of woolly fibers, which occur throughout the ventral pallidum, whereas VIP+ woolly fibers are only seen in the medial VP. ChAT, TH, and CCK immunolabeling are conspicuously poorer in VP than in adjacent regions, and thereby serve to delineate it. The complementarity between the few TH+ or CCK+ fibers, and the many SP+, ENK+, and NPY+ woolly fibers is especially striking. 3.2.6. Olfactory Tubercle Overview The olfactory tubercle is defined here as the thin subpallial region at the base of the telencephalon that receives olfactory bulb input (Figs. 2–6). Neurons The rostral and ventral olfactory tubercle contain neurons that express striatal markers, including ENK and SP. GAD+ neurons are also present in the rostral and ventromedial olfactory tubercle as expected for a striatal region. ChAT+, NPY+, VIP+, CALB+ and PARV+ neurons are scarce in this territory. Neuropil Like other striatal regions, the neuropil of the rostral ventral olfactory tubercle is rich in ENK, TH, CCK, NPY, GAD, SP, and CALB, moderate in VIP, and poor in PARV. More caudally and laterally, the olfactory tubercle has some pallidal characteristics, since it is poor in TH. 3.2.7. Lateral part of the Bed Nucleus of Stria Terminalis Overview The lateral bed nucleus of the stria terminalis (BStL) is largest at levels caudal to the nucleus accumbens (Figs. 6–8). It extends rostrally as a thin sheet along the ventricular edge of the accumbens, and can be distinguished from the accumbens by its neurochemical traits (Figs. 4–5, asterisk), as described below. Neurons The BStL is characterized by the presence of many ENK+ neurons, which are also known to be GABAergic, whereas neurons containing ChAT, NPY, SP, CALB, and PARV are scarce. By contrast, both SP+ and ENK+ neurons are abundant in nucleus accumbens. Neuropil Most of the BStL neuropil is rich in ENK but poor by comparison in VIP, although it contains smaller patches that are poor in ENK and rich in VIP. Other markers are distributed more homogeneously, and the BSTL is characterized by moderate levels of GAD and NPY, and only very low levels of ChAT, TH, CCK, and SP. The neurochemistry of the neuropil of BStL is distinctive, since it is conspicuously poor in SP, ChAT, CCK, and TH compared to the surrounding striatum, although a TH+ neuropil zone is present at caudal levels (Figs. 7D, 8D). The neuropil of the BStL and striatum express similar levels of GAD and NPY. Our results are consistent with a previous suggestion (Abellán and Medina, 2009) that the BStL has both pallidal and striatal traits, since we observed the pallidal traits of low TH and CCK, and the striatal trait of many ENK+ neurons. The BStL pallidal traits stem from its development from the pallidal sector of the lateral ventricle where it is located, and striatal neurons such as those containing ENK migrate into the pallidal territory (Abellán and Medina, 2009). The BStL expresses very low levels of CALB and PARV, in agreement with the description of Husband and Shimizu (2011). 3.2.8. External parts of the Lateral Bed Nucleus of Stria Terminalis Overview The BStL is bordered at its dorsal and ventral edges by an external zone, with the external zone border being located ventrally by the VP and dorsally by the MSt. We identify this region because it differs from the BSTL itself in its neurochemistry, but also appears distinct from the adjacent MSt and VP. Husband and Shimizu (2011) subdivided this external region into dorsal and ventral parts. Neurons The external region of the BStL is richer than the BStL in ChAT+ neurons, but poorer in ENK+ neurons. Neurons containing NPY and PARV are also scarce, although CALB+ neurons are scattered within it, as previously noted by Husband and Shimizu (2011). Neuropil The neuropil of the dorsal part of the external region is rich in ENK and VIP, moderate in ChAT, TH, CCK, and SP, but is poor in NPY and PARV and contains patches of CALB. The enrichment of the neuropil of the ventral part of the external region in CCK and VIP readily distinguishes it from the lighter immunostaining for these in the adjacent BSTL, ventral pallidum, and septum. ChAT, TH, GAD, ENK, and SP do not differentiate these two BStL regions. However, CALB expression is distributed more evenly in the ventral than in the dorsal part. 4. Discussion Previous studies of the neurochemistry of the avian basal ganglia identified the fundamental regions that are similar to the mammalian striatum, accumbens, globus pallidus, and ventral pallidum (Reiner et al., 1998; Mezey and Csillag, 2002; Reiner et al., 2004b; Balint and Csillag, 2007). The present neurochemical analysis, together with previous neurochemical and connectional studies of the basal ganglia, confirms these fundamental observations, and furthermore identifies several additional compartments and their boundaries. These data lead us to consider several possible previously unrecognized homologies between the avian and mammalian basal ganglia, as well as note that some may be unique to birds or, more broadly, to birds and reptiles. 4.1. Medial and Lateral Striatum The dorsal striatum of birds comprises the dorsalmost part of most of the lateral subpallium, and is composed of the medial and lateral striatum, formerly known as the lobus parolfactorius and paleostriatum augmentatum, respectively (Karten and Hodos, 1967; Reiner et al., 2002, 2004b). Our results illustrate striking differences in immunolabeling intensities between the medial and lateral striatum, which confirms and extends the differences noted by many other investigators (Reiner et al., 1983; Wynne and Gunturkun, 1995; Gallatioto et al., 1998; Ballint and Csillag, 2007). The histochemical differences may reflect functional differences, as for example reflected in connectional differences between the medial and lateral striatum. The avian medial striatum is connected with areas often characterized as limbic (Yamamoto and Reiner, 2005). For example, it receives its principal inputs from the prehippocampal area, olfactory cortex, caudolateral nidopallium, and posterior amygdala, and projects to the dopaminergic substantial nigra (Karten and Dubbeldam, 1973; Brauth et al., 1978; Reiner et al., 1983; Bottjer et al., 1989; Veenman et al., 1995). The lateral striatum receives inputs predominantly related to sensory and motor processing, including the hyperpallium (visual and somatosensory cortical-like areas), and lateral nidopallium (Veenman et al., 1995). The lateral striatum projects primarily to the globus pallidus, whereas the medial striatum projects primarily to the substantia nigra pars compacta and ventral tegmental area (Karten and Dubbeldam, 1973; Brauth et al., 1978; Lewis et al., 1981; Reiner et al., 1983, 1998a; Bottjer et al., 1989; Bottjer, 1993; Castro and Ball, 1994; Grisham and Arnold, 1994; Reiner et al., 1994; Medina and Reiner, 1995; Soha et al., 1996; Luo and Perkel, 1999; Sun and Reiner, 2000). As detailed below, however, our results support division of medial striatum into neurochemically distinct medial and lateral zones, as suggested by Abellán and Medina (2009), and that the lateral striatum of birds is also divided into neurochemically distinct medial and lateral zones. Most striatal regions are enriched with dopaminergic terminals, identifiable by intense granular TH immunolabeling, and a high density of medium-sized GABAergic neurons with spiny dendrites containing SP, ENK, GAD and dopamine-regulated neuronal phosphoprotein (DARPP32) (Reiner et al., 1983; Anderson and Reiner 1990a, b; Reiner et al., 1994; Veenman and Reiner, 1994). Medium-sized spiny projection neuronal perikarya containing either SP or ENK are evident in colchicine-treated material (Anderson and Reiner, 1990a, b). Although all medium-sized spiny neurons are thought to be GABAergic, their perikarya do not immunolabel well for either GAD or GABA (Veenman et al., 1994), but can be detected by in situ hybridization for GAD65 (Sun et al., 2005). Various striatal interneuron types are also characteristic of striatum, including large neurons containing ChAT, large GABAergic neurons containing parvalbumin, and medium-sized interneurons co-containing somatostatin, NPY and NADPHd (nicotinamide adenine dinucleotide phosphate-diaphorase) (Reiner et al., 1998a). Note that somatostatin, NPY and NADPHd are largely found in the same striatal interneurons in both mammals and birds, but in birds, somatostatin and NPY also occur in many medium spiny projection neurons (Anderson and Reiner, 1990b). Thus, NADPHd (which is known to represent the enzymatic activity of nitric oxide synthase) selectively identifies the interneuron cells in pigeons, while somatostatin and NPY immunolabeling do not. Other studies show that NADPHd+ neurons are present in both medial and lateral striatum in birds (Brüning, 1993; Brüning et al., 1994; Atoji et al., 2001). The lateral striatum with its paucity of cholinergic interneurons differs from medial striatum due to its relative abundance in cholinergic interneurons, as noted here and previously (Medina and Reiner, 1994), whereas both striatal sectors contain NADPHd+ interneurons and PARV+ interneurons. The predominant outputs of the medial and lateral striatum also differ, as the medial striatum largely projects to the midbrain nigral region, whereas the lateral striatum projects mainly to globus pallidus (Karten and Dubbeldam, 1973; Reiner et al., 2004b). Further details about medial and lateral striatum are considered in the next section in which the avian dorsal striatum is compared to that in mammals. 4.1.1. Comparisons of the avian striatum to striatum in mammals The avian striatum, like the mammalian striatum, develops from a Dlx1/2-rich and Nkx2.1-poor neuroepithelium (Fernandez et al., 1998; Puelles et al., 2000). The mammalian striatum contains neurochemically and hodologically distinct striosomes (or patches) dispersed within the striatal matrix, but such a segregation of patches within the striatal matrix is not obvious in birds (Karten and Dubbeldam, 1973; Brauth et al., 1978; Reiner et al., 1983, 1994, 1998a; Bottjer et al., 1989; Bottjer, 1993; Castro and Ball, 1994; Grisham and Arnold, 1994; Medina and Reiner, 1995; Soha et al., 1996; Durstewitz et al., 1999; Luo and Perkel, 1999). The matrix compartment of mammalian striatum is further divided into functional territories (termed T1–T3 by Morel et al., 2002) that are distinguished by their neuropil content of different markers, in particular CALB and PARV (Morel et al., 2002, François et al., 1994; Holt et al., 1997; Joel and Winer, 1997; Prensa et al, 2003; Riedel et al., 2002). The accumbens core comprises the T4 region of Morel et al., (2002) and is discussed with the ventral striatum below. The matrix of the caudal and lateral putamen (T1 region of Morel et al., 2002) of mammals is characterized by low levels of CALB and very high levels of PARV. It receives its dominant input from somatosensory and motor cortical areas, and for this reason is sometimes called somatomotor striatum (Morel et al., 2002; Künzle, 1975; Parent and Hazrati, 1995; Prensa 1999; Riedel et al., 2002). This compartment in mammals resembles the medial part of the lateral striatum in pigeons, which likewise contains low CALB and high PARV levels (present data). The medial part of the LSt receives a prominent input from the rostral Wulst of the hyperpallium (a motor cortical area) and projects to globus pallidus, which together form the motor output circuit of the avian basal ganglia (Veenman et al., 1995; Medina et al., 1997; Wild and Williams, 2000; Shimizu et al., 1995). Additionally, the medial LSt and mammalian somatomotor striatum have similar developmental origins (Abellán and Medina, 2009). Although the medial LSt of pigeons is certainly at least analogous to the somatomotor striatum of mammals, the immunohistochemical, connectional, and developmental similarities between it and the mammalian somatomotor striatum suggest that they are most likely homologous rather than independently derived similarities. The matrix of the rostral dorsal caudoputamen (T2 of Morel et al., 2002) is characterized by high levels of CALB and very little PARV; it receives its main inputs from associative cortical territories such as the visual, sensory-motor, cingulate, and entorhinal cortices as well as the basolateral amygdala, and projects to the globus pallidus (Mesulam, 1985; McGeorge and Faull, 1989; Parent and Hazrati, 1995; Haber and McFarland, 1999; Morel et al., 2002; Prensa et al., 2003; Riedel et al., 2002; Künzle, 2005). The matrix of the rostral ventral caudoputamen (T3 region of Morel et al., 2002) is also characterized by high levels of CALB and very little PARV, similar to the T2 region in mammals. In addition, the main inputs to T3 arise from paralimbic cortical territories such as prefrontal, insular, perirhinal and entorhinal cortices, and amygdala (Mesulam, 1985; Haber and McFarland, 1999; Morel et al., 2002; Prensa et al., 2003; Riedel et al., 2002; Künzle, 2005). Our results show that immunochemistry of the lateral part of the medial striatum of birds resembles both the rostral dorsal and rostral ventral matrix (T2 and T3) of the caudoputamen, as it, also, has high CALB/low PARV. The connections further support these comparisons with associative-like inputs arising predominantly from the caudal Wulst of the hyperpallium, a largely visual cortical area, from the caudolateral nidopallium, a limbic associative-like pallial area and limbic-like projections from periolfactory and paralimbic pallial areas and amygdalar areas (Veenman et al., 1995; Kröner and Güntürkün, 1999), and output projections to the globus pallidus (Farries et al., 2005; Abellán and Medina, 2009; Kuenzel et al., 2011). In addition to these hodological and immunohistochemical similarities, developmental parallels also support the view that the lateral part of MSt is homologous to the mammalian associative caudoputamen (Abellán and Medina, 2009; Kuenzel et al., 2011). Thus, the immunohistochemical, connectional and developmental characteristics of the lateral part of the MSt resemble both the associational and limbic striatal areas of mammals. Further work is needed to determine if separate limbic and association striatal areas can be differentiated within the lateral MSt, and to what degree the lateral MSt is homologous or independently derived from the ancestral amniote striatal organization. The projections to the medialmost part of the MSt resemble those to the mammalian T3 matrix, as medialmost MSt receives its input from such limbic-associated regions as the prehippocampal area, the pyriform cortex, prepyriform cortex, and nucleus taeniae (olfactory amygdala area), the core of the arcopallium, and the caudolateral nidopallium (Veenman et al., 1995; Atoji and Wild 2014). Moreover, medialmost MSt projects to dopamine neurons in the ventral tegmental area, as does the mammalian limbic striatum (Mezey and Csillag, 2002). The medialmost MSt of pigeons, however, shows less CALB and more PARV than the mammalian T3, and also shows fewer TH+ fibers and more SP+ woolly fibers than any of the mammalian striatal matrix areas (Voorn et al., 1989; Morel et al., 2002; Holt et al., 1997). Thus, it is uncertain if the two regions are homologous but with divergent neurochemistry, or if they are merely similar in connectivity but evolutionarily separately derived (i.e., analogous). The possibility has also been raised that medialmost MSt might correspond to a mammalian striosome-like compartment that has not been interwoven into a matrix compartment (Bálint and Csillag, 2007; Kuenzel et al., 2011). The low CALB and PARV, moderate TH, and heavy SP levels of medial MSt are consistent with this possibility, as is its limbic-associated input. The rostroventral striosomes near the ventricle appear late in neurogenesis in mammals, in agreement with the late neurogenesis of the medial striatum in birds (Tsai et al., 1981a, b; Song & Harlan, 1994). However, the medialmost MSt in birds is unique in its enrichment of woolly fibers, which are absent in the mammalian striatal striosomes and matrix. Additionally, high mu opiate receptor levels are a characteristic of striosomes, and the medial striatum possesses low levels, although slightly higher than in more lateral striatal areas (Reiner et al., 1989; Wang et al., 2007), which led Reiner et al. (1989) to suggest that neurons homologous to striosomes may be homogeneously distributed in the striatum of birds. Thus, a clear mammalian homologue of the medial MSt is uncertain and it appears to be a divergent region. Finally, the caudal and lateralmost part of the lateral striatum of pigeons expresses low levels of both PARV and CALB, and does not readily compare to any part of the mammalian striatum. Several prior studies have suggested that at least part of the caudalmost part of LSt may correspond to the central extended amygdala of mammals (Abellán and Medina, 2009; Bruce, 2012). 4.1.2. Comparisons of the avian striatum to striatum in reptiles and amphibians The reptilian striatum has a medial–lateral differentiation similar to most species that have been studied. In Caiman, the subpallial region termed the ventrolateral area is comparable to the striatum and contains heterogeneous distributions of markers that define at least three regions comparable to those identified here in pigeons (Brauth, 1984; Brauth et al., 1985; Brauth, 1988; Brauth et al., 1988). Those regions are: (1) a rostromedial region that is CALB poor but contains high levels of ENK+, SP+ and ChAT+ in neuronal perikarya and neuropil, and is thus comparable to the medial MSt of pigeons; (2) a dorsolateral region that has a CALB+, ENK+, TH+, and serotonin+ neuropil, and is thus comparable to pigeon lateral medial striatum; and (3) a ventrolateral region with a TH+ neuropil, comparable to pigeon lateral striatum. It is uncertain whether the lateral striatum of Caiman can be further divided into sectors, as we have here for pigeons. The striatum (paleostriatum augmentatum) of the turtle Pseudemys contains a medial region with higher levels of ChAT+, ENK+, NPY+, SP+, and somatostatin+ cells and neuropil and thus is comparable to avian medial striatum, and a lateral region with a similar neuronal profile but higher levels of dopaminergic terminals in the neuropil and thus is comparable to avian lateral striatum (Reiner et al., 1984; Reiner, 1987; Reiner and Oliver, 1987; Smeets et al., 1987; Powers and Reiner, 1993). In lizards, the neuropil of the medial striatum contains higher levels of ENK and somatostatin than the lateral striatum (Russchen et al., 1987; Pérez-Clausell and Fredens, 1988), but other markers (dopamine and acetylcholinesterase) did not reveal additional compartments (Smeets et al, 1986a, b). Studies of CALB and PARV localization are needed to determine if the striatum of turtles and lizards possesses two medial striatum and two lateral striatum sectors, as in pigeons. The amphibian striatum can also be divided into neurochemical compartments. In contrast to pigeons, the medial striatal region in amphibians contains higher levels of TH and ChAT and lower levels of SP and ENK compared to the lateral striatal region, although occasional patches that express high levels of ENK are present near the ventricle (Marin et al., 1997, 1998; Mühlenbrock-Lenter et al., 2005). The neuropil at the border of medial and lateral regions contains higher levels of CALB (Morona and González, 2008). Together, these data suggest that there may be three neurochemical compartments in amphibian striatum. In amphibians, however, the features of the medial and lateral striatal compartments do not correspond to those of the similarly located compartments in birds. Moreover, striatal neurons tend to be concentrated along the ventricle in amphibians, and few are located more laterally. By contrast, in birds they are uniformly distributed in striatum. Thus, the compartments in amphibians may be independently derived and unrelated to those in birds. 4.2. Ventral Striatum The avian ventral striatum consists of the nucleus accumbens, the paratubercular striatum, the reticular striatum, Stvc, and the olfactory tubercle. In addition, the lateral StM extends ventrally so that it is adjacent to ventral striatal territories, and its ventralmost part will be discussed as a possible ventral striatal territory. Each compartment is distinguished by unique combinations of neurochemical characteristics, as noted above. The olfactory tubercle is clearly similar to its mammalian namesake, but homologs of the other compartments are less straightforward. In mammals, the accumbens core and accumbens shell regions are often subdivided into medial and lateral regions based on their neurochemical properties. 4.2.1. Nucleus Accumbens: Core-like region The human and marmoset nucleus accumbens can be divided into medial and lateral compartments based on their CR+, CALB- and PARV-neuropil content (medial and lateral T4 region of Morel et al., 2002). The medial T4 region corresponds with the rodent rostral accumbens (Meredith et al., 1996) and will be considered with the shell-like regions below. The characteristics of the lateral subdivision (lateral T4 region of Morel et al., 2002) of the human and rodent accumbens core closely resemble those seen in the ventralmost lateral MSt of pigeons. The primate accumbens core contains a lateral compartment (lateral T4) that is neurochemically similar to the adjacent (T2 and T3) striatal matrix, with high levels of CALB and very low levels of PARV (Meredith et al., 1996; Morel et al., 2002). This region also contains high levels of TH+ fibers, moderate levels of ENK and SP, and low levels of NPY (Zaborsky et al., 1985; Voorn et al., 1989; Zahm and Brog, 1992; Heimer et al., 1997; Brauer et al., 2000; Riedel et al., 2002, Prensa et al., 2003). Thus, the similar characteristics and location of the ventralmost lateral MSt suggest it may be homologous to the lateral subdivision of the accumbens core compartment of humans and rodents. 4.2.2. Nucleus accumbens: Shell-like regions The accumbens shell region of rodents is distinguished from the core region by its low levels of CALB (Voorn et al., 1989; Zahm and Brog, 1992; Jongen-Rêlo et al., 1994). By this definition, the main shell region of rodents is often subdivided into medial and lateral divisions, but also includes most of the rostral pole of accumbens. The rodent rostral pole is comparable to the human medial striatal subdivision (medial T4 region of Morel et al., 2002), a medial core region of humans, and is dealt with separately below. The main part of the mammalian accumbens shell (medial and lateral divisions) contains higher levels of TH+, SP+, VIP+, and CCK+ fibers than the accumbens core (Zaborsky et al., 1985; Voorn et al., 1989; Zahm and Brog, 1992; Heimer et al., 1997; Brauer et al., 2000; Riedel et al., 2002, Prensa et al., 2003). CART+ perikarya and a CR+ neuropil are located predominantly in the medial accumbens shell (Bubser et al., 2000; Fagergren and Hurd 2007). In birds, the homologue of the mammalian accumbens shell has been identified relatively recently (Roberts et al., 2002; Reiner et al., 2004b; Ballint and Csillag, 2007; Abellán and Medina, 2009), although its boundaries remained unclear. Based on similar neurochemical similarities, both the paratubercular and Stvc regions resemble the mammalian shell of the accumbens, although the Stvc immunostains slightly less for VIP, ENK, and GAD than the paratubercular region. The Lmo4 gene has been identified as a marker for the accumbens shell in chicks and mammals (Abellán and Medina, 2009), and this Lmo4+ region in chicks roughly coincides with the paratubercular striatum plus Stvc of pigeons, consistent with a homology with the mammalian medial and lateral shell regions. In pigeons, striatal CART+ cells are primarily located in the medial part of the medial part of the paratubercular region, suggesting it may be comparable to the mammalian medial accumbens shell. The location of the mammalian medial and lateral accumbens shell also resembles the paratubercular and Stvc of pigeons, being sandwiched between the accumbens core and the ventral pallidum. Interestingly, in pigeons, the Stvc forms a cup around the caudal pole of the ventral part of the lateral MSt, much like the accumbens shell that surrounds the caudal accumbens core in mammals (Zahm and Brog, 1992), giving this homology a further topological similarity. 4.2.3 Nucleus accumbens: Rostromedial-like regions In humans, the neuropil of the medial subdivision of the accumbens core (medial T4 region of Morel et al., 2002) immunostains at very high levels for CR, moderate levels of CALB, and low levels of PARV (Morel, 2002). In rodents, this compartment appears to correspond to a low CALB region in the rostral accumbens and the striosomal patches that extend from it into the accumbens core, which label with very high levels of SP, low CALB and PARV (Voorn et al., 1989; Zahm and Brog, 1992; Zahm and Heimer, 1993; Meredith et al., 1996; Riedel et al., 2002). In rodents this region is often considered a shell-like region because of its low CALB+ neuropil and its projections to the ventral pallidum, hypothalamus, and ventral tegmental area (Zahm and Heimer, 1993). The region identified as the avian nucleus accumbens in Reiner et al. (2004b) is similarly characterized by a neuropil with very high levels of SP and CR, moderate levels of CALB, and low levels of PARV, compared to the overlying dorsal striatum. Thus, the similar combination of neurochemical staining and the rostral medial location suggests that the avian accumbens of Reiner et al. (2004b) may be comparable to the medial accumbens core of humans and to the rostral accumbens clusters and striosome-like extensions of rodents. 4.2.4. Nucleus accumbens: Connectional and functional correlations Connectional similarities further support these shell and core comparisons between mammals and pigeons. For example, in mammals, the ventral pallidum receives projections from the medial and lateral shell and core, and in pigeons, it receives projections from the paratubercular region, the Stvc, and from the ventral part of the lateral MSt (Groenewegen et al., 1999; Medina and Reiner, 1997; Bálint et al., 2011). Both the avian and mammalian accumbens shell and core exhibit largely similar efferent projections, including their major projections to the ventral pallidum, basal nucleus of Meynert, nucleus of the diagonal band, lateral hypothalamus, lateral preoptic area, subthalamic nucleus, substantia nigra and parabrachial region, although a projection from neurotensin-labeled neurons in the accumbens (likely corresponding to our Stvc) and BSTL project to the parabrachial nucleus appears to be unique to birds (Bálint et al., 2011, 2016). The accumbens core and shell of both birds and mammals are distinguished by some of their afferent projections. The hippocampal subicular and CA1 regions project predominantly to the rostral and medial parts of the shell (Groenewegen et al., 1999; Groenewegen et al., 1987; Brog et al., 1993). Similarly, the pigeon hippocampal formation projects to the accumbens and paratubercular regions (Atoji and Wild, 2004). Furthermore, the avian Stvc, ventral part of the lateral MSt (proposed accumbens shell and core), and BSTL receive projections from the dorsal arcopallium, which may be comparable in part to the mammalian projection from the pallial amygdala (Hanics et al., 2016). The subdivisions of the mammalian accumbens shell and core have distinct connections and distinct roles in processing appetitive behaviors, (e.g., Kelley, 1999; Zahm, 1999; Heimer et al., 1997). The mammalian nucleus accumbens plays a critical role in learning and regulating reward-associated motivational responses, with the shell and core regions contributing distinct functions (Mogenson et al., 1980; Swerdlow and Koob, 1987; LeMoal and Simon 1991; Heimer et al., 1991; Zahm and Brog, 1992; Kalivas et al., 1999; Haber and McFarland, 1999). The shell receives motivationally relevant information from areas such as the ventral tegmental area, the medial part of the ventral pallidum, and the prefrontal cortex. This information is integrated to determine the intensity of the response and reaches the accumbens core. The accumbens core then regulates the initiation of the motor response through projections to motor regions such as the substantia nigra, subthalamic nucleus, and pedunculopontine nuclei. Comparable studies in birds are important for resolving this possible homology. Further investigations are needed to determine if the potential accumbens shell and core subterritories that we have identified histochemically have functional interactions similar to those in mammals. 4.2.5. Striatal reticular area The striatal reticular area is distinguished by small interdigitating zones that stain lightly vs. darkly with ENK, TH, CCK, NPY, GAD, and PARV. The darker staining zones appear to represent the neuropil of the paratubercular region as it passes over the rostral pole of the ventral pallidum to fuse with the Stvc, which is located dorsal to the enlarged ventral pallidum. The interdigitating lighter staining zones may represent the rostral pole of the ventral pallidum, or an axonal tract that passes along the rostral pole of the ventral pallidum, or both. Because we could not distinguish between these possibilities, we recognize this as a distinct region that requires further investigation. 4.2.6. Comparisons of ventral striatum compartments to reptiles and amphibians A mammalian-like accumbens region associated with the rostral striatum has been identified in both reptiles and amphibians. In frogs it is characterized by very high levels of TH, SP, and ENK (Marin et al., 1998), and thus appears to be comparable to the shell-like regions of pigeons, StP and Stvc, and to the mammalian accumbens shell. In the lizard Psamodromus, two accumbens regions have been identified (Guirado et al., 1999). The rostromedial accumbens expresses high levels of TH, SP, GABA, whereas the caudolateral accumbens expresses levels more similar to the adjacent striatum, suggesting a possible shell and core division, respectively, similar to those in birds and mammals. 4.3. Olfactory tubercle The similarities between the olfactory tubercle of birds, reptiles, and mammals are striking. In birds, reptiles, and mammals, it receives a projection from, and projects to, the olfactory bulb (Martinez Garcia et al., 1991; Lohman and Smeets, 1993; Lanuza and Halpern, 1998; Atoji and Wild, 2014). Embryologically, it is formed predominantly by pallidal derivatives medially and striatal derivatives laterally (including the islands of Calleja), based on gene expression in birds and mammals (Abellán and Medina, 2009). Most neurochemical markers are richly abundant in the neuropil (ENK, TH, CCK, NPY, GAD and SP) and in cells (ENK, SP, and GAD). This region has been identified in multiple studies, and its similarity to the like-named mammalian region has been noted (Striedter et al., 1998; Puelles et al., 2000; Roberts et al., 2002; García-López et al., 2008; Reiner et al., 2004b; Abellán and Medina 2009). 4.3. Globus Pallidus and Ventral Pallidum The avian pallidum, like that of mammals, includes the globus (dorsal) pallidus and ventral pallidum, which are distinguished from striatal areas by an abundance of woolly fibers that at least contain substance P or enkephalin, and low levels of dopaminergic and cholinergic fibers and acetylcholinesterase (Young et al, 1984; Heimer et al., 1987; Medina and Reiner, 1995; Heimer et al., 1999), which are largely pallidal characteristics. Our data confirm these earlier findings. Furthermore, the pallidal nature of these groups is also indicated by their origin from comparable Nkx2.1-expressing regions during development, the medial ganglionic eminence in mammals or its equivalent in birds, as well as their expression of chicken Lhx6 and chicken Lhx7/8 (Puelles et al., 2000; Abellán and Medina, 2009). Nonetheless, in reptiles and birds the globus pallidus is not divided into internal and external pallidal segments as in mammals, since SP+ and ENK+ woolly fibers have overlapping distributions in birds, unlike in mammals, in which they are segregated to the internal and external pallidal segments, respectively. Of interest, our data indicate the existence of a prominent population of VIP+ striatal neurons that project to GP, given its enrichment in VIP+ woolly fibers. This population was not seen with colchicine treatment and will require sensitive in situ hybridization to detect. Similarly, the presence of VIP+ and NPY+ woolly fibers in VP indicates the existence of VIP+ and NPY+ striatal projection neurons projecting to VP. The present and prior studies indicate these are likely to reside in MSt and nucleus accumbens (Anderson et al., 1990b). The boundaries of the ventral pallidum have occasionally been misidentified in previous studies in birds, often locating the ventral pallidum more rostral or medial than its true location as defined here by woolly fiber localization, or mislabeling it as the accumbens shell (e.g., Kuenzel et al., 2011). Recognizing the full extent of the VP boundary is a challenge even with histochemical markers such as VIP, CCK and NPY, because they do not occupy the entire the ventral pallidum. In this regard, SP, ENK, and GAD provide useful positive markers for identifying VP, as they heavily stain the entire nucleus. TH and CCK define the VP and GP by their absence. 4.4. Intrapeduncular nucleus The intrapeduncular nucleus was once considered a pallidal region because of its location between the globus pallidus and the ventral pallidum (Karten and Dubbeldam, 1973; Medina and Reiner, 1994), and suggested to be comparable to globus pallidus internus of mammals. Indeed, it contains relatively few ChAT+ and TH+ terminals in its neuropil, giving it this pallidal characteristic. However, the intrapeduncular nucleus is devoid of SP, ENK, and VIP, and GAD woolly fibers, which is the most defining feature of the globus pallidus and ventral pallidum (Reiner et al., 2004b). Moreover, it contains neurons expressing striatal markers Lmo4, SP, and Cdh8, but has few cells expressing pallidal markers Nkx2.1, Lhx6, or Lhx7/8 (Abellán and Medina, 2009, Vicario et al., 2014). These findings suggest that INP is a dorsal striatal structure that contains a small number of cells originating from pallidal areas during development. A clear mammalian homologue is uncertain, and further studies are needed to determine if this cell group represents a part of the central extended amygdalar complex, part of the striatum, or part of the pallidum. As no clear homologue is evident in reptiles, INP may be uniquely evolved in avians. Along these lines, it is important to note that INP is highly enriched in ChAT+ neurons, as observed here and reported previously by Medina and Reiner (1994). It may be that, in the absence of intrinsic cholinergic neurons, these are the source of the cholinergic terminals in lateral striatum, and enable a role of the type of muscarinic mechanisms in motor plasticity as seen in mammals (Pisani et al., 2007). 4.5. Bed nucleus of stria terminalis: Lateral and external parts Although the location of the BSTL varied in earlier studies in pigeons, it is recognized and clearly differentiated from the striatal and accumbens regions by its low levels of TH, ChAT, NPY, SP, DARPP-32, parvalbumin and calbindin and its relative richness in ENK and calretinin (Reiner et al., 1998b; Durstewitz et al., 1999; Reiner et al., 2004b; Balint and Csillag, 2007; Husband and Shimizu, 2011; this study). Our data confirm the low levels of ChAT and SP and show that the BSTL can also be identified by lower expression levels of VIP, TH and NPY, relative to adjacent areas. The expression of VIP, ENK and CCK is more complex, with regions of higher and lower intensity within the BSTL. Our BSTL appears to correspond to what was termed the dorsomedial part of the BSTL (BSTLdm) in embryonic chicks, which contains lower levels of cLmo4 and cSP, relative to adjacent areas (Abellán and Medina, 2009). The external part of the BSTL forms the dorsal and ventral margins of the BSTL, respectively. The dorsal part of this border region displays higher levels of ChAT (or AChE), CCK, NPY, TH and VIP than the BSTL (present study). The ventral part displays higher levels of TH, CCK, and NPY, but lower levels of ChAT (or AChE). The ventral external BSTL corresponds to an area that expresses higher levels of calbindin, and the dorsal BSTL corresponds to a heterogeneous area with clusters of parvalbumin, calretinin, and calbindin neuropil, called the ventral and dorsal regions of nucleus accumbens by Husband and Shimizu (2011). Based on these characteristics, the pigeon dorsal external BSTL is comparable to the dorsal capsular and dorsal central parts of the rodent BSTL (Alheid et al., 1995). The pigeon ventral external BSTL is comparable to rodent juxtacapsular and ventral BSTL groups (Riedel et al., 2002; Alheid et al., 1995). The pigeon dorsal external BSTL appears to correspond to the dorsolateral BSTL of embryonic chicks, in which cells express cLmo4, cPax6, cCdh8. The ventral external BSTL corresponds to the ventral BSTL of embryonic chicks, with cells that express cLmo4 (Abellán and Medina, 2009). 4.5.1. Bed nucleus of stria terminalis: Comparison to reptiles and amphibians A homologue of the mammalian BSTL, a rostral part of the central extended amygdala, has been recognized in amphibians, but not yet in reptiles, although labeling patterns suggest it is present (Smeets et al., 1986a, b; Reiner et al., 1984b, 1987; Russchen et al., 1987; Marin et al., 1998). In frogs, the bed nucleus of the stria terminalis expresses low levels of TH and high levels of ENK, as in birds and mammals. 4.6. Summary The analysis of a combination of markers in the pigeon basal ganglia provides a powerful tool for identification of discrete, neurochemically distinct compartments and for comparisons to compartments in other species. This study provides a detailed description of the distribution of some of the major neurochemical markers of the basal ganglia. By combining multiple markers known to discriminate different regions of the mammalian basal ganglia and comparing their distribution over the full rostrocaudal extent of the subpallium, the present study has yielded a more comprehensive view of the heterogeneous organization of the basal ganglia in birds and a better resolution of the boundaries of the various groups. The combination of neuronal markers used here distinguished sixteen compartments in the pigeon basal ganglia. Four compartments were identified in the dorsal striatum, including a medial compartment that appears to be unique to birds based in part on its high content of woolly fibers, an associational striatal compartment, a lateral somatomotor striatal compartment, and a lateralmost compartment that is part of the central extended amygdala. Regions within the ventral striatum were identified that appear to be comparable to the rostral pole, shell, and core of the mammalian accumbens (Fig. 12). The distinct compartments align in many cases with particular sets of projections reported in various mammalian species. Comparative analyses of our results demonstrate that most basal ganglia compartments are highly conserved among tetrapods. However, there are also divergent areas that may have evolved independently in birds, notably the medialmost MSt and the INP. The compartments of the basal ganglia may have arisen from distinct genetic domains that are recognized differently by ingrowing axons. It will be important to determine how the neurochemical compartments relate to particular sets of projections, and to identify the genetic mechanisms that control their origins and development. Supported by NS-19620, NS-28721, and NS-57722 The Methodist Hospitals Endowed Professorship in Neuroscience (AR).). We are grateful for the technical assistance of Gary Henderson, Sherry Cuthbertson, Ellen Karle, and Marion Joni. Abbreviations Ac accumbens BM nucleus basalis magnocellularis BSTL bed nucleus of stria terminalis, lateral part BJ juxtacapsular part of BSTL Bv ventral part of the BSTL GP globus pallidus INP intrapeduncular nucleus LFB lateral forebrain bundle LOT lateral olfactory tractl LPS lamina pallio-subpallialis LSt lateral striatum MSt medial striatum MStM medial part of the medial striatum LStM lateral part of the medial striatum StVC ventral striatum, ventrocaudal part StP ventral striatum, paratubercular region QF quintofrontal tract TSM tractus septopallio-mesencephalicus TuO tuberculum olfactorium VP ventral pallidum Fig. 1 Lateral view of the forebrain of Columbia livia. Numbers on top refer to the figure number that illustrates transverse sections from that level. Fig. 2 Series of transverse sections through the rostral basal ganglia comparing immunostaining patterns of ChAT (A), VIP (B), ENK (C), TH(D), CCK (F), NPY G), GAD (H), and SP (I) in different compartments of the striatum and ventral pallidum. These are illustrated in a schematic diagram (E). The series is arranged from rostral (A) to caudal (I). Note that two patches of neurons, the MI (open arrowhead) and islands of cells within the LPS (open arrows) express different neurochemical patterns than the adjacent striatal matrix. Rostrocaudal level is approximately A13.50. Fig. 3 Series of transverse sections through the rostral basal ganglia at approximately A12.50, arranged as explained in Fig. 2. Open arrows - islands of cells in the lamina pallio-subpallialis. Fig. 4 Series of transverse sections through the rostral basal ganglia at approximately A11.75, arranged as explained in Fig. 2. Asterisk - thin rostral extension of the BStL; closed arrows - finger-like extensions of the GP into the medial striatum; open arrows - islands of cells within the LPS. Fig. 5 Series of transverse sections through the rostral basal ganglia at approximately A10.75, arranged as explained in Fig. 2. Asterisk - thin rostral extension of the BStL; closed arrows - finger-like extensions of the GP into the medial striatum; open arrows - islands of cells within the LPS. Fig. 6 Series of transverse sections through the rostral basal ganglia at approximately A10.00, arranged as explained in Fig. 2. Closed arrows - finger-like extensions of the GP into the medial striatum; open arrows - islands of cells in the LPS. Fig. 7 Series of transverse sections through the rostral basal ganglia at approximately A9.50, arranged as explained in Fig. 2. Closed arrows - finger-like extensions of the GP into the medial striatum; open arrows - islands of cells in the LPS. Fig. 8 Series of transverse sections through the rostral basal ganglia at approximately A9.00, arranged as explained in Fig. 2. Open arrows - islands of cells in the LPS. Fig. 9 Series of transverse sections through the rostral basal ganglia comparing immunostaining patterns of CALB (A–G) and PARV (A′–G′). Images from the 7 levels shown in Fig. 1 are arranged from rostral (A, A′) to caudal (G, G′). Fig. 10 Transverse sections through the rostral striatum immunostaining patterns of CR (A–G) and CART (H–N). Note that the neuropil of nucleus accumbens stains with high levels of CR and CART, whereas the neuropil of the medial part of the paratubercular striatum stains with high levels of CART+ but low levels of CR. Fig. 11 Images of woolly fibers stained with NPY (A), GAD (B), and SP (C) in the medial MSt (A, B) and GP (C). Note that NPY+ and GAD+ fibers (arrows), terminals, and cells (arrowheads) are evident in the medial MST (A, B) and SP+ fibers (arrows) and cells (arrowheads) are evident in the globus pallidus (C). Fig. 12 Schematic drawings summarizing the neurochemical subdivisions observed in this study, and proposed striatal and accumbens relationships. A, B, and C correspond to rostrocaudal levels A12.50, A10.75, and A9.50, respectively. Table 1 Antibodies Used in Immunohistochemistry1 Antibody Immunogen Optimal [1°] Source Mouse anti-CALB Bovine kidney calbindin-D-28K 1:500 C9848; Sigma-Aldrich, St. Louis, MO Mouse anti-CART Rat CART (54–102) conjugated to ovalbumin 1:2000 Dr. L. Thim, Novo Nordisk A/S, Bagsvaerd, Denmark Rabbit anti-CCK Sulfated CCK-8 (26–33) 1:1000 20078; Incstar, Stillwater, MN Rabbit anti-CR recombinant human calretinin 1:500 7699/3; Swant (Bellinzona, Switzerland); Mouse anti-Leu-ENK Leu5-enkephalin conjugated to bovine serum albumin 1:1000 MAS083; Sera-Lab; Crawley Down, UK; Cuello et al., 1984 Sheep anti-GAD Partially purified rat brain synaptosomes 1:3000–4000 1440-4; National Institutes of Health; Oertel et al., 1981 Rabbit anti-NPY 36 amino acid sequence 1:1,500 RAS7172N; Peninsula Labs Mouse anti-PARV Purified frog muscle parvalbumin 1:500–1:1000 P3088; Sigma-Aldrich, St. Louis, MO; Heizmann et al., 1987 Rat anti-SP Carboxylic terminal fragment of human SP 1:1000 MAS035b; Sera-Lab, Crawley Down, Sussex, UK Rabbit anti-TH TH purified from bovine adrenal medulla 1:1000–2000 TE101; Eugene Tech, Allendale, NJ; Armstrong et al., 1981 Rabbit anti-VIP Carboxyl terminal 18–28 region of VIP 1:1000 7916; Dr. J.H. Walsh, Univ. California Los Angeles 1 For abbreviations, see text. Table 2 Relative densities of neuropil immunostaining for markers examined in subregions of the subpallium Dorsal Subpallium Ventral Subpallium MStM MStL LStM LStL LPS GP INP Ac StP StR StVC VP TuO BSTL BV BD CALB 2 2 1 1 2 1 1 2 2 2 2 2 2 2 2 1 CART 4 3 3 3 3 2 2 4 4-3 3 4 4 4 4 4 4 CCK 2 2 2 3 1 1 1 3 4 3 4 2 1 2 4 2 ChAT 1 2 2 2 3 1 2 2 3 3 3 2 2 2 1 1 CR 3 2 2 3 2 4 2 4 4 3 3 3 4 4 4 3 GAD 3 3 3 3 2 4 2 3 4 3 3 4 1 3 3 2 LENK 3 2 2 2 2 2 1 4 4 3 3 4 1 4 4 1 NPY 2 2 2 2 1 1 1 4 3 3 3 4 1 2 3 1 PARV 2 1 2 1 2 1 2 1 1 1 1 1 1 1 1 1 SP 4 2 2 2 1 3 1 4 4 4 4 4 1 2 3 1 TH 2 3 3 4 1 1 2 2 4 3 4 3 1 2 1 1 VIP 1 1 2 3 1 3 1 4 4 4 3 3 3 3 4 2 Intensity of antibody expression: 1 – low; 2 – moderate; 3 – high; 4 – very high Table 3 Relative abundance of immunostained neurons for markers examined in subregions of the subpallium Dorsal Subpallium Ventral Subpallium MStM MStL LStM LStL LPS GP INP Ac StP StR StVC VP TuO BSTL BV BD CALB 2 4 1 1 1 1 1 1 1 1 1 1 3 3 3 3 CART 3 1 1 1 1 0 0 1 1 2 1 1 0 2 1 0 CCK 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 ChAT 1 2 1 1 1 3 4 2 2 2 3 4 3 3 3 3 CR 3 2 2 3 2 4 2 4 4 3 3 3 4 4 4 3 GAD 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 LENK 4 4 4 4 4 2 3 4 4 4 4 1 3 4 4 4 NPY 4 3 1 1 1 1 1 4 3 3 2 1 3 3 2 3 PARV 1 2 2 1 2 4 1 1 1 1 1 3 1 1 1 1 SP 4 4 4 4 4 2 4 4 4 4 4 1 3 1 2 2 TH 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 VIP 2 1 1 1 1 0 1 2 2 2 1 0 1 1 1 1 Abundance of labeled cells: 0 – none; 1 – negligible; 2 – few; 3 – some; 4 – many Highlights Sixteen distinct compartments were identified in the pigeon basal ganglia using multiple neurochemical markers. The striatum contains neurochemical regions comparable to the mammalian somatomotor and associational striatum. A neurochemically distinct area located in the medialmost striatum of pigeons appears to be unique to birds. The ventral striatum contains neurochemical regions similar to the mammalian accumbens core, shell, and rostral areas. Most of the main compartments of the basal ganglia were highly conserved during tetrapod evolution, yet unique avian compartments representing diversification have also evolved. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. 6.0 References 1 Abellán A Medina L 2009 Subdivisions and derivatives of the chicken subpallium based on expression of LIM and other regulatory genes and markers of neuron subpopulations during development J Comp Neurol 515 465 501 19459222 2 Alheid GF De Olmos JS Beltramino CA 1995 Amygdala and extended amygdala Paxinos G The Rat Nervous System Academic Press San Diego 495 578 3 Anderson KD Reiner A 1990a Extensive co-occurrence of substance P and dynorphin in striatal projection neurons: an evolutionarily conserved feature of basal ganglia organization J Comp Neurol 295 339 369 1693632 4 Anderson KD Reiner A 1990b Distribution and relative abundance of neurons in the pigeon forebrain containing somatostatin, neuropeptide Y, or both J Comp Neurol 299 261 282 1977774 5 Anderson KD Reiner A 1991 Striatonigral projection neurons: a retrograde labeling study of the percentages that contain substance P or enkephalin in pigeons J Comp Neurol 303 658 673 1707424 6 Armstrong DM Pickel VM Joh TH Reis DJ Miller RJ 1981 Immunocytochemical localization of catecholamine synthesizing enzymes and neuropeptides in area postrema and medial nucleus tractus solitarius of rat brain J Comp Neurol 196 505 517 6163796 7 Atoji Y Yamamoto Y Suzuki Y 2001 Distribution of NADPH diaphorase-containing neurons in the pigeon central nervous system J Chem Neuroanat 21 1 22 11173217 8 Atoji Y Wild JM 2004 Fiber connections of the hippocampal formation and septum and subdivisions of the hippocampal formation in the pigeon as revealed by tract tracing and kainic acid lesions J Comp Neurol 475 426 61 15221956 9 Atoji Y Wild JM 2014 Efferent and afferent connections of the olfactory bulb and prepiriform cortex in the pigeon (Columba livia) J Comp Neurol 522 1728 1752 24222632 10 Bálint E Csillag A 2007 Nucleus accumbens subregions: hodological and immunohistochemical study in the domestic chick (Gallus domesticus) Cell Tissue Res 327 221 230 17028892 11 Bálint E Balázsa T Zachar G Mezey S Csillag A 2016 Neurotensin: revealing a novel neuromodulator circuit in the nucleus accumbens-parabrachial nucleus projection of the domestic chick Brain Struct Funct 221 605 616 25344119 12 Bálint E Mezey S Csillag A 2011 Efferent connections of nucleus accumbens subdivisions of the domestic chicken (Gallus domesticus): an anterograde pathway tracing study J Comp Neurol 519 2922 2953 21618229 13 Barsagade VG Mazumdar M Singru PS Thim L Clausen JT Subhedar N 2010 Reproductive phase-related variations in cocaine- and amphetamine-regulated transcript (CART) in the olfactory system, forebrain, and pituitary of the female catfish, Clarias batrachus (Linn) J Comp Neurol 518 2503 2524 20503424 14 Bottjer SW 1993 The distribution of tyrosine hydroxylase immunoreactivity in the brains of male and female zebra finches J, Neurobiol 24 51 69 8093477 15 Bottjer SW Halsema KA Brown SA Miesner EA 1989 Axonal connections of a forebrain nucleus involved with vocal learning in zebra finches J Comp Neurol 279 312 326 2464011 16 Brauer K Häusser M Härtig W Arendt T 2000 The core-shell dichotomy of nucleus accumbens in the rhesus monkey as revealed by double-immunofluorescence and morphology of cholinergic interneurons Brain Res 858 151 162 10700608 17 Brauth SE 1984 Enkephalin-like immunoreactivity within the telencephalon of the reptile Caiman crocodilus Neuroscience 11 345 358 6371580 18 Brauth SE 1988 The organization and projections of the paleostriatal complex of Caiman crocodilus Schwerdtfeger WK Smeets WJAJ The Forebrain of Reptiles. Int. Symp. Recent Advances in Understanding the Structure and Function of the Forebrain in Reptiles Karger (Basel) Frankfurt 60 76 19 Brauth SE Ferguson JL Kitt CA 1978 Prosencephalic pathways related to the paleostriatum of the pigeon (Columba livia) Brain Res 147 205 221 647399 20 Brauth SE Kitt CA Gerfen CR 1988 Calcium binding protein in the basal ganglia system of a non-mammalian vertebrate: an immunohistochemical study in the reptile Caiman crocodilus Brain Res 452 367 372 3401743 21 Brauth SE Kitt CA Price DL Wainer BH 1985 Cholinergic neurons in the telencephalon of the reptile Caiman crocodilus Neurosci Lett 58 235 240 4047484 22 Brog JS Salyapongse A Deutch AY Zahm DS 1993 The patterns of afferent innervation of the core and shell in the “accumbens” part of the rat ventral striatum: immunohistochemical detection of retrogradely transported fluoro-gold J Comp Neurol 338 255 278 8308171 23 Bruce LL 2012 Evolution of the amygdala Yilmazer-Hanke D Insights into the Amygdala Nova Science New York 1 24 24 Brüning G 1993 Localization of NADPH-diaphorase in the brain of the chicken J Comp Neurol 334 192 208 7690049 25 Brüning G Wiese S Mayer B 1994 Nitric oxide synthase in the brain of the turtle Pseudemys scripta elegans J Comp Neurol 348 183 206 7529267 26 Bubser M Scruggs JL Young CD Deutch AY 2000 The distribution and origin of the calretinin-containing innervation of the nucleus accumbens of the rat Eur J Neurosci 12 1591 1598 10792437 27 Castro JM Ball GF 1994 Characterization and localization of D1 dopamine receptors in the sexually dimorphic vocal control nucleus, area X, and the basal ganglia of European starlings J, Neurobiol 25 767 780 8089655 28 Condé F Lund JS Jacobowitz DM Baimbridge KG Lewis DA 1994 Local circuit neurons immunoreactive for calretinin, calbindin D-28k or parvalbumin in monkey prefrontal cortex: distribution and morphology J Comp Neurol 341 95 116 8006226 29 Cuello AC Galfré G Milstein C 1979 Detection of substance P in the central nervous system by a monoclonal antibody Proc Natl Acad Sci USA 76 3532 3536 386341 30 Cuello AC Milstein C Coutre R Wright B Priestley JV Jarvis J 1984 Characterization and immunocytochemical application of monoclonal antibodies against enkephalins J Histochem Cytochem 32947 957 31 Durstewitz D Kroner S Gunturkun O 1999 The dopaminergic innervation of the avian telencephalon Prog Neurobiol 59 161 195 10463794 32 Erichsen JT Bingman VP Krebs JR 1991 The distribution of neuropeptides in the dorsomedial telencephalon of the pigeon (Columba livia): a basis for regional subdivisions J Comp Neurol 314 478 492 1726107 33 Fagergren P Hurd Y 2000 Human cocaine- and amphetamine-regulated transcript (CART) mRNA is highly expressed in limbic- and sensory-related brain regions J Comp Neurol 425 583 598 10975881 34 Farries MA Meitzen J Perkel DJ 2005 Electrophysiological properties of neurons in the basal ganglia of the domestic chick: conservation and divergence in the evolution of the avian basal ganglia J Neurophysiol 94 454 467 15772239 35 Fernandez AS Pieau C Repérant J Boncinelli E Wassef M 1998 Expression of the Emx-1 and Dlx-1 homeobox genes define three molecularly distinct domains in the telencephalon of mouse, chick, turtle and frog embryos: implications for the evolution of telencephalic subdivisions in amniotes Development 125 2099 2111 9570774 36 François C Yelnik J Percheron G Tande D 1994 Calbindin D-28k as a marker for the associative cortical territory of the striatum in macaque Brain Res 633 331 336 8137167 37 Furness JB Costa M Walsh JH 1981 Evidence for and significance of the projection of VIP neurons from the myenteric plexus to the taenia coli in the guinea pig Gastroenterology 80 1557 1561 6262184 38 Galatioto S Abbate F Laura R Naccari F Germanà G 1998 Morphological and immunohistochemical considerations on the basal ganglia in pigeon (Columba livia) Anat Histol Embryol 27 173 178 9652145 39 García-López M Abellán A Legaz I Rubenstein JL Puelles L Medina L 2008 Histogenetic compartments of the mouse centromedial and extended amygdala based on gene expression patterns during development J Comp Neurol 506 46 74 17990271 40 Grisham W Arnold AP 1994 Distribution of GABA-like immunoreactivity in the song system of the zebra finch Brain Res 651 115 122 7922557 41 Groenewegen HJ Wright CI Beijer AV Voorn P 1999 Convergence and segregation of ventral striatal inputs and outputs Ann N Y Acad Sci 877 49 63 10415642 42 Groenewegen HJ Vermeulen-Van der Zee E te Kortschot A Witter MP 1987 Organization of the projections from the subiculum to the ventral striatum in the rat. A study using anterograde transport of Phaseolus vulgaris leucoagglutinin Neuroscience 23 103 120 3683859 43 Guirado S Dávila JC Real MA Medina L 1999 Nucleus accumbens in the lizard Psammodromus algirus: Chemoarchitecture and cortical afferent connections J Comp Neurol 405 15 31 10022193 44 Haber SN McFarland NR 1999 The concept of the ventral striatum in nonhuman primates Ann NY Acad Sci 877 33 48 10415641 45 Hack NJ Wride MC Charters KM Kater SB Parks TN 2000 Developmental changes in the subcellular localization of calretinin J Neurosci 20 RC67 10729356 46 Hanics J Teleki G Alpár A Székely AD Csillag A 2016 Multiple amygdaloid divisions of arcopallium send convergent projections to the nucleus accumbens and neighboring subpallial amygdala regions in the domestic chicken: a selective pathway tracing and reconstruction study Brain Struct Funct 4 6 Epub ahead of print 47 Heimer L Zahm DS Churchill L Kalivas PW Wohltmann C 1991 Specificity in the projection patterns of accumbal core and shell in the rat Neuroscience 41 89 125 2057066 48 Heimer L Alheid GF de Olmos JS Groenewegen HJ Haber SN Harlan RE Zahm DS 1997 The accumbens: beyond the core-shell dichotomy J Neuropsychiatry Clin Neurosci 9 354 381 9276840 49 Heimer L de Olmos JS Alheid GF Pearson J Sakamoto N Shinoda K Marsteiner J Switzer RC III 1999 The human basal forebrain. Part II Bloom FE Bjorklund A Hokfelt T Handbook of Chemical Neuroanatomy. Vol 15. The Primate Nervous System, Part 3 Elsevier Amsterdam 57 226 50 Hervonen A Pickel VM Joh TH Reis DJ Linnoila I Kanerva L Miller RJ 1980 Immunocytochemical demonstration of the catecholamine-synthesizing enzymes and neuropeptides in the catecholamine-storing cells of human fetal sympathetic nervous system Adv Biochem Psychopharmacol 25 373 378 6160735 51 Holt DJ Graybiel AM Saper CB 1997 Neurochemical architecture of the human striatum J Comp Neurol 384 1 25 9214537 52 Husband SA Shimizu T 2011 Calcium-binding protein distributions and fiber connections of the nucleus accumbens in the pigeon (Columba livia) J Comp Neurol 519 1371 1394 21452202 53 Joel D Weiner I 1997 The organization of the basal ganglia–thalamocortical circuits: open interconnected rather than closed segregated Neuroscience 63 363 379 54 Jongen-Rêlo AL Voorn P Groenewegen HJ 1994 Immunohistochemical characterization of the shell and core territories of the nucleus accumbens in the rat Eur J Neurosci 6 1255 1264 7526940 55 Kalivas PW Churchill L Romanides A 1999 Involvement of the pallidal-thalamocortical circuit in adaptive behavior Ann N Y Acad Sci 877 64 70 10415643 56 Karten HJ Dubbeldam JL 1973 The organization and projections of the paleostriatal complex in the pigeon (Columba livia) J Comp Neurol 148 61 89 4698909 57 Karten HJ Hodos W 1967 A Stereotaxic Atlas of the Brain of the Pigeon Columba livia The Johns Hopkins University Press Baltimore 58 Kaufman DL Houser CR Tobin A 1991 Two forms of the gamma-aminobutyric acid synthetic enzyme glutamate decarboxylase have distinct intraneuronal distributions and cofactor interactions J Neurochem 56 720 723 1988566 59 Kelley AE 1999 Functional specificity of ventral striatal compartments in appetitive behaviors Ann NY Acad Sci 877 71 90 10415644 60 Krebs JR Erichsen JT Bingman VP 1991 The distribution of neurotransmitters and neurotransmitter-related enzymes in the dorsomedial telencephalon of the pigeon (Columba livia) J Comp Neurol 314 467 477 1687688 61 Kröner S Güntürkün O 1999 Afferent and efferent connections of the caudolateral neostriatum in the pigeon (Columba livia): a retro- and anterograde pathway tracing study J Comp Neurol 407 228 260 10213093 62 Kuenzel SJ Medina L Csillag A Perkel DJ Reiner A 2011 The avian subpallium: new insights into structural and functional subdivisions occupying the lateral subpallial wall and their embryological origins Brain Res 1424 67 101 22015350 63 Künzle H 1975 Bilateral projections from precentral motor cortex to the putamen and other parts of the basal ganglia. An autoradiographic study in Macaca fascicularis Brain Res 88 195 209 50112 64 Künzle H 2005 The striatum in the hedgehog tenrec: Histochemical organization and cortical afferents Brain Res 1034 90 113 15713262 65 Lanuza E Halpern M 1998 Efferents and centrifugal afferents of the main and accessory olfactory bulbs in the snake Thamnophis sirtalis Brain Behav Evol 51 1 22 9435967 66 Laverghetta AV Toledo CA Veenman CL Yamamoto K Wang H Reiner A 2006 Cellular localization of AMPA type glutamate receptor subunits in the basal ganglia of pigeons (Columba livia) Brain Behav Evol 67 10 38 16219996 67 Lázár G Calle M Roubos EW Kozicz T 2004 Immunohistochemical localization of cocaine- and amphetamine-regulated transcript peptide in the central nervous system of the frog Rana esculenta J Comp Neurol 477 324 339 15305368 68 Le Moal M Simon H 1991 Mesocorticolimbic dopaminergic network: functional and regulatory roles Physiol Rev 71 155 234 1986388 69 Lewis JW Ryan SM Arnold AP Butcher LL 1981 Evidence for a catecholaminergic projection to area X in the zebra finch J Comp Neurol 196 347 354 7217361 70 Lohman AH Smeets WJ 1993 Overview of the main and accessory olfactory bulb projections in reptiles Brain Behav Evol 41 147 55 8477339 71 Luo M Perkel DJ 1999 Long-range GABAergic projection in a circuit essential for vocal learning J Comp Neurol 403 68 84 10075444 72 Marin O Smeets WJ Gonzalez A 1997 Distribution of choline acetyltransferase immunoreactivity in the brain of anuran (Rana perezi, Xenopus laevis) and urodele (Pleurodeles waltl) amphibians J Comp Neurol 382 499 534 9184996 73 Marin O Smeets WJ Gonzalez A 1998 Basal ganglia organization in amphibians: chemoarchitecture J Comp Neurol 392 285 312 9511919 74 Martinez-Garcia F Olucha FE Teruel V Lorente MJ Schwerdtfeger WK 1991 Afferent and efferent connections of the olfactory bulbs in the lizard Podarcis hispanica J Comp Neurol 305 337 347 1709182 75 McGeorge AJ Faull RL 1989 The organization of the projection from the cerebral cortex to the striatum in the rat Neuroscience 29 503 537 2472578 76 Medina L Reiner A 1994 Distribution of choline acetyltransferase immunoreactivity in the pigeon brain J Comp Neurol 342 497 537 8040363 77 Medina L Reiner A 1995 Neurotransmitter organization and connectivity of the basal ganglia in vertebrates: implications for the evolution of basal ganglia Brain Behav Evol 46 235 258 8564466 78 Medina L Reiner A 1997 The efferent projections of the dorsal and ventral pallidal parts of the pigeon basal ganglia, studied with biotinylated dextran amine Neuroscience 81 773 802 9316028 79 Meredith GE Pattiselanno A Groenewegen HJ Haber SN 1996 Shell and core in monkey and human nucleus accumbens identified with antibodies to calbindin- D28k J Comp Neurol 365 628 639 8742307 80 Mesulam MM 1985 Patterns in behavioral neuroanatomy: Association areas, the limbic system, and hemispheric specialization Plum F Baringer JR Gilman S Principles of Behavioral Neurology Davis Philadelphia 1 192 81 Mezey SE Csillag A 2002 Selective striatal connections of midbrain dopaminergic nuclei in the chick (Gallus domesticus) Cell Tissue Res 308 35 46 12012204 82 Milner TA Pickel VM Reis DJ 1989 Ultrastructural basis for interactions between central opioids and catecholamines. I Rostral ventrolateral medulla J Neurosci 9 2114 2130 2566665 83 Mogenson GJ Jones DL Yim CY 1980 From motivation to action: functional interface between the limbic system and the motor system Prog Neurobiol 14 69 97 6999537 84 Molnar M Casini G Davis BM Bagnoli P Brecha NC 1994 Distribution of preproenkephalin mRNA in the chicken and pigeon telencephalon J Comp Neurol 348 419 432 7844256 85 Morel A Loup F Magnin M Jeanmonod D 2002 Neurochemical organization of the human basal ganglia: Anatomofunctional territories defined by the distributions of calcium-binding proteins and SMI-32 J Comp Neurol 443 86 103 11793349 86 Morona R González A 2008 Calbindin-D28k and calretinin expression in the forebrain of anuran and urodele amphibians: further support for newly identified subdivisions J Comp Neurol 511 187 220 18781620 87 Mühlenbrock-Lenter S Endepols H Roth G Walkowiak W 2005 Immunohistological characterization of striatal and amygdalar structures in the telencephalon of the fire-bellied toad Bombina orientalis Neuroscience 134 705 719 15961238 88 Oertel WH Schmechel DE Tappaz ML Kopin IJ 1981 Production of a specific antiserum to rat brain glutamic acid decarboxylase by injection of an antigen-antibody complex Neuroscience 61 2689 2700 89 Parent A Hazrati LN 1995 Functional anatomy of the basal ganglia. I The cortico-basal ganglia-thalamo-cortical loop Brain Res Rev 20 91 127 7711769 90 Pérez-Clausell J Fredens K 1988 Chemoarchitectonics in the telencephalon of the lizard Podarcis hispanica Schwerdtfeger WK Smeets WJAJ The Forebrain of Reptiles Int Symp Recent Advances in Understanding the Structure and Function of the Forebrain in Reptiles Karger (Basel) Frankfurt 85 96 91 Pickel VM Joh TH Reis DJ 1975 Ultrastructural localization of tyrosine hydroxylase in noradrenergic neurons of brain Proc Natl Acad Sci USA 72 659 663 235760 92 Pisani A Bernardi G Ding J Surmeier DJ 2007 Re-emergence of striatal cholinergic interneurons in movement disorders Trends Neurosci 30 545 553 17904652 93 Powers AS Reiner A 1993 The distribution of cholinergic neurons in the central nervous system of turtles Brain, Behav Evol 41 326 345 8324620 94 Prensa L Giménez-Amaya JM Parent A 1999 Chemical heterogeneity of the striosomal compartment in the human striatum J Comp Neurol 413 603 618 10495446 95 Prensa L Richard S Parent A 2003 Chemical anatomy of the human ventral striatum and adjacent basal forebrain structures J Comp Neurol 460 345 367 12692854 96 Puelles L Kuwana E Puelles E Bulfone A Shimamura K Keleher J Smiga S Rubenstein JL 2000 Pallial and subpallial derivatives in the embryonic chick and mouse telencephalon, traced by the expression of the genes Dlx-2, Emx-1, Nkx-2.1, Pax-6, and Tbr-1 J Comp Neurol 424 409 438 10906711 97 Reiner A 1987 The distribution of proenkephalin-derived peptides in the central nervous system of turtles J Comp Neurol 259 65 91 3294930 98 Reiner A Anderson KD 1990 The patterns of neurotransmitter and neuropeptide co-occurrence among striatal projection neurons: conclusions based on recent findings Brain Res Brain Res Rev 15 251 265 1981156 99 Reiner A Anderson KD 1993 Co-occurrence of gamma-aminobutyric acid, parvalbumin and the neurotensin-related neuropeptide LANT6 in pallidal, nigral and striatal neurons in pigeons and monkeys Brain Res 624 317 325 8252409 100 Reiner A Brauth SE Kitt CA Quirion R 1989 Distribution of mu, delta, and kappa opiate receptor types in the forebrain and midbrain of pigeons J Comp Neurol 280 359 382 2537344 101 Reiner A Carraway RE 1987 Immunohistochemical and biochemical studies on Lys8-Asn9-neurotensin8-13 (LANT6)- related peptides in the basal ganglia of pigeons, turtles, and hamsters J Comp Neurol 257 453 476 3558899 102 Reiner A Davis BM Brecha NC Karten HJ 1984a The distribution of enkephalinlike immunoreactivity in the telencephalon of the adult and developing domestic chicken J Comp Neurol 228 245 262 6207214 103 Reiner A Karle EJ Anderson KD Medina L 1994 Catecholaminergic perikarya and fibers in the avian nervous system Smeets WJAJ Reiner A Phylogeny and Development of Catecholamine Systems in the CNS of Vertebrates Cambridge Univ Cambridge 135 181 104 Reiner A Karten HJ 1985 Comparison of olfactory bulb projections in pigeons and turtles Brain Behav Evol 27 11 27 3836730 105 Reiner A Karten HJ Solina AR 1983 Substance P: localization within paleostriatal-tegmental pathways in the pigeon Neuroscience 9 61 85 6192356 106 Reiner A Krause JE Keyser KT Eldred WD McKelvy JF 1984b The distribution of substance P in turtle nervous system: a radioimmunoassay and immunohistochemical study J Comp Neurol 226 50 75 6203942 107 Reiner A Laverghetta AV Meade CA Cuthbertson SL Bottjer SW 2004a An immunohistochemical and pathway tracing study on the striatopallidal organization of area X in the male zebra finch J Comp Neurol 469 239 261 14694537 108 Reiner A Medina L Veenman CL 1998a Structural and functional evolution of the basal ganglia in vertebrates Brain Res Rev 28 235 285 9858740 109 Reiner A Oliver JR 1987 Somatostatin and neuropeptide Y are almost exclusively found in the same neurons in the telencephalon of turtles Brain Res 426 149 156 2891406 110 Reiner A Perera M Paullus R Medina L 1998b Immunohistochemical localization of DARPP32 in striatal projection neurons and striatal interneurons in pigeons J Chem Neuroanat 16 17 33 9924970 111 Reiner A Perkel DJ Bruce LL Butler AB Csillag A Kuenzel W Medina L Paxinos G Shimizu T Striedter G Wild M Ball GF Durand S Gunturkun O Lee DW Mello CV Powers A White SA Hough G Kubikova L Smulders TV Wada K Dugas-Ford J Husband S Yamamoto K Yu J Siang C Jarvis ED Avian Brain Nomenclature Forum 2004b Revised nomenclature for avian telencephalon and some related brainstem nuclei J Comp Neurol 473 377 414 15116397 112 Riedel A Härtig W Seeger G Gärtner U Brauer K Arendt T 2002 Principles of rat subcortical forebrain organization: a study using histological techniques and multiple fluorescence labeling J Chem Neuroanat 23 75 104 11841914 113 Riters LV Erichsen JT Krebs JR Bingman VP 1999 Neurochemical evidence for at least two regional subdivisions within the homing pigeon (Columba livia) caudolateral neostriatum J Comp Neurol 412 469 487 10441234 114 Roberts TF Hall WS Brauth SE 2002 Organization of the avian basal forebrain: chemical anatomy in the parrot (Melopsittacus undulatus) J Comp Neurol 454 383 408 12455005 115 Russchen FT Smeets WJ Hoogland PV 1987 Histochemical identification of pallidal and striatal structures in the lizard Gekko gecko: evidence for compartmentalization J Comp Neurol 256 329 341 2437160 116 Seroogy K Tsuruo Y Hökfelt T Walsh J Fahrenkrug J Emson PC Goldstein M 1988 Further analysis of presence of peptides in dopamine neurons. Cholecystokinin, peptide histidine-isoleucine/vasoactive intestinal polypeptide and substance P in rat supramammillary region and mesencephalon Exp Brain Res 72 523 534 2466680 117 Shimizu T Cox K Karten HJ 1995 Intratelencephalic projections of the visual wulst in pigeons (Columba livia) J Comp Neurol 359 551 572 7499547 118 Singru PS Mazumdar M Sakharkar AM Lechan RM Thim L Clausen TJ Subhedar NK 2007 Immunohistochemical localization of cocaine- and amphetamine-regulated transcript peptide in the brain of the catfish, Clarias batrachus (Linn ) J Comp Neurol 502 215 235 17348009 119 Smeets WJ Hoogland PV Lohman AH 1986a A forebrain atlas of the lizard Gekko gecko J Comp Neurol 254 1 19 3805348 120 Smeets WJ Hoogland PV Voorn P 1986b The distribution of dopamine immunoreactivity in the forebrain and midbrain of the lizard Gekko gecko: an immunohistochemical study with antibodies against dopamine J Comp Neurol 253 46 60 3540035 121 Smeets WJ Jonker AJ Hoogland PV 1987 Distribution of dopamine in the forebrain and midbrain of the red-eared turtle, Pseudemys scripta elegans, reinvestigated using antibodies against dopamine Brain, Behav Evol 30 121 142 3664261 122 Smeets WJ Marin O Gonzalez A 2000 Evolution of the basal ganglia: new perspectives through a comparative approach J Anat 196 Pt 4 501 517 10923983 123 Smith Y Kieval J Couceyro PR Kuhar MJ 1999 CART peptide-immunoreactive neurones in the nucleus accumbens in monkeys: ultrastructural analysis, colocalization studies, and synaptic interactions with dopaminergic afferents J Comp Neurol 407 491 511 10235641 124 Soha J Shimizu T Doupe AJ 1996 Development of the catecholaminergic innervation of the song system of the male zebra finch J Neurobiol 29 473 489 8656212 125 Spiro JE Dalva MB Mooney R 1999 Long-range inhibition within the zebra finch song nucleus RA can coordinate the firing of multiple projection neurons J Neurophysiol 81 3007 3020 10368416 126 Striedter GF Marchant TA Beydler S 1998 The “neostriatum” develops as part of the lateral pallium in birds J Neurosci 18 5839 5849 9671671 127 Subhedar N Barsagade VG Singru PS Thim L Clausen JT 2011 Cocaine- and amphetamine-regulated transcript peptide (CART) in the telencephalon of the catfish, Clarias gariepinus: Distribution and response to Fasting, 2-Deoxy-D-glucose, glucose, insulin, and leptin Treatments J Comp Neurol 519 1281 1300 21452197 128 Sun Z Reiner A 2000 Localization of dopamine D1A and D1B receptor mRNAs in the forebrain and midbrain of the domestic chick J Chem Neuroanat 19 211 224 11036238 129 Sun Z Wang HB Laverghetta A Yamamoto K Reiner A 2005 The distribution and cellular localization of glutamic acid decarboxylase-65 (GAD65) mRNA in the forebrain and midbrain of domestic chick J Chem Neuroanat 29 265 281 15927788 130 Swerdlow NR Koob GF 1987 Dopamine, schizophrenia, mania, and depression: Toward a unified hypothesis of cortico-striatopallido-thalamic function Behav Brain Sci 10 197 245 131 Thim L Nielsen PF Judge ME Andersen AS Diers I Egel-Mitani M Hastrup S 1998 Purification and characterization of a new hypothalamic satiety peptide, cocaine and amphetamine regulated transcript (CART), produced in yeast FEBS Lett 428 263 268 9654146 132 Veenman CL Albin RL Richfield EK Reiner A 1994 Distributions of GABAA, GABAB, and benzodiazepine receptors in the forebrain and midbrain of pigeons J Comp Neurol 344 161 189 8077457 133 Veenman CL Reiner A 1994 The distribution of GABA-containing perikarya, fibers, and terminals in the forebrain and midbrain of pigeons, with particular reference to the basal ganglia and its projection targets J Comp Neurol 339 209 250 8300906 134 Veenman CL Wild JM Reiner A 1995 Organization of the avian “corticostriatal” projection system: a retrograde and anterograde pathway tracing study in pigeons J Comp Neurol 354 87 126 7615877 135 Vicario A Abellán A Desfilis E Medina L 2014 Genetic identification of the central nucleus and other components of the central extended amygdala in chicken during development Front Neuroanat 8 90 25309337 136 Voorn P Gerfen CR Groenewegen HJ 1989 Compartmental organization of the ventral striatum of the rat: immunohistochemical distribution of enkephalin, substance P, dopamine, and calcium-binding protein J Comp Neurol 289 189 201 2478598 137 Wild JM Williams MN 2000 Rostral wulst in passerine birds. I Origin, course, and terminations of an avian pyramidal tract J Comp Neurol 416 429 450 10660876 138 Wynne B Güntürkün O 1995 Dopaminergic innervation of the telencephalon of the pigeon (Columba livia): a study with antibodies against tyrosine hydroxylase and dopamine J Comp Neurol 357 446 464 7673478 139 Yamamoto K Reiner A 2005 Distribution of the limbic system-associated membrane protein (LAMP) in pigeon forebrain and midbrain J Comp Neurol 486 221 242 15844168 140 Yan K Tank YZ Carr CE 2010 Calcium-binding protein immunoreactivity characterizes the auditory system of Gekko gecko J Comp Neurol 518 3409 3426 20589907 141 Young WS 3rd Alheid GF Heimer L 1984 The ventral pallidal projection to the mediodorsal thalamus: a study with fluorescent retrograde tracers and immunohistofluorescence J Neurosci 4 1626 1638 6374062 142 Zaborszky L Alheid GF Beinfeld MC Eiden LE Heimer L Palkovits M 1985 Cholecystokinin innervation of the ventral striatum: a morphological and radioimmunological study Neuroscience 14 427 453 3887206 143 Zahm DS 1999 Functional-anatomical implications of the nucleus accumbens core and shell subterritories Ann NY Acad Sci 877 113 128 10415646 144 Zahm DS Brog JS 1992 On the significance of subterritories in the “accumbens” part of the rat ventral striatum Neuroscience 50 751 767 1448200 145 Zahm DS Heimer L 1993 Specificity in the efferent projections of the nucleus accumbens in the rat: comparison of the rostral pole projection patterns with those of the core and shell J Comp Neurol 327 220 232 8425943

PMC005xxxxxx/PMC5135637.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 100966032 22039 Trends Immunol Trends Immunol. Trends in immunology 1471-4906 1471-4981 27743777 5135637 10.1016/j.it.2016.09.003 NIHMS818257 Article Antigen Presentation in Transplantation Alegre Maria-Luisa MD, PhD 1 Lakkis Fadi G. MD 23 Morelli Adrian E. MD, PhD 2 1 Department of Medicine, The University of Chicago, 924 E. 57th St., JFK-R312, Chicago, IL 60637 2 Thomas E. Starzl Transplantation Institute and the Department of Surgery, Immunology, Biomedical Science Tower, 200 Lothrop Street, Pittsburgh, PA 15261 3 Medicine, University of Pittsburgh School of Medicine, Biomedical Science Tower, 200 Lothrop Street, Pittsburgh, PA 15261 Correspondence: malegre@midway.uchicago.edu (M.L. Alegre) 22 9 2016 12 10 2016 12 2016 01 12 2017 37 12 831843 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. Transplantation of solid organs between genetically distinct individuals leads, in the absence of immunosuppression, to T cell-dependent transplant rejection. Activation of graft-reactive T cells relies on the presentation of transplant-derived antigens (intact donor MHC molecules or processed peptides on host MHC molecules) by mature dendritic cells (DCs). This review will focus on novel insights regarding the steps for maturation and differentiation of DCs that are necessary for productive presentation of transplant antigens to host T cells. These steps include the licensing of DCs by the microbiota, their activation and maturation following recognition of allogeneic non-self and their capture of donor cell exosomes to amplify presentation of transplant antigens. Recent Advances on Innate Immunity in Transplantation Transplantation of organs from cadaveric or living donors can cure end-stage organ failure in transplant recipients. However, with the exception of organ donation between identical twins, transplantation sets up a cascade of inflammatory events leading to recognition of the allograft (see Glossary) by the host’s immune system and to T cell-dependent transplant rejection. To prevent rejection, transplant patients take global immunosuppressive medications lifelong, which can lead to side effects as well as increased susceptibility to infections and malignancies. In order to develop more specific and less toxic therapies, a better understanding of the steps leading to robust activation of alloreactive T cells by innate immune cells is required. To this end, a large body of work has focused on the biology of alloreactive T cells and approaches to delete or suppress them. In recent years, several groups have turned their attention to the antigen-presenting cells (APCs) that initiate the activation of allogeneic T cells, revealing at least 3 new mechanisms by which innate immunity controls the quality and robustness of alloreactive T cell priming (Figure 1). First, the composition of the microbiota within the donor and the host prior to transplantation was shown to tune the capacity of APCs to prime alloreactive T cells and dictate the subsequent kinetics of graft rejection. Second, following transplantation, recognition by recipient mononuclear phagocytes of non-self determinants in the donor graft, encoded by non-MHC genes, was found to promote maturation and differentiation of these cells enhancing subsequent priming of alloreactive T cells. Third, cell-to-cell communication via exosomes was found crucial to the transfer of intact donor MHC molecules from donor dendritic cells (DCs) onto the surface of host DCs for presentation to alloreactive T cells. This review will cover these novel determinants of the quality and potency of alloantigen presentation after solid organ transplantation. Recognition of Allografts: the Basics The potency of the anti-donor response against an allograft depends on the frequency/number of host T cells that recognize alloantigens and their differentiation stage, but also on the degree of activation of innate immune cells that present such alloantigens. Alloreactive T cells include those that recognize intact donor MHC molecules on the surface of donor cells - referred in transplantation as the direct pathway of allorecognition-, and of peptides derived from polymorphic regions of the allogeneic MHC molecules or of non-MHC proteins presented by self-MHC molecules, which is known as the indirect pathway. Directly alloreactive T cells can also recognize, via the semi-direct pathway, donor MHC molecules transferred intact to host APCs [1]. This latter mechanism enables a single DC from the host to present simultaneously or not, (i) donor allopeptides loaded in self MHC molecules to indirect pathway CD4 T cells, and (ii) donor intact MHC molecules to direct pathway CD8 T cells [1] or, potentially, to directly alloreactive CD4 T cells. The semi-direct pathway provides a means by which indirectly alloreactive CD4 T cells cross-regulate positively or negatively the function of direct pathway CD8 T cells that interact with the same DCs [2–4]. The prevailing view is that the direct CD4 pathway is polyclonal and short-lived, since the number of donor passenger leukocytes is limited and decreases with time after transplantation, and that the indirect pathway is oligoclonal and long-lived [5, 6]. In most models, either pathway can by itself trigger acute rejection [7, 8]. The indirect CD4 pathway promotes alloantibody production and chronic rejection [9]. In contrast, the biological relevance of the semi-direct pathway in vivo is just beginning to be elucidated. Besides its role in cross-regulation between indirectly and directly alloreactive T cells, new evidence indicates that the semi-direct pathway may have a key role in allosensitization of direct pathway T cells in those instances where donor passenger leukocytes traffic to the graft-draining lymphoid organs in extremely low numbers (e.g. heart transplantation in mice), or are unable to migrate out of the graft due to excision of lymphatic vessels during surgery and lack of surgical blood vessel anastomoses (e.g. skin allografts)[10, 11]. Under these circumstances, soluble MHC molecules, released by the graft itself or by the relatively few donor passenger leukocytes that reach the lymphoid organs, are acquired by recipient APCs via a mechanism termed cross-dressing for presentation to directly alloreactive T cells via the semi-direct pathway. The frequency of indirect T cells against a given allopeptide is < 1:100,000, a frequency that is similar to that of any nominal peptide presented via the canonical pathway [12]. By contrast, the fraction of directly alloreactive T cells against foreign intact MHC molecules is 100-fold to 1,000 fold higher, since thymic deletion does not select for or against T cells to recognize the hundreds of MHC alleles not expressed by the graft recipient [12]. Thus, the number of unique T cell clones that are alloreactive in a given host depends on the extent of genetic disparities between the donor and the recipient that can be recognized by the host’s T cells, with disparities in MHC alleles being a critical factor in the number of potential alloreactive clones. Following transplantation, both donor and recipient APCs can activate alloreactive T cells and their state of maturation is crucial to the magnitude and quality of the T cell response. DCs are professional APCs with unique ability to present efficiently donor alloantigen to naïve T cells. Different DC subsets have been identified, including conventional, monocyte-derived, and plasmacytoid DCs [13]. Transplanted organs/tissues are populated by tissue-resident conventional DCs, which trigger activation of alloreactive T cells, directly by themselves or indirectly by transferring donor alloantigen to the recipient’s conventional DCs that reside in the graft-draining lymphoid organs [14, 15]. Monocyte-derived DCs, originated from blood monocytes, participate at a later time point by re-presenting donor alloantigen and promoting expansion of effector T cells inside the allograft [16]. Plasmacytoid DCs have been implicated in transplantation tolerance [17]. Microbiota-dependent Licensing of Innate Immune Cells for Priming of Alloimmunity Knowledge of how the microbiota can influence innate immune maturation has increased exponentially in recent years, though fine understanding of mechanisms and implications in various disease settings is still emerging. One clue that the microbiota might be important in modulating the strength of an alloresponse following transplantation came from the clinical observation that organs colonized by microbiota, such as the intestine and the lung, were more susceptible to acute and chronic rejection than organs deemed sterile such as the heart and the kidney (OPTN/SRTR database, December 2012). Innate Licensing Prior to Transplantation In addressing the role of the microbiota in alloimmunity, Lei et al. found that pre-treating both donor and recipient mice with broad-spectrum antibiotics prior to transplantation, or use of germ-free mice devoid of microbiota, resulted in prolongation of minor mismatched skin graft survival [18], suggesting that certain microbial communities promoted transplant rejection. This was ascribed to an ability of the microbiota prior to transplantation to alter signaling pathways in lymph node APCs allowing them, after transplantation, to promote enhanced proliferation and IFN-γ production by alloreactive T cells. Interestingly, reconstitution of germ-free mice with fecal microbial communities from untreated mice had a positive licensing effect on APCs. However, transfer of microbes from mice pre-treated with antibiotics did not have this effect on APCs, despite similar total bacterial loads, suggesting that different bacterial communities could affect subsequent alloreactivity differently. Mechanistically, pre-antibiotic microbiota activated the type I interferon and the NF-κB pathways in host DCs more efficiently than post-antibiotic microbiota [18], although whether all or only specific DC subsets that are affected is not yet clear. Whether changing the pre-transplantation microbiota in host or donors in the clinic could impact transplant outcomes remains to be determined. Analogies with Anti-tumor and Anti-viral Immunity The ability of the microbiota to license APCs for priming of alloreactive T cells is reminiscent of similar positive APC tuning for initiation of anti-tumoral or anti-viral immunity. Indeed, improved anti-tumor immunity in mice colonized by commensal species from the genus Bifidobacterium correlated with increased expression of genes associated with the type I interferon pathway in splenic DCs from tumor-bearing mice [19]. The microbiota was also found to poise lung DCs to produce pro-IL-1β and pro-IL-18, facilitating their migration to draining lymph nodes following intranasal influenza A virus inoculation and promoting anti-viral T cell priming [20]. Similarly, the microbiota enhanced responsiveness of splenic macrophages to type I and type II interferon signaling enabling subsequent control of systemic LCMV or mucosal influenza viral replication [21]. Splenic DCs from germ-free mice were shown to have defects not in the expression of transcription factors such as NF-κB and type-I interferon-driven IRF3, but rather in the ability of these transcription factors to bind chromatin sites, and this correlated with increased histone H3 trimethylation at several inflammatory gene start sites [22]. This suggests that the microbiota may license innate cells by triggering cell-intrinsic epigenetic modifications. Epigenetic changes induced by bacteria have also been described in T cells, by short chain fatty acids produced by fermenting bacteria from the colon, which inhibit histone deacetylases (HDACs), and promote histone acetylation of the FoxP3 locus in T cells and Treg differentiation [23–25], but whether fermenting bacteria can affect graft outcome remains to be studied. There is evidence that the microbiota can also regulate DC function through indirect mechanisms. Outer membrane vesicles (OMVs) released by commensal and pathogenic gram-negative bacteria stimulate DC maturation [26]. These vesicles carry bacterial components (e.g. LPS, DNA, RNA, proteins, enzymes) in a non-replicative way. OMVs can also promote mucosal tolerance. As example, commensal Bacteroides fragilis OMVs deliver immune-regulatory signals to intestinal DCs that promote generation of Tregs in the intestine [27]. Whether this is a mechanism by which the microbiota modulates graft survival needs to be investigated. Potential Microbiota-dependent Innate Signaling after Tissue Injury In addition to licensing innate cells at the steady state, the microbiota may also drive innate signals after tissue injury, as shown by the ability of specific intestinal commensal species, upon intestinal damage induced by dextran sulfate sodium (DSS), to drive the production of interleukin-1β by intestinal Ly6Chigh monocytes [28], a cell subset whose continuous replenishment from a pool of circulating monocytes is also dependent on the microbiota [29]. As tissue injury is a certainty following organ transplantation, which not only involves surgical sectioning of the skin, but also ischemia and reperfusion of the transplanted organ, it is conceivable that similar commensal-dependent signaling may be triggered in innate cells, perhaps helping explain the worse outcome of organs with a high commensal content, though this remains to be formerly addressed. Following transplantation, it is thought that donor DCs need to migrate to secondary lymphoid organs to prime alloreactive T cells, but whether commensal-dependent alteration of DCs facilitates such migration is not known. MyD88- and TRIF-dependent signals may help promote migration of DCs from donor skin grafts into the host’s draining lymph nodes correlating with faster skin graft rejection [30, 31], but it is not clear if these adaptors mediate signals downstream from the microbiota or downstream from damage-associated molecular patterns (DAMPs). However, as mentioned above, microbiota-dependent licensing of innate immune cells can regulate innate cell migration that contributes to initiation of anti-viral immunity [20]. Similarly, lung colonization with Staphylococcus aureus was found to promote TLR2-dependent recruitment and maturation of M2 alveolar macrophages that helped reduce influenza-mediated acute lung injury [32]. Conversely, commensal signals can also prevent migration of select innate immune cell types, as shown in the case of intestinal luminal bacteria-capturing CX3CR1hi mononuclear phagocytes whose migration to mesenteric lymph nodes is enabled in antibiotic-treated mice [33]. Preventing migration in this case is important to avoid developing an immune response to intestinal commensals. Harnessing such property in transplant settings could theoretically be favorable for graft acceptance. Impact of Tissue-specific Microbial Communities The composition of the microbiota is different in different organs, and even in different regions of a given organ or tissue depending on the pH, mucus, surfactant or oil covering, moisture and oxygen content [34]. Moreover, because the microbiota can exert local and distal effects, it is not clear whether all microbial communities throughout the body contribute to (or are sufficient for) global innate poising, and/or whether each compartment exerts location-specific effects. The fact that both donor and recipient mice had to be treated with antibiotics to achieve prolonged survival of minor mismatched skin grafts [18] suggests at least redundant effects of donor and host microbiota. In terms of tissue-specific effects, skin-resident DCs have been shown to sense and respond to alterations in local microbial communities [35] and tune the function of local T cells in a manner dependent on the IL-1R pathway [36]. In clinical transplantation, longitudinal changes in microbial composition of transplanted intestine or transplanted lung have been associated with organ-specific transplant outcome, with a relative reduction in the phylum Firmicutes in the ileal effluent being associated with acute intestinal rejection, and colonization of transplanted lungs with small conidia Aspergillus or lack of recolonization by dominant pre-transplant species such as Pseudomonas in cystic fibrosis patients being associated with chronic lung rejection [37–39]. Thus, whether skin, lung and intestinal microbial communities can tune immune responses to only transplanted skin, lung or intestine, respectively, or can cross-regulate more global alloimmunity remains to be determined. In any case, the rapid reversibility of APC licensing following antibiotic treatment and rapid acquisition of positive tuning of APCs in conventionalized germ-free mice [18] indicates a very dynamic microbial-host dialogue that may be amenable to therapeutic manipulation in settings of transplantation where the timing of initial exposure to alloantigen is known. However, more research is needed to understand the specific effects on innate cells of select microbial species, individually or as part of communities, at different locations. Ultimately, the goal will be to determine whether select species or communities could synergize with immunosuppressive therapies to promote graft acceptance. Non-self Recognition by Innate Immune Cells Whether or not properly poised by the microbiota, APCs need to differentiate from immature to mature following transplantation to be able to prime alloreactive T cells. The role of antigen-presenting DCs in initiating the primary alloimmune response within secondary lymphoid tissue is well established [10, 40, 41]. However, two fundamental questions related to DC maturation and function in alloimmunity have lingered unanswered for some time. The first is the nature of the signal required for the induction of mature DCs and the second is whether DCs continue to play a role in the alloimmune response beyond its initiation phase. These questions have been recently addressed in mouse transplantation models, leading to novel insights into the immunobiology of allograft rejection. How the Innate Immune System Senses the Allograft to Generate Mature DCs The assumption in alloimmunity has been that the release of DAMPs from the graft at the time of transplantation elicits the activation of host DCs, which in turn trigger the adaptive alloimmune response. This model, known as the “danger hypothesis”, has several shortcomings. First is the observation that DAMPs induce the generation of DCs that support T lymphocyte proliferation but fail to drive T cell differentiation into the IFN-γ-producing lymphocytes that dominate in transplant rejection - the reason being that DAMPs are not sufficient to cause DCs to produce IL-12 [42, 43]. Second, the rejection of allografts mismatched with the recipient at more than a single minor histocompatibility antigen occurs promptly in the absence of major innate signaling pathways or cytokines that mediate the actions of DAMPs [30, 44–47]. For example, Myd88 knockout, Myd88/TRIF double knockout, and intereferon-1 (IFN-1) receptor knockout recipients reject MHC-mismatched allografts without significant delay. Third, allografts parked in T lymphocyte-deficient mice are rejected when the host is replenished with T lymphocytes long after resolution of graft damage and of the resulting DAMPs [48–52]. Therefore, additional pathways that generate mature DCs in response to allotransplantation likely exist. One possible mechanism is that DCs or their precursors distinguish between self and allogeneic non-self as they would between self and microbial non-self – a phenomenon referred to as “innate allorecognition” [43]. This possibility has been formally tested in the mouse. Zecher et al showed that injecting allogeneic RAG−/− splenocytes into the ear pinnae of RAG−/− recipients elicits significantly greater swelling and infiltration of the skin with host myeloid cells than injecting syngeneic splenocytes [52]. Depletion and cell transfer experiments established that the response is independent of NK cells and, instead, is mediated by monocytes. By performing heart, kidney, and bone marrow transplants into RAG−/−γc−/− mice, which lack T, B, and innate lymphoid cells including NK cells, Oberbarnscheidt et al provided direct evidence that detection of allogeneic non-self by monocytes is necessary for initiating alloimmunity [43]. In these experiments, allogeneic but not syngeneic grafts elicited persistent differentiation of monocytes to mature DC that express IL-12, stimulate T cell proliferation and IFN-γ production, and precipitate graft rejection. Conversely, host monocyte depletion prevented rejection. Thus, innate allorecognition by monocytes seems to be the key that unlocks the differentiation of monocytes to mature, IL-12-producing DCs after allotransplantation (Figure 1). Although danger clearly amplifies this innate response, it is not sufficient to initiate it. Role of Host DCs in the Allograft: A Continuous Love Affair with the T Cell Host DCs accumulate in the graft rapidly after transplantation, replacing donor-derived DCs, and persist throughout the life of the graft [53–55]. Which host cells the DCs derive from and their function in the alloimmune response, however, have remained unclear. Recently, our groups (Morelli & Lakkis) have shown in mouse heart and kidney allografts that the vast majority of host CD11c+ DCs that replace donor DCs within few days after transplantation have phenotypic and functional features of monocyte-derived DCs (mono-DC), that they originate from Ly6Clo (non-classical) host monocytes, and that they play a key role in allograft rejection [16, 56]. First, DCs in the allograft extend dendrites into the capillary lumina of renal allografts, capture effector T cells rolling along the endothelium, and mediate their transendothelial migration into the graft in a cognate Ag-dependent fashion [56]. Second, host mono-DCs continue to make extended cognate interactions with effector T cells that infiltrated the graft [16]. These interactions arrest alloantigen-specific effectors within the graft parenchyma, increase their proliferation, and reduce their apoptosis. Depletion of graft DCs successfully delays ongoing acute rejection and prevents rejection mediated by the transfer of effector T cells into the host. How intragraft DCs that are of host origin continue to activate and engage directly alloreactive effector T cells, which constitute the majority of effectors during the alloimmune response, is not known but is likely occurring via the semi-direct (cross-dressing) pathway of alloantigen presentation (see next section: “Alloantigen Presentation by Innate Immune Cells”). It is also conceivable that some of the graft DCs migrate back to secondary lymphoid tissues where they activate additional T cells [57], but this possibility remains to be formally tested. Recognition of allogeneic non-self by monocytes, described above, is an important factor behind the differentiation of host monocytes to mono-DCs in the graft. Therefore, host DCs continue to play an essential role in alloimmunity beyond the initiation phase by propagating the effector T cell response within the graft itself, leading to full-blown graft rejection. Moreover, host DC persistence and maturation in the graft is driven by innate allorecognition. Both of the insights, innate allorecognition and extended function of host DCs in the graft, provide opportunities in the future to inhibit alloimmune responses in a cognate and potentially safe manner. The first raises the possibility that blocking the sensing mechanisms or the signaling pathways triggered by the recognition of allogeneic non-self by monocytes would constitute a novel therapeutic modality to prevent acute or chronic allograft rejection. It is known so far that initial sensing of allografts by monocytes is not dependent on MHC mismatch between the donor and recipient [43]. Instead, we (Lakkis et al) are currently exploring the possibility that SIRPα on donor tissues, which is a highly polymorphic molecule not linked to the MHC, is important for this innate allorecognition phenomenon. The second insight provides the prospect that inhibiting recipient monocyte migration to the graft or their differentiation into DCs could interrupt rejection even after T cell priming has already taken place in secondary lymphoid tissues. Indeed, Miller et al have shown that the number of intra-graft DCs directly correlates with the number of T cells that infiltrate the graft [58] and it has been show in tumor settings that intra-tumoral transfer of DCs promotes recruitment of T cells into the tumor and tumor rejection [59]. Preventing monocyte-dependent recruitment of T cells to the allograft is particularly relevant to memory T cells, which represent a substantial fraction of the alloreactive T cell repertoire in humans [60], and can be recalled and propagated by alloantigens in the graft [61]. Alloantigen Presentation by Innate Immune Cells Following their microbiota-dependent poising and allogeneic non-self-dependent maturation, DCs still need to present alloantigen to alloreactive T cells for fullblown alloimmunity to ensue. Upon transplantation, release of DAMPs induced by the ischemia and reperfusion injury incurred by all transplanted organs promotes activation of graft-resident APCs, including conventional DCs. As a result, fully activated donor DCs migrate via blood or lymphatic vessels to secondary lymphoid organs [14]. It has been generally accepted that directly alloreactive naïve T cells recognize donor (intact) MHC molecules present on the surface of donor DCs mobilized to secondary lymphoid organs. Accumulating evidence, however, has challenged this dogma [62–66]. After transplantation of fully H2-mismatched skin or cardiac allografts in mice, donor-derived living DCs are detected at very low numbers in draining lymphoid organs [10, 57]. This is because donor passenger DCs have a limited lifespan and they are recognized as non-self and eliminated by the host NK cells [66, 67]. Indeed, homing of allogeneic DCs to lymph nodes triggers recruitment and activation of host NK cells that eliminate most of the allogeneic DCs [66, 68]. Host CD8 T effector cells recognizing non-self MHC-I molecules also kill allogeneic DCs mobilized to lymph nodes [69]. Besides, most allogeneic DCs that migrated to lymph nodes fail to form stable contacts with directly alloreactive CD4 T cells [66]. Importantly, lymphatic vessels, one of the main routes used by donor DCs to migrate out of the grafts, are sectioned during the transplantation procedure and do not fully reconnect with the recipient lymphatics until approximately one week after surgery [11]. These findings raise the question of how, in certain transplant models, the relatively few donor DCs detected in graft-draining lymphoid organs elicit so efficiently the directly alloreactive T-cell response that leads to acute rejection. This issue is particularly relevant in naïve hosts, in which the percentage of naïve T cells against donor MHC molecules is much lower than in pre-sensitized recipients with anti-donor and cross-reactive T-cell memory. Numerous studies have shown that APCs can acquire MHC:peptide complexes from other leukocytes or endothelial cells [1, 70–73]. The transfer of MHC:peptide complexes and of other cell surface molecules between leukocytes has been dubbed cross-dressing, trogocytosis or cell nibbling. Following heart, kidney, or skin transplantation in mice, a percentage of host APCs in the draining lymphoid organs express donor-derived H2 molecules on the cell surface [10, 74, 75]. Also, the opposite transfer of host H2 molecules to donor DCs has been detected after bone marrow transplantation [76]. Transfer of donor MHC molecules, from donor passenger APCs or from the graft itself, to recipient APCs not only explains the semi-direct pathway, but also the efficient recognition of non-self MHC molecules by naïve T cells in the presence of extremely few donor passenger APCs, as 1 donor APC may transfer intact MHC molecules onto several host APCs. Recent evidence indicates that recipient APCs acquire donor intact MHC molecules through uptake of donor-derived extracellular vesicles (EVs) [10, 11]. Role of Extracellular Vesicles in Transplant Rejection Leukocytes, including DCs, communicate through exchange of EVs, which differ in biogenesis, composition, and size. Microvesicles and apoptotic cell blebs are between 0.2–1 µm in size and are shed from the surface membrane of living and dying cells, respectively [77]. In contrast, exosomes are EVs of endocytic origin, between 70–120 nm in diameter. Exosomes are generated as intraluminal vesicles by reverse budding of the limiting membrane of late endosomes termed multi-vesicular bodies [77, 78]. Once the limiting membrane of the multivesicular bodies fuses with the plasma membrane, it releases to the intercellular space or bodily fluids its cargo of intraluminal vesicles, which are then termed exosomes. Most studies on transfer of MHC molecules between leukocytes have focused on exosomes whereas the role of other types of EVs in the passage of functional MHC antigen between cells is very limited. The protein composition of the exosome membrane and its intraluminal cargo depends on the lineage and state of activation, transformation, or infection of the parent cell. Exosomes carry cell-specific antigens, heat shock proteins, MHC and cell adhesion molecules, cytokines, proteinases, cytoplasmic enzymes, signal transduction molecules, and molecules involved in exosome biogenesis. Exosomes also contain mRNAs, non-coding regulatory RNAs including microRNAs (miRNAs), and in some cases extra-chromosomal DNA [78]. Exosomes even have cell-independent ability to process precursor miRNAs into mature miRNAs [79]. The current view is that exosomes function as a device for horizontal dissemination of proteins, lipids, and RNAs [80]. Exosomes interact with cells via multiple ligand-receptor interactions. Exosomes may remain on the surface or be rapidly internalized by the target cell. Indirect evidence suggests that endocytosed exosomes fuse with the membrane of the endocytic vacuoles releasing their intraluminal cargo into the cytosol of the acceptor cells [81]. Thus, exosomes may “inject” their intraluminal cargo into the cytosol of the acceptor leukocytes, where the exosome-shuttled miRNAs regulate their target mRNAs [81, 82], and the exosome-derived mRNAs translate into proteins [83]. Exosomes released by activated DCs are enriched in donor MHC molecules, T cell costimulatory molecules, and adhesion molecules on the vesicle surface [84, 85]. Although at high concentration APC-derived exosomes can function as antigen-presenting vesicles for T cell clones, lines, and primed T cells, their ability to stimulate naïve T cell is limited [78]. Such capacity, nevertheless, increases substantially when the exosomes are bound to DCs [85]. After transplantation of fully H2-disparate heart or skin grafts in non-allosensitized mice, the relatively few donor migrating DCs that reach the lymphoid organs, and likely the graft itself, transfer clusters of exosomes carrying donor H2 molecules and APC-activating signals to a high number of host APCs (Figure 2) [10, 11]. Indeed, the quantity of recipient DCs bearing donor H2 molecules was approximately 100-fold higher that the number of donor passenger DCs detected in the graft-draining lymphoid organ at the same time. The clusters of donor-derived exosomes remained attached to the host DCs for presentation of donor H2 molecules to direct pathway CD8 T cells via the semi-direct pathway, or were internalized for processing and presentation of the resulting donor H2-derived allopeptides to indirect pathway CD4 T cells (Figure 2) [10, 11]. Importantly, uptake of exosomes released by fully-activated donor DCs promoted activation of the host DCs, perhaps because exosomes from activated DCs bear high content of heat shock proteins and mir-155 that are potent inducers of DC activation [81, 86]. Interestingly, exosomes released by mast cells, erythrocytes, or infected macrophages also promote DC activation [87–89]. In accordance to a role of the recipient DCs in presentation of donor MHC molecules acquired via donor-derived exosomes, depletion of host DCs after heart transplantation reduced presentation of donor H2 molecules to directly alloreactive T cells, and delayed acute rejection [10]. These findings unveil a dual role for the passage donor-derived exosomes in graft rejection, first, as the elusive ultrastructural basis behind the semi-direct pathway, and second, as a mechanism of amplification for presentation of non-self MHC molecules to T cells in lymphoid organs during allosensitization. Exosome-like EVs have also been shown to induce auto-antibodies involved in allograft vascular rejection. Serum-deprived endothelial cells release, besides apoptotic cell-derived vesicles, exosome-like EVs enriched in the LG3 fragment of the vascular extracellular matrix protein perlecam [90]. Repetitive injection of these endothelial-derived exosome-like EVs, unlike infusion of apoptotic cell vesicles from the same parent cells, triggers production of antibodies against the auto-antigen LG3 and aggravate vascular rejection in mice [90]. Extracellular Vesicles as Therapeutic Tools and Biomarkers in Transplantation Donor-derived exosomes may facilitate recognition of the allograft as non-self, but exosomes have also been shown to delay allograft rejection in some settings. Systemic administration of donor-derived exosomes produced by immature DCs plus suboptimal pharmacologic immunosuppression decreased the anti-donor response and prolonged survival of cardiac and intestinal allografts in murine models [91–93]. In such studies, however, whether the donor-specific immunosuppression was due to intrinsic properties of the donor-derived EVs, or simply to a donor-specific transfusion effect caused by delivering donor antigen through a pro-tolerogenic route (i.v.) was not explored. There is evidence that some of the immunosuppressive effects of mesenchymal stem cells are mediated by secreted factors, including EVs. Repetitive administration of exosome-enriched fractions from culture supernatants of mesenchymal stem cells, bearing high amounts of IL-10, TFG-β and HLA-G, improved the symptoms in one patient with refractory graft-versus-host disease [94]. Although these studies provide evidence for the potential use of immunosuppressive exosomes or other EVs in the clinic, the mechanisms by which certain types of exosomes regulate the immune response remain mostly unknown. Interestingly, exosomes released by CD4 Foxp3 regulatory T (Treg) cells express CD73, an ecto-enzyme that generates the immune-suppressive nucleoside adenosine [95]. Such exosomes suppressed proliferation and IL-2 and IFN-γ secretion in CD4 T cells in vitro [95]. Through exosome-mediated transfer of Let-7d miRNA, CD4 Foxp3 Treg cells reduce Th1 proliferation and IFN-γ secretion, and suppress pathogenic Th1 cell-mediated systemic disease in mice [96]. In a transplantation model, repetitive i.v. administration of exosomes purified from cultures of donor or recipient polyclonal CD4 Treg cells prolonged survival of kidney allografts in rats [97]. Although these findings are promising for future exosome-based immunosuppressive therapies, the extreme dilution in plasma of the injected EVs and their rapid clearance from circulation by phagocytes are critical hurdles for their clinical applicability. Exosomes isolated from bodily fluids represent potential protein and RNA biomarkers (Figure 2). Urine EVs, including exosomes, carry proteins derived from glomerular, tubular, prostate and bladder cells, and have been investigated as biomarkers for various kidney diseases [98]. The content of exosomes in urine augments throughout the first week after kidney transplantation [99]. In patients with delayed graft function, the levels in urine of exosome-associated neutrophil gelatinase-associated lipocalin (NGAL) are higher than those in recipients without such complication [100]. In contrast, the amounts of mRNAs for biomarkers of renal cellular injury found in urine exosomes correlate poorly with the clinical outcome of renal transplantation [99]. Exosomes isolated from sera and bronchoalveolar lavage fluid of patients undergoing acute or chronic rejection of lung allografts carry high content of collagen-V, a lung-associated self-antigen involved in lung allograft rejection [101]. Importantly, exosomes bearing high levels of collagen-V were detectable before clinical diagnosis of lung rejection [101], suggesting they could be used as predictive biomarkers. Concluding Remarks Innate immune cells are essential to activate alloreactive T cells following transplantation but it is now apparent that their activating capacity is modulated not only after, but also already before transplantation. Licensing of innate immune cells by the microbiota before transplantation and their activation by allogeneic non-self molecules expressed on donor cells after transplantation establish their activation threshold and differentiation status for subsequent alloreactive T cell priming. Concurrently, capture of donor exosomes and alloantigen processing by host DCs provide the epitopes necessary for initiation of adaptive alloimmunity. Many outstanding questions remain to be investigated (see Outstanding Questions). In particular, the role of the microbiota after transplantation and the precise mechanisms linking commensal microbes outside body barriers to modulation of alloimmunity inside the body need to be clarified. In addition, the allogeneic non-self determinants that need to be recognized by innate cells to promote their maturation and whether they can be masked to promote graft acceptance needs to be further pursued. Finally, whether release and capture of exosomes can be manipulated therapeutically to reduce allorecognition should be investigated. Resolving these and other questions will bring us one step closer to improving transplant outcome in the clinic. This work was supported by NIH RO1s AI115716 to M.-L.A, AI049466 and AI099465 to F.G.L. and HL130191 to A.E.M. Glossary Box Alloantigen polymorphic molecule for which a variant expressed by donor cells can be recognized by the adaptive immune system Allograft an organ transplanted between genetically distinct individuals of the same species Allopeptide a peptide derived from a polymorphic region of an alloantigen Alloreactive ability to recognize an alloantigen Allorecognition sensing of allogeneic non-self by either innate or adaptive immune cells Direct pathway presentation of intact donor MHC molecule by a donor cell to a T cell Exosomes nanovesicles generated in the endocytic compartment of most cell types, and released to the extracellular space or bodily fluids by fusion of multivesicular bodies with the cell membrane Extracellular vesicles generic name given to all membrane vesicles derived from cells, including microvesicles shed from the cell surface, exosomes released from the endocytic compartment and apoptotic cell-derived vesicles Indirect pathway presentation of an allopeptide by host MHC molecules to a T cell Microbiota the combination of microbes that colonize a habitat Minor mismatch a polymorphism in a non-MHC-encoded gene that can give rise to an allopeptide Semi-direct pathway presentation by host APCs of intact captured donor MHC molecules Figure 1 Tuning of Antigen-Presenting Cells (APCs) Before and After Transplantation for Optimal Priming of Alloreactive T Cells Donor cells are represented in red and host cells in blue. The donor and host microbiota license innate immune cells at the steady state for subsequent immune responses, including alloimmunity. Following transplantation, recognition by host dendritic cells (DCs) of both danger signals and allogeneic non-self determinants promotes their maturation. Danger encompasses a variety of molecules released from the damaged graft tissue as a result of ischemia reperfusion injury at the time of transplantation. Allogeneic non-self could consist of yet to be discovered allodeterminants expressed on donor tissue (including potentially both hematopoietic and parenchymal cells) that are either related or unrelated to the MHC. These allogeneic signals appear critical in particular for the induction of IL-12. Capture of donor exosomes and likely of other types of extracellular vesicles (EVs) boosts their acquisition of alloantigen for semi-direct or indirect presentation to alloreactive T cells, ultimately leading to type I adaptive immunity and rejection. Shaded boxes are described in more detail in corresponding sections of this review. Figure 2 Role of Extracellular Vesicles in Transplantation The donor is represented in red and the recipient in blue. After transplantation, donor-derived passenger antigen-presenting cells (APCs) (i.e. dendritic cells, DCs) migrate to the graft-draining lymphoid tissues, where they present donor MHC molecules to directly alloreactive T cells (i.e. direct pathway), or transfer clusters of donor exosomes (and likely other extracellular vesicles, EVs) bearing donor MHC molecules to recipient APCs. Transfer of clusters of donor exosomes bearing MHC molecules and APC-activating signals from relatively few donor migrating DCs to a higher number of recipient APCs amplifies T-cell allosensitization, and is the basis of the semi-direct pathway of allorecognition by T cells. Alternatively, EVs carrying donor MHC molecules and released by the allograft itself into blood or lymph could be a source of donor MHC molecules for recipients APCs residing in graft-draining lymphoid tissues. Donor-derived exosomes and EVs derived from donor migrating APCs undergoing apoptosis are also internalized for processing into donor-derived peptides for presentation to T cells through the indirect pathway of allorecognition. EVs released by the allograft or by graft-infiltrating leukocytes into systemic blood or bodily fluids (e.g. urine) constitute promising protein or RNA biomarkers. Outstanding Questions Box What is the fate of effector T cells that make extended, cognate contacts with DCs in the graft? Are they more likely to become potent, short-lived effectors or long-lived memory cells? Role of the microbiota in transplantation What is the impact of the microbiota after transplantation? Do different commensal species tune innate immune cells differently and how does assembly of microbial communities change these effects? What are the molecular determinants by which the microbiota tunes innate immune cells? Are there immunosuppressive commensals that could promote graft acceptance? Do organ-specific commensals tune innate immune cells locally or systemically? Role of the microbiota in transplantation What are the molecular determinants and mechanisms of allogeneic non-self recognition by innate immune cells? Extracellular vesicles and transplantation Does organ colonization alter the rate, number or quality of exosomes and other EVs produced by donor DCs? Are donor-derived EVs released mainly by the relatively few donor passenger APCs that reach the graft-draining lymphoid tissues, by the graft itself, or by both, and if released by the graft how do they avoid extreme dilution in the systemic circulation and rapid clearance by phagocytes? What is the nature of the APC-activating signals delivered by donor exosomes, and other EVs? Will future methodologies for the use of EVs as biomarkers be able to isolate EVs released exclusively by the graft or graft-infiltrating cells, and distinguish them from “background” EVs produced by host tissues? What is the effect of immunosuppressive therapies on the release and capture of exosomes and other EVs, and can capture of donor EVs by host DCs be blocked to limit presentation of alloantigens? Trends Box The composition of the microbiota tunes innate immune cells to establish their threshold of activation of adaptive alloimmunity Non-self recognition of mammalian determinants encoded by non-MHC genes in allogeneic grafts initiates activation and differentiation of host mononuclear phagocytes Capture of donor exosomes by host DCs amplifies activation of direct and indirect pathways T cells and is a major pathway for priming adaptive alloimmunity This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. References 1 Herrera OB A novel pathway of alloantigen presentation by dendritic cells J Immunol 2004 173 8 4828 4837 15470023 2 Lee RS Indirect recognition by helper cells can induce donor-specific cytotoxic T lymphocytes in vivo J Exp Med 1994 179 3 865 872 8113680 3 Wise MP Linked suppression of skin graft rejection can operate through indirect recognition J Immunol 1998 161 11 5813 5816 9834057 4 Sivaganesh S Copresentation of intact and processed MHC alloantigen by recipient dendritic cells enables delivery of linked help to alloreactive CD8 T cells by indirect-pathway CD4 T cells J Immunol 2013 190 11 5829 5838 Epub 2013 Apr 29 23630361 5 Suchin EJ Quantifying the frequency of alloreactive T cells in vivo: new answers to an old question J Immunol 2001 166 2 973 981 11145675 6 Benichou G Limited T cell response to donor MHC peptides during allograft rejection. Implications for selective immune therapy in transplantation J Immunol 1994 153 3 938 945 7517977 7 Auchincloss H Jr The role of ‘indirect’ recognition in initiating rejection of skin grafts from major histocompatibility complex class II-deficient mice Proc.Natl.Acad.Sci.USA 1993 90 3373 3377 8475083 8 Illigens BM The relative contribution of direct and indirect antigen recognition pathways to the alloresponse and graft rejection depends upon the nature of the transplant Hum Immunol 2002 63 10 912 925 12368044 9 Baker RJ Loss of direct and maintenance of indirect alloresponses in renal allograft recipients: implications for the pathogenesis of chronic allograft nephropathy J Immunol 2001 167 12 7199 7206 11739543 10 Liu Q Donor dendritic cell-derived exosomes promote allograft-targeting immune response J Clin Invest 2016 126 8 2805 2820 Epub 2016 Jun 27 27348586 11 Marino JM Donor exosomes rather than passenger leukocytes initiate alloreactive T cell responses after transplantation Science Immunology 2016 1 1 aaf8759 12 Felix NJ Allen PM Specificity of T-cell alloreactivity Nat Rev Immunol 2007 7 12 942 953 18007679 13 Geissmann F Development of monocytes, macrophages, and dendritic cells Science 2010 327 5966 656 661 20133564 14 Morelli AE Dendritic cells of myeloid lineage: the masterminds behind acute allograft rejection Curr Opin Organ Transplant 2014 19 1 20 27 24316759 15 Ochando J The Mononuclear Phagocyte System in Organ Transplantation Am J Transplant 2016 16 4 1053 1069 Epub 2016 Jan 28 26602545 16 Zhuang Q Graft-infiltrating host dendritic cells play a key role in organ transplant rejection Nat Commun 2016 7 12623 27554168 17 Ochando JC Alloantigen-presenting plasmacytoid dendritic cells mediate tolerance to vascularized grafts Nat Immunol 2006 7 6 652 662 16633346 18 Lei YM The composition of the microbiota modulates allograft rejection J Clin Invest 2016 126 7 2736 2744 Epub 2016 Jun 20 27322054 19 Sivan A Commensal Bifidobacterium promotes antitumor immunity and facilitates anti-PD-L1 efficacy Science 2015 350 6264 1084 1089 Epub 2015 Nov 5 26541606 20 Ichinohe T Microbiota regulates immune defense against respiratory tract influenza A virus infection Proc Natl Acad Sci U S A 2011 108 13 5354 5359 Epub 2011 Mar 14 21402903 21 Abt MC Commensal bacteria calibrate the activation threshold of innate antiviral immunity Immunity 2012 37 1 158 170 Epub 2012 Jun 14 22705104 22 Ganal SC Priming of natural killer cells by nonmucosal mononuclear phagocytes requires instructive signals from commensal microbiota Immunity 2012 37 1 171 186 Epub 2012 Jun 28 22749822 23 Smith PM The microbial metabolites, short-chain fatty acids, regulate colonic Treg cell homeostasis Science 2013 341 6145 569 573 Epub 2013 Jul 4 23828891 24 Furusawa Y Commensal microbe-derived butyrate induces the differentiation of colonic regulatory T cells Nature 2013 504 7480 446 450 Epub 2013 Nov 13 24226770 25 Arpaia N Metabolites produced by commensal bacteria promote peripheral regulatory T-cell generation Nature 2013 504 7480 451 455 Epub 2013 Nov 13 24226773 26 Kaparakis-Liaskos M Ferrero RL Immune modulation by bacterial outer membrane vesicles Nat Rev Immunol 2015 15 6 375 387 Epub 2015 May 15 25976515 27 Chu H Gene-microbiota interactions contribute to the pathogenesis of inflammatory bowel disease Science 2016 352 6289 1116 1120 Epub 2016 May 5 27230380 28 Seo SU Distinct Commensals Induce Interleukin-1beta via NLRP3 Inflammasome in Inflammatory Monocytes to Promote Intestinal Inflammation in Response to Injury Immunity 2015 42 4 744 755 Epub 2015 Apr 7 25862092 29 Bain CC Constant replenishment from circulating monocytes maintains the macrophage pool in the intestine of adult mice Nat Immunol 2014 15 10 929 937 Epub 2014 Aug 24 25151491 30 McKay D Simultaneous deletion of MyD88 and Trif delays major histocompatibility and minor antigen mismatch allograft rejection Eur J Immunol 2006 36 8 1994 2002 16874736 31 Goldstein DR Critical role of the Toll-like receptor signal adaptor protein MyD88 in acute allograft rejection J Clin Invest 2003 111 10 1571 1578 12750407 32 Wang J Bacterial colonization dampens influenza-mediated acute lung injury via induction of M2 alveolar macrophages Nat Commun 2106 4 33 Diehl GE Microbiota restricts trafficking of bacteria to mesenteric lymph nodes by CX(3)CR1(hi) cells Nature 2013 494 7435 116 120 Epub 2013 Jan 13 23334413 34 Findley K Topographic diversity of fungal and bacterial communities in human skin Nature 2013 498 7454 367 370 Epub 2013 May 22 23698366 35 Naik S Commensal-dendritic-cell interaction specifies a unique protective skin immune signature Nature 2015 520 7545 104 108 Epub 2015 Jan 5 25539086 36 Naik S Compartmentalized control of skin immunity by resident commensals Science 2012 337 6098 1115 1119 Epub 2012 Jul 26 22837383 37 Oh PL Characterization of the ileal microbiota in rejecting and nonrejecting recipients of small bowel transplants Am J Transplant 2012 12 3 753 762 Epub 2011 Dec 7 22152019 38 Weigt SS Colonization with small conidia Aspergillus species is associated with bronchiolitis obliterans syndrome: a two-center validation study Am J Transplant 2013 13 4 919 927 Epub 2013 Feb 7 23398785 39 Willner DL Reestablishment of Recipient-associated Microbiota in the Lung Allograft Is Linked to Reduced Risk of Bronchiolitis Obliterans Syndrome Am J Respir Crit Care Med 2013 187 6 640 647 Epub 2013 Jan 17 23328523 40 Lechler RI Batchelor JR Restoration of immunogenicity to passenger cell-depleted kidney allografts by the addition of donor strain dendritic cells J Exp Med 1982 155 1 31 41 7033437 41 Lakkis FG Immunologic ‘ignorance’ of vascularized organ transplants in the absence of secondary lymphoid tissue Nat Med 2000 6 6 686 688 10835686 42 Sporri R Reis e Sousa C Inflammatory mediators are insufficient for full dendritic cell activation and promote expansion of CD4+ T cell populations lacking helper function Nat Immunol 2005 6 2 163 170 Epub 2005 Jan 16 15654341 43 Oberbarnscheidt MH Non-self recognition by monocytes initiates allograft rejection J Clin Invest 2014 124 8 3579 3589 Epub 2014 Jul 1 24983319 44 Tesar BM TH1 immune responses to fully MHC mismatched allografts are diminished in the absence of MyD88, a toll-like receptor signal adaptor protein Am J Transplant 2004 4 9 1429 1439 15307830 45 Hutton MJ Islet allograft rejection is independent of toll-like receptor signaling in mice Transplantation 2009 88 9 1075 1080 19898202 46 Oberbarnscheidt MH Type I interferons are not critical for skin allograft rejection or the generation of donor-specific CD8+ memory T cells Am J Transplant 2010 10 1 162 167 Epub 2009 Nov 24 19951284 47 Li H Graft-versus-host disease is independent of innate signaling pathways triggered by pathogens in host hematopoietic cells J Immunol 2011 186 1 230 241 Epub 2010 Nov 22 21098219 48 Bingaman AW Vigorous allograft rejection in the absence of danger J Immunol 2000 164 6 3065 3071 10706695 49 Anderson CC Testing time-, ignorance-, and danger-based models of tolerance J Immunol 2001 166 6 3663 3671 11238605 50 Anderson CC Matzinger P Immunity or tolerance: opposite outcomes of microchimerism from skin grafts Nat Med 2001 7 1 80 87 11135620 51 Chan WF The ability of natural tolerance to be applied to allogeneic tissue: determinants and limits Biol Direct 2007 2 10 17437644 52 Zecher D An innate response to allogeneic nonself mediated by monocytes J Immunol 2009 183 12 7810 7816 Epub 19923456 53 Grau V Dynamics of monocytes/macrophages and T lymphocytes in acutely rejecting rat renal allografts Cell Tissue Res 1998 291 1 117 126 9394049 54 Penfield JG Transplant surgery injury recruits recipient MHC class II-positive leukocytes into the kidney Kidney Int 1999 56 5 1759 1769 10571784 55 Saiki T Trafficking of host- and donor-derived dendritic cells in rat cardiac transplantation: allosensitization in the spleen and hepatic nodes Transplantation 2001 71 12 1806 1815 11455262 56 Walch JM Cognate antigen directs CD8+ T cell migration to vascularized transplants J Clin Invest 2013 123 6 2663 2671 Epub 2013 May 15 23676459 57 Celli S Visualizing the innate and adaptive immune responses underlying allograft rejection by two-photon microscopy Nat Med 2011 17 6 744 749 Epub 2011 May 15 21572426 58 Miller ML Adoptive transfer of tracer alloreactive CD4+ TCR-transgenic T cells alters the endogenous immune response to an allograft Am J Transplant 2016 13821 59 Spranger S Melanoma-intrinsic beta-catenin signalling prevents anti-tumour immunity Nature 2015 523 7559 231 235 Epub 2015 May 11 25970248 60 Macedo C Contribution of naive and memory T-cell populations to the human alloimmune response Am J Transplant 2009 9 9 2057 2066 Epub 2009 Jul 16 19624567 61 Chalasani G Recall and propagation of allospecific memory T cells independent of secondary lymphoid organs Proc Natl Acad Sci U S A 2002 99 9 6175 6180 11983909 62 Chen Y The male minor transplantation antigen preferentially activates recipient CD4+ T cells through the indirect presentation pathway in vivo J Immunol 2003 171 12 6510 6518 14662851 63 Brennan TV Preferential priming of alloreactive T cells with indirect reactivity Am J Transplant 2009 9 4 709 718 19344462 64 Zheng XX The role of donor and recipient B7-1 (CD80) in allograft rejection J Immunol 1997 159 3 1169 1173 9233610 65 Mandelbrot DA Expression of B7 molecules in recipient, not donor, mice determines the survival of cardiac allografts J Immunol 1999 163 7 3753 3757 10490971 66 Garrod KR NK cell patrolling and elimination of donor-derived dendritic cells favor indirect alloreactivity J Immunol 2010 184 5 2329 2336 Epub 2010 Feb 5 20139277 67 Yu G NK cells promote transplant tolerance by killing donor antigen-presenting cells J Exp Med 2006 203 8 1851 1858 Epub 2006 Jul 24 16864660 68 Laffont S Natural killer cells recruited into lymph nodes inhibit alloreactive T-cell activation through perforin-mediated killing of donor allogeneic dendritic cells Blood 2008 112 3 661 671 Epub 2008 May 27 18505782 69 Laffont S CD8+ T-cell-mediated killing of donor dendritic cells prevents alloreactive T helper type-2 responses in vivo Blood 2006 108 7 2257 2264 Epub 2006 Jan 31 16449531 70 Wakim LM Bevan MJ Cross-dressed dendritic cells drive memory CD8+ T-cell activation after viral infection Nature 2011 471 7340 629 632 21455179 71 Smyth LA Acquisition of MHC:peptide complexes by dendritic cells contributes to the generation of antiviral CD8+ T cell immunity in vivo J Immunol 2012 189 5 2274 2282 Epub 2012 Jul 20 22821960 72 Zhang QJ Trogocytosis of MHC-I/peptide complexes derived from tumors and infected cells enhances dendritic cell cross-priming and promotes adaptive T cell responses PLoS One 2008 3 8 e3097 18769733 73 Harshyne LA Dendritic cells acquire antigens from live cells for cross-presentation to CTL J Immunol 2001 166 6 3717 3723 11238612 74 Brown K Extensive and bidirectional transfer of major histocompatibility complex class II molecules between donor and recipient cells in vivo following solid organ transplantation FASEB J 2008 22 11 3776 3784 Epub 2008 Jul 16 18632850 75 Brown K Coexpression of donor peptide/recipient MHC complex and intact donor MHC: evidence for a link between the direct and indirect pathways Am J Transplant 2011 11 4 826 831 Epub 2011 Mar 14 21401861 76 Markey KA Cross-dressing by donor dendritic cells after allogeneic bone marrow transplantation contributes to formation of the immunological synapse and maximizes responses to indirectly presented antigen J Immunol 2014 192 11 5426 5433 Epub 2014 Apr 30 24790149 77 Colombo M Biogenesis, secretion, and intercellular interactions of exosomes and other extracellular vesicles Annu Rev Cell Dev Biol 2014 30 255 289 Epub 2014 Aug 21 25288114 78 Robbins PD Morelli AE Regulation of immune responses by extracellular vesicles Nat Rev Immunol 2014 14 3 195 208 24566916 79 Melo SA Cancer exosomes perform cell-independent microRNA biogenesis and promote tumorigenesis Cancer Cell 2014 26 5 707 721 Epub 2014 Oct 23 25446899 80 Tkach M Thery C Communication by Extracellular Vesicles: Where We Are and Where We Need to Go Cell 2016 164 6 1226 1232 26967288 81 Montecalvo A Mechanism of transfer of functional microRNAs between mouse dendritic cells via exosomes Blood 2012 119 3 756 766 Epub 2011 Oct 26 22031862 82 Mittelbrunn M Unidirectional transfer of microRNA-loaded exosomes from T cells to antigen-presenting cells Nat Commun 2011 2 282 21505438 83 Valadi H Exosome-mediated transfer of mRNAs and microRNAs is a novel mechanism of genetic exchange between cells Nat Cell Biol 2007 9 6 654 659 Epub 2007 May 7 17486113 84 Segura E ICAM-1 on exosomes from mature dendritic cells is critical for efficient naive T-cell priming Blood 2005 106 1 216 223 Epub 2005 Mar 24 15790784 85 Montecalvo A Exosomes as a short-range mechanism to spread alloantigen between dendritic cells during T cell allorecognition J Immunol 2008 180 5 3081 3090 18292531 86 Thery C Molecular characterization of dendritic cell-derived exosomes. Selective accumulation of the heat shock protein hsc73 J Cell Biol 1999 147 3 599 610 10545503 87 Skokos D Mast cell-derived exosomes induce phenotypic and functional maturation of dendritic cells and elicit specific immune responses in vivo J Immunol 2003 170 6 3037 3045 12626558 88 Danesh A Exosomes from red blood cell units bind to monocytes and induce proinflammatory cytokines, boosting T-cell responses in vitro Blood 2014 123 5 687 696 Epub 2013 Dec 12 24335232 89 Bhatnagar S Exosomes released from macrophages infected with intracellular pathogens stimulate a proinflammatory response in vitro and in vivo Blood 2007 110 9 3234 3244 Epub 2007 Jul 31 17666571 90 Dieude M The 20S proteasome core, active within apoptotic exosome-like vesicles, induces autoantibody production and accelerates rejection Sci Transl Med 2015 7 318 318ra200 91 Peche H Induction of tolerance by exosomes and short-term immunosuppression in a fully MHC-mismatched rat cardiac allograft model Am J Transplant 2006 6 7 1541 1550 16827854 92 Li X Tolerance induction by exosomes from immature dendritic cells and rapamycin in a mouse cardiac allograft model PLoS One 2012 7 8 e44045 Epub 2012 Aug 29 22952868 93 Yang X Exosomes derived from immature bone marrow dendritic cells induce tolerogenicity of intestinal transplantation in rats J Surg Res 2011 171 2 826 832 Epub 2010 Jun 8 20828738 94 Kordelas L MSC-derived exosomes: a novel tool to treat therapy-refractory graft-versus-host disease Leukemia 2014 28 4 970 973 24445866 95 Smyth LA CD73 expression on extracellular vesicles derived from CD4+ CD25+ Foxp3+ T cells contributes to their regulatory function Eur J Immunol 2013 43 9 2430 2440 Epub 2013 Jul 15 23749427 96 Okoye IS MicroRNA-containing T-regulatory-cell-derived exosomes suppress pathogenic T helper 1 cells Immunity 2014 41 1 89 103 25035954 97 Yu X CD4+CD25+ regulatory T cells-derived exosomes prolonged kidney allograft survival in a rat model Cell Immunol 2013 285 1–2 62 68 Epub 2013 Jun 28 24095986 98 Salih M Urinary extracellular vesicles and the kidney: biomarkers and beyond Am J Physiol Renal Physiol 2014 306 11 F1251 F1259 Epub 2014 Apr 2 24694589 99 Peake PW A comparison of the ability of levels of urinary biomarker proteins and exosomal mRNA to predict outcomes after renal transplantation PLoS One 2014 9 2 e98644 eCollection 2014 24918752 100 Alvarez S Urinary exosomes as a source of kidney dysfunction biomarker in renal transplantation Transplant Proc 2013 45 10 3719 3723 24315007 101 Gunasekaran M Donor-Derived Exosomes With Lung Self-Antigens in Human Lung Allograft Rejection Am J Transplant 2016 13915

PMC005xxxxxx/PMC5135656.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 2984705R 2786 Cancer Res Cancer Res. Cancer research 0008-5472 1538-7445 27758883 5135656 10.1158/0008-5472.CAN-16-0715 NIHMS827037 Article Oncogenic functions of Gli in pancreatic adenocarcinoma are supported by its PRMT1-mediated methylation Wang Yan 111† Hsu Jung-Mao 1† Kang Ya’an 2 Wei Yongkun 1 Lee Pei-Chih 1 Chang Shing-Jyh 13 Hsu Yi-Hsin 1 Hsu Jennifer L. 1 Wang Hung-Ling 4 Chang Wei-Chao 45 Li Chia-Wei 1 Liao Hsin-Wei 16 Chang Shih-Shin 16 Xia Weiya 1 Ko How-Wen 16 Chou Chao-Kai 1 Fleming Jason B. 2 Wang Huamin 7 Hwang Rosa F. 8 Chen Yue 9 Qin Jun 9 Hung Mien-Chie 14610* 1 Department of Molecular and Cellular Oncology, The University of Texas MD Anderson Cancer Center, Houston, Texas, 77030, USA 2 Department of Surgical Oncology, The University of Texas MD Anderson Cancer Center, Houston, TX, 77030, USA 3 Department of Obstetrics and Gynecology, Hsinchu Mackay Memorial Hospital, Hsinchu 300, Taiwan 4 Center for Molecular Medicine and Graduate Institute of Cancer Biology, China Medical University, Taichung 404, Taiwan 5 Genomics Research Center, Academia Sinica, Taipei 115, Taiwan 6 The University of Texas Graduate School of Biomedical Sciences at Houston, Houston, Texas, 77030, USA 7 Department of Pathology, The University of Texas MD Anderson Cancer Center, Houston, TX 77030, USA 8 Department of Breast Surgical Oncology, The University of Texas MD Anderson Cancer Center, Houston, TX, 77030, USA 9 Department of Biochemistry and Molecular Biology, Baylor College of Medicine, Houston, Texas, 77030, USA 10 Department of Biotechnology, Asia University, Taichung 413, Taiwan 11 Institute of Pathology and Southwest Cancer Center, Southwest Hospital, Third Military Medical University, Chongqing 400038, China * Correspondence: Mien-Chie Hung, Department of Molecular and Cellular Oncology, University of Texas MD Anderson Cancer Center, 1515 Holcombe Blvd., Unit 108, Houston, TX 77030. Tel.: 713-792-3668. Fax: 713-794-3270. mhung@mdanderson.org † These authors contributed equally 13 11 2016 6 10 2016 1 12 2016 01 12 2017 76 23 70497058 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. The oncogenic transcription factor Gli1 is a critical effector in the Hedgehog (Hh) pathway which is necessary for the development and progression of pancreatic ductal adenocarcinoma (PDAC). While TGF-β and K-Ras are known regulators of Gli1 gene transcription in this setting, it is not understood how Gli1 functional activity is regulated. Here we report the identification of Gli1 as a substrate for the protein arginine N-methyltransferase PRMT1 in PDAC. We found that PRMT1 methylates Gli1 at R597, promoting its transcriptional activity by enhancing the binding of Gli1 to its target gene promoters. Interruption of Gli1 methylation attenuates oncogenic functions of Gli1 and sensitizes PDAC cells to gemcitabine treatment. In human PDAC specimens, the levels of both total Gli1 and methylated Gli1 were correlated positively with PRMT1 protein levels. Notably, PRMT1 regulated Gli1 independently of the canonical Hh pathway as well as the TGF-β/Kras-mediated non-canonical Hh pathway, thereby signifying a novel regulatory mechanism for Gli1 transcriptional activity. Taken together, our results identifed a new posttranslational modification of Gli1 that underlies its pivotal oncogenic functions in PDAC. Introduction The Hedgehog (Hh) signaling pathway plays critical roles in normal tissue patterning and differentiation during mammalian embryogenesis (1). Although the Hh pathway is inactivated in adults, it is reactivated during tumorigenesis in various organs, including pancreas (2). Canonically, inhibitory engagement of the Hh membrane receptor, Patched 1 (PTCH1), by Hh ligands releases Smoothened (SMO) from PTCH. Activated SMO then removes Suppressor of Fused (SuFu) from Gli transcriptional factors and allows Gli proteins to translocate into the nucleus to drive transcription of the target genes of Hh pathway. Three Gli proteins (Gli1, Gli2, and Gli3) have been identified, with Gli1 possessing the strongest transcriptional activity (3). Besides SuFu, other Gli1-binding proteins have been reported to regulate Gli1 activities. For example, we previously demonstrated that p70S6K phosphorylates Gli1 and releases Gli1 from SuFu, leading to Gli1 activation (4). DYRK1 (5) and aPKC-ι/λ (6) also enhance Gli1 activities via Gli1 phosphorylation. In contrast, protein kinase A-mediated Gli1 phosphorylation negatively regulates Gli1 functions (7). In addition, the Numb/Itch complex and the p300/CBP-associated factor interact with Gli1 and mediate Gli1 ubiquitination and degradation (8,9). Snf5 also interacts with and suppresses Gli1 (10). Together, these findings demonstrate that Gli1 activity can be regulated through various non-canonical Hh pathways and suggest that identification of novel Gli1-associated proteins may shed new light on the regulation of Gli1. The canonical Hh pathway (cHh) is known to play significant roles in pancreatic ductal adenocarcinoma (PDAC), one of the most intractable cancers, with the poorest survival rate of all cancers (11). Using a paracrine mechanism (12), PDAC cells produce Hh ligands to activate the cHh pathway in tumor-associated stromal cells but not in PDAC cells. However, Gli1 is also functionally required for PDAC cells to survive and proliferate (13,14), and high Gli1 protein levels are related to poor survival in patients with PDAC (15). Although TGF-β and Kras are known regulators of Gli1 gene transcription in PDAC (14), how the Gli1 activity is regulated in PDAC remains to be clarified. Thus, improving our understanding on the mechanism of Gli1 regulation in PDAC may lead to the development of a novel targeted therapy for PDAC. Materials and Methods Human tissues Human PDAC tissues from patients treated at The University of Texas MD Anderson Cancer Center were obtained retrospectively for immunohistochemical microarray analysis. The tissues were collected in accordance with the protocols approved by the Institutional Review Board at MD Anderson Cancer Center, and written informed consent had been obtained from all patients at the time of enrollment. PDAC xenografts were obtained as previously described (16). Antibodies and reagents The antibodies used in this study were Gli1 (#3538, Cell Signaling Technology, for western blotting and chromatin immunoprecipitation (ChIP) assay, and#sc-20687, Santa Cruz Biotechnology, for immunohistochemistry), actin (#A2066, Sigma-Aldrich), Flag (#F3165, Sigma-Aldrich), Flag M2 magnetic beads (#M8823, Sigma-Aldrich), PRMT1 (#2449, Cell Signaling Technology), and tubulin (#T5168, Sigma-Aldrich). GDC-0449, NVP-LDE225, and RAD-001 were purchased from Selleck Chemicals LLC, GANT58 and GANT61 from Tocris Bioscience, and AMI-1 from Sigma-Aldrich. The SYBR Green real-time PCR kit was obtained from Bio-Rad and TGF-β from Peprotech (#100-21). The antibody against R597-methylated Gli1 was developed using the following synthetic peptide with asymmetric dimethylation at R597 as antigen: RARYASA-[R597(aMe2)]-GGGTS (C-terminal amidation and N-terminal KLH conjugation). Other peptides used in the study included the following: Cold Gli1 peptide without R597 methylation [Gli1-R597]: RARYASA-[R597]-GGGTS; Hot Gli1 peptide with R597 monomethylation [Gli1-R597(Me)]: RARYASA-[R597(Me)] -GGGTS; Hot Gli1 peptide with R597 asymmetric dimethylation [Gli1-R597(aMe2)]: RARYASA-[R597(aMe2)]-GGGTS; Hot Gli1 peptide with R597 symmetric dimethylation [Gli1-R597(sMe2)]: RARYASA-[R597(sMe2)]-GGGTS. Plasmids Flag-tagged Gli1 was generated from pCMV10-3xFlag-Gli1 (4) and inserted into pCDH-CMV-MCS-EF1-Neo (System Biosciences). The R597 mutant was generated by using the QuikChange Multi Site-Directed Mutagenesis Kit (Agilent Technologies) with pCDH-CMV-MCS-EF1-Neo-Gli1 as a template. The pCDH-RFP-Luciferase plasmid was constructed by inserting the luciferase gene into the pCDH-RFP vector (System Biosciences). The hemagglutinin (HA) -tagged PRMT1 plasmid was generated by inserting PRMT1 complementary DNA into pCMV5. PRMT1 knockdown was carried out by shRNA with the sequence of CCGGCAGTACAAAGACTACAA (#1), GTGTTCCAGTATCTCTGATTA (#2), GCAAGTGAAGCGGAATGACTA (#3) in the vector pLKO.1. #1 shRNA-resistant PRMT1 was produced through mutation of the target sequence of #1 shRNA in PRMT1-expressing plasmid (mutating CCGGCAGTACAAAGACTACAA to TAGACAATATAAGGATTATAA) without changing amino acid. Lentiviral shRNA system in pGIPZ vector targeting Kras was purchased from Thermo Scientific. pGEX-6P-1 (GE Healthcare) was used to purify proteins for the in vitro methylation assay. Cell culture The human PDAC cell lines AsPC-1, MIA PaCa-2, and CFPAC-1 were obtained from the American Type Culture Collection (ATCC, Manassas, VA), and maintained at 37 °C in a 5% CO2 incubator in Dulbecco modified Eagle medium/F12 or RPMI 1640 plus 10% fetal bovine serum. Human PDAC-associated stromal cell line HPSC was previously described (17). All cell lines were characterized as mycoplasma negative and validated by STR DNA fingerprinting using the AmpFLSTR Identifiler kit (ThermoFisher) according to manufacturer’s instructions semiannually. The STR profiles were compared with known ATCC fingerprints (www.ATCC.org) and with the Cell Line Integrated Molecular Authentication database (CLIMA) version 0.1.200808 (http://bioinformatics.istge.it/clima/) (Nucleic Acids Research 37:D925-D932, PMCID: PMC2686526). The STR profiles matched known DNA fingerprints or were unique. Transfection and lentiviral infection The plasmids were transfected using Lipofectamine 2000 according to the manufacturer’s instructions. The cells were harvested for mRNA extraction or protein extraction after 48 h of transfection. For lentiviral infection, vector plasmids and packaging plasmids were co-transfected into 293T cells using Lipofectamine 2000, and the lentiviruses were concentrated as described previously (18). The generated viruses were used for cell infection in the presence of polybrene (EMD Millipore). After infection, stable cells were isolated by selection for resistance to puromycin or neomycin or by fluorescence-activated cell sorting. Specifically, to generate stable cells with luciferase, MIA PaCa-2 cells were infected with lentivirus containing pCDH-RFP-Luciferase followed by RFP sorting. MIA PaCa-2 luciferase-expressing stable (MIA PaCa-2-Luc) cells were then used to establish wild-type Gli1 or R597-mutantGli1 stable cells by infection with lentivirus containing pCDH-CMV-MCS-EF1-Neo-Gli1WT or pCDH-CMV-MCS-EF1-Neo-R597K under G418 selection. To generate Gli1 knockout cells, AsPC-1 cells were co-transfected with Gli1 CRISPR/Cas9 KO plasmid (Santa Cruz Biotechnology, sc-400266) and Gli1 HDR plasmid (Santa Cruz Biotechnology, sc-400266-HDR). After 48 hours, cells were selected with puromycin for 1 week and then analyzed for knockout efficiency. TCGA gene expression data sets and data processing TCGA mRNA expression (RNASeq V2 RSEM) data for lung adenocarcinoma, prostate adenocarcinoma, breast invasive carcinoma, colorectal adenocarcinoma, PDAC, liver hepatocellular carcinoma, and ovarian serous cystadenocarcinoma were downloaded from the cBioPortal Web site (19,20). Box plots showing 5th/95th percentiles and medians were generated with SigmaPlot software. Animal studies All animal experiments were approved by the Institutional Animal Care and Use Committee of MD Anderson Cancer Center. 6-week nude mice were housed under standard conditions. For orthotopic tumorigenicity assay, MIA PaCa-2 cells with luciferase only, luciferase plus wild-type Gli1, or luciferase plus R597K-mutant Gli1 were suspended in 50% Matrigel (#354230, BD Biosciences) in phosphate-buffered saline at a concentration of 5 × 106 cells/ml. The viability of the cells was > 98%. The number of mice per group was 5. General anesthesia was administered using isoflurane (#029405, Henry Schein Animal Health). A left lateral laparotomy was performed, and the spleen and distal pancreas were mobilized. Approximately 50 µl of the cells (0.25 × 106) were injected into the pancreas. The abdominal incision was closed using a surgical suture, and analgesia was administered for immediate pain relief. Palpable tumors were detected after 14 days and imaged twice a week using the Xenogen IVIS in vivo imaging system (Caliper Life Sciences) as described elsewhere (21). For subcutaneous tumorigenicity assay, 1 × 105 cells were subcutaneously injected in right flank. The resulting tumors were measured with calipers weekly, and tumor volume was determined using the formula (Length) × (Width)2, where l is the longest diameter and w is the shortest diameter. The tumors were measured with calipers every week. Data were presented as tumor volume (mean ± SD). Statistical analysis was done using the Student's t-test by the program SPSS for Windows. Statistical analyses Statistical analyses were performed with the Student’s t-test, Spearman rank correlation test, or Fisher exact test. A P value of < 0.05 was considered statistically significant. All data analyses were performed using the analytic software SPSS (IBM) for Windows. Results and Discussion PRMT1 interacts with Gli1 and promotes Gli1 transcription We first analyzed The Cancer Genome Atlas (TCGA) database and found that among of the seven deadliest cancers in 2014 in the United States (22), PDAC had the highest mean Gli1 mRNA expression (Figure 1A), which supports the reports that Gli1 plays critical roles in PDAC. Then, we profiled Gli1-associated proteins to identify potential regulators of Gli1 in PDAC cells. For this purpose, Flag-tagged Gli1 was stably transfected into MIA PaCa-2PDAC cells. Gli1-associated proteins were isolated using anti-Flag antibody and analyzed by mass spectrometry (Supplemental Figure 1A). Comparison of Flag-Gli1 and Flag-vector results profiled a total of 471 potential Gli1-binding proteins, including several well-known Gli1 interaction partners, e.g. SuFu and protein kinase A (3) (Table S1). Because Gli1 regulation in the non-canonical Hh pathway often requires enzyme-induced posttranslational modifications such as phosphorylation, ubiquitination, and acetylation (4–9), we thus specially focused on the Gli1-binding proteins with enzymatic activities (Table S2). Besides kinases, ubiquitin-conjugating enzymes, and acetyltransferases, which are responsible for modifications known to occur in Gli1 (4–8,23,24), we also noticed a methyltransferase, PRMT1, which catalyzes protein methylation (25) but has not yet been reported to be associated with Gli1. PRMT1 is a type I protein arginine methyltransferase that catalyzes asymmetric dimethylation on arginine residues. In mammalian cells, about 85% of all occurrences of asymmetric dimethylation are produced by PRMT1 (25). The arginine methylation mediated by PRMT1 positively or negatively affects protein functions depending on the biological contexts (26,27), and our recent work further revealed an important function of PRMT1 on EGFR regulation in colon cancer (28). Then, we confirmed the interaction between Gli1 and PRMT1 in AsPC-1 and MIA PaCa-2 cells (Figure 1B), but neither Gli2 nor Gli3 interacted with PRMT1 (Supplemental Figure 1B). To address the potential relationship between PRMT1 and Gli1 and the role of Gli1 in PDAC, we first knocked down PRMT1 in AsPC-1 cells by using three different small hairpin RNAs (shRNAs) and observed decreases in Gli1 protein levels. The decreases were reversed by shRNA-resistant ectopic expression of PRMT1 (Figure 1C). Knockdown of PRMT1 in MIA PaCa-2 and CFPAC-1 also decreased Gli1 protein and mRNA levels (Figures 1D and 1E). In contrast, ectopic expression of PRMT1 enhanced Gli1 expression (protein and mRNA)in MIA PaCa-2 cells (Supplemental Figure 1C). Interestingly, Gli1 protein stability did not change following such enhancement in AsPC-1 cells (Supplemental Figure 1D). These results suggested that PRMT1 upregulates Gli1 protein levels through enhanced Gli1 transcription. Microarray analysis of patient-derived PDAC tissues (n = 122) by immunohistochemistry staining also supports a positive correlation between Gli1 and PRMT1 protein expression levels (P = 0.008) (Figures 1F and 1G). PRMT1 methylates Gli1 at R597 To determine whether Gli1 is a substrate for PRMT1, we performed an in vitro methylation assay and detected a major methylation signal in a Gli1 fragment (Gli1 F2) containing amino acids 354 to 753 (Supplemental Figure 2A). We analyzed four segments of this fragment to identify potentially methylated arginine residues. Of the four subfragments, only F2-3, containing amino acids 543 to 620, produced methylation signals (Supplemental Figure 2B). When we mutated all six arginines in Gli1 F2-3 to lysine (K), only mutation of Arg597 (R597) eliminates the methylation signal (Supplemental Figure 2C). We obtained a similar result with this mutation in the full-length Gli1, in which the methylation signal was completely eliminated (Figure 2A). These results indicated that Gli1 is a substrate of PRMT1 in vitro and that R597 is the primary site methylated by PRMT1. We then purified Gli1 from AsPC-1 cells for mass spectrometry analysis, and validated dimethylated R597 also occurred in vivo (Supplemental Figure 2D). Sequence alignment of Gli1 from different mammals indicated that R597 within the RG rich motif, which is a typical feature of PRMT1 substrates (29), is highly conserved from mice to humans (Supplemental Figure 2E), suggesting a potentially important role of R597. To further characterize this methylated residue in Gli1 and to study the correlation between R597-dimethylated Gli1 (meGli1R597) and PRMT1, we developed a meGli1R597-specific antibody, which recognizes peptides with asymmetrically dimethylated R597 in Gli1 but not peptides with non-modified, monomethylated, or symmetrically dimethylated R597 (Figure Supplemental 2F). This antibody recognized only wild-type Gli1 but not the Gli1R597K mutant (Figure 2B) and therefore is suitable for detection of meGli1R597. In MIA PaCa-2 cells, ectopic expression of PRMT1 led to enhanced Gli1 methylation (Figure 2C). Since the antibody was unable to detect meGli1R597 by immunohistochemistry (data not shown), we instead performed Western blot analysis of protein lysates from 30 human PDAC xenografts maintained in mice (30). Our results indicated a positive correlation between the levels of methylated Gli1 and PRMT1 (Figure 2D), further validating our in vitro finding that Gli1 is a substrate of PRMT1. However, PRMT1 and total Gli1 or Gli1 and meGli1 only showed weak positive correlation, which might be attributed to multiple regulators of Gli1 besides PRMT1 (Figure 2D). Methylation of R597is required for regulation of Gli1 by PRMT1 Because PRMT1 increases Gli1 mRNA transcription and Gli1 is a target gene of itself (31), we asked whether Gli1R597 is required for PRMT1-enhanced Gli1 transcriptional activity. As expected, the transcripts of several Gli1 target genes, including endogenous GLI1, PTCH1, IGFBP6, CCND1, BCL2, and SNAIL1 (32), were lower in MIA PaCa-2 cells stably expressing R597K-mutant Gli1 (MIA/Gli1RK cells) than in MIA PaCa-2 cells expressing wild-type Gli1 (MIA/Gli1WT cells) (Figure 3A and Supplemental Figure 3A). Consistently, results from a transient reporter assay in 293T cells using Gli binding sequence (GliBS) reporter plasmid (31) confirmed that wild-type Gli1 harbored stronger transcriptional activity compared with R597K-mutant Gli1 (Supplemental Figure 3B). To further investigate the significance of R597 in Gli1 functions, we knocked out Gli1 in AsPC-1 cells, which have a high basal level of Gli1 protein, by CRISPR/Cas9, followed by reconstitution with wild-type Gli1 or R597K-mutant Gli1 (Figure 3B). We found that the loss of Gli1 decreased the expression of Gli1 target genes. Interestingly, restoring wild-type Gli1 expression also rescued the expression of all five target genes, but restoring the R597K mutant Gli1 only rescued PTCH1 and CCND1 (Figure 3B). Hence, the R597 site may be required for stronger transcriptional activity of Gli1, and without methylation, Gli1 transcriptional activity on some of Gli1 target genes may be suppressed. In addition, ectopic expression of PRMT1 in MIA/Gli1WT cells, but not in MIA/Gli1RK cells, enhanced Gli1 target gene expression (Figure 3C). Likewise, depletion of PRMT1 repressed the expression of Gli1 target genes in MIA/Gli1WT cells but not in MIA/Gli1RKcells (Supplemental Figure 3C), which implied that PRMT1 affects Gli1 transcriptional activities through the site of R597. Because Gli1 is a transcriptional factor, we asked whether the methylation status of Gli1R597 affects the occupancy of Gli1 on the promoter of its target genes. To this end, we examined the occupancy of Gli1 variants in natural promoters of CCND1, IGFBP6, and BCL2 (33,34) by ChIP assay. The results indicated that wild-type Gli1 occupied the promoter of BCL2 and IGFBP6 more than did the R597K-mutant Gli1, whereas both wild-type and mutant Gli1 similarly occupied the promoter of CCND1 (Supplemental Figure 3D). Similar results were observed from PCR analysis in which wild-type Gli1 promoted the transcription of BCL2 and IGFBP6 more than did the R597K-mutant Gli1 whereas both wild-type or R597K-mutant Gli1 promoted similar transcription of CCND1. These results suggested that Gli1 R597 methylation is required for Gli1 occupancy on promoters of some target genes but not for the others. Guendel et al. previously reported that methylation of BRCA1 by PRMT1 alters the occupancy of BRCA1 on its target gene promoters (35). Collectively, our results indicated that methylation of Gli1 at R597 rendered stronger transcriptional activity on some of Gli1 target genes, including SNAIL1, BCL2, and IGFBP6. Next, we specifically focused on IGFBP6 and BCL2 promoters, which were affected by Gli1 methylation status. Wild-type Gli1 and R597K-mutant Gli1 were transfected separately or together into MIA PaCa-2cells with adjustment of plasmid amounts to ensure similar levels of total Gli1 in different samples. The results indicated that the occupancy of Gli1 on the promoters is positively correlated with the expression of wild-type Gli1 but not R597K-mutant Gli1 (Figure 3D). Depletion of PRMT1 by shRNA decreased the accumulation of Gli1 to the promoters in MIA/Gli1WT but not MIA/Gli1RK cells (Figure 3E). In AsPC-1 cells, knockdown of PRMT1 reduced the binding of Gli1 to the promoters of IGFBP6 and BCL2, and rescue of PRMT1 restored the accumulation of Gli1 at the promoters (Figure 3F). Similar results were observed when we treated cells with the pan-PRMT inhibitor AMI-1 (arginine methyltransferase inhibitor 1) (36), which attenuated Gli1 methylation (Supplemental Figures 3E). Together, the results suggested that methylation of R597 in Gli1 by PRMT1 promotes Gli1 accumulation at the promoters of Gli1 target genes and enhances their transcriptions. Gli1 oncogenic functions are inhibited by loss of R597 methylation Since Gli1expression is associated with malignant transformation, we assessed the effects of methylation on Gli1 oncogenic functions. The results from bromodeoxyuridine (BrdU) incorporation assay showed that MIA/Gli1RK cells grew slower than MIA/Gli1WT cells (Supplemental Figure 4A). Cell viability (Supplemental Figure 4B), anchorage-independent growth (Figure S4C), and migration and invasion abilities (Supplemental Figures 4D and 4E) were lower in MIA/Gli1RK cells than in MIA/Gli1WT cells. In AsPC-1 cells, knockout of Gli1 inhibited cell proliferation, as measured by BrdU incorporation and cell counting. The inhibition was completely reversed by reconstitution with wild-type Gli1 but not R597K-mutant Gli1 (Supplemental Figure 4F and 4G). These results suggested that R597 methylation is necessary for full oncogenic effects of Gli1. Given that Gli1 has been implicated in resistance to gemcitabine, a chemotherapy agent widely used to treat PDAC (37–39), we investigated whether PRMT1-methylated R597 contributes to gemcitabine resistance. Indeed, even though MIA/Gli1WT cells were more resistant to gemcitabine than MIA/Gli1RK cells, the differences were marginal after PRMT1 knockdown (Figure 4A). Similar results were obtained with CFPAC-1 cells carrying wild-type or R597-mutant Gli1 (Supplemental Figure 4H). In addition, the basal level of apoptosis and the percentage of cells showing gemcitabine-induced apoptosis were higher in MIA/Gli1RK cells than in MIA/Gli1WT cells (Figure 4B). Results from an orthotopic model of human PDAC in nude mice indicated that inhibition of R597 methylation in Gli1 attenuated tumor growth (Figure 4C). Likewise, AsPC-1 cells with Gli1 knockdown formed very few subcutaneous tumors, and cells reconstituted with wild-type Gli1 formed much larger tumors than cells reconstituted with R597K-mutant Gli1 (Figure 4D and Supplemental Figure 4I). Collectively, these results indicated that inhibition of R597 methylation in Gli1 reduces the oncogenic activities of Gli1 in vitro and in vivo and that such inhibition may provide a novel strategy for reducing gemcitabine resistance in PDAC. Regulation of Gli1 by PRMT1 is a novel pathway in PDAC Since Gli1 can be regulated via a SMO-dependent (cHh) or SMO-independent (non-canonical Hh) pathway (4,14), we investigated whether the regulation of Gli1 by PRMT1 in PDAC relies on SMO. We treated MIA PaCa-2, MIA/Gli1WT, and MIA/Gli1RK cells with the SMO inhibitor GDC-0449 (marketed as Vismodegib by Roche) or NVP-LDE225 (marketed as Erismodegib by Novartis) and found that none of the cells were sensitive to the inhibitors (Supplemental Figure 5A). These results are consistent with previous studies showing that PDAC cells lack the cHh pathway (23). Gli1 is known to be transcriptionally regulated by TGF-β and Kras in PDAC (14). Thus, we also examined the possibility that TGF-β and/or Kras regulate PRMT1-mediated methylation of Gli1. To avoid interference of endogenous Gli1 transcription, we used the previously mentioned Gli1−/− with reconstitution of exogenous wild-type Gli1, which does not carry endogenous Gli1 promoters. In this cell line, knockdown of Kras had no effects on PRMT1-mediated Gli1 methylation (Supplemental Figure 5B), and TGF-β treatment, which induces Gli2 upregulation as expected (14), did not affect Gli1R597 methylation either (Supplemental Figure 5C). Therefore, regulation of Gli1 by PRMT1 is a novel pathway in PDAC. Because tumor-associated stromal cells play important roles in the tumor microenvironment of PDAC (40,41), we further investigated if the regulation of Gli1 by PRMT1 exists in PDAC-associated stromal cells. We detected both Gli1 and meGli1 in a human PDAC-associated stromal cell line, HPSC (17) and treatment of HPSC cells with PRMT inhibitor (AMI) substantially reduced meGli1 (Supplemental Figure 5D). Therefore, in tumor-associated stromal cells, methylation of Gli1 may also be dependent on PRMT1. Finally, we treated PDAC cells with GANT58 and GANT61, which interfere with the binding of Gli1 to the promoters of Gli1 target genes (42). Remarkably, GANT58 and GANT61 inhibited the viability of both MIA/Gli1WT and MIA/Gli1RK cells (Figure 5A). These results suggested that direct targeting of Gli1 may be an alternative to targeting the upstream cHh pathway for treatment of PDAC. Gli1 function is regulated by post-translational modifications, e.g., phosphorylation and acetylation (4,5,7,24,43). In this study, we have identified Gli1 as a new substrate for PRMT1 and characterized a novel post-translational modification of Gli1 in PDAC cells that was independent of cHh or TGF-β/Kras (Figure 5B). Methylation of Gli1 by PRMT1 promotes Gli1 transcriptional activity by enhancing Gli1’s accumulation on its target gene promoters, thereby activating the expression of those genes. In addition, even though R597K-mutant Gli1 retained its transcriptional activity without R597 methylation, the activity was much weaker than that of wild-type Gli1 for some of target genes. These findings implied that although R597 methylation is not absolutely required for Gli1 function, it does increase Gli1 transcriptional activities, which is sufficient to enhance the oncogenic functions of Gli1 by increasing transcription of its target genes, such as SNAIL1, BCL2, and IGFBP6. It would be of interest to comprehensively profile all the genes that are sensitive to Gli1 R597 methylation. In vivo, Gli1 functions are constitutively inhibited by SuFu (via inhibitory interaction with Gli1) or by a repressor form of Gli3 (via competitive binding to the same promoters as those bound by Gli1). However, increased Gli1 proteins and/or increased Gli1 binding to its promoters can overcome inhibition by SuFu or Gli3 repressor through either cHh pathway activation or the newly identified methylation of R597 in Gli1. Thus, methylation of R597 in Gli1 is potentially an important alternative mechanism of promoting Gli1 oncogenic functions in vivo. We also observed that the regulation of Gli1 by PRMT1 was not confined in the tumor cells, and tumor-associated stromal cell may also have this pathway, which implied that PRMT1-mediated Gli1 methylation has broader involvement in various cell types. The Hh signaling pathway plays a pivotal role in PDAC development and progression via a paracrine loop (12). However, a cHh pathway inhibitor based on this paracrine model did not produce successful results against PDAC in a clinical trial (40,41,44). In the current study, we demonstrated that elimination of Gli1 methylation by PRMT1 substantially attenuated Gli1-related oncogenic functions and sensitized PDAC cells to gemcitabine, a front-line drug for PDAC. In addition, and our result that GANT61 attenuated proliferation of MIA/Gli1WT and MIA/Gli1RK PDAC cells complements previous reports that GANT61 suppresses PDAC by inhibiting growth of cancer stem cells and promoting autophagy of tumor cells (45,46). Collectively, the findings in this study indicated that the targeting both PRMT1 and Gli1 is a viable new direction for treatment of PDAC. Supplementary Material 1 2 3 4 We thank Arthur Gelmis at the Department of Scientific Publications at MD Anderson for editing the manuscript. This work was supported in part by the following: National Institutes of Health (CA109311, CA099031, and CCSG CA016672); The University of Texas MD Anderson-China Medical University and Hospital Sister Institution Fund (to M.-C.H.); Ministry of Science and Technology, International Research-intensive Centers of Excellence in Taiwan (I-RiCE; MOST 105-2911-I-002-302); Ministry of Health and Welfare, and China Medical University Hospital Cancer Research Center of Excellence (MOHW105-TDU-B-212-134003); Center for Biological Pathways; and MD Anderson Cancer Center Sister Institute Network Fund. Figure 1 Correlation between the Expression Levels of Gli1 and PRMT1 in PDAC Cells (A) GLI1 mRNA levels according to TCGA data for patients with the seven deadliest cancers in the United States in 2014. The data are medians with the 5th and 95th percentiles and standard deviations (error bars). (B) Western blot analysis of immunoprecipitation (IP) of endogenous Gli1 and PRMT1 in AsPC-1 cells. IgG, immunoglobulin G. (C) AsPC-1 cells expressing scrambled shRNA (sh-Ctrl), PRMT1-targeting shRNA (sh-PRMT1 #1, 2, or 3), or PRMT1-targeting shRNA (sh-PRMT1 #1) with reconstituted shRNA-resistant PRMT1. (D) Western blot analysis of Gli1 and PRMT1 in three PDAC cell lines infected with control (Ct) or PRMT1 (Pr) shRNA. The results were quantified using ImageJ software, and normalized to the values for tubulin. The experiments were performed at least two times to assure reproducibility of the results. (E) mRNA expression of endogenous Gli1 and PRMT1 measured by quantitative real-time PCR in the indicated cell lines transfected with control shRNA or sh-PRMT1. The data are means with standard deviations (n = 3). *P< 0.05, **P< 0.01 (paired two-tailed Student’s t-test). (F) Representative immunohistochemistry staining of Gli1 and PRMT1 in human PDAC tissues. All immunostained slides were scanned on the ACIS III automated cellular image system for quantification by digital image analysis. The percentage of positive cells (X) and signal intensity (Y) are shown. The number from X × Y represents an arbitrary quantitative score. Tumor (T) area was labeled with dash line. The positive staining in stroma is labeled with arrowhead. (G) Analysis of correlation between Gli1 and PRMT1 levels on the basis of immunohistochemistry results for 122 human PDAC tissue samples. Protein expression was calculated from both the percentage of stained cells and the immunostaining intensity. Protein expression levels above and below the mean for all samples were categorized as high and low, respectively. There are 4 categories based on the Gli1 and PRMT1 scores on all of the immunostained slides: 1) Gli1 and PRMT1 high; 2) Gli1 and PRMT1 low; 3) Gli1 high and PRMT1 low; 4) Gli1 low and PRMT1 high. Fisher’s exact test was used to evaluate the correlation between Gli1 and PRMT1 in the 122 human tissue slides (P < 0.05). Figure 2 Methylation of Gli1 by PRMT1 (A) In vitro methylation assay with PRMT1 and wild-type (WT) or R597K-mutant Gli1. Left panel, Coomassie Blue staining. Right panel, fluorography. (B) Western blot analysis of immunoprecipitation (IP) with antibody specific to meGli1R597 (meGli1), antibody to total Gli1, and other antibodies as indicated in MIA PaCa-2 cells transfected with a plasmid carrying Flag (FL)-tagged WT Gli1 or R597K-mutant Gli1 (RK). (C) Western blot analysis of meGli1, Gli1, and PRMT1 in MIA PaCa-2 cells transfected with an empty vector or a hemagglutinin (HA)-tagged PRMT1 plasmid. (D) Western blot and correlation analysis of meGli1 and PRMT1 in human PDAC xenografts maintained in mice. Each set of samples was subjected to two independent Western blotting (upper panels), and the bands were quantified using ImageJ software. Mean expression levels were used to determine Pearson coefficients for correlation between PRMT1 and meGli1, between PRMT1 and Gli1, and between Gli1 and meGli1 (lower panels). Figure 3 R597 Methylation Positively RegulatesGli1 Transcriptional Activity (A) Left panel, Western blot analysis of meGli1 and total Gli1 in MIA PaCa-2 luciferase cells stably transfected with an empty vector (Vec), wild-type Gli1 (Gli1WT), or R597K-mutant Gli1 (Gli1RK). Right panel, mRNA expression levels, measured by quantitative real-time PCR, of Gli1 target genes in Vec-, Gli1WT (WT)-, and Gli1RK (RK)-transfected MIA PaCa-2 cells. Error bars represent SD (n = 3). *P < 0.05, **P < 0.01 (paired two-tailed Student’s t-test). (B) Left panel, Western blotting of Gli1 protein levels in AsPC-1 parental cells (PA), AsPC-1 cells with Gli1 knockout (Gli1−/−), and Gli1−/− AsPC-1 cells reconstituted with Gli1 (WT) or Gli1 RK mutant (RK). The intensity of the bands was quantified and normalized to that of tubulin. Right panel, mRNA expression of Gli1 target genes by qRT-PCR in PA, Gli1−/−, WT, and RK cells. The expression levels of target genes were normalized to that of ACTIN. Statistical significance was determined by paired, two-tailed Student’s t-test. Error bars represent SD (n = 3). *P < 0.05, **P < 0.01. (C) mRNA levels of IGFBP6 and BCL2 in MIA PaCa-2 cells transfected with wild-type Gli1 and an empty vector (WT/Vec), wild-type Gli1 and PRMT1 (WT/Prm), R597K-mutant Gli1 and an empty vector (RK/Vec), or R597K-mutant Gli1 and PRMT1 (RK/Prm). The value of WT/Prm was normalized to that of WT/Vec. The value of RK/Prm was normalized to that of RK/Vec. Error bars represent SD (n = 3). *P < 0.05 (paired two-tailed Student’s t-test). (D) Left panel, ChIP assay using Gli1 antibody for immunoprecipitation (IP) and promoter-specific primers for quantitative qRT-PCR to confirm Gli1 binding regions in promoters. IgG, immunoglobulin G; SP, primers specific to Gli1 binding regions; NSP, primers not specific to Gli1 binding regions. Middle panel, protein expression of hemagglutinin-tagged wild-type Gli1 (HA-WT) and Flag-tagged R597K-mutant Gli1 (Flag-RK) in 293T cells. Right panel, quantitative results of ChIP assay. WT, wild-type Gli1; RK, R597K-mutant Gli1. Error bars represent SD. (E) Quantitative results of ChIP assay for Gli1-bound promoters of IGFBP6 and BCL2 from qRT-PCR analyses in MIA PaCa-2 cells carrying wild-type (WT) or R597K-mutant (RK) Gli1 and infected with virus carrying control shRNA (shC) or shRNA targeting PRMT1 (shP). Error bars represent SD (n = 3). *P < 0.05, **P < 0.01 (paired two-tailed Student’s t-test). (F) Analysis of IGFBP6 and BCL2 promoters bound by Gli1 by ChIP-qPCR in AsPC-1 cells expression scrambled (sh-Ctrl) shRNA, PRMT1-targeting (sh-PRMT1 #1 and #2) shRNA, or PRMT1-targeting shRNA with reconstituted wild type PRMT1 (sh#1-Rsc). Statistical significance was determined by paired, two-tailed Student’s t-test. Error bars represent SD from triplicate experiments. *P < 0.05, **P < 0.01. Figure 4 R597 methylation positively regulates Gli1 oncogenic functions (A) Responses of MIA PaCa-2 stable clones to gemcitabine with or without PRMT1 depletion. shCtrl: control shRNA; shPRMT1: PRMT1 shRNA; Vec: MIA PaCa-2 stable clone with empty vector; WT: stable clone with wild-type Gli1; RK: stable clone with Gli1R597 mutant. Error bars represent SD (n = 4). (B) Propidium iodide staining by fluorescence-activated cell sorting to determine the percentage of apoptosis in different stale clones with or without gemcitabine. Error bars represent SD (n = 3). (C) The indicated MIA PaCa-2 stable clones were inoculated into the pancreas of 6-week nude mice. Tumor volume was measured at the indicated time points using the formula (Length) × (Width)2. Vector: MIA PaCa-2-Luc stable cells; Gli1WT: MIA PaCa-2-Luc Gli1WT stable cells; Gli1RK: MIA PaCa-2-Luc Gli1R597K stable cells. Error bars represent SD (n = 5). (D) AsPC-1 parental cells (PA), AsPC-1 cells with Gli1 knockout (Gli1−/−), or Gli1−/− AsPC-1 cells reconstituted with Gli1 (WT) or Gli1 RK mutant were subcutaneously injected into the right flank of nude mice (n = 7). Tumor volume was measured once a week. Statistical significance was determined by paired, two-tailed Student’s t-test. Error bars represent SD. *P < 0.05, **P < 0.01. Figure 5 PRMT1Regulates Gli1 via a Novel Non-Canonical Hh Pathway (A) MTT assay of the indicated MIA PaCa-2 stable cells treated with GANT58 or GANT61. The data are means (relative to the value for day 1) with standard deviations (n = 3). *P < 0.05, **P < 0.01 (paired two-tailed Student’s t-test). (B) A schematic diagram illustrating the regulation of Gli1 via SMO-dependent (cHh) or SMO-independent (non-canonical Hh) pathways in PDAC. Conflicts of Interest The authors have declared that no conflict of interest exists. References 1 Barakat MT Humke EW Scott MP Learning from Jekyll to control Hyde: Hedgehog signaling in development and cancer Trends in molecular medicine 2010 16 8 337 348 20696410 2 Ng JM Curran T The Hedgehog's tale: developing strategies for targeting cancer Nat Rev Cancer 2011 11 7 493 501 21614026 3 Robbins DJ Fei DL Riobo NA The Hedgehog signal transduction network Science signaling 2012 5 246 re6 23074268 4 Wang Y Ding Q Yen CJ Xia W Izzo JG Lang JY The crosstalk of mTOR/S6K1 and Hedgehog pathways Cancer cell 2012 21 3 374 387 22439934 5 Mao J Maye P Kogerman P Tejedor FJ Toftgard R Xie W Regulation of Gli1 transcriptional activity in the nucleus by Dyrk1 The Journal of biological chemistry 2002 277 38 35156 35161 12138125 6 Atwood SX Li M Lee A Tang JY Oro AE GLI activation by atypical protein kinase C iota/lambda regulates the growth of basal cell carcinomas Nature 2013 494 7438 484 488 23446420 7 Sheng T Chi S Zhang X Xie J Regulation of Gli1 localization by the cAMP/protein kinase A signaling axis through a site near the nuclear localization signal The Journal of biological chemistry 2006 281 1 9 12 16293631 8 Di Marcotullio L Ferretti E Greco A De Smaele E Po A Sico MA Numb is a suppressor of Hedgehog signalling and targets Gli1 for Itch-dependent ubiquitination Nature cell biology 2006 8 12 1415 1423 17115028 9 Mazza D Infante P Colicchia V Greco A Alfonsi R Siler M PCAF ubiquitin ligase activity inhibits Hedgehog/Gli1 signaling in p53-dependent response to genotoxic stress Cell death and differentiation 2013 20 12 1688 1697 24013724 10 Jagani Z Mora-Blanco EL Sansam CG McKenna ES Wilson B Chen D Loss of the tumor suppressor Snf5 leads to aberrant activation of the Hedgehog-Gli pathway Nature medicine 2010 16 12 1429 1433 11 Rahib L Smith BD Aizenberg R Rosenzweig AB Fleshman JM Matrisian LM Projecting cancer incidence and deaths to 2030: the unexpected burden of thyroid, liver, and pancreas cancers in the United States Cancer research 2014 74 11 2913 2921 24840647 12 Dimou A Syrigos K Saif MW Rationale for inhibition of the hedgehog pathway paracrine loop in pancreatic adenocarcinoma JOP : Journal of the pancreas 2011 12 1 1 5 21206093 13 Rajurkar M De Jesus-Monge WE Driscoll DR Appleman VA Huang H Cotton JL The activity of Gli transcription factors is essential for Kras-induced pancreatic tumorigenesis Proceedings of the National Academy of Sciences of the United States of America 2012 109 17 E1038 E1047 22493246 14 Nolan-Stevaux O Lau J Truitt ML Chu GC Hebrok M Fernandez-Zapico ME GLI1 is regulated through Smoothened-independent mechanisms in neoplastic pancreatic ducts and mediates PDAC cell survival and transformation Genes & development 2009 23 1 24 36 19136624 15 Marechal R Jean-Baptiste B Annabelle C Pieter D Jean Robert D Magali SM Sonic Hedgehog and Gli1 expression predict outcome in resected pancreatic adenocarcinoma Clinical cancer research : an official journal of the American Association for Cancer Research 2014 16 Kang Y Zhang R Suzuki R Li SQ Roife D Truty MJ Two-dimensional culture of human pancreatic adenocarcinoma cells results in an irreversible transition from epithelial to mesenchymal phenotype Laboratory investigation; a journal of technical methods and pathology 2015 95 2 207 222 25485535 17 Hwang RF Moore T Arumugam T Ramachandran V Amos KD Rivera A Cancer-associated stromal fibroblasts promote pancreatic tumor progression Cancer research 2008 68 3 918 926 18245495 18 Kutner RH Zhang XY Reiser J Production, concentration and titration of pseudotyped HIV-1-based lentiviral vectors Nature protocols 2009 4 4 495 505 19300443 19 Cerami E Gao J Dogrusoz U Gross BE Sumer SO Aksoy BA The cBio cancer genomics portal: an open platform for exploring multidimensional cancer genomics data Cancer discovery 2012 2 5 401 404 22588877 20 Gao J Aksoy BA Dogrusoz U Dresdner G Gross B Sumer SO Integrative analysis of complex cancer genomics and clinical profiles using the cBioPortal Science signaling 2013 6 269 pl1 23550210 21 Lang JY Hsu JL Meric-Bernstam F Chang CJ Wang Q Bao Y BikDD eliminates breast cancer initiating cells and synergizes with lapatinib for breast cancer treatment Cancer cell 2011 20 3 341 356 21907925 22 American Cancer Society Cancer Facts & Figures 2014 2014 23 Lauth M Bergstrom A Shimokawa T Tostar U Jin Q Fendrich V DYRK1B-dependent autocrine-to-paracrine shift of Hedgehog signaling by mutant RAS Nat Struct Mol Biol 2010 17 6 718 725 20512148 24 Canettieri G Di Marcotullio L Greco A Coni S Antonucci L Infante P Histone deacetylase and Cullin3-REN(KCTD11) ubiquitin ligase interplay regulates Hedgehog signalling through Gli acetylation Nature cell biology 2010 12 2 132 142 20081843 25 Nicholson TB Chen T Richard S The physiological and pathophysiological role of PRMT1-mediated protein arginine methylation Pharmacological research : the official journal of the Italian Pharmacological Society 2009 60 6 466 474 26 Thandapani P O'Connor TR Bailey TL Richard S Defining the RGG/RG motif Molecular cell 2013 50 5 613 623 23746349 27 Bedford MT Richard S Arginine methylation an emerging regulator of protein function Molecular cell 2005 18 3 263 272 15866169 28 Liao HW Hsu JM Xia W Wang HL Wang YN Chang WC PRMT1-mediated methylation of the EGF receptor regulates signaling and cetuximab response The Journal of clinical investigation 2015 29 Gayatri S Bedford MT Readers of histone methylarginine marks Biochimica et biophysica acta 2014 1839 8 702 710 24583552 30 Kim MP Evans DB Wang H Abbruzzese JL Fleming JB Gallick GE Generation of orthotopic and heterotopic human pancreatic cancer xenografts in immunodeficient mice Nature protocols 2009 4 11 1670 1680 19876027 31 Sasaki H Hui C Nakafuku M Kondoh H A binding site for Gli proteins is essential for HNF-3beta floor plate enhancer activity in transgenics and can respond to Shh in vitro Development 1997 124 7 1313 1322 9118802 32 Zhu H Lo HW The Human Glioma-Associated Oncogene Homolog 1 (GLI1) Family of Transcription Factors in Gene Regulation and Diseases Current genomics 2010 11 4 238 245 21119888 33 Xu XF Guo CY Liu J Yang WJ Xia YJ Xu L Gli1 maintains cell survival by up-regulating IGFBP6 and Bcl-2 through promoter regions in parallel manner in pancreatic cancer cells Journal of carcinogenesis 2009 8 13 19736394 34 Kasper M Schnidar H Neill GW Hanneder M Klingler S Blaas L Selective modulation of Hedgehog/GLI target gene expression by epidermal growth factor signaling in human keratinocytes Molecular and cellular biology 2006 26 16 6283 6298 16880536 35 Guendel I Carpio L Pedati C Schwartz A Teal C Kashanchi F Methylation of the tumor suppressor protein, BRCA1, influences its transcriptional cofactor function PloS one 2010 5 6 e11379 20614009 36 Cheng D Yadav N King RW Swanson MS Weinstein EJ Bedford MT Small molecule regulators of protein arginine methyltransferases The Journal of biological chemistry 2004 279 23 23892 23899 15056663 37 Kim MP Gallick GE Gemcitabine resistance in pancreatic cancer: picking the key players Clinical cancer research : an official journal of the American Association for Cancer Research 2008 14 5 1284 1285 18316544 38 Xin Y Shen XD Cheng L Hong DF Chen B Perifosine inhibits S6K1-Gli1 signaling and enhances gemcitabine-induced anti-pancreatic cancer efficiency Cancer chemotherapy and pharmacology 2014 73 4 711 719 24519751 39 Peng Z Ji Z Mei F Lu M Ou Y Cheng X Lithium inhibits tumorigenic potential of PDA cells through targeting hedgehog-GLI signaling pathway PloS one 2013 8 4 e61457 23626687 40 Rhim AD Oberstein PE Thomas DH Mirek ET Palermo CF Sastra SA Stromal elements act to restrain, rather than support, pancreatic ductal adenocarcinoma Cancer cell 2014 25 6 735 747 24856585 41 Lunardi S Muschel RJ Brunner TB The stromal compartments in pancreatic cancer: are there any therapeutic targets? Cancer letters 2014 343 2 147 155 24141189 42 Lauth M Bergstrom A Shimokawa T Toftgard R Inhibition of GLI-mediated transcription and tumor cell growth by small-molecule antagonists Proceedings of the National Academy of Sciences of the United States of America 2007 104 20 8455 8460 17494766 43 Coni S Di Magno L Canettieri G Determination of Acetylation of the Gli Transcription Factors Methods in molecular biology 2015 1322 147 156 26179046 44 Gore J Korc M Pancreatic cancer stroma: friend or foe? Cancer cell 2014 25 6 711 712 24937454 45 Xu Y An Y Wang X Zha W Li X Inhibition of the Hedgehog pathway induces autophagy in pancreatic ductal adenocarcinoma cells Oncol Rep 2014 31 2 707 712 24297612 46 Fu J Rodova M Roy SK Sharma J Singh KP Srivastava RK GANT-61 inhibits pancreatic cancer stem cell growth in vitro and in NOD/SCID/IL2R gamma null mice xenograft Cancer letters 2013 330 1 22 32 23200667

PMC005xxxxxx/PMC5135737.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 7906882 4987 J Nat Prod J. Nat. Prod. Journal of natural products 0163-3864 1520-6025 26841051 5135737 10.1021/acs.jnatprod.5b01014 NIHMS829552 Article Biochemometrics for Natural Products Research: Comparison of Data Analysis Approaches and Application to Identification of Bioactive Compounds Kellogg Joshua J. † Todd Daniel A. † Egan Joseph M. † Raja Huzefa A. † Oberlies Nicholas H. † Kvalheim Olav M. ‡ Cech Nadja B. *† † Department of Chemistry & Biochemistry, University of North Carolina Greensboro, Greensboro NC United States ‡ Department of Chemistry, University of Bergen, Bergen Norway * To whom correspondence should be addressed. Tel: 336-334-3017. Fax: 336-334-5402. nadja_cech@uncg.edu. 13 11 2016 3 2 2016 26 2 2016 03 12 2016 79 2 376386 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. A central challenge of natural products research is assigning bioactive compounds from complex mixtures. The gold standard approach to address this challenge, bioassay-guided fractionation, is often biased towards abundant, rather than bioactive, mixture components. This study evaluated the combination of bioassay-guided fractionation with untargeted metabolite profiling to improve active component identification early in the fractionation process. Key to this methodology was statistical modeling of the integrated biological and chemical datasets (biochemometric analysis). Three data analysis approaches for biochemometric analysis were compared, namely, partial least squares loading vectors, S-plots, and the selectivity ratio. Extracts from the endophytic fungi Alternaria sp. and Pyrenochaeta sp. with antimicrobial activity against Staphylococcus aureus served as test cases. Biochemometric analysis incorporating the selectivity ratio performed best in identifying bioactive ions from these extracts early in the fractionation process, yielding altersetin (3, MIC 0.23 μg/mL) and macrosphelide A (4, MIC 75 μg/mL) as antibacterial constituents from Alternaria sp. and Pyrenochaeta sp., respectively. This study demonstrates the potential of biochemometrics coupled with bioassay-guided fractionation to identify bioactive mixture components. A benefit of this approach is the ability to integrate multiple stages of fractionation and bioassay data into a single analysis. Natural products research has as its central goal the isolation and identification of bioactive constituent(s) from complex natural product mixtures. To achieve this, natural products chemists have developed a robust repertoire of techniques broadly termed “bioassay-guided fractionation”.1 Bioassay-guided fractionation is an iterative methodology that alternates between chemical fractionation and bioassays. With each stage of fractionation, the complexity of the mixture is reduced, and eventually the compound(s) responsible for the observed biological effect can be isolated and characterized. This methodology has long been the gold standard in natural products research, and has resulted in the discovery of critically important drugs, including camptothecin and taxol (paclitaxel),2,3 artemisinin,4 and vinblastine.5 Weller reported in 2012 that over 1,500 publications in ISI Web of Science employed bioassay-guided fractionation, with hundreds more citations using variants of the nomenclature.6 Despite the popularity and historical effectiveness of bioassay-guided fractionation, it has several limitations.7 The process tends to be biased towards dominant peaks in each extract or fraction and, as a result, bioactive constituents in low abundance can be overlooked.8 Furthermore, isolation of all trace constituents can be difficult, given that each chemical separation step witnesses a decrease in material. Finally, there is the potential to lose activity due to irreversible binding of mixture components to chromatographic resins or degradation during the separation process.9 In light of these limitations, new methods capable of focusing the isolation process on components most likely to be responsible for the desired biological effect are needed. Recently, there has been a great deal of interest in the application of untargeted metabolomics to study biologically active natural product mixtures.10–15 Metabolomics approaches are employed to profile multiple mixture components simultaneously, typically through the application of chromatographic analysis coupled to spectroscopic or spectrometric approaches (IR, UV, MS, or NMR detection). Such approaches can enable the detection of unstable compounds that would be lost upon purification, and consider all compounds together rather than as distinct fractions in series.7,16 Metabolomic profiling results in the generation of large data sets that include both major and minor components.11–13,15 Data-driven methods are needed to extract meaning from these complex chemical datasets, and multivariate statisticians and chemometricians have developed a number of strategies towards this goal.17,18 The most commonly employed tool in metabolomics data analysis is principal component analysis (PCA), in which a dataset is projected onto a series of latent variables, which are then mapped in two-dimensional space. Groupings of objects are discerned by their covariance, which is analyzed visually by the proximity of one object to another in the PCA scores plot.18 One limitation of metabolomics for studying natural product mixtures is the difficulty in tying identified metabolites to bioactive effects. If the end goal is determining which compounds are responsible for the biological activity of a mixture, comparing the chemical composition of different mixtures (the central goal of metabolomics) is not sufficient.14,19 There is a need to go beyond the metabolomics datasets, and to use biological assay data to inform their interpretation. To address this need is an even greater data analysis challenge than that faced in classical metabolomics, and requires the integration of both biological and chemical datasets. In 2006, chemometricians working to integrate chemical and biological data dubbed the field “biochemometrics.”20 The present report is concerned primarily with the development of effective approaches for applying biochemometrics to natural products drug discovery. Several approaches have been developed for correlating metabolite profiles with biological datasets (Table 1). Partial least squares (PLS) decomposes the spectral dataset (i.e., retention time and mass-to-charge (m/z) pairings), into uncorrelated latent variables. PLS differs from PCA in that it seeks to maximize the covariance of independent variables (spectral data from IR, UV, or MS analysis) with a dependent variable (i.e., biological activity).21 As an example of the effectiveness of this approach, Ali et al.22 utilized PLS analysis of NMR signal data to identify bioactive metabolites from marine sponges against the adenosine A1 receptor. In some cases, however, data interpretation may be difficult with PLS, because variables possessing large variance yet small correlation may mask other variables with low variance and high correlation to the dependent (response) variable. In addition, multiple PLS components are often needed to optimize the discrimination between response groups.23 Recently, Wiklund et al.24 described the S-plot as a means for interpreting orthogonal PLS (OPLS) predictive components. With an S-plot, the covariance and correlation loading variables are displayed graphically, which allows for visual identification of spectral variables that strongly correlate with a dependent biological activity variable.24 S-plots have been utilized several times for natural product research, most recently to discover immunomodulatory components from Phaleria nisidai25 and antidiabetic compounds of Cree medicinal plants.26 A limitation of the S-plot approach is that the large number of spectral variables can make visualization and interpretation of the data difficult. In addition, the S-plot relies only on the correlation and covariance of independent variables to the dependent variable, which can lead to false positives.27 As another strategy for interpreting biochemometric datasets, Kvalheim and Karstang28 developed the “target projection” component, wherein PLS components are transformed into a univariate metric that facilitates analysis and interpretation of correlative data. The variance explained by the target projection component can be calculated for each independent (spectral) variable and compared against the residual variance. The ratio between explained and residual variance of the spectral variables of the target-projection component, termed the selectivity ratio, represents a quantitative measure of each variable's power to distinguish between different groups. Variables with a high selectivity ratio have an excellent ability to separate bioactive and non-bioactive groupings. This approach has been utilized to identify clinical biomarkers from human spinal fluid (CSF) samples27 as well as for chemical fingerprint analysis of the herbal medicine Puerariae lobatae (Radix Puerariae).29 However, the selectivity ratio has not been applied to identify individual bioactive components of natural product mixtures. With this study, we compared three data analysis strategies (PLS, S-plot, and selectivity ratio) for integrating biological and chemical datasets from natural product mixtures. Our objective was to demonstrate which of these approaches would be most effective for distinguishing active and inactive compounds in the mixtures. As a case study, two endophytic fungi isolated from the botanical goldenseal [Hydrastis canadensis L. (Ranunculaceae)], were selected, namely, Alternaria sp. and Pyrenochaeta sp. Alternaria sp. was chosen because the extract from this fungus demonstrated marked antimicrobial activity in screens performed in our laboratory (data not shown), but dereplication using a UPLC-HRMS-MS/MS protocol30 identified only one primary chemical constituent, alternariol monomethyl ether (1), which is mildly active against Gram-positive bacteria. For Pyrenochaeta sp., the activity of the extract could not be correlated to known compounds in the dereplication library. Thus, these fungal extracts were a good example of mixtures containing unknown active compounds. The goal of our studies was to conduct biochemometric analysis on these fungi at an early stage of extraction and fractionation, and subsequently to verify the predictions of the biochemometric analysis with follow up isolation, structure elucidation, and biological evaluation. RESULTS AND DISCUSSION Biochemometric Analysis of Alternaria sp The first goal of these studies was to contrast various chemometric and biochemometric analysis techniques as applied to an extract from the fungus Alternaria sp. Towards this goal, chemometric profiling was first conducted on a crude extract from the Alternaria sp. fungus (AS-CR) and four fractions (AS-1 to AS-4). Untargeted metabolomic analysis of these fractions using ultra-performance liquid chromatography coupled to high-resolution mass spectrometry (UPLC-HRMS) yielded 472 total marker ions (unique retention time-m/z pairs), which were compared using principal component analysis (PCA) (Figure 1). The first three components of the PCA scores plot accounted for 93.09% of total variability of the model (component 1: 52.97%; component 2: 32.45%; component 3: 7.68%). The technical replicates (triplicate UPLC-HRMS analyses) of each sample are overlaid on the plot, indicating excellent repeatability of the chemical analysis (Figure 1). The AS-CR and AS-2 fractions group together, separated from the AS-1, AS-3, and AS-4 fractions (Figure 1), which indicates distinct chemical profiles between AS-2 and the fractions AS-1, -3, and -4. Bioactivity screening revealed complete inhibition of S. aureus strain SA1199 by the Alternaria sp. crude extract (AS-CR) as well as the second fraction (AS-2) (Table 2). At the 100 μg/mL level, the crude extract exhibited complete growth inhibition of Staphylococcus aureus, while fraction AS-2 evidenced near-complete inhibition (99.83 ± 0.01%) (Table 2). The fractions AS-1, AS-3, and AS-4 demonstrated negligible growth inhibition of S. aureus (< 1.0%). Pairing the antibacterial screening with high-resolution mass spectral data, the resulting biochemometric analytical matrix evidenced differences between the Alternaria sp. fractions based on their bioactivity. The internal cross-validated construction of the PLS model yielded four components, accounting for 100% of the independent (spectral) and dependent (bioactivity) block variation (component 1: 52.94% independent, 98.06% dependent; component 2: 10.60%, 1.79%; component 3: 29.08%, 0.16%; component 4: 7.38%, 0.00%). The PLS scores plot (Figure 2A) showed a similar clustering of fractions as the PCA analysis, with AS-CR and AS-2 separated graphically from the other, non-bioactive fractions. Also similar to the PCA analysis, the triplicate data points representing each sample are closely grouped together in the PLS scores plot, although the addition of biological variability causes a slight increase in the spread among replicates, as might be expected. PLS scores plots do not provide information regarding which specific chemical species contribute to the observed antibacterial bioactivity. To obtain this information, three distinct analytical methods were contrasted: the loadings plot from the PLS model, the multivariate transformed S-plot, and the selectivity ratio. Examination of the PLS loadings plot (Figure 2B) yielded three major metabolites that were shifted in the same direction as the bioactive fractions from the PLS scores plot (Figure 2A). Identities of these compounds were proposed using literature and high-resolution mass spectrometry data (Table 3) as: alternariol monomethyl ether (1), tenuazonic acid (2), and altersetin (3). The difference in location in the scores plot between 1, 2, and 3 was not sufficient to ascertain which of the three compounds was most responsible for the antibacterial activity of the fraction. The identifications of 1 and 3 were confirmed from the NMR spectra of the isolated compounds (Figures S2 and S4, Supporting Information, respectively) and NMR data were consistent with literature reports.31,32 Due to the low abundance and lack of bioactivity of fractions containing compound 2, this compound was not pursued for isolation, and identification is only tentative, based on matching accurate mass with literature values.29 The S-plot graphically displays the covariance and correlation of loading variables against the dependent variable as a scatter plot (Figure 2C). In an S-plot, the further a marker ion is from the origin, the greater its contribution is to the variance between bioactivity levels. For the Alternaria sp. biochemometric model, the upper right quadrant of the S-plot contributed the most to the differentiation of biologically-active versus inactive fractions, and compounds 1 and 2 were highlighted as possessing the greatest contribution to the observed bioactivity. The selectivity ratio produces a graphical representation in which the most abundant peaks correspond to marker ions that are most strongly associated with bioactivity (Figure 2D). From the selectivity ratio plot, the dominant marker ions were altersetin (3) and its sodium adduct, suggesting that this compound dominated the contributions to the antibacterial potency of the Alternaria sp. extract and fractions. In contrast to the PLS loadings plot (Figure 2B) and the S-plot (Figure 2C), alternariol monomethyl ether (1) and tenuazonic acid (2) were not among the most significant marker ions identified according to the selectivity ratio analysis (Figure 2D). Selectivity ratio analysis has an additional advantage of enabling the application of multiple independent variables, which facilitates interpretation by natural product chemists consistent with the type of instrumentation being employed. Utilizing mass spectrometry data (signal versus m/z) creates a plot similar to a mass spectrum (Figure 2D), where the x-axis is the m/z of the detected ion, and the y-axis represents how strongly associated that particular ion is with the biological activity.23 The use of chromatographic data (detector signal versus retention time) yields a similar selectivity ratio plot to that generated with mass spectrometric data, except that the x-axis represents retention time rather than m/z.33 Identification of Marker Compounds in Alternaria sp Additional purification of the most active Alternaria sp. fraction (AS-2) was conducted to investigate the accuracy of the predictions provided by the biochemometric analysis. Subfractions of AS-2 (coded AS-2-1 – AS-2-10) obtained with reversed-phase preparative scale HPLC revealed marked differences in both chemical makeup and bioactivity. UPLC-HRMS analysis of individual subfractions showed that subfraction AS-2-3 was 94% enriched in tenuazonic acid (2) (Figure 3B), subfraction AS-2-7 was 94% alternariol monomethyl ether (1) (Figure 3C), and altersetin (3) was isolated in subfraction AS-2-9 at 85% purity (Figure 3D). The other subfractions contained insignificant quantities of these compounds. From the antibacterial screening protocol, subfraction AS-2-3 only displayed 0.01 ± 0.02% inhibition of S. aureus SA1199, AS-2-7 yielded 0.08 ± 0.01% growth inhibition, and AS-2-9 inhibited bacterial growth by 99.5 ± 0.01%. These data indicate that 3 was, indeed, the most active compound from the original extract. The application of the selectivity ratio made it possible to distinguish between active and inactive ions at an early stage of the analysis (with just the crude extract and four fractions). The selectivity ratio correctly predicted that ion 3 was the most active, while the S-plot and PLS loading vectors attributed activity to ions 1, 2, and 3. Follow up antimicrobial assays on pure compounds 1 and 3 supported the data in Figure 4, indicating that 1 is weakly active against S. aureus, with a minimum inhibitory concentration (MIC) of 275 μM (75 μg/mL), while 3 possesses pronounced activity (MIC 0.59 μM (0.23 μg/mL)) (Table 4). Antimicrobial activity was also observed for both of these compounds against methicillin resistant Staphylococcus aureus (MRSA) (Table 4). Refined Biochemometric Analysis Multivariate statistical modeling increases in accuracy and precision as the sample size (number of objects) increases;33 thus, it was hypothesized that further fractionation of the bioactive Alternaria sp. fraction AS-2 would provide enhanced separation between the active and inactive marker ions in the biochemometric analysis. The incorporation of subfractions (AS-2-1 through AS-2-10) of the original active fraction (AS-2) into the biochemometric matrix yielded a more refined statistical model and subsequent analysis. The PLS scores plot (Figure 4A) distinguished the active fractions and subfractions (AS-CR, AS-2, AS-2-9, and AS-2-10) from the non-active fractions and subfractions. The increased analytical power of adding subfractions to the biochemometric matrix was reflected in changes in both the PLS loading plot (Figure 4B) and the S-plot (Figure 4C) compared to the initial matrix analysis (Figure 2B and 2C, respectively). In the expanded data matrix, both plots yielded altersetin (3) as the principal marker ion that contributed to the observed antibacterial activity, while the signals for compounds 1 and 2 were shifted downward towards a region of low covariance and correlation. The selectivity ratio (Figure 4D) maintained altersetin (3) as the principal bioactive marker ion. The inclusion of the subfractions in the selectivity ratio analysis effectively increased the abundance of the altersetin signals relative to those of other ions. With the incorporation of 15 objects spanning three different degrees of chemical complexity - crude extract, fractions, and subfractions - the PLS loadings plot and S-plot could be employed to correctly identify 3 as the antibacterial compound from the mixture. However, the selectivity ratio analysis enabled identification of the active mixture components at an earlier stage of the isolation process. Application of Biochemometrics to Pyrenochaeta sp To confirm the analytical capability of the selectivity ratio, a second bioactive endophytic fungus (Pyrenochaeta sp.) was analyzed via the same methodology. Biochemometric profiling was conducted on the crude Pyrenochaeta sp. extract (PS-CR) and four fractions (PS-1 to PS-4). UPLC-HRMS analysis yielded 659 marker ions, and bioactivity screening highlighted PS-4 as the most bioactive fraction (Table 5). The PLS scores plot (Figure 5A) separated the two active samples (PS-CR and PS-4) distinctly from the inactive samples. For Pyrenochaeta sp., the PLS loadings plot (Figure 5B) and the selectivity ratio (Figure 5C) revealed macrosphelide A (4) as the principal bioactive constituent. Subsequent isolation efforts confirmed the presence of macrosphelide A through high-resolution mass spectrometry and 1H and 13C NMR (Figure S6, Supporting Information),34 and its MIC value against S. aureus was determined to be 219 μM (75 μg/mL) (Table 4). In summary, the identification of bioactive compounds without the need for multiple bioactivity-guided isolation steps remains an important goal to improve the efficiency and productivity of natural product discovery programs. The study presented herein has utilized multivariate statistical modeling coupled to the selectivity ratio (a univariate metric) to reveal compounds from a complex chemical profile that were responsible for the observed antibacterial bioactivity. Application of this biochemometric approach has led to the identification of a minor compound (altersetin, 3) from Alternaria sp. with potent antibacterial activity against both S. aureus and MRSA. The same approach was applied with a second endophytic fungus, Pyrenochaeta sp., which revealed the bioactive compound macrosphelide A. Although these are both known compounds, they were not identified in the crude extracts because they were not included in the database of experimental UPLC-MS data used for dereplication.30 However, once it was determined (based on biochemometric analysis) that these ions were likely responsible for the biological activity of the extracts, their structures could be rapidly predicted by comparison of UPLC-MS data with published literature. Such literature searches would have been inefficient and impractical had they been conducted for all of the unique features identified by UPLC-HRMS in the Alternaria sp. and Pyrenochaeta sp. extracts (472 and 659 ions, respectively). Importantly, by using biochemometrics to integrate chemical and biological datasets, the bioactive extract compounds (3 and 4) were identified as active very early in the isolation process, after just one stage of fractionation. For the purpose of the study presented here, subsequent fractionation and isolation steps were then conducted to confirm the predictions of the biochemometric analysis. In future studies, such follow up isolation efforts might not be pursued if it was determined that the putative active compounds were of known structure and biological activity. In this way, biochemometrics could serve as a useful tool in dereplication efforts. The inclusion of biochemometrics in the dereplication process could prevent the rejection of an extract for further study based on the presence of a known (but inactive) compound, a potential pitfall of dereplication approaches that rely exclusively on chemical data. Additionally, as was the case with Alternaria sp., biochemometric analysis can point to the biological importance of a seemingly minor extract component, enabling focused efforts to rapidly solve its structure. The data presented here suggest that selectivity ratio analysis, which made better predictions than other data analysis procedures early in the fractionation process, could be a particularly effective tool for integrating biological and chemical datasets as part of dereplication efforts. Another potentially important application of the biochemometrics approach is for integrating the chemical and bioassay data obtained from multiple fractionation steps. In the process of bioassay guided fractionation, each stage of separation and bioassay data is typically considered in isolation from previous isolation steps. Using biochemometrics, it was possible to develop a model in which the active mixture components were predicted based on the data from several stages of fractionation in combination. Indeed, the study described herein shows that adding successive stages of fractionation and bioassay to the biochemometric analysis (original extract plus fractions and sub-fractions) improves the quality of the resulting selectivity plot. A future goal of our work is to apply biochemometrics using selectivity ratios to identify active components from more complex mixtures such as botanical extracts. Towards this goal, it may be necessary to conduct additional stages of purification to obtain sufficient sample size (number of objects) to accurately predict the compounds responsible for the observed biological activity. A highly complex extract could be fractionated and subjected to biochemometric analysis repeatedly, with each successive step of the fractionation and bioassay added into the dataset until a quality selectivity ratio plot could be obtained. EXPERIMENTAL SECTION General Experimental Procedures NMR spectra were acquired with a JEOL ECA-400 spectrometer (400 MHz) using DMSO-d6. Optical rotations were obtained using a Rudolph Research Autopol III polarimeter (Rudolph Research Analytical, Hackettstown, NJ, USA). UPLC-HRESIMS data were acquired using a Q Extractive Plus quadrupole-orbitrap mass spectrometer (Thermo Scientific, Waltham, MA, USA) with an electrospray ionization source coupled to an Acquity UPLC system (Waters, Milford, MA, USA). To collect UPLC-HRESIMS data, each sample was re-suspended in MeOH to a concentration of 1 mg/mL, and triplicate 3 μL injections of each sample were performed. The samples were eluted from the column (Acquity UPLC BEH C18 1.7μm, 2.1 × 50 mm, Waters) at a flow rate of 0.3 mL/min using the following binary gradient with solvent A consisting of H2O (0.1% formic acid added) and solvent B consisting of CH3CN (0.1% formic acid added): initial isocratic composition of 95:5 (A:B) for 1.0 min, increasing linearly to 0:100 over 20 min, followed by an isocratic hold at 0:100 for 1 min, gradient returned to starting conditions of 95:5 for 2 min, and held isocratically again for 1 min. The mass spectrometer was operated in the positive ionization mode over a scan range of 150–2000 with the following settings: capillary voltage set at 5 V, capillary temperature set at 300 °C, tube lens offset set at 35 V, spray voltage set at 3.80 kV, sheath gas flow set at 35, and auxiliary gas flow set at 20. Flash chromatography separations were accomplished using an automated CombiFlash RF system (Teledyne-Isco, Lincoln, NE, USA) and monitored with a PDA detector and as an evaporative light scattering detector. HPLC separations were performed on a Varian HPLC system (Agilent Technologies, Santa Clara, CA, USA) with Galaxie Chromatography Workstation software (version 1.9.3.2, Agilent Technologies). Analytical and preparative-scale HPLC separations employed a Gemini-NX C18 column (5 μm, 110 Å, 250 × 4.60 mm (analytical) or 250 × 21.20 mm (preparative); Phenomenex, Torrance, CA, USA). Unless otherwise noted, all chemicals were of spectroscopic or microbiological grade and obtained from Sigma-Aldrich (St. Louis, MO, USA). Plant Collection and Fungal Isolation Individual, asymptomatic goldenseal (Hydrastis canadensis L.) plants were collected in July 2010 from William Burch in Hendersonville, North Carolina (N 35° 24.2770, W 082° 20.9930). A voucher specimen was deposited at the herbarium of the University of North Carolina at Chapel Hill (NCU583414) and authenticated by Dr. Alan S. Weakly. Isolation of fungal endophytes was performed using methods outlined previously.35,36 Two strains: G28 (isolated from seeds), and G41 (isolated from leaf segments) were used in the present study. Axenic fungal cultures are maintained at 9 °C at the University of North Carolina at Greensboro, Department of Chemistry and Biochemistry Fungal Culture Collection. Identification of Fungal Isolates For molecular identification of fungal endophytes isolated from goldenseal, the internal transcribed spacer region of the 5.8S ribosomal RNA gene (ITS1-5-5S-ITS2) was sequenced using methods described previously.35–38 Based on BLAST search conducted with published ITS data in NCBI GenBank, strain G28 was identified as an Alternaria sp. (Pleosporales, Dothideomycetes), while strain G41 was identified as a Pyrenochaeta sp. (Pleosporales, Dothideomycetes), using cut off proxies for ITS sequence similarity outlined previously.36 The sequences from strains utilized in the present study were deposited in GenBank under accession numbers KT825854 (strain G28) and KT825855 (strain G41). Solid-state Fermentation of Fungal Cultures For chemical extraction, the fungal strains utilized in this study were grown on rice a medium.39 Briefly, seeds cultures were started on the liquid medium composed of 2% soy peptone, 2% dextrose, and 1% yeast extract (YESD). The seed culture was grown for 7 days at 22 °C with agitation, and subsequently transferred to 10 g of rice autoclaved with 25 mL of water in a 250 mL Erlenmeyer flask for screener cultures. For large-scale production of fungal cultures, four 250 mL Erlenmeyer flasks were inoculated using one seed culture for each flask. All rice cultures were allowed to grow for approximately 14–21 days prior to extraction. Extraction and Isolation Cultures of Alternaria sp. (AS) and Pyrenochaeta sp. (PS) were extracted following the established procedure.40 Briefly, to each culture flask, 60 mL of 1:1 MeOH–CHCl3 was added, chopped, and shaken overnight (~20 h) at ~100 rpm at room temperature. The sample was vacuum-filtered, 90 mL of CHCl3 and 150 mL of H2O were added, and the mixture was stirred for 30 min. This mixture was then transferred into a separatory funnel and the bottom (CHCl3) layer collected. The CHCl3 layer was evaporated to dryness, then dissolved in 100 mL of 1:1 MeOH–CH3CN and 100 mL of hexanes. The biphasic solution shaken in a separatory funnel, and the bottom layer drawn off and evaporated to yield the crude extract (CR). First-stage separations of the crude extract were conducted with normal-phase flash chromatography on a CombiFlash RF system with a 4 g silica gel column at 18 mL/min flow rate with a 40 min hexane-CHCl3-MeOH gradient, which yielded four fractions (AS-1 to AS-4 and PS-1 to PS-4, respectively) pooled based on LC-UV chromatograms. Active fractions were subjected to a second stage of purification using a reversed-phase preparative HPLC with a Gemini NX C18 column at a 21.20 mL/min flow rate. A linear CH3CN-H2O (both with 0.1% formic acid) gradient starting from 40:60 to 100:0 over 15 min was employed, with fractions collected every 0.5 min and pooled based on both UV and evaporative light scattering detector (ELSD) chromatograms. The Alternaria sp. (AS) extract yielded compounds 1 and 3, while 4 was isolated from the Pyrenochaeta sp. (PS) extract. Compound 2, tenuazonic acid, was putatively identified based upon its high resolution mass (m/z 198.1134 [M+H]+, calcd for C10H16NO3+, 198.1130), but was not present in sufficient quantities for isolation and confirmation of identity. Alternariol monomethyl ether (1): white solid, HRESIMS m/z 273.0756 [M+H]+ (calcd for C15H13O5+, 273.0763); 1H NMR (400 MHz DMSO-d6) and 13C NMR (100 MHz DMSO-d6), chemical shifts were in agreement with literature values31 and are provided as Supporting Information (Table S1 and Figure S2, Supporting Information). Altersetin (3): pale brown solid, HRESIMS m/z 400.2480 [M+H]+ (calcd for C24H34NO4+, 400.2488); 1H NMR (400 MHz DMSO-d6) and 13C NMR (100 MHz DMSO-d6), chemical shifts were in agreement with literature values31 and are provided as Supporting Information (Table S2 and Figure S4, Supporting Information). Macrosphelide A (4): yellow solid, [α]D20 = +84 (c 0.60, MeOH) HRESIMS m/z 343.1383 [M+H]+ (calcd for C16H23O8+, 343.1393); 1H NMR (400 MHz CDCl3) and 13C NMR (100 MHz CDCl3), chemical shifts were in agreement with literature values34 and are provided as Supporting Information (Table S3 and Figure S6, Supporting Information). Antibacterial Assay Antibacterial activity was assessed via growth inhibition of a laboratory strain of Staphylococcus aureus (strain SA1199)41 and methicillin-resistant S. aureus (MRSA USA300 LAC strain AH1263).42 Cultures were grown from a single colony isolate of each strain to log-phase in Müeller Hinton Broth (MHB) and plated at a final density of 1.0×106 CFU/mL. For screening, samples were assayed in triplicate at a concentration of 10 μg/mL and 100 μg/mL. Samples were dissolved in 1:1 EtOH-DMSO (v/v) and diluted in Mueller Hinton broth (MHB) to achieve the appropriate concentration, with ethanol and DMSO concentrations <2%. The positive control used for the screening procedure was chloramphenicol, at the same concentrations as the samples (10 μg/mL and 100 μg/mL). Vehicle was 2% 1:1 EtOH-DMSO. Each well was inoculated with bacterial culture, and incubated at 37 °C for 24 h. Minimum inhibitory concentration (MIC) was measured according to Clinical Laboratory Standards Institute (CLSI) standard procedures.43 Briefly, extracts or purified berberine (positive control, a known antibacterial compound from the host plant of these endophytic species, Hydrastis canadenensis L.)44 were added to 96-well plates in triplicate at concentrations ranging from 2.3 to 300 μg/mL in MHB. Vehicle (2% DMSO) served as the negative control, and DMSO content was fixed at 2% in all wells. Absorbance at 600 nm was measured after 24 h using a Synergy H1 microplate reader (Biotek, Winooski, VT, USA). The minimum inhibitory concentration (MIC) was defined as the concentration at which there was no statistically significant difference between the treatment and vehicle control. The absorbance for replicate wells containing all assay components except bacteria was subtracted from the absorbance of assay wells. Biochemometric Analysis Triplicate LC–MS datasets for each sample were individually analyzed, aligned and filtered with MZmine 2.17 software (http://mzmine.sourceforge.net/).45 Peak detection in MZmine was achieved as follows: m/z values were detected within each spectrum above a baseline, and a chromatogram was constructed for each of the m/z values that spanned longer than 0.1 min, and finally, deconvolution algorithms were applied to each chromatogram to recognize the individual chromatographic peaks. The parameters were set as follows for peak detection: noise level (absolute value) at 1 × 107, minimum peak duration 0.5 s, tolerance for m/z variation 0.05 and tolerance for m/z intensity variation 20%. Deisotoping, peak list filtering, and retention time alignment algorithm packages were employed to refine peak detection. Finally, the join align algorithm compiled a peak table according to the following parameters: the balance between m/z and retention time was set at 10.0 each, m/z tolerance was set at 0.05, and retention time tolerance size was defined as 2 min. The spectral data matrix (comprised of m/z, retention time, and peak area for each peak) was imported to Excel (Microsoft, Redmond, WA, USA) and merged with the bioactivity data set (at 100 μg/mL concentration) to form a final biochemometric analytical matrix. Triplicate datasets were included in the analysis for each sample, which consisted of three separate bioassay measurements and three separate UPLC-HRMS analyses for the same extract or fraction. Biochemometric analysis was performed using Sirius version 9.0 (Pattern Recognition Systems AS, Bergen, Norway).28,33 Initially, transformation from heteroscedastic to homoscedastic noise was carried out by a 4th root transform of the spectral variables. An internally cross-validated PLS model was constructed using 100 iterations, at a significance level of 0.05. Selectivity ratios from the final PLS model were calculated using algorithms internal to Sirius. Supplementary Material supporting information ACKNOWLEDGEMENTS This work was supported by the National Center for Complementary and Integrative Health, National Institutes of Health (grant 1R01 AT006860), and by a Biotechnology Research Grant (2011-BRG-1206) from the North Carolina Biotechnology Center. Mass spectrometry data were collected in the Triad Mass Spectrometry Facility. We thank Sochima Anika and Dr. Amninder Kaur for technical assistance and Vincent Sica for assistance with manuscript editing. Figure 1 Principal Component Analysis (PCA) scores plot of Alternaria sp. crude extract (AS-CR) and fractions AS1 – AS4, drawn with Hotelling's 95% confidence ellipse. All fractions were run in triplicate, and the resulting 472 marker ions were used to compute differences in mycochemical composition. Figure 2 Marker ion selection from a biochemometric dataset. The biochemometric dataset was obtained from the mass spectral data coupled with bacterial growth inhibition data (against S. aureus SA1199) at a concentration of 100 μg/mL (Table 2). (A) Partial least squares (PLS) scores plot, showing the grouping of bioactive and non-bioactive fractions from Alternaria sp. (AS-CR and AS-1 – AS-4). Each fraction was analyzed in triplicate via UPLC-MS and was subjected to triplicate biological assays. Thus, the replicate datapoints represent both biological and technical variability. (B) Loadings plot from the PLS analysis of biochemometric data. Variables located in the same region in the loadings plot (B) as the bioactive groups AS-CR and AS-2 in the scores plot (A) have the highest positive correlation with the dependent variable (bioactivity). Thus, three ions corresponding to alternariol monomethyl ether (1), tenuazonic acid (2), and altersetin (3) were identified from visual analysis of the loadings plot as potentially most bioactive. (C) S-plot from PLS model of antibacterial activity of Alternaria sp. extract and fractions. The upper right quadrant are the peaks with highest correlation to bioactivity, and ions 1, 2, and 3 were also identified from the S-plot. (D) The selectivity ratio analysis of the PLS model data. The ratio relates the explained variance of the variable to the residual variance. Higher values (taller lines) represent a more significant contribution to the observed bioactivity. The selectivity ratio indicates compound 3 to have the highest activity, and does not find strong correlation for compounds 1 and 2. Figure 3 UPLS-HRMS chromatograms of fraction AS-2 (A), along with selected subfractions AS-2-3 (B), AS-2-7 (C), and AS-2-9 (D) representing the semi-pure fractions of tenuazonic acid (2), alternariol monomethyl ether (1), and altersetin (3), respectively. Figure 4 Marker ion selection from the post-fractionation biochemometric dataset of Alternaria sp. The biochemometric dataset was obtained from the triplicate mass spectral data coupled with bacterial growth inhibition data (against S. aureus SA1199) at a concentration of 100 μg/mL. (A) Partial least squares (PLS) scores plot, showing the grouping of bioactive and inactive fractions from Alternaria sp. (AS-CR, AS-1 – AS-4, and AS-2-1 – AS-2-10). Each fraction was analyzed in triplicate, as shown in the scores plot. (B) Loadings plot from the PLS analysis of biochemometric data. Variable 3 was the most correlated to bioactivity, as implied by being shifted in the same direction as the bioactive samples in the scores plot. (C) S-plot from the larger PLS model of antibacterial activity of Alternaria sp. extract, fractions, and subfractions. The marker ion for 3 is distinctly separate from the others, indicating its greater contribution to the bioactivity (D) The selectivity ratio analysis of the more comprehensive PLS model data. Similar to the initial selectivity ratio analysis (Figure 2D), ion 3 displays the highest selectivity ratio. Figure 5 Identification of the bioactive principle from Pyrenochaeta sp. from the biochemometric dataset. The biochemometric dataset was obtained from the triplicate mass spectral data coupled with growth inhibition data against S. aureus (SA1199) at a concentration of 100 μg/mL. (A) The partial least squares (PLS) scores plot shows the grouping of bioactive and inactive fractions from Pyrenochaeta sp. (PS-CR, PS-1 – PS-4). Each fraction was analyzed in triplicate, as shown in the scores plot. (B) Loadings plot from the PLS analysis of biochemometric data. The ion for macrosphelide A (4) was the most correlated with the bioactive samples in the scores plot. (C) The selectivity ratio analysis of the PLS model data. Table 1 Summary of Data Analysis Methods for Biochemometric Data Sets. method methodology applications analysis output principal component analysis (PCA) • map objects and variables onto latent variables separately • identify correlations within groupings • outliers • quality control • object diversity • scores plot: summary of objects • loadings plot: summary of variables partial least squares (PLS) • incorporate objects and variables for predictive modeling • discriminating between groups • biomarker identification • scores plot: summary of objects • loadings plot: summary of variables s-plot • combine modeled covariance and correlation from pls in a scatter plot • same as pls • low correlation/intensity variables are close to origin • highly correlated variables are distanced from origin selectivity ratio • ratio of explained (predictive) and residual (uncorrelated) variance • variance developed from univariate “target projection” • variable discrimination • biomarker identification • x-axis: independent variables (spectral data, retention time, etc.) • single variable for identifying highly-correlated peaks Table 2 Antimicrobial Activity of Alternaria sp. (AS) Crude Extract (CR) and Fractions AS-1-AS-4.a sample S. aureus growth inhibition (%) chloramphenicolb 98.3 ± 0.4% AS-CR 100 ± 1% AS-1 0.00 ± 0.03% AS-2 99.83 ± 0.01% AS-3 0.14 ± 0.02% AS-4 0.76 ± 0.01% a Growth inhibition of S. aureus strain SA1199 relative to the vehicle control as measured by OD600. AS samples measured at a concentration of 100 μg/mL. Data presented as mean of triplicate analyses ± SEM. b Chloramphenicol functioned as the positive control. Table 3 Identification of Bioactive Marker Ions from Alternaria Sp. marker ion ion (molecular formula, δ (ppm)) adducts and fragments (molecular formula, δ (ppm)) tentative identification 1 273.0756 [M+H]+ (C15H13O5, 0.5) 255.0699 [M+H-H2O]+ (C15H11O4, 4.2) Alternariol monomethyl ethera 2 198.1124 [M+H]+ (C10H17NO3, 0.4) Tenuazonic acida 3 400.2480 [M+H]+ (C24H34NO4, 0.6) 422.2292 [M+Na]+ (C24H33NO4Na, 1.6) Altersetina 382.2375 [M+H-H2O]+ (C24H32NO3, 0.7) a Previously reported from cultured Alternaria spp.31 Table 4 Antimicrobial Activity of Isolated Compounds from Alternaria Sp. and Pyrenochaeta sp.a sample MIC S. aureus MIC MRSA μM μg/mL μM μg/mL berberine (+ control) 446 150 446 150 alternariol monomethyl ether (1) 275 75 NDb ND altersetin (3) 0.59 0.23 4.67 1.9 macrosphelide A (4) 219 75 ND ND a Minimum inhibitory concentrations (MIC) against S. aureus strain SA1199 and a strain of MRSA (USA300 LAC strain AH1263) are presented as mean of triplicate analyses. Berberine functioned as the positive control. b ND - not detected Table 5 Antimicrobial Activity of Pyrenochatea sp. (PS) Crude Extract (CR) and Fractions PS-1-PS-4.a sample S. aureus growth inhibition (%) chloramphenicolb 99.2 ± 0.2% PS-CR 78.2 ± 2.4% PS-1 40.0 ± 1.2% PS-2 36.3 ± 1.5% PS-3 26.4 ± 4.7% PS-4 95.8 ± 1.7% a Growth inhibition of S. aureus strain SA1199 is displayed as percent growth inhibition normalized to the vehicle control as measured by OD600. Pyrenochatea sp. samples were measured at a concentration of 100 μg/mL. Data presented as mean of triplicate analyses ± SEM. b Chloramphenicol functioned as the positive control. REFERENCES (1) Kinghorn AD Fong HHS Farnsworth NR Mehta RG Moon RC Moriarty RM Pezzuto JM Curr. Org. Chem 1998 2 597 612 (2) Wall ME Wani MC J. Ethnopharm 1996 51 239 253 (3) Oberlies NH Kroll DJ J. Nat. Prod 2004 67 129 135 14987046 (4) Tu Y Nat. Med 2011 17 1217 1220 21989013 (5) Noble RL Biochem. Cell Biol 1990 68 1344 1351 2085431 (6) Weller MG Sensors 2012 12 9181 9209 23012539 (7) Prince EK Pohnert G Anal. Bioanal. Chem 2010 396 193 197 19813005 (8) Inui T Wang Y Pro SM Franzblau SG Pauli GF Fitoterapia 2012 83 1218 1225 22766306 (9) Qui F Cai G Jaki BU Lankin DC Franzblau SG Pauli GF J. Nat. Prod 2013 76 413 419 23356207 (10) Clendinen CS Pasquel C Arjredini R Edison AS Anal. Chem 2015 87 5689 5706 (11) Kleigrewe K Almaliti J Tian IY Kinnel RB Korobeynikov A Monroe EA Duggan BM Di Marzo V Sherman DH Dorrestein PC Gerwick L Gerwick WH J. Nat. Prod 2015 78 1671 1682 26149623 (12) Wu C Du C Gubbens J Choi YH van Wezel GP J. Nat. Prod 2015 78 2355 2363 26438963 (13) Cortina NS Krug D Plaza A Revermann O Müller R Angew. Chem. Int. Edit 2012 51 811 816 (14) Sidebottom AM Johnson AR Karty JA Trader DJ Carlson EE ACS Chem. Biol 2013 8 2009 2016 23777274 (15) Hou Y Braun DR Michel CR Klassen JL Adnani N Wyche TP Bugni TS Anal. Chem 2012 84 4277 4283 22519562 (16) Yuliana ND Khatib A Choi YH Verpoorte R Phytother. Res 2010 25 157 169 (17) Okada T Afendi FM Katoh A Hirai A Kanaya S Chandra S Lata H Varma A Biotechnology for Medicinal Plants 2013 413 438 Springer-Verlag Berlin (18) Rajalahti T Kvalheim OM Int. J. Pharm 2011 417 280 290 21335075 (19) Choi YH Kim HK Linthorst HJM Hollander JG Lefeber AWM Erkelens C Nuzillard J Verpoorte R J. Nat. Prod 2006 69 742 748 16724833 (20) Martens H Bruun SW Adt I Sockalingum GD Kokhler A J. Chemometrics 2006 20 402 417 (21) Cox DG Oh J Keasling A Colson KL Hamann MT Biochim. Biophys. Acta 2014 1840 3460 3474 25151044 (22) Ali K Iqbal M Yuliana ND Lee YJ Park S Han S Lee JW Lee HS Verpoorte R Choi YH Metabolomics 2013 9 778 785 (23) Rajalahti T Arneberg R Kroksveen AC Berle M Myhr K Kvalheim OM Anal. Chem 2009 81 2581 2590 19228047 (24) Wiklund S Johansson E Sjöström L Mellerowicz EJ Edlund U Shockcor JP Gottfries J Moritz T Trygg J Anal. Chem 2008 80 115 122 18027910 (25) Kulakowski DM Wu S-B Balick MJ Kennelly EJ J. Chromatogr. A 2014 1364 74 82 25218635 (26) Shang N Saleem A Musallam L Walshe-Roussel B Badawi A Cuerrier A Arnason JT Haddad PS PLoS One 2015 10 e0135721 26263160 (27) Rajalahti T Arneberg R Berven FS Myhr K Ulvik RJ Kvalheim OM Chemom. Intell. Lab. Syst 2009 95 35 48 (28) Kvalheim OM Karstang TV Chemom. Intell. Lab. Syst 1989 7 39 51 (29) Chau F Chan H Cheung C Xu C Liang Y Kvalheim OM Anal. Chem 2009 81 7217 7225 19634860 (30) El-Elimat T Figueroa M Ehrmann BM Cech NB Pearce C Oberlies NH J. Nat. Prod 2013 76 1709 1716 23947912 (31) Hellwig V Grothe T Mayer-Bartschmid A Endermann R Geschke F Henkel T Stadler M J Antibiot 2002 55 881 892 12523821 (32) Tan N Tao Y Pan J Wang SQ Xu F She Z Lin Y-J Jones EBG Chem. Nat. Compd 2008 44 296 300 (33) Kvalheim OM Chan H Benzie IFF Szeto Y Tzang AH Mok DK Chau F Chemom. Intell. Lab. Syst 2011 107 98 105 (34) Takamatsu S Kim Y-P Hayashi M Hiraoka H NAtori M Komiyama K Omura S Journal of Antibiotics 1995 49 95 98 (35) Figueroa M Jarmusch AK Raja HA El-Elimat T Kavanaugh JS Horswill AR Cooks RG Cech NB Oberlies NH J. Nat. Prod 2014 77 1351 1358 24911880 (36) Raja HA Kaur A El-Elimat T Figueroa M Kumar R Deep G Agarwal R Faeth SH Cech NB Oberlies NH Mycology 2015 6 8 27 26000195 (37) El-Elimat T Raja HA Day CS Chen W-L Swanson SM Oberlies NH J. Nat. Prod 2014 77 2088 2098 25093280 (38) Raja HA El-Elimat T Oberlies NH Shearer CA Miller AN Tanaka K Hashimoto A Fournier J Mycologia 2015 107 845 862 26315030 (39) VanderMolen KM Raja HA El-Elimat T Oberlies NH AMB Express 2013 3 71 71 24342059 (40) El-Elimat T Raja HA Graf TN Faeth SH Cech NB Oberlies NH J. Nat. Prod 2014 77 193 199 24456525 (41) Kaatz GW Seo SM Antimicrob. Agents Chemother 1995 39 2650 2655 8592996 (42) Junio HA Todd DA Ettefagh KA Ehrmann BM Kavanaugh JS Horswill AR Cech NB J. Chromatog. B 2013 930 7 12 (43) National Committee for Clinical Laboratory Standards Medhods for dilution antimicrobial susceptibility tests for bacteria that grow aerobically 1993 Villanova, PA 3rd edition (44) Junio HA Sy-Cordero AA Ettefagh KA Burns JT Micko KT Graf TN Richter SJ Cannon RE Oberlies NH Cech NB J. Nat. Prod 2011 74 1621 1629 21661731 (45) Pluskal T Castillo S Villar-Briones A Orešič M BMC Bioinformat 2010 11 395

PMC005xxxxxx/PMC5136293.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 101525249 37848 Nanoscale Nanoscale Nanoscale 2040-3364 2040-3372 26935414 5136293 10.1039/c6nr00398b NIHMS765684 Article Anti-atherogenic Effect of Trivalent Chromium-loaded CPMV Nanoparticles in Human Aortic Smooth Muscle Cells under Hyperglycemic Conditions in vitro Ganguly Rituparna 12 Wen Amy M. 3 Myer Ashley B. 1 Czech Tori 1 Sahu Soumyadip 12 Steinmetz Nicole F. 34567 Raman Priya Dr. praman@neomed.edu 12 1 Department of Integrative Medical Sciences, Northeast Ohio Medical University, 4209 State Route 44, Rootstown, OH 44272-0095, USA 2 School of Biomedical Sciences, Kent State University, Kent, OH, USA 3 Department of Biomedical Engineering, 10990 Euclid Avenue, Case Western Reserve University, Cleveland OH, USA 4 Department of Radiology, 10990 Euclid Avenue, Case Western Reserve University, Cleveland OH, USA 5 Department of Materials Science and Engineering, 10990 Euclid Avenue, Case Western Reserve University, Cleveland OH, USA 6 Department of Macromolecular Science and Engineering, 10990 Euclid Avenue, Case Western Reserve University, Cleveland OH, USA 7 Case Comprehensive Cancer Center, 10990 Euclid Avenue, Case Western Reserve University, Cleveland OH, USA 3 6 2016 28 3 2016 03 12 2016 8 12 65426554 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. Atherosclerosis, a major macrovascular complication associated with diabetes, poses tremendous burden on national health care expenditure. Despite extensive efforts, cost-effective remedies are unknown. Therapies for atherosclerosis are challenged by lack of targeted drug delivery approach. Toward this goal, we turn to a biology-derived drug delivery system utilizing nanoparticles formed by the plant virus, Cowpea mosaic virus (CPMV). The aim herein is to investigate the anti-atherogenic potential of the beneficial mineral nutrient, trivalent chromium, loaded CPMV nanoparticles in human aortic smooth muscle cells (HASMC) under hyperglycemic conditions. A non-covalent loading protocol is established yielding CrCl3-loaded CPMV (CPMV-Cr) carrying 2,000 drugs per particle. Using immunofluorescence microscopy, we show that CPMV-Cr is readily uptaken by HASMC in vitro. In glucose (25 mM)-stimulated cells, 100 nM CPMV-Cr inhibits HASMC proliferation concomitant to attenuated proliferating cell nuclear antigen (PCNA, proliferation marker) expression. This is accompanied with attenuation in high glucose-induced phospho-p38 and pAkt expression. Moreover, CPMV-Cr inhibits expression of pro-inflammatory cytokines, transforming growth factor-β (TGF-β) and nuclear factor kappa-light-chain-enhancer of activated B cells (NF-κB), in glucose-stimulated HASMCs. Finally glucose-stimulated lipid uptake is remarkably abrogated by CPMV-Cr, revealed by Oil Red O staining. Together, these data provide key cellular evidence for an atheroprotective effect of CPMV-Cr in vascular smooth muscle cells (VSMC) under hyperglycemic conditions that may promote novel therapeutic ventures for diabetic atherosclerosis. Graphical abstract Novel trivalent chromium-loaded cowpea mosaic virus nanoparticles exhibit atheroprotective effect in vascular smooth muscle cells in a diabetic milieu. CPMV nanoparticles trivalent chromium vascular smooth muscle cells hyperglycemia anti-atherogenic effect Introduction Vascular disease accounts for increased morbidity and mortality in diabetic patients. Atherosclerosis characterized by excessive lipid deposition in the vessel wall leads to progressive luminal narrowing, limiting blood flow to target organs and triggering a multitude of cardiovascular complications.1 Diabetic patients have two- to four-fold greater risks of developing atherosclerotic complications.2–4 Despite significant advances, cost-effective strategies to alleviate risks of macrovascular complications in diabetes have remained elusive. Trivalent chromium (Cr3+) is a mineral nutrient long acclaimed for its beneficial glycemic and cardiovascular effects.5–11 Diabetic patients have low tissue Cr3+ levels, and numerous studies have indicated that suboptimal Cr3+ intake leads to elevated blood glucose, insulin and lipid levels as well as increased cardiovascular risks.12 In vivo and in vitro studies have shown that Cr3+ inhibits lipid peroxidation and attenuates secretion of pro-inflammatory cytokines including tumor necrosis factor-alpha (TGF-α), interleukin-6 (IL-6), monocyte chemoattractant protein-1 (MCP-1) and C-reactive protein (CRP),13–16 suggesting a putative role in lowering risks of vascular inflammation in diabetes. We recently reported that Cr3+ downregulates a potent pro-atherogenic protein thrombospondin-1 (TSP-1) expression and attenuates reactive oxygen species (ROS) formation coupled with an anti-proliferative effect in glucose-stimulated human aortic smooth muscle cells (HASMC) in vitro.17 These data lend support to the notion that Cr3+ modulates abnormal vascular smooth muscle cell (VSMC) function under hyperglycemic conditions, bearing significantly upon diabetic vascular disease. Current therapies for diabetic atherosclerosis are not targeted to the site of disease and therefore have major limitations coupled with reduced efficacy. While Cr3+ is a relatively inexpensive nutraceutical with low mutagenic potential,18–20 systemic administration at dosages of 2–3 mg/kg is considered toxic.18, 21, 22 Thus, tissue-targeted drug delivery systems are needed to overcome the dose-limiting toxicities. Toward this goal, we turn to a biology-inspired nanotechnology approach; specifically we are using nanoparticles derived from a plant virus. Many novel types of nanomaterials are in the developmental pipeline, and viruses have played a special role because they can be considered as Nature’s delivery systems. Based on their highly symmetrical and well-defined structures, plant viruses have advantages compared to their synthetic counterparts and they are favored over mammalian viruses owing to their higher safety profile in humans.23 The 30 nm-sized Cowpea mosaic virus (CPMV) is such a plant virus-derived nanoparticle, extensively studied for its biomedical applications in the field of cancer and cardiovascular medicine.24–28 The structure of CPMV is known to near atomic resolution, with well-established inside-out surface chemistry.29 CPMV is biocompatible and biodegradable allowing intravenous or oral delivery as purified nanoparticles or edible plant tissue,30 administered at doses of up to 100 mg/kg body weight without signs of toxicity31, 32 rendering it a suitable platform for drug delivery. While bacteriophages and other plant viruses are also being actively studied, with some examples in cancer therapy33–35 and tissue engineering,36–38 previous studies have shown that CPMV nanoparticles naturally interact with surface-expressed vimentin;27, 39, 40 this natural targeting mechanism can be utilized for disease-specific delivery. Moreover, earlier studies have reported that CPMV nanoparticles target atherosclerotic lesions and correlate with increased cell surface vimentin expression in lesions.26 The goal of the present study was to develop and investigate the anti-atherogenic potential of chromium chloride (CrCl3)-loaded CPMV nanoparticles (CPMV-Cr) in primary HASMC cultures under hyperglycemic conditions in vitro. Results Sustained Cr3+ Release from CPMV Nanoparticles Taking advantage of the RNA encapsulated within CPMV, it has previously been shown that gadolinium (Gd3+) and terbium (Tb3+) ions can be loaded through high affinity interactions with the RNA.41 Since Cr3+ has a comparable charge, we hypothesized a similar method could be used for infusing chromium within CPMV. 5 mg/mL CPMV was incubated with 50 mM CrCl3 (~60,000 equivalents) in 50 mM HEPES buffer containing 30 mM ethylenediaminetetraacetic acid (EDTA) and dialyzed against 10 mM EDTA in 10 mM HEPES. The Cr3+ concentration over 7 days was determined by Inductively Coupled Plasma Optical Emission Spectroscopy (ICP-OES), shown in Figure 1A. It was observed that after 3 days, a stable loading value of 2000 +/− 20% was reached, which is significantly higher than values of around 80 achieved previously with Gd3+ and Tb3+.41 For subsequent studies, CPMV-Cr samples were purified over 3 days and characterized using ICP-OES before use. Particle morphology examined by transmission electron microscopy (TEM) revealed that the particles remained intact and incubation with CrCl3 did not destabilize the particles (Figure 1B). Uptake of CrCl3-loaded CPMV by HASMC To confirm the cellular uptake of CPMV-Cr, primary HASMC cultures were treated with 25 mM glucose and incubated with 100 nM CrCl3-loaded CPMV. Staining was achieved using an anti-CPMV antibody followed by secondary staining. Immunofluorescence microscopy showed a significant increase in CPMV staining, as revealed by enhanced red cytoplasmic staining. In contrast, no cytoplasmic staining was detected when cells were incubated in the absence of anti-CPMV antibody. These data clearly demonstrate HASMC targeting and cellular delivery of CPMV (Figure 2). CrCl3-loaded CPMV Inhibits HASMC Proliferation under Hyperglycemic Conditions We recently reported that Cr3+ inhibits high glucose-induced HASMC proliferation in vitro.17 To investigate whether these effects are preserved in cells incubated with CrCl3-loaded CPMV nanoparticles, primary HASMCs incubated with 25 mM glucose in presence or absence of 100 nM CrCl3-loaded CPMV were used in a cell proliferation assay. As shown in Figure 3A, CPMV-Cr significantly attenuated high glucose-induced HASMC proliferation. Specifically in glucose-stimulated cells, HASMC proliferation was inhibited by 76% following incubation with CPMV-Cr (vs. glucose alone, p ≤ 0.05). Consistent with these results, immunoblotting experiments demonstrated that CPMV-Cr decreased PCNA expression, a marker of cellular proliferation, in glucose-stimulated HASMC. Densitometric quantification revealed that while high glucose increased PCNA expression by 25% compared to untreated controls, incubation with CPMV-Cr significantly attenuated glucose-induced PCNA expression (Figure 3B, 41% vs. glucose alone, p = 0.026). Next, to elucidate the signaling mechanism responsible for this anti-proliferative response, the effect of CPMV-Cr on the expression of key signaling mediators of VSMC proliferation was assessed. Briefly, glucose-stimulated HASMCs treated with or without CPMV-Cr were subjected to immunoblotting using antibodies specific for phospho-p38, total-p38, phospho-Akt and total-Akt. CrCl3-loaded CPMV attenuated activation of both p38 and Akt protein expression in glucose-stimulated cells. Specifically, while 25 mM glucose increased phospho-p38 and pAkt expression in primary HASMC cultures (76% and 66.8%, respectively), there was a statistically significant decrease in protein expression of both phospho-p38 and pAkt in cells incubated with CrCl3-loaded CPMV compared to cells treated with glucose alone (Figure 3C, 3D, 49% and 59.3%, respectively, p = 0.004). CrCl3-loaded CPMV Attenuates the Expression of Pro-inflammatory Cytokines in HASMC under Hyperglycemic Conditions Consistent with our earlier reports studying the effects of CrCl3,17 incubation with CrCl3-loaded CPMV significantly inhibited the pro-atherogenic protein thrombospondin-1 (TSP-1) expression in glucose (25 mM)-stimulated HASMC cultures in vitro. Specifically, while high glucose increased TSP-1 expression by 2.9-fold compared to untreated controls, CPMV-Cr remarkably attenuated TSP-1 expression in glucose-stimulated HASMCs (Figure. 4A, 76% vs. glucose alone, p = 0.04). Next, the effect of CPMV-Cr on the expression of pro-inflammatory cytokines, TGF-β and NF-κB, was assessed. Earlier studies have reported that high glucose upregulates TGF-β expression in VSMC as early as after 6 h of stimulation.42 Accordingly, primary HASMCs were treated with 25 mM glucose in presence or absence of CPMV-Cr and incubations were continued for 6–24 h. A significant increase in TGF-β expression was observed at both 6 h (data not shown) and 12 h of glucose stimulation and this effect was somewhat maintained up to 24 h (data not shown). Under similar experimental conditions, incubation with CPMV-Cr inhibited glucose-induced TGF-β expression. Quantification of immunoblotting images showed that while TGF-β protein expression was increased by 2.5-fold following 12 h glucose stimulation, there was a significant decrease in high glucose-induced TGF-β expression in response to CPMV-Cr (Figure 4B, 68.9% vs. glucose alone, p = 0.015). Similarly, high glucose increased NF-κB protein expression in HASMC in vitro and this effect was markedly attenuated by incubation with CPMV-Cr. Densitometric quantification of immunoblots revealed that while high glucose increased NF-κB expression by 2.2-fold compared to untreated cells, incubation with CPMV-Cr significantly inhibited NF-κB expression in glucose-stimulated HASMC (Figure 4C, 69.4% compared to cells treated with glucose alone, p = 0.005). CrCl3-loaded CPMV Inhibits Uptake of ox-LDL in Glucose-stimulated HASMC To determine the effect of CrCl3-loaded CPMV on lipid uptake by HASMCs under hyperglycemic conditions, primary HASMC cultures stimulated with 25 mM glucose with or without CrCl3-loaded CPMV were incubated with oxidized-low density lipoprotein (ox-LDL) and cellular lipid uptake was detected microscopically using Oil Red O staining. As shown in Figure 5, a significant increase in Oil Red O staining indicative of enhanced lipid uptake was observed in cells incubated with glucose alone. In contrary, CPMV-Cr remarkably abrogated high glucose-induced lipid uptake, revealed by reduced Oil Red O staining. Quantification of Oil Red O staining intensity depicted 3.36-fold increase (p = 0.002) in ox-LDL uptake in glucose-treated HASMC; CPMV-Cr, on the other hand, reduced lipid staining by ~70% in glucose-stimulated cells as compared to HASMCs treated with glucose alone (p = 0.001). Interestingly, CPMV nanoparticles in the absence of bound CrCl3 did not have any effect on the cellular uptake of ox-LDL in glucose-stimulated HASMC cultures (data not shown). Discussion The present study demonstrates novel anti-atherogenic effects of CrCl3-loaded CPMV nanoparticles in primary HASMC cultures in a diabetic milieu in vitro. These data suggest a potential application of CPMV nanoparticles as a suitable drug delivery platform for the mineral nutrient trivalent chromium in VSMCs. Accumulating literature highlight the utility of CPMV in a number of biomedical applications and nano-biotechnology including development of vaccines,43 imaging agents,25, 28, 44 therapeutics45 and chemical scaffolds.46–48 CPMV and other plant virus-based scaffolds provide unique advantages for applications in drug delivery: a particular advantage is their propagation in plants affording scalability and high degree of quality control and assurance. Being genetically encoded, these plant virus-based nanomaterials form highly monodisperse products with predictable surface morphologies. Furthermore, since plant viruses do not infect but enter mammalian cells, they provide a unique opportunity for intracellular drug delivery. CPMV, in particular, provides an interesting and unique platform because of its natural targeting to cells and tissues expressing surface vimentin.39, 49 Indeed, earlier studies have reported the ability of CPMV nanoparticles to be internalized by multiple cell types including cancer cells, endothelial cells and macrophages; in each case, cell targeting and uptake correlated with surface vimentin expression.27, 45, 50 Vimentin is a ubiquitous cytoskeletal component involved in basic cellular processes such as cell adhesion, migration, proliferation, cell-cell interactions as well as gene expression and signal transduction mechanisms.51 In its surface-expressed form, vimentin is thought to mediate endothelial interactions with circulating blood cells and migrating cells.52 Animal studies using the LDLR−/− mouse model of atherosclerosis demonstrated increased localization of CPMV nanoparticles in endothelial cells and macrophages within atherosclerotic lesions in vivo; again, CPMV targeting to the site of disease was correlated with surface vimentin expressed at the atherosclerotic lesion. On the other hand, in the absence of lesions, CPMV failed to penetrate the intact endothelium and remained associated with the endothelial cell layer in the non-lesion aorta.50 Although previous studies have demonstrated a rapid uptake of CPMV by a diverse group of cells in vivo and in vitro, interaction of CPMV with VSMC has not been previously examined. The present study provides the first evidence that CPMV nanoparticles are readily taken-up by primary HASMC cultures under diabetic conditions in vitro. Previous studies have indicated vimentin as a major intermediate filament protein expressed on VSMC that play a role in organization of the cytoskeletal framework of these cells.53 Moreover, surface-expressed vimentin on VSMCs have also been reported to possess O-GlcNAc-binding lectin-like properties.54 Accordingly, uptake of CPMV by HASMCs observed in this work suggests a role of cell-surface vimentin, as previously reported for other cell types. Smooth muscle cells (SMC) are integral cells in the vessel wall that regulate the tone and contractility of the vasculature. In a healthy blood vessel, SMCs are localized within the medial layer of the vascular wall in a quiescent, contractile state. In response to pro-atherogenic stimuli, SMCs undergo a phenotypic switching to a synthetic, proliferative phase, which predominantly resides in the intimal layer of the vessel wall.55, 56 Diabetic patients have an increased propensity for aberrant VSMC migration and proliferation,57 a hallmark of the synthetic SMC phenotype. Deregulation of VSMC phenotype from a differentiated to de-differentiated state, characterized by enhanced secretory and proliferative properties with low levels of contractile gene expression, is a major pathophysiological trigger for initiation and progression of atherosclerosis.55, 58 We recently reported that Cr3+ inhibits high glucose-induced HASMC proliferation that is specific for glucose-stimulated conditions as opposed to serum-stimulation in vitro; moreover, this anti-proliferative effect was independent of the anionic ligand (chloride or picolinate) bound to Cr3+.17 Consistent with these data, the present study demonstrates a strong anti-proliferative response to supplementation with CrCl3-loaded CPMV nanoparticles in glucose-stimulated primary HASMC cultures. Although multiple studies have suggested beneficial outcomes of Cr3+ supplementation in cardiovascular disease and dysregulated glycemic health, growing reports emphasize that long-term use of Cr3+ can result in GI disturbances, liver problems and impaired coordination or cognition; high cellular concentrations can further lead to DNA damage.22 Moreover, clinical trials using Cr3+ have yielded mixed results possibly confounded by the use of oral dose of Cr3+ viewed as being ‘too low’ and poor or inconsistent bio-distribution profiles of Cr3+ formulations.59, 60 To this end, the current work revealing an inhibitory effect of CrCl3-loaded CPMV on HASMC proliferation under hyperglycemic conditions highlight an enhanced potential of CPMV nanoparticles as an ideal platform for lesion-targeted delivery of Cr3+. Extensive literature demonstrates that the p38-MAPK pathway and AKT family of serine threonine kinases are critical regulators of cell growth and proliferation.61, 62 Previous studies have reported that activation of p38-MAPK following vascular injury promotes neointimal hyperplasia mediated via release of pro-inflammatory and fibroproliferative cytokines and growth factors.63 Multiple lines of evidence suggest that p38-MAPK pathway plays a pivotal role in diverse cellular processes including cell migration, proliferation, differentiation and cell cycle regulation. Both in vivo and in vitro studies have shown that selective inhibition of p38-MAPK attenuates aberrant VSMC proliferation limiting neointimal growth and abnormal vascular remodeling, characteristic of atherosclerosis.61, 63, 64 Earlier studies have also shown that activation of Akt and p38-MAPK by high glucose initiates downstream signaling cascades that mediate upregulation of several cell cycle-related genes including cyclin D, cyclin E and PCNA.65 In accordance with these earlier reports, the current findings that CPMV-Cr inhibits phospho-p38-MAPK and phospho-AKT expression in glucose-stimulated HASMCs suggest that ablation of VSMC growth response and cell cycle-regulated pathways may serve as important targets of CrCl3-loaded CPMV nanoparticles under hyperglycemic conditions. Atherosclerotic lesions typically have an enhanced inflammatory milieu. Upregulation of pro-inflammatory chemokines such as MCP-1 triggers migration and activation of monocytes into macrophages within the subendothelial space of the vessel wall; mature macrophages capable of ingesting atherogenic lipids, in turn contribute to plaque evolution.66 Previous studies have shown that supplementation with chromium dinicocysteinate in vivo significantly reduced CRP, MCP-1 and ICAM-1 levels in Zucker diabetic fatty rats.67 Additionally, in vivo administration of chromium niacinate and chromium picolinate lowered blood levels of TGF-α, IL-6 and CRP coupled to reduced triglyceride and cholesterol concentration in streptozotocin-treated diabetic rats.16 Furthermore, in vitro studies using isolated human blood mononuclear cells and U937 monocytes have shown that chromium chloride and chromium niacinate inhibit TGF-α, IL-6, IL-8 and MCP-1 secretion in glucose-stimulated cells.13, 15 We recently reported that Cr3+ downregulates a potent pro-atherogenic protein thrombospondin-1 (TSP-1) expression in glucose-stimulated HASMC cultures and this effect was associated with its anti-proliferative response.17 Consistent with these earlier findings, the current work demonstrated that CrCl3-loaded CPMV nanoparticles inhibit high glucose-induced TSP-1 expression in HASMCs in vitro. TSP-1 belongs to a family of extracellular matrix proteins that regulate cell-cell and cell-matrix interactions.68, 69 Earlier studies have reported increased TSP-1 expression in the injured wall70, 71 and early-stage atherosclerotic lesions,72 with enhanced expression in VSMC,73 and have implicated a role of TSP-1 in restenosis.74 Diabetic patients and diabetic animal models show enhanced TSP-1 expression;75, 76 moreover, an upregulation of TSP-1 expression was also reported in vascular cells exposed to high glucose in vitro.76, 77 Notably in the current work, TSP-1 downregulation was accompanied with attenuation in high glucose-induced TGF-β expression by CPMV-Cr. Numerous in vitro and in vivo studies have reported that TSP-1 is a critical endogenous activator of TGF-β, a potent chemotactic pro-inflammatory cytokine.78 TGF-β, a multifunctional polypeptide, regulates diverse cellular processes including cell growth, proliferation, differentiation, cell motility and apoptosis. Earlier studies have linked TGF-β with a multitude of diabetic complications including macrovascular disease and restenosis, diabetic nephropathy and diabetic cardiomyopathy.79, 80 Elevated TGF-β expression was found in diabetic patients, diabetic animal models as well as in a number of cell types (mesangial cells, VSMC) exposed to high glucose in vitro.78, 81 Previous studies have also demonstrated that Cr3+ in vivo inhibits the expression of NF-κB, a key transcription factor that mediates inflammatory and immune responses, in Zucker diabetic fatty liver.67 Earlier work suggests under certain pathological conditions, there exist direct correlations between the inflammatory response and NF-κB activation. Additionally, in vitro and in vivo studies demonstrate that NF-κB inhibition can dramatically abrogate development of atherosclerotic lesions modulated via attenuation of inflammatory and VSMC proliferative responses.82 Cogent to these reports, the present study revealing attenuation of TSP-1, TGF-β and NF-κB expression in glucose-stimulated VSMC suggest a protective role of CrCl3-loaded CPMV nanoparticles in development of fibroproliferative vascular remodeling associated with diabetes. Excessive accumulation of low density lipoproteins (LDLs) within the subendothelial spaces of the arterial vessel wall is a hallmark of atherosclerosis. While macrophages are the predominant cell-type associated with lipid uptake in the arterial intima, growing evidence demonstrate an increased ability of VSMC to incorporate aggregated LDLs, in turn contributing to foam cell formation.83 Earlier studies have shown that VSMCs express LDL receptor-related protein 1 (LRP1) that is capable of binding and internalizing aggregated LDLs, resulting in increased cholesteryl ester accumulation in VSMCs.84 LRP1 was initially identified as an endocytosis-mediated receptor for many ligands including thrombospondins, plasma lipoproteins such as apoE-enriched VLDL, lipoprotein lipase and lipoprotein lipase-triglyceride rich lipoprotein complexes.85 Previous studies have demonstrated that hypercholesterolemia upregulates LRP1 expression on VSMC, mediating enhanced cholesterol accumulation within these cells, in turn modulating collagen assembly and proteoglycan composition in VSMCs.86 Recent studies further indicate that defects in LRP1 expression in the vascular wall may promote atherosclerosis via increased VSMC proliferation.85 In the context of these earlier reports, findings from the present study that CrCl3-loaded CPMV nanoparticles inhibit high glucose-induced lipid uptake in VSMC prompt us to speculate that these effects may be mediated via modulation of LRP-1 by CPMV-Cr. Interestingly, CPMV nanoparticles in the absence of any bound Cr3+ did not have any effect on the expression of pro-inflammatory cytokines and VSMC lipid uptake stimulated by hyperglycemic conditions in vitro (data not shown). Conclusion In summary, the current study provides the first cellular evidence for important anti-atherogenic effects of CrCl3-loaded CPMV nanoparticles in VSMC under hyperglycemic conditions. Specifically using cell proliferation assay and immunoblotting experiments, we have shown that CrCl3-loaded CPMV nanoparticles inhibit smooth muscle cell proliferation and protein expression of a cell proliferation marker, PCNA, in glucose-stimulated HASMC primary cultures. This anti-proliferative effect of CPMV-Cr was further accompanied with significant attenuation in activation of high glucose-induced p38-MAPK and AKT expression, critical signaling mediators of smooth muscle cell proliferation. In addition, incubation with CPMV-Cr inhibited the expression profiles of pro-atherogenic matricellular protein (TSP-1) and pro-inflammatory cytokines (TGF-β and NF-κB) in HASMC under diabetic milieu, as shown by immunoblotting. Finally, Oil red O staining revealed that supplementation with CPMV-Cr remarkably abrogated high glucose-induced lipid uptake in HASMC cultures. Overall, these data strongly support the notion that CPMV plant virus-based nanoparticles may present a novel platform technology enabling lesion-targeted Cr3+ delivery in diabetes, which may promote increased efficacy at reduced dosage while avoiding systemic toxicity. Future in vivo studies using an atherosclerotic mouse model of diabetes to determine whether CrCl3-loaded CPMV nanoparticles may facilitate lesion-specific Cr3+ delivery and attenuate development of atherosclerotic lesions warrant further investigation. Such studies will set the stage for preclinical development of CPMV-Cr to treat atherosclerotic complications associated with diabetes. Experimental Section CPMV Purification and Cr3+ Drug Loading CPMV nanoparticles were purified from Vigna unguiculata plants infected by mechanical inoculation using a 0.1 mg/mL solution of CPMV in 0.1 M potassium phosphate buffer. Yields of 1 mg/g of infected leaf material were obtained using established procedures.87 For chromium loading, 5 mg/mL CPMV was incubated with 50 mM CrCl3 (Sigma Aldrich) and 30 mM ethylenediaminetetraacetic acid (EDTA) in 50 mM HEPES buffer, pH 7 overnight (~60,000-fold molar excess). The sample was then dialyzed against 10 mM HEPES buffer with 10 mM EDTA using 12–14 kDa molecular weight cut-off tubing (Spectrum Laboratories) over 3 days. Some precipitation from the presence of NaOH in the buffer was observed over time and removed by a clearing spin at 10,000 rpm for 10 min. The release profile was studied by taking aliquots of the sample over a period of 7 days; the amount of drug in the sample was then determined using ICP-OES (see below). UV/visible Spectroscopy CPMV concentration was determined using a NanoDrop 2000 Spectrophotometer (Thermo Scientific) with the reported extinction coefficient for CPMV (ε260 nm = 8.1 mg−1 mL cm−1). Inductively Coupled Plasma Optical Emission Spectroscopy (ICP-OES) Chromium content in CPMV was measured by ICP-OES (Agilent 730 Axial ICP-OES), as determined by the emission spectral line at 267.716 nm. Transmission Electron Microscopy (TEM) After sample dialysis, CPMV-Cr was imaged by TEM. Samples were diluted to a concentration of 0.1 mg/mL CPMV in deionized water then applied to a formvar/carbon-coated copper grid for 5 minutes. The grid was negatively stained with 2% (w/v) uranyl acetate for 2 minutes and imaged at 200 kV using a Zeiss Libra 200FE transmission electron microscope. Cell Culture Primary HASMC cultures (purchased from Cambrex) at passages 6–14 grown in DMEM/F12 media supplemented with 10% FBS and 1% penicillin-streptomycin (Cellgro) were used in all experiments. Confluent cells were placed in low glucose (5 mM)-0.2% FBS DMEM overnight (16–17 h) followed by incubation with 25 mM glucose in the presence or absence of 100 nM CrCl3-loaded CPMV nanoparticles. Immunocytochemistry About 70% confluent HASMCs grown on cover slips in six-well clusters were treated as described above. Cells were fixed in a solution containing 4% paraformaldehyde and 0.2% Triton X-100 for 20 min at 25°C and blocked in 5% donkey serum for 1 h at room temperature. This was followed by overnight incubation at 4°C with rabbit anti-CPMV primary antibody (Pacific Immunology) or no primary antibody (used as negative control). Cells were then incubated with Alexa Fluor® 594 secondary antibody (1:1500) for 1 h at room temperature followed by washing in PBS (3X, 5 mins each). Cover slips were mounted using DAPI-containing vectashield mounting media (Vector Lab) for identification of cell nuclei. The mean red fluorescence intensity indicative of CPMV staining was measured for each image. Proliferation of Cultured Smooth Muscle Cells Primary HASMC cultures were plated (1500–2000 cells/well) on 96-well tissue culture plates in 10% FBS and 1% penicillin-streptomycin containing DMEM/F12 media. After allowing for overnight cell growth, HASMCs were treated with 25 mM glucose in presence or absence of 100 nM CrCl3-loaded CPMV. Cell incubations were continued for 72 hours followed by measurement of cellular proliferation using WST-1 reagent (Cayman Chemicals), as we previously reported.17 Preparation of Whole Cell Lysate and Immunoblotting Whole cell lysates were prepared and proteins were measured (BioRad), as we reported earlier.17 Equal amounts of proteins were resolved on 8% SDS-polyacrylamide gels and transferred on PVDF membranes. Membranes were blocked in 5% nonfat dry milk for 45 mins-1 h at room temperature followed by overnight incubation at 4°C with different antibodies: anti-PCNA (1:300, Abcam), anti-phospho-Akt (1:1000, Cell Signaling), anti-total-Akt (1:1000, Cell Signaling), anti-phospho- and total-p38 (1:500, Cell Signaling), anti-TSP-1, (1:500; clone AB11, Thermo Fisher), anti-TGF-β (1:250, Bioss), anti-NF-κB (1:500, Bioss) and anti-β-actin (loading control, Cell Signaling). Equal protein loading was also confirmed by staining membranes with Ponceau S. Densitometric quantifications were conducted using ImageJ and Adobe photoshop softwares and results were expressed as fold-increase vs. Controls. Lipid Uptake Assay About 60–70% confluent HASMC cultures, seeded on cover slips in six-well plates, were placed overnight in serum-free low-glucose (5 mM) DMEM media. This was followed by the addition of 25 mM glucose and cell incubations were continued in presence or absence of 100 nM CrCl3-loaded CPMV for 19 h; ox-LDL (50 µg/ml) was then added to the media and incubations were continued for an addition 4–5 h. Following this, media was aspirated and cells were rinsed in PBS followed by formalin fixation for 10 mins; cells were rinsed once again in PBS (1 min) and 60% isopropanol (15 sec). This was followed by staining of cells with (0.5%) Oil Red O solution for 15 mins at room temperature; cells were destained with 60% isopropanol for 15 sec followed by PBS washing. Nuclei were stained with Harris hematoxylin and cover slips were mounted on glycerin jelly. Images were acquired using Olympus BX40 microscope at 40X magnification. Statistical Analyses All experiments were repeated three-to-five times with two replicates for each treatment within an independent experiment. For microscopic experiments, six to ten images were collected for each individual treatment within each independent experiment. Images were quantified using ImageJ software. Results are presented as fold-increase vs. controls. Values are means ± SEM. Significant differences between the mean values were determined using unpaired Student’s t test, with p ≤ 0.05 being considered statistically significant. This work was supported in parts by grants from Diabetes Action Research and Education Foundation (DAREF) award (to PR), Northeast Ohio Medical University start-up funds (to PR), National Institutes of Health NHLBI R21 HL121130 (to NFS) and NHLBI F31 HL129703 (to AMW). Figure 1 Loading and release of chromium in CPMV A) After CPMV incubation with 50 mM CrCl3, dialysis was performed against 10 mM HEPES with 10 mM EDTA. Quantification of chromium in CPMV-Cr over time was measured by ICP-OES. Error bars show standard deviation of three measurements (it should be noted that all data points were measured in triplicates and error bars are shown for each time point, significant experimental variation was only observed for data collected on day 5). B) Transmission electron microscopy of CPMV loaded with CrCl3 indicated that the particle structure remains unchanged after modification. Figure 2 Uptake of CrCl3-loaded CPMV nanoparticles by HASMCs in vitro Primary HASMCs were grown on cover slips and acclimatized overnight in low glucose (5 mM) DMEM. Cells were incubated for 24 hrs with 100 nM CrCl3-loaded CPMV nanoparticles in presence of 25 mM glucose. Fixed cells were incubated overnight with or without anti-CPMV primary antibody at 4°C followed by secondary incubation with Alexa Fluor® 555 conjugated anti-rabbit IgG for 1 h at room temperature. Shown are representative immunofluorescence images from three independent experiments. Images were captured at 40X (upper panel) and 60X (lower panel) magnification, depicting cytosolic CPMV staining (red) and DAPI (blue) stained nuclei. Figure 3 CrCl3-loaded CPMV attenuates HASMC proliferation, inhibits PCNA, p38 and pAKT expression in glucose-stimulated HASMC A) Primary HASMCs were grown in 96-well plates and incubated in 25 mM glucose containing serum free DMEM in presence or absence of 100 nM CPMV-Cr. Cell proliferation was assessed 72 h later using WST-1 reagent as described in Methods. Results are expressed as fold-change vs. glucose stimulation. Values are expressed as means ± S.E. (n=3–4); *p ≤ 0.05 vs. Glucose. B-D) Following overnight pre-incubation in low glucose (5 mM)-0.2% FBS-containing DMEM, primary HASMC cultures were incubated in 25 mM glucose with or without 100 nM CrCl3-loaded CPMV nanoparticles for 24 hrs. Whole cell lysates were prepared at endpoint and subjected to immunoblotting, as described in Methods. Shown are representative immunoblots depicting (B) PCNA, (C) p38 and (D) pAkt protein expression (upper panels); densitometric quantification of immunoblots from 3–5 independent experiments is shown in the bar graphs (lower panels). Results are expressed as fold-increase vs. Controls. All values are expressed as means ± S.E. (n=3–5); *p ≤ 0.03 vs. Control, # p ≤ 0.03 vs. Glucose. Figure 4 CrCl3-loaded CPMV attenuates TSP-1, TGF-β and NF-kB expression in glucose-stimulated HASMC Primary HASMC cultures were acclimatized overnight in low glucose (5 mM)-0.2% FBS-containing DMEM. This was followed by stimulation with 25 mM glucose in presence or absence of 100 nM CrCl3-loaded CPMV nanoparticles for 24 hrs. For (B) cell incubation were continued for 12 h. Whole cell lysates were prepared at end point and utilized in immunoblotting experiments using antibodies against TSP-1, TGF-β and NF-κB. Shown are the representative immunoblots (upper panels); densitometric quantification of immunoblots from 3–5 independent experiments are shown in the bar graphs (lower panels). Results are expressed as fold-increase vs. Controls. All values are expressed as means ± S.E. (n=3–5); *p ≤ 0.026 vs. Control, ♦ p ≤ 0.04 vs. Glucose. Figure 5 CrCl3-loaded CPMV inhibits uptake of oxLDL in glucose-stimulated HASMC A) Representative light microscopic images taken at 40X magnification showing Oil Red O stained oxidized LDL particles taken up by primary HASMCs and hematoxylin stained blue nuclei. HASMCs grown on cover slips were fixed and stained by Oil Red O followed by hematoxylin counter staining. B) Red intensity measurement from 10–20 different microscopic fields. Results are expressed as means ± S.E. (n=5); *p ≤ 0.002 vs. Control, **p ≤ 0.0012 vs. Glucose. References 1 Lusis AJ Atherosclerosis Nature 2000 407 233 241 11001066 2 Haffner SM Mykkanen L Festa A Burke JP Stern MP Insulin-resistant prediabetic subjects have more atherogenic risk factors than insulin-sensitive prediabetic subjects: implications for preventing coronary heart disease during the prediabetic state Circulation 2000 101 975 980 10704163 3 Abate N Obesity and cardiovascular disease. Pathogenetic role of the metabolic syndrome and therapeutic implications J. Diabetes Complications 2000 14 154 174 10989324 4 Isomaa B Almgren P Tuomi T Forsen B Lahti K Nissen M Taskinen MR Groop L Cardiovascular morbidity and mortality associated with the metabolic syndrome Diabetes Care 2001 24 683 689 11315831 5 Lau FC Bagchi M Sen CK Bagchi D Nutrigenomic basis of beneficial effects of chromium(III) on obesity and diabetes Mol. Cell. Biochem 2008 317 1 10 18636317 6 Lee NA Reasner CA Beneficial effect of chromium supplementation on serum triglyceride levels in NIDDM Diabetes Care 1994 17 1449 1452 7882815 7 Anderson RA Chromium in the prevention and control of diabetes Diabetes Metab 2000 26 22 27 10705100 8 Hummel M Standl E Schnell O Chromium in metabolic and cardiovascular disease Horm. Metab. Res 2007 39 743 751 17952838 9 Martin J Wang ZQ Zhang XH Wachtel D Volaufova J Matthews DE Cefalu WT Chromium picolinate supplementation attenuates body weight gain and increases insulin sensitivity in subjects with type 2 diabetes Diabetes Care 2006 29 1826 1832 16873787 10 Mossop RT Effects of chromium III on fasting blood glucose, cholesterol and cholesterol HDL levels in diabetics Cent. Afr. J. Med 1983 29 80 82 6883494 11 Machalinski B Walczak M Syrenicz A Machalinska A Grymula K Stecewicz I Wiszniewska B Dabkowska E Hypoglycemic potency of novel trivalent chromium in hyperglycemic insulin-deficient rats J. Trace Elem. Med. Biol 2006 20 33 39 16632174 12 Wang ZQ Cefalu WT Current concepts about chromium supplementation in type 2 diabetes and insulin resistance Curr. Diab. Rep 2010 10 145 151 20425574 13 Jain SK Lim G Chromium chloride inhibits TNFalpha and IL-6 secretion in isolated human blood mononuclear cells exposed to high glucose Horm. Metab. Res 2006 38 60 62 16477544 14 Jain SK Kannan K Chromium chloride inhibits oxidative stress and TNF-alpha secretion caused by exposure to high glucose in cultured U937 monocytes Biochem. Biophys. Res. Commun 2001 289 687 691 11726202 15 Jain SK Rains JL Croad JL High glucose and ketosis (acetoacetate) increases, and chromium niacinate decreases, IL-6, IL-8, and MCP-1 secretion and oxidative stress in U937 monocytes Antioxid. Redox Signal 2007 9 1581 1590 17665966 16 Jain SK Rains JL Croad JL Effect of chromium niacinate and chromium picolinate supplementation on lipid peroxidation, TNF-alpha, IL-6, CRP, glycated hemoglobin, triglycerides, and cholesterol levels in blood of streptozotocin-treated diabetic rats Free Radic. Biol. Med 2007 43 1124 1131 17854708 17 Ganguly R Sahu S Chavez RJ Raman P Trivalent chromium inhibits TSP-1 expression, proliferation, and O-GlcNAc signaling in vascular smooth muscle cells in response to high glucose in vitro Am. J. Physiol. Cell Physiol 2015 308 C111 C122 25354527 18 Lamson DW Plaza SM The safety and efficacy of high-dose chromium Altern. Med. Rev 2002 7 218 235 12126463 19 Hathcock JN Safety limits for nutrients J. Nutr 1996 126 2386S 2389S 8811802 20 Rhodes MC Hebert CD Herbert RA Morinello EJ Roycroft JH Travlos GS Abdo KM Absence of toxic effects in F344/N rats and B6C3F1 mice following subchronic administration of chromium picolinate monohydrate Food Chem. Toxicol 2005 43 21 29 15582192 21 Katz SA Salem H The toxicology of chromium with respect to its chemical speciation: a review J. Appl. Toxicol 1993 13 217 224 8326093 22 Eastmond DA Macgregor JT Slesinski RS Trivalent chromium: assessing the genotoxic risk of an essential trace element and widely used human and animal nutritional supplement Crit. Rev. Toxicol 2008 38 173 190 18324515 23 Manchester M Singh P Virus-based nanoparticles (VNPs): platform technologies for diagnostic imaging Adv. Drug Deliv. Rev 2006 58 1505 1522 17118484 24 Wen AM Wang Y Jiang K Hsu GC Gao H Lee KL Yang AC Yu X Simon DI Steinmetz NF Shaping bio-inspired nanotechnologies to target thrombosis for dual optical-magnetic resonance imaging J. Mater. Chem. B 2015 3 6037 6045 25 Lewis JD Destito G Zijlstra A Gonzalez MJ Quigley JP Manchester M Stuhlmann H Viral nanoparticles as tools for intravital vascular imaging Nat. Med 2006 12 354 360 16501571 26 Plummer EM Thomas D Destito G Shriver LP Manchester M Interaction of cowpea mosaic virus nanoparticles with surface vimentin and inflammatory cells in atherosclerotic lesions Nanomedicine (Lond) 2012 7 877 888 22394183 27 Steinmetz NF Cho CF Ablack A Lewis JD Manchester M Cowpea mosaic virus nanoparticles target surface vimentin on cancer cells Nanomedicine (Lond) 2011 6 351 364 21385137 28 Steinmetz NF Ablack AL Hickey JL Ablack J Manocha B Mymryk JS Luyt LG Lewis JD Intravital Imaging of Human Prostate Cancer Using Viral Nanoparticles Targeted to Gastrin-Releasing Peptide Receptors Small (Weinheim an der Bergstrasse, Germany) 2011 7 1664 1672 29 Wen AM Shukla S Saxena P Aljabali AA Yildiz I Dey S Mealy JE Yang AC Evans DJ Lomonossoff GP Steinmetz NF Interior engineering of a viral nanoparticle and its tumor homing properties Biomacromolecules 2012 13 3990 4001 23121655 30 Rae CS Khor IW Wang Q Destito G Gonzalez MJ Singh P Thomas DM Estrada MN Powell E Finn MG Manchester M Systemic trafficking of plant virus nanoparticles in mice via the oral route Virology 2005 343 224 235 16185741 31 Kaiser CR Flenniken ML Gillitzer E Harmsen AL Harmsen AG Jutila MA Douglas T Young MJ Biodistribution studies of protein cage nanoparticles demonstrate broad tissue distribution and rapid clearance in vivo Int. J. Nanomedicine 2007 2 715 733 18203438 32 Singh P Prasuhn D Yeh RM Destito G Rae CS Osborn K Finn MG Manchester M Bio-distribution, toxicity and pathology of cowpea mosaic virus nanoparticles in vivo J. Control. Release 2007 120 41 50 17512998 33 Gandra N Abbineni G Qu X Huai Y Wang L Mao C Bacteriophage bionanowire as a carrier for both cancer-targeting peptides and photosensitizers and its use in selective cancer cell killing by photodynamic therapy Small 2013 9 215 221 23047655 34 Wen AM Ryan MJ Yang AC Breitenkamp K Pokorski JK Steinmetz NF Photodynamic activity of viral nanoparticles conjugated with C60 Chem. Commun. (Camb) 2012 48 9044 9046 22858632 35 Lockney DM Guenther RN Loo L Overton W Antonelli R Clark J Hu M Luft C Lommel SA Franzen S The Red clover necrotic mosaic virus capsid as a multifunctional cell targeting plant viral nanoparticle Bioconjug. Chem 2011 22 67 73 21126069 36 Lee DY Lee H Kim Y Yoo SY Chung WJ Kim G Phage as versatile nanoink for printing 3-D cell-laden scaffolds Acta Biomater 2016 29 112 124 26441128 37 Wang J Yang M Zhu Y Wang L Tomsia AP Mao C Phage nanofibers induce vascularized osteogenesis in 3D printed bone scaffolds Adv. Mater 2014 26 4961 4966 24711251 38 Zhao X Lin Y Wang Q Virus-based scaffolds for tissue engineering applications Wiley Interdiscip. Rev. Nanomed. Nanobiotechnol 2015 7 534 547 25521747 39 Koudelka KJ Destito G Plummer EM Trauger SA Siuzdak G Manchester M Endothelial targeting of cowpea mosaic virus (CPMV) via surface vimentin PLoS Pathog 2009 5 e1000417 19412526 40 Steinmetz NF Mauger J Sheng H Bensussan A Marcic I Kumar V Braciak T Two domains of vimentin are expressed on the surface of lymph node, bone and brain metastatic prostate cancer lines along with the putative stem cell marker proteins CD44 and CD133 Cancers 2011 3 2870 2885 24212937 41 Prasuhn DE Jr Yeh RM Obenaus A Manchester M Finn MG Viral MRI contrast agents: coordination of Gd by native virions and attachment of Gd complexes by azide-alkyne cycloaddition Chem. Commun. (Camb) 2007 1269 1271 17356779 42 Lindschau C Quass P Menne J Guler F Fiebeler A Leitges M Luft FC Haller H Glucose-induced TGF-beta1 and TGF-beta receptor-1 expression in vascular smooth muscle cells is mediated by protein kinase C-alpha Hypertension 2003 42 335 341 12939231 43 Lizotte PH Wen AM Sheen MR Fields J Rojanasopondist P Steinmetz NF Fiering S In situ vaccination with cowpea mosaic virus nanoparticles suppresses metastatic cancer Nat. Nanotechnol 2015 in press 44 Leong HS Steinmetz NF Ablack A Destito G Zijlstra A Stuhlmann H Manchester M Lewis JD Intravital imaging of embryonic and tumor neovasculature using viral nanoparticles Nat. Protoc 2010 5 1406 1417 20671724 45 Yildiz I Lee KL Chen K Shukla S Steinmetz NF Infusion of imaging and therapeutic molecules into the plant virus-based carrier cowpea mosaic virus: cargo-loading and delivery J. Control. Release 2013 172 568 578 23665254 46 Steinmetz NF Hong V Spoerke ED Lu P Breitenkamp K Finn MG Manchester M Buckyballs meet viral nanoparticles: candidates for biomedicine J. Am. Chem. Soc 2009 131 17093 17095 19904938 47 Wen AM Steinmetz NF The aspect ratio of nanoparticle assemblies and the spatial arrangement of ligands can be optimized to enhance the targeting of cancer cells Adv. Healthc. Mater 2014 3 1739 1744 24729309 48 Kaltgrad E O'Reilly MK Liao L Han S Paulson JC Finn MG On-virus construction of polyvalent glycan ligands for cell-surface receptors J. Am. Chem. Soc 2008 130 4578 4579 18341338 49 Koudelka KJ Rae CS Gonzalez MJ Manchester M Interaction between a 54-kilodalton mammalian cell surface protein and cowpea mosaic virus J. Virol 2007 81 1632 1640 17121801 50 Plummer EM Thomas D Destito G Shriver LP Manchester M Interaction of cowpea mosaic virus nanoparticles with surface vimentin and inflammatory cells in atherosclerotic lesions Nanomedicine 2012 7 877 888 22394183 51 Ivaska J Pallari HM Nevo J Eriksson JE Novel functions of vimentin in cell adhesion, migration, and signaling Exp. Cell Res 2007 313 2050 2062 17512929 52 Nieminen M Henttinen T Merinen M Marttila-Ichihara F Eriksson JE Jalkanen S Vimentin function in lymphocyte adhesion and transcellular migration Nat. Cell Biol 2006 8 156 162 16429129 53 Wang R Li Q Tang DD Role of vimentin in smooth muscle force development Am. J. Physiol. Cell Physiol 2006 291 C483 C489 16571866 54 Ise H Kobayashi S Goto M Sato T Kawakubo M Takahashi M Ikeda U Akaike T Vimentin and desmin possess GlcNAc-binding lectin-like properties on cell surfaces Glycobiology 2010 20 843 864 20332081 55 Lacolley P Regnault V Nicoletti A Li Z Michel JB The vascular smooth muscle cell in arterial pathology: a cell that can take on multiple roles Cardiovasc. Res 2012 95 194 204 22467316 56 Rensen SS Doevendans PA van Eys GJ Regulation and characteristics of vascular smooth muscle cell phenotypic diversity Neth. Heart J 2007 15 100 108 17612668 57 Faries PL Rohan DI Takahara H Wyers MC Contreras MA Quist WC King GL Logerfo FW Human vascular smooth muscle cells of diabetic origin exhibit increased proliferation, adhesion, and migration J. Vasc. Surg 2001 33 601 607 11241133 58 Rivard A Andres V Vascular smooth muscle cell proliferation in the pathogenesis of atherosclerotic cardiovascular diseases Histol. Histopathol 2000 15 557 571 10809377 59 Cefalu WT Hu FB Role of chromium in human health and in diabetes Diabetes Care 2004 27 2741 2751 15505017 60 Althuis MD Jordan NE Ludington EA Wittes JT Glucose and insulin responses to dietary chromium supplements: a meta-analysis Am. J. Clin. Nutr 2002 76 148 155 12081828 61 Graf K Xi XP Yang D Fleck E Hsueh WA Law RE Mitogen-activated protein kinase activation is involved in platelet-derived growth factor-directed migration by vascular smooth muscle cells Hypertension 1997 29 334 339 9039124 62 Muto A Fitzgerald TN Pimiento JM Maloney SP Teso D Paszkowiak JJ Westvik TS Kudo FA Nishibe T Dardik A Smooth muscle cell signal transduction: implications of vascular biology for vascular surgeons J. Vasc. Surg 2007 45 Suppl A A15 A24 17544020 63 Zhan Y Kim S Izumi Y Izumiya Y Nakao T Miyazaki H Iwao H Role of JNK, p38, and ERK in platelet-derived growth factor-induced vascular proliferation, migration, and gene expression Arterioscler. Thromb. Vasc. Biol 2003 23 795 801 12637337 64 Kanda Y Nishio E Kuroki Y Mizuno K Watanabe Y Thrombin activates p38 mitogen-activated protein kinase in vascular smooth muscle cells Life Sci 2001 68 1989 2000 11388701 65 Guo R Li W Liu B Li S Zhang B Xu Y Resveratrol protects vascular smooth muscle cells against high glucose-induced oxidative stress and cell proliferation in vitro Med. Sci. Monit. Basic Res 2014 20 82 92 24971582 66 Tabas I Garcia-Cardena G Owens GK Recent insights into the cellular biology of atherosclerosis J. Cell. Biol 2015 209 13 22 25869663 67 Jain SK Croad JL Velusamy T Rains JL Bull R Chromium dinicocysteinate supplementation can lower blood glucose, CRP, MCP-1, ICAM-1, creatinine, apparently mediated by elevated blood vitamin C and adiponectin and inhibition of NFkappaB, Akt, and Glut-2 in livers of zucker diabetic fatty rats Mol. Nutr. Food Res 2010 54 1371 1380 20306473 68 Adams JC Thrombospondins: multifunctional regulators of cell interactions Annu. Rev. Cell Dev. Biol 2001 17 25 51 11687483 69 Bornstein P Thrombospondins as matricellular modulators of cell function J. Clin. Invest 2001 107 929 934 11306593 70 Raugi GJ Mullen JS Bark DH Okada T Mayberg MR Thrombospondin deposition in rat carotid artery injury Am. J. Pathol 1990 137 179 185 1695483 71 Sajid M Hu Z Guo H Li H Stouffer GA Vascular expression of integrin-associated protein and thrombospondin increase after mechanical injury J. Investig. Med 2001 49 398 406 72 Riessen R Kearney M Lawler J Isner JM Immunolocalization of thrombospondin-1 in human atherosclerotic and restenotic arteries Am. Heart J 1998 135 357 364 9489988 73 Miano JM Vlasic N Tota RR Stemerman MB Smooth muscle cell immediate-early gene and growth factor activation follows vascular injury. A putative in vivo mechanism for autocrine growth Arterioscler. Thromb 1993 13 211 219 8427857 74 Pallero MA Talbert Roden M Chen YF Anderson PG Lemons J Brott BC Murphy-Ullrich JE Stainless steel ions stimulate increased thrombospondin-1-dependent TGF-beta activation by vascular smooth muscle cells: implications for instent restenosis J. Vasc. Res 2010 47 309 322 20016205 75 Bayraktar M Dundar S Kirazli S Teletar F Platelet factor 4, beta-thromboglobulin and thrombospondin levels in type I diabetes mellitus patients J. Int. Med. Res 1994 22 90 94 8020643 76 Stenina OI Krukovets I Wang K Zhou Z Forudi F Penn MS Topol EJ Plow EF Increased expression of thrombospondin-1 in vessel wall of diabetic Zucker rat Circulation 2003 107 3209 3215 12810612 77 Raman P Krukovets I Marinic TE Bornstein P Stenina OI Glycosylation mediates up-regulation of a potent antiangiogenic and proatherogenic protein, thrombospondin-1, by glucose in vascular smooth muscle cells J. Biol. Chem 2007 282 5704 5714 17178709 78 Wang S Lincoln TM Murphy-Ullrich JE Glucose downregulation of PKG-I protein mediates increased thrombospondin1-dependent TGF-{beta} activity in vascular smooth muscle cells Am. J. Physiol. Cell Physiol 2010 298 C1188 C1197 20164378 79 Chamberlain J Transforming growth factor-beta: a promising target for anti-stenosis therapy Cardiovasc. Drug Rev 2001 19 329 344 11830751 80 Yokoyama H Deckert T Central role of TGF-beta in the pathogenesis of diabetic nephropathy and macrovascular complications: a hypothesis Diabet. Med 1996 13 313 320 9162605 81 Zhu Y Usui HK Sharma K Regulation of transforming growth factor beta in diabetic nephropathy: implications for treatment Semin. Nephrol 2007 27 153 160 17418684 82 Jeong IK Oh da H Park SJ Kang JH Kim S Lee MS Kim MJ Hwang YC Ahn KJ Chung HY Chae MK Yoo HJ Inhibition of NF-kappaB prevents high glucose-induced proliferation and plasminogen activator inhibitor-1 expression in vascular smooth muscle cells Exp. Mol. Med 2011 43 684 692 21975282 83 Allahverdian S Pannu PS Francis GA Contribution of monocyte-derived macrophages and smooth muscle cells to arterial foam cell formation Cardiovasc. Res 2012 95 165 172 22345306 84 Llorente-Cortes V Martinez-Gonzalez J Badimon L LDL receptor-related protein mediates uptake of aggregated LDL in human vascular smooth muscle cells Arterioscler. Thromb. Vasc. Biol 2000 20 1572 1579 10845874 85 Basford JE Moore ZW Zhou L Herz J Hui DY Smooth muscle LDL receptor-related protein-1 inactivation reduces vascular reactivity and promotes injury-induced neointima formation Arterioscler. Thromb. Vasc. Biol 2009 29 1772 1778 19729608 86 Llorente-Cortes V Otero-Vinas M Sanchez S Rodriguez C Badimon L Low-density lipoprotein upregulates low-density lipoprotein receptor-related protein expression in vascular smooth muscle cells: possible involvement of sterol regulatory element binding protein-2-dependent mechanism Circulation 2002 106 3104 3110 12473559 87 Cho CF Shukla S Simpson EJ Steinmetz NF Luyt LG Lewis JD Molecular targeted viral nanoparticles as tools for imaging cancer Methods Mol. Biol 2014 1108 211 230 24243252

PMC005xxxxxx/PMC5136315.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 101132535 30097 Mol Cancer Ther Mol. Cancer Ther. Molecular cancer therapeutics 1535-7163 1538-8514 27638858 5136315 10.1158/1535-7163.MCT-16-0366 NIHMS814215 Article Oncogenic Receptor Tyrosine Kinases Directly Phosphorylate Focal Adhesion Kinase (FAK) as a Resistance Mechanism to FAK-kinase Inhibitors Marlowe Timothy A. 1 Lenzo Felicia L. 2 Figel Sheila A. 3 Grapes Abigail T. 2 Cance William G. 34* 1 Department of Pharmacology & Therapeutics, Roswell Park Cancer Institute, Buffalo, NY 2 Department of Cell Stress Biology, Roswell Park Cancer Institute, Buffalo, NY 3 Department of Surgical Oncology, Roswell Park Cancer Institute, Buffalo, NY 4 FAKnostics, LLC, Buffalo, NY * To whom correspondence should be addressed: william.cance@roswellpark.org Reprint Requests: William G. Cance, Elm & Carlton Streets, Buffalo, NY 14263 3 9 2016 16 9 2016 12 2016 01 12 2017 15 12 30283039 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. Focal adhesion kinase (FAK) is a major drug target in cancer and current inhibitors targeted to the ATP-binding pocket of the kinase domain have entered clinical trials. However, preliminary results have shown limited single-agent efficacy in patients. Despite these unfavorable data, the molecular mechanisms which drive intrinsic and acquired resistance to FAK-kinase inhibitors are largely unknown. We have demonstrated that receptor tyrosine kinases (RTKs) can directly bypass FAK-kinase inhibition in cancer cells through phosphorylation of FAK’s critical tyrosine 397 (Y397). We also showed that HER2 forms a direct protein-protein interaction with the FAK-FERM-F1 lobe, promoting direct phosphorylation of Y397. Additionally, FAK-kinase inhibition induced two forms of compensatory RTK reprogramming: 1) the rapid phosphorylation and activation of RTK signaling pathways in RTKHigh cells and 2) the long-term acquisition of RTKs novel to the parental cell line in RTKLow cells. Finally, HER2+ cancer cells displayed resistance to FAK-kinase inhibition in 3D–growth assays using a HER2 isogenic system and HER2+ cancer cell lines. Our data indicate a novel drug resistance mechanism to FAK-kinase inhibitors whereby HER2 and other RTKs can rescue and maintain FAK activation (pY397) even in the presence of FAK-kinase inhibition. These data may have important ramifications for existing clinical trials of FAK inhibitors and suggest that individual tumor stratification by RTK expression would be important to predict patient response to FAK-kinase inhibitors. FAK inhibitors Drug Resistance Growth factors and receptors Molecular modeling Novel mechanisms Introduction Focal adhesion kinase (FAK) is considered a major cancer drug target due to its vast overexpression in 80% of all solid tumors both at the protein and mRNA level (1–3). Additionally, it has been shown to play a critical role in multiple aspects of tumor progression such as proliferation, survival, invasion, metastasis, angiogenesis, cancer stem cell maintenance, and recently, immune cell suppression (4–11). Multiple studies using transgenic mouse models have shown a defect in tumor growth, invasion, and metastasis with specific FAK deletion (12–15). In further support, FAK overexpression has been clinically correlated with higher tumor stage, metastasis, and in some tumors (breast, lung, endometrium, liver, GI system, and esophagus) is associated with poor prognosis (16). FAK functions as a dual kinase and scaffolding protein where it forms complexes with various oncogenic proteins such as receptor tyrosine kinases (RTKs), integrins, and tumor suppressor proteins (17–20). Upon extracellular stimuli, FAK is activated by autophosphorylation of critical residue tyrosine 397 (Y397) through the FAK-kinase domain (21). Phosphorylated Y397 subsequently serves as a canonical SH2-domain docking site for SRC-family kinases, PI3K, and GRB7 (22–24). These signaling molecules then lead to the full activation of FAK through phosphorylation of effector residues (Y861 & Y925) as well as the activation of downstream oncogenic pathways ERK and AKT (25–27). In an attempt to inhibit FAK in cancer, numerous groups have developed FAK-kinase inhibitors which bind to the ATP-binding pocket of the FAK-kinase domain and block catalysis (28–31). Although very potent (low nM) and specific inhibitors of FAK-kinase activity have been discovered, limited efficacy has been observed in Phase I/II clinical trials with lack of mechanistic knowledge explaining this phenomenon (32–34). Furthermore, there are no current molecular markers available that successfully predict the response of FAK-kinase inhibitors in patients, nor are there biomarkers that would predict resistance to FAK-kinase inhibition. As there are 18 reported clinical trials involving FAK-kinase inhibitors (35), the ability to predict both response and resistance to FAK-kinase inhibitors is critical to clinical development. Along with the canonical FAK activation pathway through autophosphorylation, it is also known that FAK can be activated independently of FAK catalytic activity (19). PDGF- and EGF-stimulated FAK phosphorylation and cell migration occurs independently of FAK-kinase activity in a genetically modified FAK kinase-dead (K454R) model system. However, the tyrosine 397 site via the Y397F mutation was shown to be indispensable for FAK-dependent migration. Despite these data, it is still unknown how Y397 is activated by RTKs and whether multiple oncogenic RTKs can directly phosphorylate FAK at Y397 as a resistance mechanism to FAK-kinase inhibitors. In this report, we have investigated whether multiple oncogenic RTKs could re-activate FAK and therefore cause resistance to FAK-kinase inhibitors. We found that multiple classes of RTKs directly phosphorylated FAK at Y397 independently of FAK catalytic activity. FAK-kinase inhibitor treatment induced rapid RTK activation, leading to FAK reactivation as well as MAPK/AKT activation. Additionally, the initial presence of HER2 activity predicted resistance to FAK-kinase inhibitors in vitro. Together, these studies have identified a novel drug resistance mechanism to FAK-kinase inhibitors through transphosphorylation by RTKs. We call this “oncogenic protection” of FAK due to the variety of oncogenic RTKs that can phosphorylate Y397 to protect a cancer cell’s ability to maintain signaling through SH2 domain pathways. Furthermore, we believe that these results have identified a fragile point in FAK signaling pathways that can be exploited with precision-targeted therapeutics directed at FAK Y397 opposed to simply targeting the intrinsic FAK-kinase activity. Materials and Methods Cell Culture Cell lines MDA-MB-231, MDA-MB-453, and MCF7 (courtesy of Katerina Gurova), FAK−/− MEFs (courtesy of Duško Ilić), SYF−/− MEFs (courtesy of Irwin Gelman), H292 (courtesy of Pamela Hershberger), MCF7/HER2 tet-off cells (courtesy of Andrei Bakin) were cultured at 37°C with 5% CO2. H292 cells were cultured in RPMI 1640 (GIBCO, Life Technologies), 10% HI FBS (GIBCO, Life Technologies), 1% Pen Strep (GIBCO, Life Technologies), MEM Nonessential Amino Acids (cellgro, Mediatech, Inc.), Sodium Pyruvate (cellgro, Mediatech, Inc.), and HEPES (GIBCO, Life Technologies), and 0.2% Normocin (InvivoGen). Remaining cell lines were cultured in DMEM (GIBCO, Life Technologies), 10% HI FBS (GIBCO, Life Technologies), 1% Pen Strep (GIBCO, Life Technologies), and 0.2% Normocin (InvivoGen). Cells lines were not submitted for cell authentication services however were instead validated for proper HER2, EGFR, and FAK expression via western blotting and were tested for mycoplasma infection via the MycoAlert™ Mycoplasma Detection Kit (Lonza). Creation of HER2/HER3 stably transduced FAK−/− MEFs HER2 and HER3 retrovirus was prepared from 293T-Phoenix cells transfected with either pLXSN-HER2 or pLXSN-HER3 vector (courtesy of Dr. H. Shelton Earp III) similarly as described (36). Cultured FAK−/− MEFs were transduced with retrovirus and selected for HER2High/HER3High expression using FACS (Roswell Park Flow Cytometry Core). Antibodies used for selection were Alexa Fluor® 488 anti-human erbB2/HER-2 (cat # 324410, BioLegend) and APC-conjugated anti-hErbB3/Her3 (cat # FAB3481A, R&D Systems). Stable cells were subsequently transiently transfected with various HA-FAK constructs (WT, Y397F, K454R) using LipoD293™ (SignaGen® Laboratories) and processed for western blotting analysis. Protein Purification His-tagged avian FAK-FERM (AA 31-405) in modified pET vector (provided by Dr. Michael Eck) was expressed in BL21 (DE3) E. coli (Life Technologies) and purified on Ni-NTA resin (Thermo Scientific) using buffers as described (37). His-tagged human FAK-CD (AA 677-1052) was cloned into pET15b vector and similarly purified on Ni-NTA resin (Thermo Scientific). GST-FAK-NT (AA 1-415), GST-FAK-KD (AA 416-676), GST-FAK-CD (AA 677-1052), GST-FAK-NT1 (AA 1-126), GST-FAK-NT2 (AA 127-243), and GST-FAK-NT3 (244-415) protein in pGEX-4T1 vector were expressed in BL21 (DE3) E. coli and purified on Glutathione Sepharose 4B resin (GE Healthcare) as previously described (20, 38). GST-HER2-ICD (AA 676-1255), GST-HER2-ICD1 (AA 676-801), GST-HER2-ICD2 (AA 802-1029), and GST-HER2-ICD3 (AA 1030-1255) constructs were designed into the pGEX-4T1 vector, sequenced by the Roswell Park Sequencing Core, expressed in BL21 (DE3) E. coli, and purified on Glutathione Sepharose 4B resin (GE Healthcare). HER2-ECD protein (cat#BMS362) was purchased from eBiocience. Recombinant HER2-ICD (cat#PV3366), EGFR (cat#PV3872), Tie2 (cat#PV3628), EphA2 (cat#PV3688), and FGFR4 (cat#P3054) were purchased from Life Technologies. In vitro kinase assay Purified FAK-FERM (100ng) or FAK-CD (100ng) were incubated with purified HER2 (100ng) or additional RTK for 30 min in the presence or absence of ATP (10µM) in standard tyrosine kinase buffer: 20mM HEPES (pH=7.3), 5mM MgCl2, 5mM MnCl2, 2mM DTT, 0.1mg/mL BSA, and 0.1mM Na3VO4. Proteins were resolved on 4–20% gradient gels and probed for phosphorylated FAK (Y397), FAK (Y925), HER2 (Y1248), FAK-FERM, FAK-CD, and HER2/RTK using standard western blotting techniques. Pull Down Assays Purified GST-FAK or GST constructs were incubated in NP40 buffer plus 0.1% BSA with purified HER2-ECD and HER2-ICD. GST or GST-FAK constructs were pulled down using Glutathione Sepharose 4B (GE Healthcare Life Sciences) and washed three times with NP40 buffer. Proteins were eluted off of beads in 2X laemmli buffer (BioRad) and boiled. Samples were resolved on 4–20% gradient gels and probed for HER2-ECD or HER2-ICD using standard western blotting techniques and antibodies as described below. Secondary gels were run and stained with SimplyBlue™ SafeStain to confirm protein loading. PathScan® RTK Signaling Antibody Array Kit (Chemiluminescent Readout) Cell lysates were collected using NP40 lysis buffer containing protease (Roche) and phosphatase inhibitors (Roche) and were incubated on profiler slides according to manufacturer’s instructions (Cell Signaling) (map in Sup Methods). Slides were imaged via chemilumienescence on film and dot intensities were analyzed using ImageJ densitometry software and graphed in Graphpad Prism 6. Phosphorylation (relative to 0h) was quantified for each protein by subtracting the negative control dots within each panel, respectively. Subsequently, values were divided by signal obtained at 0h to obtain phosphorylation levels relative to 0h. Matrigel-on-top 3D growth assay Cells were plated at a density of 1,000 cells per well in a 1:50 solution of matrigel:complete DMEM culture media on top of a base matrix composed of 1:1 matrigel:DMEM media. Doxycycline (1µg/mL) was added for MCF7-HER2 Tet-Off cells. Defactinib was added at various concentrations the day after initial plating. Cell proliferation was evaluated after 5 days using CellTiter AQueous One Solution Cell Proliferation Assay (Promega) according to manufacturer’s instructions. Viability was plotted relative to DMSO control and IC50’s were calculated using Dose-response – Inhibition nonlinear regression algorithm (log(inhibitor) vs. response) in Graphpad Prism 6. Molecular Modeling Protein Data Bank files 2AL6 (crystal structure of FAK FERM domain), 3RCD (crystal structure of HER2 kinase domain), 4RIW (crystal structure of EGFR kinase domain), and 2J0J (crystal structure of FAK FERM-Kinase domains) were downloaded and utilized for all modeling experiments. “Active” surface residues determined from CPORT studies were utilized as input residues for restraints to drive the HADDOCK docking process (39). Restraints were only set to “active” residues that corresponded to FAK FERM F1 lobe and HER2 Kinase N-lobe sub-domains that were found to be required for HER2-FAK binding in experimental assays. EGFR restraints were solely based from CPORT results. The HADDOCK docking protocol was performed similarly as described (40). The top-ranking HADDOCK cluster (based on HADDOCK score and Z-score) was selected and visualized using PyMOL software (Incentive version). Images were ray-traced and saved for publication-quality purposes. Statistical analysis Comparisons between two groups were made using a Students t test (GraphPad Prism6). Data were considered significant when p<0.05. Two-way ANOVA and Tukey’s multiple comparison test were used to calculate significance when comparing multiple groups within the same experiment (GraphPad Prism6). Results Oncogenic Receptor Tyrosine Kinases HER2 and EGFR reactivate FAK in FAK-kinase inhibited cancer cells and activate AKT/ERK independent of FAK-kinase activity FAK has been shown to be primarily activated at Y397 by autophosphorylation through the FAK-kinase domain. Whereas some data suggest that RTKs can signal through FAK independently of FAK-kinase activity (19), it is still unclear how RTKs activate FAK and whether this alternative pathway plays a role in cancer cell resistance to FAK-kinase inhibitors. Thus, we hypothesized that oncogenic RTKs could directly transphosphorylate FAK Y397 as a drug resistance mechanism to bypass FAK-kinase inhibition. In order to test this, in vitro studies were carried out in HER2+ MDA-MB-453 and SkBr3 breast cancer cells as well as EGFR+ H292 and A549 lung cancer cells (Fig. 1, Supplementary Fig. S1, S2, and Table S1) treated with three different FAK-kinase inhibitors (defactinib, PF-228, and PF-271), two of which have been in clinical trials (28, 29, 32, 34, 41, 42). We used 10µM of FAK-kinase inhibitor treatment to ensure full inactivation of the FAK enzyme in a variety of cell lines (Supplementary Fig. S3 and S4). HER2 activation by Heregulin (HER2-activating ligand) reactivated FAK Y397 phosphorylation after treatment with all three FAK-kinase inhibitors in both MDA-MB-453 and SkBr3 cell lines, suggesting that FAK Y397 phosphorylation could be maintained independent of its intrinsic kinase activity. While the effects of EGF stimulation on FAK reactivation were not as robust as HRG stimulation, EGF stimulation slightly reactivated FAK pY397 after treatment with PF-228 in H292 cells. Nonetheless, growth factor stimulation activated both AKT and ERK pathways in the presence of all three FAK-kinase inhibitors in all four cell lines, indicating that FAK-kinase activity was dispensable for downstream pathway activation. Conversely, FAK-depletion using FAK-null MEFs or FAK siRNA in MDA-MB-453 cells partially diminished HER2-dependent activation of AKT/ERK, indicating that FAK total protein, however not kinase activity, could regulate downstream AKT/ERK pathways (Supplementary Fig. S5). We also noticed that even in the absence of HRG or EGF stimulation, all three FAK kinase inhibitors stimulated phosphorylation of HER2/AKT in MDA-MB-453 cells as well as HER2/AKT/ERK in SkBr3 cells and EGFR/AKT/ERK in A549 cells (Fig. 1). These data indicated a novel drug resistance pathway to FAK-kinase inhibitors through the reactivation of FAK Y397 and downstream AKT/ERK pathways by oncogenic RTKs. To verify that the compensatory increases in HER2, AKT, and ERK after FAK-kinase inhibitor treatment were in fact due to FAK-kinase inhibition and not off-target kinase effects of the 10µM dose, we revisited our dose-titration experiments using clinical-stage FAK inhibitor, defactinib (Supplementary Fig. S3 and S4). In MDA-MB-453 and SkBr3 cells, once FAK pY397 levels were decreased by defactinib treatment (0.3–1.0µM and 0.004µM respectively), compensatory increases in pHER2, pAKT, and pERK were also observed. In MDA-MB-453 cells, maximal FAK inhibition (10–30µM) caused maximal compensatory increases in pAKT and pERK. Intriguingly, at higher doses of defactinib (3µM) in SkBr3 cells, pHER2 and pAKT levels were maximally increased and FAK pY397 levels were paradoxically restored. In H292 and A549 cells, compensatory increases in pEGFR and pERK were observed after low dose defactinib treatment (0.001–1µM). We also noticed that 0.01–0.3µM defactinib treatment in H292 cells induced paradoxical increases in FAK pY397 that were in alignment with increases in pEGFR, suggesting that FAK-kinase inhibition led to FAK hyperphosphorylation by EGFR. To further confirm the FAK-specificity of these compensatory increases, we performed FAK siRNA studies in SkBr3 cells using all three FAK-kinase inhibitors (Supplementary Fig. S6). FAK knockdown by siRNA itself induced the phosphorylation of AKT, ERK, and the reactivation of FAK pY397. In addition, FAK siRNA reduced the relative increase in pAKT and pERK after FAK-kinase inhibitor treatment. In all, these data supported that the compensatory increases in pHER2, pAKT, and pERK upon FAK-kinase inhibition were indeed due to the specific inhibition of FAK. HER2 and EGFR phosphorylate kinase-dead FAK to maintain cell migration and invasion Next, we used genetic studies in FAK-null MEFs to confirm the phenomenon of RTK-driven Y397 phosphorylation (Fig. 2A, 2B, and Supplementary Fig. S7). Both WT-FAK and K454R-FAK (kinase-dead) were found to be re-phosphorylated at Y397 upon stimulation of both HER2/HER3-stably expressing and EGFR+ FAK-null MEFs. Additionally, studies in SRC/YES/FYN (SYF)-null MEFs confirmed FAK Y397 phosphorylation by HER2 independently of known SYF adaptor proteins (Supplementary Fig. S8). These data demonstrated that Y397 can still be phosphorylated by RTKs independently of both FAK and SRC activity, consistent with the RTK-dependent drug resistance we observed with all three FAK-kinase inhibitors. Because HER2 reactivated kinase-dead FAK phosphorylation at Y397 in MEFs, we then hypothesized that HER2 could maintain FAK-dependent biological functions as well. To test this, we performed transwell migration and invasion assays utilizing HER2/HER3-stably expressing FAK-null MEFs that were transiently transfected with WT-FAK, K454R-FAK, or Y397F-FAK. As shown in Figure 2C, HRG-stimulated migration was enhanced by not only WT-FAK, but K454R-FAK as well. Conversely, only mutation of Y397 directly (Y397F-FAK) was found to completely block HRG-stimulated migration. Intriguingly, basal cell migration (-HRG) was fully inhibited by K454R mutation, suggesting that basal migration, but not HER2-stimulated migration, is regulated by FAK-kinase activity. These findings were confirmed in cell invasion assays as well (Figure 2D). Kinase-dead FAK (K454R) moderately reduced HRG-stimulated invasion compared to WT-FAK, however only Y397F-FAK completely abrogated FAK-enhanced cell invasion. These data showed that HER2 can maintain FAK-dependent biological functions (migration and invasion) under kinase inhibition as a result of the FAK transphosphorylation phenomenon. HER2 directly binds to FAK as a structural mechanism to promote Y397 phosphorylation Because FAK was reactivated by RTKs independently of both FAK and SRC activity, we hypothesized that RTKs were directly binding to FAK to promote phosphorylation. Although HER2 has been previously shown to co-localize and form a complex with FAK in cells, it is still unknown whether HER2 and FAK form a direct protein-protein interaction (43, 44). To further confirm the mechanism of direct FAK activation by HER2, we performed a series of GST pull-down assays with purified recombinant proteins. GST-FAK constructs were cloned according to FAK-NT (N-terminus), -KD (Kinase Domain), or -CD (C-terminal Domain) regions and fusion-proteins were incubated with either HER2-ICD (Intracellular Domain) or -ECD (Extracellular Domain) proteins (Fig. 3A and Supplementary Fig. S9A). Significant binding was detected between GST-FAK-NT and HER2-ICD, but minimal binding was observed for FAK-KD and –CD regions. As expected, no binding was detected between GST-FAK constructs and HER2-ECD protein. To further define the binding region between HER2 and FAK, we performed similar assays with GST-FAK-NT1, NT2, and NT3 proteins cloned according to the F1, F2, and F3 lobes of the FAK N-terminal FERM domain (Fig. 3B and Supplementary Fig. S9B). The FAK NT1 (or F1 lobe) region, but not NT2 or NT3, was found to bind to HER2-ICD. Additionally, we tested various HER2 segments with GST-HER2-ICD, ICD1, ICD2, and ICD3 proteins cloned according to whole ICD, kinase domain N-lobe, kinase domain C-lobe, and c-terminal tail regions of the HER2 ICD (Fig. 3B and Supplementary Fig. S9C). Both the HER2-ICD and ICD1 (or kinase N-lobe) regions, but not ICD2 or ICD3, were found to bind to FAK-NT. These data confirmed the direct interaction of FAK and HER2 between the FAK FERM F1 lobe and HER2 kinase N-lobe. We then examined the x-ray crystal structure of the FAK-FERM domain (PDB 2AL6) and observed that Y397, located within the flexible FERM-Kinase linker region, binds back onto the structured FERM F1 lobe where HER2 interacts (Supplementary Fig. S10). Since Fl lobe binding could provide proximity of HER2 to Y397, we utilized HADDOCK protein-protein docking to develop a structural model for HER2-FAK binding and putative Y397 transphosphorylation (40). FAK FERM (PDB 2AL6) and HER2 kinase domain (PDB 3RCD) crystal structures were computationally analyzed for active surface residues likely to be involved in protein-protein interactions using the program CPORT (39). Subsequently, CPORT-predicted residues that overlapped with experimentally validated amino acids from GST pull-down assays were set as restraints to drive the HER2-FAK docking model. Intriguingly, HADDOCK produced a HER2-FAK structural model similar to a prototypical kinase-substrate interaction (Fig. 3C). As observed in Fig. 3D, the FAK FERM F1 lobe makes contact with the HER2 N-lobe to orient the flexible Y397 FERM-linker segment close to the HER2 substrate-binding region and ATP-binding pocket. HADDOCK docking was also performed similarly with the EGFR kinase domain (PDB 4RIW) to approximate a model between EGFR and FAK (Supplementary Fig. S11). In agreement with the HER2 model, the FAK FERM domain was predicted to bind EGFR through the F1 lobe, facilitating orientation of Y397 into the EGFR ATP-binding pocket. These data provided structural rationale for the HER2-FAK interaction, where binding of HER2 (and potentially other RTKs) to the F1 lobe promotes proximity to Y397, allowing HER2 to directly phosphorylate Y397. HER2, EGFR, and additional RTKs directly phosphorylate FAK FERM at Y397 Direct HER2-FAK binding assays as well as our HADDOCK docking model suggested that HER2, and potentially other RTKs, could directly activate FAK at Y397 (Fig. 3). To evaluate direct phosphorylation and experimentally validate our HADDOCK model, we tested whether HER2 could directly phosphorylate FAK using an in vitro kinase assay with purified HER2-ICD, FAK-FERM, and FAK-CD proteins (Fig. 4A and 4B). In agreement with our HADDOCK model, HER2 directly phosphorylated FAK-FERM specifically at Y397. However, no phosphorylation was detected at FAK-CD Y925, confirming the specificity of the kinase reaction. Intriguingly, multiple RTKs (EGFR, Tie2, EphA2, FGFR4) were also found to directly phosphorylate FAK at Y397 (Fig. 4C), indicating a redundant FAK activation pathway by multiple RTKs. These data supported a direct Y397 phosphorylation mechanism by HER2 and other RTKs, where HER2 directs interacts with the FAK-FERM F1 lobe to phosphorylate Y397. Small molecule inhibition of FAK kinase activity induces compensatory RTK reprogramming As mentioned previously, even in the absence of growth factor stimulation, all three FAK-kinase inhibitors stimulated the phosphorylation of RTK/AKT/ERK pathways as a drug resistance mechanism (Fig. 1). Given our data of FAK reactivation by RTKs, we hypothesized that cancer cells might induce novel RTK signatures (rapid vs. long-term depending on initial RTK expression) in response to FAK-kinase inhibitors as a mechanism to maintain FAK and AKT/ERK signaling. To test this, we performed time course experiments in RTKHigh cell lines (MDA-MB-453, SkBr3, H292) as well as RTKLow cell lines (MDA-MB-231, MDA-MB-468) and evaluated changes in phosphorylation patterns. We focused on defactinib due to its current involvement in multiple ongoing clinical trials and its recently failed Phase II clinical trial (34). Additionally, we used a dose of 1µM, which is in alignment with patient serum concentrations of defactinib achieved with the recommended phase II dose of 425 mg BID (42). In MDA-MB-453 cells, defactinib treatment induced rapid activation of HER2, EGFR, AKT, and ERK within 15 min, with maximal activation of HER2 and EGFR coming at 1 and 4 hours, respectively (Fig. 5A and Supplementary Fig. S12A). When FAK-kinase inhibition was lost and pY397 levels were restored at 48–72 hours, pHER2 and pEGFR levels also returned to basal levels. In addition, we utilized RTK arrays to assess a broader kinome response to FAK-kinase inhibition (Fig. 5, Supplementary Fig. S13, S14, and Table S2). In H292 cells, we found consistent rapid activation of RTK pathways as observed in MDA-MB-453 cells, with increases in EGFR, HER2, ERK, and AKT from 15 min to 4 hours post treatment. Immunoblots were used to confirm our findings of EGFR, AKT, and ERK activation after defactinib treatment (Fig. 5B and Supplementary Fig. S12B). Intriguingly, S6 ribosomal protein (S6 RP), a marker of increased translation of mRNA transcripts encoding for proteins regulating cell cycle progression, was rapidly activated at 15–30 minutes in two of three RTKHigh cell lines. These data showed that RTKHigh cancer cells are involved in a rapid feedback loop, where FAK-kinase inhibition induces corresponding RTK activation to compensate for loss of FAK Y397 phosphorylation. Conversely, in RTKLow cell lines, the RTK reprogramming in response to defactinib treatment occurred at longer timepoints (48 and 72 hours). In MDA-MB-231 triple-negative (ER/PR/HER2−) breast cancer cells, defactinib induced the expression of HER2 and EGFR as well as the activation of other RTKs such as FGFR4 and EphA2 at 72 hours. Immunoblots were used to confirm to total protein increases in both EGFR and HER2 after defactinib treatment (Fig. 5C and Supplementary Fig. 12C). These data demonstrated that RTKLow cancer cells require longer time points to undergo kinome changes, most likely due to the requirement to express new RTKs. Additionally, these data showed that the selective pressure of FAK-kinase inhibition is able to drive triple-negative breast cancer cells to express HER2. RTK positivity predicts resistance to FAK kinase inhibitors Based on our findings of rapid vs. long-term RTK reprogramming in RTKHigh vs. RTKLow cells, we predicted that cancer cells driven by oncogenic RTKs would be more resistant to FAK inhibition than those that are not driven by RTKs. We utilized 3D matrigel-on-top growth assays to measure of the efficacy of FAK kinase inhibitors, as the 3D-environment of these assays were shown to recapitulate the cellular requirement for FAK (45, 46). Indeed, HER2-addicted breast cancer cell lines MDA-MB-453 (IC50 ND) and SkBr3 (IC50 > 10µM) showed no changes in viability in response to defactinib treatment in 3D growth assays, whereas defactinib treatment decreased the viability of HER2− cell line MDA-MB-231 (IC50 = 0.281µM) in a dose-responsive manner (Fig. 6A). We confirmed these findings using the MCF7-HER2 Tet-Off system, where HER2 expression was modulated by the simple removal or addition of doxycycline in this isogenic system (Fig. 6B). HER2High (-Dox) cells showed minimal response to defactinib (IC50 = 1.58µM), whereas HER2Low (+Dox) cells were quite responsive to defactinib treatment (IC50 = 0.052µM). Together, these data showed that HER2 expression levels can distinguish cancer cells that may or may not respond to FAK-kinase inhibitors and demonstrated that RTKs such as HER2 can drive resistance to FAK-kinase inhibitors. Discussion The data described in this report have identified a novel mechanism of drug resistance that may explain some of the failures of FAK-kinase inhibitors in clinical trials. Here, we have shown that oncogenic RTKs, such as EGFR or HER2, directly phosphorylated FAK at Y397 to rescue kinase-inhibited FAK in cancer cells. Additionally, HER2 formed a direct protein-protein interaction complex with FAK nearby this critical residue Y397. We also have shown that RTKHigh cell lines displayed rapid resistance to FAK-kinase inhibitors due to upregulation of RTK signaling pathways while RTKLow cells were initially sensitive to the inhibitors. However, RTKLow cell lines, such as MDA-MB-231, started to express HER2 and EGFR protein after longer incubations with defactinib, explaining a possible acquired-resistance mechanism to FAK-kinase inhibitors. An overview of our findings is summarized in Supplementary Fig. S15. In all, we have identified a common resistance mechanism to FAK-kinase inhibitors, whereby RTKs can bypass FAK-kinase inhibition to re-phosphorylate FAK Y397. In our analysis, we noticed that RTKHigh cell lines showed a trend whereby RTK activation led to increased FAK pY397, but we also observed that different RTKHigh cell lines had different capacities to re-phosphorylate FAK at Y397. Additionally, EGFR+ cells line, A549, showed minimal reduction of pY397 by FAK-kinase inhibitors and compensatory increases in pEGFR, suggesting an innate resistance to these inhibitors. Also observed was a temporal difference between cell lines in their ability to induce compensatory RTK reprogramming after defactinib treatment. To explain these results, we propose that the ability of RTKs to transphosphorylate FAK at Y397 is very cell line-dependent, where multiple variables (i.e. RTK levels, mutation status, adaptor protein levels) can affect both the levels and kinetics of Y397 phosphorylation. Future experiments will be designed to evaluate other cellular factors that may drive Y397 phosphorylation and FAK-kinase inhibitor resistance. Of additional notice was the strong compensatory activation of HER2/AKT/ERK pathways as a drug resistance mechanism to FAK-kinase inhibitors. We acknowledge that high-dose kinase inhibitor treatment has the potential for off-target kinase effects. Nonetheless, our data with low concentrations of defactinib and FAK-specific siRNA suggest that the compensatory RTK reprogramming after FAK-kinase inhibitor treatment was FAK-specific. Kinase inhibitor off-target kinase effects are generally inhibitory (due to the conserved ATP-binding region) and not agonistic in nature. In addition, the chemical structures of all three FAK-kinase inhibitors (defactinib, PF-271, PF-228) are different and therefore display different off-target kinase effects as observed (28, 29, 41). However, all three FAK-kinase inhibitors still activated HER2/EGFR and AKT/ERK in a similar manner, supporting the notion that these compensatory increases are FAK-specific. Further, as mentioned previously, the recommended phase II dose of defactinib is 425 mg BID, and at steady state dosing (day 15) the average serum Cmax was greater than 1.0 ug/mL, or 1.95 uM, indicating that our experimental doses are not only appropriate for cell culture studies, but are also in line with circulating levels of drug in patients (42). Therefore, we believe our data describe a clinically relevant drug resistance mechanism to FAK-kinase inhibitors at concentrations utilized in current clinical trials. Surprisingly, we found that multiple RTKs from different family-members could directly phosphorylate FAK FERM at Y397. Additionally, we found a wide array of upregulated RTKs in response to the FAK-kinase inhibitor in widespread clinical trials, defactinib. These findings suggest there are redundant transphosphorylation mechanisms where cancer cells can hijack multiple different RTK pathways (depending on cellular context and genomic accessibility) to maintain FAK Y397 pathway activation during FAK-kinase inhibition. Also, FAK depletion by siRNA in SkBr3 cells induced the hyperactivation of FAK Y397 despite lower levels of total FAK, implying cancer cell dependency on FAK Y397 phosphorylation. Therefore, we propose FAK Y397 phosphorylation as a fragile point in cancer cells, required to maintain cancer cell motility and proliferation. We have termed our observation “oncogenic protection of FAK” because of the ability of different oncogenic RTKs to phosphorylate and protect this fragile Y397 site. In fact, several other groups have shown the requirement of FAK for tumor initiation and progression (12, 13). Additionally, another group has implicated the FERM domain in binding of FAK to RTKs in cells, but did not show a direct interaction (19). We have demonstrated a direct interaction of FAK and HER2 and further hypothesize that multiple RTKs bind to the FAK-F1 lobe as a means to directly phosphorylate Y397. Although our direct kinase and cellular data support the redundant phosphorylation of FAK Y397 by multiple RTKs, we acknowledge the known promiscuity issue with in vitro kinase assays (47). Future cellular evaluation of novel RTK-FAK signaling complexes as well as in vivo drug resistance models will be helpful as we seek to use precision RTK biomarkers of a patient’s tumor to predict their response/resistance to FAK kinase inhibitors. Several recent reports have characterized a novel drug resistance mechanism to kinase inhibitors involving rapid dynamic kinase signaling called kinome reprogramming (48–50). These groups have identified global changes in cancer cell kinome expression and phosphorylation events after kinase inhibitor treatment. Interestingly, MEK-inhibitor and HER2-inhibitor treatment induced global upregulation of redundant intracellular signaling nodes, RTKs, and MAPK/AKT signaling pathways. Our data describe a novel and previously uncharacterized form of compensatory RTK reprogramming to FAK-kinase inhibitors through the re-phosphorylation of FAK Y397 but are in agreement with these studies where other kinase inhibitors induced global changes in kinase expression/activity. It remains to be determined whether FAK-kinase inhibition induces both whole kinome as well as non-kinase forms of drug resistance. At the time this manuscript was prepared, there were 18 reported clinical trials involving FAK inhibitors on clinicaltrials.gov (35). Although FAK is considered a promising drug target for cancer therapy, questions still remain on which function of FAK (kinase vs. scaffold) is optimal to target and under which molecular context (17). Several pre-clinical studies have shown that molecular markers, such as MerlinLow expression, can be utilized to stratify tumor sub-types which may be responsive to FAK-kinase inhibitors (46). Unfortunately, clinical trial results of the COMMAND (Control Of Mesothelioma with MAiNtenance Defactinib) trial have shown limited efficacy and no difference in control vs. defactinib-treated patients with MerlinLow tumors (34). As such, it is important to identify new molecular markers, such as HER2 and other RTKs that may have the ability to predict which population of patients will or will not respond to FAK-kinase inhibitor therapy. Finally, the data presented here have several direct clinical implications regarding FAK-kinase inhibitors. First, we show that expression of RTKs in cancer cells, such as HER2, can cause resistance to FAK-kinase inhibitors due to re-phosphorylation of Y397. This suggests an opportunity to retrospectively analyze tumor samples from patients who have received FAK-kinase inhibitors to determine the levels of specific RTKs such as HER2 and EGFR and correlate them with clinical response. Second, our data suggest that RTKLow cancers will be initially sensitive to FAK-kinase inhibitors and that RTK expression can be induced as a mechanism of acquired resistance. This suggests the need to regularly monitor tumors for changes in their kinome that could predict acquired resistance to the FAK-kinase inhibitor. Third, these studies provide a molecular rationale for combination therapy of FAK-kinase inhibitors with RTK inhibitors such as lapatinib or erlotinib. However, we predict eventual resistance to combination therapy, due to the observed upregulation of diverse RTKs in RTK reprogramming assays. As an alternative, the discovery and development of FAK-scaffold inhibitors that directly inhibit FAK Y397 may provide a more durable response to FAK-based therapy (51) by blocking the fragile Y397 site to globally prevent its phosphorylation. As we further understand the dynamic nature that surrounds FAK and its complex interactome, we believe that more precision-targeted therapeutics will be elucidated. Supplementary Material 1 2 3 We would like to thank Drs. Vita Golubovskaya, Elena Kurenova, and Shelton Earp for scientific discussions regarding the research topic, Mr. Thomas Mathers for pharmaceutical industry advice, the Roswell Park Flow Cytometry Core for performing FACS analysis, and Mr. George Clinton for creative influences. Funding: This work was funded by the NCI (R01CA065910) awarded to W.G. Cance, a Ruth L. Kirschstein National Research Service Award (NRSA) Institutional Research Training Grant (T32CA009072) awarded to T.A. Marlowe, and a NCI Cancer Center Support Grant for Roswell Park Cancer Institute (P30CA016056). Abbreviations FAK Focal Adhesion Kinase RTKs Receptor Tyrosine Kinases HER2 Human Epidermal Growth Factor Receptor 2 EGFR Human Epidermal Growth Factor Receptor FERM 4.1 Ezrin Radixin Moesin Fig. 1 HER2 and EGFR phosphorylate FAK Y397 independently of FAK-kinase activity Immunoblots showing phosphorylation of FAK and downstream molecules in HER2+ breast cancer cells (A) MDA-MB-453 and (C) SkBr3, as well as EGFR+ lung cancer cells (B) H292 and (D) A549. Cancer cells were serum-starved overnight followed by drug treatment with FAK-kinase inhibitor (defactinib, PF-228, PF-271) or HER2/EGFR inhibitor (Lapatinib) for 1 hour and stimulated with either Heregulin-β1 (HRG) or Epidermal Growth Factor (EGF) for 30 min. Images shown are representative of three independent experiments. Densitometry and statistical analysis is found in Supplementary Fig. S1, S2, and Table S1. Fig. 2 HER2 and EGFR phosphorylate kinase-dead FAK to maintain cell migration and invasion Immunoblots showing phosphorylation of FAK in (A) HER2+/HER3+ MEFFAK-/FAK-, and (B) EGFR+ MEFFAK-/FAK- cells. MEFs transfected with FAK constructs (WT, K454R-kinase dead, and Y397F) were serum-starved overnight followed by drug treatment with FAK-kinase inhibitor (defactinib) for 1 hour and stimulated with either Heregulin-β1 (HRG) or Epidermal Growth Factor (EGF) for 30 min. Both HER2 and EGFR activation partially reactivated kinase-dead FAK phosphorylation at Y397. Images shown are representative of three independent experiments. Densitometry and statistical analysis is found in Supplementary Fig. S7. (C) Transwell migration and (D) invasion assays performed in transfected HER2+/HER3+ MEFFAK-/FAK- cells. Transfected MEFs were serum-starved overnight and placed into upper chamber to allow chemotaxis towards BSA (-HRG) or HRG (+HRG) placed into the lower chamber. Assays were performed for 3 hours (migration) or 24 hours (invasion). Western blot panels in upper right demonstrate levels of transfected FAK. Two-way ANOVA statistical analysis was performed for results from three independent experiments. Asterisks represent multiple comparison corrected p-values < 0.05 in comparison to pcDNA. Fig. 3 HER2 directly binds FAK as a structural mechanism to promote Y397 phosphorylation (A) GST pull-down assays demonstrating direct binding of FAK-NT (N-terminal domain) to HER2-ICD (Intracellular domain). GST, GST-FAK-NT, GST-FAK-KD (Kinase domain), and GST-FAK-CD (C-terminal domain) proteins were incubated with HER2-ECD (Extracellular domain) or HER2 ICD (Intracellular domain) proteins and resulting immunoblots represent the bound fraction. (B) Upper panel: GST pull-down assays showing direct binding of FAK-NT1 (FERM-F1 lobe) to HER2-ICD. GST-FAK-NT1, GST-FAK-NT2, and GST-FAK-NT3 were cloned according to FAK-FERM-F1, -F2, and -F3 lobes, respectively. Lower panel: GST pull-down assays showing direct binding of HER2-ICD1 (kinase N-lobe) to FAK-NT (FERM). GST-HER2-ICD, -ICD1, -ICD2, and -ICD3 were cloned according to HER2-ICD, kinase N-lobe, kinase C-lobe, and c-terminal tail regions, respectively. Images shown are representative of three independent experiments. (C) Structural model of the HER2-FAK interaction as determined by HADDOCK docking studies. The N-lobe (cyan) and C-lobe (magenta) of the HER2 kinase domain (PDB 3RCD) as well as the F1 lobe (blue), F2 lobe (green), F3 lobe (red), and FERM-kinase linker (yellow) of the FAK FERM domain (PDB 2AL6) are depicted. The model shown represents the top-ranked cluster in HADDOCK. (D) Zoomed inset of the HER2-FAK model showing proximity of the binding interface to Y397 of the FERM-kinase linker region and orientation of Y397 within the HER2 substrate-binding region. All images were created in PyMOL Fig. 4 HER2, EGFR and additional RTKs directly phosphorylate FAK at Y397 (A) Direct in vitro kinase assay between HER2-ICD and FAK-FERM domain purified proteins showing direct phosphorylation of FAK Y397 by HER2 with the addition of ATP. (B) Direct in vitro kinase assay between HER2, SRC, and FAK-CD domain purified proteins showing the direct phosphorylation of Y925 only by SRC (positive control) but not HER2 with the addition of ATP. (C) Direct in vitro kinase assay between EGFR, Tie2, EphA2, and FGFR4 and FAK-FERM domain purified proteins showing direct phosphorylation of FAK Y397 by RTKs with the addition of ATP. Images shown are representative of three independent experiments. Fig. 5 Cancer cells demonstrate rapid and long-term compensatory RTK reprogramming to defactinib (A) Timecourse experiment where HER2+ breast cancer cells (MDA-MB-453) were treated with FAK-kinase inhibitor (defactinib) for the indicated timepoints. Note; defactinib induced EGFR/HER2 and AKT/ERK phosphorylation within 15 min of treatment. RTK arrays were performed (right panel) with matched immunoblot confirming results. (B) Timecourse experiment where EGFR+ lung cancer cells (H292) were treated with defactinib showing rapid compensatory RTK reprogramming. RTK arrays (right panel) showed activation of EGFR, ERK, and S6RP within 30 min. (C) Timecourse experiment where triple-negative breast cancer cells (MDA-MB-231) were treated with defactinib showing long-term (72h) RTK reprogramming in MDA-MB-231 cells. Note, initially HER2− cells were found to have low levels of HER2 at 2–4 hours and high levels at 48–72 hours after defactinib treatment. RTK arrays (right panel) showed upregulation of additional RTKs, FGFR4 and EphA2, 72 hours after treatment. Immunoblot images shown are representative of three independent experiments. Densitometry and statistical analysis is found in Fig. S12. RTK array results represent results from one RTK array chip (Supplementary Table S2). Fig. 6 HER2-positive cancer cells are resistant to FAK-kinase inhibition Matrigel on-top 3D growth assays with (A) a panel of breast cancer cell lines and (B) MCF7-HER2 isogenic line (Tet-Off) treated with titrating doses of defactinib. HER2 positivity confers resistance to defactinib treatment (IC50 values shown below graph). Viability values shown are averaged results from triplicate measurements in three independent experiments. Error bars shown represent 95% confidence intervals. The immunoblot right of (B) shows HER2 knockdown with 1µg/mL doxycycline treatment. Competing interests: We declare no financial conflicts of interest. W.G. Cance is Chief Scientific Officer of FAKnostics, LLC, a company focused on FAK biomarkers and therapeutics. References 1 Weiner TM Liu ET Craven RJ Cance WG Expression of focal adhesion kinase gene and invasive cancer Lancet 1993 342 1024 1025 8105266 2 Owens LV Xu L Craven RJ Dent GA Weiner TM Kornberg L Overexpression of the focal adhesion kinase (p125FAK) in invasive human tumors Cancer Research 1995 55 2752 2755 7796399 3 Lark AL Livasy CA Calvo B Caskey L Moore DT Yang X Overexpression of focal adhesion kinase in primary colorectal carcinomas and colorectal liver metastases: immunohistochemistry and real-time PCR analyses Clin Cancer Res 2003 9 215 222 12538472 4 Hanahan D Weinberg Robert A Hallmarks of Cancer: The Next Generation Cell 2011 144 646 674 21376230 5 Schaller MD Borgman CA Cobb BS Vines RR Reynolds AB Parsons JT pp125fak a structurally distinctive protein-tyrosine kinase associated with focal adhesions Proceedings of the National Academy of Sciences of the United States of America 1992 89 5192 5196 1594631 6 Hanks SK Calalb MB Harper MC Patel SK Focal adhesion protein-tyrosine kinase phosphorylated in response to cell attachment to fibronectin Proceedings of the National Academy of Sciences of the United States of America 1992 89 8487 8491 1528852 7 Siesser PMF Hanks SK The Signaling and Biological Implications of FAK Overexpression in Cancer Clin Cancer Res %R 101158/1078-0432CCR-06-0456 2006 12 3233 3237 8 McLean GW Carragher NO Avizienyte E Evans J Brunton VG Frame MC The role of focal-adhesion kinase in cancer - a new therapeutic opportunity Nat Rev Cancer 2005 5 505 515 16069815 9 Zhao X Guan JL Focal adhesion kinase and its signaling pathways in cell migration and angiogenesis Advanced drug delivery reviews 2011 63 610 615 21118706 10 Luo M Fan H Nagy T Wei H Wang C Liu S Mammary Epithelial-Specific Ablation of the Focal Adhesion Kinase Suppresses Mammary Tumorigenesis by Affecting Mammary Cancer Stem/Progenitor Cells Cancer Res 2009 69 466 474 19147559 11 Serrels A Lund T Serrels B Byron A McPherson RC von Kriegsheim A Nuclear FAK controls chemokine transcription, Tregs, and evasion of anti-tumor immunity Cell 2015 163 160 173 26406376 12 McLean GW Komiyama NH Serrels B Asano H Reynolds L Conti F Specific deletion of focal adhesion kinase suppresses tumor formation and blocks malignant progression Genes Dev 2004 18 2998 3003 15601818 13 Lahlou H Sanguin-Gendreau V Zuo D Cardiff RD McLean GW Frame MC Mammary epithelial-specific disruption of the focal adhesion kinase blocks mammary tumor progression Proc Natl Acad Sci U S A 2007 104 20302 20307 18056629 14 Provenzano PP Inman DR Eliceiri KW Beggs HE Keely PJ Mammary epithelial-specific disruption of focal adhesion kinase retards tumor formation and metastasis in a transgenic mouse model of human breast cancer Am J Pathol 2008 173 1551 1565 18845837 15 Pylayeva Y Gillen KM Gerald W Beggs HE Reichardt LF Giancotti FG Ras- and PI3K–dependent breast tumorigenesis in mice and humans requires focal adhesion kinase signaling The Journal of Clinical Investigation 2009 119 252 266 19147981 16 Golubovskaya VM Kweh FA Cance WG Focal adhesion kinase and cancer Histol Histopathol 2009 24 503 510 19224453 17 Cance WG Kurenova E Marlowe T Golubovskaya V Disrupting the scaffold to improve focal adhesion kinase-targeted cancer therapeutics Science Signaling 2013 6 e10 18 Zachary I Rozengurt E Focal adhesion kinase (p125FAK): a point of convergence in the action of neuropeptides, integrins, and oncogenes Cell 1992 71 891 894 1458538 19 Sieg DJ Hauck CR Ilic D Klingbeil CK Schaefer E Damsky CH FAK integrates growth-factor and integrin signals to promote cell migration Nat Cell Biol 2000 2 249 256 10806474 20 Golubovskaya VM Finch R Cance WG Direct Interaction of the N-terminal domain of Focal Adhesion Kinase with the N-terminal transactivation domain of p53 J Biol Chem 2005 280 25008 25021 15855171 21 Mitra SK Hanson DA Schlaepfer DD Focal adhesion kinase: in command and control of cell motility Nat Rev Mol Cell Biol 2005 6 56 68 15688067 22 Schaller MD Hildebrand JD Shannon JD Fox JW Vines RR Parsons JT Autophosphorylation of the focal adhesion kinase, pp125FAK, directs SH2-dependent binding of pp60src Molecular & Cellular Biology 1994 14 1680 1688 7509446 23 Chen HC Appeddu PA Isoda H Guan JL Phosphorylation of tyrosine 397 in focal adhesion kinase is required for binding phosphatidylinositol 3-kinase Journal of Biological Chemistry 1996 271 26329 26334 8824286 24 Han DC Guan JL Association of focal adhesion kinase with Grb7 and its role in cell migration J Biol Chem 1999 274 24425 24430 10446223 25 Calalb MB Zhang X Polte TR Hanks SK Focal adhesion kinase tyrosine-861 is a major site of phosphorylation by Src Biochemical & Biophysical Research Communications 1996 228 662 668 8941336 26 Schlaepfer DD Hunter T Evidence for in vivo phosphorylation of the Grb2 SH2-domain binding site on focal adhesion kinase by Src-family protein-tyrosine kinases Molecular & Cellular Biology 1996 16 5623 5633 8816475 27 Xia H Nho RS Kahm J Kleidon J Henke CA Focal adhesion kinase is upstream of phosphatidylinositol 3-kinase/Akt in regulating fibroblast survival in response to contraction of type I collagen matrices via a beta 1 integrin viability signaling pathway J Biol Chem 2004 279 33024 33034 15166238 28 Slack-Davis JK Martin KH Tilghman RW Iwanicki M Ung EJ Autry C Cellular characterization of a novel focal adhesion kinase inhibitor J Biol Chem 2007 282 14845 14852 17395594 29 Roberts WG Ung E Whalen P Cooper B Hulford C Autry C Antitumor activity and pharmacology of a selective focal adhesion kinase inhibitor, PF-562,271 Cancer Res 2008 68 1935 1944 18339875 30 Tanjoni I Walsh C Uryu S Tomar A Nam JO Mielgo A PND-1186 FAK inhibitor selectively promotes tumor cell apoptosis in three-dimensional environments Cancer Biol Ther 2010 9 764 777 20234191 31 Shi Q Hjelmeland AB Keir ST Song L Wickman S Jackson D A novel low-molecular weight inhibitor of focal adhesion kinase, TAE226, inhibits glioma growth Mol Carcinog 2007 46 488 496 17219439 32 Infante JR Camidge DR Mileshkin LR Chen EX Hicks RJ Rischin D Safety, pharmacokinetic, and pharmacodynamic phase I dose-escalation trial of PF-00562271, an inhibitor of focal adhesion kinase, in advanced solid tumors J Clin Oncol 2012 30 1527 1533 22454420 33 Gan H Nebot N Soria JC Arkenau HT Blagden S Plummer R Pharmacokinetics/pharmacodynamics (PK/PD) of GSK2256098, a Focal Adhesion Kinase (FAK) Inhibitor, in Patients with Advanced Solid Tumors 2012 24th EORTC-NCI-AACR Symposium on ‘Molecular Targets and Cancer Therapeutics’ 6 to 9 November 2012 Dublin, Ireland 34 Verastem, Inc Placebo Controlled Study of VS-6063 in Subjects With Malignant Pleural Mesothelioma (COMMAND) ClinicalTrials.gov [Internet] 2000 cited 2016 April 24 Bethesda MD National Library of Medicine (US) Available from: http://clinicaltrials.gov/ct2/show/record/NCT01870609 NLM Identifier: NCT01870609 35 U.S. National Institutes of Health ClinicalTrials.gov [Internet] 2000- cited 2016 April 24 Bethesda MD National Library of Medicine (US) Search: “FAK inhibitor”. Available from: http://clinicaltrials.gov/ct2/results?term=FAK+inhibitor&Search=Search 36 Sartor CI Zhou H Kozlowska E Guttridge K Kawata E Caskey L Her4 mediates ligand-dependent antiproliferative and differentiation responses in human breast cancer cells Mol Cell Biol 2001 21 4265 4275 11390655 37 Ceccarelli DF Song HK Poy F Schaller MD Eck MJ Crystal structure of the FERM domain of focal adhesion kinase J Biol Chem 2006 281 252 259 16221668 38 Zheng D Kurenova E Ucar D Golubovskaya V Magis A Ostrov D Targeting of the protein interaction site between FAK and IGF-1R Biochemical and Biophysical Research Communications 2010 388 3015 39 de Vries SJ Bonvin AM CPORT: a consensus interface predictor and its performance in prediction-driven docking with HADDOCK PLoS One 2011 6 e17695 21464987 40 de Vries SJ van Dijk M Bonvin AM The HADDOCK web server for data-driven biomolecular docking Nat Protoc 2010 5 883 897 20431534 41 National Center for Biotechnology Information PubChem Compound Database; CID=25117126 accessed Aug. 9, 2016 https://pubchem.ncbi.nlm.nih.gov/compound/25117126 42 Jones SF Siu LL Bendell JC Cleary JM Razak AR Infante JR A phase I study of VS-6063, a second-generation focal adhesion kinase inhibitor, in patients with advanced solid tumors Invest New Drugs 2015 33 1100 1107 26334219 43 Benlimame N He Q Jie S Xiao D Xu YJ Loignon M FAK signaling is critical for ErbB-2/ErbB-3 receptor cooperation for oncogenic transformation and invasion J Cell Biol 2005 171 505 516 16275754 44 Wang SE Xiang B Zent R Quaranta V Pozzi A Arteaga CL Transforming growth factor beta induces clustering of HER2 and integrins by activating Src-focal adhesion kinase and receptor association to the cytoskeleton Cancer Res 2009 69 475 482 19147560 45 Shibue T Brooks MW Inan MF Reinhardt F Weinberg RA The outgrowth of micrometastases is enabled by the formation of filopodium-like protrusions Cancer Discov 2012 2 706 721 22609699 46 Shapiro IM Kolev VN Vidal CM Kadariya Y Ring JE Wright Q Merlin deficiency predicts FAK inhibitor sensitivity: a synthetic lethal relationship Science translational medicine 2014 6 237ra68 47 Manning BD Cantley LC Hitting the target: emerging technologies in the search for kinase substrates Sci STKE 2002 2002 e49 48 Duncan JS Whittle MC Nakamura K Abell AN Midland AA Zawistowski JS Dynamic reprogramming of the kinome in response to targeted MEK inhibition in triple-negative breast cancer Cell 2012 149 307 321 22500798 49 Stuhlmiller TJ Miller SM Zawistowski JS Nakamura K Beltran AS Duncan JS Inhibition of Lapatinib-Induced Kinome Reprogramming in ERBB2-Positive Breast Cancer by Targeting BET Family Bromodomains Cell Rep 2015 11 390 404 25865888 50 Kampen KR Ter Elst A Mahmud H Scherpen FJ Diks SH Peppelenbosch MP Insights in dynamic kinome reprogramming as a consequence of MEK inhibition in MLL-rearranged AML Leukemia 2014 28 589 599 24240200 51 Golubovskaya VM Nyberg C Zheng M Kweh F Magis A Ostrov D A Small molecule Inhibitor, 1,2,4,5-benzenetetraamine tetrahydrochloride, targeting the Y397 site of Focal Adhesion Kinase decreases tumor growth J Med Chem 2008 51 7405 7416 18989950

PMC005xxxxxx/PMC5136321.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 101132535 30097 Mol Cancer Ther Mol. Cancer Ther. Molecular cancer therapeutics 1535-7163 1538-8514 27784795 5136321 10.1158/1535-7163.MCT-16-0320 NIHMS827073 Article The potential roles of long non-coding RNAs (lncRNAs) in glioblastoma development Liu Shuang 1 Mitra Ramkrishna *23 Zhao Ming-ming 1 Fan Wenhong 4 Eischen Christine M. 2 Yin Feng 1 Zhao Zhongming 356 1 Department of Neurosurgery, Navy General Hospital, PLA. Beijing, 100048, China. 2 Department of Cancer Biology, Sidney Kimmel Cancer Center, Thomas Jefferson University, Philadelphia, PA 3 Department of Biomedical Informatics, Vanderbilt University Medical Center, Nashville, TN 37203, USA 4 Department of Recombinant Drugs, National Institutes for Food and Drug Control, Beijing, 100050, China. 5 Department of Cancer Biology, Vanderbilt University Medical Center, Nashville, TN 37212, USA 6 Center for Precision Health, School of Biomedical University, The University of Texas Health Science Center at Houston, Houston, TX 77030, USA * Current Address Ramkrishna Mitra, Thomas Jefferson University, Department of Cancer Biology, Sidney Kimmel Cancer Center, Philadelphia, PA Correspondence to: Shuang Liu, Department of Neurosurgery, Navy General Hospital, PLA. Beijing, 100048, China. Phone: 86-010-68780201, Fax: 86-010-68780201, shuangff@sina.com; or Feng Yin, Department of Neurosurgery, Navy General Hospital, PLA. Beijing, 100048, China, yinf897@163.com; or Zhongming Zhao, Department of Biomedical Informatics, Vanderbilt University Medical Center, 2525 West End Avenue, Suite 600, Nashville, TN 37203, Phone: 615-343-9158, Fax: 615-936-8545, zhongming.zhao@vanderbilt.edu. 9 11 2016 26 10 2016 12 2016 01 12 2017 15 12 29772986 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. Long non-coding RNA (lncRNA) may contribute to the initiation and progression of tumor. In this study, we first systematically compared lncRNA and mRNA expression between GBM and paired normal brain tissues using microarray data. We found 27 lncRNA and 82 mRNA significantly up-regulated in GBM, as well as 198 lncRNA and 285 mRNA significantly down-regulated in GBM. We identified 138 co-expressed lncRNA-mRNA pairs from these differentially expressed lncRNA and genes. Subsequent pathway analysis of the lncRNA-paired genes indicated that EphrinB-EPHB, p75-mediated signaling, TNF alpha/NF-κB, and ErbB2/ErbB3 signaling pathways might be altered in GBM. Specifically, lncRNA RAMP2-AS1 had significant decrease of expression in GBM tissues and showed co-expressional relationship with NOTCH3, an important tumor promoter in many neoplastic diseases. Our follow up experiment indicated that (1) an overexpression of RAMP2-AS1 reduced GBM cell proliferation in vitro and also reduced GBM xenograft tumors in vivo; (2) NOTCH3 and RAMP2-AS1 co-expression rescued the inhibitory action of RAMP2-AS1 in glioblastoma cells; and (3) RNA pull-down assay revealed a direct interaction of RAMP2-AS1 with DHC10, which may consequently inhibit, as we hypothesize, the expression of NOTCH3 and its downstream signaling molecule HES1 in GBM. Taken together, our data revealed that lncRNA expression profile in GBM tissue was significantly altered; and RAMP2-AS1 might play a tumor suppressive role in GBM through an indirect inhibition of NOTCH3. Our results provided some insights into understanding the key roles of lncRNA-mRNA co-regulation in human GBM and the mechanisms responsible for GBM progression and pathogenesis. Glioblastoma Glioma stem cells LncRNA-mRNA co-expression network lncRNA-RAMP2-AS1 NOTCH3 Introduction Gliomas account for approximately 30% of primary tumors in the brain (1), and can be categorized into four grades (I-IV) by the histopathological classification according to The World Health Organization (WHO) (2). Among the four grades, glioblastoma (GBM, grade IV) has the most dismal prognosis with an overall survival of less than 14 months (1). Despite multimodal and aggressive treatments that include surgical resection, local radiotherapy and systemic chemotherapy, the outcome of GBM patients remains poor (3, 4). To improve treatment efficacy, a better understanding of glioma pathogenesis at the genetic and molecular levels is urgently needed. Molecular profiling of normal and tumor tissues has revealed that long non-coding RNA (lncRNA) is dysregulated in many human malignancies, including prostate (5), colorectal (6), breast (7), bladder (8), liver (9) and brain cancers (10). Multiple lncRNA genes are postulated to function as oncogenes and tumor suppressors and regulate many hallmarks of cancer. In particular, the regulatory roles of lncRNA in expression, activity and localization of protein-coding genes have attracted much attention (11). Recent evidence has indicated that lncRNA may play important roles in glioma pathogenesis (12, 13). For example, it has been reported that lncRNA may regulate biological processes in glioma, such as cellular proliferation and apoptosis, which contribute to tumorigenesis (14). Aberrant expression of lncRNA has also been implicated for clinical phenotypes (15) and patient prognosis of GBM (16), which could be further exploited as potential diagnostic and therapeutic targets (17). An increasing number of lncRNAs have been characterized in cancer including GBM, lncRNA H19 has been shown to promote glioma angiogenesis and invasion (18, 19). Another study reported that knockdown of lncRNA XIST exerted tumor-suppressive functions in human glioblastoma stem cells by up-regulating the microRNA -152 (20). lncRNA HOTAIR was found to be a cell cycle-regulator and essential for proliferation in human glioblastoma (21, 22). Although these lncRNAs were found their roles in some glioma paradigm, systematic investigation of lncRNA in GBM has not been reported yet (23). In this study, we performed a systematic analysis of lncRNAs with their paired mRNAs in GBM, from which the we pinpointed several lncRNAs whose aberrant expression in GBM specimens potentially altered GBM-specific pathways. Specifically, we studied the function of a lncRNA, RAMP2-AS1, in GBM both in vitro and in vivo. Our data indicated that RAMP2-AS1 might contribute to GBM through targeting the NOTCH3 signaling pathway. Our study provides some important insights of lncRNAs into the mechanisms responsible for GBM pathogenesis. Materials and Methods Acquisition of clinical specimens and ethical standards Glioblastoma specimens were obtained from 20 GBM patients who underwent surgical treatment at Navy General Hospital, China, from January 2011 to December 2013 (Table S1). Glioma was diagnosed according to the 2007 WHO Classification of Tumors of the Central Nervous System. Normal samples were obtained from the normal brain tissues around the tumors (approximately 2-3 cm from the tumor border) of the same GBM patients. Written informed consent of the patients was provided by their legal surrogates to permit surgical procedures and use of resected tissues. This study was approved by the Specialty Committee on Ethics of Biomedicine Research, Navy General Hospital of PLA, China (permission number: 0506-2006). Microarray expression profiles for lncRNA and mRNA The Agilent human lncRNA and mRNA array V4.0 was designed with four identical arrays per slide (4 × 180K format), with each array containing probes interrogating approximately 41,000 and 34,000 human lncRNAs and mRNAs, respectively. These lncRNA and mRNA target sequences were merged from multiple databases, such as 23,898 from GENCODE (V19) (24), 14,353 from Human LincRNA Catalog (25), 7760 from RefSeq, 5627 from UCSC (http://genome.ucsc.edu/), 13,701 from NRED (ncRNA Expression Database, http://nred.matticklab.com/cgi-bin/ncrnadb.pl), 21,488 from LNCipedia (http://www.lncipedia.org/), 1038 from H-InvDB (http://www.h-invitational.jp/hinv/ahg-db/news.jsp), 3019 from lncRNAs-a (Enhancer-like) (25), 1053 from Antisense ncRNA pipeline (26), 407 Hox ncRNAs, 962 UCRs (27), and 848 from Chen Ruisheng lab (Institute of Biophysics, Chinese Academy of Science, http://www.ibp.cas.cn/nRNA). Each RNA sequence was detected by probes and repeated twice. The array also contained 4974 control probes from Agilent. Data normalization of the 2 channel ratios was achieved using an intensity-dependent “Lowess” module implemented in the R programming language. Differential expression (DE) of lncRNA/gene was defined according to the following criteria: > 2.0 fold-change and P <0.05. We have submitted the data of microArray to GEO. The GEO accession number is GSE77452 (Differentially expressed lncRNAs and genes in GBM compared to matched normal brain samples). RNA extraction, labeling and hybridization Total RNA was extracted from 3 pairs of snap-frozen GBM specimens and matched noncancerous tissues using Trizol (Invitrogen, California), and purified with mirVana miRNA Isolation Kit (Ambion, Austin, TX, USA) according to manufacturer's protocol. The amplified cRNA was purified using the RNA Clean-up Kit (MN). The cDNA, labeled with a fluorescent dye (Cy5 and Cy3-dCTP), was produced by Eberwine's linear RNA amplification method and subsequent enzymatic reaction. Labeled controls and test samples, labeled with Cy5-dCTP and Cy3-dCTP, were dissolved in 80 μL hybridization solution containing 3×SSC, 0.2% SDS, 5×Denhardt's solution and 25% formamide. Arrays hybridization was performed in an Agilent Hybridization Oven overnight at a rotation speed of 20 rpm and a temperature of 42°C, then washed with two consecutive solutions (0.2% SDS and 2× SSC) at 42°C and room temperature for 5 minutes. lncRNA-gene co-expression network construction Pre-processed RNA-Seq data of 19 different human normal tissues including brain was downloaded from lncRNA function database (28). The downloaded expression profiling data consisted of 13,249 lncRNA and 20,447 protein-coding genes. We excluded lncRNA or genes that had low variance (belonged to lowest quartile). Furthermore, an lncRNA or mRNA gene was excluded from the analysis if >25% samples had missing values. We computed Pearson correlation coefficient for all possible lncRNA-gene, lncRNA-lncRNA, and gene-gene pairs. A pair was considered as significantly co-expressed if the absolute correlation score was >0.75 and correlation P<0.05 (P values adjusted by Benjamini-Hochberg multiple test correction method) (29). Quantitative real-time polymerase chain reaction (qRT-PCR) validation Total RNA was extracted from 3 GBM and 3 matched normal brain samples using TRIzol reagent (Invitrogen Life Technologies), and then reverse-transcribed using Fermentas RT reagent Kit (Perfect Real Time) according to the manufacturer's instructions. Two microgram of total RNA was converted to cDNA according to the manufacturer's protocol. Expression of lncRNA was measured by real-time RT-PCR using SYBR Premix Ex Taq on MX3000 instrument. The primers used in this study are shown in Table S2. PCR was performed in a total reaction volume of 8 μl, including 5 μl 2×PCR master mix (Superarray), 0.5 μl of PCR Forward Primer (10μM), 2 μl of cDNA, and diluted to 8 μl with double distilled water. The quantitative real-time RT-PCR reaction was set at an initial denaturation step of 10 minutes at 95°C; and 95°C (10 seconds), 60°C (60 seconds), 95°C (10 seconds) in a total 40 cycles, with a final step heating slowly from 60 to 99°C. All samples normalized to GAPDH to calculate relative lncRNA concentrations. Cell lines and cell culture The human glioblastoma cell lines U87 and U251 were obtained from the American Type Culture Collection. These cell lines were purchased between 2006 and 2010 and authenticated by morphologic and growth curve analysis before this study beginning. U87 and U251 cell lines were derived from human glioblastoma specimens and cultured in D/F12 medium supplemented with 10% fetal bovine serum (FBS), (Hyclone, USA). Lentiviral infection and gene transfection Lentivirus (LV) containing RAMP2-AS1 (P34516) and NOTCH3 full-length sequence were obtained from Shanghai GenePharma Co., Ltd. Viral particles were harvested 48 hours after cotransfection of the lentiviral vector (or the control LV vector) and the packaging vectors into HEK293T cells using Lipofectamine 2000 (Life Technologies Corporation, Carlsbad, CA, USA). U87 and U251 cells were then infected with the LV (LV-RAMP2-AS1 or LV-NOTCH3), or the control virus (NC). The infection ratio was determined by fluorescence labeling and real-time RT-PCR. Cell cycle distribution U87 and U251 cells (1×105 cells) were plated in 60-mm culture plates, and the cells were infected with LV-RAMP2-AS1, LV- NOTCH3 and control virus respectively. After 96h, the cells were trpsinized, fixed in 70% ethanol, washed once with PBS, and then labeled with propidium iodide (Sigma-Aldrich, USA) in the presence of RNase A (Sigma-Aldrich) for 30 min in the dark (50g/mL). Samples were run on a FACScan flow cytometer (Becton-Dickinson, USA), and the percentages of cells within each phase of the cell cycle were analyzed using Cell Quest software. Cell proliferation analysis Glioblastoma cells (2×103 cells per well) were placed into 96-well plates. Cells were infected with LV-RAMP2-AS1 and LV controls for 24, 48 and 72 hours, respectively. Thereafter, CCK-8 reagent was added to the cells. The cells were further cultured for two hours, and the optical density (OD) at 450 nm was measured by a microplate reader according to the manufacturer's instructions (Thermo Fisher Scientific, USA). Immunofluorescent staining Cells were washed three times with ice-cold PBS and fixed with 4% paraformaldehyde-PBS. Cell were then incubated with 0.1% Triton-PBS for 15 minutes, 1% bovine serum albumin-PBS for 10 minutes, and then treated with goat anti-human Ki67 monoclonal antibody (1:1000, Santa Cruze, Texas, USA at 4°C overnight. FITC labeled mouse anti-goat IgG (1:1000, Santa Cruz, Texas, USA) was used for visualization. RNA pull-down assay Biotin-labeled RNAs were in vitro transcribed using the Biotin RNA Labeling Mix and T7 RNA polymerase (Ambion, USA), and purified with the RNeasy Mini Kit (QIAGEN, Germany) on-column digestion of DNA. To prepare the glioblastoma cell-U87 nuclear extract, frozen U87 cells were homogenized using a dounce homogenizer with 15-20 strokes in nuclear isolation buffer (250 mM sucrose, 10mM Tris-HCL [PH 7.5], 1mM EDTA with protease inhibitors). Nuclear pellets were collected by centrifugation at 1,000 × g for 10 min, resuspended in 1 ml RNA immunoprecipitation buffer (150mM Nacl, 20 mM Tris[PH7.4], 1mM EDTA, 0.5% Triton X-100 with protease inhibitors and RNaseOUT). The lysates were mechanically sheared again using a dounce homogenizer with 15–20 strokes. Nuclear membrane and other debris were pelleted by centrifugation at 12,000 rpm for 10 min. The folded sense or anti-sense RNAs (1 ug) were added into 2 mg pre-cleared nuclear lysates (supplemented with 0.2 mg/ml heparin, 0.2 mg/ml yeast tRNA and 1 mM DTT) and incubated at 4_C for 1 hr. Sixty microliters of washed Streptavidin-coupled Dynabeads (Invitrogen, USA) were added to each binding reaction and further incubated at 4°C for 1 hr. Beads were washed briefly five times with RIP buffer and heated at 70°C for 10 min in 13 LDS loading buffer, and the retrieved proteins were visualized by SDS-PAGE and silver staining. The unique protein bands shown in the sense RNA pull-down were identified by Mass Spectrometry. Western blot Protein lysates were prepared as previously described (30). The protein samples were resolved by SDS-PAGE and transferred onto PVDF membranes (Roche, Basel, Switzerland). The membranes were then incubated with the following antibodies: Rabbit anti-NOTCH3 (Cell Signaling Technology, USA), mouse anti-HES1 (R&D, USA), mouse anti-DHC10 (R&D, USA) and mouse anti-P16/P21/P27 (Santa Cruz, USA). Chemiluminescence antibody-labeled protein bands were detected using a G:BOX F3 (Syngene, Cambridge, UK). Xenografts U87 and U251 (3×108 cells) infected with LV-RAMP2-AS1 or LV control were injected (subcutaneous) into the upper right flanks of nude mice. Thirty days after injection, the xenograft were harvested and volumes measured according to the formula: V (mm3) = (a)×(b2/2), where a is the largest diameter of tumor and b is the smallest diameter of tumor. For conducting intracerebral xenografted tumors, nude mice were anesthetized and stereotaxically inoculated in the right striatum (Bregmaanteroposterior: −0.5mm, mediolateral: +2mm, dorsoventral: −3mm) with 1×108 tumor cells derived from the subcutaneous tumors. After 3 weeks of intracerebral implantation, the mice were anesthetized deeply and perfused with saline that contained 100 U/ml of heparin (Sigma-Aldrich, USA). They were then fixed with 4% of poly-formaldehyde that was prepared in PBS. Next, the paraffin sections (4μm) of the xenografted tumors were analyzed by H&E staining. Experiments involving mice were performed in accordance with an approved Institutional Animal Care and Use Committee protocol (permission number: 0308-2013). Statistical analysis The difference in results between two groups was examined by two-tailed Student t-test, a P value <0.05 was considered statistically significant. The mean ± S.D. values were displayed in the figures. All statistical analyses were performed and graphs were generated using the SPSS software (version 12.0; Windows platform), or R software package (version 3.2.2). Detail survival analysis using a cohort of 20 GBM patient samples (in-house data) or 150 GBM patient samples (extracted from The Cancer Genome Atlas [TCGA] project) was provided in Supplemental data 1. Results Systematic identification of lncRNA in GBM We first performed microarray expression profiling of lncRNA and mRNA in 3 human GBM and matched normal brain tissue samples. These analyses yielded 27 up- and 198 down-regulated lncRNA, as well as 82 up- and 285 down-regulated protein-coding genes in GBM compared to normal brain tissues of the same patients that passed the criteria of >2 fold-change and P <0.05 (Fig. 1A). Next, we mapped the aberrantly expressed lncRNA and mRNA onto an lncRNA-mRNA co-expression network, which was constructed based on significant expression correlation (absolute Pearson correlation coefficient >0.75, P <0.05 after adjusted by Benjamini-Hochberg [BH] multiple testing correction method) (31) among all the lncRNA and protein-coding genes across 19 human normal tissues, including brain (32). In the network, there were three types of co-expression links: between a coding RNA and a coding RNA, between a non-coding RNA and a coding RNA, and between a non-coding RNA and a non-coding RNA (Fig. 1B). For each type of co-expression link, we summarized dysregulation patterns of associated molecules, number of dysregulated links, lncRNA, and genes in Supplemental Table S3 (33). To illustrate functional importance of the dysregulated lncRNA, we performed pathway and Gene Ontology (GO) term enrichment analyses of their co-expressed genes that were also differentially expressed in GBM. This approach facilitated prioritizing lncRNA whose aberrant expression potentially alters the expression of genes that belong to GBM specific pathways and biological processes (BPs). We identified 10 significantly enriched pathways (adjusted P <0.05, hypergeometric test followed by BH multiple test correction) using the Pathway Commons database as the pathway annotation dataset (34), which is embedded into the software WebGestalt (28, 35) (Supplementary Table S4). Intriguingly, most of the inferred pathways are linked to functions related to GBM pathogenesis, such as EphrinB-EPHB pathway, p75(NTR)-mediated signaling, TNF alpha/NF-kB, Neurotrophic factor-mediated Trk receptor signaling, and ErbB2/ErbB3 signaling events (Fig. 1C). The GO functional annotations were also informative in the context of GBM development. Specifically, we obtained 11 BP terms (non-redundant terms that belong to level 5 or more), among which most are closely linked to the biology of neurogenesis, neuron regeneration and differentiation. The result suggested that these relevant biological functions might be altered in the origin and development of GBM (Supplementary Table S5). From the GBM relevant pathways and GO BP terms, we pinpointed 4 protein-coding genes and their co-expressed lncRNAs (ENSG00000197291-NOTCH3, ENSG00000226645-ARHGAP32, ENSG00000261684-NKIRAS1, and ENSG00000234741-SRC) (Supplemental Table S6) and performed real-time RT-PCR using 3 pairs of snap-frozen GBM specimens and matched noncancerous tissues to verify whether these important pathway-linked molecules were indeed dysregulated in GBM. Our real-time RT-PCR confirmed aberrant expression patterns of all the lncRNA (except ENSG00000226645) and genes that were observed in the microarray data (Fig. 2), suggesting that these lncRNAs and genes were dysregulated in GBM patient samples and that they might have important roles in GBM pathology. For example, in the co-expression network, lncRNA ENSG00000197291 (RAMP2-AS1) was linked to NOTCH3, a key regulator of many signaling pathways involved in cancer stem cells (CSCs) differentiation and development of glioma (36-39). It has been shown that activation of NOTCH3 promotes invasive glioma in a tissue-specific manner (40). The results from both the microarray and real-time RT-PCR showed that, when compared with normal brain tissue, the expression level of RAMP2-AS1 in GBM significantly decreased (P<0.05), whereas the expression of NOTCH3 significantly increased (P<0.05, Fig. 2). Low RAMP2-AS1 expression correlates with poor prognosis in GBM Considering the lower expression of RAMP2-AS1 in GBM than normal brain tissue, we further examined whether RAMP2-AS1 expression correlated with patient survival. We separated 20 GBM patient samples into two groups by using the median expression of RAMP2-AS1. Our analysis revealed that the patients with low expression of RAMP2-AS1 had significantly shorter survival times (P <0.001, log-rank test; Fig. 3A) than those with a high expression level of RAMP2-AS1 (detail procedure in Supplementary data1). We also assessed 5-year survival by dividing a larger cohort of 150 GBM patients extracted from TCGA project into two groups based on the median RAMP2-AS1 expression (Supplementary data1). The result showed a similar trend to what we observed in our patient samples, though statistical significance is moderate (P =0.094, log-rank test; Fig. 3B). RAMP2-AS1 regulates the cell cycle progression mediated by NOTCH3 To determine the exact effects of RAMP2-AS1 on the growth of GBM cells, we generated GBM cell lines with over-expression of RAMP2-AS1 by infection with the LV-RAMP2-AS1. The immunofluorescence and qRT-PCR analysis showed that the infection efficiency of LV-RAMP2-AS1 in U251 and U87 were greater than 90% (Figs. 4A and 4B). U87 and U251 cells were infected with LV-RAMP2-AS1 only, or co-infected with LV-NOTCH3 at the same time. At 96h after treatment, flow cytometry was performed to examine the cell cycle. We found that RAMP2-AS1 overexpression could block the GBM cell cycle progress and NOTCH3 co-expression rescued the cell cycle arrest (Fig.5A). The proteins which are relative to cell cycle, such as P21, P27 and P16 were prepared for Western blot analysis. The result showed that RAMP2-AS1 overexpression decreased the expression of P21 significantly and NOTCH3 could restore P21 to basal levels (Fig.5D). RAMP2-AS1 inhibits GBM cell growth Cell-counting kit-8 assays indicated that cell proliferation was reduced in both U87 and U251 cells when RAMP2-AS1 was over-expressed (P<0.05, Fig. 5B). Consistent with the decrease in cell proliferation, we observed a significant lower level of Ki67 expression (0.2808±0.091 and 0.3911±0.185236) in U87 and U251 cells, which had increased RAMP2-AS1 expression, than that of the control cells (0.59189±0.1734 and 0.5282±0.2050) (P<0.05, Fig. 5C). To further evaluate the effects of RAMP2-AS1 on GBM growth in vivo, RAMP2-AS1-overexpressing U87 and U251 cells and the negative control cells (those were infected with control virus) were injected into the flanks of nude mice. Tumors were allowed to grow for 30 days and then they were extracted and measured for mean volumes. The xenografts of U87 and U251 cells with overexpressed RAMP2-AS1(269±80.56mm3 and 403±62.17mm3) were significantly smaller than those GBM tumors formed from control cells (1355±76.32mm3 and 1447±85.95mm3) (P <0.05, Figure 6A, 6B). Furthermore, the H&E staining of intracranial implanted tumors showed that the subcutaneous tumors still remained the pathologic characters of GBM, such as cellular heterogeneity, rapid proliferation, angiogenesis and extensive invasion (Figure 6 A). The results above indicated that RAMP2-AS1 inhibited the growth of GBM cells both in vitro and in vivo, suggesting its tumor suppressive role in GBM. RAMP2-AS1 may regulate the proliferation of GBM through NOTCH3 signaling Our co-expression network analysis suggested that RAMP2-AS1 expression correlated with NOTCH3 expression (Fig. 1). To determine whether NOTCH3 is regulated by RAMP2-AS1, we measured the expression of NOTCH3 after over-expression of RAMP2-AS1. Expression of NOTCH3 protein and its downstream signaling molecule, HES1, decreased significantly in both U87 and U251 cells which were infected by LV-RAMP2-AS1 when we compared with those cells expressing control vector (P= 0.039738, Fig. 5D). To determine the specificity of NOTCH3 regulation, we overexpressed NOTCH3 together with RAMP2-AS1 in U87 and U251 cells, and did not observe the reduction of HES1 expression by Western blot analysis (Fig. 5D). Furthermore, we observed that the growth inhibition of U87 and U251, both in vitro and in vivo induced by RAMP2-AS1 over-expression, was rescued by NOTCH3 over-expression (Fig. 5B, Fig.5C and Fig. 6). These results suggested that the RAMP2-AS1-mediated down-regulation of NOTCH3 likely contributed to the reduction of GBM proliferation. RAMP2-AS1 interacts with DHC10 to regulate NOTCH3 expression We next investigated how RAMP2-AS1 regulates NOTCH3 expression. First, we performed an RNA pull-down experiment using nuclear extracts of U87 cells to identify the proteins interacting with RAMP2-AS1. Several additional bands were present in the SDS-PAGE silver staining analysis of the fraction precipitated with biotin-labeled RAMP2-AS1 compared to an antisense control (Fig. 7A). Proteins in these specific bands were identified by mass spectrometry, one of which was human dynein heavy chain 10 (DHC10), a known microtubule-associated protein (41). We confirmed the specific interaction between DHC10 and RAMP2-AS1 by immunoblotting (Fig. 7B). Next, qRT-PCR confirmed that DNAH10siRNA in combination with RAMP2-AS1 could restore NOTCH3 expression associated with RAMP2-AS1 depletion (Fig. 7C). Put together, our data suggested that RAMP2-AS1 might inhibit the expression of NOTCH3 mediated by DNAH10 in GBM pathology. Discussion GBM is the most common and aggressive type of primary brain tumor in humans. It accounts for 52% of all parenchymal brain tumor cases and 20% of all intracranial tumors (42). Although the mechanisms of GBM occurrence and development have been extensively investigated during the past two decades, the pathogenesis of this disease is still ill defined, and the gene regulation involved in this disease remains largely unclear. Increasing lines of evidence suggested that lncRNA may be important factors in controlling gene expression (43). Hence, we examined the expression profiles of both lncRNAs and mRNAs in GBM tissue and matched control samples (Fig.1A). Subsequently, we constructed lncRNA-mRNA co-expression network based on publicly available, matched lncRNA and mRNA expression profiling data and mapped the differently expressed lncRNAs and mRNAs in GBM into this unique reference network. This approach was taken to examine which lncRNA-mRNA pairs have been aberrantly expressed in GBM and have potential roles in the origination and development of GBM. Our results may better reflect the genetic variations of GBM due to the ruling out the difference at the individual level. Thus, this co-expression network analysis could generate useful hypotheses for further study of the functional roles of the differentially expressed lncRNAs in the development of GBM. Importantly, our systems biology approach found that the abnormally expressed lncRNAs were mostly involved in the pathways and biological functions that are related to GBM pathogenesis, indicating the effectiveness of our approach. While the functional roles of most of the lncRNAs in the co-expression network have not been previously studied, the mRNAs that have co-expression relationship with those lncRNAs have been previously reported to have important roles in various tumors including prostate cancer, glioma, breast cancer and squamous cell carcinoma of the lung (39, 44-45). Therefore, according to the functional roles of these mRNAs, we could prioritize the lncRNAs that might play key roles in GBM. Our investigation of lncRNA-mRNA network revealed that NOTCH3, a key gene in the progression of GBM, had co-expression relationship with the lncRNA RAMP2-AS1. The expression of RAMP2-AS1 was significantly reduced in GBM. Notch signaling has a critical function in the specification, proliferation, and survival of stem/progenitor cells in a number of tissues, including the central and peripheral nervous systems (46). The pathway is widely implicated in neoplasia, and in most contexts it promotes neoplastic growth (47-52). It has been shown that activation of NOTCH3 pathway promotes murine T-acute lymphoblastic leukemia (T-All), similar to that has been seen in humans (50, 53, 54). Notch is also thought to play important role in poorly differentiated tumor cells. Inhibition of this pathway may deplete “cancer stem cells”, which are resistant to radiation and standard chemotherapies (55-58). Small molecules targeting Notch have shown great promise in preclinical testing of several tumor models. Based on such studies, phase I clinical trials for leukemia and breast cancer have been initiated using gamma-secretase inhibitors that block the activation of Notch receptors (59, 60). Similar to leukemia and breast, Notch signaling pathway has also been shown to promote proliferation of glioma cells and play different roles in primary and secondary GBM (61, 62); however, the mechanisms of Notch regulation in GBM has not yet clear. In our study, we found that RAMP2-AS1 expression was significantly reduced in primary GBM tissues compared with normal brain tissues and lower RAMP2-AS1 expression was correlated with poor overall survival of GBM patients (Fig. 3). We utilized a cohort of 150 GBM patient samples from TCGA and a cohort of 20 patient samples from our in house data. Although the survival analysis of a large cohort of 150 GBM patients extracted from TCGA showed that the patients with low expression of RAMP2-AS1 had only moderately shorter survival times than those with a high expression level of RAMP2-AS1 (P =0.094, log-rank test; Fig. 3B, Supplementary data 1), the survival analysis of our in-house 20 primary GBM revealed significant survival difference between the two groups. Several factors might contribute to the difference between the two datasets: specimens background (ethnic background, gender, age, etc.), sample heterogeneity and treatment (e.g., all the 20 in-house GBM samples were primary GBM and the patients had experienced the same treatment), and statistical power. Recently, Pastori et al. applied single molecule sequencing to quantify the expression of lncRNAs in GBM. They reported the top 100 up- and down- regulated lncRNAs in GBM versus control samples, and RAMP2-AS1 was not found in that list (63). Pastori et al. used epilepsy samples as control, while we used matched normal brain tissue samples. This different selection of control tissues may partially explain why RAMP2-AS1 showed different result in the two studies, though further investigation is warranted. Our co-expression network analysis of lncRNA-mRNA genes revealed that RAMP2-AS1 was significantly correlated with NOTCH3, which was significantly up-regulated in GBM (Fig.2). This is consistent with many previous studies in which Notch signaling acts as a tumor promotor (45, 46). Our study also showed that over-expression of RAMP2-AS1 inhibited the proliferation of GBM cells in vitro and in vivo in part by down-regulating NOTCH3 and its downstream molecule, HES1(Fig.5, Fig.6). These results indicated that RAMP2-AS1 might contribute to GBM through its regulation of NOTCH3. Furthermore, although we did not detect the direct interaction between RAMP2-AS1 and NOTCH3, our RNA pull-down experiment indicated another protein, DHC10 had direct interaction with RAMP2-AS1. For all the results we had, we speculated that RAMP2-AS1 might interact with NOTCH3 mediated by DHC10 (Fig.7). While our results are still preliminary, the findings are promising toward setting us future research direction focusing on lncRNA-mRNA synergistic regulation in the pathogenesis of GBM. It is our belief that the lncRNA-mRNA pairs highlighted in EphrinB-EPHB, p75-mediated signaling, TNF alpha/NF-κB and ErbB2/ErbB3 signaling pathways warrant future investigation in GBM or its related phenotype. Supplementary Material 1 Acknowledgments This work was supported by the National Nature Science Foundation of China (Grants No. 81172384 and No.30873029). Z.M. Zhao was partially supported by National Institutes of Health (NIH) grants (R01LM011177, R21CA196508 and R03CA167695). The funders had no role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript. We thank the CapitalBio Corporation for performing the microarray detection of lncRNA & mRNA and Dr. Zheng Wang (Chinese Academy of Medical Sciences) for his assistance in establishing animal models. We thank the three reviewers' constructive comments which helped us substantially improve the quality of this work in revision. Figure 1 Flowchart for deciphering functional lncRNA(s) in GBM. The framework has the following major steps. A, Identification of differentially expressed (DE) lncRNAs and genes in GBM compared to matched normal samples. Volcano plots show DE probes for lncRNAs and mRNA genes. X-axis and Y-axis show fold-change and −log(P values), respectively. Red, green, and grey dots represent up-, down-, and no significant expression change, respectively. B, Construction of lncRNA-gene co-expression network. Red and green circle nodes denote up- and down-regulated genes, respectively. Red and green v-shaped nodes denote up- and down-regulated lncRNAs, respectively. Up- or down-regulation was measured in GBM patient samples compared to the matched normal brain tissues. Edges denote significant expression correlation between nodes. C, Detection of GBM-specific pathways and Gene Ontology biological process terms enriched in lncRNA-gene co-expression network. D, Selection of lncRNAs and genes of interest to perform follow-up experiments in-vitro and in-vivo to determine lncRNA's pathogenic potential in GBM. Additionally, Kaplan-Meier analysis assessed the ability of the selected lncRNA to predict survival of GBM patients using an in-house as well as a large cohort of GBM patient samples available in The Cancer Genome Atlas (TCGA) project. Figure 2 Real-time RT-PCR verification of lncRNA and mRNA expression. Four lncRNAs and four mRNAs were selected to examine their expression between GBM and normal control by real-time RT-PCR. Figure 3 Decreased expression of RAMP2-AS1 in GBM correlates with poor clinical outcome of GBM patients. A, Kaplan-Meier survival curves using the data from 20 GBM patients with respect to RAMP2-AS1 expression. B, A large cohort of 150 GBM patient samples from TCGA was evaluated by Kaplan-Meier analysis with respect to RAMP2-AS1 expression. Figure4 The infection effect of LV-RAMP2-AS1 on GBM cells. A, After 96h of LV-RAMP2-AS1(5×108TU/ml) infection, U87 and U251 cells were observed under fluorescence microscope. B, After 96h of LV-RAMP2-AS1 infection, qRT-PCR were performed to analyze the expression of RAMP2-AS1in U87 and U251 cells. Experiments were performed in triplicate in three times. The values shown are mean ± SD. *P <0.05 vs. control group. Figure 5 RAMP2-AS1 inhibits the growth of GBM cells. A, Flow cytometry was performed to examine the S arrest in U87 and U251 cells after treatment with control LV vector LV-NC, LV-RAMP2-AS1(RAMP2-AS1) and LV- NOTCH3 combined with LV-RAMP2-AS1 (Notch3+RAMP2-AS1) for 96h. B, U87 and U251 GBM cells which were infected with above vectors respectively and subjected to CCK-8 cell proliferation assays. C, Ki67 expression of U87 and U251 GBM cells which were infected with above vectors respectively were detected by immunofluorescence (left). The Ki67 data is presented as the ratio of Ki67 positive cells (light blue) to total DAPI positive cells (light blue U dark blue) (right). D, The expression of protein NOTCH3, HES-1 and P21 in U87 and U251 cells was determined by Western blot after 72h of above vectors infection. All above experiments were performed in triplicate in three times. The values shown are mean ± SD (n=5 in each group). *P <0.05 vs. control group. Figure 6 RAMP2-AS1 inhibits the growth of GBM in vivo. A, U87 and U251 GBM cells which were infected with LV-NC, LV-RAMP2-AS1 and LV- NOTCH3 combined with LV-RAMP2-AS1 were injected subcutaneously into nude mice. After 30 days tumors were removed and photographed and their volumes were measured. There were five mice in each group. *: P <0.05 each group vs. the NC group (right). The tumor cells derived from above subcutaneous xenografts were implanted into right striatum of nudes and analyzed by H&E staining (left). B, The volume of subcutaneous xenografts of the groups mentioned above. Bars are mean±SD. *p<0.05, n=5 per group. Figure7 RAMP2-AS1 interacts with DHC10. A, Silver-stained SDS-PAGE gel analysis of proteins in nuclear extract of U87 cells that are bound to biotinylated RAMP2-AS1 or its antisense. The highlighted regions were analyzed by mass spectrometry, identifying DHC10 as a protein unique to RAMP2-AS1 (left). Immunoblotting analysis of proteins in nuclear extract of U87 cells that are bound to biotinylated RAMP2-AS1 or its antisense using an anti-DHC10 antibody (right). B, NOTCH3 expression in U87 cells receiving LV, LV-RAMP2-AS1, scramble siRNA(scrb si), or DHC10siRNA in combination as indicated (Error bars represent SD, *p<0.05). Disclosure of Potential Conflicts of Interest The authors declare no conflict of interest. References 1 Wen PY Kesari S Malignant gliomas in adults. N Engl J Med 2008 359 492 507 18669428 2 Louis DN Ohgaki H Wiestler OD Cavenee WK Burger PC Jouvet A The 2007 WHO classification of tumours of the central nervous system. Acta Neuropathol 2007 114 97 109 17618441 3 Omuro A DeAngelis LM Glioblastoma and other malignant gliomas: a clinical review. JAMA 2013 310 1842 1850 24193082 4 Taylor LP Diagnosis, treatment, and prognosis of glioma: five new things. Neurology 2010 75 S28 S32 21041768 5 Prensner JR Iyer MK Sahu A Asangani IA Cao Q Patel L The long noncoding RNA SChLAP1 promotesaggressive prostate cancer and antagonizes the SWI/SNF complex. Nat Genet 2013 45 1392 1398 24076601 6 Marín-Béjar O Marchese FP Athie A Sánchez Y González J Segura V Pint lincRNA connects the p53 pathway with epigenetic silencing by the Polycomb repressive complex 2. Genome Biol 2013 14 R104 24070194 7 Gupta RA Shah N Wang KC Kim J Horlings HM Wong DJ Long non-coding RNA HOTAIR reprograms chromatin state to promote cancer metastasis. Nature 2010 464 1071 1076 20393566 8 Zhang Q Su M Lu G Wang J The complexity of bladder cancer: long noncoding RNAs are on the stage. Mol Cancer 2013 12 101 24006935 9 Braconi C Valeri N Kogure T Gasparini P Huang N Nuovo GJ Expression and functional role of a transcribed noncoding RNA with an ultraconserved element in hepatocellular carcinoma. Proc Natl Acad Sci USA 2011 108 786 791 21187392 10 Zhou Y Zhong Y Wang Y Zhang X Batista DL Gejman R Activation of p53 by MEG3 non-coding RNA. J Biol Chem 2007 282 24731 24742 17569660 11 Zong X Tripathi V Prasanth KV RNA splicing control: Yet another gene regulatory role for long nuclear noncoding RNAs. RNA Biol 2011 6 968 977 12 Bian EB Li J Xie YS Zong G Zhao B LncRNAs: new players in gliomas, with special emphasis on the interaction of lncRNAs With EZH2. J Cell Physiol 2014 230 496 503 13 Sun YZ Wang Z Zhou D Long non-coding RNAs as potential biomarkers and therapeutic targets for gliomas. Med Hypotheses 2013 81 319 321 23688743 14 Wang P Ren Z Sun P Overexpression of the long non-coding RNA MEG3 impairs in vitro glioma cell proliferation. J Cell Biochem 2012 113 1868 1874 22234798 15 Zhang X Sun S Pu JK Tsang AC Lee D Man VO Long non-coding RNA expression profiles predict clinical phenotypes in glioma. Neurobiol Dis 2012 48 1 8 22709987 16 Zhang XQ Sun S Lam KF Kiang KM Pu JK Ho AS A long non-coding RNA signature in glioblastoma multiforme predicts survival. Neurobiol. Dis 2013 58 123 131 23726844 17 Amit D Matouk IJ Lavon I Birman T Galula J Abu-Lail R Transcriptional targeting of glioblastoma by diphtheria toxin-A driven by both H19 and IGF2-P4 promoters. Int. J. Clin. Exp. Med 2012 5 124 135 22567173 18 Jiang X Yan Y Hu M Chen X Wang Y Dai Y Increased level of H19 long noncoding RNA promotes invasion, angiogenesis and stemness of glioblastoma cells. J Neurosurg 2016 124 129 36 26274999 19 Jia P Cai H Liu X Chen J Ma J Wang P Long non-coding RNA H19 regulates glioma angiogenesis and the biological behavior of glioma-associated endothelial cells by inhibiting microRNA-29a. Cancer Lett 2016 38 359 69 20 Yilong Y Jun M Yixue X Ping W Zhen L Jing L Knockdown of long non-coding RNA XIST exerts tumor-suppressive functions in human glioblastoma stem cells by up-regulating miR-152. Cancer letters 2015 359 75 85 25578780 21 Kailiang Z Xiaotian S Xuan Z Lei H Luyue C Zhendong S Long non-coding RNA HOTAIR promotes glioblastoma cell cycle progression in an EZH2 dependent manner. Oncotarget 2014 6 537 46 22 Chiara P Philipp K Clara P Veronica P Claude-Henry V Jann NS The bromodomain protein BRD4 controls HOTAIR, a long noncoding RNA essential for glioblastoma proliferation. PNAS 2015 112 8326 31 26111795 23 Wang KC Chang HY Molecular mechanisms of long non-coding RNAs. Mol Cell 2011 6 904 914 24 Benjamini Y Hochberg Y Controlling the False Discovery Rate - a Practical and Powerful Approach to Multiple Testing. J Roy Stat Soc B Met 1995 57 289 300 25 Jiang Q Ma R Wang J Wu X Jin S Peng J LncRNA2Function: a comprehensive resource for functional investigation of human lncRNAs based on RNA-seq data. BMC genomics 2015 16 Suppl 3 S2 26 Liao Q Liu C Yuan X Kang S Miao R Xiao H Large-scale prediction of long non-coding RNA functions in a coding-non-coding gene co-expression network. Nucleic acids research 2011 39 3864 3878 21247874 27 Cerami EG Gross BE Demir E Rodchenkov I Babur O Anwar N Pathway Commons, a web resource for biological pathway data. Nucleic acids research 2011 39 D685 690 21071392 28 Zhang B Kirov S Snoddy J WebGestalt: an integrated system for exploring gene sets in various biological contexts. Nucleic acids research 2005 33 W741 748 15980575 29 Yen WC Fischer MM Axelrod F Bond C Cain J Cancilla B Targeting Notch signaling with a Notch2/Notch3 antagonist (tarextumab) inhibits tumor growth and decreases tumor-initiating cell frequency. Clin Cancer Res 2015 21 2084 2095 25934888 30 Alqudah MA Agarwal S Al-Keilani MS Sibenaller ZA Ryken TC Assem M NOTCH3 is a prognostic factor that promotes glioma cell proliferation, migration and invasion via activation of CCND1 and EGFR. PloS One 2013 8 e77299 24143218 31 Benjamini Y Hochberg Y Controlling the False Discovery Rate - a Practical and Powerful Approach to Multiple Testing. J Roy Stat Soc B Met 1995 57 289 300 32 Jiang Q Ma R Wang J Wu X Jin S Peng J LncRNA2Function: a comprehensive resource for functional investigation of human lncRNAs based on RNA-seq data. BMC genomics 2015 16 Suppl 3 S2 33 Liao Q Liu C Yuan X Kang S Miao R Xiao H Large-scale prediction of long non-coding RNA functions in a coding-non-coding gene co-expression network. Nucleic acids research 2011 39 3864 3878 21247874 34 Cerami EG Gross BE Demir E Rodchenkov I Babur O Anwar N Pathway Commons, a web resource for biological pathway data. Nucleic acids research 2011 39 D685 690 21071392 35 Mitra R Edmonds MD Sun J Zhao M Yu H Eischen CM Reproducible combinatorial regulatory networks elucidate novel oncogenic microRNAs in non-small cell lung cancer. RNA 2014 20 1356 68 25024357 36 Yen WC Fischer MM Axelrod F Bond C Cain J Cancilla B Targeting Notch signaling with a Notch2/Notch3 antagonist (tarextumab) inhibits tumor growth and decreases tumor-initiating cell frequency. Clin Cancer Res 2015 21 2084 2095 25934888 37 Alqudah MA Agarwal S Al-Keilani MS Sibenaller ZA Ryken TC Assem M NOTCH3 is a prognostic factor that promotes glioma cell proliferation, migration and invasion via activation of CCND1 and EGFR. PloS One 2013 8 e77299 24143218 38 Saito N Fu J Zheng S Yao J Wang S Liu DD A high Notch pathway activation predicts response to γ secretase inhibitors in proneural subtype of glioma tumor-initiating cells. Stem Cells 2014 32 301 312 24038660 39 Xu P Yu S Jiang R Kang C Wang G Jiang H Differential expression of Notch family members in astrocytomas and medulloblastomas. Pathol Oncol Res 2009 15 703 710 19424825 40 Pierfelice TJ Schreck KC Dang L Asnaghi L Gaiano N Eberhart CG Notch3 activation promotes invasive glioma formation in a tissue site-specific manner. Cancer research 2011 71 1115 1125 21245095 41 Arai E Gotoh M Tian Y Sakamoto H Ono M Matsuda A Alterations of the spindle checkpoint pathway in clinicopathologically aggressive CpG island methylator phenotype clear cell renal cell carcinomas. Int J Cancer 2015 137 2589 606 26061684 42 Wick W Weller M Weiler M Batchelor T Yung AW Platten M Pathway inhibition: emerging molecular targets for treating glioblastoma. Neuro Oncol 2011 6 566 579 43 Khachane AN Harrison PM Mining mammalian transcript data for functional long non-coding RNAs. PLoS One 2010 4 E10316 44 Sarma P Bag I Ramaiah MJ Kamal A Bhadra U Pal Bhadra M Bisindole-PBD regulates breast cancer cell proliferation via SIRT-p53 axis. Cancer Biol Ther 2015 16 1486 1501 26192233 45 Deng G Ju X Meng Q Yu ZJ Ma LB Emodin inhibits the proliferation of PC3 prostate cancer cells in vitro via the Notch signaling pathway. Mol Med Rep 2015 12 4427 4433 26081222 46 Kim JS Kim ES Liu D Lee JJ Behrens C Lippman SM Activation of insulin-like growth factor 1 receptor in patients with non-small cell lung cancer. Oncotarget 2015 6 16746 16756 25944691 47 Pierfelice TJ Schreck KC Eberhart CG Gaiano N Notch, neural stem cells, and brain tumors. Cold Spring Harb Symp Quant Biol 2008 73 367 75 19022772 48 Weng AP Ferrando AA Lee W Morris JP Silverman LB Sanchez-Irizarry C Activating mutations of NOTCH1 in human T cell acute lymphoblastic leukemia. Science 2004 306 269 71 15472075 49 Sun J Gong X Purow B Zhao Z Uncovering miroRNA and transcription factor mediated regulatory networks in glioblastoma. PLoS Comput Biol 2012 8 7 e1002488 22829753 50 Fan X Matsui W Khaki L Stearns D Chun J Li YM Notch pathway inhibition depletes stem-like cells and blocks engraftment in embryonal brain tumors. Cancer Res 2006 66 7445 7452 16885340 51 O'Neil J Calvo J McKenna K Krishnamoorthy V Aster JC Bassing CH Activating Notch1 mutations in mouse models of T-ALL. Blood 2006 107 781 785 16166587 52 Wang Z Li Y Banerjee S Sarkar FH Exploitation of the Notch signaling pathway as a novel target for cancer therapy. Anticancer Res 2008 28 3621 3630 19189643 53 Rizzo P Osipo C Foreman K Golde T Osborne B Miele L Rational targeting of Notch signaling in cancer. Oncogene 2008 27 5124 5131 18758481 54 Lee SY Kumano K Masuda S Hangaishi A Takita J Nakazaki K Mutations of the Notch1 gene in T-cell acute lymphoblastic leukemia: analysis in adults and children. Leukemia 2005 19 1841 1843 16079893 55 Ellisen LW Bird J West DC Soreng AL Reynolds TC Smith SD TAN-1, the human homolog of the Drosophila notch gene, is broken by chromosomal translocations in T lymphoblastic neoplasms. Cell 1991 66 649 661 1831692 56 Wicha MS Targeting breast cancer stem cells. Breast 2009 18 S56 58 19914544 57 Talora C Campese AF Bellavia D Felli MP Vacca A Gulino A Notch signaling and diseases: an evolutionary journey from a simple beginning to complex outcomes. Biochim Biophys Acta 2008 1782 489 497 18625307 58 Wang J Wakeman TP Lathia JD Hjelmeland AB Wang XF White RR Notch Promotes Radioresistance of Glioma Stem Cells. Stem Cells 2009 28 17 28 59 Kvinlaug BT Huntly BJ Targeting cancer stem cells. Expert Opin Ther Targets 2007 11 915 927 17614760 60 Papayannidis C DeAngelo DJ Stock W Huang B Shaik MN Cesari R A phase 1 study of the novel gamma-secretase inhibitor PF-03084014 in patients with T-cell acute lymphoblastic leukemia and T-cell lymphoblastic lymphoma. Blood Cancer J 2015 5 e350 26407235 61 Wei P Walls M Qiu M Ding R Denlinger RH Wong A Evaluation of selective gamma-secretase inhibitor PF-03084014 for its antitumor efficacy and gastrointestinal safety to guide optimal clinical trial design. Mol Cancer Ther 2010 9 1618 1628 20530712 62 Marie-Therese S Karina K Hans SP The functional role of Notch signaling in human gliomas. Neuro-Oncology 2010 12 199 211 20150387 63 Alqudah MA Agarwai S AI-keilani MS Sibenaller ZA Ryken TC Assem M Notch3 is a prognostic factor that promotes glioma cell proliferation, migration and invasion via activation of CCND1 and EGFR. PLoS One 2013 8 10 e77299 24143218

PMC005xxxxxx/PMC5136323.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 2985117R 4816 J Immunol J. Immunol. Journal of immunology (Baltimore, Md. : 1950) 0022-1767 1550-6606 27849168 5136323 10.4049/jimmunol.1601488 NIHMS824598 Article Overexpression of soluble Fas ligand following AAV gene therapy prevents retinal ganglion cell death in chronic and acute murine models of glaucoma Krishnan Anitha * Fei Fei *† Jones Alexander * Busto Patricia § Marshak-Rothstein Ann § Ksander Bruce R. *‡ Gregory-Ksander Meredith *‡ * Schepens Eye Research Institute, Massachusetts Eye and Ear Infirmary, Department of Ophthalmology, Harvard Medical School, Boston, MA, USA † Department of Ophthalmology, Xijing Hospital, Fourth Military Medical University, Peoples Republic of China § Department of Medicine, University of Massachusetts Medical School, Worcester, MA, USA Corresponding author: Meredith Gregory-Ksander, PhD, 20 Staniford Street, Boston, MA.02114, 617-912-7455, Meredith_gregory@meei.harvard.edu ‡ These authors contributed equally to this work 22 10 2016 14 11 2016 15 12 2016 15 12 2017 197 12 46264638 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. Glaucoma is a multifactorial disease resulting in the death of retinal ganglion cells (RGCs) and irreversible blindness. Glaucoma-associated RGC cell death depends on the pro-apoptotic and proinflammatory activity of membrane-bound FasL (mFasL). In contrast to mFasL, the natural soluble FasL cleavage product (sFasL) inhibits mFasL-mediated apoptosis and inflammation and is therefore a mFasL antagonist. DBA/2J (D2) mice spontaneously develop glaucoma and predictably RGC destruction is exacerbated by expression of a mutated membrane-only FasL (mFasL) gene that lacks the extracellular cleavage site. Remarkably, one time intraocular adeno-associated virus-mediated gene delivery of sFasL (AAV2.sFasL) provides complete and sustained neuroprotection in both the chronic D2 and acute microbead-induced models of glaucoma, even in the presence of elevated intraocular pressure (IOP). This protection correlated with inhibition of glial activation, reduced production of TNFα, and decreased apoptosis of RGCs and loss of axons. These data indicate that cleavage of FasL under homeostatic conditions, and the ensuing release of sFasL, normally limits the neurodestructive activity of FasL. The data further support the notion that sFasL, and not mFasL, contributes to the immune privileged status of the eye. Introduction Glaucoma, a leading cause of blindness worldwide, is a complex multifactorial disease characterized by the progressive loss of retinal ganglion cells (RGCs) (1). While elevated intraocular pressure (IOP) is a well-recognized risk factor for the development of glaucoma, and remains the only modifiable disease-associated parameter, recent studies demonstrate that reduction of IOP alone does not prevent disease progression in all patients and that RGC destruction can continue even after IOP has been successfully lowered (2, 3). Moreover, the high incidence of normal tension glaucoma (4, 5, 6) and the absence of neurodegeneration in some patients with elevated IOP (2, 7), indicates that IOP-independent mechanisms also participate in the development and progression of glaucoma. As a result, current research is focused on developing neuroprotective therapies to complement conventional IOP lowering-drugs with the goal of preventing optic nerve degeneration and preserving vision. Growing evidence in clinical and experimental studies over the past decade strongly suggest the involvement of the immune system in glaucoma (8, 9, 10). In experimental models of glaucoma, treatment with anti-inflammatory drugs such as minocycline (11) or treatment with antibodies that inhibit TNFα (12, 13) resulted in increased RGC survival. Enhanced RGC survival was also observed in mouse models of glaucoma where the infiltration of macrophages, present in both human and experimental models of glaucoma (13, 14, 15) was inhibited through the use of Mac1−/− mice (13) or ocular irradiation (15). Together these studies demonstrate a critical role for inflammation in the pathogenesis of glaucoma. Fas Ligand (FasL) is a 40 kDa type II transmembrane protein of the TNF family, originally identified by its capacity to induce apoptosis in Fas receptor positive cells (16) and mediate activation induced cell death in T cells (17–19). However, there is increasing evidence that activation of the Fas receptor can also result in inflammation (20–23). FasL can be expressed as a membrane-bound protein (mFasL), that is both pro-apoptotic and proinflammatory, or cleaved and released as a soluble protein (sFasL) that inhibits both the apoptotic and inflammatory activity of mFasL (24–26). In vitro, we demonstrated that mFasL, but not sFasL, stimulated purified peritoneal macrophages and neutrophils to secrete the proinflammatory mediators MIP and IL1β (27, 28). As a result of its pro-apoptotic and proinflammatory functions, FasL is a potentially destructive molecule, whose expression is normally tightly regulated. One exception is in the eye where FasL is constitutively expressed in the cornea and retina (29). Importantly, in the context of experimental glaucoma, FasL expression is increased on resident microglial cells (27, 30) and the Fas/FasL signaling pathway is required for RGC death (27). However, in the eye, the form, as well as, the level of FasL expression is critical in determining the outcome of FasL expression (31, 32). We previously demonstrated that transduced ocular tumor cells expressing high levels of mFasL terminate immune privilege and induce a potent inflammatory response (31). By contrast, immune privilege was maintained and inflammation prevented in the presence of high levels of sFasL (31). Using a TNFα-induced “normal-tension” model of glaucoma we previously demonstrated that gene-targeted mice expressing only the membrane form of FasL exhibited accelerated RGC death and that a single intravitreal injection of recombinant sFasL provided short-term protection to RGCs in these mFasL-only mice (27). Together these data revealed a critical neurotoxic effector function for mFasL and neuroprotective function of sFasL in glaucoma. In the current study, we further explored the neurodestructive function of mFasL and neuroprotective function of sFasL in the development and progression of glaucoma in two of the most widely used mouse models of elevated-IOP-induced glaucoma, the spontaneous genetic-based D2 (DBA/2J) mouse model and the microbead-induced mouse model. Our data reveal that mFasL participates in the development and progression of glaucoma through activating both apoptotic and non-apoptotic signaling pathways. In addition, AAV-mediated gene delivery of sFasL (AAV2.sFasL) to the retina provided long-term protection to RGCs and optic nerve axons in both the D2 chronic mouse model of glaucoma and the acute mouse model of microbead-induced elevated IOP. Together, these data reveal the pleotropic effects of sFasL on glial activation, inflammation, and apoptosis of RGCs and provide proof-of-principal that AAV2.sFasL can provide complete and sustained neuroprotection of RGCs in both chronic and acute mouse models of glaucoma, even in the presence of ongoing elevated IOP. Materials and Methods Animals All animal experiments were approved by the Institutional Animal Care and Use Committee at Schepens Eye Research Institute and were performed under the guidelines of the Association of Research in Vision and Ophthalmology (Rockville, MD). The DBA/2J-Gpnmb +/SjJ mice (D2-Gp) were purchased from Jackson Laboratories (Bar Harbor, ME). DBA/2J mice expressing only membrane FasL (this cleavage site deleted-mouse line was designated as D2-ΔCS) were produced by crossing the ΔCS founder mice (27) to DBA/2J mice purchased from Jackson Laboratories for 10 generations. After 10 generations, D2-ΔCS mice heterozygous for the ΔCS mutation were intercrossed to produce D2.ΔCS homozygous for the ΔCS mutation (D2-ΔCS) mice and WT littermates (D2). All three genotypes (D2, D2-ΔCS, and D2-Gp), were housed and maintained under cyclic light (12L-30 lux:12D) conditions in an AAALAC approved animal facility at the Schepens Eye Research Institute and equal number of males and females were used in each study. For the microbead studies, 8 week old C57BL/6J WT mice were purchased from Jackson laboratories (Bar Harbor, ME). Viral Vector Construction The Adeno-associated vectors, scAAV2-CB6-PI-eGFP and scAAV2-CB6-PI-sFasLecto, were prepared at the Gene Therapy Center (Dr. Gaungping Gao, University of Massachusetts Medical School, Worcester MA). The seed plasmid pAAVsc CB6 PI sFasLecto was made by isolating a 582 bp fragment corresponding to the full cDNA sequence of sFasL ecto 5”GCSF from pcDNA3-sFasL-ecto plus 5′ mGCSF (31). The 582 bp fragment was ligated (T4 ligase NEB #M0202) to a BamHI, HindIII double digested and Antartic phosphatase treated pAAVsc-CB6-PI plasmid provided by the Gene Therapy Center, UMMS. An aliquot of the ligation mixture was transfected onto E.Coli competent cells and plated onto LB-Agar plus 50 ug/ml Carbenicillin. Colonies were picked and grown on LB-Carbenicillin media, and plasmid mini-preps were checked by restriction enzyme and sequencing. A large scale of the plasmid was prepared from 1 liter of bacterial culture and isolated using a ZR Plasmid gigaprep kit (#D4056, Zymo Research, Irvine CA) IOP measurements IOP was measured with a rebound TonoLab tonometer (Colonial Medical Supply, Espoo, Finland), as previously described (27). Mice were anesthetized by 3% isoflurane in 100% oxygen (induction) followed by 1.5% isoflurane in 100% oxygen (maintenance) delivered with a precision vaporizer. IOP measurement was initiated within 2 to 3 min after animals lost toe pinch reflex or tail pinch response. Anesthetized mice were placed on a platform and the tip of the pressure sensor was placed approximately 1/8 inch from the central cornea. Average IOP was displayed automatically after 6 measurements after elimination of the highest and lowest values. This machine-generated mean was considered as one reading, and six readings were obtained for each eye. For D2.Gp, D2, and D2.ΔCS mice, IOPs were measured once a month. For the microbead study, baseline IOPs were obtained one day before microbead injection and the IOPs were then measured every three days after microbead injection. All IOPs were taken at the same time of day (between 10:00 and 12:00 hours) due to the variation of IOP throughout the day. Intravitreal injections The intravitreal injections, just posterior to the limbus-parallel conjunctival vessels, were performed as previously described (27). Mice received a 1μl intravitreal injection into both eyes containing AAV2.sFasL (3x1012 PFU/ml) or AAV2.eGFP (3x1012 PFU/ml) and sterile physiologic saline was used as a vehicle control. Induction of elevated IOP Mice were anesthetized by intraperitoneal injection of a mixture of ketamine (100 mg/kg; Ketaset; Fort Dodge Animal Health, Fort Dodge, IA) and xylazine (9 mg/kg; TranquiVed; Vedco, Inc., St. Joseph, MO) supplemented by topical application of proparacaine (0.5%; Bausch & Lomb, Tampa, FL). Elevation of IOP was induced unilaterally by injection of polystyrene microbeads (FluoSpheres; Invitrogen, Carlsbad, CA; 15-μm diameter) to the anterior chamber of the right eye of each animal under a surgical microscope, as previously reported (33, 34). Briefly, microbeads were prepared at a concentration of 5.0 × 106 beads/mL in sterile physiologic saline. The right cornea was gently punctured near the center using a sharp glass micropipette (World Precision Instruments Inc., Sarasota, FL). A small volume (2 μL) of microbeads was injected through this preformed hole into the anterior chamber followed by injection of an air bubble via the micropipette connected with a Hamilton syringe. Any mice that developed signs of inflammation (clouding of the cornea, edematous cornea etc) were excluded from the study. Quantification of retinal ganglion cells The neural retina was isolated and fixed in 4% paraformaldehyde for 2 hour at room temperature. Retinal flat mounts were incubated at 4 °C overnight with a primary antibody against an RGC-specific marker, β-III-tubulin (Millipore, Billerica, MA) (33, 34). An Alexa Fluor 594–conjugated secondary antibody (Invitrogen) was used as secondary antibody. Nuclei were counterstained with DAPI (Vector Stain). 60X oil-immersion was used and sixteen non-overlapping images were taken, with four-five images within each quadrant. All cells in ganglion cell layer positively labeled by the β-III-tubulin antibody were counted. Retinal areas were measured using Image J software, and the RGC density/mm2 retina was calculated. Quantification of optic nerve axons For quantification of axons, optic nerves were dissected and fixed in karnovsky’s reagent (50% in phosphate buffer) overnight. Semi-thin cross-sections of the nerve were taken at 1.0 mm posterior to the globe and stained with 1% p-phenylenediamine (PPD) for evaluation by light microscopy. Ten non-overlapping photomicrographs were taken at 100× magnification covering the entire area of the optic nerve cross-section. The total area of each optic nerve cross-section was determined with Image J software and this value was used to estimate the total axon density per optic nerve cross-section. TUNEL staining Apoptotic cells were evaluated by TUNEL staining. Eyes were enucleated, fixed in 10% formalin. Eyes were processed and 20 μm paraffin sections mounted on slides were used for staining. The sections were de-paraffinized before staining for TUNEL according to the manufacturer’s protocol (In Situ Cell Death Detection Kit; Roche Applied Science, Mannheim Germany, 11684795910) as previously described (25). The sections were mounted using VECTARSHIELD Mounting Medium with DAPI (Vector Laboratories, H-1200). Sections were washed, cover slipped and examined by confocal laser scanning microscopy ((Leica Microsystems and processed using SP6 software). Quantitative RT-PCR At the time point for euthanasia anesthetized mice were perfused through the left ventricle with 10 ml of 1xPBS. Following perfusion, the eyes were enucleated and RNA was isolated from the neural retina using Qiagen RNeasy mini kit (Cat # 74104) according to the manufacturer’s protocol. RNA was treated with DNAse (Cat # AM222, Invitrogen) to ensure no contamination of genomic DNA. 500ng of RNA was used to prepare cDNA using the iScript cDNA synthesis kit from Bio-Rad (Cat # 170-8890) according to the manufacturer’s protocol. cDNA was treated with RNaseH (18021-014, Invitrogen) to ensure absence of single stranded RNA. Quantitative real time PCR was performed using the master mix from Invitrogen (Cat # 4472903) using the manufacturer’s protocol in 10μl total volume in duplicates. Relative expression to house keeping gene β-actin was quantified using the formula: Relative expression =10,000 x 1/2^(Avg. Gene cT- Avg. β-actin cT). All of the primers used are listed in Table I. Western blot analysis To assess total FasL expression in the posterior segments of D2, D2-Gp, and D2-ΔCS mice, protein lysates (20μg/sample) were prepared from posterior eye cups (neural retina, choroid, and sclera) as previously described (27). To assess over expression of sFasL in the neural retina of AAV2 treated eyes, protein lysates (5ug/sample) were prepared from the neural retina only. Proteins were separated on 12% Tris-glycine gels (Invitrogen, Carlsbad CA) and transferred to polyvinylidene difluoride membranes (Invitrogen, Carlsbad CA). The membranes were probed for Fas ligand using a polyclonal rabbit anti-Fas ligand antibody (25) followed by an IRDye secondary antibody (Li-Cor Biotechnology, Lincoln NE). Mouse β actin was used to assess equal loading. Each blot was developed using the LI-COR Odyssey imaging system (LI-COR Biotechnology, Lincoln NE). Densitometry was performed using Image Studio software. L5178Y-R tumor transfectants over expressing sFasL or mFasL (25, 27) were used as positive controls and a protein lysate prepared from a posterior eyecup prepared from FasL-KO mice (28, 35) was used as a negative control. Statistics “All data are presented as mean ± SEM and Graph Pad Prism 6 (La Jolla, CA, USA) was used to perform statistical analysis of the data For comparisons in the D2.ΔCS studies, one-way ANOVA and Tukeys multiple comparison test were used. For comparisons in the D2-AAV2 studies, two-way ANOVA and Tukey’s multiple-comparison test were used. For comparisons in the microbead-model studies, two-way ANOVA and Sidak’s multiple-comparison test were used. A P-value of less than 0.05 was considered significant. Results Accelerated glaucoma in D2.ΔCS mice DBA/2J (D2) mice spontaneously develop age-related elevated intraocular pressure (IOP) due to mutations in Gpnmb and Tyrp1 genes that trigger iris stromal atrophy and pigment dispersion, respectively (36, 37). As a result, D2 mice develop elevated IOP by approximately 6–8 months of age, followed by the loss of RGCs and nerve fibers. We previously demonstrated that D2 mice expressing only mFasL (D2-ΔCS) displayed marked thinning of the nerve fiber layer at 5 months of age that was not seen in D2 mice that expressed WT FasL (27). Early thinning of the nerve fiber layer suggested that the development of glaucoma was accelerated. However, our previous study did not include a quantitative analysis of RGCs, optic nerve axons or apoptosis in the neural retina. To confirm the accelerated development of glaucoma in D2-ΔCS mice, we performed quantitative analysis of RGCs and axons and correlated this data with accelerated apoptosis in the neural retina. In addition, we now have age- and strain-matched control D2-Gp (DBA/2JGpnmb+/+) mice that do not develop glaucoma. These control D2-Gp congenic mice express a normal Gpnmb gene and develop a very mild form of iris disease due to the Tyrp1 mutation, but do not develop elevated IOP or glaucoma. To perform our analysis, we compared disease progression in young and old D2-ΔCS mice with D2 littermates that express WT FasL (positive control) and the D2-Gp mice that do not develop elevated IOP or glaucoma (negative control) (38). IOP was measured using rebound tonometry at 3, 6 and 9 months of age (Figure 1A). D2 and D2-ΔCS mice displayed elevated IOP beginning at 6 months of age, as compared with the negative control D2-Gp mice. However, there were no significant differences in IOP between D2 and D2-ΔCS mice at any age, indicating the ΔCS mutation does not affect IOP. To determine whether the loss of RGCs and/or axons was accelerated in D2-ΔCS mice, groups of mice were euthanized at 3 and 6 months of age and retinal flat mounts were stained with the RGC-specific marker βIII tubulin to assess RGC density and optic nerve sections were stained with paraphenylenediamine (PPD) to assess axon density. As reported previously, the loss of RGCs and axons in D2 mice is age-related and only develops in mice >8 months old (36, 37, 39). Therefore, at 3 and 6 months of age D2 displayed no significant loss of RGCs or axons, as compared with D2-Gp mice (Figure 1B, C, D, E). By contrast, D2-ΔCS mice developed early loss of RGCs and axons at 6 months, indicating expression of non-cleavable FasL in D2-ΔCS mice results in accelerated development of glaucoma. Numerous studies demonstrate that RGCs die by apoptosis during glaucoma (40–42). To demonstrate the accelerated death of RGCs in D2-ΔCS mice was due to apoptosis, TUNEL staining was performed on retinal sections from D2, D2-ΔCS, and D2-Gp mice at 3 months (Figure 2A, B) and 6 months (Figure 2C, D) of age. TUNEL+ cells were detected in the RGC layer of D2-ΔCS at both 3 and 6 months of age, as compared with D2 and D2-Gp mice. Moreover, at 6 months of age, the TUNEL positive cells detected in D2-ΔCS mice were no longer restricted to the RGC layer (Figure 2C). Together, these results indicate that D2-ΔCS mice display early death of RGCs via apoptosis, resulting in accelerated development of glaucoma. Expression of Fas, FasL and FADD in D2-ΔCS mice While previous studies demonstrated FasL was upregulated in the retina following elevated IOP (25, 26) the levels of sFasL and mFasL were not assessed. To determine the effects of elevated IOP on components of the Fas-induced apoptosis pathway in D2 and D2-ΔCS mice, the neural retina was analyzed by quantitative RT-PCR for the mRNA levels of Fas, FasL, and the Fas activating death domain (FADD) and Western blots were performed to quantitate protein expression. The Western blot for FasL was prepared from lysates of the posterior segment of the eye (retina, choroid, and sclera) which detected the smaller 26 kD sFasL band, as well as, two larger 34 kD and 38 kD mFasL bands (Figure 3D). The two mFasL bands (34 and 38 kD) were reported previously and represent differential glycosylation of mFasL (25, 27, 43, 44). The two forms of FasL (mFasL and sFasL) were present in the posterior segment from all three groups of mice (D2-Gp, D2, and D2-ΔCS mice). As an important specificity control for the Western blot, posterior eye segment lysates from FasL KO mice confirmed the loss of sFasL (26 kD band) and mFasL (34 and 38 kD bands). As expected, no sFasL was detected in the eye tissue from D2-ΔCS mice (Figure 3D), proving the knock-in mutation of the FasL cleavage site prevents expression of sFasL in the retina. There was no significant difference in Fas expression (mRNA and protein levels) between D2-Gp, D2, and D2-ΔCS mice at 3 or 6 months of age (Figure 3A and 3E). While there was a trend toward reduced Fas mRNA expression in the D2-ΔCS mice at 6 months, the difference was not statistically significant (P=0.15, Figure 3A). These data indicate that Fas expression in all three groups of mice remained constant and therefore cannot account for the accelerated development of glaucoma at 6 months in D2-ΔCS mice. In contrast with the constant Fas levels, there was a significant age-related increase in mFasL and FADD mRNA between 3–6 months in D2-ΔCS mice, which coincided with the beginning of RGC death at 6 months of age. Moreover, in D2 and D2-Gp mice, which displayed no RGC loss at 6 months, the levels of mFasL and FADD (mRNA and protein levels) were significantly reduced as compared with D2-ΔCS mice (Figure 3B, D and 3C, F). Previous work by our laboratory (27) and Ju et al., (30) demonstrated that within the neural retina, FasL is primarily expressed on microglia in the ganglion cell layer and inner nuclear layer of non-glaucomatous eyes. In glaucomatous eyes, FasL is expressed on microglia (and/or infiltrating macrophages) at higher levels and coincided with apoptosis of RGCs (27, 30). Thus the increased FasL expression in glaucomatous D2-ΔCS mice may account for the trend toward reduced Fas expression (Figure 3A) as some Fas+ cells are induced to undergo apoptosis. Increased engagement of the remaining Fas may also account for the increase in FADD. Together, these data indicate the accelerated disease development in D2-ΔCS mice coincided with an increase in the expression of the neurotoxic mFasL and FADD, the adapter protein required for the formation of the death-inducing signaling complex during apoptosis. Importantly, our Western blot analysis of mFasL and sFasL in the posterior segment of the eye showed that sFasL was the predominant form of FasL expressed in the retina of mice that possess the WT allele for FasL (D2-Gp and D2 mice). Therefore, in non-diseased retinas, under homeostatic conditions, most FasL is cleaved to release neuroprotective sFasL. It follows that the development of glaucoma in D2 mice coincides with a shift from the production of neuroprotective sFasL to neurodestructive mFasL. Muller cell activation in D2-ΔCS mice Muller cells, which span all retinal layers, constitute the principal glial cells of the retina and provide support to retinal neurons (45). In the normal retina, glial fibrillary acidic protein (GFAP) is expressed in the nerve fiber layer by astrocytes and also in the end feet of Muller cells (29, 39). However, in glaucoma, GFAP expression is markedly increased in Muller cells as a result of reactive gliosis, which is considered a key indicator of neuroinflammation (30, 46, 47). To monitor Muller cell activation, GFAP mRNA expression in the retina was quantified by RT-PCR, and protein levels were assessed by immunohistochemical staining using an anti-GFAP antibody. D2, D2-ΔCS, and D2-Gp mice were examined at 3 and 6 months of age. At 3 months of age, GFAP expression was restricted to astrocytes and Muller cell-end-feet in the nerve fiber layer in all three groups of mice and there was no significant difference in expression of GFAP mRNA among the 3 strains of mice (Figure 4A, B). However, by 6 months of age, expression of GFAP in D2-ΔCS mice was no longer limited to astrocytes and Muller cell-end-feet in the nerve fiber layer, but extended throughout the whole length of retinal Muller cells. This correlated with a modest but significant increase in GFAP mRNA levels in the D2-ΔCS mice, as compared to the D2 and D2-Gp groups (Figure 4 A, B). The increase in GFAP coincided with a significant increase in TNFα mRNA levels in D2-ΔCS mice as compared to D2 and D2-Gp mice (Figure 4C). TNFα is an immediate early gene that is rapidly transcribed in response to signals of stress and inflammation and is often used as a measure of macrophage and glial activation (48–52). In regards to glaucoma, a variety of rodent animal models have implicated glial-cell produced TNFα in the pathogenesis of elevated IOP-induced glaucoma (12, 13, 53, 54). These models include: rat episcleral vein cauterization model (13), mouse laser-induced photocoagulation of the limbus model (12), and the spontaneous D2 model (53, 54). These data indicate that accelerated loss of RGCs in D2-ΔCS mice that express high levels of mFasL coincides, not only with the induction of apoptosis of RGCs, but also with glial activation and the induction of the immediate early gene, TNFα. Intravitreal delivery of sFasL using AAV2 The membrane-bound form of FasL is responsible for both the induction of apoptosis and for proinflammatory cytokine production and therefore the exacerbated development of glaucoma in the D2-ΔCS mice can be explained by the increased expression of uncleaved mFasL. However, sFasL has been shown to serve as an mFasL antagonist that blocks FasL-induced apoptosis and inflammation (25, 26). Moreover, we previously found that a single intravitreal injection of recombinant sFasL protein can protect the nerve fiber layer in the short-term model of TNFα-induced glaucoma (27). Therefore, it was important to determine whether sFasL could also block the effects of mFasL in the widely accepted mouse models of spontaneous and induced elevated IOP glaucoma models. Previous studies demonstrated that an intravitreal injection of AAV2 vectors in mice results primarily in transduction of long-lived RGCs and cells of the inner nuclear layer (57–59). As a result, gene delivery using AAV2 can provide long-term consistent expression within the inner retina. To evaluate which retinal cells were transduced, an AAV2 vector expressing only eGFP (AAV.eGFP) was injected into the vitreous of two month-old D2 mice. Confocal microscopy was used to assess eGFP expression in retinal whole mounts at 2 weeks post injection and showed uniform eGFP distribution throughout the retina (Figure 5A). Cross sectional reconstruction of the retina showed strong eGFP expression throughout the ganglion cell layer and inner nuclear layers (Figure 5B). To determine whether we could maintain long-term sFasL expression in the retina of D2 mice, we used AAV2 vectors to deliver a gene encoding sFasL to the RGCs of D2 mice via a single intravitreal injection. Western blot analysis of cells isolated from the neural retina showed a significant increase in sFasL expression as early as 2 weeks post injection, as compared to D2 mice treated with either AAV2.eGFP or uninjected control mice (Figure 5C). Moreover, the increase in sFasL in the retina was sustained at a consistent level throughout the length of the study (8 months) following a single injection. Monthly monitoring of IOP revealed no significant differences in IOP levels at any age between D2 uninjected control mice and D2 mice that received either AAV2.eGFP or AAV2.sFasL treatment (Figure 5D). Furthermore, representative slit lamp images taken at 9 months of age showed no differences in the severity of either iris pigment dispersion or iris atrophy between D2 uninjected control mice and D2 mice that received either AAV2.eGFP or AAV2.sFasL treatment (Figure 5D). Taken together, these results demonstrate that a single intravitreal injection of AAV2.sFasL into D2 mice produces long-term stable expression of high levels of sFasL within the inner retina, which has no effect on the development of: iris pigment dispersion, iris atrophy, and elevated IOP. Intravitreal delivery of AAV2.sFasL prevents loss of RGCs and axons and provides long-term protection in D2 mice To determine whether over expression of sFasL prevents the loss of RGCs and axons in D2 mice, mice received an intravitreal injection of AAV2.sFasL at 2 months of age and were euthanized 8 months later at 10 months of age. RGC density was measured in retinal whole mounts stained with anti-βIII tubulin antibody and axon density was measured in optic nerve sections. As expected in D2-uninjected mice and D2-AAV2.eGFP mice at 10 months of age, there was a significant decrease in both RGC and axon density as compared to the D2-Gp negative control mice that do not develop glaucoma (Figure 6A–D). By contrast, D2 mice treated with AAV2.sFasL displayed no significant loss in either RGC or axon density as compared with D2-Gp control mice (Figure 6B, D). To demonstrate that AAV2.sFasL treatment provides long-term protection and does not just delay the death of RGCs and axons, additional groups of mice were followed for up to 15 months of age. Quantification of RGCs and axons revealed that even at 15 months of age, the AAV2.sFasL treated D2 mice continued to display significant protection of RGCs and axons when compared with D2-uninjected mice and AAV2.eGFP mice (Figure 6E–H). In conclusion, these data demonstrate that overexpression of sFasL resulting from a single intravitreal injection provides complete and long-term protection to the RGCs and axons in D2 mice, even in the presence of elevated IOP. sFasL prevents activation of pro-inflammatory Muller cells in D2 mice As we previously demonstrated (Figure 4A, B, C), loss of RGCs in glaucomatous D2 mice coincided with activation of Muller cells that express GFAP and the induction of TNFα. To determine whether the absence of glaucoma in AAV2.sFasL treated D2 mice coincided with reduced Muller cell activation, GFAP expression in Muller cells was determined by immunohistochemical staining and RT-PCR was used to measure GFAP and TNFα mRNA expression in the neural retina of 10 month old: D2-untreated, D2-AAV2.eGFP, D2-AAV2.sFasL, and D2-Gp mice. As expected, D2-untreated and D2-AAV2.eGFP mice that develop glaucoma displayed activated pro-inflammatory Muller cells that express high levels of GFAP as seen in confocal images (Figure 7A) and RT-PCR revealed a significant increase in GFAP and TNFα mRNA expression (Figure 7B, C). By contrast, in D2-AAV2.sFasL mice that do not develop glaucoma, Muller cells express significantly less GFAP and TNFα mRNA levels were equivalent to those seen in D2-Gp normal control mice (Figure 7A, B, C). These data demonstrate that in D2-AAV2.sFasL mice the sFasL-mediated protection of RGCs and axons coincides with preventing activation of Muller cells and induction of the immediate early gene, TNFα. sFasL prevents upregulation of apoptotic genes and downregulation of pro-survival genes in D2 mice In glaucoma, both extrinsic and intrinsic pathways of apoptosis have been identified as mediators of RGC apoptosis (12, 27, 60–62). To determine whether treatment with sFasL inhibited the induction of both extrinsic and intrinsic pathways of apoptosis, qRT-PCR was performed on retinal tissue from D2, D2-Gp, D2-AAV2.eGFP, and D2-AAV2.sFasL mice at 10 months of age to assess the expression of apoptotic genes and pro-survival genes. At 10 months of age, there were two distinct gene expression patterns in untreated D2 mice as compared with sFasL treated D2 mice. In untreated D2 mice, pro-apoptotic genes in both intrinsic (BAX) and extrinsic (Fas, FADD) pathways of apoptosis were upregulated, while anti-apoptotic genes (cFLIP, Bcl2, cIAP-2) were down regulated. By contrast, mice treated with sFasL prevented both the induction of pro-apoptotic genes and downregulation of anti-apoptotic genes (Figure 7D and E). In D2 and D2-AAV.eGFP mice, the expression of pro-apoptotic genes Fas, FADD, and BAX were elevated in comparison to D2-Gp controls (Figure 7D). However, D2 mice treated with AAV2.sFasL prevented the upregulation of these genes (Fas, FADD, and BAX), which were equal to the levels detected in the D2-Gp control mice (Figure 7D). Similarly, the levels of the pro-survival genes (cFLIP, Bcl2, and cIAP-2) (Figure 7E) were down regulated in D2 and D2-AAV2.eGFP mice, but D2 mice treated with AAV2.sFasL prevented the downregulation of these genes, which were equal to the levels detected in the D2-Gp control mice. While mRNA expression does not always predict protein expression, these data indicate that (i) genes in both the intrinsic and extrinsic pathways of apoptosis are induced in glaucomatous D2 mice, which is in agreement with previous studies (53, 54, 61, 62), and (ii) treatment with sFasL prevents the induction of both of these pathways. Together, these results suggest that sFasL acts as an antagonist and prevents the binding of mFasL to the Fas receptor. Therefore, sFasL is preventing the mFasL-induced up-regulation of pro-apoptotic genes and the mFasL-induced down regulation of anti-apoptotic genes. sFasL prevents loss of RGCs and axons in a microbead-induced model of glaucoma To demonstrate that the neuroprotective effect of sFasL was not limited to the D2 model of glaucoma, we tested whether sFasL prevented RGC loss in a microbead-induced mouse model of elevated IOP, where an injection of microbeads into the anterior chamber of the eye obstructs the aqueous humor outflow and induces elevated IOP (33, 34). C57BL/6 mice received an intravitreal injection of AAV2.eGFP or AAV2.sFasL followed three weeks later by an anterior chamber injection of microbeads or saline as a negative control. At four weeks post microbead or saline injection, Western blot analysis confirmed the overexpression of sFasL in the neural retina of AAV2.sFasL treated mice as compared to the AAV2.eGFP treated mice (Figure 8A). As previously demonstrated (33,34) a single anterior chamber injection of 15μm polystyrene microbeads resulted in elevated IOP for up to 21 days in both AAV2.eGFP and AAV2.sFasL mice as compared to saline controls (Figure 8B). IOPs peaked at 11 days post microbead injection and there was no significant difference in the time course or magnitude of the microbead-induced elevated IOP between AAV2.sFasL or AAV2.eGFP treated mice, indicating sFasL did not affect IOP. Quantification of RGC density at 4 weeks post microbead injection revealed a significant decrease in RGC density in microbead-injected AAV2.eGFP treated mice as compared to the saline-injected control (Figure 8C, D). However, treatment with AAV2.sFasL protected RGCs and the RGC density in microbead-injected AAV2.sFasL mice was equal to the RGC density in the saline-injected controls (Figure 8C, D). Similar results were observed in the optic nerve where AAV2.sFasL treatment afforded complete protection of axons as compared to the AAV2.eGFP treated control group (Figure 8E, F). Taken together, these results demonstrate that AAV2.sFasL provides complete neuroprotection in both the chronic D2 and the acute microbead-induced models of elevated IOP. Discussion FasL is constitutively expressed in a variety of cells in the eye and other immune privileged sites where it has been implicated in the deletion of activated Fas-expressing effector cells and therefore the prevention of inflammation (29, 64). However, this protective activity of FasL has been challenged by numerous examples of FasL-associated ocular and CNS pathologies (22, 27, 30, 65, 66). For example, FasL has been shown to play a major role in animal models of glaucoma where it contributes to the death of Fas+ RGCs (27, 30). This apparent dichotomy, constitutive expression of FasL in non-diseased eyes and the association of FasL with numerous ocular pathologies, can be explained by the isoform of expressed FasL. Membrane-bound FasL is clearly proapoptotic and proinflammatory, but can be cleaved to release a soluble molecule that is incapable of mediating either cell death or apoptosis and in fact has been found to antagonize the activity of mFasL (24–26). However, the actual role of membrane vs soluble FasL at immune privileged sites has not been carefully examined. Herein, we show that under homeostatic conditions, FasL in the eye is primarily cleaved to the soluble form and is therefore anti-inflammatory. Exactly how the balance between sFasL and mFasL shifts during diseases such as glaucoma is unclear, but our data demonstrate that mFasL is definitely the pathogenic isoform. We have explored the role of mFasL vs sFasL in gene-targeted ΔCS mice where the FasL cleavage site has been mutated and can no longer be cleaved by metalloproteinases (27, 67). In ΔCS mice, total FasL expression is controlled by the endogenous FasL promoter as in WT mice, but in ΔCS mice, cells that normally express FasL can only make mFasL and not sFasL. As a result, ΔCS mice develop accelerated and exacerbated destruction of retinal cells in both spontaneous and inducible models of glaucoma (27; current study). These data point to a key role for mFasL in glaucoma pathogenesis. The current study now demonstrates that sFasL can prevent the development of glaucoma. We previously demonstrated that a single intravitreal administration of recombinant sFasL protein provided short-term protection to RGCs in a TNFα-induced “normal-tension” model of glaucoma (28). However, clinically glaucoma is a slow progressing chronic disease, requiring a treatment that can provide long-term protection. To further explore the function of sFasL in glaucoma, we constructed AAV-vectors that incorporated the genes for GFP or sFasL and injected them into the vitreous cavity of WT mice. GFP expression was monitored in vivo by fundus examination and confirmed in vitro by confocal microscopy of retinal flat mounts. We found that GFP was expressed in the RGC layer and inner nuclear layer for at least 8 months following a single injection. Likewise, Western blots confirmed sFasL expression in the neural retina of AAV-sFasL mice extended over the same time period. Importantly AAV-sFasL gene therapy resulted in complete long-term neuroprotection of RGCs and axons of the optic nerve, even in the continued presence of elevated IOP, in both spontaneous and inducible models of glaucoma. Therefore the increased disease severity in ΔCS mice most likely reflects excessive mFasL expression, as well as, the absence of sFasL-dependent inhibition. Fas ligand is a 40-kDa Type II transmembrane protein that was initially identified as an inducer of apoptosis in Fas expressing cells (16). Using a TNFα-induced “normal-tension” model of glaucoma we previously demonstrated that gene-targeted mice expressing only mFasL exhibited accelerated RGC death. However, in addition to the proapoptotic effects of mFasL, the results of our current study indicate that mFasL also plays a role earlier in the pathogenesis of glaucoma via non-apoptotic pathways. FasL is known to directly and rapidly promote the transcription and secretion of a number of proinflammatory cytokines and neutrophil chemokines (28, 31, 68) that likely account for the massive neutrophil influx following the injection of FasL-transfected cells and tumors (31). FasL can also induce the processing of IL-1β and IL-18 through an inflammasome independent caspase 8-dependent mechanism (68). GFAP is a well-known marker of glial activation and is significantly increased in retinal astrocytes and muller cells following elevated IOP (46, 47, 69). As indicated by RT-PCR and immunofluorescence, we have now shown a significant increase in the expression of GFAP that occurs much earlier in D2-ΔCS mice than D2-WT mice. This early upregulation of GFAP indicates accelerated activation of glial cells in D2-ΔCS mice. Moreover, the accelerated glial activation in D2-ΔCS mice also correlated with an earlier induction of TNFα when compared to D2-WT mice. There is substantial evidence that TNFα is induced following elevated IOP (10, 12, 13, 53, 54,70) and is required for the death of RGCs (12, 13). Using the TNFα-induced model of glaucoma (12), we previously demonstrated that TNFα increased the expression of mFasL on retinal microglia and/or infiltrating macrophages and that Fas-FasL signaling was required for TNFα-mediated death of RGCs (28). While retinal microglia are also a cellular source for TNFα following elevated IOP (12, 13), in the absence of Fas or FasL, RGCs are protected from apoptosis (27). Taken together with the current study, these data suggest that TNFα and mFasL may form a positive feed-forward-loop where TNFα upregulates mFasL expressed on retinal microglia and mFasL induces apoptosis of RGCs while at the same time feeding back on the microglia and/or infiltrating macrophages to upregulate TNFα, thus leading to chronic neuroinflammation. The ability of AAV2.sFasL to prevent induction of TNFα, as well as, the apoptosis of RGCs suggest that sFasL directly blocks FasL-induced apoptosis of RGCs, while at the same time inhibits the TNFα-FasL positive feed-forward- loop, effectively shutting down neuroinflammation. The eye has long been considered an immune privileged site as evidenced by the high acceptance rate of HLA unmatched allogeneic corneal transplants in uncomplicated or “low-risk” patients in which the recipient graft bed shows no sign of inflammation and/or neovascularization (71, 72). However, corneal immune privilege is not universal as demonstrated by the high rejection rate of allogeneic corneal transplants in “high-risk” patients in which donor tissue is transplanted into a recipient with on-going corneal inflammation and/or neovascularization (73). FasL was first implicated in maintaining immune privilege when it was shown that FasL-deficient corneal allografts were readily rejected (29, 74). However, overexpression of mFasL on the donor cornea, as a result of viral transfection, resulted in accelerated transplant rejection associated with innate-mediated inflammation (75). Together these data suggest that the soluble isoform may be responsible for the acceptance of corneal allografts in low-risk patients and raise the possible application of sFasL gene therapy for facilitating acceptance of corneal allografts in high-risk patients. Over the past decade gene therapy has been tested in animal models of glaucoma as a treatment for glaucoma, delivering: neurotrophins (58), anti-apoptotic genes (57), and transcription factors (76) to the retina. There is substantial evidence in both human and experimental glaucoma animal models that RGCs die by apoptosis, but the molecular mechanisms that trigger apoptosis are not well understood (40–42). However, it is clear that simply blocking the apoptotic pathway in RGCs does not prevent completely neurodegeneration in glaucoma. This was illustrated in studies by Libby et al., who demonstrated that RGC somata of Bax-deficient D2 mice survived indefinitely in the presence of elevated IOP, while the axons continued to degenerate distal to the optic nerve head (61). In a related study by McKinnon et al., treatment with an adeno-associated virus (AAV) vector expressing baculoviral IAP repeat-containing protein-4 (BIRC4), a potent inhibitor of Caspase-3, -7, and -9, was only able to protect 50% of optic nerve axons (77), demonstrating that blocking only the apoptotic pathway in the RGCs is not sufficient to prevent glaucoma completely. Several hypotheses have been proposed regarding the mechanism(s) that induce axonal degeneration and death of RGCs including oxidative stress (78), ischemia (79), excitotoxicity (80), and mitochondrial dysfunction (8). Although targeting these pathways has been successful in experimental models of glaucoma, these strategies have generally failed to translate into clinical success. This failure may be due to multiple reasons, including the lack of a suitable animal model, as well as, the fact that glaucoma is a multifactorial disease where numerous factors interact during disease progression. Therefore, the most effective neuroprotective strategy should target more than one pathway, work long-term, and show efficacy in more than one animal model of glaucoma. AAV viral vectors have proven to be very effective at delivering transgenes to the retina and providing long-term gene expression that persists in the adult retina for over a year (81, 82). In particular, the AAV2 serotype specifically targets the cells of the ganglion cell layer and inner nuclear layer following an intravitreal injection (81). A recent report by Sullivan et al., demonstrated that AAV-mediated delivery of erythropoietin (AAV.EPOR76E) protected RGCs and axons in the D2 model of glaucoma (83). However, an additional study was performed to determine if the AAV.EPOR76E could also protect RGC axons in the mouse microbead occlusion model of elevated IOP and while the AAV.EPOR76E treated group appeared to have fewer degenerating axons when compared to the untreated group, there was no significant difference in the total number of axons between groups (84). In the current study using AAV2.sFasL, we provide the first direct evidence of a virus-mediated gene therapy that provides complete and long-term protection, out to 15 months of age, of both RGCs and axons of the optic nerve in the chronic D2 mouse model of glaucoma. Treatment with AAV2.sFasL inhibited multiple pathways activated in the pathogenesis of glaucoma in the D2 mice including activation of retinal glial cells, induction of TNFα, and death of RGCs and axons, even in the presence of elevated IOP. Moreover, the efficacy of this gene therapy was even further validated in a second mouse microbead occlusion model of elevated IOP, again revealing complete protection of RGCs and axons. Exactly how sFasL attenuates mFasL activity remains to be determined. The sFasL may simply bind to Fas but fail to effectively crosslink the receptor, while at the same time preventing activation of the receptor complex by the membrane-bound form. Alternatively, sFasL may actively engage the receptor and lead to its internalization. Further studies will also need to address how the appropriate balance between sFasL an mFasL is maintained under normal conditions, and then disrupted under pathological conditions. Grant support: The study was supported by the NIH Core grant P30EY003790 (Schepens) and NIH grants AR055634 (AMR) and EY021543 (MGK). The authors would like to thank Philip Seifert from morphology core at Schepens Eye Research Institute for processing and PPD staining of optic nerves. We would also like to thank Dr. Guangping Gao and the Gene Therapy center and Vector Core at the University of Massachusetts Medical School for producing the AAV2.sFasL and AAV2.eGFP vectors. Non-standard abbreviations RGC retinal ganglion cell IOP intraocular pressure FasL Fas ligand sFasL soluble Fas ligand mFasL membrane-bound Fas ligand AAV adeno-associated virus PPD p-phenylenediamine GFAP Glial fibrillary acidic protein cFLIP cellular FADD-like IL-1β-converting enzyme-inhibitory protein FADD Fas-associated death domain BAX bcl-2-like protein 4 Bcl2 B-cell lymphoma 2 cIAP-2 cellular inhibitor of apoptosis protein-2 Fig 1 Accelerated loss of RGCs and axons in D2-ΔCS mice (A) IOP measurements were taken by rebound tonometry in D2-Gp, D2, and D2-ΔCS mice at 3, 6, and 9 months of age (N=16 D2-GP; N=22 D2, N=22 D2-ΔCS). Data is presented as mean IOP ± SEM. (B) Representative confocal images of retinal flat-mounts isolated from D2-Gp, D2, and D2-ΔCS mice at 3 and 6 months of age, stained with β-III tubulin (red) a RGC-specific marker and DAPI a nuclear stain (blue) (Scale bar, 75μm). (C) Quantification of β-III tubulin positive RGCs, represented as RGC density/mm2 retina. N=10 eyes per group. (D) Representative photomicrographs of PPD optic nerve cross sections taken from D2-Gp, D2 and D2-ΔCS at 3 and 6 months of age (Scale bar, 10μm). (E) Quantification of healthy axons, represented as axon density (104)/mm2 ON. N=10 optic nerves per group. Ns, non-significant, ****P<0.0001. Fig 2 Accelerated apoptosis coincides with RGC loss in the absence of sFasL Representative TUNEL staining in paraffin embedded retinal sections taken from D2-Gp, D2 and D2-ΔCS mice at (A) 3 and (C) 6 months of age (Scale bar, 100μm). TUNEL = red, DAPI, nuclear marker = blue. GCL, ganglion cells layer; INL, inner nuclear layer; ONL, outer nuclear layer. White arrowhead = TUNEL positive cells in GCL, white arrow = TUNEL positive cells in INL. TUNEL positive cells in the GCL were quantitated at (B) 3 months and (D) 6 months of age, shown as the number of TUNEL positive cells in the GCL/retinal section (9 sections/retina), N=8 per group. ****P<0.0001. Fig 3 Expression of Fas, FasL, and FADD in D2-ΔCS mice Quantitative RT-PCR was performed on the neural retina isolated from D2-GP, D2, and D2-ΔCS mice at 3 and 6 months of age to quantitate mRNA levels of: (A) Fas, (B) FasL, and (C) FADD. N=6 per group. Representative Western blots from protein lysates (20μg/sample) prepared from posterior eye cups isolated from D2-GP, D2, and D2-ΔCS mice for (D) FasL at 3 and 6 months (actin is green, FasL is red), (E) Fas at 6 months (actin is green, Fas is red), and (F) FADD at 6 months (actin is green, FADD is green). Protein lysates prepared from L5178Y-R tumor cells transfected with sFasL or mFasL vectors (31) were used as positive controls for FasL and the protein lysate from posterior eyecups of FasL-KO mice (27, 35) was used as a negative control for FasL. Densitometry analysis of mFasL (34 and 38 kD bands), sFasL (26 kD band), Fas (43 kD), and FADD (28 kD) is the average of 3 independent experiments (3 independent blots consisting of 1 posterior segment per group, per experiment) Error bars indicate SEM; N.D.- not detected (below the level of detection by densitometry). *P<0.05, **P<0.01, ****P<0.0001. Fig 4 Expression of GFAP and TNFα in the retina of D2-ΔCS mice (A) Representative confocal microscopy images of paraffin embedded retinal sections taken from D2-Gp, D2, and D2-ΔCS mice at (A) 3 and 6 months of age and stained for GFAP (red) and DAPI (blue). (white arrow=GFAP in muller cells, Scale bar, 100μm). Quantitative RT- PCR was performed on the neural retina isolated from D2-GP, D2, and D2-ΔCS mice at 3 and 6 months of age to quantitate mRNA levels of (B) GFAP and (C) TNFα. N=6 per group. Error bar indicates SEM. **<0.01 and ***P<0.001. Fig 5 Over expression of sFasL in D2 mice D2 mice received one intravitreal injection of AAV2.sFasL or AAV2.eGFP as a control at 2 months of age. (A) Retinal flat-mount showing expression of AAV2.eGFP throughout the retina (Magnification, 5x and 40x). (B) Cross sectional 3-D reconstruction of the retina. GCL-ganglion cell layer, INL-inner nuclear layer. (C) Western blot from neural retinal lysates (5μg/sample, sFasL is red band, actin is green band) showing overexpression of sFasL (26 kD) at both 2wk and at 8 months post AAV2 injection. N=3 per group. Densitometry is the average of 3 independent experiments (3 independent blots consisting of 1 neural retina per group, per experiment). (D) IOP measurements were taken at 3, 5, 7, and 9 months of age by rebound tonometry in D2 (uninjected), D2-AAV2.eGFP, and D2-AAV2.sFasL. Mean IOP for the D2-Gp non-glaucomatous control group is 12 mmHg ±3 SEM, identified as an solid line on the IOP graph. Data is presented as mean IOP ± SEM IOP (N=10 per group). Representative slit lamp images taken at 9 months of age show pigment dispersion and iris stromal atrophy in D2-Gp, D2 (uninjected), D2-AAV.eGFP, and D2-AAV.sFasl mice. Ns, non-significant, *P<0.05. Fig 6 Intravitreal AAV2.sFasL protects RGCs and axons in D2 mice D2 mice received one intravitreal injection of AAV2.sFasL or AAV2.eGFP as a control at 2 months of age. At 10 and 15 months of age, AAV2.sFasL and AAV2.eGFP treated mice, in addition to age-matched D2-uninj (uninjected) and D2-Gp mice were euthanized and retinas and optic nerves processed for analysis of RGC and axon density. (A) Representative confocal images of retinal flat-mounts isolated from 10 month old D2-Gp, D2 uninjected, D2-AAV2.eGFP and D2-AAV2.sFasL mice, stained with β-III tubulin (red) an RGC-specific marker and DAPI a nuclear stain (blue). (Scale bar, 50μm). (B) Quantification of β-III tubulin positive RGCs, represented as RGC density/mm2 retina. N=10 per group (C) Representative photomicrographs of PPD stained optic nerve cross sections taken from 10 month old D2-Gp, D2 uninjected, D2-AAV2.eGFP and D2-AAV2.sFasL mice (Scale bar, 10μm). (D) Quantification of healthy axons, represented as axon density (104)/mm2. N=10 per group (E) Representative confocal images of retinal flat-mounts isolated from 15 month old D2-Gp, D2 uninjected, D2-AAV2.eGFP and D2-AAV2.sFasL mice, stained with β-III tubulin (red) an RGC-specific marker and DAPI a nuclear stain (blue) (Scale bar, 75μm). (F) Quantification of β-III tubulin positive RGCs, represented as RGC density/mm2 retina. N=5 per group. (G) Representative photomicrographs of PPD optic nerve cross sections taken from 15 month old D2-Gp, D2 uninjected, D2-AAV2.eGFP and D2-AAV2.sFasL mice (Scale bar, 10μm). (H) Quantification of healthy axons, represented as axon density (104)/mm2. N=5 per group. *P<0.05, **P<0.01, ***P<0.001 ****P<0.0001 Fig 7 Muller glial cell activation and induction of inflammatory and apoptotic mediators in the retina of D2 mice treated with AAV2.sFasL (A) Representative confocal microscopy images of paraffin embedded retinal sections taken from D2-Gp, D2-uninj. D2-AAV2-eGFP, and D2-AAV2-sFasL mice at 10 months of age and stained for GFAP (red) and DAPI (blue) (Scale bar, 100μm). Quantitative RT- PCR was performed on the neural retina isolated from D2-Gp, D2-uninj., D2-AAV2-eGFP, and D2-AAV2-sFasL mice at 10 months of age to quantitate mRNA levels of (B) GFAP, (C) TNFα, (D) pro-apoptotic mediators Fas, FADD, and BAX, and (E) anti-apoptotic mediators cFLIP, Bcl2, and cIAP2. N=5–6 per group. Error bar indicates SEM. *P<0.05, **<0.01, ***P<0.001, ****<0.0001. Fig. 8 Pre treatment with AAV2.sFasL protects RGC cell death in microbead-induced mouse model of elevated IOP in C57/BL6 mice (A) Western blot from neural retina lysates (5ug/sample) showing overexpression of sFasL at 26 kD (Red band) and actin (Green band) in saline and microbead injected B6.AAV2.eGFP and B6.AAV2.sFasL mice at 4 weeks post microbead or saline injections. (B) IOP measurements were taken by rebound tonometry from B6.AAV2.eGFP and B6.AAV2.sFasL mice treated with saline or microbeads. Data is presented as mean IOP ± SEM, (N=10 per group). At 28 days post microbead injection the neural retina and optic nerve were processed for quantification of RGC and axon density. (C) Representative confocal microscopic images from retinal flat-mounts stained with βIII tubulin (red), an RGC-specific marker and DAPI a nuclear stain (blue) at 4 weeks post microbead or saline injections (Scale bar, 75μm). (D) Quantification of β-III tubulin positive RGCs, represented as RGC density/mm2 retina (E) Representative photomicrographs of optic nerve cross sections stained with PPD at 4 weeks post microbead or saline injections (Scale bar, 10μm). (F) Quantification of healthy axons, represented as axon density (104)/mm2. N=5 per group. *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001. Table I List of RNA primers used for Real time PCR Primers Sequence 1) Forward TNF-α 5′-GGG ACA GTG ACC TGG ACT GT-3′ 2) Reverse TNF-α 5′-CTC CCT TTG CAG AAC TCA GG-3′ 3) Forward β-actin 5′-TGT TAC CAA CTG GGA CGA CA-3′ 4) Reverse β-actin 5′-CTT TTC ACG GTT GGC CTT AG-3′ 5) Forward C-FLIP 5′-TTC TGA TAT AGG GTC CTG C-3′ 6) Reverse C-FLIP 5′-TCA CCA GAT CCA AGA AAC TC-3′ 7) Forward FADD 5′-CAA GCT GAG TGT AAC TGA AG-3′ 8) Reverse FADD 5′-TTA AAA GGC ATC AGC AAG AG-3′ 9) Forward GFAP 5′-GGC GCT CAA TGC TGG CTT CA-3′ 10) Reverse GFAP 5′-TCT GCC TCC AGC CTC AGG TT-3′ 11) Forward BAX 5′-AGG GTT TCA TCC AGG ATC GAG CAG-3′ 12) Reverse BAX 5′-ATC TTC TTC CAG ATG GTG AGC GAG-3′ 13) Forward BCL-2 5′TTG TGG CCT TCT TTG AGT TCG GTG-3′ 14) Reverse BCL-2 5′ GGT GCC GGT TCA GGT ACT CAG TCA-3′ 15) Forward TRADD GAA GTT CCC GGT TTC CTC TC 16) Reverse TRADD GAG GGC AGG ATC TCT CAG TG 17) Forward c-IAP-2 TGT CAG CCA AGT TCA AGC TG 18) Reverse c-IAP-2 ATC TTC CGA ACT TTC TCC AGG G 19) Forward Fas-L TGG GTA GAC AGC AGT GCC AC 20) Reverse Fas-L GCC CAC AAG ATG GAC AGG G 21) Forward Fas-R GTC CTG CCT CTG GTG CTT GCT G 22) Reverse Fas-R CAG GTT GGC ATG GTT GA 1 Casson RJ Chidlow G Wood JP Crowston JG Goldberg I 2012 Definition of glaucoma; clinical and experimental concepts Clin Exp Ophthalmol 40 4 341 349 22356435 2 Kass MA Heuer DK Higginbotham EJ Johnson CA Keltner JL Miller JP Parrish RK Wilson MR Gordon MO 2002 The Ocular Hypertension Treatment Study Arch Ophthalmol 120 701 713 12049574 3 Heijl A Leske C Bengtsson B Hyman L Bengtsson B Hussein M 2002 Reduction of intraocular pressure and glaucoma progression Arch Ophthalmol 120 1268 1279 12365904 4 Mudumbai RC 2013 Clinical update on normal tension glaucoma Sem Ophthalmol 28 3 173 179 5 Sommer A Tielsch JM Katz J Quigley HA Gottsch JD Javitt J Singh K 1999 Relationship between intraocular pressure and primary open angle glaucoma among white and black aemricans-The Baltimore Eye Survey Arch Ophthalmol 109 1090 1095 6 Song BJ Caprioli J 2014 New directions in the treatment of normal tension glaucoma Indian J Ophthalmol 62 5 529 537 24881596 7 Leibowitz HM Krueger DE Maunder LR Milton RC Kini MM Kahn HA Nickerson RJ pool J Colton TL Ganley JP Loewenstein JI Dawber TR 1980 The Framingham Eye Study monograph: an ophthalmological and epidemiological study of cataract, glaucoma, diabetic retinopathy, macular degeneration, and visual acuity in a general population of 2631 adults, 1973–1975 Surv Ophthalmol 24 suppl 335 610 7444756 8 Tezel G 2009 The Role of Glia, Mitochondria, and the Immune System in Glaucoma Invest Ophthalmol Vis Sci 50 1001 1012 19244206 9 Rieck J 2013 The Pathogenesis of Glaucoma in the Interplay with the Immune System Invest Ophthalmol Vis Sci 54 2393 2409 23539162 10 Yang X Luo C Cai J Powell DW Yu D Kuehn MH Tezel G 2011 Neurodegenerative and inflammatory pathway components linked to TNF-alpha/TNFR1 signaling in the glaucomatous human retina Invest Ophthalmol Vis Sci 52 8442 8454 21917936 11 Bosco A Romero CO Breen KT Chagovetz AA Steele MR Ambati BK Vetter ML 2015 Minocycline upregulates pro-survival genes and downregulates pro-apoptotic genes in experimental glaucoma Dis Mod Mech 8 5 443 455 12 Nakazawa T Nakazawa C Matsubara A Noda K Hisatomi T She H Michaud N Hafezi-Moghadam A Miller JW Benowitz LI 2006 Tumor necrosis factor-alpha mediates oligodendrocyte death and delayed retinal ganglion cell loss in a mouse model of glaucoma J Neurosci 26 49 12633 12641 17151265 13 Roh M Zhang Y Murakami Y Thanos A Lee SC Vavvas DG Benowitz LI Miller JW 2012 Etanercept, a widely used inhibitor of tumor necrosis factor-α (TNFα), prevents retinal ganglion cell loss in a rat model of glaucoma PLoS One 7 7 e40065 22802951 14 Yuan L Neufeld AH 2001 Activated microglia in the human glaucomatous optic nerve head J Neurosci Res 64 5 523 32 11391707 15 Howell GR Soto I Zhu X Ryan M Macalinao DG Sousa GL Caddle LB MacNicoll KH Barbay JM Porciatti V Anderson MG Smith RS Clark AF Libby RT John SWM 2012 Radiation treatment inhibits monocyte entry into the optic nerve head and prevents neuronal damage in a mouse model of glaucoma J Clin Invest 122 4 1246 1261 22426214 16 Suda T Takahashi T Golstein P Nagata S 1993 Molecular cloning and expression of the Fas ligand, a novel member of the tumor necrosis factor family Cell 75 6 1169 1178 7505205 17 Ju ST Panka DJ Cui H Ettinger R El-Khatib M Sherr DH Stanger BZ Marshak-Rothstein A 1995 Fas (CD95)/FasL interactions required for programmed cell death after T-cell activation Nature 373 444 448 7530337 18 Brunner T Mogli RJ LaFace D Yoo NJ Mahboubi A Echeverri F Martin SJ Force WR Lynch DH Ware CF Green DR 1995 Cell-autonomous Fas (CD95)/Fas-ligand interaction mediates activation-induced apoptosis in T-cell hybridomas Nature 373 441 444 7530336 19 Dhein J Walczak H Baumler C Debatin KM Krammer PH 1995 Autocrine T-cell suicide mediated by APO-1/(Fas/CD95) Nature 373 6513 438 441 7530335 20 Wajant H Pfizenmaier K Scheurich P 2003 Non-apoptotic Fas signaling Cytokine and Growth Factor Rev 14 53 66 12485619 21 Lambert C Landau AM Desbarats J 2003 Fas-beyond death: A regenerative role for Fas in the nervous system Apoptosis 8 551 562 14574061 22 Saas P Boucraut J Quiquerez AL Schnuriger V Perrin G Desplat-Jego S Bernard D Walker PR Dietrich PY 1999 CD95 (Fas/Apo-1) as a receptor governing astrocyte apoptotic or inflammatory responses: A key role in brain inflammation J Immunol 162 2326 2333 9973511 23 Choi C Benveniste EN 2004 Fas ligand/Fas system in the brain: regulator of immune and apoptotic responses Brain Res Rev 44 1 65 81 14739003 24 Tanaka M Itai T Adachi M Nagata S 1998 Downregulation of Fas ligand by shedding Nat Med 4 1 31 36 9427603 25 Hohlbaum AH Moe S Marshak-Rothstein A 2000 Opposing effects of transmembrane and soluble Fas ligand expression on inflammation and tumor cell survival J Exp Med 191 1209 1219 10748238 26 Suda T Hashimoto H Tanaka M Ochi T Nagata S 1997 Membrane Fas ligand kills human peripheral blood T lymphocytes, and soluble Fas ligand blocks the killing J Exp Med 186 2045 2050 9396774 27 Gregory MS Marshak-Rothstein A Hackett CG Abernathy EF Moody KS Hobson M Lee K Chen DF Ksander BR 2011 The form of Fas Ligand determines retinal ganglion cell survival during glaucoma PLOS ONE 6 3 e17659 21479271 28 Hohlbaum AM Gregory MS Ju ST Marshak-Rothstein A 2001 Fas ligand engagement of resident peritoneal macrophages in vivo induces apoptosis and the production of neutrophil chemotactic factors J Immunol 167 6217 6224 11714783 29 Griffith TS Brunner T Fletcher SM Green DR Ferguson TA 1995 Fas ligand-induced apoptosis as a mechanism of immune privilege Science 270 1189 1192 7502042 30 Ju KR Kim HS Kim JH Lee NY Park CK 2006 Retinal glial cell responses and Fas/FasL activation in rats with chronic ocular hypertension Brain Res 1122 209 221 17045251 31 Gregory MS Repp A Hohlbaum AM Marshak-Rothstein A Ksander BR 2002 Membrane Fas ligand activates innate immunity and terminates ocular immune privilege J Immunol 169 2727 2735 12193747 32 Gregory MS Saff RR Marshak-Rothstein A Ksander BR 2007 Control of ocular tumor growth and metastatic spread by soluble and membrane Fas ligand Cancer Res 67 11951 11958 18089826 33 Chen H Wei X Cho KS Chen GM Sappington R Calkins DJ Chen DF 2011 Optic neuropathy due to microbead-induced elevated intraocular pressure in the mouse Invest Ophthalmol Vis Sci 52 36 44 20702815 34 Dordea AC Bray MA Allen K Logan DJ Fei F Malhotra R Gregory MS Carpenter AE Buys ES 2016 An open-source computational tool to automatically quantify immunolabeled retinal ganglion cells Exp Eye Res 147 50 6 27119563 35 Karray S Kress C Cuvellier S Hue-Beauvais C Damotte D Babinet C Levi-Strauss M 2004 Complete loss of Fas ligand gene causes massive lymphoproliferation and early death, indicating a residual activity of gld allele J Immunol 172 4 2118 25 14764677 36 John SWM Smith RS Savinova OV Hawes NL Chang B Turnbull D Davisson M Roderick TH Heckenlively JR 1998 Essential iris atrophy, pigment dispersion, and glaucoma in DBA/2J mice Invest Ophthalmol Vis Sci 39 951 962 9579474 37 Libby RT Li Y Savinova OV Barter J Smith RS Nickells RW John SWM 2005 Susceptibility to neurodegeneration in glaucoma is modified by Bax gene dosage Plos Genetics 1 17 26 16103918 38 Howell GR Libby RT Marchant JK Wilson LA Cosma IM Smith RS Anderson MG John SW 2007 Absence of glaucoma in DBA/2J mice homozygous for wild-type versions of Gpnmb and Tyrp1 BMC Genet 8 45 17608931 39 Howell GR Libby RT Jakobs TC Smith RS Phalan FC Barter JW Barbay JM Marchant JK Mahesh N Porciatti V Whitmore AV Masland RH John SW 2007 Axons of retinal ganglion cells are insulted in the optic nerve early in DBA/2J glaucoma J cell Biol 179 1523 1537 18158332 40 Quigley HA Nickells RW Kerrigan LA Pease ME Thibault DJ Zack DJ 1995 Retinal ganglion cell death in experimental glaucoma and after axotomy occurs by apoptosis Invest Ophthalmol Vis Sci 36 5 774 786 7706025 41 Kerrigan LA Zack DJ Quigley HA Smith SD Pease ME 1997 TUNEL-positive ganglion cells in human primary open-angle glaucoma Arch Ophthamol 115 8 1031 1035 42 Levkovitch-Verbin H Dardik R Vander S Nisgav Y Kalev-Landoy M Melamed S 2006 Experimental glaucoma and optic nerve transection induce simultaneous upregulation of proapoptotic and prosurvival genes Invest Ophthalmol Vis Sci 47 6 2491 2497 16723461 43 Martinez-Lorenzo MJ Alava MA Gamen S Kim KJ Chuntherapai A 1998 Involvement of APO2 ligand/TRAIL in activation-induced death of Jurkat and human peripheral blood T cells Eur J Immunol 28 2714 2725 9754559 44 Powell WC Fingelton B Wilson CL Boothby M Matrisian LM 1999 The metalloproteinase matrilysin proteolytically generates active soluble Fas ligand and potentiates epithelial cell apoptosis Curr Biol 9 1441 1447 10607586 45 Bringmann A Pannicke T Grosche J Francke M Wiedemann P Skatchkov SN Osborne NN Reichenbach A 2006 Muller cells in the healthy and diseased retina Prog Retin Eye Res 25 397 424 16839797 46 Wang X Tay SS Ng YK 2000 An immunohistochemical study of neuronal and glial cell reactions in retinae of rats with experimental glaucoma Exp Brain Res 132 476 484 10912828 47 Tanihara H Hangai M Sawaguchi S Abe H Kageyama M Nakazawa F Shirasawa E Honda Y 1997 Up-regulation of glial fibrillary acidic protein in the retina of primate eyes with experimental glaucoma Arch Ophthalmol 115 752 756 9194727 48 Falvo JV Tsytsykova AV Goldfeld AE 2010 Transcriptional control for the TNF gene Curr Dir Autoimmun 11 27 60 20173386 49 Zhu W Downey JS Gu J Padova F Gram H Han J 2000 Regulation of TNF expression by multiple mitogen-activated protein kinase pathways J Immunol 64 12 6349 6358 50 Izquierdo E V Cuevas D Fernandez-Arroyo S Riera-Borrull M Orta-Zavalza E Joven J Rial E Corbi AL Escribese MM 2015 Reshaping of human macrophage polarization through modulation of glucose catabolic pathways J Immuunol 195 2442 2451 51 Tonari M Kurimoto T Horie T Sugiym T Ikeda T Oku H 2012 Blocking endothelin-B receptors rescues rential ganglion cells from optic nerve injury through suppression of neuroinflammation Invest Ophthalmol Vis Sci 53 3490 3500 22562513 52 Ahmad S Fatteh N El-Sherbiny N Naime M Ibrahim AS El-Sherbini AM El-Shafey SA Khan S Fulzele S Gonzales J Liou GI 2013 Potential role of A2A adenosine receptor in traumatic optic neuropathy J Neruoimmunol 264 54 64 53 Zhou X Li F Kong L Chodosh J Cao W 2009 Anti-inflammatory effect of pigment epithelium-derived factor in DBA.2J mice Mol Vis 15 438 450 19247457 54 Howell GR Macalinao DG Sousa GL Walden M Soto I Kneeland SC Barbay JM King BL Marchany JK Hibbs M Stevens B Barres BA Clark AF Libby RT John SWM 2011 Molecular clustering identifies complement and endothelin induction as early events in a mouse model of glaucoma J Clin Invest 121 1429 1444 21383504 55 Yuan L Neufeld AH 2000 Tumor necrosis factor-alpha: a potentially neurodestructive cytokine produced by glia in the human glaucomatous optic nerve head Glia 32 42 50 10975909 56 Tezel G Li LY Patil RV Wax MB 2001 TNF-alpha and TNF-alpha receptor-1 in the retina of normal and glaucomatous eyes Invest Ophthalmol Vis Sci 42 1787 1794 11431443 57 Zho Y Pernet V Hauswirth WW Di Polo A 2005 Activation of the extracellular signal-regulated kinase 1/2 pathway by AAV gene transfer protects retinal ganglion cells in glaucoma Mol Ther 12 402 412 15975850 58 Pease ME Zack DJ Berlinicke C Bloom K Cone F Wang Y Klein RL Hauswirth WW Quigley HA 2009 Effect of CNTF on retinal ganglion cell survival in experimental glaucoma Invest Ophthalmol Vis Sci 50 2194 200 19060281 59 Ju WK Kim KY Duong-Polk KX Lindsey JD Ellisman MH Weinreb RN 2010 Increased optic atrophy type 1 expression protects retinal ganglion cells in a mouse model of glaucoma Mol Vis 16 1331 1342 20664796 60 Almasieh M Wilson AM Morquette B Vargas JLC Di Polo A 2012 The molecular basis of retinal ganglion cell death in glaucoma Prog Ret Eye Res 31 152 181 61 Libby RT Li Y Savinova OV Barter J Smith RS Nickelles RW John SW 2005 Susceptibility to neurodegeneration in glaucoma is modified by Bax gene dosage PLoS Genet 1 1 17 26 16103918 62 Zhou X Li F Kong L Tomita H Li C Cao W 2005 Involvement of inflammation, degradation, and apoptosis in a mouse model of glaucoma J Biol Chem 280 35 31240 31248 15985430 63 Yang Z Quigley HA Pease ME Yang Y Qian J Valenta D Zack DJ 2007 Changes in Gene Expression in Experimental Glaucoma and Optic Nerve Transection: The Equilibrium between Protective and Detrimental Mechanisms Invest Ophthalmol Vis Sci 48 5539 5548 18055803 64 Ferguson TA Apte RS 2008 Angiogenesis in eye disease: immunity gained or immunity lost? Semin Immunopathol 30 111 119 18297288 65 Vernet-der Garabedian B Dere P Bailly Y Mariani J 2013 Innate immunity in the Grid2Lc/+ mouse model of cerebeller neurodegeneration: glial CD95/CD95L plays a non-apoptotic role in persistent neuron loss-associated inflammatory reactions in the cerebellum J Neuroinflamm 10 65 66 Sabelko KA Kelly KA Nahm MH Cross AH Russell JH 1997 Fas and Fas ligand enhance the pathogenesis of experimental allergic encephalomyelitis, but are not essential for immune privilege in the central nervous system J Immunol 159 3096 3099 9317103 67 Matsumoto H murakami Y Kataoka K Notomi S Mantopoulos D Trochons G Miller JW Gregory MS Ksander BR Marshak-Rothstein A Vavvas DG 2015 Membrane-bound and soluble Fas ligands have opposite functions in photoreceptor cell death following separation from the retinal pigment epithelium Cell Death Dis 6 e1986 26583327 68 Bossaller L Chiang PI Schmidt-Lauber C Ganesan S Kaiser WJ Rathinam VAJ Mocarski ES Subramanian D Green DR Silverman N Fitzgerald KA Marsha-Rothstein A Latz E 2012 Cutting Edge: FAS (CD95) mediates noncanonical IL-1b and IL-18 maturation via casapse-8 in an RIP3-independent manner J Immunol 189 5508 5512 23144495 69 Kim IB Kim KY Joo CK Lee MY Oh SJ Chung JW Chun MH 1998 Reaction of Muller cells after increased intraocular pressure in the rat retina Exp Brain Res 121 419 424 9746148 70 Yang X Hondur G Tezel G 2016 Antioxidant treatment limits neuroinflammation in experimental glaucoma Invest Ophthalmol Vis Sci 57 4 2344 2354 27127934 71 Batchelor JR Casey TA Gibbs DC 1976 HLA matching and corneal grafting Lancet 1 551 554 55836 72 Price FW Whitson WE Marks RG 1991 Progression of visual acuity after penetrating keratoplasty Ophthalmology 98 1177 1185 1923353 73 Sano Y Ksander BR Streilein JW 1995 Fate of orthotopic corneal allografts in eyes that cannot support anterior chamber-associated immune deviation induction Invest Ophthalmol Vis Sci 36 11 2176 2185 7558710 74 Stuart PM Griffith TS Usui N Pepose J Yu X Ferguson TA 1997 CD95 ligand (FasL)-induced apoptosis is necessary for corneal graft survival J Clin Invest 99 3 396 402 9022072 75 Sano Y Yamada J Ishino Y Adachi W Kawasaki S Suzuki T Kinoshita S Okuyama T Azuma N 2002 Non-cleavable mutant Fas ligand transfection of donor cornea abrogates ocular immune privilege Exp Eye Res 75 475 483 12387794 76 Stankowska DL Minton AZ Rutledge MA Muelle BH Phatak NR He S Ma HY Forster MJ Yorio T Krishnamoorthy RR 2015 Neuroprotective effects of transcription factor Brn3b in an ocular hypertension rat model of glaucoma Invest Ophthalmol Vis Sci 56 2 893 907 25587060 77 McKinnon SJ Lehman DM Tahzib NG Ransom NL Reitsamer HA Liston P LaCase E Li Q Korneluk RG Hauswirth WM 2002 Baculoviral IAP repeat-containing-4 protects optic nerve axons in a rat glaucoma model Mol Therapy 5 6 780 787 78 Tezel G Yang X Cai J 2005 Proteomic identification of oxidatively modified retinal proteins in a chronic pressure-induced rat model of glaucoma Invest Ophthalmol Vis Sci 46 3177 3187 16123417 79 Pache M Flammer J 2006 A sick eye in a sick body? Systemic findings in patients with primary open-angle glaucoma Survey Ophthalmol 51 3 179 212 80 Nucci C Tartaglione R Rombola L Morrone LA Fazzi E Bagetta G 2005 Neurochemical evidence to implicate elevated glutamate in the mechanisms of high IOP-induced retinal ganglion cell death in rat Neurotox 26 5 935 941 81 Ali RR Reichel MB De Alwis M Kanuga N Kinnon C Levinsky RJ Hunt DM Bhattachatya SS Thrasher AJ 1998 Adeno-associated virus gene transfer to mouse retina Human Gen Ther 9 1 81 86 82 Dudus L Anand V Acland GM Chen SJ Wilson JM Fisher KJ Maguire AM Bennet J 1999 Persistent transgene product in the retina, optic nerve, and brain after intraocular injection of rAAV Vis Res 39 150 2545 53 10396623 83 Sullivan TA Geisert EE Hines-Beard J Rex TS 2001 Systemic adeno-associated virus-mediated gene therapy preserves reitnal ganglion cells and visual function in DBA/2J glaucomatous mice Human Gen Ther 22 1191 1200 84 Bond WS Hines-Beard J Golden Merry YL Davis M Farooque A Sappington RM Calkins DJ Rex TS 2016 Virus-mediated EpoR76E therapy slows optic nerve axonopathy in experimental glaucoma Mol Ther 24 2 230 239 26502777

PMC005xxxxxx/PMC5136333.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 101479409 34798 Cancer Prev Res (Phila) Cancer Prev Res (Phila) Cancer prevention research (Philadelphia, Pa.) 1940-6207 1940-6215 27658890 5136333 10.1158/1940-6207.CAPR-15-0254 NIHMS826975 Article A Randomized Phase IIb Trial of myo-Inositol in Smokers with Bronchial Dysplasia Lam Stephen 1 Mandrekar Sumithra J. 2 Gesthalter Yaron 3 Allen Ziegler Katie L. 2 Seisler Drew K. 2 Midthun David E. 2 Mao Jenny T. 4 Aubry Marie Christine 2 McWilliams Annette 5 Sin Don D. 6 Shaipanich Tawimas 1 Liu Gang 3 Johnson Evan 3 Bild Andrea 7 Lenburg Marc E. 3 Ionescu Diana N. 1 Mayo John 8 Yi Joanne (Eunhee) 2 Tazelaar Henry 9 Harmsen William S. 2 Smith Judith 10 Spira Avrum E. 3 Beane Jennifer 3 Limburg Paul J. 2 Szabo Eva 10 for the Cancer Prevention Network. 1 British Columbia Cancer Agency, Vancouver, British Columbia, Canada 2 Mayo Clinic, Rochester, MN 3 Boston University Medical Center, Boston, MA 4 New Mexico Veteran’s Health Care System, Albuquerque, NM 5 Fiona Stanley Hospital, Palmyra DC, Western Australia 6 St. Paul’s Hospital, Vancouver, British Columbia, Canada 7 University of Utah, Salt Lake City, Utah 8 Vancouver General Hospital, Vancouver, British Columbia, 9 Mayo Clinic, Scottsdale, AZ 10 Division of Cancer Prevention, National Cancer Institute, National Institutes of Health Address for Correspondence: Stephen Lam, M.D., British Columbia Cancer Agency, 10-111 675 West 10th Avenue, Vancouver, British Columbia, V5Z 4E6, Canada; Telephone: 604-675-8094; slam2@bccancer.bc.ca 9 11 2016 22 9 2016 12 2016 01 6 2017 9 12 906914 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. Previous preclinical studies and a phase I clinical trial suggested myo-inositol may be a safe and effective lung cancer chemopreventive agent. We conducted a randomized, double blind, placebo-controlled, phase IIb study to determine the chemopreventive effects of myo-inositol in smokers with bronchial dysplasia. Smokers with ≥ 1 site of dysplasia identified by autofluorescence bronchoscopy-directed biopsy were randomly assigned to receive oral placebo or myo-inositol, 9 g once/day for two weeks, and then twice/day for 6 months. The primary endpoint was change in dysplasia rate after six months of intervention on a per participant basis. Other trial endpoints reported herein include Ki-67 labeling index, blood and bronchoalveolar lavage fluid (BAL) levels of pro-inflammatory, oxidant/anti-oxidant biomarkers, and an airway epithelial gene-expression signature for phosphatidylinositol 3-kinase (PI3K) activity. Seventy four (n=38 myo-inositol, n=36 placebo) participants with a baseline and 6-month bronchoscopy were included in all efficacy analyses. The complete response and the progressive disease rates were 26.3% versus 13.9% and 47.4% versus 33.3%, respectively, in the myo-inositol and placebo arms (p=0.76). Compared with placebo, myo-inositol intervention significantly reduced IL-6 levels in BAL over 6 months (p=0.03). Among those with a complete response in the myo-inositol arm, there was a significant decrease in a gene-expression signature reflective of PI3K activation within the cytologically-normal bronchial airway epithelium (p=0.002). The heterogeneous response to myo-inositol suggests a targeted therapy approach based on molecular alterations is needed in future clinical trials to determine the efficacy of myo-inositol as a chemopreventive agent. myo-Inositol bronchial dysplasia chemoprevention INTRODUCTION Lung cancer is the most common cause of cancer death worldwide, with an estimated 1.8 million new cases and 1.6 million deaths in 2012(1) and causes more deaths in the United States than colorectal cancer, breast cancer, and prostate cancer combined.(2) Former heavy smokers retain an elevated risk for lung cancer even years after they stop smoking.(3, 4) Therefore, a strategy to prevent lung cancer in addition to smoking cessation is needed. Chemoprevention involves the use of dietary or pharmaceutical interventions to slow or reverse the progression of premalignancy to invasive cancer.(5, 6) In addition to efficacy, safety is a critical consideration since the intervention is given to individuals who are at risk for cancer, but otherwise are in apparent good health. myo-Inositol is found in a wide variety of foods such as whole grains, seeds, and fruits. It is a source of several second messengers including diacylglycerol and is required by human cells for growth and survival in culture. Pre-clinical studies show myo-inositol inhibits carcinogenesis by 40% to 50% in both the induction and post-initiation phases.(7, 8) Mechanistically, the phosphatidylinositol 3-kinase (PI3K) pathway that is activated in the bronchial epithelial cells of smokers with dysplasia was found to be inhibited by myo-inositol and associated with regression of bronchial dysplasia.(9) The low toxicity and promising pre-clinical and phase I clinical trial data led to the current randomized, double-blind, placebo-controlled, phase IIb clinical trial in smokers with bronchial dysplasia, who are at high risk for lung cancer.(10) MATERIALS AND METHODS Clinical Trial Protocol The Review of Ethics Boards of the BCCA and the University of British Columbia, and the Mayo Clinic and University of New Mexico Institutional Review Boards approved this study. Written informed consent was obtained from all participants. The clinical trial registration number was NCT00783705. Study participant recruitment and eligibility From November 3, 2008 to August 9, 2013, current and former smokers between the ages of 45 to 79 years from the Greater Vancouver area who had a ≥ 30 pack-year smoking history were recruited through the community outreach networks, television programs, radio broadcasts, and local newspapers. In March of 2010 recruitment was extended to participants in Rochester, Minnesota and Albuquerque, New Mexico. A former smoker was defined as a person who had not smoked for at least one year, verified by urinary cotinine below 100 ng/ml. Eligibility criteria for randomization to study drug included ≥ 1 site of histologically-confirmed bronchial dysplasia on baseline bronchoscopy, no evidence of lung cancer (stage 0/I curatively treated non-small cell lung cancer with all therapy completed ≥6 months prior to randomization allowed), and normal organ and marrow function. Bronchoscopic Procedures Prior to July 8, 2009, 41 participants were screened for bronchoscopy using C-reactive protein (CRP) level in plasma. Those with CRP≥1.25 mg/L were offered autofluorescence bronchoscopy to localize areas of dysplasia using the Onco-LIFE device (Novadaq Technologies Corp., Richmond, BC, Canada) as described previously.(11, 12) The use of CRP was based on our preliminary study that the prevalence of dysplasia was higher among those with elevated CRP. Biopsy samples were taken from areas that were abnormal under white light and/or autofluorescence examination and at least one control biopsy was obtained from bronchial mucosa with normal fluorescence in an upper or lower lobe of the lung. After July 8, 2009, the CRP eligibility criterion was removed as the prevalence of dysplasia in the high versus low CRP group was not sufficiently higher to justify its use. An additional 407 participants who met the age and smoking criteria were offered a bronchoscopy with biopsy of all abnormal sites under white-light or autofluorescence examination and 6 pre-determined sites in the main carina, both upper and lower lobes and the right middle lobe before and after treatment. The median number of biopsy samples obtained per participant was 6 (range=1–14) in the myo-inositol group and 7 (range=1–14) in the placebo group. Bronchoalveolar lavage (BAL) from the right upper lobe or the left upper lobe was performed using 20 mL aliquots of normal saline as described previously.(13) BAL cells were separated from the fluid by centrifugation at 1500 rpm for 5 minutes and fluids were stored at −80°C until cytokine evaluation. Bronchial brushing was performed in two separate subsegmental bronchi that had not been lavaged or biopsied using a 1.7 mm diameter bronchial cytology brush (Hobbs Medical, Stafford Springs, CT). The brushes were retrieved and immediately immersed in RNALater and kept frozen at −80° C until RNA extraction and gene expression profiling using RNA-Seq. The biopsy samples were fixed in buffered formalin, embedded in paraffin, and serial sections were obtained. Sections 1, 6, and 13 were stained with hematoxylin-eosin and used to determine histopathologic classification. They were systematically reviewed by two pulmonary pathologists (MA, DI) who were blinded to intervention assignments. All biopsy samples were classified according to World Health Organization (WHO) criteria (normal, basal cell hyperplasia, metaplasia, mild/moderate/severe dysplasia, or carcinoma in-situ).(14) One grade difference in sample classification between the two pathologists was resolved by a third pathologist (JY) to reach a final diagnosis. Spiral Chest Computed Tomography (CT) CT scans were performed at BCCA as previously described using a 16 detector CT scanner at 120 kVp, 0.5 second rotation time, pitch 1.25 and 40 mA. Images were reconstructed at 1 mm slice width at 1 mm spacing.(15, 16) The CT scans were performed before and at the end of the 6 month intervention. The site, size and appearance of nodules ≥ 1mm in diameter were recorded. All scans were reviewed by an experienced chest radiologist (JM) without knowledge of the intervention assignment. Randomization Participants were randomly assigned to receive either myo-inositol (Tsuno Food Industries Co., Ltd., Wakayama, Japan) at a dose of 9 grams orally (with water or juice) once/day for 2 weeks and then twice/day for 6 months, or placebo. The placebo powder sachet was visually identical to the active compound sachet. A dynamic allocation procedure was used to balance marginal distributions of the specified stratification factors: smoking status (current versus former), prior stage 0/I lung cancer (yes versus no) and number of dysplastic lesions at baseline (1 versus >1). All study personnel were blinded to the study codes, as was confirmed by independent review. Follow Up The participants were interviewed by telephone at week 2, months 1, 2, 4, 5, and 7 to 8 and were seen in person at months 3 and 6 for monitoring of compliance and drug-related adverse events. Compliance was determined from an agent diary and by unused sachet counts at each follow-up visit. Compliance was defined as ingestion of ≥ 80% of the planned doses. Smoking status was confirmed by measuring urinary cotinine at months 3 and 6. Toxicity was monitored according to the NCI Common Terminology Criteria for Adverse Events (CTCAE version 3.0). Fasting blood glucose was measured at baseline and at months 1, 3, and 6. No dose modification was made for grade 1 toxicity. For grade 2 toxicity or intolerable grade 1 adverse events (AEs) that were possibly, probably, or definitely related to study agent, study agent was stopped for up to 2 weeks. If the AE resolved, the study agent was resumed at once/day dosing for one week and then increased to full dose if there was no AE recurrence. If the AE recurred, the participant was taken off the study permanently. For grade 3 or 4 AEs that were deemed possibly, probably, or definitely related to study agent, participants were taken off study permanently. When grade 3 or 4 adverse events occurred that were judged unlikely related to the study agent, intervention was stopped for up to 2 weeks and then resumed at the same dose as prior to the AE after the AE resolved. If the AE recurred, the participant was taken off study permanently. Participants underwent a second autofluorescence bronchoscopy with BAL and bronchial brushings after 6 months on study agent, and biopsies were obtained from the same sites that were biopsied at baseline as well as any new areas that displayed abnormal fluorescence. The bronchoscopist was blinded to the intervention assignment. Biomarker Analysis Ki-67 Expression In Bronchial Biopsies Unstained bronchial biopsy specimens (5-micron) were mounted on silanized glass slides (HistoBond, Marienfeld-Germany) and Ki-67 expression was determined by the method of Shi et al. (17) The primary Ki-67 antibody (1:250, Lab Vision, Fremont, CA.), and the biotinylated secondary antibody (1:500, Vector Lab) were used, followed by ABC method (Vectastain ABC Elite Kit, Vector Lab, Burlingame, CA) with diaminobenzidine (Sigma, St. Louis, MO) used as chromogen. For a negative control, the primary antibody was omitted. The percentage of positively stained cells was determined by counting a total of 100 cells in the most positively stained area in the tissue section. BAL and Plasma Biomarkers The potential anti-oxidative and anti-inflammatory effects of myo-inositol were determined in distal BAL and plasma collected before and after intervention. The biomarkers interrogated using ELISA were: 1) pro-inflammatory proteins: C-reactive protein, interleukin-6, and CCL-2 (all from R&D Systems, Minneapolis, MN); 2) oxidant/antioxidants: myeloperoxidase (MPO, R&D Systems Minneapolis, MN), nitrotyrosine (Hycult Biotech, Plymouth Meeting, PA) and glutathione (Millipore-Calbiochem, San Diego, CA); and 3) pneumoproteins: Clara cell protein-16 (CC-16, Biovendor, Asheville, NC), surfactant protein-D (SFTPD, R&D Systems, Minneapolis, MN) and CCL18 (R&D Systems, Minneapolis, MN). Laboratory personnel were blinded to the intervention assignment. All measurements were performed in duplicate. Assessing PI3K Activity Based on Airway Gene Expression Matched pre- and post-treatment bronchial brushings of cytologically normal epithelium were collected during bronchoscopy from 72 subjects (n=144 samples). Total RNA was extracted using the miRNeasy Mini Kit (Qiagen). Sequencing libraries were prepared using the Illumina® TruSeq® RNA Kit v2 and multiplexed in groups of six using the Illumina® TruSeq® Paired-End Cluster Kit. Each sample was sequenced on an Illumina® HiSeq® 2500 to generate paired-end 100 nucleotide reads. Demultiplexing and creation of FASTQ files were performed using Illumina CASAVA. Alignment and gene-level counts were generated using RSEM (v1.2.10) (18) and hg19 and Ensembl v74 annotation. Six samples were removed from downstream analyses based on data quality as assessed using RSeQC (v2.3.3) (19) and single nucleotide variant calls to verify that paired samples were derived from the same subject. Genes were filtered out based on a modified version of the mixture model in the SCAN.UPC package (20); a gene was included in downstream analyses if the mixture model classified it as “ “signal” in at least 15% of the samples. All subsequent analyses were conducted using R (v3.0.0). Linear modeling to identify treatment effects within complete responders (n=10 subjects in the myo-inositol arm or n=5 subjects in the placebo arm) and progressors (n=15 subjects in the myo-inositol arm or n=11 subjects in the placebo arm) was performed using the edgeR (v3.4.2) (21) and limma (v23.18.13) packages (22) using normalized voom-transformed data (23). To identify relationships between the effects of myo-inositol treatment and PI3K-pathway activation, the moderated t-statistics from the linear modeling were used to create rankings of genes for each group. GSEA (Gene Set Enrichment Analysis) (24) was used determine if genes that we previously reported to be either increased or decreased with PIK3CA overexpression in vitro (9) are significantly enriched at the extremes of the ranked lists. The RNA-Seq data from this study is available from the NCBI Gene Expression Omnibus (GEO) under accession [GEO ID pending]. Outcomes The primary endpoint of the study was defined as change in dysplasia rate on a per-participant basis, with the per-lesion analysis specified as a secondary endpoint. Secondary endpoints reported herein include change in Ki-67 labeling index in bronchial biopsies, changes in biomarker levels in BAL and plasma, and effect on airway epithelial gene expression signature for PI3K activity. We also determined whether baseline biomarker measurements were related to progression/regression of dysplasia at 6 months. Safety and AE profiles of participants enrolled in both intervention arms were also closely monitored. Statistical Design and Analyses The sample size for this trial was calculated as follows. The placebo group was expected to have a 20% complete response rate based on three previous NCI sponsored chemoprevention trials with similar but not identical eligibility criteria.(12, 25, 26) In our previously reported pilot study in 20 subjects, the complete response rate was 67%.(10) Assuming at least a 30% difference in the dysplasia response rates between the myo-inositol and placebo arms (20%–50%), a sample size of 50 evaluable participants per intervention arm provided 80% power (2-sided chi-square test with continuity correction; alpha=0.05). If the difference in the response rates between the treatment groups was at least 35%, the study had 90% power to detect a significant difference. Assuming a 10% drop out rate, we planned to enroll a total of 110 participants to have 100 evaluable participants. Descriptive statistics were used to summarize participant characteristics and pathologic evaluations of the bronchial biopsy examinations. Comparison between groups was done with the Wilcoxon rank sum test for continuous variables. Pearson’s chi-square test with continuity correction or Fisher’s exact test, as appropriate for small expected cell sizes, was used to compare categorical variables. Response rates were calculated on a per-site and a per-participant basis. For the lesion-specific analysis, complete response (CR) was defined as the regression of a dysplastic lesion of any grade to one classified as being either hyperplastic or normal. Progressive disease (PD) was defined as appearance of lesions that were classified as mild dysplasia or worse, irrespective of whether the site was biopsied at baseline, or worsening of the dysplastic lesion present at baseline by two or more grades (e.g., mild dysplasia to severe dysplasia or worse). Post-intervention biopsies of metaplasia or lower that were either metaplastic or lower at baseline, or not biopsied at baseline were graded as not applicable (NA). Dysplastic lesions that were not classified as complete response, progressive disease, or not applicable were referred to as stable disease (SD). For the participant specific analysis, CR was defined as regression of all dysplastic lesions found at baseline to lesions that were no worse than hyperplasia, as defined by the site analysis at 6 months and the appearance of no new dysplastic lesions that were graded as mild dysplasia or worse. PD was defined as progression of one or more sites by 2 more grades as defined for the lesion-specific analysis above, or the appearance of new dysplastic lesions that were mild dysplasia or worse at 6 months. Partial response (PR) was defined as regression of some but not all of the dysplastic lesions with the appearance of no new lesions that were graded as mild dysplasia or worse. Stable disease (SD) referred to participants who did not have a CR, PR, or PD. For participant-specific analysis, PR and SD were combined into one response category because minor changes like one grade change are prone to grading error. We also categorized participants as regression or stable (CR, PR, or SD) vs. progressive disease (PD) when doing the lesion specific analysis. In the comparison of treatment arms for the participant-specific assessment of response (CR vs. SD/PR vs. PD) from baseline to 6 month, a Kruskal Wallis test was used, accounting for the ordered nature of response. A multiple variable logistic regression model was used to assess the odds of progression (as opposed to CR, SD, or PR) with the variables of treatment arm, gender, smoking status (current vs. former) and maximal histologic grade in the participant (mild vs. moderate vs. severe). Generalized estimating equations (GEE) were used for the lesion specific analyses due to the varying number of lesions per participant. The association of baseline histologic grade of a lesion and baseline Ki-67 staining was also assessed using a linear regression GEE model. The percent change in Ki-67 labeling index from bronchial biopsies with dysplasia at baseline to 6 months post-intervention was compared between intervention arms using a GEE model. Biomarker expression levels (at baseline and change from baseline) were compared between and within arms using a Wilcoxon rank sum or sign rank test. The trends in response rates by baseline marker levels (divided into quartiles) were tested using a Cochran Armitage test for trend (1-sided test). Logistic regression models were used to evaluate the impact of baseline biomarker levels (divided into quartiles) on participant level response (CR versus others; CR/PR versus others) unadjusted and adjusted for intervention arm. All P values are two-sided, unless otherwise noted. A two-sided P value less than .05 was considered statistically significant. No adjustments for multiple comparisons were done for the secondary endpoints as this was largely an exploratory exercise. SAS version 9.3 (SAS, Inc., Cary, NC) was used for all analyses. RESULTS After 448 participants were screened and 85 randomized to receive placebo (41) or myo-inositol (44), the trial was closed due to slow accrual. All 85 participants were included in the baseline and AE analyses. Eleven of the 85 participants did not have a follow-up bronchoscopy due to AEs, loss to follow-up, or refusal. Therefore, 74 participants (38 myo-inositol, 36 placebo) were included in the efficacy analyses (Figure 1). Clinical Characteristics Baseline characteristics of the 85 participants are shown in Table 1. There was no statistically significant difference in median age, race, body mass index, gender, self-reported smoking status, prior NSAID use, number of biopsies obtained, and number of dysplastic lesions or severity of dysplasia at baseline. Based on the baseline cotinine measurement, 6 participants in the myo-inositol group and 1 participant in the placebo group with a self-reported former smoking status were reclassified as current smokers. There was no significant difference in study arms based on this reclassification (p=0.36). Effects of myo-inositol on Histopathology of Bronchial Biopsies Participant-specific Analysis There was no statistically significant difference in response, categorized as complete response (CR) vs. partial or stable response (PR/SD) vs. progressive disease (PD), between intervention arms, p=0.76 (Figure 2). The CR rate was 26.3% in the myo-inositol group and 13.9% in the placebo group and the PD rates were 47.4% and 33.3%, respectively. The response rates in the 23 participants with a maximum histopathology grade of mild dysplasia at baseline were also not significantly different (p=0.34) with CR rates of 46.2% versus 9.1% and PD rates of 38.5% versus 36.7% in the myo-inositol and placebo arms, respectively (Figure 2). In a similar assessment of 50 participants with a maximum histopathology grade of moderate/severe dysplasia at baseline, there was no significant difference in the response rates between the intervention arms (p=0.27; Figure 2). There was no significant difference in the efficacy of myo-inositol by smoking status (Table 2). A multiple variable logistic regression model adjusting for gender, smoking status and maximum histopathology grade at baseline showed no significant difference in the odds of progression between the intervention arms. Lesion-specific Analysis In the per-lesion analyses, a total of 267 lesions in 38 participants assigned to the myo-inositol arm were biopsied at baseline (31.8% dysplastic) and 265 lesions were biopsied post-intervention (20.4% dysplastic). In the placebo arm, a total of 258 lesions in 36 participants were biopsied at baseline (34.5% dysplastic) and 243 lesions were biopsied post-intervention (23.9% dysplastic). The per-lesion response rates in the myo-inositol arm were 10.2% with CR, 15.9% with SD, and12.5% with PD, and for the placebo arm the corresponding rates were 7.4%, 22.6%, and 10.3%. CT Detected Lung Nodules Sixty-two BCCA participants had CT prior to intervention. Of the 27 participants with no lung nodules, only 4 underwent a repeat CT scan at Month 6. Among the 39 participants with pre- and post-treatment CT data, 18 participants with 53 lesions in the myo-inositol group and 17 participants with 49 lesions in the placebo group had one or more non-calcified lung nodules at baseline, with mean (SD) sizes of 5.9 (3.8) mm and 4.1 (2.2) mm, respectively. A GEE analysis of % change in the CT nodule size from baseline to 6 months showed no difference between the intervention arms (p=0.6). Based on a nodule specific GEE analysis, there was also no significant difference between the arms (p=0.91) when looking at the nodules categorized as progressed (myo-inositol=12.5%; placebo=3.5%) versus regressed/stable (myo-inositol=87.5%; placebo=96.5%), with progression defined as any increase from baseline or appearance of new nodules and regression/stable defined as decrease from baseline and no new nodules. Biomarker Analyses Ki-67 in Bronchial Biopsies Ki-67 labeling index data were available from baseline bronchial dysplasia and site-matched, post-intervention biopsy samples from 65 participants (n=33 and n=32 in the myo-inositol and placebo arms, respectively). The mean percent change in Ki-67 labeling index in the bronchial biopsies with dysplasia from baseline to 6 months in the myo-inositol arm was −22.8% compared to −6.2% in the placebo arm, which was not significantly different between the intervention arms (p=0.34). BAL and Plasma Biomarkers Compared with placebo, treatment with myo-inositol significantly reduced IL-6 levels in BAL over 6 months (p=0.03) and produced borderline significant effects on BAL glutathione and myeloperoxidase (p=0.06 for both) (Table 3). There were no significant effects of myo-inositol on any of the plasma biomarkers (data not shown). To determine whether any of the biomarkers predicted CR or CR/PR, we performed a series of exploratory analyses using baseline biomarker levels (in quartiles) in BAL and plasma. Irrespective of treatment status, increased baseline levels of CC16 in both BAL and plasma were associated with CR (Supplementary Table 1) and CR/PR (Supplementary Table 2). Reduced plasma levels of MPO were also significantly associated with CR/PR but not with CR. PI3K-Associated Gene Expression Within the Cytologically Normal Bronchial Epithelium In order to test the previously developed hypothesis that PI3K-associated gene expression within the cytologically normal epithelium decreases with clinical response to myo-inositol (9), we evaluated PI3K-activity-related gene expression pre- vs. post-therapy amongst patients receiving myo-inositol who achieved a complete response. Overall, there was a significant inverse relationship between PI3K activity and response to myo-inositol, which was expected given our previous results which show myo-inositol inhibition of PI3K signaling (9). Specifically, we found that in participants with a complete response to myo-inositol (n=10), those genes that increase upon treatment are inversely enriched in decreased genes from our in vitro signature for PI3K activity (p = 0.002); this results suggests a decrease in PI3K activity in this group. This decrease in PI3K activity was not found among the complete responders in the placebo arm, nor among the subjects with progressive disease in either treatment arm. Agent Compliance Participants in the myo-inositol and placebo groups took 67.9% ± 3.8% and 79.8% ± 3.1% of the prescribed doses, respectively. There was no significant difference in response rates (CR vs. SD/PR vs. PD) between the intervention arms within compliant and non-compliant participants (Table 4). Adverse Events All adverse events (AEs) regardless of attribution were collected. At least one AE, regardless of grade or attribution, was reported by 39/44 (89%) participants in the myo-inositol arm and 34/41 (83%) participants in the placebo arm (p=0.45 for comparison between arms). Most AEs were classified as grade 1 (77.3%), with progressively fewer grade 2 (19.6%), and grade 3 (2.7%) adverse events reported. No grade 4 AEs were reported. Seven (8%) participants (n=4, myo-inositol and n=3, placebo) experienced a total of 9 grade 3 adverse events (dyspnea, dizziness, pain, arthralgia, and bilateral cataracts in the myo-inositol arm; syncope, rash, and peripheral neuropathy in the placebo arm) all of which were felt to be unlikely to be related to the study agents, except for syncope, which was possibly related. Participants in the myo-inositol arm reported a higher incidence of gastrointestinal AEs compared to the placebo arm (59% myo-inositol arm; 43% placebo arm; p=0.16). One Serious Adverse Event was reported: a grade 2 coronary artery calcification in the myo-inositol arm, deemed unrelated to study intervention. DISCUSSION This is the first phase IIb chemoprevention trial to examine the safety and efficacy of myo-inositol for lung cancer chemoprevention. Following on promising preclinical data and our previous small phase IIa trial showing a high bronchial dysplasia reversion rate, this trial was designed to evaluate the efficacy of a longer (6 month) myo-inositol intervention (10). Although safety and tolerability were established, there was no overall statistically significant effect on bronchial dysplasia, albeit in a study that only achieved three-quarters of its planned accrual. The goal of phase II chemoprevention trials is to identify agents that have clinically meaningful effects. Our sample size calculation was based on a ≥30% better complete response (CR) rate in the myo-inositol group versus placebo. A post-hoc sample size calculation showed that with 35 participants per group, a difference of 35% (20% versus 55%) in the dysplasia response rates, with a power of 80% and a 2-sided error rate of 0.05, would have been detectable. Thus, even with the smaller than anticipated accrual, this study had sufficient power to detect a meaningful treatment effect. myo-Inositol would likely not be considered for a phase III trial based on the 12% improvement in CR rates over placebo observed in this trial. It is possible, and perhaps even likely, that decreased progression to higher grades of dysplasia would be more predictive of cancer prevention than dysplasia regression. In colon cancer prevention, the most effective trial model has examined progression to new polyps rather than regression of sporadic polyps (27, 28). In bronchial dysplasia, such studies would be significantly larger than our current trial and thus even more difficult to perform. Nevertheless, our data suggest potential differential effects in subpopulations within the studied cohort. A numerically higher, but statistically not significant percentage of CR was observed in the myo-inositol arm as compared to placebo (26.3% versus 13.9%); this was balanced by a non-statistically significant increase in the percentage of PD with myo-inositol (47.4% versus 33.3%). Similarly, a higher but statistically not significant mean percent change in Ki-67 expression level in the bronchial biopsies with dysplasia from baseline to 6 months was observed in the myo-inositol arm compared to the placebo arm (−22.8% compared to −6.2%, p=0.34). Furthermore, treatment with myo-inositol significantly reduced levels of IL-6, a pro-inflammatory biomarker in BAL (p=0.0317). Taken together, this consistent modulation of multiple markers suggests that there could be a subpopulation within the entire cohort that may experience benefit from myo-inositol. There is a precedent for differential responses to chemopreventive interventions within participant subgroups from the previously reported phase IIb iloprost trial, which showed a significant regression of dysplasia in former, but not current, smokers (29). In our study, there was no significant interaction of smoking status and treatment arm using participant reported or urinary cotinine verified smoking status. The effect of short-term changes in the smoking status on bronchial dysplasia in 30% of the participants could not be evaluated in a phase II study of this size. Thus we were unable to identify a myo-inositol-sensitive subgroup, if one exists. The PI3K/Akt signaling pathway regulates diverse cellular function including proliferation and survival. We have previously demonstrated significantly increased pAkt in dysplastic lesions versus hyperplastic/metaplastic lesions before myo-inositol treatment (30). Following myo-inositol treatment, significant decrease in pAkt was observed in dysplastic (P < 0.01) but not hyperplastic/metaplastic lesions (P > 0.05). In vitro, myo-inositol decreased endogenous and tobacco carcinogen-induced activation of Akt in immortalized human bronchial epithelial cells, which decreased cell proliferation and induced a G(1)-S cell cycle arrest.(30) These results show that the phenotypic progression of premalignant bronchial lesions from smokers correlates with increased activation of Akt and that it is a target of myo-inositol. Similarly, examination of gene expression in cytologically normal bronchial epithelial cells from participants in the same phase IIa study showed that myo-inositol was associated with reduction in PI3k activity among those smokers who had regression of their premalignant lesions (9). This finding is confirmed in the current study. Among those smokers who had a complete response to myo-inositol, the gene-expression signature of PI3K activity was reduced in bronchial epithelial cells post-treatment, whereas there was no significant reduction in this signature among the clinical non-responders to myo-inositol nor in the placebo arm regardless of clinical response. While the sample size was limited and varied between response subgroups, these results suggest that PI3k associated gene expression could serve as an intermediate biomarker of therapeutic efficacy for myo-inositol. Whether selection of subjects with activation of the PI3K pathway would lead to a higher complete response rate requires further investigation. The advantage of using bronchial dysplasia for phase II chemoprevention trials is that these lesions can be localized and biopsied using white light and autofluorescence bronchoscopy for histopathology confirmation. The presence of dysplasia is a known risk marker for lung cancer both in the central airways and the peripheral lung.(31–33) However, there has been a change in the lung cancer cell type distribution worldwide. The prevalence of centrally located squamous cell carcinomas has been steadily decreasing and replaced by an increase in adenocarcinomas,(34) which are usually located in the peripheral lung beyond the range of sampling by standard flexible bronchoscopes. This is reflected in a steady decline in proportion of smokers found to have bronchial dysplasia in the last decade.(35) This contributed to the reported difficulty in identifying participants for the current clinical trial. Alternative intermediate endpoint biomarkers, such as CT detected non-calcified lung nodules are needed for future phase II lung cancer chemoprevention trials.(36) CT scan was done in this study to rule out lung cancer prior to starting treatment with myo-inositol or placebo. However, the study population was selected for the presence of central lung bronchial dysplasia rather than presence of peripheral lung nodules, and thus was not sufficiently powered to robustly examine the effect of myo-inositol on CT-detected lung nodules. In summary, despite a reduction in pro-inflammatory cytokine and oxidant level in BAL, the slight but statistically insignificant increase in CR and reduction in Ki-67 labeling index in bronchial biopsies after treatment with myo-inositol compared to placebo was accompanied by an increase in PD rate of similar magnitude. This suggests that a targeted approach based on an understanding of the underlying molecular alterations (9) is probably needed to determine if myo-inositol has a role as a chemopreventive agent in a more defined cohort. Supplementary Material 1 2 Research Support: Supported by a contract from the National Cancer Institute (N01CN35000). The authors gratefully acknowledge the following people for their technical assistance in participant recruitment, bronchoscopy, sample processing, data management, and immunohistochemistray assays: Sharon Gee, Jennifer Kidd, Myles McKinnon, Bimmie Kalan, Lori Bergstrom, Meghan Muse, Hanqiao Liu, Huiqing Si, Cindy Fitting, Mary Fredericksen, Cindy Beinhorn, Barbara Greguson, Christopher Zima, and Dr. Steven Kye. We thank Fumi Tsuno, Tsuno Food Industries Co., Ltd., Wakayama, Japan for supplying the myo-inositol to NCI DCP. The authors also gratefully acknowledge Colleen Garvey, Sharon Kaufman, and Karrie Fursa for their assistance with study design, administration, and manuscript preparation. Figure 1 CONSORT Diagram Figure 2 Change in Bronchial Dysplasia, by Intervention Arm Numbers shown represent the proportion of participants with each bronchial dysplasia change state. No participants with a partial response were observed among those with Mild Dysplasia at baseline in the myo-inositol arm. Table 1 Baseline Characteristics for Randomized Participants Receiving Study Intervention myo-Inositol (N=44) Placebo (N=41) Total (N=85) p value Age, years 0.521   Median 58.5 58.0 58.0   Range (45.0–75.0) (46.0–79.0) (45.0–79.0) Race 0.592   White 42 (95.5%) 38 (92.7%) 80 (94.1%)   Asian 2 (4.8%) 3 (7.3%) 5 (5.9%) Body mass index, kg/m2 0.621   Median 27.2 26.0 26.5   Range (21.1–36.3) (21.0–35.2) (21.0–36.3) Gender, N (%) 0.352   Female 10 (22.7%) 13 (31.7%) 23 (27.1%)   Male 34 (77.3%) 28 (68.3%) 62 (72.9%) Smoking status, N (%) 0.852   Current 27 (61.4%) 26 (63.4%) 53 (62.4%)   Former 17 (38.6%) 15 (36.6%) 32 (37.6%) Prior NSAID use 0.492   No 28 (63.6%) 29 (70.7%) 57 (67.1%)   Yes 16 (36.4%) 12 (29.3%) 28 (32.9%) Dysplastic lesions identified 0.742   1 Dysplastic lesion 22 (50.0%) 19 (46.3%) 41 (48.2%)   >1 Dysplastic lesions 22 (50.0%) 22 (53.7%) 44 (51.8%) Mucosal biopsies obtained 0.621   Median 6.0 7.0 7.0   Range (1.0–14.0) (1.0–14.0) (1.0–14.0) Most advanced histology, N (%) 0.112   Mild dysplasia 15 (34.1%) 13 (31.7%) 28 (32.9%)   Moderate dysplasia 28 (63.6%) 22 (53.7%) 50 (58.8%)   Severe dysplasia 1 (2.3%) 6 (14.6%) 7 (8.2%) 1 Kruskal Wallis 2 Chi-Square No registered participants reported prior lung cancer No registered participants reported Hispanic or Latino ethnicity Table 2 Response to myo-Inositol by Smoking Status Showing No Difference Between myo-Inositol and Placebo Smoking Status Participant- level Response myo-Inositol (N=38) Placebo (N=36) P-value† Current smokers N=23 N=23 0.1571 PD 13 (56.5%) 11 (47.8%) SD or PR 8 (34.8%) 11 (47.8%) CR 2 (8.7%) 1 (4.3%) Former smokers N=15 N=13 0.881 PD 5 (33.3%) 1 (7.7%) SD or PR 2 (13.3%) 8 (61.5%) CR 8 (53.3%) 4 (30.8%) 1 Kruskal Wallis test, considering responses as ordered; progressive (PD), then stable/partial response (SD/PR), then complete response (CR). Table 3 Change in Biomarker Levels in Bronchoalveolar Lavage Fluid from baseline (pre) to 6 Months Post-Intervention Difference in post – pre levels; Median (range) myo-Inositol Placebo P* IL-6 (pg/ml) −0.68 (−2.78 to 0.34) −0.27 (−1.54 to 1.89) 0.03 GSH (umol/L) −0.246 (−1.98 to 0.48) −0.558 (−1.13 to 0.33) 0.06 MPO (ng/mL) −3.457 (−8.17 to 2.77) −1.154 (−8.43 to 1.21) 0.06 CC16 (ng/mL) −66.962 (−127.56 to 52.26) −54.323 (−144.65 to 140.84) 0.10 SFTPD (ng/ml) −12.387 (−47.12 to 5.72) 7.213 (−7.62 to 23.79) 0.22 CCL18 (ng/ml) −121.470 (−1234.7 to 1122.9) 10.702 (−684.95 to 683.25) 0.63 CCL-2 (pg/ml) −9.2465 (−44.32 to 8.62) 9.413 (−52.58 to 54.6) 0.58 * Wilcoxon Rank-Sum P-Value (for difference in absolute change by arm) Abbreviations: CC-16, Clara cell protein-16; CRP, c-reactive protein; GSH, total glutathione; IL-6, interleukin-6; MPO, myeloperoxidase, SFTPD, surfactant protein D CCL18, chemokine ligand 18 previously known as pulmonary and activation-regulated chemokine ; CCL-2, chemokine ligand 2 previously known as monocyte chemotactic protein 1 Table 4 Participant-Specific Analysis of Response by Intervention Arm Based on Agent Compliance Compliance Status Participant- level Response myo-Inositol (N=38) Placebo (N=35) P-value† Compliant * N=25 N=28 PD 12 (48.0%) 9 (32.1%) 0.73 SD or PR 7 (28.0%) 16 (57.2%) CR 6 (24.0%) 3 (10.7%) Non-Compliant * N=13 N=7 PD 6 (46.2%) 3 (37.5%) 0.91 SD or PR 3 (23.1%) 3 (37.5%) CR 4 (30.8%) 2 (25.0%) † Kruskal Wallis test. * defined as the number of sachets taken by the participant divided by the number of sachets that should have been taken by the participant based on a 6-month intervention. Clinical Trial Registration: www.clinicaltrials.gov; NCT00783705 The authors have no conflict of interest REFERENCES 1 Stewart BaW CP The World Cancer Report 2014: International Agency for Research on Cancer 2014 2 Siegel R Ma J Zou Z Jemal A Cancer statistics, 2014 CA Cancer J Clin 2014 64 9 29 24399786 3 Tong L Spitz MR Fueger JJ Amos CA Lung carcinoma in former smokers Cancer 1996 78 1004 1010 8780538 4 Peto R Darby S Deo H Silcocks P Whitley E Doll R Smoking, smoking cessation, and lung cancer in the UK since 1950: combination of national statistics with two case-control studies Bmj 2000 321 323 329 10926586 5 Sporn MB Newton DL Chemoprevention of cancer with retinoids Fed Proc 1979 38 2528 2534 488376 6 Szabo E Mao JT Lam S Reid ME Keith RL Chemoprevention of lung cancer: Diagnosis and management of lung cancer, 3rd ed: American College of Chest Physicians evidence-based clinical practice guidelines Chest 2013 143 e40S e60S 23649449 7 Estensen RD Wattenberg LW Studies of chemopreventive effects of myo-inositol on benzo[a]pyrene-induced neoplasia of the lung and forestomach of female A/J mice Carcinogenesis 1993 14 1975 1977 8403228 8 Wattenberg LW Estensen RD Chemopreventive effects of myo-inositol and dexamethasone on benzo[a]pyrene and 4-(methylnitrosoamino)-1-(3-pyridyl)-1-butanone-induced pulmonary carcinogenesis in female A/J mice Cancer Res 1996 56 5132 5135 8912846 9 Gustafson AM Soldi R Anderlind C Scholand MB Qian J Zhang X Airway PI3K pathway activation is an early and reversible event in lung cancer development Sci Transl Med 2010 2 26ra5 10 Lam S McWilliams A LeRiche J MacAulay C Wattenberg L Szabo E A phase I study of myo-inositol for lung cancer chemoprevention Cancer Epidemiol Biomarkers Prev 2006 15 1526 1531 16896044 11 Lam S Kennedy T Unger M Miller YE Gelmont D Rusch V Localization of bronchial intraepithelial neoplastic lesions by fluorescence bronchoscopy Chest 1998 113 696 702 9515845 12 Lam S leRiche JC McWilliams A Macaulay C Dyachkova Y Szabo E A randomized phase IIb trial of pulmicort turbuhaler (budesonide) in people with dysplasia of the bronchial epithelium Clin Cancer Res 2004 10 6502 6511 15475437 13 Lam S Leriche JC Kijek K Phillips D Effect of bronchial lavage volume on cellular and protein recovery Chest 1985 88 856 859 4064774 14 Travis W Histological typing of lung and pleural tumors 1999 Berlin-Heidelberg Springer-Verlag 15 McWilliams A Mayo J MacDonald S leRiche JC Palcic B Szabo E Lung cancer screening: a different paradigm Am J Respir Crit Care Med 2003 168 1167 1173 12882756 16 McWilliams AM Mayo JR Ahn MI MacDonald SL Lam SC Lung cancer screening using multi-slice thin-section computed tomography and autofluorescence bronchoscopy J Thorac Oncol 2006 1 61 68 17409828 17 Shi SR Key ME Kalra KL Antigen retrieval in formalin-fixed, paraffin-embedded tissues: an enhancement method for immunohistochemical staining based on microwave oven heating of tissue sections J Histochem Cytochem 1991 39 741 748 1709656 18 Li B Dewey CN RSEM: accurate transcript quantification from RNA-Seq data with or without a reference genome BMC Bioinformatics 2011 8 4 12 323 21816040 19 Wang L Wang S Li W RSeQC:quality control of RNA-seq experiments Bioinformatics 2012 8 15 28 16 2184 2185 Epub 2012 Jun 27 22743226 20 Kusko RL Brothers Ii JF Tedrow J Pandit K Huleihel L Integrated Genomics Reveals Convergent Transcriptomic Networks Underlying COPD and IPF Am J Respir Crit Care Med 2016 4 22 [Epub ahead of print] 21 Robinson MD McCarthy DJ Smyth GK edgeR: a Bioconductor package for differential expression analysis of digital gene expression data Bioinformatics 2010 1 1 26 1 139 140 Epub 2009 Nov 11 19910308 22 Ritchie ME Phipson B Wu D Hu Y Law CW Shi W Smyth GK limma powers differential expression analyses for RNA-sequencing and microarray studies Nucleic Acids Res 2015 4 20 43 7 e47 Epub 2015 Jan 20 25605792 23 Law CW Chen Y Shi W Smyth GK voom: Precision weights unlock linear model analysis tools for RNA-seq read counts Genome Biol 2014 2 3 15 2 R29 24485249 24 Subramanian A Tamayo P Mootha VK Mukherjee S Ebert BL Gillette MA Paulovich A Pomeroy SL Golub TR Lander ES Mesirov JP Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles Proc Natl Acad Sci U S A 2005 10 25 102 43 15545 15550 Epub 2005 Sep 30 16199517 25 Lam S MacAulay C Le Riche JC Dyachkova Y Coldman A Guillaud M A randomized phase IIb trial of anethole dithiolethione in smokers with bronchial dysplasia J Natl Cancer Inst 2002 94 1001 1009 12096085 26 Lam S Xu X Parker-Klein H Le Riche JC Macaulay C Guillaud M Surrogate end-point biomarker analysis in a retinol chemoprevention trial in current and former smokers with bronchial dysplasia Int J Oncol 2003 23 1607 1613 14612933 27 Baron JA Cole BF Sandler RS Haile RW Ahnen D Bresalier R A randomized trial of aspirin to prevent colorectal adenomas N Engl J Med 2003 348 891 899 12621133 28 Sandler RS Halabi S Baron JA Budinger S Paskett E Keresztes R A randomized trial of aspirin to prevent colorectal adenomas in patients with previous colorectal cancer N Engl J Med 2003 348 883 890 12621132 29 Keith RL Blatchford PJ Kittelson J Minna JD Kelly K Massion PP Oral iloprost improves endobronchial dysplasia in former smokers Cancer Prev Res (Phila) 2011 4 793 802 21636546 30 Han W Gills JJ Memmott RM Lam S Dennis PA The chemopreventive agent myoinositol inhibits Akt and extracellular signal-regulated kinase in bronchial lesions from heavy smokers Cancer Prev Res (Phila) 2009 4 2 4 370 376 19336734 31 Ishizumi T McWilliams A MacAulay C Gazdar A Lam S Natural history of bronchial preinvasive lesions Cancer Metastasis Rev 2010 29 5 14 20112052 32 Jeremy George P Banerjee AK Read CA O'Sullivan C Falzon M Pezzella F Surveillance for the detection of early lung cancer in patients with bronchial dysplasia Thorax 2007 62 43 50 16825337 33 Lam S Szabo E Preinvasive Endobronchial Lesions: Lung Cancer Precursors and Risk Markers? Am J Respir Crit Care Med 2015 12 15 192 12 1411 1413 26669471 34 Lortet-Tieulent J Soerjomataram I Ferlay J Rutherford M Weiderpass E Bray F International trends in lung cancer incidence by histological subtype: adenocarcinoma stabilizing in men but still increasing in women Lung Cancer 2014 84 13 22 24524818 35 Jen RLS Lam S Roth JA Cox JD Hong WK Detection and Treatment of Preneoplastic Lesions Lung Cancer 2014 Fourth John Wiley & Sons, Inc. Lung Cancer, 4rd Edition. 36 Pinsky PF Nath PH Gierada DS Sonavane S Szabo E Short- and long-term lung cancer risk associated with noncalcified nodules observed on low-dose CT Cancer Prev Res (Phila) 2014 7 1179 1185 24755313

PMC005xxxxxx/PMC5136335.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 101150042 30118 Mol Cancer Res Mol. Cancer Res. Molecular cancer research : MCR 1541-7786 1557-3125 27658423 5136335 10.1158/1541-7786.MCR-16-0233 NIHMS827055 Article MUC1-C Represses the Crumbs Complex Polarity Factor CRB3 and Downregulates the Hippo Pathway Alam Maroof 1 Bouillez Audrey 1 Tagde Ashujit 1 Ahmad Rehan 1+ Rajabi Hasan 1 Maeda Takahiro 1 Hiraki Masayuki 1 Suzuki Yozo 1* Kufe Donald 1 1 Dana-Farber Cancer Institute, Harvard Medical School, Boston, MA + Present address: College of Medicine, King Saud University, Riyadh, Saudi Arabia * Present address: Department of Gastroenterological Surgery, Osaka Police Hospital, Kitayama-Cho 10-31 Tennoji, Osaka City, Osaka 543-0035, Japan 4 11 2016 22 9 2016 12 2016 01 12 2017 14 12 12661276 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. Apical-basal polarity and epithelial integrity are maintained in part by the Crumbs (CRB) complex. The C-terminal subunit of MUC1 (MUC1-C) is a transmembrane protein that is expressed at the apical border of normal epithelial cells and aberrantly at high levels over the entire surface of their transformed counterparts. However, it is not known if MUC1-C contributes to this loss of polarity that is characteristic of carcinoma cells. Here it is demonstrated that MUC1-C downregulates expression of the Crumbs complex CRB3 protein in triple-negative breast cancer (TNBC) cells. MUC1-C associates with ZEB1 on the CRB3 promoter and represses CRB3 transcription. Notably, CRB3 activates the core kinase cassette of the Hippo pathway, which includes LATS1 and LATS2. In this context, targeting MUC1-C was associated with increased phosphorylation of LATS1, consistent with activation of the Hippo pathway, which is critical for regulating cell contact, tissue repair, proliferation and apoptosis. Also shown is that MUC1-C-mediated suppression of CRB3 and the Hippo pathway is associated with dephosphorylation and activation of the oncogenic YAP protein. In turn, MUC1-C interacts with YAP, promotes formation of YAP/β-catenin complexes and induces the WNT target gene MYC. These data support a previously unrecognized model in which targeting MUC1-C in TNBC cells (i) induces CRB3 expression, (ii) activates the CRB3-driven Hippo pathway, (iii) inactivates YAP, and thereby (iv) suppresses YAP/β-catenin-mediated induction of MYC expression. Implications These findings demonstrate a previously unrecognized role for the MUC1-C oncoprotein in the regulation of polarity and the Hippo pathway in breast cancer. MUC1 CRB3 Hippo YAP triple-negative breast cancer MYC Introduction Epithelia are comprised of a laterally connected layer of cells with apical-basal polarity. The epithelial stress response is associated with loss of apical-basal cell polarity and disruption of cell-cell adhesion (1, 2). Loss of polarity thereby constitutes an early step in the epithelial to mesenchymal transition (EMT), whereby epithelial cells acquire invasive and migratory properties. Apical-basal polarity is maintained by: (i) the Scribble complex (Scrib, Dlg, Lgl) responsible for establishing the basolateral membrane domain, (ii) the PAR complex (Cdc42, PAR3/ASIP, PAR6, atypical protein kinase C) at the apical-lateral junctions between cells, and (iii) the Crumbs complex (Crb, PALS, PATJ, Lin7), which defines the apical membrane (2). Cell polarity is thus maintained as a result of mutual interactions among the PAR, CRB and SCRIB complexes (3). Three orthologs of the Drosophila Crb protein are expressed in human tissues, named CRB1, CRB2, and CRB3, of which CRB3 is predominantly found in epithelial cells and suppresses epithelial tumor progression (2). Additionally, CRB3 functions as a tumor suppressor by activating the Hippo signaling network (mammalian orthologs SAV1, MST1/2 and LATS1/2) (4, 5). The Hippo pathway has been linked to cancer stem-like cells, invasion, DNA repair and therapeutic resistance (4, 6–8). Moreover, the Hippo pathway intersects with the canonical WNT/β-catenin pathway (4, 9). The major downstream effectors of the Hippo pathway are the transcriptional co-activator with PDZ-binding motif (TAZ) and the Yes-associated protein (YAP) (10). TAZ activates Dishevelled and represses β-catenin (9); whereas, YAP binds directly to β-catenin (11). CRB3-mediated activation of the Hippo kinase cascade results in the downstream phosphorylation and inactivation of YAP (10, 12). Phosphorylated YAP is retained in the cytoplasm and thereby inhibits the WNT/β-catenin pathway by binding to β-catenin and restricting its translocation to the nucleus (10, 12, 13). However, in response to CRB3 suppression, the Hippo pathway is inactivated, leading to dephosphorylation and activation of YAP (4). In turn, localization of the YAP/β-catenin complex in the nucleus promotes the induction of certain WNT target genes (10, 11, 14, 15). Mucin 1 (MUC1) is a transmembrane protein that is widely overexpressed in breast and other carcinomas (16). MUC1 undergoes an autocleavage process, resulting in an extracellular N-terminal subunit (MUC1-N) and a transmembrane C-terminal subunit (MUC1-C), which form complexes at the apical membranes of normal epithelial cells (16, 17). With loss of cell polarity, the oncogenic MUC1-C subunit is expressed over the entire cell membrane, and interacts with receptor tyrosine kinases (RTKs), such as EGFR, which are typically positioned at the basal-lateral borders (16, 17). As a consequence of this association with RTKs, MUC1-C upregulates RTK signaling by promoting activation of the downstream PI3K→AKT and MEK→ERK pathways (16–21). In addition, MUC1-C is imported into the nucleus, where it interacts with transcription factors, such as NF-κB p65 and ZEB1, among others (16, 17, 22, 23). The MUC1-C cytoplasmic domain binds directly to NF-κB p65 and increases occupancy of NF-κB on the promoters of its target genes (22). In this context, MUC1-C/NF-κB p65 complexes activate the ZEB1 promoter and increase ZEB1 expression (23). Further, MUC1-C associates with ZEB1 to repress the miR-200c gene, which encodes a tumor suppressor that reverses EMT (23). MUC1-C also binds directly to β-catenin, stabilizes β-catenin/TCF4 complexes and thereby promotes activation of WNT target genes, such as CCDN1 and MYC (24–28). To our knowledge, there is no available evidence supporting involvement of MUC1-C in the regulation of apical-basal polarity. The present results demonstrate that MUC1-C represses the CRB3, HUGL2 and PATJ polarity factors in TNBC cells, indicating that MUC1-C is of importance to the loss of cell polarity. Based on the role of CRB3 in activating the Hippo pathway, the present work has focused on the downstream effects of MUC1-C-mediated CRB3 suppression. We show that MUC1-C represses CRB3 transcription and downregulates the Hippo pathway. In addition, we show that MUC1-C activates YAP and forms a complex with YAP/β-catenin that activates the MYC promoter. Our findings thus demonstrate that targeting MUC1-C activates the CRB3→Hippo tumor suppressor cascade. Materials and Methods Cell culture Human MDA-MB-231, BT-20 and MCF-7 breast cancer cells were cultured in DMEM (Dulbecco’s Modified Eagle’s Medium) growth medium containing 10% heat-inactivated fetal bovine serum, 100 units/ml penicillin, 100 μg/ml streptomycin and 2 mM L-glutamine. Human BT-549 breast cancer cells were grown in RPMI1640 medium with heat-inactivated fetal bovine serum, antibiotics, L-glutamine and 10 μg/ml insulin. Authentication of the cells was confirmed by short tandem repeat (STR) analysis. MDA-MB-231 and BT-549 cells were infected with lentiviral vectors that express a MUC1shRNA (MISSION shRNA TRCN0000122938; Sigma, St Louis, MO) or scrambled control shRNA (CshRNA; Sigma) (29). BT-20 and MCF-7 cells were transfected with a pHR-CMV vector expressing MUC1-C or with an empty vector. Cells were also infected with lentiviral vectors expressing a SNAIL1 shRNA (MISSION shRNA TRCN0000063822; Sigma), a ZEB1 shRNA (MISSION shRNA TRCN0000017565; Sigma) or a scrambled control shRNA vector (CshRNA; Sigma). Cells were also infected with lentivirus vectors expressing a tetracycline-inducible MUC1shRNA (tet-MUC1shRNA), as described (30). MUC1shRNA (MISSION shRNA TRCN0000122938; Sigma) or a control scrambled CshRNA (Sigma) was inserted into the pLKO-tet-puro vector (Addgene, Plasmid #21915). The viral vectors were produced in HEK293T cells as previously described (30, 31). Cells expressing tet-MUC1shRNA or tet-CshRNA were selected for growth in 1–3 μg/ml puromycin. Cells were (i) treated with doxycycline (DOX; Sigma), and (ii) transfected with a CRB3 siRNA (sc43698) or a control siRNA (sc37007) (Santa Cruz Biotechnology). Immunoprecipitation and immunoblot analysis Whole cell lysates were prepared using NP-40 lysis buffer containing protease inhibitor cocktail (Thermo Scientific). Nuclear and cytosolic lysates were prepared using the NucBuster nuclear protein extraction kit (Millipore). Soluble proteins were immunoprecipitated with anti-MUC1-C (NeoMarker) or a control IgG. Immunoprecipiates and lysates not subjected to precipitation were analyzed by immunoblotting with anti-MUC1-C (NeoMarker), anti-CRB3 (Abcam), anti-HUGL2 (Genetex), anti-PATJ, anti-SNAIL1 (Santa Cruz Biotechnology), anti-CDC42, anti-ZEB1, anti-phospho-LATS1, anti-LATS1, anti-phospho-YAP, anti-YAP, anti-HDAC1 (Cell Signaling Technology) and anti-β-actin (Sigma). Immunoreactive complexes were detected using horseradish peroxidase-conjugated secondary antibodies (GE Healthcare) and an enhanced chemiluminescence (ECL) detection system (Perkin Elmer Health Sciences). Quantitative real time, reverse transcriptase PCR qRT-PCR analysis was performed on cDNA synthesized from total RNA using the Superscript III cDNA synthesis system (Life Technologies). cDNA samples were then amplified using the SYBR green qPCR assay kit (Applied Biosystems) and the ABI Prism 7300 Sequence Detector (Applied Biosystems)(32). qPCR primers used for detection of CRB3, HUGL2, PATJ, CDC42, CTGF, CYR61, MYC, MUC1 and GAPDH are listed in Supplemental Table S1. Statistical significance was determined by the Student’s t-test. Analysis of CRB3 promoter activity Cells were cultured in six-well plates followed by transfection with an empty vector, pCRB3-Luc and, as an internal control, SV-40-Renilla-Luc (Promega) in the presence of Superfect transfection reagent (Qiagen). After 48 h, the transfected cells were lysed with passive lysis buffer and the lysates were analyzed using the Dual Luciferase Assay system (Promega). Chromatin immunoprecipitation (ChIP) assays Soluble chromatin was prepared from 2–3 × 106 cells as described (27) and precipitated with anti-ZEB1, anti-β-catenin (Cell Signaling Technology), or a control nonimmune IgG (Santa Cruz Biotechnlogy). For re-ChIP assays, ZEB1 or β-catenin complexes from the primary ChIP were released and reimmunoprecipitated with anti-MUC1-C (NeoMarker) or anti-YAP (Cell Signaling Technology). The SYBR green qPCR kit (Applied Biosystems) was used for the qPCR analyses with the ABI Prism 7300 Sequence Detector (Applied Biosystems) as described (30, 33). Primer pairs used for the CRB3 and MYC promoters and a control GAPDH region are listed in Supplemental Table S2. Relative fold enrichment was calculated as described (34). Protein binding assays GST-tagged YAP, GST-YAP(1–160) and GST-YAP(161–504) were generated by PCR amplification of the GST-YAP plasmid (Addgene) and subcloning into the pGEX-5X-1 expression vector (GE Healthcare). GST-β-catenin was expressed from pGEX-β-catenin (Addgene) and cleaved with thrombin to isolate purified β-catenin. MUC1-CD, MUC1-CD(1–45) and MUC1-CD(46–72) peptide fragments were prepared by expressing GST-fusion proteins and cleavage of the GST tag with thrombin as described (23). GST and GST fusion proteins bound to glutathione beads were incubated with purified proteins. The adsorbates were analyzed by immunoblotting with anti-MUC1-C cytoplasmic domain antibodies CD1 (35) and CT2 (NeoMarker). Results MUC1-C downregulates CRB3 expression BT-549 and MDA-MB-231 Basal B triple-negative breast cancer (TNBC) cells have low to undetectable levels of CRB3, HUGL2 and PATJ expression as compared to that found for CDC42 (Fig. 1A and Fig. 1B). Notably, however, stable silencing of MUC1-C in BT-549 cells was associated with upregulation of CRB3, HUGL2 and PATJ, but not CDC42, mRNA (Fig. 1A, left) and protein (Fig. 1A, right). We also found that silencing MUC1-C in MDA-MB-231 cells results in induction of CRB3, HUGL2 and PATJ expression (Fig. 1B, left and right), indicating that MUC1-C downregulates multiple effectors of cell polarity. The following studies have focused on CRB3, which in addition to its role in establishing the apical membrane, activates the Hippo pathway (2). Interestingly and in contrast to the results obtained with Basal B TNBC cells, we found that Basal A BT-20 TNBC cells express CRB3 (Fig. 1C, left and right). Moreover, we found that ectopic expression of MUC1-C in BT-20 cells suppresses CRB3 levels (Fig. 1C, left and right). Constitutive expression of CRB3 in epithelial MCF-7 cells was also downregulated by a MUC1-C-mediated mechanism (Supplemental Fig. S1A, left and right). To further assess the regulation of CRB3 in a setting of transiently targeting MUC1-C, we established BT-549 cells transduced to express a tetracycline-inducible MUC1 shRNA (tet-MUC1shRNA). Treatment of BT-549/tet-MUC1shRNA cells with doxycycline (DOX) for 7 days resulted in suppression of MUC1-C and induction of CRB3 expression (Supplemental Fig. S1B, left and right). In contrast, silencing CRB3 had little if any effect on MUC1-C expression (Supplemental Fig. S1D, left and right). The MUC1-C cytoplasmic domain includes a CQC motif that is necessary for the formation of MUC1-C homodimers and their nuclear localization (Fig. 1D) (36–38). Accordingly, we treated BT-549 cells with GO-203, a cell penetrating peptide that targets this CQC motif or, as a control, an inactive peptide CP-2 (18) (Fig. 1D). We found that treatment with GO-203, but not CP-2, results in upregulation of CRB3 mRNA and protein (Fig. 1E, left and right). A similar response was observed when MDA-MB-231 cells were treated with GO-203 (Fig. 1F, left and right). As further evidence, we found that GO-203 blocks MUC1-C-induced CRB3 downregulation in MCF-7/MUC1-C cells (Supplemental Fig. S1C, left and right). These findings support the notion that MUC1-C promotes loss of cell polarity, at least in part, by downregulating CRB3 in a cell context-dependent manner. CRB3 is repressed by a ZEB1-dependent mechanism MUC1-C drives ZEB1 expression in BT-549, MDA-MB-231 and MCF-7/MUC1-C cells (23). In turn, MUC1-C binds to ZEB1 and contributes to repression of the ZEB1 target gene, miR-200c (23). ZEB1 also represses transcription of the CRB3 gene (39). In this context and to investigate if MUC1-C suppresses CRB3 by a ZEB1-mediated mechanism, we stably silenced ZEB1 in BT-549 cells and found upregulation of CRB3 mRNA and protein (Fig 2A, left and right). Similar results were obtained with ZEB1 silencing in MDA-MB-231 cells (Fig. 2B, left and right). Moreover, we found that MUC1-C-induced downregulation of CRB3 in MCF-7/MUC1-C cells is ZEB1-dependent (Fig. 2C, left and right). SNAIL1, another transcriptional repressor, binds to the same target CRB3 promoter region as ZEB1 (39). Silencing MUC1-C in MDA-MB-231 cells was associated with a partial decrease in SNAIL1 expression (Fig. 2D, left); however, silencing SNAIL1 had no apparent effect on CRB3 expression (Fig. 2D, right). Additionally, SNAIL1 was upregulated in MCF-7/MUC1-C cells (Fig. 2E, left) and silencing SNAIL1 had no detectable effect on CRB3 expression in these cells (Fig. 2E, right). These findings supported the premise that MUC1-C suppresses CRB3 by a ZEB1-dependent, SNAIL1-independent mechanism. MUC1-C interacts with ZEB1 on the CRB3 gene promoter ZEB1 represses the CRB3 promoter by binding to five E-Box elements (5′-CACCTG-3′) located within 1410 bp upstream to the transcription start site (39)(Fig. 3A). To assess involvement of MUC1-C in suppression of the CRB3 promoter, we transfected MDA-MB-231 cells with a CRB3 promoter-luciferase reporter (pCRB3-Luc) containing the proximal 5 E-boxes (Fig. 3A). Silencing MUC1-C was associated with an increase in pCRB3-Luc activity as compared to that found in MDA-MB-231/CshRNA cells (Fig. 3B, left). Similar results were obtained in studies of BT-549 cells (Fig 3B, middle). In addition, overexpression of MUC1-C in MCF-7/MUC1-C cells decreased pCRB3-Luc activity (Fig. 3B, right), indicating that MUC1-C contributes to repression of the CRB3 promoter. To assess occupancy on the CRB3 promoter, we performed ChIP and re-ChIP studies on chromatin from MDA-MB-231 cells. The results demonstrated that ZEB1 occupies the CRB3 promoter (Fig. 3C, left) in a complex with MUC1-C (Fig. 3C, right). Moreover, targeting MUC1-C with GO-203 was associated with a decrease in occupancy of ZEB1 (Fig. 3D, left) and ZEB1/MUC1-C complexes (Fig. 3D, right) on the CRB3 promoter. Consistent with these results, we found that MUC1-C increases occupancy of ZEB1 (Fig. 3E, left) and ZEB1/MUC1-C complexes (Fig. 3E, right) on the CRB3 promoter in MCF-7/MUC1-C cells. These findings and those with the pCRB3-Luc reporter support a model in which MUC1-C promotes ZEB1 occupancy on the CRB3 promoter and thereby inactivates CRB3 transcription. MUC1-C-induced downregulation of CRB3 results in inactivation of the Hippo pathway CRB3 activates the core kinase cassette of the Hippo pathway, which includes MST1/2 and LATS1/2 (5). Activated p-LATS1/2 phosphorylates YAP, which in turn leads to retention of p-YAP in the cytoplasm and its degradation (10, 40). Consistent with activation of the Hippo pathway, silencing MUC1-C in MDA-MB-231 and BT-549 cells was associated with increased p-LATS1 levels (Fig. 4A, left and middle). By contrast, p-LATS1 levels were decreased with overexpression of MUC1-C in MCF-7/MUC1-C cells (Fig. 4A, right). Increases in p-LATS1 were in turn associated with upregulation of p-YAP in the cytosolic fraction of MDA-MB-231/MUC1shRNA (Fig. 4B, left) and BT-549/MUC1shRNA (Fig. 4B, middle) cells. Moreover and in concert with decreases in p-LATS1 in MCF-7/MUC1-C cells, p-YAP was downregulated in the cytosolic fraction (Fig. 4B, right). A concomitant decrease of the transcriptionally active form of non-phosphorylated YAP was observed in the nuclear fraction of MDA-MB-231/MUC1shRNA, as compared with that in MDA-MB-231/CshRNA, cells (Fig. 4C, left). Similar results were obtained in the response to MUC1-C silencing in BT-549 cells (Fig. 4C, middle). In addition, increased levels of nuclear YAP were detectable in MCF/MUC1-C cells (Fig. 4C, right). Activated YAP forms a complex with β-catenin that localizes to the nucleus (10, 11, 14, 15). Moreover, MUC1-C stabilizes β-catenin and promotes activation of WNT/β-catenin/TCF4 target genes (25–27). In concert with these pathways, nuclear β-catenin levels were decreased in MDA-MB-231 and BT-549 cells silenced for MUC1-C as compared to that in control cells (Fig. 4D, left and middle). In contrast, nuclear β-catenin was increased in MCF-7/MUC1-C cells (Fig. 4D, right), supporting a model in which both MUC1-C and YAP may drive nuclear β-catenin signaling. MUC1-C binds directly to YAP and β-catenin MUC1-C, like YAP, binds directly to β-catenin (24, 25). We therefore asked if MUC1-C forms a complex with YAP and β-catenin. Coimmunoprecipitation studies using lysates from MDA-MB-231 cells showed that MUC1-C associates with YAP (Fig. 5A, left). Similar results were obtained with BT-549 and MCF-7/MUC1-C cells (Fig. 5A, middle and right). To further define the interaction between MUC1-C and YAP, we incubated GST-YAP with purified MUC1-C cytoplasmic domain (MUC1-CD) and found direct binding (Fig. 5B). We also found that GST-YAP(1–160), but not GST-YAP(161–504), binds to MUC1-CD (Fig 5C). In addition, GST-YAP binds to purified MUC1-CD(1–45) and not MUC1-CD(46–72) (Fig 5D). Based on these findings and our previous work that MUC1-CD binds to β-catenin (24, 25), we incubated GST-YAP with purified MUC1-CD and β-catenin. Interestingly, the results demonstrate that MUC1-CD promotes the formation of trimolecular complexes containing both YAP and β-catenin (Fig. 5E). MUC1-C associates with YAP and β-catenin on the MYC promoter YAP activates Hippo target genes by forming transcriptional complexes with TEAD family members and with β-catenin/TCF4 (4). Based on the findings that MUC1-C activates YAP, we first asked if MUC1-C signaling is linked to YAP/TEAD-driven genes, such as CTGF and CYR61. Indeed, we found that targeting MUC1-C suppresses CTGF and CYR61 expression (Fig. 6A), indicating that MUC1-C promotes activation of this arm of YAP-activated transcription. The interaction between YAP and β-catenin also promotes activation of their respective target genes (11, 14, 15, 41). Recent work has demonstrated that MUC1-C drives MYC expression by forming complexes with β-catenin on the MYC promoter and increasing β-catenin occupancy (27, 28). Other studies have supported a role for YAP/β-catenin in activating MYC expression (42). These and the above findings that MUC1-C forms a complex with YAP and β-catenin invoked the possibility that MUC1-C could associate with YAP and β-catenin on the MYC promoter. To investigate this notion, we performed ChIP and re-ChIP studies of the MYC promoter in MDA-MB-231 cells. We found that β-catenin occupies the MYC promoter (Fig. 6B) and that β-catenin associates with both MUC1-C and YAP (Fig. 6B). To determine whether the formation of these complexes is MUC1-C-dependent, we targeted MUC1-C with GO-203 treatment. The results showed that targeting MUC1-C is associated with decreased β-catenin occupancy on the MYC promoter (Fig. 6C). Moreover, in re-ChIP studies, we found that GO-203 treatment is associated with decreases in MUC1-C and YAP occupancy (Fig. 6C). In addition, targeting MUC1-C with GO-203 (Fig. 6D, left and right) or silencing (Fig. 6E, left and right) resulted in the suppression of MYC mRNA and protein. As confirmation that the observed MYC downregulation is indeed YAP-dependent, we transiently silenced YAP in MDA-MB-231 cells and found suppression of MYC expression (Fig. 6F). Discussion Epithelia are comprised of a single layer of cells that have a unique structure in which the apical surface faces (i) the external environment in the respiratory and gastrointestinal tracts, or (ii) a lumen in ducts of specialized organs, such as the mammary gland (Fig. 7A)(2). Apical-basal polarity is however disrupted in the response to stress and is associated with induction of a proliferation and survival program (43). MUC1 evolved in mammals to afford protection of epithelia (16, 44). In this way, the MUC1-N subunit forms a physical mucous barrier that protects against agents, such as toxins and microorganisms among others, that damage the apical surface (Fig. 7A) (16). In turn, the MUC1-C subunit functions in transducing signals that promote the growth and survival response (Fig. 7A) (16). The present studies provide evidence that MUC1-C is of importance to the loss of epithelial cell polarity. Our results demonstrate that MUC1-C drives downregulation of (i) the Scribble complex HUGL2 protein, which is necessary for the establishment of the basolateral domains, and (ii) the Crumbs complex CRB3 and PATJ proteins, which define the apical membrane (2). The Scribble and Crumbs complexes are necessary for maintaining apical-basal polarity and epithelial integrity. These findings do not exclude the possibility that MUC1-C regulates other effectors necessary for cell polarity. In this regard, MUC1-C (i) epigenetically suppresses CDH1 expression, which disrupts adherens junctions, and (ii) interacts directly with β-catenin and p120, promoting their import from the lateral cell membrane to the nucleus (16, 17). Collectively, these findings support a previously unidentified model in which MUC1-C induces loss of polarity and integrates that response with self-renewal and stemness (21). In concert with this model, MUC1-C is expressed over the entire surface of breast cancer cells and is substantially upregulated (45), supporting the premise that TNBC cells have appropriated and exploited MUC1-C to confer properties of EMT, invasion and self-renewal (Fig. 7B) (21, 23). MUC1-C activates the inflammatory NF-κB pathway through interactions with TAK1 and the IKKs (33, 46). In addition, MUC1-C binds directly to NF-κB p65 and increases NF-κB occupancy on the promoters of its target genes, including MUC1 itself in an autoinductive circuit (22). The MUC1-C→NF-κB pathway also drives the ZEB1 gene (23), which encodes a transcriptional repressor that promotes EMT (47). In turn, MUC1-C forms a complex with ZEB1 and confers ZEB1 occupancy on miR-200c promoter and thereby its suppression (23). The present results further demonstrate that MUC1-C/ZEB1 complexes occupy the CRB3 promoter and, like miR-200c, inhibit CRB3 transcription (Fig. 7B, left). A similar mechanism of suppression applies to the HUGL2 promoter, linking MUC1-C/ZEB1 complexes to the downregulation of the CRB3 and HUGL2 genes. ZEB1 also plays a role in suppressing expression of the Pals1-associated tight junction protein (39); however, further studies will be needed to assess whether MUC1-C is also involved in this pathway. Indeed and as mentioned above, the present work has focused on MUC1-C-induced downregulation of CRB3 as a potential mechanism by which MUC1-C integrates loss of polarity and EMT with the Hippo pathway. By extension, the Hippo pathway regulates proliferation, apoptosis and differentiation (4, 5, 48), all of which have also been ascribed to MUC1-C signaling (16, 17). In concert with CRB3 functioning as an activator of the Hippo pathway, we found that targeting MUC1-C in TNBC cells is associated with increases in p-LATS1. In further support of MUC1-C as an inhibitor of the Hippo pathway, ectopic expression of MUC1-C in MCF-7 cells decreased p-LATS1. These findings support a model in which MUC1-C constitutively suppresses CRB3 and the Hippo pathway in mesenchymal TNBC cells. In addition, overexpression of MUC1-C in luminal MCF-7 cells is necessary for induction of these responses, indicating that this MUC1-C function is at least in part dependent on cell context. In this regard, MUC1-C activates the inflammatory NF-κB pathway in mesenchymal breast cancer cells, which also have increased YAP activity (49). By contrast, in luminal MCF-7 cells, ectopic MUC1-C expression is necessary for activating the NF-κB pathway (21, 23). These findings and the present results provide evidence that MUC1-C links the NF-κB inflammatory response to loss of polarity and suppression of the Hippo pathway (Figs. 7B, left). One outcome of an activated Hippo pathway is p-LATS1-mediated phosphorylation of YAP, resulting in the retention of p-YAP in the cytoplasm (10). By contrast, activated YAP localizes to the nucleus, where it functions by associating with transcription factors, such as TEAD (TEA domain family member) and β-catenin/TCF4, to activate genes involved in tumor induction and progression (4, 10). Another mechanism by which CRB3 can direct Hippo pathway function is by sequestering p-YAP at cell junctions through formation of complexes with certain cell polarity proteins, such as PALS, PATJ and AMOT. This sequestration also prevents access of p-YAP to phosphatases that counter the effect of the Hippo pathway kinases on YAP phosphorylation (4, 6). The present studies demonstrate that MUC1-C downregulates the Hippo pathway and thereby induces the nuclear localization of activated YAP (Figs. 7B, right). In the nucleus, YAP functions as a co-activator involved in the regulation of transcriptional programs that determine organ size and promote cellular proliferation (10). YAP has also been linked to transformation by activating transcription of mitogenic and anti-apoptotic genes through interactions with TEAD or with β-catenin/TCF4 (10, 11, 13–15). MUC1-C activates the WNT/β-catenin pathway by stabilizing β-catenin and driving WNT target genes, such as CCND1 and MYC (25–28). The present work extends this interaction by demonstrating that the MUC1-C cytoplasmic domain binds directly to YAP and promotes the in vitro formation of YAP/β-catenin complexes. In addition, we found that (i) MUC1-C associates with β-catenin and YAP on the MYC promoter, and (ii) targeting MUC1-C decreases occupancy of both β-catenin and YAP (Fig. 7B, right). In concert with these findings, targeting MUC1-C resulted in the downregulation of MYC, which in turn affects multiple pathways involved in cell growth, proliferation and survival (50). In summary, our findings provide convincing evidence for a previously unidentified role for MUC1-C in promoting loss of apical-basal polarity and linking this response to regulation of the Hippo pathway and YAP activation (Fig. 7B). These studies have largely focused on TNBC cells; however, we note that the results may be more broadly applicable to other types of carcinomas with overexpression of MUC1-C and downregulation of CRB3. Our findings also provide support for the notion that MUC1-C is an attractive therapeutic target. In this regard, GO-203 has completed Phase I clinical evaluation and has been formulated in polymeric nanoparticles for the treatment of patients with TNBC and other MUC1-C-expressing cancers (51). Supplementary Material 1 2 Financial Support: Research reported in this publication was supported by the National Cancer Institute of the National Institutes of Health under award numbers CA97098 and CA166480 and the Lung Cancer Research Foundation. Abbreviations MUC1 mucin 1 MUC1-C MUC1 C-terminal transmembrane subunit CRB3 mammalian homolog 3 of Drosophila Crumbs (Crb) TNBC triple-negative breast cancer YAP Yes-associated protein TAZ the transcriptional co-activator with PDZ-binding motif TSG Tumor Suppressor Gene ChIP chromatin immunoprecipitation DOX doxycycline TEAD transcriptional enhancer activator domain family of transcription factors Figure 1 MUC1-C downregulates CRB3 expression A, BT-549 cells infected with lentiviruses to stably express a control scrambled shRNA (CshRNA) or a MUC1shRNA were analyzed for MUC1, CRB3, HUGL2, PATJ and CDC42 mRNA levels by qRT-PCR (left). The results are expressed as relative mRNA levels (mean±SD of three determinations) as compared with that obtained for BT-549/CshRNA cells (assigned a value of 1). Lysates from the BT-549/CshRNA and BT-549/MUC1shRNA cells were immunoblotted with the indicated antibodies (right). B, MDA-MB-231/CshRNA and MDA-MB-231/MUC1shRNA cells were analyzed for MUC1, CRB3, HUGL2, PATJ and CDC42 mRNA levels by qRT-PCR (left). The results are expressed as relative mRNA levels (mean±SD of three determinations) as compared with that obtained for MDA-MB-231/CshRNA cells (assigned a value of 1). Lysates from the MDA-MB-231/CshRNA and MDA-MB-231/MUC1shRNA cells were immunoblotted with the indicated antibodies (right). C, BT-20 cells stably expressing an empty vector or MUC1-C were analyzed for CRB3 mRNA levels by qRT-PCR (left). The results are expressed as relative CRB3 mRNA levels (mean±SD of three determinations) as compared to that obtained for BT-20/vector cells (assigned a value of 1). Lysates from the BT-20/vector and BT-20/MUC1-C cells were immunoblotted with the indicated antibodies (right). D, Schematic showing the MUC1-C subunit and the amino acid (aa) sequence of the cytoplasmic domain (CD). ED, extracellular domain; TM, transmembrane domain. The CQC motif is necessary and sufficient for MUC1-C homodimerization. D-amino acid sequences are shown for GO-203 and CP-2. E, BT-549 cells were treated with 5 μM GO-203 or CP-2 each day for 3 days and then analyzed for CRB3 mRNA levels by qRT-PCR (left). Lysates from BT-549 cells treated as above were immunoblotted with the indicated antibodies (right). F, MDA-MB-231 cells were treated with 5 μM GO-203 or CP-2 each day for 3 days and then analyzed for CRB3 mRNA levels by qRT-PCR (left), while lysates from similarly treated cells were immunoblotted with the indicated antibodies (right). Figure 2 MUC1-C suppresses CRB3 expression by a ZEB1-dependent mechanism A, BT-549 cells infected with lentiviruses to stably express a control scrambled shRNA (CshRNA) or a ZEB1shRNA were analyzed for CRB3 mRNA levels by qRT-PCR (left). The results are expressed as relative CRB3 mRNA levels (mean±SD of three determinations) as compared with that obtained for BT-549/CshRNA cells (assigned a value of 1). Lysates from the BT-549/CshRNA and BT-549/ZEB1shRNA cells were immunoblotted with the indicated antibodies (right). B, MDA-MB-231 cells stably expressing a control scrambled shRNA (CshRNA) or a ZEB1 shRNA were analyzed for CRB3 mRNA levels by qRT-PCR (left). The results are expressed as relative CRB3 mRNA levels (mean±SD of three determinations) as compared with that obtained for MDA-MB-231/CshRNA cells (assigned a value of 1). Lysates from the MDA-MB-231/CshRNA and MDA-MB-231/ZEB1shRNA cells were immunoblotted with the indicated antibodies (right). C, MCF-7/MUC1-C cells stably expressing a control scrambled shRNA (CshRNA) or a ZEB1 shRNA were analyzed for CRB3 mRNA levels by qRT-PCR (left). Lysates from MCF-7/MUC1-C/CshRNA and MCF-7/MUC1-C/ZEB1shRNA cells were immunoblotted with the indicated antibodies (right). D, Lysates from MDA-MB-231/CshRNA and MDA-MB-231/MUC1shRNA (left) and from MDA-MB-231/CshRNA and MDA-MB-231/SNAIL1shRNA (right) cells were immunoblotted with the indicated antibodies. E, Lysates from MCF-7/vector and MCF-7/MUC1-C cells (left) were immunoblotted with the indicated antibodies. Lysates from MCF-7/MUC1-C cells infected with lentiviruses to stably express a control scrambled shRNA (CshRNA) or a SNAIL1shRNA were immunoblotted with the indicated antibodies (right). Figure 3 MUC1-C represses CRB3 promoter activation by a ZEB1-dependent mechanism A, Schema of CRB3 luciferase reporter plasmid. B, MDA-MB-231/CshRNA and MDA-MB-231/MUC1shRNA cells (left), BT-549/CshRNA and BT-549/MUC1shRNA cells (middle) and MCF-7/Vector and MCF-7/MUC1-C cells (right) were transfected with the empty Luc vector or pCRB3-Luc. Cells were also transfected with the SV-40-Renilla-Luc plasmid as an internal control. Luciferase activity was measured at 48 h following transfection. The results are expressed as relative luciferase activity (mean±SD of three determinations) compared with that obtained from cells transfected with the empty Luc vector (assigned a value of 1). C (left), Soluble chromatin from the indicated MDA-MB-231 cells was precipitated with anti-ZEB1 or a control IgG. The final DNA precipitates were amplified by qPCR with pairs of primers for the ZEB1 binding region in CRB3 promoter region. Results are expressed as the relative fold enrichment (mean±SD of three determinations) compared with that obtained for the IgG control (assigned a value of 1). For re-ChIP analysis, soluble chromatin from MDA-MB-231 cells (right) was first precipitated with anti-ZEB1, then released and reimmunoprecipitated with anti-MUC1-C. The results are expressed as the relative fold enrichment (mean±SD of three determinations) compared with that obtained with the IgG control (assigned a value of 1). D, MDA-MB-231 cells were treated with 5 μM GO-203 or CP-2 each day for 3 days. Soluble chromatin was precipitated with anti-ZEB1 or a control IgG (left). For re-ChIP analysis, complexes were released and re-immunoprecipitated with anti-MUC1-C (right). Results are expressed as the relative fold enrichment (mean±SD of three determinations) compared with that obtained for the IgG control (assigned a value of 1). E, Soluble chromatin from MCF-7/vector and MCF-7/MUC1-C cells was precipitated with anti-ZEB1 or a control IgG (left). For re-ChIP analysis, complexes were released and re-immunoprecipitated with anti-MUC1-C (right). Results are expressed as the relative fold enrichment (mean±SD of three determinations) compared with that obtained for the IgG control (assigned a value of 1). Figure 4 MUC1-C-mediated CRB3 repression increases nuclear accumulation of YAP A, Lysates from MDA-MB-231/CshRNA and MDA-MB-231/MUC1shRNA cells (left), BT-549/CshRNA and bt-549/MUC1shRNA (middle) and MCF-7/Vector and MCF-7/MUC1-C (right) were immunoblotted with the indicated antibodies. B, Cytosolic fraction from MDA-MB-231/CshRNA and MDA-MB-231/MUC1shRNA cells (left), BT-549/CshRNA and BT-549/MUC1shRNA (middle) and MCF-7/Vector and MCF-7/MUC1-C (right) were immunoblotted with the indicated antibodies. C, Nuclear fraction lysates from MDA-MB-231/CshRNA and MDA-MB-231/MUC1shRNA cells (left), BT-549/CshRNA and BT-549/MUC1shRNA (middle) and MCF-7/Vector and MCF-7/MUC1-C (right) were immunoblotted with the indicated antibodies. D, Nuclear fraction lysates from MDA-MB-231/CshRNA and MDA-MB-231/MUC1shRNA cells (left), BT-549/CshRNA and BT-549/MUC1shRNA (middle) and MCF-7/Vector and MCF-7/MUC1-C (right) were immunoblotted with the indicated antibodies. Figure 5 Binding of MUC1-C and YAP A, Lysates from MDA-MB- 231 (left), BT-549 (middle) and MCF-7/MUC1-C (right) cells were precipitated with anti-MUC1-C or a control IgG. The precipitates were immunoblotted with the indicated antibodies. B, GST and GST-YAP were incubated with purified MUC1-C cytoplasmic domain (MUC1-CD). The adsorbates were immunoblotted with anti-MUC1-C. Input of the GST proteins was assessed by Coomassie blue staining. C, GST-YAP(1–160) and GST-YAP(161–504) were incubated with purified MUC1-CD. The adsorbates were immunoblotted with anti-MUC1-C. Input of the GST proteins was assessed by Coomassie blue staining. D, GST and GST-YAP were incubated with MUC1-CD(1–45) (left) or MUC1-CD(46–72) (right). The adsorbates and purified MUC1-CD proteins were immunoblotted with anti-MUC1-CD (CD1, left; CT2, right) antibodies. Input of the GST proteins was assessed by Coomassie blue staining. E, GST and GST-YAP were incubated with purified MUC1-CD and/or with purified β-catenin. The adsorbates were immunoblotted with anti-MUC1-CD. Input of the GST proteins was assessed by Coomassie blue staining. Figure 6 MUC1-C associates with YAP and β-catenin on the MYC promoter A, The indicated MDA-MB-231 cells were analyzed for CTGF and CYR61 mRNA levels by qRT-PCR. The results are expressed as relative mRNA levels (mean±SD of three determinations) as compared with that obtained for MDA-MB-231/CshRNA cells (assigned a value of 1). B, Soluble chromatin from the indicated MDA-MB-231 cells was precipitated with anti-β-catenin or a control IgG (left). The final DNA precipitates were amplified by qPCR with pairs of primers for the β-catenin binding region in the MYC promoter region. Results are expressed as the relative fold enrichment (mean±SD of three determinations) compared with that obtained for the IgG control (assigned a value of 1). For separate re-ChIP analysis, soluble chromatin from MDA-MB-231 cells was first precipitated with anti-β-catenin, then released and reimmunoprecipitated with anti-MUC1-C (middle) or anti-YAP (right). The results (mean±SD of three determinations) are expressed as the relative fold enrichment compared with that obtained with the IgG control (assigned a value of 1). C, MDA-MB-231 cells were treated with 5 μM GO-203 or CP-2 each day for 3 days. Soluble chromatin was precipitated with anti-β-catenin or a control IgG (left). For separate re-ChIP analysis, soluble chromatin from treated MDA-MB-231 cells was first precipitated with anti-β-catenin, then released and reimmunoprecipitated with anti-MUC1-C (middle) or anti-YAP (right). The results (mean±SD of three determinations) are expressed as the relative fold enrichment compared with that obtained with the IgG control (assigned a value of 1). D, MDA-MB-231 cells treated with 5 μM GO-203 or CP-2 each day for 3 days were analyzed for MYC mRNA by qRT-PCR (left). The results are expressed as relative MYC mRNA levels (mean±SD of three determinations) as compared with that obtained for cells treated with CP2 peptide (assigned a value of 1). Lysates from MDA-MB-231 cells treated with 5 μM GO-203 or CP-2 were immunoblotted with the indicated antibodies (right). E, MDA-MB-231 cells stably expressing a control scrambled shRNA (CshRNA) or a MUC1 shRNA were analyzed for MYC mRNA levels by qRT-PCR (left). The results are expressed as relative MYC mRNA levels (mean±SD of three determinations) as compared with that obtained for MDA-MB-231/CshRNA cells (assigned a value of 1). Lysates from the MDA-MB-231/CshRNA and MDA-MB-231/MUC1shRNA cells were immunoblotted with the indicated antibodies (right). F, MDA-MB-231 cells were transiently transfected with a YAPsiRNA or a control siRNA (CsiRNA). Lysates were immunoblotted with the indicated antibodies. Figure 7 Proposed schema depicting the role of MUC1-C in repression of CRB3 expression, downregulation of the Hippo pathway and activation of YAP in TNBC cells A, The MUC1-N/MUC1-C complex is positioned at the apical borders of normal polarized epithelial cells (16, 17). MUC1-N protects the apical surface by forming a physical mucous barrier (16). The inactive MUC1-C subunit is poised to respond to stress signals, such as those induced by inflammation, toxins and microorganisms (16). Expression of CRB3 is associated with maintaining polarity, activating the Hippo pathway and downregulating YAP by retention of p-YAP in the cytoplasm. B, With transformation and irreversible loss of polarity, MUC1 is upregulated at the cell membrane and MUC1-N is shed from the cell surface. In turn, MUC1-C forms homodimers that are transported into the nucleus by an importin-β-mediated mechanism (16, 17). In the nucleus, MUC1-C forms a complex with ZEB1 on the CRB3 promoter and represses CRB3 transcription (left). Downregulation of CRB3 expression is associated with suppression of the Hippo pathway and activation of YAP (left). Notably, MUC1-C/ZEB1 complexes also repress miR-200c expression with the induction of EMT (23). MUC1-C also binds directly to β-catenin, stabilizes β-catenin and promotes the formation of MUC1-C/YAP/β-catenin complexes in the nucleus, which associate with TCF4 and drive MYC expression (right). Based on this model and the present results, targeting MUC1-C with GO-203 in TNBC cells blocks MUC1-C homodimerization and thereby its nuclear import with upregulation of CRB3 and activation of the Hippo tumor suppressor pathway. Potential Conflict of Interest: The authors declare competing financial interests: D.K. holds equity in Genus Oncology and is a consultant to the company. The other authors disclosed no potential conflicts of interest. 1 Thiery JP Acloque H Huang RY Nieto MA Epithelial-mesenchymal transitions in development and disease Cell 2009 139 5 871 90 19945376 2 Martin-Belmonte F Perez-Moreno M Epithelial cell polarity, stem cells and cancer Nat Rev Cancer 2012 12 1 23 38 3 Moreno-Bueno G Portillo F Cano A Transcriptional regulation of cell polarity in EMT and cancer Oncogene 2008 27 55 6958 69 19029937 4 Harvey KF Zhang X Thomas DM The Hippo pathway and human cancer Nat Rev Cancer 2013 13 4 246 57 23467301 5 Meng Z Moroishi T Guan KL Mechanisms of Hippo pathway regulation Genes Dev 2016 30 1 1 17 26728553 6 Mo JS Park HW Guan KL The Hippo signaling pathway in stem cell biology and cancer EMBO Rep 2014 15 6 642 56 24825474 7 Pefani DE O’Neill E Hippo pathway and protection of genome stability in response to DNA damage FEBS J 2015 8 Maugeri-Sacca M De Maria R Hippo pathway and breast cancer stem cells Crit Rev Oncol Hematol 2016 99 115 22 26725175 9 Varelas X Miller BW Sopko R The Hippo pathway regulates Wnt/beta-catenin signaling Dev Cell 2010 18 4 579 91 20412773 10 Moroishi T Hansen CG Guan KL The emerging roles of YAP and TAZ in cancer Nat Rev Cancer 2015 15 2 73 9 25592648 11 Heallen T Zhang M Wang J Hippo pathway inhibits Wnt signaling to restrain cardiomyocyte proliferation and heart size Science 2011 332 6028 458 61 21512031 12 Pan D The hippo signaling pathway in development and cancer Dev Cell 2010 19 4 491 505 20951342 13 Kim M Jho EH Cross-talk between Wnt/beta-catenin and Hippo signaling pathways: a brief review BMB Rep 2014 47 10 540 5 25154721 14 Imajo M Miyatake K Iimura A Miyamoto A Nishida E A molecular mechanism that links Hippo signalling to the inhibition of Wnt/beta-catenin signalling EMBO J 2012 31 5 1109 22 22234184 15 Tao J Calvisi DF Ranganathan S Activation of beta-catenin and Yap1 in human hepatoblastoma and induction of hepatocarcinogenesis in mice Gastroenterology 2014 147 3 690 701 24837480 16 Kufe D Mucins in cancer: function, prognosis and therapy Nature Reviews Cancer 2009 9 12 874 85 19935676 17 Kufe D MUC1-C oncoprotein as a target in breast cancer: activation of signaling pathways and therapeutic approaches Oncogene 2013 32 9 1073 81 22580612 18 Raina D Kosugi M Ahmad R Dependence on the MUC1-C oncoprotein in non-small cell lung cancer cells Mol Cancer Ther 2011 10 5 806 16 21421804 19 Raina D Uchida Y Kharbanda A Targeting the MUC1-C oncoprotein downregulates HER2 activation and abrogates trastuzumab resistance in breast cancer cells Oncogene 2014 33 26 3422 31 23912457 20 Alam M Ahmad R Rajabi H Kharbanda A Kufe D MUC1-C oncoprotein activates ERK→C/EBPβ-mediated induction of aldehyde dehydrogenase activity in breast cancer cells J Biol Chem 2013 288 43 30829 903 21 Alam M Rajabi H Ahmad R Jin C Kufe D Targeting the MUC1-C oncoprotein inhibits self-renewal capacity of breast cancer cells Oncotarget 2014 5 9 2622 34 24770886 22 Ahmad R Raina D Joshi MD Kawano T Kharbanda S Kufe D MUC1-C oncoprotein functions as a direct activator of the NF-κB p65 transcription factor Cancer Res 2009 69 7013 21 19706766 23 Rajabi H Alam M Takahashi H MUC1-C oncoprotein activates the ZEB1/miR-200c regulatory loop and epithelial-mesenchymal transition Oncogene 2014 33 13 1680 9 23584475 24 Yamamoto M Bharti A Li Y Kufe D Interaction of the DF3/MUC1 breast carcinoma-associated antigen and β-catenin in cell adhesion J Biol Chem 1997 272 19 12492 4 9139698 25 Huang L Chen D Liu D Yin L Kharbanda S Kufe D MUC1 oncoprotein blocks GSK3β-mediated phosphorylation and degradation of β-catenin Cancer Res 2005 65 22 10413 22 16288032 26 Rajabi H Ahmad R Jin C MUC1-C oncoprotein induces TCF7L2 transcription factor activation and promotes cyclin D1 expression in human breast cancer cells J Biol Chem 2012 287 13 10703 13 22318732 27 Bouillez A Rajabi H Pitroda S Inhibition of MUC1-C suppresses MYC expression and attenuates malignant growth in KRAS mutant lung adenocarcinomas Cancer Res 2016 76 6 1538 48 26833129 28 Tagde A Rajabi H Bouillez A MUC1-C drives MYC in multiple myeloma Blood 2016 127 21 2587 97 26907633 29 Kharbanda A Rajabi H Jin C Alam M Wong K Kufe D MUC1-C confers EMT and KRAS independence in mutant KRAS lung cancer cells Oncotarget 2014 5 19 8893 905 25245423 30 Hasegawa M Takahashi H Rajabi H Functional interactions of the cystine/glutamate antiporter, CD44v and MUC1-C oncoprotein in triple-negative breast cancer cells Oncotarget 2016 7 11 11756 69 26930718 31 Hiraki M Suzuki Y Alam M MUC1-C stabilizes MCL-1 in the oxidative stress response of triple-negative breast cancer cells to BCL-2 inhibitors Sci Rep 2016 6 32 Sonawane P Cho HE Tagde A Metabolic characteristics of 13-cis-retinoic acid (isotretinoin) and anti-tumour activity of the 13-cis-retinoic acid metabolite 4-oxo-13-cis-retinoic acid in neuroblastoma Br J Pharmacol 2014 171 23 5330 44 25039756 33 Takahashi H Jin C Rajabi H MUC1-C activates the TAK1 inflammatory pathway in colon cancer Oncogene 2015 34 40 5187 97 25659581 34 Wang Q Carroll JS Brown M Spatial and temporal recruitment of androgen receptor and its coactivators involves chromosomal looping and polymerase tracking Mol Cell 2005 19 5 631 42 16137620 35 Panchamoorthy G Rehan H Kharbanda A Ahmad R Kufe D A monoclonal antibody against the oncogenic mucin 1 cytoplasmic domain Hybridoma 2011 30 6 531 5 22149278 36 Leng Y Cao C Ren J Nuclear import of the MUC1-C oncoprotein is mediated by nucleoporin Nup62 J Biol Chem 2007 282 27 19321 30 17500061 37 Raina D Ahmad R Rajabi H Panchamoorthy G Kharbanda S Kufe D Targeting cysteine-mediated dimerization of the MUC1-C oncoprotein in human cancer cells Int J Oncol 2012 40 5 1643 9 22200620 38 Raina D Agarwal P Lee J Characterization of the MUC1-C cytoplasmic domain as a cancer target PLoS One 2015 10 8 e0135156 26267657 39 Aigner K Dampier B Descovich L The transcription factor ZEB1 (deltaEF1) promotes tumour cell dedifferentiation by repressing master regulators of epithelial polarity Oncogene 2007 26 49 6979 88 17486063 40 Zhao B Wei X Li W Inactivation of YAP oncoprotein by the Hippo pathway is involved in cell contact inhibition and tissue growth control Genes Dev 2007 21 21 2747 61 17974916 41 Rosenbluh J Nijhawan D Cox AG beta-Catenin-driven cancers require a YAP1 transcriptional complex for survival and tumorigenesis Cell 2012 151 7 1457 73 23245941 42 Shao DD Xue W Krall EB KRAS and YAP1 converge to regulate EMT and tumor survival Cell 2014 158 1 171 84 24954536 43 Vermeer PD Einwalter LA Moninger TO Segregation of receptor and ligand regulates activation of epithelial growth factor receptor Nature 2003 422 6929 322 6 12646923 44 Duraisamy S Kufe T Ramasamy S Kufe D Evolution of the human MUC1 oncoprotein Int J Oncology 2007 31 671 7 45 Kufe D Inghirami G Abe M Hayes D Justi-Wheeler H Schlom J Differential reactivity of a novel monoclonal antibody (DF3) with human malignant versus benign breast tumors Hybridoma 1984 3 3 223 32 6094338 46 Ahmad R Raina D Trivedi V MUC1 oncoprotein activates the IκB kinase β complex and constitutive NF-κB signaling Nat Cell Biol 2007 9 1419 27 18037881 47 Brabletz S Brabletz T The ZEB/miR-200 feedback loop--a motor of cellular plasticity in development and cancer? EMBO Rep 2010 11 9 670 7 20706219 48 McCaffrey LM Macara IG Epithelial organization, cell polarity and tumorigenesis Trends Cell Biol 2011 21 12 727 35 21782440 49 Cordenonsi M Zanconato F Azzolin L The Hippo transducer TAZ confers cancer stem cell-related traits on breast cancer cells Cell 2011 147 4 759 72 22078877 50 Dang CV MYC on the path to cancer Cell 2012 149 1 22 35 22464321 51 Hasegawa M Sinha RK Kumar M Intracellular targeting of the oncogenic MUC1-C protein with a novel GO-203 nanoparticle formulation Clin Cancer Res 2015 21 10 2338 47 25712682

PMC005xxxxxx/PMC5136470.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 2985117R 4816 J Immunol J. Immunol. Journal of immunology (Baltimore, Md. : 1950) 0022-1767 1550-6606 27807193 5136470 10.4049/jimmunol.1600604 EMS70248 Article PTPN22 is a critical regulator of Fcγ receptor mediated neutrophil activation Vermeren Sonja * Miles Katherine * Chu Julia Y. * Salter Donald † Zamoyska Rose ‡ Gray Mohini * * MRC / UoE Centre for Inflammation Research, Queen’s Medical Research Institute, University of Edinburgh, United Kingdom † Institute for Genetics and Molecular Medicine, University of Edinburgh, United Kingdom ‡ Institute of Immunology and Infection Research, Ashworth Laboratories; University of Edinburgh, United Kingdom Authors for correspondence: Sonja.Vermeren@ed.ac.uk and Mohini.Gray@ed.ac.uk, MRC / UoE Centre for Inflammation Research, Queen’s Medical Research Institute, 47 Little France Crescent, Edinburgh EH16 4TJ, U.K. Phone (+44) 131 2429100, Fax (+44) 131 2426578 19 10 2016 02 11 2016 15 12 2016 15 6 2017 197 12 47714779 This file is available to download for the purposes of text mining, consistent with the principles of UK copyright law. Neutrophils act as a first line of defense against bacterial and fungal infections but they are also important effectors of acute and chronic inflammation. Genome wide association studies have established that the gene encoding the protein tyrosine phosphatase PTPN22 makes an important contribution to susceptibility to autoimmune disease, notably rheumatoid arthritis. Although PTPN22 is most highly expressed in neutrophils, its function in these cells remains poorly characterized. We show here that neutrophil effector functions, including adhesion, production of reactive oxygen species and degranulation induced by immobilized immune complexes were reduced in Ptpn22-/- neutrophils. Tyrosine phosphorylation of Lyn and Syk was altered in Ptpn22-/- neutrophils. On stimulation with immobilized immune complexes, Ptpn22-/- neutrophils manifested reduced activation of key signaling intermediates. Ptpn22-/- mice were protected from immune complex mediated arthritis, induced by the transfer of arthritogenic serum. In contrast, in vivo neutrophil recruitment following thioglycollate induced peritonitis and in vitro chemotaxis were not affected by lack of PTPN22. Our data suggest an important role for PTPN22-dependent dephosphorylation events, that are required to enable full Fcγ receptor induced activation, pointing to an important role for this molecule in neutrophil function. Introduction Neutrophils are the most abundant peripheral blood leukocytes in humans. As part of the innate immune system they provide an immediate response to infection or injury. Neutrophils are rapidly activated by a variety of stimuli, including bacterial peptides, complement and immune complexes. Autoimmune diseases, including rheumatoid arthritis (RA), are associated with the generation of immune complexes, that accumulate in synovial fluid or are deposited on articular cartilage surfaces. They engage and activate neutrophils via Fcγ receptors (FcγRs) (1, 2). Severe inflammation follows neutrophil degranulation, releasing a plethora of degradative enzymes and other inflammatory mediators (3). The ensuing release of ROS and proteases degrades articular cartilage, whilst secreted chemokines attract further immune cells into the joint, driving chronic inflammation (4). Thus neutrophilic inflammation forms a crucial part of the inflammatory response, which needs to be resolved in a timely manner to minimize host damage. Protein tyrosine phosphatase nonreceptor 22 (PTPN22) is a leukocyte-restricted phosphatase, which is associated with an increased risk in a range of autoimmune diseases, notably RA. The single missense nucleotide polymorphism (SNP), C1858T encoding an R620W substitution is the single most important non MHC gene contributor to RA susceptibility, and the second most important for juvenile idiopathic arthritis according to candidate gene and genome wide association studies (5, 6). Although expression of PTPN22 is highest in the neutrophil (7), its function in these myeloid cells remains largely unknown. In T cells, PTPN22 has been shown to suppress T cell receptor (TCR) signaling, for instance by dephosphorylating key tyrosine residues within the activation loops of the Src family kinases (SFKs) Lck and Fyn and the TCR adapter Zap-70. At least in T cells, PTPN22 cooperates with the C-terminal Src kinase (Csk); their physical interaction is critical to their synergistic regulatory function. On a protein level, the disease-associated R620W variant (R619W in the mouse) affects one of four proline-rich regions in the C-terminus of PTPN22. This disrupts PTPN22 binding to Csk (8, 9). The K/BxN serum transfer arthritis model of arthritis is induced by administration of arthritogenic serum from arthritic KRN x NOD donors. This bypasses the need for an adaptive immune system-driven break in self-tolerance. It results in a transient, but rapidly evolving inflammatory arthritis, that reproduces many of the hallmarks of RA (10, 11). In combination with a range of experimental approaches, including genetic lineage depletion and reconstitutions, this disease model has helped to elucidate the important contribution of innate immune cells, notably neutrophils, to the effector phase of rheumatoid arthritis (12–14). We present here an analysis of PTPN22 function in the neutrophil, concentrating on FcγR signaling due to its prominent role in autoimmune diseases. By performing functional assays with isolated neutrophils from PTPN22 deficient mice and by analysing inflammation in K/BxN serum transfer arthritis, we demonstrate that PTPN22 regulates FcγR neutrophilic inflammation. Methods Unless otherwise stated, materials were obtained from Sigma. Antibodies Antibodies directed against phosphotyrosine (PY1000), phospho-Syk (Y525/526), phospho-pan Src (Y527 and Y416), phospho-Akt (S473), phospho-p38 (T180, Y182) and phospho-Erk (T202, Y204) were from Cell Signaling Technology. Anti-Syk (clone 5F5) and anti-Lyn (clone LYN-01) were from Biolegend. Anti-Ly6G (clone RC6-8C5) was obtained from R&D Systems. Anti-BSA and anti-lactoferrin antibodies were from Sigma and an antibody against β-COP was a gift from Nick Ktistakis (The Babraham Institute, Cambridge). HRP-conjugated secondary antibodies were from Santa Cruz Biotechnology and Biorad. Fluorescently-conjugated antibodies for flow cytometry were obtained from eBioscience (F4/80, GR1), Biolegend (CD11b, CD11a, CD16/32, Ly6G, CD62L, CD19) and BD (Ly6C). PTPN22 mouse model Generation of the PTPN22 mouse has been previously described (15). Experimental mice were housed in individually ventilated cages in a specific and opportunistic pathogen free small animal barrier unit at the University of Edinburgh. All animal work was approved by United Kingdom Home Office Project license PPL60/4567. Analysis of peripheral blood Peripheral blood was sampled from the superficial temporal vein into sodium citrate. Leukocytes were stained using fluorescently conjugated antibodies, mixed with flow-check fluorospheres (Beckman Coulter) to determine cell numbers and red blood cells were lysed (BD FacsLyse). Samples were analysed by flow cytometry using a LSR Fortessa (BD). K/BxN serum transfer arthritis model The K/BxN serum transfer model was induced using pooled arthritogenic serum as previously described (10, 11, 16). Mice were scored for 20 days. Each limb was scored as; 0, normal; 1, erythema or swelling in a single digit; 2, erythema and swelling in two or more joints; 3, swelling of the entire paw including the hock joint. The sum of the score for each limb (giving a maximum possible score of 12) was taken as a measure of the extent of arthritis at that time-point. Formalin fixed samples were prepared for histology by decalcification in EDTA. 3 μm sections were stained with hematoxillin and eosin. Immunostaining for Ly6G was performed according to standard procedures without antigen retrieval employing a secondary HRP-coupled antibody followed by a DAB reaction. Analysis of histological specimen occurred in a blinded fashion by a histopathologist, using an Olympus BX51 microscope. Images were acquired using a Micropublisher 3.3RTV camera and Q Imaging software using 2x, 20x and 40x objectives. Thioglycollate induced sterile peritonitis Sterile peritonitis was induced by injection of matured 4% Brewer’s thioglycollate (BD) at a concentration of 20 ml/kg. Mice were sacrificed 3 hours after induction and peritoneal cells harvested following a peritoneal flush. Cells were stained with fluorescently conjugated antibodies and mixed with flow-check fluorospheres (Beckman Coulter) to determine cells numbers, before being analysed by fluorescence activated flow sorting using the LSR Fortessa (BD). Neutrophil purification Neutrophils were isolated from bone marrows of 12 to 14 week old sex- and age-matched mice using discontinuous Percoll gradients (GE Healthcare Amersham, Uppsala, Sweden) as previously described (17, 18), using endotoxin-free reagents throughout. Immobilized immune complexes and immobilization of adhesive proteins For immobilized immune complexes (ICs), dishes were coated overnight at 4ºC with endotoxin and fatty acid free BSA in PBS (100 μg/ml), blocked with 1% fat free milk in PBS, and incubated with rabbit anti-BSA (1:2000). Control surfaces were treated identically, but not incubated with antibody. Some assays were carried out with insoluble immune complexes, which had been generated as described (19). Human fibrinogen (150 μg/ml), polyRGD (20 μg/ml) or, as a control, heat-inactivated fetal calf serum (HI-FCS) were adsorbed onto tissue culture grade plastics overnight at 4ºC and for 3 hours at room temperature, respectively. All surfaces were extensively washed with PBS prior to performing any assays. ROS production assays ROS production was measured by chemiluminescence in a Synergy H1 plate reader (BioTek) employing a luminol-based assay in luminescence grade 96-well plates (Nunc, Thermo Scientific) essentially as previously described (20). Measurements were started immediately following cell stimulation and light emission was recorded. Data output was relative light units per second (RLU/sec). Chemotaxis assays For chemotaxis assays, neutrophils were resuspended in HBSS supplemented with 15mM Hepes pH 7.4 and 0.05% fatty acid and endotoxin free BSA. Dunn chambers were assembled as described (21), and chemotaxis assays were carried out employing MIP2 (R&D Systems) as chemoattractant. Cells were monitored by time-lapse imaging for 30 minutes using an inverted RMDIB microscope (Leica) equipped with temperature controlled chamber, automated stage (Prior), Orca camera (Hamamatsu) and Micromanager image acquisition software (Fiji). Paths of all individual neutrophils were tracked using the ‘manual tracking’ plug-in into ImageJ. Tracks were subsequently analysed using the ‘Chemotaxis Tool’ (Ibidi) plug-in into Image J. Transendothelial migration assays TEM assays were carried out as previously described (22). Briefly, 6.5μm transwell inserts with 3μm polycarbonate membranes (Costar, Corning, UK) were coated overnight with 2μg/ml fibronectin, and seeded with 5 x 104 bEnd5 cells grown in DMEM supplemented with 10% heat inactivated fetal bovine serum (both Life Technologies) as described. Confluent bEnd5 cells were stimulated for 16 hrs with 5nM TNFα and 5 x 105 neutrophils were added into washed inserts that had been placed into wells of 24 well plates in the presence of 0, 1 or 3nM MIP2 and allowed to migrate towards the chemoattractant. Transmigrated neutrophils were labelled for GR1 and 8 randomly chosen fields of view / 24 well were photographed for counting [20x magnification using an Evos cell imaging system (AMG)]. Adhesion assays Neutrophils were allowed to adhere to immobilized IC or blocked surfaces at 37ºC in 96 well plates. Dishes were then subjected to rapid orbital shaking for 30 seconds before contents were flicked sharply out of wells prior to cell fixation. Following extensive washing, 4 random images were taken from each well for analysis of cell spreading (20x magnification; Evos system). Degranulation assays Gelatinase granule release after plating neutrophils onto an immobilized IC coated surface, or following stimulation with fMLF and cytochalasin B was detected by in-gel zymography as previously described (20). Lactoferrin release was assayed by making use of an antibody to human lactoferrin that had previously been shown to cross-react with mouse protein essentially as described (23, 24). Analysis of protein phosphorylation Neutrophils were plated onto immobilized IC coated or control treated dishes for stimulation. For analysis of phosphorylation of PKB, Erk and p38 MAPK, this was done as previously described (20, 25). For analysis of Lyn and Syk, non-adherent cells and scraped, adherent cells were combined for lysis in ice-cold 100mM NaCl, 30mM Hepes pH7.4, 20mM NaF, 1mM EGTA, 1% Triton X-100, 1mM benzamidine, 10μg/ml aprotinin, 1mM PMSF, 1mM Na3VO4 as well as mammalian protease inhibitor cocktail and phosphatase inhibitor cocktail 2. After pelleting detergent-insoluble material, proteins of interest were immunoprecipitated using protein G agarose and antibodies as indicated. Carefully washed beads were boiled in sample buffer and immunoprecipitated proteins and lysate controls separated SDS-PAGE. Proteins were transferred to Immobilon membrane (Millipore) for Western blotting using antibodies as indicated. Cytokine release assays Cytokine release assays were carried out using insoluble HSA-anti HSA immune complexes as stimulation. Insoluble ICs were prepared by titrating the point of equivalence between antigen and antibody and monitoring the point of equivalence by measuring the absorbance at 450nM as previously described (19). Neutrophils were stimulated as indicated and cultured in round bottom 96 well plates (Corning) in Dulbecco’s PBS supplemented with Ca2+ and Mg2+, 1g/L glucose and 4mM sodium bicarbonate in a humidified atmosphere at 37°C and 5% CO2. After 6 hours, supernatants were harvested for cytokine analysis by ELISA (R&D research) according to manufacturer’s instructions. Statistical Analysis For kinetic experiments, the area under the graph was used for analysis. Where data met the assumptions for parametric tests, two-tailed Student’s t-tests were applied. Otherwise, the non-parametric Mann-Whitney was test used. P-values <0.05 were considered statistically significant. Results Ptpn22-/- mice have already been described. In line with previous reports, we found them to be fertile and without overt deleterious or autoimmune phenotype (15, 26). Ptpn22-/- mice were characterized by mild peripheral blood neutrophilia (Fig 1A), complementing previous observations of increased B and T cell numbers in Ptpn22-/- mice (Hasegawa et al 2004, Maine et al 2015). To test whether this mild neutrophilia was a result of inadequate activation of Ptpn22-/- neutrophils, we analysed cell surface L-selectin / CD62L on freshly drawn peripheral blood neutrophils from control and Ptpn22-/- mice. This was similar, suggesting that the neutrophil activation status was not affected by the absence of PTPN22 (Fig 1B). PTPN22 regulates neutrophil adhesion to immobilized immune complexes Given the documented role of PTPN22, neutrophils and immune complexes (IC) in autoimmune driven inflammatory arthritis, we analysed IC-induced effector functions with purified, bone marrow derived neutrophils. When plating the cells onto surfaces that had been coated with immobilized ICs, fewer Ptpn22-/- than wild-type control neutrophils were found to adhere firmly (Fig 2A). In contrast, both Ptpn22-/- and wild-type (WT) control neutrophils were equally able to spread once attached to the immobilized ICs (Fig 2B,C). Similar results were obtained with neutrophils that had been allowed to bind to immobilized ICs for a shorter time (10 minutes; not shown). Neutrophils of both genotypes displayed very similar levels of surface FcγRII/III (Fig 2D). Since FcγRs and β2 integrins are known to cross-talk extensively (27, 28), we also characterised cell surface integrins. We observed no difference in cell surface expression of the common leukocyte β2 integrins LFA-1 / CD11a and noted a similar increase in cell surface Mac-1 / CD11b following stimulation of Ptpn22-/- and wild-type control neutrophils with fMLF (Fig 2D; supplemental Fig 1). These data suggest that reduced adhesion of Ptpn22-/- neutrophils was not due to reduced cell surface FcγR or integrin receptor density. PTPN22 regulates adhesion- and FcγR-dependent ROS formation FcγR engagement initiates specific neutrophil effector functions, including ROS production and degranulation. To test whether Ptpn22-/- neutrophils were able to generate ROS, we stimulated cells with the phorbol ester PMA (Fig 3A,B) or with a formylated peptide (fMLF) (Fig 3C,D). We noted no differences to control neutrophils. In contrast, ROS formation by Ptpn22-/- neutrophils that had been plated onto immobilized ICs was significantly lower than that seen with wild-type control neutrophils (Fig 3E,F). In-line with well-established cross-talk between FcγRs and integrins (27, 28), a similarly reduced generation of ROS was also observed when neutrophils were plated on the β2 integrin ligand fibrinogen in the presence of TNFα (Fig 3G,H) or onto the synthetic pan integrin ligand polyArgGlyAsp (Fig 3I,K). In conclusion, Ptpn22-/- neutrophils were characterized by reduced ROS production following FcγR and integrin-dependent stimulations. PTPN22 regulates FcγR-dependent neutrophil degranulation We next tested the ability of Ptpn22-/- neutrophils to degranulate by measuring the release of lactoferrin and gelatinase, components of neutrophil secondary and tertiary granules, respectively. As observed with ROS production, Ptpn22-/- neutrophils degranulated less efficiently than controls when plated onto immobilized ICs, but not when they were stimulated with fMLF in the presence of cytochalasin B (Fig 4 A-C). These data argue that PTPN22 is required for optimal neutrophil effector functions involved in generating an inflammatory environment following stimulation of FcγR (and integrins), and likely functioning in intracellular signaling events downstream of receptor engagement. To address whether PTPN22 may also be involved in cytokine production, we analysed release of TNFα, MIP2 (a murine equivalent of IL-8) and IL-1β by neutrophils that had been stimulated with immune complexes. The extent of cytokine release triggered by IC-stimulation of neutrophils detected in these assays was very low and any reductions observed with Ptpn22-/- neutrophils did not reach significance (Supplemental Fig 2). PTPN22 functions to activate signaling processes downstream of FcγRs PTPN22 is a protein tyrosine phosphatase. Signaling downstream of FcγRs (and integrins) involves significant tyrosine phosphorylation. Receptor-proximal key signaling proteins regulated include the major neutrophil SFKs Lyn, Hck and Fgr, and the spleen tyrosine kinase (Syk). We analysed tyrosine phosphorylation of Syk immunoprecipitated from control and Ptpn22-/- lysates from neutrophils that had been plated onto immobilized ICs (IgG-BSA) or onto BSA (Fig 5A). Mock-stimulated Ptpn22-/- neutrophils were characterized by increased Syk tyrosine phosphorylation (221.4 ± 16.8% compared to control) as well as a second, slightly slower migrating band (indicated by * in Fig 5A) than wild-type controls. No differences were apparent between Syk tyrosine phosphorylation of activated wild-type and Ptpn22-/- neutrophils. Probing with an antibody that specifically recognises the activating phosphotyrosine residues Y519/Y520 in mouse Syk, revealed no differences between genotypes, with no signal in basal control or Ptpn22-/- neutrophils. This suggested that Syk tyrosine residue(s) other than Y519/Y520 are dephosphorylated either directly or indirectly by PTPN22 in the neutrophil. Lyn has been shown to act as an inhibitory SFK that phosphorylates Syk in B cells (29, 30). We therefore analysed Lyn in control and Ptpn22-/- neutrophils. As a result of differential splicing Lyn is known to run as a doublet (31), however, at least with the antibody we employed, Lyn migrated as three distinct bands in basal and stimulated control cells as well as in basal Ptpn22-/- cells. In lysates from stimulated Ptpn22-/- neutrophils, we noticed a fourth Lyn band (Fig 5B, arrow). Given the short time-frame of the stimulation employed, this was likely caused by a post-translational modification of Lyn. We therefore immunoprecipitated Lyn for analysis with phosphorylation specific antibodies. This revealed increased total tyrosine phosphorylation of Ptpn22-/- neutrophils (150.1 ± 6.2% compared to control), and slight increases for both the activating Y527 residue (124.5 ± 6.1%), and the inhibitory Y416 residue (118 ± 2.7%). In addition, phosphorylated Lyn from Ptpn22-/- neutrophils migrated slower than Lyn from controls. This was particularly apparent for the phospho-Y527 (Fig 5B, arrows). Taken together, our observations suggested that PTPN22 regulates Lyn phosphorylation in the neutrophil. To further address the implications of PTPN22-mediated dephosphorylation events, we also analyzed the activity status of some key signaling intermediates that are activated downstream of SFK / Syk signaling following activation of FcγRs. We probed for phospho-PKB (also known as Akt), a read-out for the activity status of phosphoinositide 3-kinase, phospho-extracellular signal regulated kinase (Erk), and phospho-p38 mitogen activated protein (MAP) kinase. All of these signaling intermediates were found to be mildly hypophosphorylated in Ptpn22-/- neutrophils (Fig 5C-F), indicating that many signaling events downstream of FcγR engagement are influenced by PTPN22 in the neutrophil. Finally, phosphotyrosine blots of lysates from control and Ptpn22-/- neutrophils that had been plated onto BSA or immobilized immune complexes showed reduced tyrosine phosphorylation in basal and activated Ptpn22-/- lysates (supplemental Fig 3). Together, these in vitro data suggested that PTPN22 phosphatase performs activating dephosphorylation events in the neutrophil. Ptpn22-/- mice are protected from K/BxN serum transfer arthritis Given the association of the human PTPN22 SNP, C1858T with rheumatoid arthritis, we tested whether the in vitro observations were relevant in vivo. K/BxN serum was administered to control and Ptpn22-/- mice to induce serum transfer arthritis. Ptpn22-/- mice were significantly protected from inflammatory joint disease compared to age- and sex-matched wild-type controls. This was independent of the sex of the recipient animals, and the trend was observed with various doses of arthritogenic serum (Fig 6 and table I). Ptpn22-/- mice were characterized by slower progression of the disease and also reduced overall severity (Fig 6A). The clinical observations in the Ptpn22-/- mice were confirmed in histological sections taken near the height of inflammation (on day 5) which showed a reduced inflammatory cell infiltrate and few joint erosions when compared to wild-type controls (Fig 6B). Very few neutrophils were apparent in ankle joint sections at this time-point. Therefore, neutrophil infiltration was further evaluated at an earlier time-point (day 2), when neutrophil influx was expected to be maximal. Neutrophils in joint sections were visualized by their distinct morphology and by Ly6G staining (Fig 6C, red arrows). In contrast to controls and in keeping with the milder clinical disease, very few neutrophils were apparent in ankle joints of Ptpn22-/- mice. PTPN22 does not regulate neutrophil migration or recruitment Neutrophils have been shown to organise their own recruitment to the inflamed joint in K/BxN serum transfer arthritis. This is regulated by a series of chemokines and cytokines released by neutrophils and other synovial cells, that mediate recruitment of consecutive waves of neutrophils to the inflamed joint (32, 33). Since neutrophil migration into the inflamed joint is an absolute requirement for the generation of inflammation in this model (12–14), we analysed chemotaxis of Ptpn22-/- and control neutrophils. We observed no differences in terms of Euclidian or total accumulated distances travelled or indeed in the chemotactic directionality (the ratio between the Euclidian and the total accumulated distance) in Dunn chamber chemotaxis towards MIP-2, indicating that Ptpn22-/- neutrophils undergo normal chemotaxis (Fig 7A-D). We further addressed whether the absence of PTPN22 might affect transendothelial migration (TEM), a key step required to enable neutrophils to leave the blood stream. TEM was assessed using an in vitro transwell chemotaxis assay, where neutrophils migrated through a transwell membrane supporting a monolayer of TNFα-stimulated murine endothelial cells. Ptpn22-/- deficient neutrophils migrated through these activated endothelial cells as efficiently as wild-type controls (Fig 7E), suggesting that PTPN22 does not regulate neutrophil TEM either. Finally, to test more definitively leukocyte recruitment in vivo we utilized the thioglycollate-induced sterile peritonitis model. Leukocyte recruitment was analysed at an early timepoint after induction of peritonitis, when neutrophil recruitment peaks ahead of the recruitment of other leukocytes (34). Again, and in line with the in vitro observations above, there was no detectable defect in neutrophil recruitment to the inflamed peritoneum of Ptpn22-/- deficient mice (Fig 7F). These observations demonstrate that PTPN22 does not regulate neutrophil chemotaxis, TEM or recruitment per se. Rather, PTPN22 influences the activation of neutrophils in response to ICs. Discussion Our data show a significant protection of Ptpn22-/- mice from immune complex-driven K/BxN serum transfer arthritis. This model tests the effector phase of arthritis, which depends on neutrophil activation for its clinical expression. Protection was observed with both males and females under a number of experimental conditions (Fig 6 and Table I). In contrast to our results, Wang et al (35) and Maine et al (36) noted no protection from joint protection using the K/BxN model in Ptpn22-/- mice. We suspect that these discrepancies may be the result of differences in the experimental conditions. Both studies induced arthritis by injecting larger amounts of arthritogenic serum than ourselves, inducing a severe and prolonged arthritis. In contrast, we aimed for sub-maximal transient arthritis, that would enable the detection of subtle differences between experimental cohorts. It is conceivable that the degree of neutrophil stimulation induced in these studies overrode the protection afforded by the lack of PTPN22. In addition, it is possible that differences in housing conditions would lead to variations in the microbiome of the mice used for the different studies. The gastrointestinal microbiota has recently been recognized to regulate immunity and inflammatory responses to in vivo challenges (37). Interestingly, K/BxN arthritis depends heavily on the microbiome, with the T cell independent effector phase of K/BxN serum transfer arthritis being regulated by gut microbiota (38, 39). In vitro, with Ptpn22-/- neutrophils, we observed normal chemotaxis and transendothelial migration towards MIP2. In vivo, normal levels of neutrophil recruitment in the thioglycollate induced model of peritonitis was also apparent in Ptpn22-/- mice. This argues against a role for PTPN22 in neutrophil migration or recruitment (Fig 7). In contrast, few neutrophils were observed in the ankle joints of Ptpn22-/- mice following the induction of K/BxN arthritis (Fig 6). One explanation for this difference is that alternative mechanisms may be utilized by neutrophils migrating to different sites. Neutrophils have been shown to promote their own recruitment to inflamed joints in the K/BxN serum transfer arthritis model, by generating chemokines and inciting other cells to do the same (32, 33). Therefore, an alternative explanation maybe that Ptpn22-/- neutrophils were less efficient at promoting a suitable inflammatory environment to drive forward further waves of neutrophil recruitment, as seen with the K/BxN serum transfer arthritis. In vitro, neutrophil activation by immobilized ICs was defective, whilst stimulation induced by fMLF or PMA was not (Fig 3, 4). This indicates that PTPN22 is likely to be preferentially involved in inflammation, induced following integrin and/or FcγR-dependent neutrophil activation. Our observations are reminiscent of recent reports with Syk and SFK triple knock-out neutrophils, that provided complete protection from K/BxN serum transfer arthritis (40, 41). In these reports, Syk along with Lyn, Hck and Fgr were shown to be dispensable for neutrophil recruitment. Yet they played a crucial role in FcγR / integrin mediated neutrophil effector functions, including the generation of chemokines, that are required to generation inflammation (40, 41). We analysed IC induced chemokine generation by Ptpn22-/- neutrophils (Supplemental Fig 2). Overall cytokine secretion in these experiments was low and any differences between genotypes did not reach significance. It is likely such differences will be more pronounced with inflammatory neutrophils, such as those within the synovial fluid of the arthritic joint, than with unprimed, bone marrow derived neutrophils. It will be interesting to address these issues in more depth in the future. Interestingly, our data suggest that both Lyn and Syk tyrosine phosphorylation are affected in Ptpn22-/- neutrophils (Fig 5). We do not yet know the identity of the affected phosphorylation sites, nor whether these were direct events. It will be interesting to identify the exact nature of neutrophil PTPN22 substrates in the future. Given that the protection from K/BxN serum transfer arthritis afforded by Ptpn22-/- mice was larger than that of any individual SFK (Hck, Fgr or Lyn) (41), it is interesting to speculate that PTPN22 may contribute to regulating several SFKs, as well as other potential substrates. Such a function would be in keeping with our current understanding of PTPN22 function in T cells, where PTPN22 has been shown to dephosphorylate tyrosine residues within the SFKs Lck and Fyn; ZAP-70, a Syk family protein and also immune receptor tyrosine based activation motifs (ITAMS) in TCR associated proteins (26, 42, 43). Whilst these substrates are not expressed in neutrophils, some of their family members are. The findings presented here contrast sharply with observations by others who analyzed the role of PTPN22 in lymphocytes. In T cells, PTPN22 functions as a negative regulator (44). For instance, T cell receptor signaling and cell adhesion, mediated by Rap dependent integrin inside-out signaling were increased in Ptpn22-/- mice (15). The activating role of PTPN22 downstream of FcγRs might be specific to the neutrophil, suggesting that PTPN22 expression might individually modulate various leukocytes that contribute to inflammation. A missense SNP in PTPN22 (PTPN22-C1858T) predisposes carriers to a number of autoimmune diseases, most notably RA, SLE and type 1 diabetes (43, 45, 46). However, it protects against Behcets disease and bowel inflammation secondary to Crohn’s disease. Both of these diseases are characterized by significant neutrophil mediated chronic inflammation (47, 48). Our understanding of PTPN22 function and biochemistry, in particular in the neutrophil, lags behind studies of its function in T cells. Indeed, the effects of the R620W mutation is still being debated, with suggestions varying between the mutation being activating, inactivating and a hypomorph (45). Moreover, observations in the mouse do not always tally with those in humans, with studies in human cells split between those suggesting the R620W variant is a gain of function (43, 49), and others suggesting that it is a loss of function (50). In contrast, most studies conducted in the mouse are more suggestive of a loss of function mutation (51, 52). Furthermore, genetic background, as well as the presence of other, potentially as yet unidentified susceptibility alleles may play a role. R619W Ptpn22 knock-ins largely replicated the phenotype of Ptpn22-/- lymphocytes, suggestive of a loss of function mutation. One study suggested R619W PTPN22 is prone to degradation (52). However in a separately derived knock-in, protein stability was not impaired and mice kept on a mixed 129 x C57Bl/6 background, which is more prone to autoimmunity than pure C57Bl/6 developed mild, spontaneous autoimmune disease (51). In ascribing to PTPN22 an activating function, our work is suggestive, at least in the neutrophil, of a gain of function with R620W PTPN22. In the context of the results described here, it is hard to explain how a loss of function mutation would trigger enhanced inflammation in immune receptor dependent contexts, such as the K/BxN model for arthritis. It will be very interesting to test this hypothesis in the future by analyzing neutrophils derived from a R619W Ptpn22 point mutation knock-in mouse. As noted above, such mice have already been generated, but their analysis to date has been restricted to cell types outwith the neutrophil (51, 52). One recent study analyzed effector functions in human neutrophils derived from healthy donors and from RA patients that were or were not homozygous carriers of the SNP (53). Whilst that study did not test any IC-dependent events, the authors reported that neutrophils from homozygous carriers generate increased ROS, release more calcium in response to other stimulations and undergo faster transendothelial migration than those from matched control donors, in general agreement with the notion that R620W PTPN22 acts as a gain of function mutation in neutrophils. Supplementary Material Supplemental Data Acknowledgements We thank Adriano Rossi, Ian Dransfield and Moira Whyte for helpful discussions, Nick Ktistakis for antibodies and Kev Dhaliwal for access to equipment. Work in MG’s laboratory was supported by the Arthritis-Research-UK [20035] and the Medical Research Council [MR/J009555/1]. Work in SV’s laboratory was supported by a University of Edinburgh Chancellor’s Fellowship and a Wellcome Trust ISSF award. Figure 1 Ptpn22-/- mice have mild neutrophilia. (A) Ptpn22-/- mice have mild neutrophilia. Bloods from 17 Ptpn22-/- and 22 matched control mice were labeled for CD11b and Ly6G and analysed by flow cytometry. Data (means ± SEM) were pooled from four separate experiments; **, p<0.01. (B) Cell surface L-selectin was analysed in unstimulated neutrophils from bloods of 8 wild-type and 7 Ptpn22-/- mice. Neutrophils were gated for analysis of CD62L staining; MFIs are plotted. Figure 2 Ptpn22-/- mice have a defect in adhering to immobilized ICs. Neutrophils isolated from bone marrows of Ptpn22-/- (ko) and matched wild-type control (wt) mice were allowed to adhere for 20 minutes to wells that had been coated with immobilized ICs (IgG-BSA) or BSA as control, followed by washing and fixing. Adherent and spread cells were counted in four randomly photographed fields of view of each well. (A) Total number of all adhered cells. (B) Percentage of spread cells out of total. (A,B) Pooled data (means ± SEM) from three separate experiments are presented; **, p < 0.01. (C) Representative examples of Ptpn22-/- and matched control neutrophils allowed to adhere to immobilized ICs for 20 minutes. (D) Surface FcγRII/III, LFA-1 and Mac-1 of purified bone marrow derived neutrophils from control and Ptpn22-/- mice were analyzed by flow cytometry. Neutrophils were or were not stimulated with fMLF at 37°C before being labelled with FITC-conjugated anti-GR1 and PE-conjugated anti-FcγRII/III, anti-Mac1 or anti-LFA1. GR1 positive cells were gated and PE staining was measured. Results were analysed using FlowJo software. Black lines, wild-type neutrophils; grey lines, Ptpn22-/- neutrophils. Full lines, unstimulated cells and broken lines, stimulated cells. Experiments were performed with cells from four animals / genotype and representative examples are shown. Please see supplemental Fig 1 for a graphical representation of all experiments. Figure 3 Ptpn22-/- neutrophils are characterized by impaired immune receptor induced ROS production. Neutrophils isolated from bone marrows of Ptpn22-/- (ko) and matched wild-type control (wt) mice were used to characterize ROS production following stimulation with PMA (A,B), with fMLF (C,D), or following plating of cells into wells coated with immobilized IC (IgG-BSA) or as a control, BSA (E,F), into fibrinogen coated wells in the presence (stimulated) of absence (control) of TNFα (G, H), or into wells coated with the synthetic integrin ligand poly-Arg-Gly-Asp (polyRGD) or as a control, heat inactivated fetal calf serum (FCS) (I,K). Data shown (means ± range) are representative individual experiments (A,C,E,G,I). Total light emissions (means ± SEM) expressed as percentage of the response obtained with stimulated control neutrophils pooled from a minimum of three separate experiments (B,D,F,H,K). Figure 4 Ptpn22-/- neutrophils are characterized by reduced immune complex induced degranulation. (A,B) Gelatinase release by control (wt) and Ptpn22-/- (ko) neutrophils plated into wells coated with immobilized ICs (IgG-BSA) or as a control, BSA, or that had been stimulated by fMLF in the presence of cytochalasin B (fMLF + CB) was assessed by in-gel zymography. A representative experiment (A) and pooled data (means ± SEM) from five separately performed experiments (B) are presented. (C) Lactoferrin release by control and Ptpn22-/- neutrophils plated into wells coated with immobilized ICs (IgG-BSA) or as a control, BSA, or that had been stimulated by fMLF in the presence of cytochalasin B (fMLF + CB) was assayed by ELISA. Pooled data from three separate experiments are presented. Data presented are normalized to activated wild-type. Figure 5 Neutrophil PTPN22 regulates Lyn and Syk tyrosine phosphorylation which affects the activity status of many signaling cascades. Neutrophils isolated from bone marrows of Ptpn22-/- (ko) and matched wild-type control (wt) mice were allowed to adhere to immobilized ICs (IgG-BSA) or, as a control, BSA for 12 minutes. (A,B) Lysates were prepared as detailed in Materials and Methods for Western blotting using the indicated antibodies; in addition, Syk (A) or Lyn (B) were immunoprecipitated for analysis of phosphotyrosine as well as probing with antibodies detecting specific tyrosine phosphorylations in Syk / SFK proteins as indicated. Representative sample from at least 3 separately performed experiments are shown. (C-F) Lysates were prepared and subjected to immunoblotting with antibodies specific for phospho-PKB, phospho-Erk and phospho-p38 MAPK (C-F). Blots were quantitated using Image J software. (C) Representative blots; (D-F) Densitometry from five separate experiments (means ± SEM) was normalized to activated controls; *, p<0.05; **, p<0.01. Figure 6 Ptpn22-/- mice are protected from K/BxN arthritis. K/BxN serum transfer arthritis was induced in Ptpn22-/- (ko) and matched control (wt) mice. (A) Clinical scores (Means ± SEM) of the experimental groups were plotted over 20 days. Four experiments were carried out with 5-8 animals in each experimental group; a representative experiment is shown; experimental groups behaved significantly different, p = 0.02. (B) Wax sections of decalcified joints were H&E stained to visualize joint erosion and leukocyte infiltration on day 5 after serum transfer. Representative examples from sections obtained with right rear joints of arthritogenic serum injected control (wt) and Ptpn22-/- (ko) mice are shown (scale bar, 500μm); panels wt’ and ko’ are higher power images of the boxed areas in panels wt and ko (scale bar, 50μm). (C) Wax sections of decalcified joints were stained with an antibody specific for Ly6G and counterstained with hematoxylin to visualize neutrophil infiltration at day 2 following administration of arthritogenic serum. A red arrow in Ptpn22-/- points out a rare neutrophil (scale bar, 25μm). Figure 7 Neutrophil chemotaxis and recruitment are not affected in Ptpn22-/- mice. (A-D) Neutrophils isolated from bone marrows of Ptpn22-/- (ko) and matched wild-type control (wt) mice were allowed to chemotax towards 10nM MIP2 in Dunn chambers. Cell migration was monitored by time-lapse imaging, and individual cells were tracked using a cell tracking plug-in into Image J. (A) Tracks obtained in experiments carried out with three separate preparations of bone marrow derived neutrophils analysed were pooled for these spider plots. The source of chemoattractant is indicated (*). (B-D) The coordinates for the tracks shown in (A) were analysed using statistics features of the Ibidi chemotaxis plug-in into Image J as detailed in Materials and Methods. Parameters presented (means ± SEM) are, directionality (B), total accumulated (C) and Euclidian distances (D); differences were not statistically significant. (E) Ptpn22-/- (ko) and control (wt) neutrophils were allowed to migrate towards indicated concentrations of MIP2 in transwells supporting a monolayer of TNFα stimulated mouse endothelial cells. Data (means ± SEM) expressed as percentage of control cells migrating towards 1nM MIP2 from two pooled experiments carried out with separately prepared neutrophils are presented. (F) Peritonitis was induced in Ptpn22-/- (ko) and wild-type control (wt) mice by injection of thioglycollate containing broth. Peritoneal flushes were analyzed to enumerate numbers (means ± SEM) of peritoneal neutrophils. A representative experiment performed with 6 controls and 5 Ptpn22-/- mice is shown. Differences were not statistically significant. Table I Ptpn22-/- mice are protected from K/BxN arthritis. Irrespective of the gender of the recipient and of disease induction with a single, or repeat injections with arthritogenic serum, Ptpn22-/- mice were characterized by slower disease progression and reduced peak amplitude. Recipient N Treatment Mean time to peak Mean maximal score ± SD Time at peak intensity Median clinical score wt female 9 75μl on day 0 6.2 days 8.6 (±1.9) 2 days 3.8 ko female 9 75μl on day 0 6.8 days 5.0 (±1.9) 3.4 days 2.3 wt female 4 100μl on day 0, 100μl on day 2 5.8 days 10.25 (±1.6) 4 days 8.25 ko female 3 100μl on day 0, 100μl on day 2 7 days 6.0 (±1.7) 3.7 days 3.0 wt male 6 100μl on day 0 4.2 days 9.8 (±1.6) 2.3 days 4.0 ko male 6 100μl on day 0 6.7 days 6.5 (±2.9) 3.3 days 3.3 Authorship Contributions: SV and MG initiated the study and designed the experiments; SV, KM and YJC performed experiments; SV, KM, YJC, DS and MG analyzed experiments and interpreted the data; RZ provided experimental tools and advice. SV, MG and RZ wrote the manuscript. 1 Quayle JA Watson F Bucknall RC Edwards SW Neutrophils from the synovial fluid of patients with rheumatoid arthritis express the high affinity immunoglobulin G receptor, Fc gamma RI (CD64): role of immune complexes and cytokines in induction of receptor expression Immunology 1997 91 266 273 9227327 2 Robinson JJ Watson F Bucknall RC Edwards SW Role of Fc gamma receptors in the activation of neutrophils by soluble and insoluble immunoglobulin aggregates isolated from the synovial fluid of patients with rheumatoid arthritis Annals of the rheumatic diseases 1994 53 515 520 7944636 3 Nathan C Neutrophils and immunity: challenges and opportunities Nature reviews Immunology 2006 6 173 182 4 Wright HL Moots RJ Edwards SW The multifactorial role of neutrophils in rheumatoid arthritis Nature reviews Rheumatology 2014 10 593 601 24914698 5 Viatte S Plant D Raychaudhuri S Genetics and epigenetics of rheumatoid arthritis Nature reviews Rheumatology 2013 9 141 153 23381558 6 Hinks A Cobb J Marion MC Prahalad S Sudman M Bowes J Martin P Comeau ME Sajuthi S Andrews R Brown M Dense genotyping of immune-related disease regions identifies 14 new susceptibility loci for juvenile idiopathic arthritis Nature genetics 2013 45 664 669 23603761 7 Chien W Tidow N Williamson EA Shih LY Krug U Kettenbach A Fermin AC Roifman CM Koeffler HP Characterization of a myeloid tyrosine phosphatase, Lyp, and its role in the Bcr-Abl signal transduction pathway The Journal of biological chemistry 2003 278 27413 27420 12764153 8 Cloutier JF Veillette A Association of inhibitory tyrosine protein kinase p50csk with protein tyrosine phosphatase PEP in T cells and other hemopoietic cells The EMBO journal 1996 15 4909 4918 8890164 9 Gregorieff A Cloutier JF Veillette A Sequence requirements for association of protein-tyrosine phosphatase PEP with the Src homology 3 domain of inhibitory tyrosine protein kinase p50(csk) The Journal of biological chemistry 1998 273 13217 13222 9582365 10 Korganow AS Ji H Mangialaio S Duchatelle V Pelanda R Martin T Degott C Kikutani H Rajewsky K Pasquali JL Benoist C From systemic T cell self-reactivity to organ-specific autoimmune disease via immunoglobulins Immunity 1999 10 451 461 10229188 11 Kouskoff V Korganow AS Duchatelle V Degott C Benoist C Mathis D Organ-specific disease provoked by systemic autoimmunity Cell 1996 87 811 822 8945509 12 Monach PA Nigrovic PA Chen M Hock H Lee DM Benoist C Mathis D Neutrophils in a mouse model of autoantibody-mediated arthritis: critical producers of Fc receptor gamma, the receptor for C5a, and lymphocyte function-associated antigen 1 Arthritis and rheumatism 2010 62 753 764 20191628 13 Wipke BT Allen PM Essential role of neutrophils in the initiation and progression of a murine model of rheumatoid arthritis Journal of immunology 2001 167 1601 1608 14 Kim ND Chou RC Seung E Tager AM Luster AD A unique requirement for the leukotriene B4 receptor BLT1 for neutrophil recruitment in inflammatory arthritis The Journal of experimental medicine 2006 203 829 835 16567386 15 Brownlie RJ Miosge LA Vassilakos D Svensson LM Cope A Zamoyska R Lack of the phosphatase PTPN22 increases adhesion of murine regulatory T cells to improve their immunosuppressive function Science signaling 2012 5 ra87 23193160 16 Gray M Miles K Salter D Gray D Savill J Apoptotic cells protect mice from autoimmune inflammation by the induction of regulatory B cells Proceedings of the National Academy of Sciences of the United States of America 2007 104 14080 14085 17715067 17 Ellson CD Davidson K Ferguson GJ O'Connor R Stephens LR Hawkins PT Neutrophils from p40phox-/- mice exhibit severe defects in NADPH oxidase regulation and oxidant-dependent bacterial killing The Journal of experimental medicine 2006 203 1927 1937 16880254 18 Ferguson GJ Milne L Kulkarni S Sasaki T Walker S Andrews S Crabbe T Finan P Jones G Jackson S Camps M PI(3)Kgamma has an important context-dependent role in neutrophil chemokinesis Nature cell biology 2007 9 86 91 17173040 19 Fossati G Bucknall RC Edwards SW Insoluble and soluble immune complexes activate neutrophils by distinct activation mechanisms: changes in functional responses induced by priming with cytokines Annals of the rheumatic diseases 2002 61 13 19 11779751 20 Gambardella L Anderson KE Jakus Z Kovacs M Voigt S Hawkins PT Stephens L Mocsai A Vermeren S Phosphoinositide 3-OH kinase regulates integrin-dependent processes in neutrophils by signaling through its effector ARAP3 Journal of immunology 2013 190 381 391 21 Zicha D Dunn GA Brown AF A new direct-viewing chemotaxis chamber Journal of cell science 1991 99 Pt 4 769 775 1770004 22 Bixel MG Li H Petri B Khandoga AG Khandoga A Zarbock A Wolburg-Buchholz K Wolburg H Sorokin L Zeuschner D Maerz S CD99 and CD99L2 act at the same site as, but independently of, PECAM-1 during leukocyte diapedesis Blood 2010 116 1172 1184 20479283 23 Mocsai A Ligeti E Lowell CA Berton G Adhesion-dependent degranulation of neutrophils requires the Src family kinases Fgr and Hck Journal of immunology 1999 162 1120 1126 24 Abdel-Latif D Steward M Macdonald DL Francis GA Dinauer MC Lacy P Rac2 is critical for neutrophil primary granule exocytosis Blood 2004 104 832 839 15073033 25 Gambardella L Anderson KE Nussbaum C Segonds-Pichon A Margarido T Norton L Ludwig T Sperandio M Hawkins PT Stephens L Vermeren S The GTPase-activating protein ARAP3 regulates chemotaxis and adhesion-dependent processes in neutrophils Blood 2011 118 1087 1098 21490342 26 Hasegawa K Martin F Huang G Tumas D Diehl L Chan AC PEST domain-enriched tyrosine phosphatase (PEP) regulation of effector/memory T cells Science 2004 303 685 689 14752163 27 Mocsai A Abram CL Jakus Z Hu Y Lanier LL Lowell CA Integrin signaling in neutrophils and macrophages uses adaptors containing immunoreceptor tyrosine-based activation motifs Nature immunology 2006 7 1326 1333 17086186 28 Abram CL Lowell CA Convergence of immunoreceptor and integrin signaling Immunological reviews 2007 218 29 44 17624942 29 Keshvara LM Isaacson CC Yankee TM Sarac R Harrison ML Geahlen RL Syk- and Lyn-dependent phosphorylation of Syk on multiple tyrosines following B cell activation includes a site that negatively regulates signaling Journal of immunology 1998 161 5276 5283 30 Pereira S Lowell C The Lyn tyrosine kinase negatively regulates neutrophil integrin signaling Journal of immunology 2003 171 1319 1327 31 Yi TL Bolen JB Ihle JN Hematopoietic cells express two forms of lyn kinase differing by 21 amino acids in the amino terminus Mol Cell Biol 1991 11 2391 2398 2017160 32 Chou RC Kim ND Sadik CD Seung E Lan Y Byrne MH Haribabu B Iwakura Y Luster AD Lipid-cytokine-chemokine cascade drives neutrophil recruitment in a murine model of inflammatory arthritis Immunity 2010 33 266 278 20727790 33 Sadik CD Kim ND Iwakura Y Luster AD Neutrophils orchestrate their own recruitment in murine arthritis through C5aR and FcgammaR signaling Proceedings of the National Academy of Sciences of the United States of America 2012 109 E3177 3185 23112187 34 Pan D Amison RT Riffo-Vasquez Y Spina D Cleary SJ Wakelam MJ Page CP Pitchford SC Welch HC P-Rex and Vav Rac-GEFs in platelets control leukocyte recruitment to sites of inflammation Blood 2015 125 1146 1158 25538043 35 Wang Y Shaked I Stanford SM Zhou W Curtsinger JM Mikulski Z Shaheen ZR Cheng G Sawatzke K Campbell AM Auger JL The autoimmunity-associated gene PTPN22 potentiates toll-like receptor-driven, type 1 interferon-dependent immunity Immunity 2013 39 111 122 23871208 36 Maine CJ Marquardt K Cheung J Sherman LA PTPN22 controls the germinal center by influencing the numbers and activity of T follicular helper cells Journal of immunology 2014 192 1415 1424 37 Kamada N Seo SU Chen GY Nunez G Role of the gut microbiota in immunity and inflammatory disease Nature reviews Immunology 2013 13 321 335 38 Wu HJ Ivanov II Darce J Hattori K Shima T Umesaki Y Littman DR Benoist C Mathis D Gut-residing segmented filamentous bacteria drive autoimmune arthritis via T helper 17 cells Immunity 2010 32 815 827 20620945 39 Lee H Jin BE Jang E Lee AR Han DS Kim HY Youn J Gut-residing Microbes Alter the Host Susceptibility to Autoantibody-mediated Arthritis Immune Netw 2014 14 38 44 24605079 40 Elliott ER Van Ziffle JA Scapini P Sullivan BM Locksley RM Lowell CA Deletion of Syk in neutrophils prevents immune complex arthritis Journal of immunology 2011 187 4319 4330 41 Kovacs M Nemeth T Jakus Z Sitaru C Simon E Futosi K Botz B Helyes Z Lowell CA Mocsai A The Src family kinases Hck, Fgr, and Lyn are critical for the generation of the in vivo inflammatory environment without a direct role in leukocyte recruitment The Journal of experimental medicine 2014 211 1993 2011 25225462 42 Stanford SM Krishnamurthy D Falk MD Messina R Debnath B Li S Liu T Kazemi R Dahl R He Y Yu X Discovery of a novel series of inhibitors of lymphoid tyrosine phosphatase with activity in human T cells Journal of medicinal chemistry 2011 54 1640 1654 21341673 43 Vang T Liu WH Delacroix L Wu S Vasile S Dahl R Yang L Musumeci L Francis D Landskron J Tasken K LYP inhibits T-cell activation when dissociated from CSK Nature chemical biology 2012 8 437 446 22426112 44 Zheng J Petersen F Yu X The role of PTPN22 in autoimmunity: learning from mice Autoimmunity reviews 2014 13 266 271 24189282 45 Bottini N Peterson EJ Tyrosine phosphatase PTPN22: multifunctional regulator of immune signaling, development, and disease Annu Rev Immunol 2014 32 83 119 24364806 46 International Consortium for Systemic Lupus Erythematosus, G Harley JB Alarcon-Riquelme ME Criswell LA Jacob CO Kimberly RP Moser KL Tsao BP Vyse TJ Langefeld CD Nath SK Genome-wide association scan in women with systemic lupus erythematosus identifies susceptibility variants in ITGAM, PXK, KIAA1542 and other loci Nature genetics 2008 40 204 210 18204446 47 Baranathan V Stanford MR Vaughan RW Kondeatis E Graham E Fortune F Madanat W Kanawati C Ghabra M Murray PI Wallace GR The association of the PTPN22 620W polymorphism with Behcet's disease Annals of the rheumatic diseases 2007 66 1531 1533 17660222 48 Barrett JC Hansoul S Nicolae DL Cho JH Duerr RH Rioux JD Brant SR Silverberg MS Taylor KD Barmada MM Bitton A Genome-wide association defines more than 30 distinct susceptibility loci for Crohn's disease Nature genetics 2008 40 955 962 18587394 49 Vang T Congia M Macis MD Musumeci L Orru V Zavattari P Nika K Tautz L Tasken K Cucca F Mustelin T Autoimmune-associated lymphoid tyrosine phosphatase is a gain-of-function variant Nature genetics 2005 37 1317 1319 16273109 50 Zikherman J Hermiston M Steiner D Hasegawa K Chan A Weiss A PTPN22 deficiency cooperates with the CD45 E613R allele to break tolerance on a non-autoimmune background Journal of immunology 2009 182 4093 4106 51 Dai X James RG Habib T Singh S Jackson S Khim S Moon RT Liggitt D Wolf-Yadlin A Buckner JH Rawlings DJ A disease-associated PTPN22 variant promotes systemic autoimmunity in murine models The Journal of clinical investigation 2013 123 2024 2036 23619366 52 Zhang J Zahir N Jiang Q Miliotis H Heyraud S Meng X Dong B Xie G Qiu F Hao Z McCulloch CA The autoimmune disease-associated PTPN22 variant promotes calpain-mediated Lyp/Pep degradation associated with lymphocyte and dendritic cell hyperresponsiveness Nature genetics 2011 43 902 907 21841778 53 Bayley R Kite KA McGettrick HM Smith JP Kitas GD Buckley CD Young SP The autoimmune-associated genetic variant PTPN22 R620W enhances neutrophil activation and function in patients with rheumatoid arthritis and healthy individuals Annals of the rheumatic diseases 2014

PMC005xxxxxx/PMC5136472.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 101150042 30118 Mol Cancer Res Mol. Cancer Res. Molecular cancer research : MCR 1541-7786 1557-3125 27604276 5136472 10.1158/1541-7786.MCR-16-0108 EMS69963 Article Olaparib, alone or in combination with ionizing radiation, exacerbates DNA damage in normal tissues, as revealed by a new p21 reporter mouse McMahon Michael Frangova Tania G. Henderson Colin J. Wolf C. Roland 1 a School of Medicine, Ninewells Hospital and Medical School, University of Dundee, Dundee DD1 9SY, Scotland, United Kingdom a Corresponding author: C. Roland Wolf, c.r.wolf@dundee.ac.uk 28 9 2016 07 9 2016 12 2016 01 6 2017 14 12 11951203 This file is available to download for the purposes of text mining, consistent with the principles of UK copyright law. Many drugs targeting the DNA damage response are being developed as anti-cancer therapies, either as single agents or in combination with ionising radiation or other cytotoxic agents. Numerous clinical trials in this area are either in progress or planned. However, concerns remain about the potential of such treatments to increase toxicity to normal tissues. In order to address this issue, we have created a novel reporter mouse line through the simultaneous incorporation of multiple reporters, β-galactosidase and firefly luciferase, into the DNA damage-inducible p21 locus. We show that in situ β-galactosidase staining facilitates high fidelity mapping of p21 expression across multiple organs and tissues at single-cell resolution, whereas the luciferase reporter permits non-invasive bioluminescent imaging of p21 expression. Using this model, we have studied the capacity of a number of DNA damaging agents, including ionizing radiation, cisplatin, and etoposide to induce p21 expression in normal tissues. We have also studied the PARP inhibitor olaparib alone or in combination with ionizing radiation as well as cisplatin. A single exposure to olaparib alone caused DNA damage to cells in the mucosal layer lining mouse large intestine. It also exacerbated DNA damage induced in this organ and the kidney by co-administered ionizing radiation. These studies suggest that olaparib might be a carcinogen in man and illustrate the power of our new model to evaluate the safety of new therapeutic regimens involving combination therapies. DNA damage CDKN1A olaparib PARP inhibitors anticancer drugs Introduction Genomic instability is a defining characteristic which distinguishes tumours from normal cells(1). For this reason, inhibitors of DNA damage response (DDR) pathways have become an important emerging class of anticancer drugs. Work in this area was given added impetus by the discovery that inhibitors of PARP1, a protein involved in single-strand DNA repair(2), are over 1000-fold more toxic to BRCA-deficient cells than normal cells in culture (3,4). This synthetic lethality suggested that PARP inhibitors, and perhaps other DDR-targeted agents, would make safe and efficacious anti-cancer drugs. One such inhibitor, olaparib(5), has been shown to be well-tolerated in clinical trials and exhibits favourable activity against ovarian(6) tumours bearing defective BRCA alleles. On this basis, it was approved in 2014 for the monotherapy of this disease (7). Olaparib is now in clinical trials in combination with DNA damaging agents such as ionizing radiation (IR), cisplatin, gemcitabine, temozolomide and irinotecan (https://clinicaltrials.gov/ct2/results?term=OLAPARIB&pg=1) in the hope that it will synergize, and enhance the efficacy of these DNA-damaging agents selectively in tumour cells (8). In addition, on the basis that the safety of olaparib has already been firmly established, it has been proposed as a chemopreventive agent for use in healthy individuals carrying germline BRCA mutations who are at high-risk of developing ovarian and/or breast cancer (9), It has also been proposed as an agent to alleviate the symptoms of chronic inflammation (10). However, the safety of olaparib and other DDR-targeted drugs has not yet been adequately assessed in vivo. The major concern is that they might accelerate the background rate of DNA damage in normal tissues (1). Unfortunately, the size of this risk remains unknown at present because sophisticated pre-clinical safety models with which to measure DNA damage in vivo have not been available (1). To fulfil this medical need, we created a new and improved p21 reporter mouse line. The cell-cycle inhibitor protein p21, the product of the Cyclin-Dependent Kinase Inhibitor 1A (CDKN1A) gene, accumulates in vivo in response to many forms of DNA damage(11,12) and, as such, is an excellent in vivo biomarker. However, previous p21 reporter lines have limitations that make them unsuitable for this purpose. They either do not track p21 expression with sufficient fidelity(13) or they do not allow the resolution of expression down to individual tissues and cells (14). Our model overcomes these problems and allows us to monitor p21 expression, and DNA damage, with a high level of fidelity and resolution. Using this model, we found that a single exposure to a clinically-relevant dose of olaparib was sufficient to cause DNA damage in the large intestine of mice. We also show that olaparib exacerbates the damaging effects of IR in the large intestine and kidney. These data demonstrate that olaparib is genotoxic in mice with important possible implications for its clinical use in both disease prevention and cancer treatment. Materials and Methods Detailed descriptions of chemicals and y-irradiation treatments, olaparib pharmacokinetics, immunoblots, relative quantitation of mRNA species, and clinical chemical analyses can be found at SI Materials and Methods. p21 reporter mice For the generation of p21 reporter mice, a T2A-LacZ-loxP-T2A-Fluc-loxP cassette was inserted between the penultimate and STOP codons in exon 3 of p21. The LoxP sites were inserted in case it proved necessary to remove the luciferase reporter sequence. The positive selection marker (puromycin resistance (PuroR)) was flanked by FRT sites to allow for removal after the successful generation of transgenic mice. The targeting vector was generated using BAC clones from the C57BL/6J RPCIB-731 BAC library and electroporated into TaconicArtemis C57BL/6NTac ES cell line Art B6/3.6. Positive clones were verified by PCR and Southern blot before being injected into blastocysts from superovulated BALB/c mice. Blastocysts were injected into pseudopregnant NMRI females, and the chimerism of offspring was evaluated by coat color. Highly chimeric mice were bred with C57BL/6 females mutant for the gene encoding Flp recombinase (C57BL/6-Tg(CAG-Flpe)2 Arte). Germline transmission was identified by the presence of black C57BL/6 offspring (G1). The gene encoding Flp was removed by further breeding to Flp− partners after successful verification of PuroR removal. All animal work was carried out in accordance with the Animal Scientific Procedures Act (1986) and after local ethical review. All mice were kept under standard animal house conditions, with free access to food and water, and 12h light/12h dark cycle. Data in this paper was obtained using female mice of between 14 and 38 weeks of age. In any given experiment, mice were age-matched to within one month of each other. Mouse genotyping Ear biopsies of mice 4–8 weeks old were incubated at 50 °C for 4–5 h in lysis buffer containing 75 mM NaCl, 25 mM EDTA, 1% (w/v) SDS and 100 μg/ml (39 U/mg) proteinase K (Sigma). The concentration of NaCl in the reaction was raised to 0.6 M and a chloroform extraction was performed. Two volumes of isopropyl alcohol were added to the extracted supernatant to precipitate genomic DNA (gDNA). 40 μl TE buffer (10 mM Tris, 1 mM EDTA, pH 8.0) was added to the pellet, and subsequently gDNA was dissolved overnight at 37 °C. The typical PCR sample consisted of a 25-μl volume containing 10 pmol of the primers (4155_54_for 5'-GCTACTTGTGCTGTTTGCACC-3'; 4155_55_rev 5'-TCAAGGCTTTAGGTTCAAGTACC-3'). Each reaction also contained 1.25 U Taq DNA polymerase (Thermo-Scientific) with 10 mM dNTPs, buffer and 25 mM MgCl2. The following PCR conditions were applied: 5 min, 95 °C initial denaturation; 30 s, 95 °C cyclic denaturation; 30 s, 60 °C cyclic annealing; 1 min, 72 °C cyclic elongation for a total of 35 cycles, followed by a 10-min 72 °C elongation step. All PCR protocols were developed by TaconicArtemis. PCR amplification products were analyzed by agarose gel electrophoresis. In vivo luciferase imaging Imaging was performed at baseline and after chemical-dosing and/or irradiation. Reporter mice were injected ip with 5 µl/g body weight RediJect d-Luciferin (30 mg/ml, Caliper) and anaesthetized by isofluorane before being transferred into the IVIS Lumina II imaging chamber (Caliper) for bioluminescent imaging. Luminescent images (5 sec, f/stop 1.2, binning 4) and gray-scale images (2 sec, f/stop 16, binning 2) were acquired. Photon fluxes in Regions Of Interest (ROI) were quantified using the LivingImage (R) Software, version 4.3.1 (Caliper). ROI were defined as an oval from just below the forepaws to just above the genitals of each mice. Photon fluxes are expressed as photons/sec/cm2/sr. Luminescent images were rendered using the fire LUT in ImageJ, and superimposed on gray-scale photographs. Tissue harvesting and processing for cryo-sectioning Mice were euthanized by exposure to rising concentrations of CO2. The median lobe of the liver was fixed in 10% Neutral Buffered Formalin. The proximal 2-4 cm of duodena were fixed in 4% (w/v) paraformaldehyde. All other organs, including sagittally-cut kidney, the stomach, the proximal 2 cm of the large intestine, the thymus, the spleen, the left lung lobe, the heart and the brain were fixed in Mirsky’s fixative (National Diagnostics). Tissues were stored at 4°C, formalin-fixed ones for 4 hours, Mirsky-fixed ones overnight, before being transferred into 30% (w/v) sucrose for 24 hours. Embedding was carried out in Shandon M-1 Embedding Matrix (Thermo Scientific) in a dry ice-isopentane bath. Sectioning was performed on an OFT5000 cryostat (Bright Instrument Co). With the exception of lung and brain sections, all sections were cut at 10 μm thickness with a chamber temperature of -20°C. Lung sections were cut at 12 μm thickness with a chamber temperature of -23°C. Brain sections were cut at 20 μm thickness with a chamber temperature of -23°C. In situ β-gal staining Sections were thawed at room temperature and rehydrated in PBS supplemented with 2 mM MgCl2 for 5 minutes before being incubated overnight at 37°C in X-gal staining solution (PBS (pH 7.4) containing 2 mM MgCl2, 0.01% (w/v) sodium deoxycholate, 0.02% (v/v) Igepal CA630, 5 mM potassium ferricyanide, 5 mM potassium ferrocyanide and 1 mg/ml 5-bromo-4-chloro-3-indolyl β-D-galactopyranosidise). On the following day, slides were washed in PBS, counterstained in Nuclear Fast Red (Vector Laboratories) for 5 minutes, washed twice in distilled water and dehydrated through 70% and 95% ethanol before being incubated in Histoclear (VWR) for 3 minutes, air-dried and mounted in DPX mountant (Sigma). Sections were evaluated by an expert, independent clinical pathologist (Dr Shaun Walsh, Department of Pathology, Ninewells Hospital, Dundee). Dr Walsh was blinded in the case of sections from olaparib-treated mice (Fig 5). Brain sections were additionally examined by a neuroscientist (Dr John Sharkey, University of Dundee). Images were captured using a Zeiss 12 megapixel digital camera attached to a Zeiss Axio Scope.A1 microscope, and controlled by means of the AxioVision v4.5 software (Carl Zeiss). Images were rendered in ImageJ. Tissue harvesting and processing for γ-H2AX staining Mice were euthanized by exposure to rising concentrations of CO2. Organs were fixed in Gurr buffer overnight before transfer into 70% (v/v) ethanol. The following day, organs were dehydrated and embedded in paraffin and subsequently sectioned at 5 μm thickness using a Shandon Finesse 325 microtome. γ-H2AX staining Sections were deparaffinised in xylene and rehydrated through a graded series of 100 – 50% alcohol solutions, followed by immersion in MilliQ water. Antigen retrieval was performed using 0.01 M citrate buffer, pH 6.0. Sections were blocked in a 4% (w/v) solution of BSA in 1x TBS-T (0.01 M Tris, 150 mM NaCl, and 0.1% (v/v) Tween-20) and then incubated with 1:250 dilution of rabbit anti-γH2AX (Cell Signalling) in blocking buffer overnight at 4°C. After rinsing in 1x TBS-T, primary antibodies were detected with a 10 μg/ml solution of Alexa Fluor 488-conjugated goat anti-rabbit IgG in blocking buffer for 2 h. This solution also contained a 1:1000 dilution of the DNA counterstain DRAQ5 (Abcam). Finally, sections were mounted with Mowiol mounting medium (0.1 M Tris, pH 8.5, 10% (w/v) Mowiol 4-88, 25% (v/v) glycerol and 10 μg/ml 1,4-diazabicyclo(2,2,2)octane (DABCO)). Images were captured by confocal microscopy using the Leica S5 microscope equipped with a 40x PlanApochromat objective. Images were rendered in ImageJ. Results A new model for monitoring p21 expression at single-cell resolution We generated a mouse line in which an open-reading frame encoding T2A-β-gal-T2A-luciferase was inserted between the penultimate amino acid-encoding codon and the STOP codon in exon 3 of the p21 gene (Fig 1A). T2A peptides promote a phenomenon known as ribosome skipping that allows multiple proteins to be expressed from a single mRNA(15). We therefore anticipated that three separate polypeptides, p21, β-gal, and luciferase would be expressed from this engineered allele. The β-gal protein was tagged with a nuclear localisation sequence to target it to the nucleus. Moreover, because p21 expressed from the reporter allele contains the T2A amino acid sequence at its C-terminus, it can be distinguished from p21 expressed from the wild type (wt) allele based on electrophoretic mobility. Reporter mice were fertile and produced offspring at the expected Mendelian frequencies. To establish whether three separate peptides were being correctly processed and expressed from the reporter transcript, liver, lung and large intestine lysates from irradiated mice were immunoblotted for p21, β-gal and luciferase proteins (Fig 1B). Samples of all three possible genotypes were run to confirm antibody specificity. Two bands of the predicted molecular weights for the p21 expressed from the wt allele and the T2A-tagged p21 expressed from the reporter allele were observed by western blotting (Fig 1B). Surprisingly, the level of p21 expression off the reporter allele was significantly lower than off the wt allele. Both β-gal and luciferase were detected as single proteins with higher expression levels in animals homozygous for the reporter allele. These data demonstrated that the 2A strategy for the expression of multiple reporters off a single allele worked well. However, because of the reduced expression of p21 from the reporter allele, heterozygous mice were used for all subsequent experiments, in order to minimise any unanticipated and undesirable perturbations of the DNA damage signalling. In fact, the reduced expression of p21 in the reporter line does not seem to interfere with normal responses to DNA damage as heterozygous mice express approximately half the amount of native p21 protein found in wild-type 6 h after IR treatment (Fig 1B). Given that the heterozygous reporter allele only contains one native p21 allele rather than two, this is the result we should expect if the p21 response to IR is similar in wild-type and heterozygous mice. Similarly, equivalent levels of p53 protein are found in mice of all three genotypes 6 h after exposure to IR (Fig 1B). B-gal is expressed in the same organs, tissues, and cell types as p21 An essential attribute of any reporter system is that the pattern of expression of the reporter proteins faithfully mimic the expression of the endogenous gene. We anticipated that this would be the case in our model, as all the coding and regulatory elements that might control p21 expression are retained in the reporter allele. To confirm this, we exposed reporter mice to 4 Gy of IR and analysed expression of p21, p53, β-gal and luciferase at different time-points post-exposure (Fig 1C). Reporter activity was induced by IR concomitantly with p21 in all tissues examined. Alongside the induction of p21, a rapid and substantial accumulation of p53 protein was observed in the large intestine – but not the liver – of irradiated mice (Fig 1C), as previously described (16). We also observed a modest accumulation of p53 in lung lysates (Fig 1C). Notably, the extent of induction of p21 protein and both reporters (Fig 1C & Fig S2) exceeded that of p21 mRNA (Fig 1C), consistent with expression of p21 being determined by post-transcriptional as well as transcriptional mechanisms (11). Increased levels of reporter protein persisted longer than p21 itself. This is explained by the short p21 half-life of approximately 90 min (11) relative to luciferase (3-4 h) (17) and β-gal (>24 h) (18). As a more stringent test of the fidelity of the reporters we measured β-gal activity at single-cell resolution across an extensive panel of organs harvested from control or irradiated mice, and compared these data with previously reported in vivo patterns of p21 expression. In view of the extent of this analysis, only representative images of β-gal staining are shown (Fig 2, S3 – S7 and SI Appendix.) These analyses showed that reporter expression patterns were the same as the current relatively limited information on p21 expression patterns in mouse tissues (see SI Appendix). Moreover, Fig S8 displays a high-resolution image of a liver section from an irradiated mouse that demonstrates both the single-cell resolution achievable with this technology and the nuclear localization of the β-galactosidase protein. The p21 reporter is sensitive to multiple forms of DNA damage prior to overt organ toxicity The data above demonstrate that the new mouse model is a bone fide p21 reporter and that it responds to DNA double-strand breaks caused by IR. Next, we confirmed that the model also responds to other types of DNA damage. We treated reporter mice with 0-, 1-, 3- or 10 mg/kg of cisplatin, a chemical that causes primarily DNA inter- and intra-strand crosslinks. Significant increases in the β-gal signal were observed 24 h later only at the 10 mg/kg dose of cisplatin (Fig 3 and data not shown). The pattern of induction was markedly dissimilar to that of IR. Whereas the major effect of IR, whose dose-limiting toxicity is gastro-intestinal syndrome(19), was observed in the GI tract (Fig 2), the major induction of p21 by cisplatin, which is primarily nephrotoxic(20), was observed in the kidney. Also, unlike IR, the effects of cisplatin on p21 expression in the lung were much less marked. Both agents do however elicit significant β-gal activity in hepatocytes. Importantly, the induction of reporter activity preceded overt organ damage, as clinical markers of kidney- (urea) and liver damage (alanine- and aspartate aminotransferases (Fig S9)) were not increased in these experiments. Mice were also treated with etoposide, a topoisomerase II poison that triggers DNA double-strand breaks (21). In this case, increased reporter activity was observed predominantly in the large- and small intestines with little effect in the kidney, lung, or liver (Fig S10). This pattern of induction is consistent with clinical experience that mucositis is one of the two major dose-limiting toxicities of this drug (22). These data collectively demonstrate that these reporter mice provide an early and sensitive measure of DNA damage/toxicity in vivo. In vivo bioluminescent detection of p21 activation We confirmed that the luciferase reporter facilitated the non-invasive luminescent imaging of p21 induction/DNA damage. For example, we successfully measured basal as well as dose-dependent, IR-inducible luminescence in living mice (Fig 4A & B – see Fig S11 for expanded images of mice). However, the majority of the signal in both control and treated animals emanated from the skin. This was evidenced by the intense luminescence from the tail, face, paws, and genitals of the mice. In addition to these hairless ‘hotspots’, areas of high luminescence were also observed on the torso. However, these areas did not correspond to any underlying visceral tissues but were found, on closer examination, to coincide with areas where a knap of the mouse’s fur exposed the underlying skin. Although visceral organs do luminesce intensely, as evidenced by opening up the abdominal cavity (Fig S12), this cannot be observed non-invasively as these signals are attenuated by adsorption by overlying tissue and masked by the skin luminescence. Olaparib causes DNA damage alone and in combination with IR Having validated and characterized the reporter line, we next determined whether the PARP inhibitor olaparib induces DNA damage in mouse tissues when dosed alone or in combination with IR. Pilot experiments were performed to establish the pharmacokinetic properties of olaparib so that doses were chosen to reflect the level of drug exposure in patients (Figs 5A & S13). The Cmax values for olaparib vary substantially between cancer patients, with a mean value of approximately 10 μg/ml (6). The pharmacokinetic data suggested that a dose of 75 mg/kg olaparib would be required to attain a similar level of exposure, and was used for subsequent experiments. Reporter mice were then treated with olaparib or vehicle 30 min prior to exposure to 0-, or a low dose of 1 Gy of IR. Subsequent staining for β-gal activity in liver, lungs, small intestine and brain showed that olaparib alone had no effect on p21 expression in these tissues (data not shown). In contrast, mice which had been exposed to one single dose of inhibitor had elevated levels of β-gal activity in the large intestine, primarily in the apical, terminally-differentiated epithelial cells (Fig 5B), suggesting an increased level of DNA damage in this tissue. Olaparib pre-treatment also significantly exacerbated subsequent IR-induced DNA damage in both large intestine (Fig 5B) and kidney (Fig 5C). Surprisingly, when these experiments were repeated with cisplatin (3 mg/ml) in place of IR, no potentiation of DNA damage was observed in the kidney, or in the large- and small intestines (data not shown). In order to confirm the reporter mouse studies that olaparib alone caused DNA damage in the large intestine, we exposed wt mice to this compound and stained large intestines for γ-H2AX, an established marker of DNA damage. Significant γ-H2AX staining in cells lining the mucosa of the large intestine were observed 4h after olaparib treatment (Fig 6A), but not at any other time point (1-, 2-, 8-, or 24 h (data not shown)). The γ-H2AX staining was largely focal indicating that it arises from DNA damage signalling (Fig 6B). We also observed a minority of cells displaying a more intense and pan-nuclear stain, the significance of which is less clear (23) but which may indicate the presence of clustered DNA lesions (24). Discussion A new p21 reporter mouse model has been developed that permits the measurement of p21 expression non-invasively in live mice (luciferase reporter) and, post-mortem, in all tissues at single-cell resolution (β-gal reporter). This has been achieved by incorporating both reporters into the p21 gene locus to create a polycistronic mRNA where each protein is translated individually by flanking each coding region with foot and mouth virus 2A sequences. It should be noted, however, that the expression of p21 from this locus was significantly reduced relative to the wild type allele. The reason for this remains unclear. One possible explanation is that translation of the polycistronic reporter mRNA is hindered by the presence of multiple T2A sequences, which cause translational pausing(15). Alternatively, it is possible that the insertion of the reporter cassette might affect mRNA secondary structure/stability (25). We have demonstrated the versatility of the p21 reporter model by showing that it responds to both double-strand breaks (IR and etoposide) and inter- and intra-strand DNA crosslinks (cisplatin). Further work is ongoing to determine the full spectrum of DNA lesions to which it responds. The reporter is active in most organs examined (GI Tract, liver, lungs, kidneys, female reproductive tract) and in both proliferative (e.g large intestine crypt cells) and post-mitotic (e.g alveolar) cells. An advantage of our system is that, because β-gal is a stable protein, our reporter response is durable and lasts at least 24 h after damage, unlike classic markers of DNA damage, such as γ-H2AX, whose presence is often transient and easily missed. The reporter mice were created to enable DNA damage to be measured in vivo in response to drugs and chemicals. For this first study, we investigated the effects of olaparib primarily because of the current interest in combining it with other DNA damaging agents in the treatment of cancer (1), and also because it is being considered as an agent to be used prophylactically in prevention of breast and ovarian cancer in individuals carrying BRCA mutations (26). We discovered that a single exposure to a clinically relevant dose of olaparib causes DNA damage in the large intestine of mice. Additionally, it exacerbated DNA damage induced by IR treatment in this tissue and also the kidney. Both of these findings have significant implications for the clinical use of this drug. We also note that such detrimental effects may become more pronounced, and affect other tissues, after repeat chronic dosing with olaparib. Our finding that olaparib alone induced DNA damage in the colon suggests that the use of this drug prophylactically may increase the risk of developing colorectal cancer. This concern might seem to have been allayed by the recent report that mice chronically exposed to olaparib suffered no adverse effects (26). However, in these studies, the mice had an average lifespan of only 38 weeks due to a mammary-specific BRCA1 deletion that caused rapid breast tumour development. This lifespan may well have been too short to observe sporadic colorectal tumours, which generally present in aged mice(27). Based on our findings, the possibility that long-term exposure to olaparib is carcinogenic warrants further investigation. At present, it is unclear whether olaparib causes DNA damage in the gut because it inhibits PARP1/2 or whether it is an off-target effect; olaparib, has broad activity and inhibits multiple PARPs in addition to PARP1/2, including PARP3/4/12/15 and 16 (28). It may also target additional proteins unrelated to the PARP family (8). However, olaparib toxicity in haploid ES cells is due to PARP1 inhibition (29). If the damage that we observed is also due to an on-target effect, it is difficult to envisage olaparib, or PARP inhibitors more generally, being developed for purposes other than the treatment of cancer. Our finding that olaparib exacerbates IR-elicited DNA damage in normal tissues such as the large intestine and kidney, is also of significant potential clinical importance as it suggests that combination therapies involving olaparib and DNA damaging agents will increase toxicity to normal tissues. Initial clinical trial data support this contention with more adverse effects reported in patients treated with olaparib and chemotherapy than olaparib alone (30). The olaparib-induced radiosensitisation of normal tissues is particularly concerning in light of the fact that tumours are generally radiosensitised by olaparib by a factor of less than two (31). Interestingly, we observed no evidence of exacerbated DNA damage when olaparib and cisplatin were used in combination. A possible explanation is that the major DDR pathway involved in repair of cisplatin crosslinks is nucleotide excision repair (32). Unlike for the homologous recombination repair pathway involved in repairing IR-induced damage, there is no compelling evidence for a biochemical or genetic interaction between nucleotide excision repair and the PARP pathway. Therefore, the lack of an interaction between olaparib and cisplatin in our model may reflect the fact that these two drugs do not synergize together and that the rationale for combining them in the clinic is not well-founded (33). Indeed, the optimal PARP inhibitor-chemotherapy drug combination for cancer treatment remains to be determined. It will also be challenging to define the optimal doses and scheduling of combined drugs so as to minimize side-effects while maximizing efficacy. These issues could be addressed by pre-clinical studies using the p21 reporter model, especially if it were crossed onto genetically-engineered mouse models for specific cancers to permit anti-tumour efficacy to be measured alongside normal tissue toxicity. Finally, it is worthy of note that while our motivation in developing the p21 reporter mouse was to provide a tool for pre-clinical drug safety testing, accumulation of DNA damage and p21 protein is a symptom of many degenerative diseases associated with aging, such as Hutchinson-Gifford Progeria Syndrome and Parkinson’s disease, and indeed normal aging itself(34). For this reason, monitoring of p21 levels using our reporter might catalyse research into the basic science underpinning these devastating disorders and, as an efficacy biomarker, would expedite testing of drugs to alleviate their symptoms. Supplementary Material SI Figures 1-3 SI Figures 4-5 SI Figures 6-7 SI Figures 8-9 SI Figure 10 SI Figures 11-13 Supporting Information Acknowledgement We are grateful to Dr Shaun Walsh (Department of Pathology, Ninewells Hospital, Dundee) for advice in interpretation of tissue sections. 1 This work was funded by a European Research Council Advanced Investigator Award (number 294533) and a CRUK Programme grant (C4639/A10822), both awarded to CRW. Figure 1 Validation of the p21 reporter allele. A) The gene structures of the wild-type- and reporter p21 alleles. B) Immunoblots of pooled protein samples prepared from the liver, lung, and large intestine (LI) of groups of three wild-type (+/+), heterozygous reporter (+/r) or homozygous reporter (r/r) mice exposed to a single 4 Gy dose of IR for 6 h. Quantification of these blots can be found in Fig S1. C) Heterozygous reporter mice were exposed to a single-dose of 4 Gy of IR. At the indicated time-points post-treatment, protein and RNA were prepared from livers, lungs and large intestines (LI). The graphs depict relative p21 mRNA levels (x ± SD, n = 3 mice). Also shown are immunoblots of pooled protein samples from each group of three mice. Quantification of blot data can be found in Fig S2. Figure 2 β-gal staining of fore-and glandular stomachs, small intestine (SI) or large intestine (LI) from untreated and irradiated p21 reporter mice. Triplicate mice were sham-irradiated (-IR) or exposed to a single dose of 4 Gy IR (+IR) and sacrificed 24 h later. Representative images are shown. Scale-bar = 100 μm. ME, muscularis externae; SE, squamous epithelium; F, foveolar mucus-secreting cells; P, Parietal and chief cells; LP, lamina propria; MM, muscularis mucosae. Figure 3 β-gal staining of various organs from control and cisplatin-treated p21 reporter mice. Mice were treated with vehicle or 10 mg/kg cisplatin and sacrificed 24 h later. Representative images are shown. Scale-bar = 200 μm. Figure 4 Bioluminescent in vivo imaging of reporter mice following IR. A) Five heterozygous reporter mice were pre-imaged. Two of the mice were sham-irradiated whereas the remaining three were exposed to a single dose of 4 Gy of IR. Sham- (0 Gy) and irradiated mice (4 Gy) were re-imaged 6- and 24 h after exposure. Quantification of photon fluxes at the 24 h time-point is shown in the accompanying bar chart (x ± SD, n = 3 or 2 mice). B) 12 mice were in vivo imaged and randomly assigned to receive 0-, 1-, 2-, or 4 Gy of IR. Mice were in vivo imaged 6- and 24 h later (x ± SD, n = 3 mice). Figure 5 Effect of olaparib, alone and combined with IR, on p21 reporter activity. A) Heterozygous reporter mice were dosed with 50 mg/kg or 100 mg/kg olaparib and drug concentrations in plasma determined at various time-points thereafter (x ± SD, n = 3 mice). B & C) p21 reporter mice were treated in triplicate with 75 mg/kg olaparib or with vehicle 30 min prior to exposure to 0- or 1 Gy IR. Representative images of in situ β-gal activity in large intestines (B) and kidneys (C). Scale bar = 50 μm for large intestine micrographs. For kidney panoramic micrographs, scale bar = 1000 μm and 100 μm for all other kidney images. Figure 6 Olaparib causes DNA damage in the mucosal lining of the large intestine. A) The micrographs display γ-H2AX staining in the large intestines of three separate untreated wild-type mice (-olaparib), or three separate mice that were euthanized 4 h after treatment with 75 mg/kg olaparib. Scale-bar = 50 μm. B) A higher resolution image demonstrating the primarily focal nature of the γ-H2AX signal in large intestines of olaparib-treated wild-type mice. The arrow points to a cell displaying a more pan-nuclear γ-H2AX signal. Scale-bar = 10 μm. Disclosure of conflict of interest: The authors confirm that no potential conflict of interest exists. 1 Lord CJ Ashworth A The DNA damage response and cancer therapy Nature 2012 481 287 94 22258607 2 Ashworth A Lord CJJ Reis-Filho JSS Genetic interactions in cancer progression and treatment Cell 2011 145 30 8 Elsevier Inc. cited 2011 Apr 3 [Internet] 21458666 3 Farmer H McCabe N Lord CJ Tutt ANJ Johnson Da Richardson TB Targeting the DNA repair defect in BRCA mutant cells as a therapeutic strategy Nature 2005 434 917 21 15829967 4 Bryant HE Schultz N Thomas HD Parker KM Flower D Lopez E Specific killing of BRCA2-deficient tumours with inhibitors of poly(ADP-ribose) polymerase Nature 2005 434 913 7 15829966 5 Menear Ka Adcock C Boulter R Cockcroft X Copsey L Cranston A 4-[3-(4-cyclopropanecarbonylpiperazine-1-carbonyl)-4-fluorobenzyl]-2H-phthalazin-1-one: a novel bioavailable inhibitor of poly(ADP-ribose) polymerase-1 J Med Chem 2008 51 6581 91 [Internet] 18800822 6 Fong PC Boss DS Yap Ta Tutt A Wu P Mergui-Roelvink M Inhibition of Poly(ADP-Ribose) Polymerase in Tumors from BRCA Mutation Carriers N Engl J Med 2009 361 123 34 19553641 7 Deeks ED Olaparib: First Global Approval Drugs 2015 75 231 40 10.1007/s40265-015-0345-6 [Internet] 25616434 8 Ekblad T Camaioni E Schüler H Macchiarulo A PARP inhibitors: Polypharmacology versus selective inhibition FEBS J 2013 280 3563 75 23601167 9 Sandhu SK Yap Ta de Bono JS Poly(ADP-ribose) polymerase inhibitors in cancer treatment: a clinical perspective Eur J Cancer 2010 46 9 20 cited 2014 Apr 16 Elsevier Ltd [Internet] 19926276 10 Kapoor K Singla E Sahu B Naura AS PARP inhibitor, olaparib ameliorates acute lung and kidney injury upon intratracheal administration of LPS in mice Mol Cell Biochem 2014 400 153 62 [Internet] 10.1007/s11010-014-2271-4 25404465 11 Warfel Na El-Deiry WS p21WAF1 and tumourigenesis: 20 years after Curr Opin Oncol 2013 25 52 8 cited 2014 Apr 1 [Internet] 23159848 12 Grek CL Townsend DM Tew KD The impact of redox and thiol status on the bone marrow: Pharmacological intervention strategies Pharmacol Ther 2011 129 172 84 Elsevier Inc. [Internet] 10.1016/j.pharmthera.2010.09.008 20951732 13 Vasey DB Wolf CR MacArtney T Brown K Whitelaw CBa p21-LacZ reporter mice reflect p53-dependent toxic insult Toxicol Appl Pharmacol 2008 227 440 50 cited 2013 Apr 12 [Internet] 18215733 14 Tinkum KL Marpegan L White LS Sun J Herzog ED Piwnica-Worms D Bioluminescence imaging captures the expression and dynamics of endogenous p21 promoter activity in living mice and intact cells Mol Cell Biol 2011 31 3759 72 cited 2013 Apr 12 [Internet]. Available from: http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=3165732&tool=pmcentrez&rendertype=abstract 21791610 15 de Felipe P Luke Ga Hughes LE Gani D Halpin C Ryan MD E unum pluribus: multiple proteins from a self-processing polyprotein Trends Biotechnol 2006 24 68 75 cited 2013 Feb 28 [Internet] 16380176 16 Midgley CA Owens B Briscoe CV Thomas DB Lane DP H P Coupling between gamma irradiation, p53 induction and the apoptotic response depends upon cell type in vivo J Cell Sci 1995 108 1843 8 7657708 17 Leclerc G Boockfor F Faught W Frawley L Development of a destabilized firefly luciferase enzyme for measurement of gene expression Biotechniques 2000 29 590 1 10997273 18 Stack J Whitney M Rodems S Pollok B A ubiquitin-based tagging system for controlled modulation of protein stability Nat Biotechnol 2000 18 1298 302 11101811 19 Kirsch DG Santiago PM di Tomaso E Sullivan JM Hou W-S Dayton T P53 Controls Radiation-Induced Gastrointestinal Syndrome in Mice Independent of Apoptosis Science 2010 327 593 6 20019247 20 Dhar S Kolishetti N Lippard SJ Farokhzad OC Targeted delivery of a cisplatin prodrug for safer and more effective prostate cancer therapy in vivo Proc Natl Acad Sci U S A 2011 108 2 7 21 Pommier Y Leo E Zhang H Marchand C DNA Topoisomerases and Their Poisoning by Anticancer and Antibacterial Drugs Chem Biol 2010 17 421 33 Elsevier Ltd [Internet]. Available from: http://linkinghub.elsevier.com/retrieve/pii/S1074552110001614 20534341 22 Hainsworth JD Greco FA Etoposide: Twenty years later 1995 23 Solier S Pommier Y The nuclear γ-H2AX apoptotic ring: Implications for cancers and autoimmune diseases Cell Mol Life Sci 2014 71 2289 97 24448903 24 Meyer B Voss KO Tobias F Jakob B Durante M Taucher-Scholz G Clustered DNA damage induces pan-nuclear H2AX phosphorylation mediated by ATM and DNA-PK Nucleic Acids Res 2013 41 6109 18 23620287 25 Pessi G Blumer B Haas D lacZ fusions report gene expression, don’t they? Microbiology 2000 147 1993 5 26 To C Kim EH Royce DB Williams CR Collins RM Risingsong R The PARP inhibitors, veliparib and olaparib, are effective chemopreventive agents for delaying mammary tumor development in BRCA1-deficient mice Cancer Prev Res 2014 7 698 707 27 Fleet JC Animal models of gastrointestinal and liver diseases. New mouse models for studying dietary prevention of colorectal cancer AJP Gastrointest Liver Physiol 2014 307 G249 59 [Internet] 10.1152/ajpgi.00019.2014 28 Wahlberg E Karlberg T Kouznetsova E Markova N Macchiarulo A Thorsell A-G Family-wide chemical profiling and structural analysis of PARP and tankyrase inhibitors Nat Biotechnol 2012 30 283 8 22343925 29 Pettitt SJ Rehman FL Bajrami I Brough R Wallberg F Kozarewa I A genetic screen using the PiggyBac transposon in haploid cells identifies Parp1 as a mediator of olaparib toxicity PLoS One 2013 8 e61520 cited 2014 Apr 16 [Internet]. Available from: http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=3636235&tool=pmcentrez&rendertype=abstract 23634208 30 Oza AM Cibula D Benzaquen AO Poole C Mathijssen RHJ Sonke GS Olaparib combined with chemotherapy for recurrent platinum-sensitive ovarian cancer: a randomised phase 2 trial Lancet Oncol 2015 16 87 97 [Internet]. Available from: http://linkinghub.elsevier.com/retrieve/pii/S1470204514711350 25481791 31 Powell C Mikropoulos C Kaye SB Nutting CM Bhide SA Newbold K Pre-clinical and clinical evaluation of PARP inhibitors as tumour-specific radiosensitisers Cancer Treat Rev 2010 36 566 75 Elsevier Ltd [Internet] 10.1016/j.ctrv.2010.03.003 20409643 32 Martin LP Hamilton TC Schilder RJ Platinum Resistance: The Role of DNA Repair Pathways Clin Cancer Res 2008 14 1291 5 [Internet] 10.1158/1078-0432.CCR-07-2238 18316546 33 Murai J Zhang Y Morris J Ji J Takeda S Doroshow JH Rationale for poly(ADP-ribose) polymerase (PARP) inhibitors in combination therapy with camptothecins or temozolomide based on PARP trapping versus catalytic inhibition J Pharmacol Exp Ther 2014 349 408 16 [Internet] 24650937 34 Garinis GA Van Der Horst GTJ Vijg J Hoeijmakers JHJ DNA damage and ageing : new-age ideas for an age-old problem Nat Cell Biol 2008 10 1241 7 18978832

PMC005xxxxxx/PMC5136504.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 101084138 22395 Infect Genet Evol Infect. Genet. Evol. Infection, genetics and evolution : journal of molecular epidemiology and evolutionary genetics in infectious diseases 1567-1348 1567-7257 27789390 5136504 10.1016/j.meegid.2016.10.018 NIHMS828868 Article Primate Immunodeficiency Virus Classification and Nomenclature: Review Foley Brian T. 1* Leitner Thomas 1 Paraskevis Dimitrios 2 Peeters Martine 34 1 Theoretical Biology and Biophysics Group, T-6 Mail Stop K710, Los Alamos National Laboratory, Los Alamos, NM 87545 USA 2 National and Kapodistrian University of Athens Department of Hygiene,Epidemiology and Medical Statistics,Medical School, Athens, Greece 3 UMI233-TransVIHMI, Institut de Recherche pour le Developpement (IRD), INSERM U1175, University of Montpellier, Montpellier, France 4 IBC, Computational Biology Institute, 34095 Montpellier, France * Corresponding author btf@lanl.gov (505) 665-1970 11 11 2016 24 10 2016 12 2016 01 12 2017 46 150158 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. The International Committee for the Taxonomy and Nomenclature of Viruses does not rule on virus classifications below the species level. The definition of species for viruses cannot be clearly defined for all types of viruses. The complex and interesting epidemiology of Human Immunodeficiency Viruses demands a detailed and informative nomenclature system, while at the same time it presents challenges such that many of the rules need to be flexibly applied or modified over time. This review outlines the nomenclature system for primate lentiviruses and provides an update on new findings since the last review was written in 2000. Immunodeficiency HIV Lentivirus Nomenclature Classification Introduction 1.1 Primate Immunodeficiency Virus nomenclature history The first complete genome of a human immunodeficiency virus, named HTLV-III at that time, was sequenced in 1984(Ratner et al., 1985). By 1992 it was clear that there was considerable diversity between HIV-1 M group viruses, that HIV-2 was also infecting humans, that chimpanzees carried viruses similar to HIV-1(Huet et al., 1990), and that sooty mangabeys carried viruses similar to HIV-2(Sanchez-Pescador et al., 1985)(Myers et al., 1992)(Hirsch et al., 1989). In subsequent years analyses of tens of thousands of virus genomic sequences have provided great insight into the origins and spread of these viruses. The rules for naming viruses, established and overseen by the International Committee for the Taxonomy and Classification of Viruses (ICTV), only go to the species level, with nomenclatures of subspecies, strains, isolates, serotypes and similar subspecies designations left up to the research community that studies each virus(Fauquet and Fargette, 2005). The HIV research community worked on names informally until a formal meeting was held in September 1999 (Robertson et al., 2000), where a set of rules for naming primate lentiviruses (HIVs and SIVs) was agreed upon. 1.2 Lentiviruses and Primate Immunodeficiency Viruses The broad picture of lentiviruses is that they have only been detected in a small subset of mammals, and that the phylogeny of the viruses is similar enough to the phylogeny of the host species to suggest a very long time of co-evolution of these viruses with their hosts. However, there are many exceptions where the virus phylogeny disagrees with host phylogeny due to recent and/or ancient cross species transfer events, and also due to ancient or recent recombination events between different virus lineages. The non-primate lentiviruses include bovine immunodeficiency viruses found in both domestic cattle and wild buffalo, Equine Infectious Anemia Virus, Small Ruminant Lentiviruses of goats and sheep, and feline immunodeficiency viruses. The feline immunodeficiency viruses have been extensively studied and have been found in domestic housecats, New World felines including pumas, bobcats and lynx, and Old World felines including lions, leopards and Pallas’ cats. Outside the primate immunodeficiency viruses (PIVs), less effort has been made to standardize the classification and nomenclature, but in cases where there has been some effort made, it generally follows a similar set of criteria as have been used for the PIVs. This review will only cover the nomenclature of PIVs, with a special emphasis on HIV-1. 1.3 PIV species specificity, coevolution For the most part, the lentiviruses do not easily cross species boundaries, and there are several virus-host protein interactions which either partially or completely block the ability of lentivirus from one host to infect another host species. Because of this, it makes sense to name each lineage of simian lentivirus for the host it infects, such as SIV_Gorilla (SIVgor) for simian immunodeficiency viruses isolated from gorillas. Short sequence or isolate names are nice for labeling phylogenetic trees and other purposes, so the convention is to use a 3-letter code for the primate species from which the virus was isolated. The exact line between species and subspecies of primates, as with most other organisms, is in some cases difficult to define. For some groups, the geographic range is more important than genetic distances, while for others the genetics is more important. For example, zoo keepers have reported that Pan troglodytes schweinfurthi, Pan troglodytes verus and Pan troglodytes troglodytes can produce fertile offspring with Pan paniscus and thus the species (paniscus vs troglodytes) and subspecies of chimpanzees are better defined by their geographic ranges. A similar situation exists for baboons and African green monkeys. In other cases such as the greater spot-nosed monkey (Cercopithecus nictitans) and lesser spot-nosed monkey (Cercopithecus petaurista) the species ranges overlap but mating preferences and genetic distances justify separate species designations(Tosi et al., 2005). The three letter code chosen to represent each primate species or subspecies from which a PIV has been isolated is usually derived from the common name (cpz for chimpanzee, for example) but sometimes derived from the Latin name, ex. SIVasc from Cercopithecus ascanius (red tailed guenon). In some cases a single species or subspecies of primate can harbor two or more very distinct lineages of SIV, in which case sequential numbers are added to the name (HIV-1 and HIV-2, SIVmnd-1, SIVmnd-2). Table 1 is a comprehensive list of primate species which have been tested by serology or virus sequencing. The HIV Databases at LANL maintain a listing of codes used for the sequenced viruses here: http://www.hiv.lanl.gov/content/sequence/HelpDocs/subtypes.html 1.4 Human Immunodeficiency Viruses There have been several transfers of nonhuman primate immunodeficiency viruses into humans, from chimpanzees, gorillas and sooty mangabeys. Viruses in the HIV-1 M and N groups are most closely related to viruses found in wild chimpanzees of the Pan troglodytes troglodytes subspecies from south eastern Cameroon(Keele et al., 2006, Van Heuverswyn et al 2007). HIV-1 O and P group viruses are most closely related to viruses found in western lowland gorillas in Cameroon(D’arc et al., 2015). Each group of HIV-1 viruses seems to be the result of a separate cross-species transmission event from chimpanzee or gorilla into humans, followed by different degrees of epidemic spread within humans. The HIV-1 M group spread far more extensively in humans than did N, O, or P group viruses. The AIDS pandemic is primarily caused by viruses in HIV-1 M group. HIV-2 viruses similarly have been introduced into humans several times, from sooty mangabeys in West Africa, of which only two transfers resulted in significant spread within humans in West Africa. The viruses from each of the sooty mangabey to human transfers were initially named as “subtypes” of HIV-2 and the subtype (A, B, C) designations remain in use today even though they are equivalent to the “groups” of HIV-1, and not equivalent to the subtypes which have been defined within the HIV-1 M group (http://www.hiv.lanl.gov/content/sequence/HelpDocs/subtypes-more.html accessed May 26, 2016). The HIV-2 clades should be referred to as “groups” and not “subtypes”. 2 HIV-1, SIVcpz, SIVgor Because the AIDS pandemic is primarily caused by viruses in the HIV-1 M group, far more research has been conducted on this group of viruses. The global spread of HIV-1 M group viruses has also resulted in a much more complex and interesting epidemiological history, than those of the other HIV-1 groups or HIV-2. Likewise the extra interest in this group has also resulted in increased searching and sampling of chimpanzees and gorillas aimed at tracing the origins of the HIV-1 viruses and a better understanding of the natural history of these viruses than any of the other primate lentiviruses (Keele 2006 (D’arc 2015) (Li 2012). It is now clear that the HIV-1 M and N groups came from chimpanzees in Cameroon, while the HIV-1 O and P groups came from western lowland gorillas in Cameroon (D’arc et al., 2015; Keele et al., 2006; Li et al., 2012; Sharp and Hahn, 2010). The diversity of HIV-1 O group viruses in humans is nearly equivalent to the diversity of HIV-1 M group viruses, and one of the earliest proven AIDS cases in humans was with O group virus infecting a Norwegian sailor and his family in 1970(Wertheim and Worobey, 2009) (Froeland 1988). Despite apparently similar timing of the origins, the O group has not spread around the world in humans and evolved into subtypes, while the HIV-1 M group has. It is possible that the reason for the limited spread of group O versus M was due to stochastics associated with the ignition and generation of HIV-1 epidemic at the early stage(Faria et al., 2014)(Hogan et al., 2016). 2.1 HIV-1 M group 2.1.1 Subtypes, Subsubtypes and local outbreak clusters Very early in the study of HIV and AIDS it was noted that there was very great diversity between viruses isolated from patients sampled in the developed world, and patients sampled in Africa (Smith 1988). Although the very first HIV-1 isolates to be cloned and sequenced were from the developed world, they were named “subtype B” in the 1992 HIV Database Compendium (Myers 1992). The 1992 Database Compendium named subtypes A through D for gag and A though E for envelope sequences, noting that subtype E viruses were classified as A in the gag gene. Although this publication established the subtype nomenclature for the HIV-1 M group, there is no discussion of how the letters were assigned to each clade. Researchers studying serological reactions to various isolates in human infections and in animals experimentally infected with cloned isolates had been using the term “subtype” for at least 4 years (Cheng-Mayer 1988)(Neurath 1990) but those serological groups do not all coincide with the genotypes. By 1994 subtype F had been detected among orphans infected in an outbreak in Romania, and in Brazil more subtype F, as well as B/F recombinant virus was found (Cernescu et al., 1994)(Morgado et al., 1994). Subtype G was also described in 1994 (Bobkov 1994). Between 1992 and 2000 subtypes A though K were established and many circulating recombinant forms were described (Triques et al., 2000). Although there have been a few nearly complete genomes of HIV-1 M group viruses sequenced since 2000 which are unique from the established subtypes, and thus hint that other subtypes of HIV-1 exist, in no case have 3 or more of the same clade been identified to qualify for naming a new subtype. Conversely, a set of related viruses which was initially designated as subtype I was renamed as a circulating recombinant form because several regions of its genome fell within the established A–H clades even though other regions were unique enough to warrant a new subtype designation (Gao et al., 1998; Kostrikis et al., 1995)(Paraskevis et al., 2001)(Nasioulas et al., 1999). Indeed these unique regions were later found to be more closely related to subytpes J and K once those subtypes were described (Montavon et al., 2002). Although it is not completely incorrect to refer to the unique regions in CRF01_AE as “E” or the unique regions in CRF04_cpx as “I”, it is highly recommended to refer to these lineages, and subgenomic regions derived from these lineages, as CRF01_AE and CRF04_cpx. Differences between the subtypes of the HIV-1 M group include many other attributes in addition to the DNA and amino acid sequence diversity and distances. Some subtypes account for millions of infections in dozens of countries while other subtypes have a very limited epidemiology. Some of the subtypes apparently began spreading earlier than others and are thus more diverse, while others followed different epidemiological trajectories. Subtype A is very diverse and six major subclades have been designated subsubtypes A1 through A6(Vidal et al., 2009, 2006). Subsubtype A1 spread considerably across Africa especially in eastern Africa and A2 to some extent in central Africa. The A3, A4 and A5 subsubtypes each have few sequences described and no pure A5 strains have been identified, A5 is only found in large regions of the CRF26_A5U genomes (Meloni et al., 2004)(Vidal et al., 2006)(Vidal et al., 2009). Another group of divergent subtype A viruses, labeled as subsubtype A6 in figure 3, has spread quite widely among IDU in the regions of the former Soviet Union and surrounding areas(Bobkov et al., 1994; Lapovok et al., 2014). This group of viruses have been also named as AFSU corresponding the geographic area (former Soviet Union countries) of their endemicity. Subtype B is closely related to subtype D, suggesting that these clades share a more recent common ancestor than the other subtypes. However subtype B was defined and named as its own subtype long before the addition of subsubtypes to the nomenclature scheme and is grandfathered in as a separate subtype. A similar relationship, although deeper, exists between subtypes F and K. The assignment of subsubtypes and of other subdivisions within a subtype creates some difficult issues. Soon after HIV-1 subtype B was detected in Thailand in the 1990s, it was noted that the Thai epidemic seemed to be spreading from one introduction of virus, and this became known as the B’ (B-prime) or Thai-B strain(Kalish et al., 1995). Numerous other such founder effects have resulted in similar subclades or local epidemic strains in other parts of the world. It is acceptable and indeed good practice to give labels to any subclades or local strains in order to clearly convey concepts about the epidemiology of a subset of viruses. However, it is obvious that there can be no clear definitions put in place to standardize the nomenclature of these lineages. It would be useful to name these subclades according to the geographic area(s) and/or risk groups where they are endemic. The Subtyping Distance tool (SUDI; http://www.hiv.lanl.gov/content/sequence/SUDI/sudi.html) and the PhyloPlace tool (http://www.hiv.lanl.gov/content/sequence/phyloplace/PhyloPlace.html) were created in order to assist researchers in determining whether or not a group of related sequences represents a true “sub-subtype”, a new subtype, or just some variants of an already labeled subtype. The HIV and SIV nomenclature committee has not defined a cut-off rule on the branching index value for determining when a single virus isolate sequence or group of sequences should be named as a subtype or variant (Fig 4). However, a subtype specific branching index evaluation was reported in Hraber et al (Hraber et al., 2008). While an overall cut-off at 0.66 can be used as a rule of thumb, the cut-off for certain subtype associations varies across the genome and between subtypes. Applying a universal cut-off is complicated by the inclusion of new subtype reference sequences, and thus relies on a fixed set of reference sequences for meaningful results. For instance, we may have a small region of a complex recombinant genome that is “H-like” or falls close to the “crown group” of subtype H sequences that were chosen as the subtype H reference set of sequences, in phylogenetic analysis. If the branching index is greater than 0.66 it makes a lot of sense to label this small region of the genome as “H” or “H-like” rather than labeling it as “U” or “undetermined” subtype. As another example, we may have a group of related viruses such as the virus isolates now labeled as subtype K and we note that in some regions of the genome they are F-like and in other regions of the genome they are C-like. In this case we are concerned with whether we should name the viruses as a new subtype (subtype K) or name them as a circulating recombinant form. Even though at least one region of the subtype K genomes fall in between the F1 and F2 subsubtypes, in no place do the subtype K genomes have a branching index of more than 0.66 for either F1 or F2, so they were named as subtype K. An important point in this decision, is that the genomic region where subtype K sequences fall in between F1 and F2 is a region where F1 and F2 are very diverse from one another, a region of low signal to noise ratio. Historically different clades have been named as subtypes based on partial genomic analysis. Moreover, at the early time of subtype assignment, currently existing tools and nomenclature proposal were not available. Thus some of the existing classification rules are not fulfilled by some subtypes (e.g. subtypes D and K). No matter the previous issues, our recommendation for current assignments is that they should be according to the criteria specified in the existing nomenclature proposal. Isolates of HIV-1 M group virus which are not related to any one of the defined subtypes or circulating forms are labeled as “U” for unique or unsubtyped. Because the HIV-1 M group pandemic began with a Chimpanzee cross-species transmission to human in central Africa it is rather common to find such unique viruses in or near central Africa(Mokili et al., 2002). To date, the sampling and sequencing of nearly complete genomes of HIV-1 isolated from central Africa has been rather sparse in comparison to the number of infections there, and this contributes to a lack of identifying more subtypes. A recent paper studying published genomes suggests that more subtypes may be defined in the future(Tongo et al., 2015). The ongoing PANGEA_HIV project (http://www.pangea-hiv.org/) may also make contributions in this area(Pillay et al., 2015). In recent years several research groups have named local subclades of CRF01_AE viruses found in various cities, or risk groups within cities in China, with sequential numbers(Zeng et al., 2016)(Feng et al., 2013). It is difficult to ascertain whether the authors of different papers are using the same numbering system such that “cluster 1” in one paper is the same local epidemic as “clade 1” in another publication. The papers most often do not provide a table clearly linking each virus isolate to each cluster or subclade, and the GenBank entries also lack clear annotation, so it is very difficult to determine which sequences belong to each subclade. Most of the analyses are based on small genomic regions rather than complete genomes in order to obtain enough isolate sequences to find the clusters. Using those same genomic regions and subtype B from the USA, similar clusters or subclades are notable, although they are not always geographically ordered. Another difficulty with this fine classification within a subtype or CRF, is that the geographical location is often not specified below the country level, so it is not possible to search GenBank or the HIV database for, for example, “New York” or “California” isolate sequences. Although it can be useful to give names to local clusters in order to discuss them in a paper, no expectations can be made for those names to be useful beyond the one study being published. Authors of papers describing sequences are encouraged to provide annotation in GenBank that clearly identifies the virus sample date, genotype, location and any other pertinent information. 2.1.2 Circulating Recombinant Forms Whenever an individual is infected by more than one HIV-1 subtype, whether it is a dual infection before an effective immune response to the first virus is mounted, or a superinfection after the first virus has been partially suppressed by the host immune response, recombinants between the two virus lineages may soon be formed. The recombination is not due to strand breakage and re-joining as in most dsDNA recombination events. It is due to template switching during the reverse transcription process. Two copies of the viral RNA genome are packaged in each virus particle, so a single cell must be dually infected such that one copy of each of the two genomes is packaged, and the reverse transcriptase switches between the two templates during the reverse transcription process. This process leads to a recombination rate on par with the high substitution rate (Neher and Leitner, 2010) Quite early in the studies of complete genome sequences from isolates of the HIV-1 M group viruses, it was noted that viruses from Thailand formed a clade with some of the viruses from Africa in gag and pol regions of the genome but formed their own distinct clade in the env gene region (Carr et al., 1996)(McCutchan et al., 1992). The env gene genotype had already been named “subtype E” (Myers et al., 1992). Some debate was carried out as to whether the envelope gene of these viruses had just evolved very rapidly compared to the gag-pol region of the genome, or alternatively if recombination had been the cause (Anderson et al., 2000)(Robertson et al., 2000). The name CRF01_AE was adopted by the majority vote of the HIV and SIV nomenclature committee(Robertson et al., 2000). The CRF01_AE genotype was, and remains, somewhat problematic because no full length non-recombinant “genotype E” virus has yet been sampled and sequenced. Thus the env region could be called “U” for “untyped” if it was not already grandfathered in as “subtype E”. Moreover it has been shown that it contains a small genomic fragment closely related to subtype G suggesting that CRF01_AE has a complicated pattern of mosaicism consisting of subtypes A, G, and E.(Magiorkinis et al., 2002). The CRF02_AG genomes are somewhat similar to CRF01_AE with regard to both of these circulating recombinant forms originating many decades ago. It is clear that these viruses had already spread quite substantially before HIV was discovered. Careful analyses of subregions of HIV-1 genomes suggests that recombinations were ongoing in the early years before the subtypes each expanded, and it has been suggested that for example CRF02_AG may predate the origin of subtype G(Abecasis et al., 2007)(Zhang et al., 2010). As stated in the HIV nomenclature proposal, the “pure subtypes” were not intended to imply that no recombinations had contributed to their formation, but only that they had spread extensively and have a relatively uniform phylogenetic signal or origin across their genomes (Robertson et al., 2000). The initial HIV classification reflects in fact sampling history rather than evolutionary history of the virus. The situation is different in recent epidemics. For example, the third recombinant form to be discovered, CRF03_AB from the Ukraine, was the result of a very recent recombination event that took place very shortly before it was discovered (Liitsola et al., 1998)(Liitsola et al., 2000). Although the exact individual who was dual infected with subtypes A and B to create the CRF03_AB virus that then spread among IDU in the region was never identified, isolates of viruses very closely related to the parental strains were found among IDU in the same geographical region and sequenced(Zarandia et al., 2006). By the time of the 1999 HIV and SIV Nomenclature Committee meeting, four circulating recombinant forms had been identified and published and it was clear that more were in press or in the process of being characterized(Robertson et al., 2000)(http://www.hiv.lanl.gov/content/sequence/HIV/REVIEWS/nomenclature/Nomen.html accessed May 26, 2016). The committee established and agreed that new subtypes and CRFs would require at least three independent isolates of the same form. The complete genome sequence must be determined for at least two of the isolates. The third isolate has to be sequenced in gene regions which prove that it belongs to the same new subtype or recombinant form. 2.1.3 Unique Intersubtype Recombinants Overall, dual infections with more than one subtype of virus are rather rare except in certain regions of the world where two or more subtypes or CRFs co-circulate, in population groups with high HIV prevelance and high risk behavior, for example within an IDU community, or an MSM network. In central and western Africa many intersubtype recombinant viruses are assumed to have arisen via sexual transmissions, because opioid injection drug use is not very common. The number of recombinant viruses that are sequenced is influenced not only by the number of dual-infected people per year, but also by the length of time that two or more subtypes have been circulating in the local area, and the percentage of infected people who are sampled and sequenced. 2.1.4 Intrasubtype Recombinants With more than 30 million people infected with HIV today, and less than 100,000 infected people sampled for DNA sequencing (excluding drug resistance testing, where the sequences are not made public), a considerable proportion of sequences are unique enough to fall on their own branch within a subtype, rather than being part of a subclade within the subtype. Even when studies have sampled infections in rather small villages the infections have most often been found to be surprisingly diverse, with most transmissions coming into the villages from surrounding cities(Grabowski et al., 2014)(Carrel et al., 2014). 2.1.5 Small Genome Regions In a recent paper, Marcel Tongo et al. propose using lowercase letters to indicate regions of a virus genome that are somewhat similar to, but fall outside the known diversity of a given subtype(Tongo et al., 2015). These regions are currently listed as “U” for untyped or unclassified. One of the problems with this proposal, is that there can be many different reasons why a small region is not classifiable. In many cases, it is very likely that recombination has confounded our ability to reconstruct an accurate history. On the other side of the same issue, some researchers have argued that only regions of the genome which give 100% bootstrap support for being in the crown group of a given subtype should be labeled with that subtype, and other regions should be listed as “U” for uncertain or untyped. For most small regions of the genome, less than 100 or in some cases 200 bases long, obtaining solid bootstrap support for the correct classification is problematic. Indeed, the length of the genomic fragment that gives proper subtype signal varies across the genome. 3 HIV-2, SIV-SMM, SIV-Mac HIV-2 is the result of cross species transmissions from sooty mangabeys (Cercocebus atys) to humans. To date at least nine different lineages representing nine different cross species transfers to humans have been described, resulting in HIV-2 groups A–I(Ayouba et al., 2013)(Santiago et al., 2005). HIV-2 groups A and B have spread in western Africa and have also been detected in India and other countries. Dual infections with HIV-2 groups A and B have been reported, and at least one HIV-2 A/B recombinant virus has spread to several individuals to become listed as HIV-2 CRF01_AB(Ibe et al., 2010). Dual infections with HIV-1 and HIV-2 have been reported many times, but not recombination between these two distantly related viruses, and for most reports of dual infection there is only serological and not genetic data for confirming the dual infection. Although HIV-2 groups A and B have spread to many humans, the groups C through I of HIV-2 have mostly been detected and sequenced in just one or two individuals and technically do not meet the nomenclature criteria to be named as a group. However, because of the relative importance of cross species transmission events, it is worthwhile to note each event, even when it does not result in significant human to human spread of the virus. The SIVsmm also crossed species to infect macaques in captivity in the USA several times in the 1960s and 1970s(Apetrei et al., 2005). Several lineages of the virus then evolved in captive macaques and/or sooty mangabeys co-housed with macaques in different primate research centers in the USA. A few of these viruses were isolated in the 1980s and became very important research reagents for studying immune deficiency in macaques and testing vaccines in a nonhuman primate model(Fischer et al., 2012)(http://www.hiv.lanl.gov/content/sequence/HIV/REVIEWS/KUIKEN2000/Kuiken.html accessed June 7, 2016). For virus lineages where the transfer from sooty mangabeys to captive macaques occurred unintentionally they have been labeled as SIVstm for stump tailed macaque(Novembre et al., 1992), SIVmac for rhesus macaque(Apetrei et al., 2005), and SIVmne for Macaca nemestrema (pig tailed macaque)(Kimata et al., 1998). Laboratory made chimeras between SIVmac and HIV-1, most often with the envelope gene derived from HIV-1, are common in vaccine research, and these are called SHIVs for simian-human immunodeficiency virus. 4 PIV Details and peculiarities The number of primate immunodeficiency viruses isolated and sequenced from nonhuman primates has increased dramatically in recent years since the development of protocols for isolating viruses from freshly collected fecal samples allowed non-invasive sampling of wild animals(Santiago et al., 2003). It has long been assumed that the ancient virus host coevolution of these viruses has resulted in a decrease in pathogenicity over time to the point where the viruses are nonpathogenic to their natural hosts. Indeed there is evidence that many of the viruses are far less pathogenic to the natural host than HIV-1 is to humans or SIVsmm and SIVmac are to macaques. However, it was always known that the chimpanzee named Marilyn, from which the first SIVcpz was isolated, had died with infections in 1985 after giving birth to stillborn twins(Gao et al., 1999). More recent studies have determined that at least some of the natural infections are pathogenic but typically with long latency periods not unlike humans infected with HIV-1 or HIV-2(Keele et al., 2009). Several studies specifically focused on developing highly pathogenic SIVs and SHIVs in order to produce vaccine challenge viruses such that large improvements in the clinical outcome of vaccinated animals could be seen in short periods of time(Haddrick et al., 2001)(McCormick-Davis et al., 1998). Although these viruses prove that there can be great differences in pathogenicity between viruses that are very similar to each other, the majority of studies of rates of disease progression in humans infected with HIV-1 M group viruses have shown that it is human genotype rather than viral genotype that is most often the determining factor(McLaren and Carrington, 2015)(Limou and Zagury, 2013). Highly active antiretroviral therapies in humans have made studies of disease progression rates in untreated humans unethical(INSIGHT START Study Group et al., 2015)(Wagner et al., 2016). Estimating the length of time that PIVs have coevolved with their natural hosts is problematic due to the lack of solid fossil record or datable events deep in the phylogenies. Molecular phylogenetic methods are hampered by the problem of saturation(Duchêne et al., 2015). Although no New World monkeys are known to be infected with PIVs, felines around the world are infected with feline immunodeficiency viruses with no greater diversity than is found in the African PIVs(Troyer et al., 2011)(O’Brien et al., 2012). Endogenous lentiviruses have been found in lemurs, rabbits, ferrets and weasels and estimated to have become endogenous as much as ten million years ago(Han and Worobey, 2012). Although other estimates of PIV origins have been made with populations of primates separated for long periods of time, these methods can be confounded by not knowing that the animals which became isolated from each other were carrying just one lineage of virus at the time of isolation(Wertheim and Worobey, 2009)(Worobey et al., 2010)(Ayouba et al., 2015). Cross species transmissions of PIVs are not limited to primate to human transfers. It is clear for example that baboons have repeatedly been infected in the wild with SIVver from the vervet species of African Green Monkeys, Chlorocebus pygerythrus(van Rensburg et al., 1998)(Jin et al., 1994). The SIV lineage infecting chimanzees and gorillas is apparently the result of a very ancient recombination event, with the pol gene of SIVcpz somewhat related to virus from red capped mangabeys while the env gene is more closely related to virus from greater spot-nosed monkeys(Etienne et al., 2013)(Bailes, 2003) (Leitner et al., 2007). Fighting and predation between primate species is rather common, so the relative lack of cross species transmission events is a testament to how well the primate APOBEC system, and other innate and adaptive immune functions protect against these viruses(Puvvada and Patel, 2013)(Bibollet-Ruche et al., 2004)(Etienne et al., 2015)(D’arc et al., 2015). 5 Conclusions Although there are some inconsistencies and confusing issues in the nomenclature of HIVs and PIVs, these problems are mostly a reflection either of changes in the state of knowledge about the diversity of these viruses over time, or inherent biological quirks such as recombination and cross species transmissions. The goals of the classification and nomenclature system are both to make biological sense and to make it possible for humans and man-made databases to organize and understand the data. Moreover classification of viral sequences is crucial to understand the complex epidemiology of HIV-1 epidemic and also to monitor temporal and spatial changes on a global and regional scale. There have been proposals to better organize the CRFs such as grouping related recombinants together(Zhang et al., 2010). The future seems certain to bring challenges to the existing system. One confusing issue may come up if more subtypes in the HIV-1 M group are discovered and classified, such that the A, B, C, D naming of subtypes overtakes the M, N, O, P naming of groups. Despite some complexities or difficulties with naming primate lentiviruses, it is very clear that the current system works well in comparison to the classification of many or even most other viruses(Fauquet and Fargette, 2005). For many viruses, the lack of standards set by the International Committee for the Taxonomy of Viruses (ICTV) below the species level is a barrier to accurate classification. One problem is that there is no biologically solid way to define what constitutes a “species” of virus. For the lentiviruses a classification based partly on host makes good sense because in general the cross species events are quite rare in comparison to coevolution within a virus-host pair. But for other viruses such as the influenza A virus it is very common for a virus that is very clearly a human virus to be isolated from a pig, or vise versa. In cases where the current rules for nomenclature and classification of lentiviruses are violated by names that were “grandfathered in” before these rules were standardized, the problem of rules broken is less severe than the confusion which would be caused by renaming well-known groups now. Most challenges with the current system are local in nature, such as describing the epidemiology of local outbreaks, and do not have an impact on the HIV-1 M group pandemic. It is completely acceptable to use names for local strains as has already been done many times, but authors should not expect those names to become incorporated into the official nomenclature. Supplementary Material 1 2 Funding: This work was supported by the National Institutes of Health, NIH Contract AAI12007-001-01001 Figure 1 A maximum likelihood phylogenetic tree constructed from the nearly complete genomes (gag through nef genes) of primate immunodeficiency viruses. The plot of transitions and transversions vs estimated evolutionary distances shows that the distances across the alignment are far beyond saturation of silent site mutations and thus underestimated in comparison to distances within a recent expansion such as the HIV-1 M group of viruses. Clades in color represent viruses identified in humans after relatively recent transfer from chimpanzee, gorilla, and sooty mangabey reservoirs. Figure 2 The cover of the 1992 HIV Database Compendium illustrated the diversity of HIV-1 M group viruses that had been sequenced to date, with a phylogenetic tree built from envelope gene sequences including 3 sequences from each subtype, plus one from “subtype E” which later became known as the CRF01_AE circulating recombinant form. Figure 3 Maximum likelihood phylogenetic tree constructed using nearly complete HIV-1 M group genomes (gag through nef genes). Sequences were aligned with the HIV Databases GeneCutter tool (http://www.hiv.lanl.gov/content/sequence/GENE_CUTTER/cutter.html) and a maximum likelihood phylogenetic tree was constructed using the GTR model of evolution plus site-specific rates of evolution using IQ-tree(Nguyen et al., 2015). Figure 4 Schematic picture of the branching index. Here the association of sequence X to the subtype cluster S is investigated. The letters a and b represent genetic distances that depend on the position of the node of sequence X (white circle) at the bold branch. The branching index is defined as a/(a + b) and can take values between 0 and 1 , where 0 means no support and 1 means solid support for belonging to the subtype(Wilbe et al., 2003). Table 1 Virus species Primate species Comments Primate&species ASC Red-tailed Guenon Also dubbed SIVrtg. There are at least 3 SIVasc species infecting the two subspecies of Red-tailed Guenons, but complete genomes not yet available for all 3. Cercopithecus ascanius schmidti BAB Baboon Baboon infected in the wild with vervet SIV. Papio spp. BLC Bioko Black Colobus Cercopithecus,satanas,satanas BKM Black Mangabey Lophocebus aterrimus BLU Blue Monkey Cercopithecus,mitis COL Colobus'Monkey Colobus guereza CPZ Chimpanzee SIVcpz infects Pan troglodytes troglodytes and Pan troglodytes schweinfurthi subspecies. Pan paniscus does not seem to have a SIV. Pan troglodytes troglodytes or P. t. schweinfurthii DEB DeBrazza’ s Monkey Cercopithecus neglectus DEN Dent s Monkey Cercopithecus denti DRL Drill Monkey Mandrillus leucophaeus GOR Gorilla Gorilla gorilla GRV Grivet Formerly a subspecies of African Green Monkey. Chlorocebus aethiops GSN Greater Spotnosed Monkey Cercopithecus nictitans KRC Kibale Red Colobus Procolobus [Piliocolobus] rufomitratus tephrosceles LST L’Hoest ’ s Monkey Cercopithecus lhoesti MAC Macaque Rhesus Macaque naturally infected by SIVsmm. Macaca mulatta MND-1 Mandrill Mandrills are infected with two types of SIV nearly as diverse from each other as HIV-1 and HIV-2 are from each other. Mandrillus sphinx MNE Pig-tailed Macaque Pig-tailed Macaque infected in captivity by SIVsmm. Macaca nemestrina MON Mona Monkey Cercopithecus mona  MUS-1  MUS-2 MUS-3 Mustached Monkey At least 3 distinct SIV species infect Mustached Monkeys. Cercopithecus cephus OLC Olive Colobus Procolobus verus PAT Patas Monkey Erythrocebus patas PRG Preuss s Guenon Preuss ’ s Red-eared Guenon from Bioko. Cercopithecus preussi insularis RCM Red-capped Mangabey Cercocebus torquatus torquatus REG Red-eared Guenon Red-eared Guenon from Bioko. Also dubbed SIVery. Cercopithecus erythrotis erythrotis SAB Sabaeus Monkey Also called Green Monkey. Chlorocebus sabaeus SMM Sooty Mangabey SMM is the species for both Sooty Mangabey and macaques experimentally infected with SMM. Cercocebus atys atys STM Stump-tailed Macaque Stump-tailed Macaque infected in captivity by SIVsmm. Macaca arctoides SUN Sun-tailed Monkey Cercopithecus solatus SYK Sykes Monkey Cercopithecus albogularis TAL Talapoin Monkey Miopithecus ogouensis or Miopithecus talapoin TAN Tantalus Monkey Formerly a subspecies of African Green Monkey. Chlorocebus tantalus TRC Tshuapa Red Colobus Piliocolobus tholloni VER Vervet Formerly a subspecies of African Green Monkey. Chlorocebus pygerythrus WCM White-crowned Mangabey White-crowned Mangabey infected in captivity by SIVver. Cercocebus torquatus lunulatus WOL Wolf s Monkey Cercopithecus wolfi WRC Western Red Colobus Procolobus verus This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. Abecasis AB Lemey P Vidal N de Oliveira T Peeters M Camacho R Shapiro B Rambaut A Vandamme AM 2007 Recombination confounds the early evolutionary history of human immunodeficiency virus type 1: subtype G is a circulating recombinant form J Virol 81 8543 51 10.1128/JVI.00463-07 17553886 Anderson JP Rodrigo AG Learn GH Madan A Delahunty C Coon M Girard M Osmanov S Hood L Mullins JI 2000 Testing the hypothesis of a recombinant origin of human immunodeficiency virus type 1 subtype E J Virol 74 10752 65 11044120 Apetrei C Kaur A Lerche NW Metzger M Pandrea I Hardcastle J Falkenstein S Bohm R Koehler J Traina-Dorge V Williams T Staprans S Plauche G Veazey RS McClure H Lackner AA Gormus B Robertson DL Marx PA 2005 Molecular epidemiology of simian immunodeficiency virus SIVsm in U.S. primate centers unravels the origin of SIVmac and SIVstm J Virol 79 8991 9005 10.1128/JVI.79.14.8991-9005.2005 15994793 Ayouba A Akoua-Koffi C Calvignac-Spencer S Esteban A Locatelli S Li H Li Y Hahn BH Delaporte E Leendertz FH Peeters M 2013 Evidence for continuing cross-species transmission of SIVsmm to humans: characterization of a new HIV-2 lineage in rural Côte d’Ivoire AIDS 27 2488 91 10.1097/01.aids.0000432443.22684.50 23939239 Ayouba A Njouom R Chia JE Ahuka-Mundeke S Kfutwah A Ngole EM Nerrienet E Delaporte E Peeters M 2015 Molecular characterization of a new mosaic Simian Immunodeficiency Virus in a naturally infected tantalus monkey (Chlorocebus tantalus) from Cameroon: a challenge to the virus-host co-evolution of SIVagm in African green monkeys Infect Genet Evol 30 65 73 10.1016/j.meegid.2014.12.005 25500294 Bailes E 2003 Hybrid Origin of SIV in Chimpanzees Science (80- ) 300 1713 1713 10.1126/science.1080657 Bibollet-Ruche F Bailes E Gao F Pourrut X Barlow KL Clewley JP Mwenda JM Langat DK Chege GK McClure HM Mpoudi-Ngole E Delaporte E Peeters M Shaw GM Sharp PM Hahn BH 2004 New simian immunodeficiency virus infecting De Brazza’s monkeys (Cercopithecus neglectus): evidence for a cercopithecus monkey virus clade J Virol 78 7748 62 10.1128/JVI.78.14.7748-7762.2004 15220449 Bobkov A Cheingsong-Popov R Garaev M Rzhaninova A Kaleebu P Beddows S Bachmann MH Mullins JI Louwagie J Janssens W 1994 Identification of an env G subtype and heterogeneity of HIV-1 strains in the Russian Federation and Belarus AIDS 8 1649 55 7888112 Carr JK Salminen MO Koch C Gotte D Artenstein AW Hegerich PA St Louis D Burke DS McCutchan FE 1996 Full-length sequence and mosaic structure of a human immunodeficiency virus type 1 isolate from Thailand J Virol 70 5935 43 8709215 Carrel M Eron JJ Emch M Hurt CB 2014 Spatial epidemiology of recently acquired HIV infections across rural and urban areas of North Carolina PLoS One 9 e88512 10.1371/journal.pone.0088512 24520392 Cernescu CE Tardei G Necula A Ruta SM Pau CP 1994 The serologic significance of F viral genotype for human immunodeficiency virus type 1 epidemic J Infect Dis 170 1043 4 7930705 D’arc M Ayouba A Esteban A Learn GH Boué V Liegeois F Etienne L Tagg N Leendertz FH Boesch C Madinda NF Robbins MM Gray M Cournil A Ooms M Letko M Simon VA Sharp PM Hahn BH Delaporte E Mpoudi Ngole E Peeters M 2015 Origin of the HIV-1 group O epidemic in western lowland gorillas Proc Natl Acad Sci U S A 112 E1343 52 10.1073/pnas.1502022112 25733890 Duchêne S Ho SYW Holmes EC 2015 Declining transition/transversion ratios through time reveal limitations to the accuracy of nucleotide substitution models BMC Evol Biol 15 36 10.1186/s12862-015-0312-6 25886870 Etienne L Bibollet-Ruche F Sudmant PH Wu LI Hahn BH Emerman M 2015 The Role of the Antiviral APOBEC3 Gene Family in Protecting Chimpanzees against Lentiviruses from Monkeys PLoS Pathog 11 e1005149 10.1371/journal.ppat.1005149 26394054 Etienne L Hahn BH Sharp PM Matsen FA Emerman M 2013 Gene loss and adaptation to hominids underlie the ancient origin of HIV-1 Cell Host Microbe 14 85 92 10.1016/j.chom.2013.06.002 23870316 Faria NR Rambaut A Suchard MA Baele G Bedford T Ward MJ Tatem AJ Sousa JD Arinaminpathy N Pépin J Posada D Peeters M Pybus OG Lemey P 2014 HIV epidemiology. The early spread and epidemic ignition of HIV-1 in human populations Science 346 56 61 10.1126/science.1256739 25278604 Fauquet CM Fargette D 2005 International Committee on Taxonomy of Viruses and the 3,142 unassigned species Virol J 2 64 10.1186/1743-422X-2-64 16105179 Feng Y He X Hsi JH Li F Li X Wang Q Ruan Y Xing H Lam TTY Pybus OG Takebe Y Shao Y 2013 The rapidly expanding CRF01_AE epidemic in China is driven by multiple lineages of HIV-1 viruses introduced in the 1990s AIDS 27 1793 802 10.1097/QAD.0b013e328360db2d 23807275 Fischer W Apetrei C Santiago ML Li Y Gautam R Pandrea I Shaw GM Hahn BH Letvin NL Nabel GJ Korber BT 2012 Distinct evolutionary pressures underlie diversity in simian immunodeficiency virus and human immunodeficiency virus lineages J Virol 86 13217 31 10.1128/JVI.01862-12 23055550 Gao F Bailes E Robertson DL Chen Y Rodenburg CM Michael SF Cummins LB Arthur LO Peeters M Shaw GM Sharp PM Hahn BH 1999 Origin of HIV-1 in the chimpanzee Pan troglodytes troglodytes Nature 397 436 41 10.1038/17130 9989410 Gao F Robertson DL Carruthers CD Li Y Bailes E Kostrikis LG Salminen MO Bibollet-Ruche F Peeters M Ho DD Shaw GM Sharp PM Hahn BH 1998 An isolate of human immunodeficiency virus type 1 originally classified as subtype I represents a complex mosaic comprising three different group M subtypes (A, G, and I) J Virol 72 10234 41 9811767 Grabowski MK Lessler J Redd AD Kagaayi J Laeyendecker O Ndyanabo A Nelson MI Cummings DAT Bwanika JB Mueller AC Reynolds SJ Munshaw S Ray SC Lutalo T Manucci J Tobian AAR Chang LW Beyrer C Jennings JM Nalugoda F Serwadda D Wawer MJ Quinn TC Gray RH 2014 The role of viral introductions in sustaining community-based HIV epidemics in rural Uganda: evidence from spatial clustering, phylogenetics, and egocentric transmission models PLoS Med 11 e1001610 10.1371/journal.pmed.1001610 24595023 Haddrick M Brown CR Plishka R Buckler-White A Hirsch VM Ginsberg H 2001 Biologic studies of chimeras of highly and moderately virulent molecular clones of simian immunodeficiency virus SIVsmPBj suggest a critical role for envelope in acute AIDS virus pathogenesis J Virol 75 6645 59 10.1128/JVI.75.14.6645-6659.2001 11413332 Han GZ Worobey M 2012 Endogenous lentiviral elements in the weasel family (Mustelidae) Mol Biol Evol 29 2905 8 10.1093/molbev/mss126 22522310 Hirsch VM Olmsted RA Murphey-Corb M Purcell RH Johnson PR 1989 An African primate lentivirus (SIVsm) closely related to HIV-2 Nature 339 389 92 10.1038/339389a0 2786147 Hogan CA Iles J Frost EH Giroux G Cassar O Gessain A Dion M-J Ilunga V Rambaut A Yengo-Ki-Ngimbi A-É Behets F Pybus OG Pépin J 2016 Epidemic History and Iatrogenic Transmission of Blood-borne Viruses in Mid-20th Century Kinshasa J Infect Dis 10.1093/infdis/jiw009 Hraber P Kuiken C Waugh M Geer S Bruno WJ Leitner T 2008 Classification of hepatitis C virus and human immunodeficiency virus-1 sequences with the branching index J Gen Virol 89 2098 107 10.1099/vir.0.83657-0 18753218 Huet T Cheynier R Meyerhans A Roelants G Wain-Hobson S 1990 Genetic organization of a chimpanzee lentivirus related to HIV-1 Nature 345 356 9 10.1038/345356a0 2188136 Ibe S Yokomaku Y Shiino T Tanaka R Hattori J Fujisaki S Iwatani Y Mamiya N Utsumi M Kato S Hamaguchi M Sugiura W 2010 HIV-2 CRF01_AB: first circulating recombinant form of HIV-2 J Acquir Immune Defic Syndr 54 241 7 10.1097/QAI.0b013e3181dc98c1 20502347 INSIGHT START Study Group Lundgren JD Babiker AG Gordin F Emery S Grund B Sharma S Avihingsanon A Cooper DA Fätkenheuer G Llibre JM Molina J-M Munderi P Schechter M Wood R Klingman KL Collins S Lane HC Phillips AN Neaton JD 2015 Initiation of Antiretroviral Therapy in Early Asymptomatic HIV Infection N Engl J Med 373 795 807 10.1056/NEJMoa1506816 26192873 Jin MJ Rogers J Phillips-Conroy JE Allan JS Desrosiers RC Shaw GM Sharp PM Hahn BH 1994 Infection of a yellow baboon with simian immunodeficiency virus from African green monkeys: evidence for cross-species transmission in the wild J Virol 68 8454 60 7966642 Kalish ML Baldwin A Raktham S Wasi C Luo CC Schochetman G Mastro TD Young N Vanichseni S Rübsamen-Waigmann H 1995 The evolving molecular epidemiology of HIV-1 envelope subtypes in injecting drug users in Bangkok, Thailand: implications for HIV vaccine trials AIDS 9 851 7 7576318 Keele BF Jones JH Terio KA Estes JD Rudicell RS Wilson ML Li Y Learn GH Beasley TM Schumacher-Stankey J Wroblewski E Mosser A Raphael J Kamenya S Lonsdorf EV Travis DA Mlengeya T Kinsel MJ Else JG Silvestri G Goodall J Sharp PM Shaw GM Pusey AE Hahn BH 2009 Increased mortality and AIDS-like immunopathology in wild chimpanzees infected with SIVcpz Nature 460 515 9 10.1038/nature08200 19626114 Keele BF Van Heuverswyn F Li Y Bailes E Takehisa J Santiago ML Bibollet-Ruche F Chen Y Wain LV Liegeois F Loul S Ngole EM Bienvenue Y Delaporte E Brookfield JFY Sharp PM Shaw GM Peeters M Hahn BH 2006 Chimpanzee reservoirs of pandemic and nonpandemic HIV-1 Science 313 523 6 10.1126/science.1126531 16728595 Kimata JT Mozaffarian A Overbaugh J 1998 A lymph node-derived cytopathic simian immunodeficiency virus Mne variant replicates in nonstimulated peripheral blood mononuclear cells J Virol 72 245 56 9420221 Kostrikis LG Bagdades E Cao Y Zhang L Dimitriou D Ho DD 1995 Genetic analysis of human immunodeficiency virus type 1 strains from patients in Cyprus: identification of a new subtype designated subtype I J Virol 69 6122 30 7666516 Lapovok I Kazennova E Laga V Vasilyev A Utegenova A Abishev A Dzissyuk N Tukeev M Bobkova M 2014 Short communication: molecular epidemiology of HIV type 1 infection in Kazakhstan: CRF02_AG prevalence is increasing in the southeastern provinces AIDS Res Hum Retroviruses 30 769 74 10.1089/AID.2013.0291 24873898 Leitner T Dazza MC Ekwalanga M Apetrei C Saragosti S 2007 Sequence diversity among chimpanzee simian immunodeficiency viruses (SIVcpz) suggests that SIVcpzPts was derived from SIVcpzPtt through additional recombination events AIDS Res Hum Retroviruses 23 1114 8 10.1089/aid.2007.0071 17919106 Li Y Ndjango JB Learn GH Ramirez MA Keele BF Bibollet-Ruche F Liu W Easlick JL Decker JM Rudicell RS Inogwabini BI Ahuka-Mundeke S Leendertz FH Reynolds V Muller MN Chancellor RL Rundus AS Simmons N Worobey M Shaw GM Peeters M Sharp PM Hahn BH 2012 Eastern chimpanzees, but not bonobos, represent a simian immunodeficiency virus reservoir J Virol 86 10776 91 10.1128/JVI.01498-12 22837215 Liitsola K Holm K Bobkov A Pokrovsky V Smolskaya T Leinikki P Osmanov S Salminen M 2000 An AB recombinant and its parental HIV type 1 strains in the area of the former Soviet Union: low requirements for sequence identity in recombination. UNAIDS Virus Isolation Network AIDS Res Hum Retroviruses 16 1047 53 10.1089/08892220050075309 10933619 Liitsola K Tashkinova I Laukkanen T Korovina G Smolskaja T Momot O Mashkilleyson N Chaplinskas S Brummer-Korvenkontio H Vanhatalo J Leinikki P Salminen MO 1998 HIV-1 genetic subtype A/B recombinant strain causing an explosive epidemic in injecting drug users in Kaliningrad AIDS 12 1907 19 9792392 Limou S Zagury JF 2013 Immunogenetics: Genome-Wide Association of Non-Progressive HIV and Viral Load Control: HLA Genes and Beyond Front Immunol 4 118 10.3389/fimmu.2013.00118 23750159 Magiorkinis G Paraskevis D Magiorkinis E Vandamme AM Hatzakis A 2002 Reanalysis of the HIV-1 circulating recombinant form A/E (CRF01_AE): evidence of A/E/G recombination J Acquir Immune Defic Syndr 30 124 9 12048373 McCormick-Davis C Zhao LJ Mukherjee S Leung K Sheffer D Joag SV Narayan O Stephens EB 1998 Chronology of genetic changes in the vpu, env, and Nef genes of chimeric simian-human immunodeficiency virus (strain HXB2) during acquisition of virulence for pig-tailed macaques Virology 248 275 83 10.1006/viro.1998.9300 9721236 McCutchan FE Hegerich PA Brennan TP Phanuphak P Singharaj P Jugsudee A Berman PW Gray AM Fowler AK Burke DS 1992 Genetic variants of HIV-1 in Thailand AIDS Res Hum Retroviruses 8 1887 95 10.1089/aid.1992.8.1887 1489577 McLaren PJ Carrington M 2015 The impact of host genetic variation on infection with HIV-1 Nat Immunol 16 577 83 10.1038/ni.3147 25988890 Meloni ST Kim B Sankalé JL Hamel DJ Tovanabutra S Mboup S McCutchan FE Kanki PJ 2004 Distinct human immunodeficiency virus type 1 subtype A virus circulating in West Africa: sub-subtype A3 J Virol 78 12438 45 10.1128/JVI.78.22.12438-12445.2004 15507630 Mokili JLK Rogers M Carr JK Simmonds P Bopopi JM Foley BT Korber BT Birx DL McCutchan FE 2002 Identification of a novel clade of human immunodeficiency virus type 1 in Democratic Republic of Congo AIDS Res Hum Retroviruses 18 817 23 10.1089/08892220260139567 12167273 Montavon C Toure-Kane C Nkengasong JN Vergne L Hertogs K Mboup S Delaporte E Peeters M 2002 CRF06-cpx: a new circulating recombinant form of HIV-1 in West Africa involving subtypes A, G, K, and J. J. Acquir Immune Defic Syndr 29 522 30 Morgado MG Sabino EC Shpaer EG Bongertz V Brigido L Guimaraes MD Castilho EA Galvão-Castro B Mullins JI Hendry RM 1994 V3 region polymorphisms in HIV-1 from Brazil: prevalence of subtype B strains divergent from North American/European prototype and detection of subtype F AIDS Res Hum Retroviruses 10 569 76 10.1089/aid.1994.10.569 7522493 Myers G MacInnes K Korber B 1992 The emergence of simian/human immunodeficiency viruses AIDS Res Hum Retroviruses 8 373 86 1571197 Nasioulas G Paraskevis D Magiorkinis E Theodoridou M Hatzakis A 1999 Molecular analysis of the full-length genome of HIV type 1 subtype I: evidence of A/G/I recombination AIDS Res Hum Retroviruses 15 745 58 10.1089/088922299310836 10357470 Neher RA Leitner T 2010 Recombination rate and selection strength in HIV intra-patient evolution PLoS Comput Biol 6 e1000660 10.1371/journal.pcbi.1000660 20126527 Nguyen LT Schmidt HA von Haeseler A Minh BQ 2015 IQ-TREE: a fast and effective stochastic algorithm for estimating maximum-likelihood phylogenies Mol Biol Evol 32 268 74 10.1093/molbev/msu300 25371430 Novembre FJ Hirsch VM McClure HM Fultz PN Johnson PR 1992 SIV from stump-tailed macaques: molecular characterization of a highly transmissible primate lentivirus Virology 186 783 7 1733112 O’Brien SJ Troyer JL Brown MA Johnson WE Antunes A Roelke ME Pecon-Slattery J 2012 Emerging viruses in the Felidae: shifting paradigms Viruses 4 236 57 10.3390/v4020236 22470834 Paraskevis D Magiorkinis M Vandamme AM Kostrikis LG Hatzakis A 2001 Re-analysis of human immunodeficiency virus type 1 isolates from Cyprus and Greece, initially designated “subtype I”, reveals a unique complex A/G/H/K/? mosaic pattern J Gen Virol 82 575 80 10.1099/0022-1317-82-3-575 11172098 Pillay D Herbeck J Cohen MS de Oliveira T Fraser C Ratmann O Brown AL Kellam P PANGEA-HIV Consortium 2015 PANGEA-HIV: phylogenetics for generalised epidemics in Africa Lancet Infect Dis 15 259 61 10.1016/S1473-3099(15)70036-8 25749217 Puvvada MP Patel SS 2013 Role of trim5α in the suppression of cross-species transmission and its defence against human immunodeficiency virus Curr HIV Res 11 601 9 24606328 Ratner L Haseltine W Patarca R Livak KJ Starcich B Josephs SF Doran ER Rafalski JA Whitehorn EA Baumeister K 1985 Complete nucleotide sequence of the AIDS virus, HTLV-III Nature 313 277 84 2578615 Robertson DL Anderson JP Bradac JA Carr JK Foley B Funkhouser RK Gao F Hahn BH Kalish ML Kuiken C Learn GH Leitner T McCutchan F Osmanov S Peeters M Pieniazek D Salminen M Sharp PM Wolinsky S Korber B 2000 HIV-1 nomenclature proposal Science 288 55 6 10766634 Sanchez-Pescador R Power MD Barr PJ Steimer KS Stempien MM Brown-Shimer SL Gee WW Renard A Randolph A Levy JA 1985 Nucleotide sequence and expression of an AIDS-associated retrovirus (ARV-2) Science 227 484 92 2578227 Santiago ML Bibollet-Ruche F Bailes E Kamenya S Muller MN Lukasik M Pusey AE Collins DA Wrangham RW Goodall J Shaw GM Sharp PM Hahn BH 2003 Amplification of a complete simian immunodeficiency virus genome from fecal RNA of a wild chimpanzee J Virol 77 2233 42 12525658 Santiago ML Range F Keele BF Li Y Bailes E Bibollet-Ruche F Fruteau C Noë R Peeters M Brookfield JFY Shaw GM Sharp PM Hahn BH 2005 Simian immunodeficiency virus infection in free-ranging sooty mangabeys (Cercocebus atys atys) from the Taï Forest, Côte d’Ivoire: implications for the origin of epidemic human immunodeficiency virus type 2 J Virol 79 12515 27 10.1128/JVI.79.19.12515-12527.2005 16160179 Sharp PM Hahn BH 2010 The evolution of HIV-1 and the origin of AIDS Philos Trans R Soc Lond B Biol Sci 365 2487 94 10.1098/rstb.2010.0031 20643738 Tongo M Dorfman JR Martin DP 2015 High Degree of HIV-1 Group M (HIV-1M) Genetic Diversity within Circulating Recombinant Forms: Insight into the Early Events of HIV-1M Evolution J Virol 90 2221 9 10.1128/JVI.02302-15 26656688 Tosi AJ Detwiler KM Disotell TR 2005 X-chromosomal window into the evolutionary history of the guenons (Primates: Cercopithecini) Mol Phylogenet Evol 36 58 66 10.1016/j.ympev.2005.01.009 15904856 Triques K Bourgeois A Vidal N Mpoudi-Ngole E Mulanga-Kabeya C Nzilambi N Torimiro N Saman E Delaporte E Peeters M 2000 Near-full-length genome sequencing of divergent African HIV type 1 subtype F viruses leads to the identification of a new HIV type 1 subtype designated K. AIDS Res Hum Retroviruses 16 139 51 10.1089/088922200309485 Troyer JL Roelke ME Jespersen JM Baggett N Buckley-Beason V MacNulty D Craft M Packer C Pecon-Slattery J O’Brien SJ 2011 FIV diversity: FIV Ple subtype composition may influence disease outcome in African lions Vet Immunol Immunopathol 143 338 46 10.1016/j.vetimm.2011.06.013 21723622 van Rensburg EJ Engelbrecht S Mwenda J Laten JD Robson BA Stander T Chege GK 1998 Simian immunodeficiency viruses (SIVs) from eastern and southern Africa: detection of a SIVagm variant from a chacma baboon J Gen Virol 79 Pt 7 1809 14 10.1099/0022-1317-79-7-1809 9680146 Vidal N Bazepeo SE Mulanga C Delaporte E Peeters M 2009 Genetic characterization of eight full-length HIV type 1 genomes from the Democratic Republic of Congo (DRC) reveal a new subsubtype, A5, in the A radiation that predominates in the recombinant structure of CRF26_A5U AIDS Res Hum Retroviruses 25 823 32 10.1089/aid.2008.0283 19678767 Vidal N Mulanga C Bazepeo SE Lepira F Delaporte E Peeters M 2006 Identification and molecular characterization of subsubtype A4 in central Africa AIDS Res Hum Retroviruses 22 182 7 10.1089/aid.2006.22.182 16478401 Wagner GJ Linnemayr S Ghosh-Dastidar B Currier JS Hoffman R Schneider S 2016 Supporting Treatment Adherence Readiness through Training (START) for patients with HIV on antiretroviral therapy: study protocol for a randomized controlled trial Trials 17 162 10.1186/s13063-016-1287-3 27009061 Wertheim JO Worobey M 2009 Dating the age of the SIV lineages that gave rise to HIV-1 and HIV-2 PLoS Comput Biol 5 e1000377 10.1371/journal.pcbi.1000377 19412344 Wilbe K Salminen M Laukkanen T McCutchan F Ray SC Albert J Leitner T 2003 Characterization of novel recombinant HIV-1 genomes using the branching index Virology 316 116 25 14599796 Worobey M Telfer P Souquière S Hunter M Coleman CA Metzger MJ Reed P Makuwa M Hearn G Honarvar S Roques P Apetrei C Kazanji M Marx PA 2010 Island biogeography reveals the deep history of SIV Science 329 1487 10.1126/science.1193550 20847261 Zarandia M Tsertsvadze T Carr JK Nadai Y Sanchez JL Nelson AK 2006 HIV-1 genetic diversity and genotypic drug susceptibility in the Republic of Georgia AIDS Res Hum Retroviruses 22 470 6 10.1089/aid.2006.22.470 16706626 Zeng H Li T Wang Y Sun B Yang R 2016 The Epidemic Dynamics of Four Major Lineages of HIV-1 CRF01_AE Strains After Their Introduction into China AIDS Res Hum Retroviruses 32 420 6 10.1089/AID.2015.0212 26830205 Zhang M Foley B Schultz AK Macke JP Bulla I Stanke M Morgenstern B Korber B Leitner T 2010 The role of recombination in the emergence of a complex and dynamic HIV epidemic Retrovirology 7 25 10.1186/1742-4690-7-25 20331894

PMC005xxxxxx/PMC5136519.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 8007161 20873 Int J Cosmet Sci Int J Cosmet Sci International journal of cosmetic science 0142-5463 1468-2494 27261203 5136519 10.1111/ics.12348 NIHMS793378 Article Molecular basis of retinol anti-aging properties in naturally aged human skin in vivo Shao Yuan He Tianyuan Fisher Gary J. Voorhees John J. Quan Taihao * Department of Dermatology, University of Michigan Medical School, Ann Arbor, Michigan, USA * To whom correspondence should be addressed: Department of Dermatology, University of Michigan Medical School, 1301 Catherine, Medical Science I, Room 6447, Ann Arbor, Michigan 48109-0609, Telephone: (734) 165-2403, Facsimile: (734) 647-0076, thquan@umich.edu 10 6 2016 04 7 2016 2 2017 01 2 2018 39 1 5665 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. Synopsis OBJECTIVE Retinoic acid has been shown to improve the aged-appearing skin. However, less is known about the anti-aging effects of retinol (ROL, vitamin A), a precursor of retinoic acid, in aged human skin in vivo. This study aimed to investigate the molecular basis of retinol anti-aging properties in naturally aged human skin in vivo. METHODS Sun-protected buttock skin (76±6 years old, n=12) was topically treated with 0.4% ROL and its vehicle for seven days. The effects of topical ROL on skin epidermis and dermis were evaluated by immunohistochemistry, in situ hybridization, Northern analysis, real-time RT-PCR, and Western analysis. Collagen fibrils nanoscale structure and surface topology were analysed by atomic force microscopy. RESULTS Topical ROL shows remarkable anti-aging effects through three major types of skin cells: epidermal keratinocytes, dermal endothelial cells and fibroblasts. Topical ROL significantly increased epidermal thickness by stimulating keratinocytes proliferation and upregulation of c-Jun transcription factor. In addition to epidermal changes, topical ROL significantly improved dermal extracellular matrix (ECM) microenvironment; increasing dermal vascularity by stimulating endothelial cells proliferation and ECM production (type I collagen, fibronectin, and elastin) by activating dermal fibroblasts. Topical ROL also stimulates TGF-β/CTGF pathway, the major regulator of ECM homeostasis, and thus enriched the deposition of ECM in aged human skin in vivo. 0.4% topical ROL achieved similar results as seen with topical retinoic acid, the biologically active form of ROL, without causing noticeable signs of retinoid side effects. CONCLUSION 0.4% topical ROL shows remarkable anti-aging effects through improvement of the homeostasis of epidermis and dermis by stimulating the proliferation of keratinocytes and endothelial cells, and activating dermal fibroblasts. These data provide evidence that 0.4% topical ROL is a promising and safe treatment to improve naturally aged human skin. Retinoid Skin Aging TGF-β ECM Vascularity Introduction Human skin, like all other organs, undergoes natural aging process as a consequence of the passage of time [1]. In addition, human skin, unlike other organs, continuously experience harmful stress and damage from environmental sources such as solar UV irradiation [2]. As skin changes are among the most visible signs of aging, skin is central in the social and visual experience. As such, skin appearance has a significant emotional and psychological impact on our life quality. In clinically, aged skin has a significant pathological impact on many age-related skin diseases, notably impaired wound healing [3, 4] and promotion of cancer development in elderly [5, 6]. Histological and ultrastructural studies have revealed skin undergoes remarkable morphologic changes with aging, which are primarily characterized by thinner epidermis and dermis with reduced numbers of epidermal keratinocytes and dermal stromal cells, respectively [7, 8]. This age-related thinning of epidermis and dermis is the major driving force for the aged-appearing skin in elderly. It is well documented that age-related thinning of the epidermis significantly impairs skin function such as impaired skin barrier function that forms a barrier to prevent water loss and protect against environmental insults. In addition to epidermis, aged skin dermis became thin largely due to loss of collagen, the major structural protein in skin. Age-related thinning of the dermis increases fragility [9, 10], impaired vasculature support [11, 12], poor wound healing [3, 4], and altered tissue microenvironment that promotes cancer development [5, 6]. Anti-aging preventative approaches have been proposed for the purpose of rejuvenating aged human skin and preventing age-related skin diseases. However, safe and effective therapies to reverse the atrophy of aged skin do not exist currently. Topical retinoic acid has been shown to improve clinical features of aged-appearance by reduction in fine wrinkling, increased smoothness, and diminished hyperpigmentation [13]. However, the underlying cellular and molecular mechanisms are remaining to be fully elucidated. Moreover, in contrast to retinoic acid, there are limited studies describing the molecular basis of retinol (ROL, vitamin A), a precursor of retinoic acid, anti-aging effects in human skin in vivo. As human skin has the capacity to convert retinol to its biologically active metabolite retinoic acid, we investigated retinol anti-aging effects in aged human skin in vivo. Our results demonstrate that topical application of 0.4% ROL shows remarkable anti-aging effects in aged human skin through improvement of both epidermis and dermis and enhanced TGF-β/CTGF pathway. These data reveal molecular basis of retinol anti-aging effects and further suggest that topical retinol is a promising and safe treatment to improve naturally aged human skin. Materials and methods Procurement of human tissue samples and retinol treatment in human skin in vivo All procedures involving human volunteers were approved by the University of Michigan Institutional Review Board, and all volunteers provided written informed consent. Full thickness punch biopsies (4 mm) were obtained from aged (76±6 years) healthy sun-protected buttock skin after topical treatment of vehicle (70% ethanal, 30% polyethylene glycol, and 0.05% butylated hydroxytoluene) and retinol (ROL, vitamin A, 0.4%) [14]. Vehicle and retinol were applied topically to sun-protected buttock skin for 7days under occlusion to prevent loss and exposure to light. Immunohistology Immunohistology was performed as described previously [15]. Briefly, skin OCT-embedded cryo-sections (7μm) were fixed in paraformaldehyde. Subsequently, the slides were incubated for 1 hour at room temperature with normal control serum followed by incubation of primary antibodies against Ki 67, CD31, type I procollagen, fibronectin, TGF-b1, and CTGF/CCN2 (Santa Cruz Biotechnology, CA, USA). To verify antibody specificity, antibodies were incubated with antigenic peptides (Santa Cruz Biotechnology, CA, USA) for 30 minutes prior to addition to skin sections. All sections were lightly counterstained with hematoxylin, and were mounted with mounting media (Vector, Laboratories, CA, USA). The intensity of positive staining was quantified by computerized image analysis (Image-pro Plus software, version 4.1, Media Cybernetics, MD, USA). RNA isolation and Northern blot analysis Total RNA from human skin was extracted with a commercial kit (RNeasy Midi Kit, Qiagen, Chatsworth, CA) according to the manufacturer’s protocol. Samples of total RNA (30–50 μg) were resolved by 1.2% agarose electrophoresis, transferred to nylon membranes, and hybridized with TGF-β1 [16] and CCN2 CTGF/CCN2 [17] cDNA probes labeled with [32P]dCTP by random priming. Each blot was stripped and re-hybridized with 36B4 internal control gene transcript to monitor the sample load in each lane. The intensities of each band were quantified by STORM PhosphorImager (Molecular Dynamics, Sunnyvale, CA, USA) and normalized to the 36B4 gene transcript. In situ hybridization pCDNA3.1 (Invitrogen, Carlsbad, CA) plasmids containing human TGF-β1 and CTGF/CCN2 were linearized with NotI and BamHI. Digoxigenin-containing sense and antisense riboprobes were synthesized using SP6 and T7 ribonucleic polymerase. Frozen skin sections (5 μm) were mounted, fixed, treated, and hybridized as previously described [17]. The hybridization signals were detected immunohistochemically by alkaline phosphatase-conjugated antidigoxigenin antibody. RNA isolation and quantitative real-time RT-PCR Total RNA was extracted from human skin by using a commercial kit (RNeasy mini kit, Qiagen, Chatsworth, CA, USA), as previously described [18]. 200ng total RNA was reverse transcribed using Taqman Reverse Transcription kit (Applied Biosystems, Foster City, CA, USA). Real-time RT-PCR was performed using a Taqman Universal PCR Master Mix kit (Applied Biosystems, Foster City, CA, USA) and 7700 Sequence Detector (Applied Biosystems, Foster City, CA, USA). All PCR primers and probes were purchased from Applied Biosystems (Assays-on-Demand™ Gene Expression Products, Applied Biosystems, Foster City, CA, USA). Target gene mRNA levels were normalized to the housekeeping gene 36B4 (a ribosomal protein used as an internal control for quantitation) levels as an internal control. Western blot analysis Whole cell proteins were extracted from human skin dermis. The epidermis was removed by cutting off epidermis at a depth of 1 mm using cryostat. Equal amounts of protein (~50μg/lane) were resolved by 6–12% gradient sodium dodecyl sulfate-polyacrylamide (SDS) gel electrophoresis and transferred to polyvinylidene difluoride membranes. Membranes were then blocked with PBST (0.1% Tween 20 in PBS) containing 5% nonfat milk for one hour at room temperature. Smad7 primary antibody was generously provided by Dr. Ten Dijke (Leiden University Medical Centrum, Leiden, Netherlands). Primary antibody was incubated for one hour at room temperature. The membranes were washed three times with PBST solution and incubated with rabbit secondary antibody for one hour at room temperature. After washing three times with PBST, the membranes were developed with ECF (Vistra ECF Western blotting system, GE Health Care, Piscataway, NJ, USA) following the manufacturer’s protocol. The membranes were scanned with a STORM PhosphorImager (Molecular Dynamics, Sunnyvale, CA, USA), the intensities of each band were quantified using ImageQuant software (GE Health Care, Piscataway, NJ, USA) and normalized to β-actin (control). Atomic force microscopy (AFM) imaging Human skin biopsies were embedded in OCT, and cryo-sections (10 μm) were attached onto the microscope cover glass (1.2 mm diameter, Fisher Scientific Co., Pittsburgh, PA). These AFM samples were allowed to air dry for at least 24 hours before imaging them to AFM analysis. Images were taken by Dimension Icon AFM system (Bruker-AXS, Santa Barbara, CA, USA) using a silicon AFM probe (PPP-BSI, force constant 0.01–0.5N/m, resonant frequency 12–45kHz, NANOSENSORS™, Switzerland). AFM was conducted at the Electron Microbeam Analysis Laboratory (EMAL), University of Michigan College of Engineering, and analyzed using Nanoscope Analysis software (Nanoscope Analysis v120R1sr3, Bruker-AXS, Santa Barbara, CA, USA). Statistical analysis Comparisons between groups were determined with the Student’s t-test. All p values are two-tailed, and considered significant when p<0.05. Results Topical ROL increases epidermal thickness and dermal vascularity by proliferation of epidermal keratinocytes and dermal endothelial cells, respectively, in aged human skin in vivo As thin epidermis is a characteristic feature of aged human skin, we found that epidermal thickness was greatly improved following ROL treatment (Fig. 1A). Quantitative morphometric analyses revealed that epidermal thickness was increased 2.1-fold by ROL treatment (Fig. 1B). Consistent with thickened epidermis, keratinocyte proliferation, assessed by Ki67 immunostaining (Fig. 1C upper right panel), was increased 12-fold by ROL treatment (Fig. 1D). Interestingly, in addition to epidermal changes, we noticed topical ROL markedly increased the proliferation of dermal stromal cells (Fig. 1C lower right panel, Fig. 1E). No Ki67 positive cell was observed in vehicle-treated aged skin dermis (Fig. 1C lower left panel, Fig. 1E). Ki67 positive staining cells were particularly strong in papillary dermis. To confirm the identity of proliferating dermal stromal cells, we performed double-label immunofluorescence staining. As papillary dermis contains rich vascular networks, we found the majority of Ki67 positive cells were stained with CD31, a marker of endothelial cells (Fig. 1F). This result was further supported by increased prominence of blood vessels, confirmed by CD31 immunostaining (Fig. 1G). Quantitative analyses revealed that blood vessel formation was increased 3.8-fold in ROL-treated aged human skin (Fig. 1H). Together, these data indicate that topical ROL increases epidermal thickness and dermal angiogenesis by stimulating the proliferation of epidermal keratinocytes and dermal endothelial cells, respectively, in aged human skin in vivo. ROL improves dermal ECM microenvironment by stimulating the expression of type I collagen, fibronectin, and tropoelastin in aged human skin in vivo As thin dermis is a prominent feature of aged skin dermis, we analyzed the expression of major ECM proteins by immunohistochemistry. Type I collagen is the precursor of mature type I collagen fibrils. Compared with vehicle treated aged skin, ROL treatment increased type I collagen staining, the major structural protein in skin (Fig. 2A, right panels). Type I procollagen positive staining was particularly strong in epidermal and dermal junction (Fig. 1A lower right panel). Quantitative analysis indicated that the amount of staining was increased 3-fold by topical ROL treatment (Fig. 2B). In addition to type I collagen, topical ROL markedly increased the expression of fibronectin (FN) (Fig. 2C, right panels) and tropoelastin (Fig. 2E, right panels). Quantification indicated that fibronectin (Fig. 2D) and elastin (Fig. 2F) staining were increased 2.2-fold and 4-fold, respectively. These data indicate that topical ROL improves dermal ECM microenvironment by stimulating the major ECM proteins in aged human skin in vivo. Elevated epidermal-specific c-Jun transcription factor by topical ROL in aged human skin in vivo Next, we explored the potential mechanisms by which topical ROL stimulates keratinocytes proliferation in aged human skin. It has been reported that AP-1 transcription factor plays a major role in keratinocytes proliferation in response to growth factors, cytokines, and various stimuli [2]. As AP-1 complex largely comprised of c-Jun/c-fos, we examined the expression c-Jun and c-Fos in aged human skin after topical ROL treatment. We found that topical ROL markedly increased epidermal-specific c-Jun protein positive staining (Fig. 3A right panel) along with significant epidermal thickness. No c-Jun positive cell was observed in the dermis. Quantitative analyses revealed that c-Jun positive staining was increased 2.3-fold in ROL-treated aged human skin (Fig. 3A right panel). In contrast, no change of c-Fos protein expression was observed by topical ROL treatment (Fig. 3B). Western analysis further confirmed that while c-fos was constitutively expressed (Fig. 3D), c-Jun protein was elevated 3-fold (Fig. 3C) by topical ROL treatment. These data suggest that topical ROL elevates epidermal-specific c-Jun transcription factor, which in turn drives epidermal keratinocytes proliferation in aged human skin in vivo. Elevated TGF-β/CTGF pathway by topical ROL in aged human skin in vivo Next, we explored the potential mechanisms by which ROL stimulates dermal ECM production in aged human skin. We first examined whether topical ROL is able to activate dermal fibroblasts, the major ECM producing cells in skin, by staining with HSP47, the marker of fibroblast activation [19]. We observed that topical ROL markedly increased HSP47 positive staining, suggesting the activation of dermal fibroblasts in aged human skin in vivo. (Fig 4A, right panel). Quantitative analyses revealed that HSP47 positive staining was increased 3.8-fold in ROL-treated aged human skin (Fig. 4A right panel). It has been well-documented that activated dermal fibroblasts increase ECM production through TGF-β pathway, the major regulator of ECM production [20]. Therefore, we investigated the possible involvement of TGF-β pathway in increased ECM production by topical ROL in aged human skin. To test this possibility, we examined the effect of topical ROL on TGF-β pathway components such as TGF-β receptors, ligands, and Smads proteins. Interestingly, topical ROL significantly upregulated TGF-β1 mRNA and down-regulated the inhibitory Smad7, while other TGF-β pathway components remained unchanged (Fig. 4B). To further confirm these results, we performed immunostaining of TGF-β1 and its downstream mediator connective tissue growth factor (CTGF). In situ hybridization indicated that intense staining of TGF-β1 (Fig. 4C) and CTGF/CCN2 (Fig. 4D). Quantification indicated that TGF-β1 (Fig. 4C right panel) and CTGF (Fig. 4D right panel) dermal staining were increased 3.1-fold and 2.8-fold, respectively, following topical ROL treatment. Northern analysis further confirmed that TGF-β1 (Fig. 4E) and CTGF/CCN2 (Fig 4F) transcription were increased 2.6-fold and 3.8-fold, respectively, after topical ROL treatment. Consistent with mRNA expression, immunostaining of TGF-β1 (Fig. 4G) and CTGF/CCN2 (Fig. 4H) were increased 2.7-fold and 2.8-fold, respectively, after topical ROL treatment. We further confirmed that ROL treatment significantly reduced Smad7 mRNA (Fig. 4I) and protein (Fig. 4J), a potent inhibitor of TGF-β signaling. These data suggest that topical ROL stimulates dermal fibroblasts ECM production through upregulation of TGF-β/CTGF pathway in aged human skin. Topical ROL increases the deposition of mature collagen in aged human skin in vivo Having found that ROL is able to stimulate collagen production, we next assessed the collagen fibrils nanoscale structure and surface topology by atomic force microscopy (AFM). In vehicle-treated aged skin, collagen fibrils appeared disorganized and fragmented (Fig. 5A, left panel). However, topical ROL-treated aged skin revealed highly organized, dense bundles of collagen fibrils, with characteristic banded structure (D-spacing) representing the staggered alignment of individual collagen molecules within fibrils (Fig. 5A, right panel). Three-dimensional topographical AFM images further indicated that vehicle-treated aged skin dermis was much rougher and uneven (Fig. 5B, left panel). In contract, ROL-treated aged skin dermis was smooth and flattened (Fig. 5B, right panel). Quantitative analysis of AFM data indicated that the average roughness (a measure of fibril organization) of dermal collagen fibrils in ROL-treated aged skin is 48% less than in vehicle-treated aged skin dermis (Fig. 5B, 16 nm vs. 31 nm). These findings indicate that ROL treatment stimulates synthesis of procollagen, which is processed into mature collagen in aged human skin. Discussion Retinoids is the term used for the group of vitamin A derivatives [21]. Retinoic acid is the biologically active form of vitamin A which is also known as retinol (ROL), a precursor to retinoic acid. Human skin has the capacity to convert ROL to its biologically active metabolite retinoic acid. When topically applied to human skin, it penetrates and is sequentially converted to retinaldehyde and then to retinoic acid. One important benefit of topical ROL is that compared to retinoic acid, ROL induces fewer signs of side effects, which is characterized by erythema, scaling, dryness, and pruritis [22, 23]. Our results indicate that although a much higher concentration of ROL (0.4%) is needed to achieve similar results as seen with topical retinoic acid, ROL induced the same histological changes (epidermal hyperplasia and dermal ECM production) as retinoic acid without causing retinoids irritation, the most common side effect of retinoids. Mechanistically, it appears that topical ROL increases epidermal thickness by stimulating keratinocytes proliferation, which involves epidermal-specific upregulation of c-Jun transcription factor. Topical ROL also significantly elevates ECM production, which involves enhanced TGF-β/CTGF pathway in the dermis. Additionally, topical ROL significantly increased dermal vascularity which may have a significant impact on the homeostasis of epidermis and dermis. Although topical retinoids have been widely used in the treatment of skin aging, the precise mechanism(s) by which retinoids lead to visible improvement in aged skin is not fully understood. One significant finding of our study is that topical ROL improves aged dermal ECM microenvironment through enhancement of TGF-β/CTGF pathway in the dermis, the principle regulator of ECM homeostasis [20]. ROL is able to enhance TGF-β/CTGF pathway through two independent pathways: increases of the expression of TGF-β1/CTGF and suppression of inhibitory Smad7 of TGF-β signaling (Fig. 3). In primary human skin fibroblasts, the major ECM-producing cells in skin, the expression of major ECM proteins is regulated by TGF-β/CTGF pathway [7, 20]. For example, neutralization of endogenous TGF-β or knockdown of CTGF substantially reduced the expression of type I procollagen and fibronectin mRNA and protein. In contrast, overexpression of inhibitory Smad7, which is significantly reduced by topical ROL (Fig. 3B, I, J), abolished TGF-β or CTGF stimulation of ECM production. Importantly, we have previously reported that TGF-β/CTGF pathway is significantly reduced in dermal fibroblasts, in aged (80+ years) human skin in vivo [7, 20, 24, 25]. This impaired TGF-β/CTGF pathway in aged human skin is largely caused by reduced expression of TGF-β1 and CTGF. As impaired TGF-β/CTGF pathway significantly contributes to thinned dermis, our data suggest that the ability of topical ROL to improve aged skin dermis involves enhanced TGF-β/CTGF pathway. Reduced vasculature and epidermal thinning contribute significantly to skin fragility and impaired wound healing in aged skin [26, 27]. AP-1 transcription factor plays a major role in keratinocytes proliferation [13]. We found that while c-Fos is constitutively expressed in human skin, topical ROL elevates epidermal-specific c-Jun transcription factor. These data suggest that topical ROL may stimulate keratinocytes proliferation by epidermal-specific upregulation c-Jun transcription factor in aged human skin. Our study added another new piece of information that topical ROL improves dermal microenvironment through not only stimulation of ECM production, but also expansion of vasculature by proliferation of endothelial cells in aged human skin (Fig. 1C, F, G). A reduction of the cutaneous vasculature has been reported in aged human skin [27]. Increased vascularity of the dermis by topical ROL could increase skin blood flow and provide a more suitable microenvironment for the homeostasis of skin epidermal and dermal homeostasis. In contrast, it is also conceivable that proliferation of epidermal keratinocytes and restoration of ECM production by topical ROL may provide a more favorable microenvironment for the proliferation of endothelial cells and dermal vasculature. Indeed epidermal keratinocytes are important source of vascular endothelial growth factor (VGEF), a potent angiogenic factor [27]. Additionally, it has been reported that enhanced dermal ECM production resulted in stimulation of endothelial cell proliferation [19]. Clearly, additional studies are warranted to uncover the precise molecular mechanism(s) by which topical ROL stimulates the proliferation of epidermal keratinocytes and dermal endothelial cells in aged human skin. The molecular mechanisms by which retinoids improve aged human skin have been difficult to investigate largely due to lack of appropriate in vitro models. For example, cells in both the epidermis and dermis contain all the proteins and receptors that mediate the biological effects of vitamin A metabolites in the skin. However, primary keratinocytes or dermal fibroblasts are minimally responsive to RA treatment in monolayer culture [14, 28]. This lack of responsiveness is due, at least in part, to low levels of nuclear retinoid receptors, which mediate expression of retinoid-regulated genes [29, 30]. Skin equivalent cultures could be a useful model to investigate retinoids regulation of collagen homeostasis [14]. Skin equivalent cultures are composed of stratified, differentiated keratinocytes (model epidermis) on top of a type I collagen lattice, in which dermal fibroblasts are embedded (model dermis). Treatment of skin equivalent cultures with retinoic acid significantly increased the number of keratinocyte layers and dermal response, similar to the effect of topical application of RA to human skin in vivo. These data suggest that skin equivalent cultures could be a useful model for investigating mechanisms by which retinoids improve aged-appearing skin in human. Together, our findings extend current knowledge of retinoids anti-aging mechanisms in aged human skin. We propose a working model for the molecular basis of retinol anti-aging properties in aged human skin (Fig. 5C). Topical application of 0.4% ROL to aged human skin leads to remarkable skin changes in both epidermis and dermis through affecting three major types of skin cells, epidermal keratinocytes, dermal endothelial cells and fibroblasts. Topical ROL significantly increases epidermal thickness by stimulating epidermal keratinocytes proliferation, which involves c-Jun transcription factor, a major deriving force for keratinocyte proliferation. In addition to epidermal changes, topical ROL significantly improves dermal ECM microenvironment; increasing dermal blood vessel formation by stimulating endothelial cells proliferation and ECM production by activating fibroblasts. Topical ROL also stimulates TGF-β/CTGF pathway, the major regulator of ECM homeostasis, and thus increased the deposition of mature collagen in aged human skin in vivo. Increased dermal vasculatures could increase skin blood flow and provide a more active microenvironment for both epidermal keratinocyte proliferation and dermal fibroblasts activation. On the other hand, the proliferation of epidermal keratinocytes may also promote dermal blood vessel growth through stimulation of VEGF expression [27]. Additionally, the restoration of dermal ECM may provide a better, more permissive environment for the proliferation of dermal endothelial cells and epidermal keratinocytes, and activation of dermal fibroblasts (TGF-β/CTGF pathway) [19]. We propose that coupling of the proliferation of keratinocytes and endothelial cells, and dermal fibroblasts activation forms a self-enforcing environment, which might explain the remarkable anti-aging effects of ROL in aged human skin. Proposed model extends current understanding of the retinol anti-aging mechanisms in human skin. We thank Suzan Rehbine for the procurement of tissue specimens and Diane Fiolek for graphic and administrative assistance. We thank Electron Microbeam Analysis Laboratory (EMAL), University of Michigan College of Engineering for AFM. This work was supported by the National Institute of Health (AG019364 to GJ Fisher and T Quan; ES014697 to T Quan), Dermatology Foundation Research grant (to T Quan), and a grant from the Johnson &Johnson Corporation. Abbreviations ROL retinol ECM extracellular matrix AFM atomic force microscopy FN fibronectin Figure 1 Topical ROL increases epidermal thickness and dermal vascularity by proliferation of epidermal keratinocytes and dermal endothelial cells, respectively, in aged human skin in vivo. OCT-embedded skin sections (7μm) were obtained from aged (76±6 years) healthy sun-protected buttock skin after topical treatment of vehicle and 0.4% retinol for seven days. (A) H&E staining. Representative images of twelve individuals (N=12). White arrows indicate epidermal thickness. Bars=100μm. (B) Quantification of epidermal thickness (μm). (C) Ki67 immunostaining. 3.0× enlargement of the boxed region is shown to lower panels. Bars= 100μm. Representative images of twelve individuals (N=12). Quantification of epidermal (D) and dermal (E) Ki 67 immunostaining. (F) Ki67 and CD31 co-immunofluoresce staining. Skin sections were co-immunofluoresce stained with Ki67 and CD31, a marker of endothelial cells. Arrows indicate double stained cells. Representative of five individuals. Bar=50μm. (G) CD31 immunostaining. Representative images of twelve individuals (N=12). Bar=100μm. (H) Quantification of CD31 immunostaining. All immunostainings were quantified by computerized image analysis (Image-pro Plus software, version 4.1, Media Cybernetics, MD) and data are expressed as mean±SEM, *p<0.05. N=12. Figure 2 ROL improves dermal ECM microenvironment in aged human skin in vivo. OCT-embedded skin sections (7μm) were obtained from aged (76±6 years) healthy sun-protected buttock skin after topical treatment of vehicle and 0.4% retinol for seven days. (A) Type I procollagen immunostaining. 3× enlargement of the boxed region is shown to lower panels. Representative images of twelve individuals (N=12). Arrows indicate positive cells. Bars=100μm. (B) Quantification of type I procollagen. (C) Fibronectin immunostaining. 2.5× enlargement of the boxed region is shown to lower panels. Representative images of twelve individuals (N=12). Arrows indicate positive cells. Bars= 100μm. (D) Quantification of fibronectin. (E) Tropoelastin immunostaining. 3.0× enlargement of the boxed region is shown to lower panels. Representative images of twelve individuals (N=12). Bars=100μm. (F) Quantification of tropoelastin. All immunostainings were quantified by computerized image analysis (Image-pro Plus software, version 4.1, Media Cybernetics, MD) and data are expressed as mean±SEM, *p<0.05. N=12. Figure 3 Elevated epidermal-specific c-Jun transcription factor by topical ROL in aged human skin in vivo. OCT-embedded skin sections (7μm) were obtained from aged (76±6 years) healthy sun-protected buttock skin after topical treatment of vehicle and 0.4% retinol for seven days. (A) c-Jun immunostaining. Representative images of twelve individuals (N=12). Bars=100μm. (B) c-Fos immunostaining. Representative images of twelve individuals (N=12). Bars=100μm. All immunostainings were quantified by computerized image analysis (Image-pro Plus software, version 4.1, Media Cybernetics, MD) and data are expressed as mean±SEM, *p<0.05. N=12. (C) c-Jun protein. (D) c-Fos protein. Protein levels were determined by Western analysis. Protein levels were normalized to β-actin (loading control). Insets show representative Western blots. Mean ± SEM, N=6, *p < 0.05. Figure 4 Topical ROL stimulates TGF-β/CTGF pathway, the major regulator of ECM homeostasis, in aged human skin in vivo. OCT-embedded skin sections (7μm) were obtained from aged (76±6 years) healthy sun-protected buttock skin after topical treatment of vehicle and 0.4% retinol for seven days. (A) HSP47 immunostaining. Representative images of twelve individuals (N=12). Bars=100μm. (B) TGF-β pathway components mRNA levels. Total RNA was prepared from human skin samples. mRNA levels were determined by real-time RT-PCR. mRNA levels were normalized to 36B4 (internal housekeeping gene control). Mean ± SEM, N=12, *p<0.05. (C) TGF-β1 (D) CTGF/CCN2 in situ hybridization. Representative images of twelve individuals (N=12). Bars= 100μm. (E) TGF-β1 (F) CTGF/CCN2 Northern analysis. The intensities were quantified and normalized using 36B4 as loading control. Insets show representative Northern blots. Data are expressed as mean±SEM, *p<0.05. N=12. (G) TGF-β1 (H) CTGF/CCN2 immunostaining. Representative images of twelve individuals (N=12). Bars= 100μm. (I) Smad7 Northern analysis. The intensities were quantified and normalized using 36B4 as loading control. Inset shows representative Northern blots. Data are expressed as mean±SEM, *p<0.05. N=12. (J) Smad7 protein. Smad7 protein levels were determined by Western analysis. Protein levels were normalized to β-actin (loading control). Insets show representative Western blots. Mean ± SEM, N =6, *p < 0.05. All positive staining was quantified by computerized image analysis (Image-pro Plus software, version 4.1, Media Cybernetics, MD). Data are expressed as mean±SEM, *p<0.05. N=12. Figure 5 Deposition of mature collagen is increased by topical ROL in aged human skin in vivo. OCT-embedded skin sections (7μm) were obtained from aged (76±6 years) healthy sun-protected buttock skin after topical treatment of vehicle and 0.4% retinol for seven days. (A) Nanoscale collagen fibrils were imaged by AFM. Representative AFM images were shown. N=12. The blue arrows indicate intact collagen fibrils and red arrows heads indicate damaged collagen fibrils. (B) Three dimensional collagen fibrils. Dermal roughness was analyzed using Nanoscope Analysis software (Nanoscope_Analysis_v120R1sr3, Bruker-AXS, Santa Barbara, CA). All results are expressed as the mean ± SEM, N=12, *p < 0.05. Bars=100 nm. (C) Proposed model for topical ROL exerts anti-aging effects in aged human skin by proliferation of epidermal keratinocytes and dermal endothelial cells and activation of dermal fibroblasts (see Discussion for details). 1 Quan T Skin connective tissue aging and dermal fibroblasts Dermal Fibroblasts: Histological Perspectives, Characterization and Role in Disease 2013 31 55 2 Fisher GJ Mechanisms of photoaging and chronological skin aging Arch Dermatol 2002 138 11 1462 70 12437452 3 Thomas DR Burkemper NM Aging skin and wound healing Clin Geriatr Med 2013 29 2 xi xx 23571044 4 Worley CA Aging skin and wound healing Dermatol Nurs 2006 18 3 265 6 16856682 5 Bissell MJ Hines WC Why don’t we get more cancer? A proposed role of the microenvironment in restraining cancer progression Nat Med 2011 17 3 320 9 21383745 6 Bissell MJ Kenny PA Radisky DC Microenvironmental regulators of tissue structure and function also regulate tumor induction and progression: the role of extracellular matrix and its degrading enzymes Cold Spring Harb Symp Quant Biol 2005 70 343 56 16869771 7 Quan T Fisher GJ Role of Age-Associated Alterations of the Dermal Extracellular Matrix Microenvironment in Human Skin Aging: A Mini-Review Gerontology 2015 61 5 427 34 25660807 8 Yaar M Eller MS Gilchrest BA Fifty years of skin aging J Investig Dermatol Symp Proc 2002 7 1 51 8 9 Smith JG Jr Alterations in human dermal connective tissue with age and chronic sun damage J Invest Dermatol 1962 39 347 50 13993162 10 Uitto J Bernstein EF Molecular mechanisms of cutaneous aging: connective tissue alterations in the dermis J Investig Dermatol Symp Proc 1998 3 1 41 4 11 Jacob MP Extracellular matrix remodeling and matrix metalloproteinases in the vascular wall during aging and in pathological conditions Biomed Pharmacother 2003 57 5–6 195 202 12888254 12 Marastoni S Extracellular matrix: a matter of life and death Connect Tissue Res 2008 49 3 203 6 18661343 13 Fisher GJ Molecular mechanisms of photoaging in human skin in vivo and their prevention by all-trans retinoic acid Photochem Photobiol 1999 69 2 154 7 10048311 14 Quan T Retinoids suppress cysteine-rich protein 61 (CCN1), a negative regulator of collagen homeostasis, in skin equivalent cultures and aged human skin in vivo Exp Dermatol 2011 20 7 572 6 21488975 15 Quan C Dermal fibroblast expression of stromal cell-derived factor-1 (SDF-1) promotes epidermal keratinocyte proliferation in normal and diseased skin Protein Cell 2015 6 12 890 903 26296527 16 Quan T Ultraviolet irradiation alters transforming growth factor beta/smad pathway in human skin in vivo J Invest Dermatol 2002 119 2 499 506 12190876 17 Quan T Connective tissue growth factor: expression in human skin in vivo and inhibition by ultraviolet irradiation J Invest Dermatol 2002 118 3 402 8 11874477 18 Quan C Age-associated reduction of cell spreading induces mitochondrial DNA common deletion by oxidative stress in human skin dermal fibroblasts: implication for human skin connective tissue aging J Biomed Sci 2015 22 62 26215577 19 Quan T Enhancing structural support of the dermal microenvironment activates fibroblasts, endothelial cells, and keratinocytes in aged human skin in vivo J Invest Dermatol 2013 133 3 658 67 23096713 20 Quan T Reduced expression of connective tissue growth factor (CTGF/CCN2) mediates collagen loss in chronologically aged human skin Journal of Investigative Dermatology 2010 130 2 415 424 19641518 21 Blomhoff R Transport and metabolism of vitamin A Nutr Rev 1994 52 2 Pt 2 S13 23 8202278 22 Mukherjee S Retinoids in the treatment of skin aging: an overview of clinical efficacy and safety Clin Interv Aging 2006 1 4 327 48 18046911 23 Darlenski R Surber C Fluhr JW Topical retinoids in the management of photodamaged skin: from theory to evidence-based practical approach Br J Dermatol 2010 163 6 1157 65 20633013 24 Quan T Elevated cysteine-rich 61 mediates aberrant collagen homeostasis in chronologically aged and photoaged human skin Am J Pathol 2006 169 2 482 90 16877350 25 Fisher GJ Reduction of fibroblast size/mechanical force down-regulates TGF-beta type II receptor: implications for human skin aging Aging Cell 2016 15 1 67 76 26780887 26 Zouboulis CC Makrantonaki E Clinical aspects and molecular diagnostics of skin aging Clin Dermatol 2011 29 1 3 14 21146726 27 Chung JH Eun HC Angiogenesis in skin aging and photoaging J Dermatol 2007 34 9 593 600 17727362 28 Elder JT Differential regulation of retinoic acid receptors and binding proteins in human skin J Invest Dermatol 1992 98 5 673 9 1314862 29 Di W Keratinocyte-specific retinoid regulation of human cellular retinoic acid binding protein-II (hCRABPII) gene promoter requires an evolutionarily conserved DR1 retinoic acid-responsive element J Invest Dermatol 1998 111 6 1109 15 9856825 30 Fisher GJ Immunological identification and functional quantitation of retinoic acid and retinoid X receptor proteins in human skin J Biol Chem 1994 269 32 20629 35 8051161

PMC005xxxxxx/PMC5136774.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 0372354 646 Ann Surg Ann. Surg. Annals of surgery 0003-4932 1528-1140 25371118 5136774 10.1097/SLA.0000000000001001 NIHMS830796 Article Temporal Patterns of Circulating Inflammation Biomarker Networks Differentiate Susceptibility to Nosocomial Infection Following Blunt Trauma in Humans Namas Rami A. MD *† Vodovotz Yoram PhD *†‡ Almahmoud Khalid MD * Abdul-Malak Othman MD * Zaaqoq Akram MD § Namas Rajaie MD *¶ Mi Qi PhD || Barclay Derek BS * Zuckerbraun Brian MD * Peitzman Andrew B. MD * Sperry Jason MD, MPH * Billiar Timothy R. MD *† * Department of Surgery, University of Pittsburgh, Pittsburgh, PA † Center for Inflammation and Regenerative Modeling, McGowan Institute for Regenerative Medicine, University of Pittsburgh, Pittsburgh, PA ‡ Department of Computational and Systems Biology, University of Pittsburgh, Pittsburgh, PA § Department of Critical Care Medicine, University of Pittsburgh, Pittsburgh, PA ¶ Department of Internal Medicine, Division of Rheumatology, Wayne State University, Detroit, MI || Department of Sports Medicine and Nutrition, University of Pittsburgh, Pittsburgh, PA Reprints: Timothy R. Billiar, MD, Department of Surgery, University of Pittsburgh, F1281 Presbyterian University Hospital, 200 Lothrop St, Pittsburgh, PA 15213. billiartr@upmc.edu 19 11 2016 1 2016 05 12 2016 263 1 191198 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. Background Severe traumatic injury can lead to immune dysfunction that renders trauma patients susceptible to nosocomial infections (NI) and prolonged intensive care unit (ICU) stays. We hypothesized that early circulating biomarker patterns following trauma would correlate with sustained immune dysregulation associated with NI and remote organ failure. Methods In a cohort of 472 blunt trauma survivors studied over an 8-year period, 127 patients (27%) were diagnosed with NI versus 345 trauma patients without NI. To perform a pairwise, case-control study with 1:1 matching, 44 of the NI patients were compared with 44 no-NI trauma patients selected by matching patient demographics and injury characteristics. Plasma obtained upon admission and over time were assayed for 26 inflammatory mediators and analyzed for the presence of dynamic networks. Results Significant differences in ICU length of stay (LOS), hospital LOS, and days on mechanical ventilation were observed in the NI patients versus no-NI patients. Although NI was not detected until day 7, multiple mediators were significantly elevated within the first 24 hours in patients who developed NI. Circulating inflammation biomarkers exhibited 4 distinct dynamic patterns, of which 2 clearly distinguish patients destined to develop NI from those who did not. Mediator network connectivity analysis revealed a higher, coordinated degree of activation of both innate and lymphoid pathways in the NI patients over the initial 24 hours. Conclusions These studies implicate unique dynamic immune responses, reflected in circulating biomarkers that differentiate patients prone to persistent critical illness and infections following injury, independent of mechanism of injury, injury severity, age, or sex. dynamic network analysis high-mobility group protein B1 injury severity score intensive care unit interleukin multiple organ failure nosocomial infection Traumatic injury, often accompanied by hemorrhage, is the leading cause of death in patients younger than 45 years, and it represents a significant source of morbidity and mortality for all ages.1 In patients who survive beyond the initial few hours after injury, early and sustained multiple organ failure (MOF) and delayed nosocomial infection (NI) are leading causes of late death2–4 and contribute to prolonged and resource-intensive hospital stays.5,6 It is currently unknown why some patients develop NI while other patients—with apparently similar demographics and injury characteristics—do not. One key mechanism by which traumatic injury contributes to an increased susceptibility to NI is through impairment of host defense mechanisms,7,8 in particular, dysregulation of the trauma-induced inflammatory response. However, the precise mechanisms underlying this impairment remain unclear,9,10 nor is it clear how dysregulated inflammation predisposes some patients to NI. Though inflammation is generally well-regulated, the posttraumatic inflammatory response can be either greater or lesser than that required to respond to the initial insult, resulting in immune dysregulation.11–13 Rather than promoting resolution and healing, this dysregulated inflammatory response may result in secondary damage to various tissues and organs,11,14,15 coining the term “second hit,” which renders trauma patients vulnerable to organ dysfunction and NI. Inflammatory chemokines, cytokines, free radicals, and damage-associated molecular pattern molecules are the key regulators of the host inflammatory response to tissue injury after trauma.11,16 Previous studies17–20 have suggested an exaggerated systemic inflammatory response in patients who develop organ dysfunction. A recent genome-wide study examining the transcriptomic response within circulating leukocytes in severely injured humans identified distinct patterns between patients that followed either uncomplicated or complicated 28-day clinical course. In both patient cohorts, there was a simultaneous upregulation of transcripts associated with innate immunity, and downregulation in the expression of genes linked to adaptive immunity. This pattern was exaggerated both in terms of magnitude and duration in patients with complications. Thus, the magnitude of the immune response is one factor associated with complications after trauma. However, biomarker patterns that yield mechanistic insights or that predict MOF or NI have not been studied rigorously. To address the phenotypic complexity of the trauma-induced inflammation associated with higher susceptibility to NI, we analyzed data retrospectively from a large cohort of blunt trauma survivors (472 patients) studied over an 8-year period. To reduce confounding variables, we derived stringently matched subcohorts of NI and no-NI patients that still reflected the primary demographic and injury characteristics of the original large cohort. Our analyses revealed that early, persistent changes in postinjury inflammation manifest in unique biomarker patterns postinjury in patients prone to develop infections. We propose that a combination of traditional parameters of inflammation such as white blood cell differential combined with a novel subset of circulating biomarkers measured upon presentation and over the initial 24 hours could be used to stratify trauma intensive care unit (ICU) patients for more or less intensive care and monitoring and hence optimize resource utilization and outcomes. MATERIALS AND METHODS Patient Enrollment and Data Collection All human sampling was carried out after approval by the University of Pittsburgh institutional review board, and informed consent was obtained from each patient or next of kin as per institutional review board regulations. Patients eligible for enrollment in the study were at least 18 years of age, admitted to the ICU after being resuscitated, and, per treating physician, were expected to live more than 24 hours. Reasons for ineligibility were isolated head injury or brain death criteria, pregnant women, and penetrating trauma. Laboratory results and other basic demographic data were recorded in the database via direct interface with electronic medical record. Three plasma samples, starting with the initial blood draw upon arrival, were assayed within the first 24 hours after injury and then from days 1 to 7 after injury. Clinical data, including Injury Severity Score (ISS), Abbreviated Injury Scale (AIS) score, Marshall Multiple Organ Dysfunction (MOD) score,21 ICU LOS, hospital LOS, and days on mechanical ventilation were collected from hospital inpatient electronic trauma registry database. AIS-05 (according to the updated 2005 injury code)22 and ISS23 scores were calculated for each patient by a single trauma surgeon after attending radiology evaluations were finalized. Study Design and Case Identification The study period ran from January 1, 2004, to May 1, 2012, during which time 493 blunt trauma patients were admitted to the emergency department of the Presbyterian University hospital, a level I trauma center which receives trauma patients directly from the accident scene or transfers from regional hospitals. This relatively small number of patients enrolled during the study period is attributed to the willingness of patients to consent and the aforementioned stringent selection criteria (see earlier). The overall demographics, mechanism of injury, comorbidities, and outcomes of the 493 patients are shown in Table S1, Supplemental Digital Content, http://links.lww.com/SLA/A677. After excluding 21 trauma in-hospital nonsurvivors, clinical data from 472 blunt trauma survivors were analyzed for the presence of NI. Using the United States Centers for Disease Control clinical criteria for diagnosis of NI,24 we identified 127 blunt trauma patients (prevalence = 27%) with NI. To perform a pairwise, retrospective case-control study with 1:1 matching, we initially sought to avoid confounding factors related to type of mechanism of injury by selecting the predominant mechanism of injury in both subcohorts. Accordingly, we selected patients in the motor vehicle accident (MVA) category, as MVA was the predominant mechanism of injury in NI and no-NI trauma patients (65.4% and 53.9%, respectively). Next, we further excluded patients who received blood transfusions or underwent emergency surgical procedures within the first 24 hours after trauma to avoid the confounding impact of these interventions on the overall inflammatory response. As a result of this stringent selection criteria, 44 NI trauma patients were identified and subsequently analyzed for their inflammatory biomarker profile. We next sought to select a control subcohort (no-NI) according to the following matching criteria: age (±5 years), sex, ISS calculated on hospital discharge (±5 points), similar mechanism of injury (MVA), and no history of blood transfusions or major surgical intervention within 24-hours after trauma. A control was required to be without evidence of NI at any time during hospitalization in the ICU. Accordingly, a computer-generated list of potential controls was obtained from a database including 345 trauma patients (see earlier). According to these patient-matching criteria, 44 trauma patients without NI were selected and subsequently compared their inflammatory biomarker profile to the 44 NI trauma patients (Table S3, Supplemental Digital Content, http://links.lww.com/SLA/A679). Analysis of Inflammation Biomarkers Blood samples were collected into citrated tubes via central venous or arterial catheters within 24 hours of admission and daily up to 7 days after injury. The blood samples were centrifuged, and plasma aliquots were stored in cryoprecipitate tubes at − 80°C for subsequent analysis of inflammatory mediators. The human inflammatory MILLIPLEX MAP Human Cytokine/Chemokine Panel-Pre-mixed 24-Plex (Millipore Corporation, Billerica, MA) and Luminex 100 IS (Luminex, Austin, TX) were used to measure plasma levels of interleukin (IL)-1β, IL-1 receptor antagonist (IL-1RA), IL-2, soluble IL-2 receptor-α (sIL-2Rα), IL-4, IL-5, IL-6, IL-7, IL-8 (CCL8), IL-10, IL-13, IL-15, IL-17A, interferon (IFN)-γ, IFN-γ inducible protein (IP)-10 (CXCL10), monokine induced by gamma interferon (MIG; CXCL9), macrophage inflammatory protein (MIP)-1α (CCL3), MIP-1β (CCL4), monocyte chemotactic protein (MCP)-1 (CCL2), granulocyte-macrophage colony stimulating factor (GM-CSF), Eotaxin (CCL11), and tumor necrosis factor alpha (TNF-α). The Luminex system was used in accordance to manufacturer’s instructions. High-mobility group protein B1 (HMGB1) measurement was performed using ELISA (Shino-Test Corp, Kanagawa, Japan, distributed by IBL international, Toronto, Ontario, Canada) according to the manufacturer’s instructions. Statistical Analysis All data were analyzed using SigmaPlot 11 software (Systat Software, Inc, San Jose, CA). Statistical difference between NI and no-NI groups was determined by either Student t test or χ2 test as appropriate. Group-time interaction of plasma inflammatory mediators’ levels between NI and no-NI groups was determined by 2-way analysis of variance (ANOVA). To quantify the differences between the statistically significant mediators, we calculated the area under the curve (AUC) using the mean values for each time point, then calculating NI/no-NI AUC fold change. P < 0.05 was considered statistically significant for all analyses. Data-driven Modeling: Dynamic Network Analysis (DyNA) The goal of this analysis was to gain insights into dynamic changes in network connectivity of the posttraumatic inflammatory response to NI and no-NI over time. The mathematical formulation of this method is essentially to calculate the correlation among variables by which we can examine their dependence. To do so, inflammatory mediator networks were created in adjacent 8-hour time periods (0–8 hours, 8–16 hours, and 16–24 hours) using MAT-LAB (The MathWorks, Inc, Natick, MA). Connections in the network were created if the correlation coefficient between 2 nodes (inflammatory mediators) was greater or equal to a threshold of 0.7. RESULTS Characteristics of NI and No-NI Subcohorts: Demographics and Outcomes Hypothesizing that distinct inflammation biomarker patterns could characterize patients who develop NI as a consequence of a unique immune phenotype, we identified 127 patients with NI and 345 patients without NI (Table S2, Supplemental Digital Content, http://links.lww.com/SLA/A678). Overall, males were predominant in both NI and no-NI subcohorts (66.1% and 71.3%, respectively) with no statistical difference in mean age (P = 0.7) between the 2 subcohorts. However, ISS was statistically significantly higher in the NI cohort (P < 0.001) than in the no-NI cohort. Interestingly, we observed a statistically significantly longer ICU LOS (P < 0.001), hospital LOS (P < 0.001), and days on mechanical ventilator (P <0.001) in the NI cohort when compared with the no-NI cohort. Description of Nosocomial Infections in NI Subcohort Overall, 127 of the ICU patients studied had at least 1 nosocomial infection, 15 cases (12%) had multiple infections: 2 nosocomial infections developed in 14 of these cases and 3 or more nosocomial infections developed in 1 case. During their ICU stay, the 127 identified case patients developed 141 episodes of nosocomial infections (1.1 episodes per patient) with an overall infection rate (number of infections per 100 admissions) of 30%. The sites of infection were as follows: 79 episodes of pneumonia (56%), 39 urinary tract infections (UTIs, 28%), 15 bloodstream infections (11%), 5 wound infections (3%), 2 empyema (1.6%), and 1 Clostridium difficile infection (0.4%). Of the 79 episodes of pneumonia, 66 were primary; the other 13 were complicated by the following nosocomial infections: 7 UTIs, 4 bloodstream infections, and 2 wound infections. To establish the diagnosis of suspected hospital-acquired pneumonia (HAP), we used clinical criteria that included new or progressive pulmonary infiltrates on radiograph after 48 hours of hospital admission, and 1 or more of the following: fever, leukocytosis, or leukopenia, an increase in purulent endotracheal secretions. Clinically suspected pneumonia was diagnosed when a bacterial cell culture of 10,000 or more colony forming units per milliliter of bronchoalveolar lavage fluid was grown. Differences in Circulating Inflammatory Mediators in NI Versus No-NI Patients We first sought to determine if there was a statistically significant difference in the levels of inflammatory mediators in either subcohort relative to baseline values. To establish baseline, inflammation biomarker data were obtained from 12 healthy volunteers matched to both subcohorts based on age and sex distribution (healthy volunteers; age: 46 ± 2.1; 8 men and 4 women) with no history of recent infections or existing comorbidities. This analysis showed that even upon admission, multiple inflammation biomarkers were statistically significantly different in both NI and no-NI sub-cohorts over the baseline levels (Figure S1, Supplemental Digital Content, http://links.lww.com/SLA/A673). We next hypothesized that dynamics of circulating inflammatory mediators would differ in patients with NI from patients without NI, especially with regard to early (<24 hours) inflammatory responses that might predispose to NI. We tested this hypothesis by analyzing an extensive time course of plasma inflammation biomarkers from onset of injury and up to 7 days after injury in both subcohorts. The overall analysis of the inflammation biomarkers over the 7 days’ course after injury between NI and no-NI subcohorts showed that multiple circulating levels of cytokines and chemokines were significantly higher in the NI cohort when compared with the no-NI cohort over the 7 day course (Figure S2, Supplemental Digital Content, http://links.lww.com/SLA/A673). Comparison of Stringently Matched NI and No-NI Subgroups Injury severity,25 the additional trauma of early surgical procedures,26 and blood component transfusions27 are all variables known to modulate the inflammatory response in blunt trauma patients. To determine if the susceptibility to NI was indeed associated with a unique temporal pattern of circulating inflammatory biomarkers independent of these aforementioned confounders, we matched 88 patients (44 from each sub-cohort) stringently according to age, sex distribution, and ISS and with no transfusions or major surgical interventions within the first 24-hours after injury (Table S3, Supplemental Digital Content, http://links.lww.com/SLA/A679). Among 44 NI patients, 27 were male and 17 were female (age: 48 ± 3, ISS: 26.3 ± 1.7), whereas among 44 no-NI patients, 27 were male and 17 were female (age: 47.3 ± 2.3, ISS: 26 ± 0.9). The extent of initial hypoperfusion was estimated by comparing blood lactate levels upon admission and over time in the NI and no-NI subgroups. Lactate levels were elevated in both groups upon admission being slightly but statistically significantly higher at 4 and 8 hours in the no-NI group and then gradually normalized over the initial 24-hours after injury (Fig. 1A). Importantly, these subgroups retained the key demographic and injury characteristics of the original, large sub-cohorts. Further supporting the validity of examining these stringently matched subgroups were the clinical outcomes: the NI group exhibited a prolonged ICU LOS (P < 0.001), hospital LOS (P < 0.001), and days on mechanical ventilation (P < 0.001) when compared with the no-NI group (Table S3, Supplemental Digital Content, http://links.lww.com/SLA/A679), independent of injury severity and major interventions within the first 24-hours after injury. Site of Infection and Mean Day of Diagnosis in NI Patients During their ICU stay, the 44-patient NI group developed 47 episodes of NI (1.1 infections per patient, as in the overall cohort): 22 episodes of pneumonia (47%), 14 UTIs (30%), 9 bloodstream infections (19%), and 2 wound infections (4%). Of the 22 episodes of pneumonia, 20 were primary; the other 2 were complicated with wound infections. The average time to the development of NI was 7 days after injury, and the mean day of diagnosis for each type of infection was as follows: pneumonia 6 ± 1 days, UTI 9 ± 2 days, bloodstream infections 7 ± 2 days, and wound infections 11 ± 4 days. Comparison of Injuries by Body Region in NI Versus No-NI Subgroups We next sought to calculate the AIS in NI and no-NI subgroups to identify whether specific body region injuries could be associated with an enhanced susceptibility to develop NI after trauma. The analysis of the injury patterns factored into the AIS revealed statistically significant differences in the head (P = 0.04) and chest (P = 0.01) regions in the NI group when compared with no-NI group (Fig. 1B). The clinical course, functional outcome, and total hospital LOS of trauma patients can be influenced substantially by the severity of brain injury after trauma.28 The Glasgow Coma Scale is one of the most common tools used for gradation of central nervous system injury severity using clinical observations.29 This analysis showed no statistical difference in the mean Glasgow Coma Scale score upon presentation (P = 0.08) between NI and no-NI subgroups (11.8 ± 0.7 and 13.6 ± 0.5, respectively). Multiple Organ Dysfunction and Circulating Leukocyte Patterns in NI and No-NI Patients The NI and no-NI subgroups differed in their degree of MOD, as indicated by the Marshall MOD score, a well-validated index of dysfunction in multiple organ systems,30,31 which was calculated at each time point in which inflammation biomarkers were assessed. This analysis suggested that a similar level of organ dysfunction was present in NI and no-NI groups on the first day. However, by day 2, the NI group exhibited a statistically significantly higher degree of organ dysfunction (P < 0.001) and then up to day 7 after injury when compared with the no-NI group (lesser degree of organ dysfunction) (Fig. 1C). This difference seems to be due to not only an early increase in organ dysfunction in the NI group but also a persistence in organ dysfunction that rapidly resolves in the no-NI group. We next assessed whether leukocyte populations could differ between the NI and no-NI subgroups after injury. Figure S2, Supplemental Digital Content, http://links.lww.com/SLA/A674, shows that the NI group exhibited statistically significantly higher total leukocyte counts at time of presentation and at days 8, 9, and 10 after injury when compared with no-NI group (Figure S2A, Supplemental Digital Content, http://links.lww.com/SLA/A674). Polymorphonuclear neutrophils (PMN) percentages were statistically significantly higher in the NI group at time of presentation and days 8, 9, 11, and 12 after injury when compared with no-NI group (Figure S2B, Supplemental Digital Content, http://links.lww.com/SLA/A674). Moreover, total lymphocyte percentages were statistically significantly higher in no-NI group at multiple time points including the initial blood draw when compared with NI subgroup (Figure S2C, Supplemental Digital Content, http://links.lww.com/SLA/A674). In addition, monocyte percentages were statistically significantly higher in no-NI group at days 8, 9, 10, 11, 12, and 14 after injury when compared with NI group (Figure S2D, Supplemental Digital Content, http://links.lww.com/SLA/A674). Unique Dynamic Patterns of Inflammation Biomarkers Emerge in NI Patients After Injury We next sought to determine if levels of circulating inflammation biomarkers over 7 days after injury would differentiate trauma patients who would develop NI from stringently matched no-NI patients. Accordingly, at least 3 plasma samples were obtained in the first 24 hours after injury, including upon arrival as well as daily up to day 7 after injury, which corresponds to the mean day of diagnosis of NI. Significant differences were observed in multiple inflammation biomarkers upon admission over the baseline (healthy volunteers) in both the NI and no-NI subgroups (Figure S3, Supplemental Digital Content, http://links.lww.com/SLA/A675). To determine difference in levels of inflammatory mediators between the NI and no-NI subgroups from time of admission and over the 7 days’ course after injury, the biomarker data were analyzed using 2-way ANOVA (see Materials and Methods). This extensive analysis revealed statistically significantly higher circulating levels of multiple cytokines and chemokines as well as HMGB1 in the NI group when compared with the no-NI group (Figure S3, Supplemental Digital Content, http://links.lww.com/SLA/A675). Importantly, this analysis revealed 4 distinct biomarker patterns; one that NI and no-NI patients share in common and 3 that clearly distinguish patients destined to develop NI from those who will not. These patterns are depicted qualitatively in Figure 2, and detailed data are shown in Figures S3A, S3B, S3C, and S3D, Supplemental Digital Content, http://links.lww.com/SLA/A675. In the first dynamic pattern, MIG/CXCL9, IL-10, IL-8/CCL8, and Eotaxin/CCL11 were elevated upon presentation in both NI and no-NI groups and then declined steadily within 24 hours and leveled off between days 1 and 7 after injury (Fig. 2A). In the second pattern, MCP-1/CCL2, IL-6, HMGB1 (Fig. 3), and IL-1RA were elevated upon presentation to a significantly greater degree in the NI group. Levels of these mediators declined sharply within 24 hours followed by moderate oscillations up to day 7 after injury (Fig. 2B). In the third pattern, mediators that were initially elevated minimally in both NI and no-NI patients began to increase early in the NI group in the second 12 hours after injury and then remained elevated to the end of the sampling period. This group included IL-7, IL-5, IL-17A, IL-4, IL-13, MIP-1α/CCL3, MIP-1β/CCL4, IFN-γ, IL-15, sIL-2Rα, GM- CSF, and IP-10/CXCL10 (Fig. 2C). Finally, in the fourth pattern, IFN-α, IL-1β, IL-2, and TNF-α were low on admission with minor but gradual elevation after 24 hours and up to day 7 after injury (Fig. 2D) only in the NI group. The aforementioned biomarker patterns suggested that the posttraumatic inflammatory response in patients who developed NI diverges early after injury from similarly injured patients who experience an uncomplicated clinical course. Accordingly, we sought to examine the total inflammatory mediator production across all time points in the first 24 hours by calculating the AUC in NI and no-NI subgroups. The AUC was calculated for each biomarker and expressed as fold change difference between NI and no-NI subgroups (see Materials and Methods). Subsequently, the biomarkers were ranked according to their fold values (from highest to lowest fold change). This analysis suggested that multiple biomarkers expressed an increased fold change over the first 24 hours after injury in the NI group when compared with the no-NI group (Table S4, Supplemental Digital Content, http://links.lww.com/SLA/A680). Thus, both the levels and patterns of mediator accumulation in the circulation, identified upon admission and over the first 24 hours, distinguish patients whowill develop a sustained systemic inflammatory response, persistent organ dysfunction, and susceptibility to NI from similarly injured patients that will not. Furthermore, the dynamic changes identify a profile of mediators present within hours of injury, known to be derived from both immune and non-immune cells, which gives way to mediators associated predominantly with lymphoid cells. Different Dynamic Networks of Systemic Inflammation Inferred in NI Versus No-NI Patients On the basis of these findings, we hypothesized that the differences in the early dynamic, systemic inflammatory response between NI and no-NI could be explained, at least in part, by differential network connectivity among inflammatory mediators. This analysis was achieved by examining the time-dependent evolution of cytokine networks inferred from correlated changes in circulating inflammatory mediators. In this study, we wished not only to determine which networks were present at specific time intervals but also to assess the total degree of connectivity at each of these intervals. Figure 4 shows the detailed DyNA results for NI and no-NI over 3 different time periods following presentation (0–8 hours, 8–16 hours, and 16–24 hours). This analysis suggested that the connectivity among central nodes (≥6 connections) evolved rapidly in the NI group, whereas the no-NI group exhibited substantial reduction of network connectivity (<6 nodes) over the initial 24 hours after injury. Finally, we sought to go beyond an examination of inflammatory mediators and assess the global state of inflammatory networks, by quantifying the degree of network connectivity as a function of time in NI and no-NI subgroups. The NI group exhibited a higher network density at multiple time points (Figure S4A, Supplemental Digital Content, http://links.lww.com/SLA/A676), whereas the no-NI group had initially a high-network density within 8 to 16 hours after injury, which declined over the 16- to 24-hour time period (Figure S4B, Supplemental Digital Content, http://links.lww.com/SLA/A676). Taken together, these analyses suggest a higher, coordinated degree of activation of both innate and lymphoid pathways in NI patients as compared with no-NI, a difference that can be observed as early as the first 8 hours after injury. DISCUSSION Traumatic injury results in biochemical and physiological changes that are in part dependent on the nature and magnitude of the accompanying inflammatory response.32 Though properly regulated, self-resolving inflammation allows for timely recognition and effective responses to injury, yet, if in excess, inflammation can go awry resulting in immune dysregulation and subsequently impairment of host physiological functions.11 This impairment is manifested by an early excessive and sustained systemic inflammatory response, which eventually becomes associated with persistent critical illness and a simultaneous persistent immune suppression with enhanced susceptibility to delayed infection.33 We found that the magnitude and pattern of early circulating inflammatory biomarkers distinguished between similarly injured patients who would either improve or experience persistent organ dysfunction and delayed NI. The observed patterns suggest that patients prone to follow a complicated course could be predicted on the basis of biomarker analysis performed at early time points. The patterns also provide insights into the unique nature of the immune-inflammatory response in patients that follow a more complicated course after injury. Multiple prior studies have reported a strong association between injury severity and an increased susceptibility to NI after injury, which in turn contributes to higher in-hospital morbidity.7,34 In our analysis of the overall patient cohort, we noted that patients with NI had a higher injury severity, which was associated with elevations in multiple circulating inflammatory mediators over the 7 days’ course after trauma when compared with the no-NI cohort. Although these results support the findings of these prior studies, they also pose a conundrum: Are changes in inflammation after injury related only to the severity of injury or to some other factors that predispose otherwise similarly injured patients to develop NI? We were also concerned about the potential confounding effect of invasive patient management on the inflammatory response. Our analyses using stringently matched cohorts recapitulate the clinical outcomes of the overall cohort, despite comparable degree of injury (ISS ~26). The notion that the magnitude of injury was comparable in these subcohorts is further supported by the observation that both groups had similar elevations in admission lactate levels and a subset of circulating inflammatory biomarkers (Pattern A, Figure 3A, Supplemental Digital Content, http://links.lww.com/SLA/A675). On the basis of the observed changes in circulating biomarker levels, we identified 4 distinct temporal patterns from time of admission and through 7 days after injury in the NI group. Pattern A is virtually identical in the 2 cohorts and is remarkable for the early elevations in the chemokines MIG/CXCL9, IL-8/CXCL8, and Eotaxin. This observation suggest that these mediators represent a common component of the immune response to injury, which is aimed at modifying leukocyte trafficking. Moreover, AUC analysis showed that approximately half of the biomarkers that had a higher fold change in the NI group exhibit dynamics that suggest 2 distinct patterns, which are unique between the subgroups in this early time frame. In particular, the much higher levels of IL-6, MCP-1/CCL2, HMGB1, and IL-1RA (Pattern B) upon admission in the NI group suggest that a program is activated early in a subset of patients that is independent of mediator levels in Pattern A. What drives the differential expression of the Pattern B mediators is unclear; however, the early peak suggests that these mediators are produced in response to preexisting factors or variables associated with injury mechanism/pattern or prehospital management that are not revealed in our analysis. It is unclear, for example, if this pattern reflects the higher incidence of thoracic trauma in the NI group. In addition, we also suggest that the elevation in Pattern B mediators gives way to marked increase in Pattern C mediators after 12 hours of injury. These mediators reflect cytokines that, for the most part, are either produced by lymphoid cells or involved in the regulation of lymphocyte responses. In fact, both the AUC and time course analysis of inflammatory mediators show that the greatest fold changes occur in IL-7, IL-5, IL-17A, and IL-4, which correspond to Pattern C. These unique patterns (B and C) suggest that the phenotypic nature of the postinflammatory response that predisposes trauma patients to develop NI could be mediated by an early, robust proinflammatory signal through activation of the innate immune system. In addition, these patterns also suggest that the divergence of the response includes engagement of lymphocyte responses early in the clinical course, which correlate with development of MOD and persist until onset of infection. This could reflect the activation of the innate lymphoid cell lineage with a concomitantly strong Th2 or type 2 immune signal, both of which persists over the 7 days’ postinjury time course. We speculate that the gradual elevation of cytokines such as IL-1β and TNF-α seen in Pattern D could reflect the response to microbes in the early phases of infection, though further studies are needed to test this hypothesis. Several studies have reported on the pivotal role of the cellular component of the innate and T cell-mediated immune responses that precedes the development of NI and MOD following traumatic injury.35,36 In this study, both the total leukocyte counts and PMN percentage were significantly higher in patients who developed NI on admission and after 1 week of injury versus no-NI patients and elevated PMN corresponded with the mean day of NI diagnosis. Furthermore, our analysis revealed that the NI group exhibited sustained lymphopenia upon admission and up to 14 days after injury when compared with the no-NI group. Recent studies have demonstrated the association between prolonged lymphopenia and the development of NI and MOD,37,38 suggesting that lymphocyte depletion, at least in part, plays a detrimental role in the evolution of nosocomial sepsis-induced MOD after traumatic injury. HMGB1 has emerged as a prototypical damage-associated molecular pattern in the response to sterile injury, including hemorrhagic shock and ischemia/reperfusion injury.39 In addition, HMGB1 promotes processes required for host defense, tissue repair, and regeneration, including chemotaxis, angiogenesis, maturation of dendritic cells, and recruitment and proliferation of stem cells.40 Other reports suggest that HMGB1 amplifies the inflammatory response by binding endogenous and exogenous inflammatory mediators, such as, cytokines or endotoxins.41,42 In agreement with prior findings,43 our results show an early elevation of HMGB1 in patients who went on to develop infection, signifying a potential role of HMGB1 in earliest stages of the trauma-induced inflammatory response. Data-driven computational modeling methods have emerged as adjuncts to in vitro and in vivo studies to help define the dynamic, multidimensional inflammatory response after injury.44,45 These methods are primarily based on associations among data variables and include logistic regression techniques that allow monitoring of either static or dynamic changes as well as more recent tools that enable graphical views of network interconnectivity.44,45 Utilizing one such tool, DyNA,46 in parallel with 2-way ANOVA and AUC analysis, we suggest the presence of larger and more highly connected networks of systemic inflammation biomarkers in the NI group versus the no-NI group. These phenotypes were observed as early as we measured and persisted throughout the 24 hours’ time course after injury. These findings suggest that, although mounting an adequately robust inflammatory response is essential for effective restoration of homeostasis after injury,47 an overly exuberant and sustained immune response may be detrimental. Thus, to achieve complete resolution of trauma-induced inflammation, it is necessary to turn off both inflammatory mediator production and inflammatory cell accumulation, in particular activated PMNs, in a timely fashion. Importantly, DyNA also suggested that the trauma-induced inflammation in the NI group was primarily driven by IL-7, IL-4, IL-2, IL-13, IL-15, and IL-1β over the initial 24 hours after injury; these inflammatory mediators were persistently elevated in the NI group throughout 7 days after injury when compared with the no-NI group. In contrast, IL-13, IL-15, and IFN-α were the most connected nodes in the no-NI group over the initial 24 hours after injury. Collectively, these results imply a role for innate-like lymphoid cells and T-cells in both early and late inflammation in patients predisposed to develop NI. This hypothesis must be confirmed with additional studies. We recognize that there are several limitations in this study. First, this study was performed at a single level I trauma center and thus may not be generalizable or pertinent to other centers with differing admission demographics, injury characteristics, or management practices. This issue warrants additional, similar studies in other trauma centers to validate the results suggested from this study. Another important limitation is the number of inflammatory mediators analyzed, which was limited to the number of analytes we could measure using commercially available Luminex beadsets. Further studies examining a larger panel of inflammatory mediators are warranted. We also note that the diagnosis of HAP is difficult to establish, as there are no reliable tools to determine whether the patient has HAP, particularly in trauma patients with multiple other possible reasons such as lung contusion, acute respiratory distress syndrome, and transfusion-related acute lung injury with abnormal pulmonary gas exchange and abnormal chest radiograph on presentation. Furthermore, when interpreting data from registry-based and institutional series, the limitations of retrospective analyses must be considered. These limitations include patient selection bias and misclassification or information bias as a result of the retrospective aspect. In this study, we tried to overcome this limitation by selecting controls (ie, no-NI) through a computer-generated list with 1:1 matching based on stringent selection criteria (see Material and Methods). Finally, we note that data-driven modeling relies on available data and as such depends on the quality of those data. One of the key drawbacks of purely data-driven modeling techniques for monitoring of biological processes is their input-output nature, which does not provide any knowledge of the internal state of the process. In addition, an input-output system description cannot deal with physical system interconnections. Hence, these methods do not provide any direct mechanistic information about the system; rather they are based on association among data variables in some fashion or another. In multiple prior studies of trauma and sepsis, we have leveraged the counterpart to data-driven modeling, namely mechanistic modeling, to overcome some of these limitations.45 CONCLUSIONS This study demonstrates unique inflammatory biomarker patterns, particularly in the early events after injury, which emerge in patients prone to develop infections, suggesting that specific inflammatory biomarkers can potentially predict or drive processes that increase the susceptibility to infection. In addition, our study suggests the diagnostic value of these temporal patterns coupled with network analysis within the initial 24 hours after injury and based on admission characteristics that could allow early patient stratification and allocation of resource intensive care resource allocation. Finally, and in line with recent calls for data-driven patient stratification within the context of Precision Medicine, our study highlights the power of computational analysis for gaining insights into complex patient cohorts.48 FIGURE 1 A, Plasma lactate levels on admission and over 7 days after injury in NI and no-NI patients after trauma. Both subgroups had elevated lactate levels on admission, being statistically significantly higher in no-NI group at 4 and 8 hours after injury. *P < 0.05 by 2-way ANOVA. B, AIS of NI and no-NI subgroups. Head/neck and thoracic injuries were assessed as part of the AIS and were significantly higher in the NI group as than in the no-infection group. *P < 0.05 by Student t test. C, MOD score in NI vs no-NI subgroups over 7 days’ time course after injury. Initially, both subgroups had higher degree of MOD at day 1, and by day 2 the NI group had persistently higher degree of MOD when compared to no-NI group. *P < 0.05 by 2-way ANOVA. FIGURE 2 Temporal dynamic patterns of inflammation biomarkers in NI and no-NI subgroups. A, Pattern A includes MIG, IL-10, IL-8, and Eotaxin; B, Pattern B includes MCP-1, IL-6, and IL-1RA; C, Pattern C includes IL-7, IL-5, IL-17A, IL-4, IL-13, MIP-1α, IFN-γ, IL-15, sIL-2Rα, IP-10, GM-CSF, and MIP-1β; D, Pattern D includes IFN-α, IL-1β, IL-2, and TNF-α. Actual concentration levels for each biomarker are provided in Figure S3, Supplemental Digital Content, http://links.lww.com/SLA/A680. FIGURE 3 Plasma HMGB1 levels within 24 hours after trauma in NI and no-NI subgroups compared to healthy volunteers. HMGB1 levels were statistically significantly elevated within 4 hours of injury in NI group when compared with no-NI group. Both groups had elevated levels of HMGB1 at multiple time points when compared with healthy volunteers. *P <0.05 by 2-way ANOVA between NI and no-NI subgroups; †P <0.05 by 1-way ANOVA vs baseline (healthy volunteers). FIGURE 4 DyNA of inflammation biomarkers in NI and no-NI subgroups suggests differential network connectivity within 24 hours after injury. DyNA at 0 to 8 hours suggested that IL-7/IL-15/IL-13/IL-4/IL-1β/ IL-2/IFN-γ/IL-17A were highly connected in the NI group (A), whereas the no-NI group (D) exhibited a lesser degree of connected nodes: IL-15/IFN-α/IL-13 with 2 isolated networks. DyNA at 8 to 16 hours suggested that the NI group (B) retained the connectivity among IL-7/IL-15/IL-5/IL-4/IL-1β/IL-2/IL-13, whereas the no-NI group (E) continued to exhibit a lesser degree of connections with 1 main network: IL-13/IL-15/IL-5/IL-4/IFN-γ/ IL-17A/IFN-α and 2 isolated networks. DyNA at 16 to 24 hours revealed that the NI group (C) had an increasing connectivity: IL-7/IL-15/IL-5/IL-13/IL-17A/IL-4/IL-2/IL-1β/IFN-γ/GM-CSF. In contrast, the no-NI group (F) exhibited a substantial reduction of network connectivity with 4 isolated networks when compared with NI group. Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal’s Web site (www.annalsofsurgery.com). Disclosure: This work was supported by National Institutes of Health grant P50-GM-53789. The authors declare no conflicts of interest. 1 Kauvar DS Lefering R Wade CE Impact of hemorrhage on trauma outcome: an overview of epidemiology, clinical presentations, and therapeutic considerations J Trauma 2006 60 S3 S11 16763478 2 Klevens RM Edwards JR Richards CL Jr Estimating health care-associated infections and deaths in US hospitals, 2002 Public Health Rep 2007 122 160 166 17357358 3 Turrini RN Santo AH Nosocomial infection and multiple causes of death J Pediatr (Rio J) 2002 78 485 490 14647729 4 Harbrecht BG Zenati MS Doyle HR Hepatic dysfunction increases length of stay and risk of death after injury J Trauma 2002 53 517 523 12352490 5 Reed D Kemmerly SA Infection control and prevention: a review of hospital-acquired infections and the economic implications Ochsner J 2009 9 27 31 21603406 6 Bochicchio GV Joshi M Knorr KM Impact of nosocomial infections in trauma: does age make a difference? J Trauma 2001 50 612 617 11303154 7 Jamulitrat S Narong MN Thongpiyapoom S Trauma severity scoring systems as predictors of nosocomial infection Infect Control Hosp Epidemiol 2002 23 268 273 12026152 8 Chaudry IH Ayala A Mechanism of increased susceptibility to infection following hemorrhage Am J Surg 1993 165 59S 67S 8439001 9 Gebhard F Huber-Lang M Polytrauma-pathophysiology and management principles Langenbecks Arch Surg 2008 393 825 831 18431593 10 Keel M Trentz O Pathophysiology of polytrauma Injury 2005 36 691 709 15910820 11 Nathan C Points of control in inflammation Nature 2002 420 846 852 12490957 12 Bochicchio GV Napolitano LM Joshi M Persistent systemic inflammatory response syndrome is predictive of nosocomial infection in trauma J Trauma 2002 53 245 250 12169929 13 Adrie C Pinsky MR The inflammatory balance in human sepsis Intensive Care Med 2000 26 364 375 10872127 14 Namas R Ghuma A Hermus L The acute inflammatory response in trauma(hemorrhage and traumatic brain injury: current state and emerging prospects Libyan J Med 2009 4 97 103 21483522 15 Matzinger P The danger model: a renewed sense of self Science 2002 296 301 305 11951032 16 Nathan C Sporn M Cytokines in context J Cell Biol 1991 113 981 986 2040651 17 Cuschieri J Bulger E Schaeffer V Early elevation in random plasma IL-6 after severe injury is associated with development of organ failure Shock 2010 34 346 351 20844410 18 Sperry JL Friese RS Frankel HL Male gender is associated with excessive IL-6 expression following severe injury J Trauma 2008 64 572 578 18332795 19 Xiao W Mindrinos MN Seok J A genomic storm in critically injured humans J Exp Med 2011 208 2581 2590 22110166 20 Jastrow KM Gonzalez EA McGuire MF Early cytokine production risk stratifies trauma patients for multiple organ failure J Am Coll Surg 2009 209 320 331 19717036 21 Marshall JC Cook DJ Christou NV Multiple organ dysfunction score: a reliable descriptor of a complex clinical outcome Crit Care Med 1995 23 1638 1652 7587228 22 Gennarelli TA Wodzin E Gennarelli TA Wodzin E AIS 2005 Barrington, IL Association for the Advancement of Automotive Medicine 2008 23 Baker SP O’Neill B Haddon W Jr The injury severity score: a method for describing patients with multiple injuries and evaluating emergency care J Trauma 1974 14 187 196 4814394 24 Rosenthal VD Maki DG Jamulitrat S International Nosocomial Infection Control Consortium (INICC) report, data summary for 2003–2008, issued June 2009 Am J Infect Control 2010 38 95 104 20176284 25 Ertel W Keel M Bonaccio M Release of anti-inflammatory mediators after mechanical trauma correlates with severity of injury and clinical outcome J Trauma 1995 39 879 885 7474003 26 Desborough JP The stress response to trauma and surgery Br J Anaesth 2000 85 109 117 10927999 27 Dunne JR Malone DL Tracy JK Allogenic blood transfusion in the first 24 hours after trauma is associated with increased systemic inflammatory response syndrome (SIRS) and death Surg Infect (Larchmt) 2004 5 395 404 15744131 28 Bulger EM Nathens AB Rivara FP Management of severe head injury: institutional variations in care and effect on outcome Crit Care Med 2002 30 1870 1876 12163808 29 Jennett B Bond M Assessment of outcome after severe brain damage Lancet 1975 1 480 484 46957 30 Marshall JC Measuring organ dysfunction in the intensive care unit: why and how? Can J Anaesth 2005 52 224 230 15753490 31 Sauaia A Moore EE Johnson JL Validation of postinjury multiple organ failure scores Shock 2009 31 438 447 18838942 32 Hoover L Bochicchio GV Napolitano LM Systemic inflammatory response syndrome and nosocomial infection in trauma J Trauma 2006 61 310 316 16917443 33 Moore FA Moore EE Evolving concepts in the pathogenesis of postinjury multiple organ failure Surg Clin North Am 1995 75 257 277 7899997 34 El-Masri MM Joshi M Hebden J Use of the injury severity score to predict nosocomial bloodstream infections among critically ill trauma patients AACN Clin Issues 2002 13 367 372 12151990 35 Hietbrink F Koenderman L Rijkers G Trauma: the role of the innate immune system World J Emerg Surg 2006 1 15 16759367 36 Santucci CA Purcell TB Mejia C Leukocytosis as a predictor of severe injury in blunt trauma West J Emerg Med 2008 9 81 85 19561712 37 Le Tulzo Y Pangault C Gacouin A Early circulating lymphocyte apoptosis in human septic shock is associated with poor outcome Shock 2002 18 487 494 12462554 38 Felmet KA Hall MW Clark RS Prolonged lymphopenia, lymphoid depletion, and hypoprolactinemia in children with nosocomial sepsis and multiple organ failure J Immunol 2005 174 3765 3772 15749917 39 Tsung A Sahai R Tanaka H The nuclear factor HMGB1 mediates hepatic injury after murine liver ischemia-reperfusion J Exp Med 2005 201 1135 1143 15795240 40 Klune JR Dhupar R Cardinal J HMGB1: endogenous danger signaling Mol Med 2008 14 476 484 18431461 41 Abraham E Unraveling the role of high mobility group box protein 1 in severe trauma Crit Care 2009 13 1004 19930620 42 Sha Y Zmijewski J Xu Z HMGB1 develops enhanced proinflammatory activity by binding to cytokines J Immunol 2008 180 2531 2537 18250463 43 Cohen MJ Brohi K Calfee CS Early release of high mobility group box nuclear protein 1 after severe trauma in humans: role of injury severity and tissue hypoperfusion Crit Care 2009 13 R174 19887013 44 Janes KA Yaffe MB Data-driven modelling of signal-transduction networks Nat Rev Mol Cell Biol 2006 7 820 828 17057752 45 Vodovotz Y Billiar TR In silico modeling: methods and applications to trauma and sepsis Crit Care Med 2013 41 2008 2014 23863232 46 Mi Q Constantine G Ziraldo C A dynamic view of trauma(hemorrhage-induced inflammation in mice: principal drivers and networks PLoS One 2011 6 e19424 21573002 47 Namas R Ghuma A Torres A An adequately robust early TNF-alpha response is a hallmark of survival following trauma(hemorrhage PLoS One 2009 4 e8406 20027315 48 Hansen J Iyengar R Computation as the mechanistic bridge between precision medicine and systems therapeutics Clin Pharmacol Ther 2013 93 117 128 23212109

PMC005xxxxxx/PMC5136776.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 101617579 41986 J Hypertens (Los Angel) J Hypertens (Los Angel) Journal of hypertension : open access 2167-1095 27928516 5136776 10.4172/2167-1095.1000218 NIHMS823715 Article Antioxidant and Antiproliferative Activities of Purslane Seed Oil Guo Gai 1 Yue Li 1 Fan Shaoli 1 Jing Siqun 1* Yan Liang-Jun 2 1 College of Life Sciences and Technology, Xinjiang University, Shengli Road 14, Urumqi, Xinjiang 830046, China 2 Department of Pharmaceutical Sciences, UNT System College of Pharmacy, University of North Texas Health Science Center at Fort Worth, 3500 Camp Bowie Boulevard, Fort Worth, TX 76107, USA Corresponding author: Siqun Jing, College of Life Sciences and Technology, Xinjiang University, Shengli Road 14, Urumqi, Xinjiang 830046, China, Tel: +86 991 8582554; Fax: +86 991 2339267; jingsiqun@163.com 18 10 2016 25 4 2016 6 2016 05 12 2016 5 2 218This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. The aim of this study was to evaluate the antioxidant and antiproliferative activities of PSO in vitro and its application in horse oil storage. We determined the reducing power of PSO and its scavenging effects on hydroxyl (•OH) and 1,1-diphenyl-2-picrylhydrazyl radicals (DPPH•) and tested its stabilizing effects on horse oil storage. The results showed that PSO had remarkable, dose-dependent antioxidant activities, and it effectively prevented horse oil lipid oxidation. We treated cervical cancer HeLa cells, esophageal cancer Eca-109 cells and breast cancer MCF-7 cells with PSO using non-neoplastic monkey kidney Vero cells as controls. The results indicate that PSO significantly inhibited tumor cell growth in a time- and dose-dependent fashion. Our studies suggest that PSO may be used as a substitute for synthetic antioxidants in food preservation and may be potentially useful as a food and cosmetic ingredient. Meanwhile, the oxidative stress can cause hypertension, so PSO is expected to develop a health care products for the prevention and mitigation hypertensive symptoms. Portulaca oleracea L. seed oil (PSO) Antioxidant Anti-proliferation Storage application Introduction Purslane (Portulaca oleracea L.) is a widely distributed weed that is extensively used not only as an edible plant but also as a traditional Chinese herbal medicine [1]. Both the leaves and seeds of purslane can be consumed orally or applied topically to soothe a skin allergy [2]. Many studies have demonstrated various pharmacological effects of this plant, such as antibacterial [3,4] hypoglycemic, [5] anti-hypoxia, [6] antioxidant effects, [7] antitumor activity, [8] and neuroprotective effects [9]. As a medicinal and edible wild plant, purslane is generally known as a “longevity food” due to its reputation as a “natural antibiotic”. Purslane contains many compounds, including flavonoids, [10] alkaloids, [11] omega-3 fatty acids, noradrenaline, alkaloids, coumarins, flavonoids, polysaccharides, and other active ingredients. In particular, purslane seeds are reportedly more effective in antioxidation than those from other herbs [12]. In previous studies, we have extracted purslane seed oil (PSO) with a 17.68% yield using an ultrasound-assisted enzyme hydrolysis combined with a Soxhlet extraction method. We then analyzed the fatty acid profile and content of the oil using a Gas Chromatography-Mass Spectrometer (GC-MS) [13]. Analysis of the PSO showed that alpha-linolenic acid reached 40.2570% followed by linoleic acid (29.4308%) and oleic acid (15.6103%). Saturated fatty acids represent 13.9455% of the total oil, while monounsaturated fatty acids and polyunsaturated fatty acids (PUFAs) account for 16.2877% and 69.6878%, respectively. Moreover, the linolenic acid content (40.2570%) in PSO is much higher than in camellia seed (0.27%), [14] grape seed (7.3%), [15] and olive (6.09%) oils [14] but is slightly lower than in flaxseed oil (41.22%) [12]. The content of linoleic acid (29.4308%) is much higher than in most other vegetable oils, such as flaxseed (15.44%), camellia seed (7.26%), grape seed (11.4%), and olive (0.56%) oils. It is well known that linolenic acid is an omega-3 fatty acid while linoleic acid is an omega-6 fatty acid, and both are the essential fatty acids that play important roles in human growth and development as well as in disease prevention [16]. Additionally, oils that are rich in omega-3 fatty acids are most likely beneficial to human health [17]. PSO is expected to show superior antioxidant activity and antitumor effects due to its high omega-3 fatty acid content. Thus, it is a good candidate as both a health food and a cosmetic ingredient. However, there are few studies regarding the composition of fatty acids in purslane and PSO [13,18,19]. Moreover, to the best of our knowledge, there is no published research on PSO's antioxidant activity or antiproliferative effect on cancer cell lines. Therefore, in this study, we tested PSO's antioxidant activity on free radical scavenging and its inhibitory effects on tumor cell proliferation. Additionally, we also tested PSO as a preserving agent in horse oil storage to explore whether it can be used as a food-preserving agent. Materials and Methods Fresh, mature purslane seeds were provided by Xinjiang Yuansen Agriculture Science and Technology Development Co., Ltd. PSO containing 40.2570% alpha-linolenic acid and 29.4308% linoleic acid was obtained by an ultrasound-assisted enzyme hydrolysis combined with a Soxhlet extraction method. The optimal preparation conditions of PSO were as follows: [13] for the hydrolysis process, 2% complex enzyme was used (the ratio of neutral protease to cellulase was 1:1), the liquid-solid ratio was 5:1, pH was 5.0, and the hydrolysis time was 2 h; for the sonication process, 40 W ultrasonic power was used, an ultrasonic bath temperature was set to 55°C, and the sonication time was 15 min. Petroleum ether was used as a solvent for the Soxhlet extraction. Fresh, untreated horse fat was purchased from a local market in Yili, Xinjiang, China. Liquid horse oil was obtained from the horse fat using a steam melting method followed by a refining process consistent with our previous studies [20]. Horse fat provides the raw material of which 31.08% of the lipids are unsaturated fatty acids, notably, palmitoleic acid (3.71%) and oleic acid (27.37%). The standard tertiary butylhydroquinone (TBHQ) that was used in vitro studies was purchased from Sigma Chemical Co. (St. Louis, MO, USA). In vitro antioxidant potential of PSO TBHQ is a frequently used synthetic antioxidant; therefore, it was used as a reference material for evaluating the antioxidant activity of PSO in this study. The antioxidant activity of PSO was determined by various methods, such as 1,1-diphenyl-2-picryl hydrazyl (DPPH) radical scavenging, hydroxyl radical scavenging, and reducing power assays. PSO was dissolved in an anhydrous ethanol to form a series of final concentrations. Equal concentrations of TBHQ and PSO were used in each experiment. All tests were conducted in triplicate, and the mean values were plotted. Hydroxyl radical (•OH) scavenging activity assay The phenanthroline-Fe2+ oxidation method previously described by Xiao et al. was used to measure hydroxyl radical scavenging activity [21]. Briefly, 4 mL of sodium phosphate buffer (pH 7.4) was added to a test tube and mixed with 1.5 mL of 5 mmol /L phenanthroline solution. Next, 1 mL of 7.5 mmol /L FeSO4 solution and 1 mL of a series of concentrations (0.5, 1, 2, 3, 4, 5, 6, 7 mg/mL) of PSO sample solution were added to the solution in sequence. Finally, 1.5 mL of double distilled water and 1.0 mL of 0.1% H2O2 were added. The absorbance of the final solutions was measured at 536 nm with a UV-visible spectrophotometer following incubation at 37 for 60 min. Deionized water and TBHQ were used as blank and positive controls, respectively. Antioxidant value was expressed as IC50, the concentration of the sample that caused 50% inhibition of hydroxyl radical formation. DPPH• radical scavenging activity assay The scavenging activity of DPPH free radicals was measured according to the method reported by Ting et al, with slight modifications [22]. Briefly, 2 mL of 2×10-4 mol/mL DPPH was added to 2 mL series of concentrations (0.1, 0.3, 0.5, 1, 2, 3, 6, 10, 15, 20, 25, 30 mg/mL) of PSO in sequence. The reaction mixture was incubated for 30 min at room temperature in the dark, after which absorbance was measured at 517 nm using a spectrophotometer (TU-1900 PuXiTongYong, Beijing, China). Distilled water was used instead of the sample solutions as a control. TBHQ was used as a positive control. A mixture with an equal volume of distilled water and anhydrous ethanol was used as a blank control. The results were expressed as the amount of sample necessary to scavenge 50% of DPPH• radicals (IC50). Determination of reducing power The reducing power of PSO was examined using the Prussian blue method [23]. Briefly, 1 mL of each PSO sample solution (0.01, 0.05, 0.1, 0.3, 0.5, 1.2, 2.4, 3.6, 4.8, 6, 7.2 mg/mL) was added to a solution containing 2.5 mL of phosphate buffer (pH 6.6) and 2.5 mL of 1% K3Fe(CN)6. The buffered solutions were then stored at 50 for 20 min, after which 2.5 mL of 10% trichloroacetic acid (TCA) was added. Next, 2.5 mL of distilled water and 2.5 mL of 0.1% FeCl3 were mixed with 2.5 mL of the previous mixtures. After 10 min, the absorbance value (A) was measured at 700 nm with a UV-visible spectrophotometer. The reference absorbance value (A0) was given by a blank reagent control. Deionized water was used as the blank control, and TBHQ was used as the positive control. Antiproliferative effect of PSO in vitro HeLa, Eca-109, MCF-7, and Vero cells were obtained from the Xinjiang University Xinjiang Biological Resources Gene Engineering Key Laboratory (Urumqi, China). The in vitro antiproliferative activity of PSO was determined by measuring 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) dye absorbance in living cells (HeLa, Eca-109, and MCF-7) with Vero (normal) cell as controls. Briefly, cells were seeded in the 96-well, flat-bottomed plates containing 100 μL of a cell suspension with a known concentration per well and allowed to adhere at 37 in a humidified atmosphere containing 5% CO2. Usually, 5×104 cells were seeded per well. PSO was dissolved in dimethyl sulfoxide (DMSO) and then filtered with filter membranes (0.45 μm and 0.22 μm) to achieve sterilization. 200 μL PSO at concentrations of 12.5 μg/mL, 50 μg/mL, 200 μg/mL, 800 μg/mL, 1600 μg/mL and 3200 μg/mL were added to their respective wells. In total, 20 μL of the MTT solution (5 mg/ml; Sigma-Aldrich, MO, USA) was then added at 24 h, 48 h, or 72 h for dyeing, and the cells were incubated for another 4 h at 37. After the incubation, the cell suspensions were centrifuged at 800 rpm for 10 min, and the supernatants were replaced by 200 μL DMSO to solubilize the formazan crystals formed in viable cells. Absorbance at 570 nm was measured using a microplate ELISA reader (Model 550, Bio-Rad, USA). The results were expressed as a percentage of control proliferation (100%). The IC50 value was expressed as the concentration of PSO that inhibited the growth of cells by 50% [8]. Effect of PSO on the oxidative stability of horse oil during storage The peroxide value (POV) is an indicator of lipid oxidation. The POV was determined by the Schall Oven method [24]. Briefly, an oil sample was incubated in a digital electric heating blast oven at 63 ± 1 constant temperature, and the POV values was measured once every 24 h, according to the National Standard of the People's Republic of China (GB/T 5538-2005/ISO3960:2001) [25]. The lower the POV, the stronger the oxidative stability of the sample. The POV (mmol/kg) of the sample was calculated by the following equation: POV(mmol/kg)=1000(V−V0)C2m, where V is the volume of sodium thiosulfate (Na2S2O3) for the measurement (mL); V0 is the volume of Na2S2O3 for the blank test (mL); C is the concentration of Na2S2O3 solution (mol/L); and m is the weight of the sample (g). To investigate the effect of PSO on horse oil storage stability, we added 0.5% (w/w) PSO to 30 g of horse oil. After thoroughly mixing the solution, the subsequent operations were performed according to the Schall Oven method. A blank test without added PSO and TBHQ was necessary. The TBHQ-added group was used as a positive control. To investigate the effect of the dose of PSO on horse oil storage stability, we added PSO with different proportions of 0.05%, 0.25% and 0.5% to horse oil. After thoroughly mixing the solution, the subsequent operations were performed according to the Schall Oven method. The group without PSO added was regarded as a control. Statistical analysis All results were expressed as the mean ± SD. The data were analyzed statistically using ANOVA. Statistical calculations were conducted using Graphpad Prism 5.0 (Graphpad Software Inc., San Diego, CA, USA). Values of p < 0.05 were considered significantly different. Results and Discussion Evaluation of PSO antioxidant activity in vitro Hydroxyl free radical scavenging activity of PSO Among the tested free radicals, hydroxyl free radicals are the most active and toxic. Thus, the hydroxyl free radical scavenging capacity can be used as an indicator of antioxidant activity. As shown in Figure 1a, the hydroxyl free radical scavenging capacity of PSO was enhanced at higher concentrations of PSO in a nearly linear relationship until a concentration of 4 mg/mL. The IC50 values of PSO and TBHQ were 1.388 ± 0.2033 mg/mL and 2.193 ± 0.1014 mg/mL, respectively. Moreover, at the same concentrations, the ranking of hydroxyl radical scavenging ability of PSO, TBHQ, almond oil, and grape seed oil was PSO > TBHQ> almond oil (IC50 2.53 mg/mL) > grape seed oil (IC50 6.66 mg/mL) [26]. Therefore, the hydroxyl free radical scavenging activity of PSO was the strongest among the tested samples. In the literature, grape seed oil is reported to contain linoleic (65.0%), linolenic (1.5%), oleic (17.0%), palmitic (8.0%), stearic (4.4%) and arachidonic (0.6%) acids; [27] while oleic (63-78%) and linoleic (12-27%) acids are the major fatty acids in almond oil [28]. It is clear that the predominant fatty acid in PSO is linolenic omega-3 fatty acid. Therefore, we reason that the stronger antioxidant activity of PSO may be attributed to its higher content of linolenic acid. Indeed, previous research indicates that an increased amount of omega-3 PUFAs may enhance antioxidant activity [29,30]. However, whether the observed effect is due to the omega-3 or to the other fatty acids in PSO needs to be further evaluated. DPPH• free radical scavenging capacity of PSO The DPPH• scavenging ability of PSO was enhanced when the oil concentration was increased (Figure 1b). A strong linear relationship is observed within the range of PSO concentrations from 3-20 mg/mL. Moreover, it is worth mentioning that PSO's antioxidant activity could only be detected at concentrations at or above 3 mg/mL. The IC50 values of DPPH• radicals scavenging of PSO and TBHQ were 11.16 ± 0.07075 mg/mL and 0.3783 ± 0.07886 mg/mL, respectively, and the result indicated that the DPPH• free radical scavenging capacity of PSO was weaker than that of TBHQ. Additionally, the DPPH• free radical scavenging capacity of PSO was higher than that of walnut oil (IC50 147.0 mg/mL) [31] and weaker than that of flaxseed oil (IC50 3.31 mg/mL) [32]. It has been reported that the major fatty acids in walnut oil are linoleic (60.42-65.77%), oleic (13.21-19.94%) and linolenic (7.61-13%) acids [31]. The main fatty acid components of flaxseed oil are alpha-linolenic (41.22%), linoleic (15.44%), and oleic (28.2%) acids. Obviously, the stronger DPPH• radical scavenging ability of PSO may be due to its higher content of linolenic omega-3 fatty acid. Reducing power of PSO Many studies have demonstrated that the activity of an antioxidant is closely related to its reducing power: the greater the reducing power, the stronger the antioxidant activity. Therefore, the reducing power can reflect the antioxidant activity [33]. The results in Figure 1c showed that the reducing power increased as the concentration of PSO increased, and there was a strong positive linear relationship. However, the reducing power of PSO was weaker than that of TBHQ. In summary, we observed a strong antioxidant activity of PSO, suggesting that PSO is a likely candidate as a food and cosmetic ingredient. In vitro inhibitory effect of PSO on tumor cell proliferation MTT assay for tumor cell proliferation inhibition The antiproliferative effect of PSO on MCF-7, Eca-109 and HeLa cells was evaluated by an MTT assay, and the IC50 values were derived from the dose-response curves (Figure 2). PSO induced a significant dose- and time-dependent decrease in the proliferation rate of HeLa, MCF-7, and Eca-109 cells. The IC50 values of PSO against MCF-7, HeLa, and Eca-109 cells were 1566 ± 0.01691 μg/mL, 1844 ± 0.0217 μg/mL and 5366 ± 0.03851 μg/mL, respectively. PSO showed a stronger inhibitory effect on the proliferation of MCF-7 cells. Growth inhibition of HeLa cells As shown in Figure 3, PSO inhibited the growth of HeLa cells in a dose- and time-dependent manner with IC50 values of 11634 ± 0.02706 μg/mL, 1844 ± 0.0217 μg/mL and 1179 ± 0.01989 μg/mL, respectively, after 24 h, 48 h, and 72 h. The same dual-dependent relationship was found by Cao et al. concerning essential oil from Artemisia lavandulaefolia, [34] Thus, PSO may significantly inhibit the proliferation of HeLa cells. Effect on proliferation of Eca-109 cells As shown in Figure 4, PSO inhibited the growth of Eca-109 cells in a dose- and time-dependent manner with IC50 values of 65540 ± 0.03675 μg/mL, 5366 ± 0.03851 μg/mL, and 3048 ± 0.03686 μg/mL, respectively, after 24 h, 48 h and 72 h. Therefore, PSO may significantly inhibit the proliferation of Eca-109 cells. Effect on proliferation of MCF-7 cells As shown in Figure 5, PSO inhibited the growth of MCF-7 cells in a dose- and time-dependent manner with IC50 values of 3179 ± 0.02242 μg/mL, 1566 ± 0.01691 μg/mL, and 1064 ± 0.01413 μg/mL, respectively, after 24 h, 48 h and 72 h. PSO significantly inhibited the proliferation of Eca-109 cells. Growth inhibition of vero cells As shown in Figure 6, the antiproliferative effect of PSO on Vero cells has a slightly increasing trend in a concentration- and time-dependent manner. However, the inhibitory rate of PSO against Vero cells was only 19.25 ± 1.2162% at a concentration of 1600 μg/mL after a 72 h incubation time. The inhibitory rate against MCF-7 cells was up to 54.51 ± 1.1738% (p<0.05), while against HeLa cells, the rate was up to 54.42 ± 1.7466% (p<0.05). For Eca-109 cells, the rate was up to 44.33 ± 2.3405% (p<0.05) under the same conditions. Therefore, our results indicate that PSO had less cytotoxic effect on the normal Vero cells. Overall, the growth inhibition of Vero cells by PSO was much weaker than that of MCF-7, HeLa, and Eca-109 cells. In summary, the above in vitro studies clearly show that the inhibitory effect of PSO on MCF-7 cells was stronger than that on HeLa and Eca-109 cells. This is the first report that demonstrates inhibition of the breast cancer cell growth in vitro by PSO. This report suggests a potential therapeutic role of PSO in the treatment of breast cancer. Further research on the mechanism of PSO inhibition of MCF-7 cell proliferation remains to be conducted. Application of PSO to horse oil storage Due to the adverse health effects of synthetic antioxidants, such as TBHQ and butylated hydroxyanisole (BHA), there has been a considerable increase in demand for isolating naturally occurring bioactive molecules for the food and pharmaceutical industries [35]. In this study, we compared the antioxidant effect of PSO to the synthetic antioxidant TBHQ in horse oil storage. As shown in Figure 7a, the addition of both PSO and TBHQ showed improved protection against auto-oxidation of horse oil. The POVs were much lower than those of horse oil alone, but PSO was weaker than TBHQ in inhibiting lipid peroxidation. This implies that we can further improve horse oil stabilization if PSO and TBHQ are combined, producing a synergistic effect [36] while minimizing the amount of synthetic antioxidant. Additionally, we investigated the dose-dependent effect of PSO on storage stability of horse oil. We found that the lipid antioxidant capacity of PSO increased with the dose. This is significantly different from the control group, which showed a significant increasing trend the POV in the absence of PSO (Figure 7b). This result suggests that PSO can be used as a natural preservative in the food and cosmetics industries. Conclusions With increased knowledge of the important bioactive molecules in seed oils, such as lipids and pigments, seed oils that possess anti-oxidation and anti-tumor potential have gained significant attention for the treatment of tumors and other cancer-related problems [37,38]. We tested the anti-oxidation and anti-tumor cell proliferation activities of PSO. Our study showed that PSO had remarkable free radical scavenging capabilities, as its hydroxyl free radical scavenging activity is stronger than that of TBHQ. PSO also displayed a stronger antiproliferative effect on MCF-7 cells than on either HeLa or Eca-109 cells with a dual-dependent relationship in a time- and dose-dependent manner. Moreover, PSO protects horse oil against lipid oxidation in a dose-dependent manner. Our findings suggest that PSO can be useful as a health food for the prevention and mitigation hypertensive symptoms and cosmetic ingredient. Studies on the mechanisms of anti-oxidation and anti-tumor cell proliferation are currently underway. This study was funded by Xinjiang Scientific Research and Innovation Projects for Postgraduates (No. XJGRI2014037. [2014]). Figure 1a Antioxidant activity of PSO in vitro. Hydroxyl radical scavenging ability. Figure 1b Antioxidant activity of PSO in vitro. DPPH• radicals scavenging ability. Figure 1c Antioxidant activity of PSO in vitro reducing power. Figure 2 The curve of PSO inhibition of MCF-7, HeLa and Eca-109 cell proliferation for 48 h. Figure 3 Time curve of HeLa cell proliferation as a function of different concentrations of PSO. Figure 4 Time curve of Eca-109 cells proliferation as a function of different concentrations of PSO. Figure 5 Time curve of MCF-7 cells proliferation as a function of PSO concentration. Figure 6 Time curve of vero cells proliferation as a function of different concentrations of PSO. Figure 7a Effect of PSO on the oxidative stability of horse oil during storage. Effect of PSO and other antioxidants on horse oil storage stability. Figure 7b Effect of PSO on the oxidative stability of horse oil during storage. Effect of the PSO dose on storage stability of horse oil. Conflict of interest: The authors declare that they have no conflicts of interest 1 Liang X Tian J Li L Gao J Zhang Q 2014 Rapid determination of eight bioactive alkaloids in Portulaca oleracea L. by the optimal microwave extraction combined with positive-negative conversion multiple reaction monitor (+/-MRM) technology Talanta 120 167 172 24468356 2 Ghazanfar S 1994 Handbook of Arabian Medicinal Plants CRC Press Boca Raton, Florida, United States 3 Zhang X Ji Y Qu Z Xia JC Wang L 2002 Experimental studies on antibiotic functions of Portulaca oleracea L. in vitro Chin J Microecol 14 277 280 4 Dong CX Hayashi K Lee JB Hayashi T 2010 Characterization of structures and antiviral effects of polysaccharides from Portulaca oleracea L Chem Pharm Bull (Tokyo) 58 507 510 20410633 5 Gu J Zheng Z Yuan J Zhao B Wang C 2014 Comparison on hypoglycemic and antioxidant activities of the fresh and dried Portulaca oleracea L. in insulin-resistant HepG2 cells and streptozotocin-induced C57BL/6J diabetic mice J Ethnopharmacol 161 214 223 25523372 6 Wanyin W Liwei D Lin J Hailiang X Changquan L 2012 Ethanol extract of Portulaca oleracea L. protects against hypoxia-induced neuro damage through modulating endogenous erythropoietin expression J Nutr Biochem 23 385 391 21543202 7 Silva R Carvalho IS 2014 In vitro antioxidant activity, phenolic compounds and protective effect against DNA damage provided by leaves, stems and flowers of Portulaca oleracea (Purslane) Nat Prod Commun 9 45 50 24660460 8 Zhao R Gao X Cai Y Shao X Jia G 2013 Antitumor activity of Portulaca oleracea L. polysaccharides against cervical carcinoma in vitro and in vivo Carbohydr Polym 96 376 383 23768576 9 Abdel Moneim AE 2013 The neuroprotective effects of purslane (Portulaca oleracea) on rotenone-induced biochemical changes and apoptosis in brain of rat CNS Neurol Disord Drug Targets 12 830 841 23844694 10 Zhu H Wang Y Liu Y Xia Y Tian T 2010 Analysis of flavonoids in Portulaca oleracea L. by UV–Vis spectrophotometry with comparative study on different extraction technologies Food Anal Method 3 90 97 11 Xiang L Xing D Wang W Wang R Ding Y 2005 Alkaloids from Portulaca oleracea L Phytochemistry 66 2595 2601 16203019 12 El-Sayed MI 2011 Effects of Portulaca oleracea L. seeds in treatment of type-2 diabetes mellitus patients as adjunctive and alternative therapy J Ethnopharmacol 137 643 651 21718775 13 Liu L Xie C Yue L Jing S Cai Y 2014 Preparation of Portulaca oleracea L. seed oil by ultrasound-assisted enzyme hydrolysis combined with Soxhlet extraction method and the analysis of its fatty acids Food Ferment Ind 40 218 222 14 Fangfang A Jun B Dan Z Yang Y Yizeng L 2013 Comparison and analysis of fatty acids between oil-tea camellia seed oil and other vegetable oils China Oils Fats 38 77 80 15 Zhang L Mou D Du Y 2007 Technology of extracting grape seed oil via supercritical fluid J Chin Cereal Oil Assoc 22 60 62 65 16 Dunbar BS Bosire RV Deckelbaum RJ 2014 Omega 3 and omega 6 fatty acids in human and animal health: an African perspective Mol Cell Endocrinol 398 69 77 25458696 17 Asif M 2011 Health effects of omega-3,6,9 fatty acids: Perilla frutescens is a good example of plant oils Orient Pharm Exp Med 11 51 59 21909287 18 Simopoulos AP Salem N Jr 1986 Purslane: a terrestrial source of omega-3 fatty acids N Engl J Med 315 833 19 Uddin MK Juraimi AS Ali ME Ismail MR 2012 Evaluation of antioxidant properties and mineral composition of Purslane (Portulaca oleracea L.) at different growth stages Int J Mol Sci 13 10257 10267 22949859 20 Jing S Aibaila R Li Y 2012 Study on process optimizing of refining process and antibacterial effect of hurse oil Sci Technol Food Ind 33 291 294 298 21 Xiao J Sun J Yao L Zhao Q Wang L 2012 Physicochemical characteristics of ultrasonic extracted polysaccharides from cordyceps cephalosporium mycelia Int J Biol Macromol 51 64 69 22575391 22 Ting H Hsu Y Tsai C Lu F Chou M 2011 The in vitro and in vivo antioxidant properties of seabuckthorn (Hippophae rhamnoides L.) seed oil Food Chem 125 652 659 23 Miao L Wu H Qiu N Pang F 2010 Assessment on antioxidant activity of pomegranate seed oil in vivo China Oils Fats 35 37 40 24 Ke H Chun C Mouming Z Hao M 2011 Comparative of Rancimat method and Schaal oven method for the determination of oxidation stability of peanut oil and peanut butter Food Ferment Ind 37 145 148 25 The State Administration of Quality Supervision Inspection and Quarantine of People's Republic of China. Food Safety National Standards Determination of Peroxide Value of Animal and Plant Lipid China Standard Press Beijing 2015 26 Kong H Zhang J 2008 Study on the Scavenging free radical capacity of almond oil and grape seed oil Gansu Sci Technol 3 57 58 27 Kikalishvili BIu Zurabashvili DZ Turabelidze DG Zurabashvili ZA Giorgobiani IB 2012 Fatty acids of grape seed oil and its biological activity as 1,0% and 2,5% food-additive Georgian Med News 47 50 28 Kodad O Socias I Company R 2008 Variability of oil content and of major fatty acid composition in almond (Prunus amygdalus Batsch) and its relationship with kernel quality J Agric Food Chem 56 4096 4101 18461963 29 Simopoulos AP 2004 Omega-3 fatty acids and antioxidants in edible wild plants Biol Res 37 263 277 15455656 30 Uddin MK Juraimi AS Hossain MS Nahar MA Ali ME 2014 Purslane weed (Portulaca oleracea): a prospective plant source of nutrition, omega-3 fatty acid, and antioxidant attributes Sci World J 2014 951019 31 Bouabdallah I Bouali I Martinez-Force E Albouchi A Perez Camino MC 2014 Composition of fatty acids, triacylglycerols and polar compounds of different walnut varieties (Juglans regia L.) from Tunisia Nat Prod Res 28 1826 1833 25142132 32 Long JJ Fu YJ Zu YG Li J Wang W 2011 Ultrasound-assisted extraction of flaxseed oil using immobilized enzymes Bioresour Technol 102 9991 9996 21890349 33 Zhang J Yuan K Zhou WL Zhou J Yang P 2011 Studies on the active components and antioxidant activities of the extracts of Mimosa pudica Linn. from southern China Pharmacogn Mag 7 35 39 21472077 34 Zhang LM Lv XW Shao LX Ma YF Cheng WZ 2013 Essential oil from Artemisia lavandulaefolia induces apoptosis and necrosis of HeLa cells Zhong Yao Cai 36 1988 1992 25090687 35 Ito N Fukushima S Hagiwara A Shibata M Ogiso T 1983 Carcinogenicity of butylated hydroxyanisole in F344 rats J Natl Cancer Inst 70 343 352 6571941 36 Vattem DA Ghaedian R Shetty K 2005 Enhancing health benefits of berries through phenolic antioxidant enrichment: focus on cranberry Asia Pac J Clin Nutr 14 120 130 15927928 37 Hosokawa M Kudo M Maeda H Kohno H Tanaka T 2004 Fucoxanthin induces apoptosis and enhances the anti-proliferative effect of the PPARa ligand, trogitazone, on colon cancer cells Biochim Biophys Acta 1675 113 119 15535974 38 Kohno H Suzuki R Yasui Y Hosokawa M Miyashita K 2004 Pomegranate seed oil rich in conjugated linolenic acid suppresses chemically induced colon carcinogenesis in rats Cancer Sci 95 481 486 15182427

PMC005xxxxxx/PMC5136800.txt

LICENSE: This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. 2985108R 3923 Gut Gut Gut 0017-5749 1468-3288 26045134 5136800 10.1136/gutjnl-2014-307649 NIHMS830235 Article Identification of an anti-inflammatory protein from Faecalibacterium prausnitzii, a commensal bacterium deficient in Crohn's disease Quévrain E. 123 Maubert M. A. 1234 Michon C. 56 Chain F. 56 Marquant R. 1310 Tailhades J. 1310 Miquel S. 56 Carlier L. 1310 Bermúdez-Humarán L. G. 56 Pigneur B. 123 Lequin O. 1310 Kharrat P. 56 Thomas G. 123 Rainteau D. 1234 Aubry C. 56 Breyner N. 56 Afonso C. 7 Lavielle S. 1310 Grill J.-P. 123 Chassaing G. 1310 Chatel J. M. 56 Trugnan G. 1234 Xavier R. 8 Langella P. 56 Sokol H. 12359 Seksik P. 1239 (1) Sorbonne Universités, UPMC Univ Paris 06, LBM, 27 rue de Chaligny, F-75012, Paris, France. (2) INSERM-ERL 1157 and Inflammation-Immunopathology-Biotherapy Department (DHU i2B), CHU Saint-Antoine 27 rue de Chaligny, F-75012 Paris, France. (3) CNRS, UMR 7203 LBM, F-75005, Paris, France (4) APHP, Hôpital Saint Antoine - Département PM2 Plateforme de Métabolomique, Peptidomique et dosage de Médicaments, F-75012 Paris, France (5) INRA, UMR1319 Micalis, F-78350 Jouy-en-Josas, France. (6) AgroParisTech, UMR Micalis, F-78350 Jouy-en-Josas, France. (7) Université de Rouen, UMR 6014 COBRA / IRCOF, F-76130 Mont Saint Aignan, France (8) Center for Systems Biology, Massachusetts General Hospital, Boston, Massachusetts, USA. (9) APHP, Hôpital Saint Antoine – Service de Gastroentérologie et nutrition, F-75012 Paris, France (10) Ecole Normale Supérieure- PSL Research University, Département de Chimie 24 rue Lhomond, F-75005 Paris, France Corresponding author: Pr. Philippe SEKSIK, Service de Gastroentérologie et Nutrition, Hôpital St-Antoine, 184 rue du Faubourg St-Antoine, 75571 Paris CEDEX 12, France. Tel: (+33) 149283162, Fax: (+33) 149283188, philippe.seksik@sat.aphp.fr 17 11 2016 04 6 2015 3 2016 01 3 2017 65 3 415425 This file is available for text mining. It may also be used consistent with the principles of fair use under the copyright law. Background Crohn's disease (CD) associated dysbiosis is characterized by a loss of Faecalibacterium prausnitzii, whose culture supernatant exerts an anti-inflammatory effect both in vitro and in vivo. However, the chemical nature of the anti-inflammatory compounds has not yet been determined. Methods Peptidomic analysis using mass spectrometry was applied to F. prausnitzii supernatant. Anti-inflammatory effects of identified peptides were tested in vitro directly on intestinal epithelial cell lines and on cell lines transfected with a plasmid construction coding for the candidate protein encompassing these peptides. In vivo, the cDNA of the candidate protein was delivered to the gut by recombinant Lactic Acid Bacteria to prevent DNBS-colitis in mice. Results The seven peptides, identified in the F. prausnitzii culture supernatants, derived from a single Microbial Anti-inflammatory Molecule (MAM), a protein of 15 kDa and comprising 53% of nonpolar residues. This last feature prevented the direct characterization of the putative anti-inflammatory activity of MAM-derived peptides. Transfection of MAM cDNA in epithelial cells led to a significant decrease in the activation of the NF-κB pathway with a dose-dependent effect. Finally, the use of a food-grade bacterium, Lactococcus lactis, delivering a plasmid encoding MAM was able to alleviate DNBS-induced colitis in mice. Conclusion A 15kDa protein with anti-inflammatory properties is produced by F. prausnitzii, a commensal bacterium involved in CD pathogenesis. This protein is able to inhibit the NF-κB pathway in intestinal epithelial cells and to prevent colitis in an animal model. microbiota dysbiosis Crohn’s disease Faecalibacterium prausnitzii anti-inflammatory protein Background Evidence from immunological, microbiological and genetic studies implicates abnormal host-microbial interactions in the pathogenesis of Crohn’s Disease (CD) [1] [2] [3] [4] [5]. Dysbiosis characterized by a reduction in bacterial biodiversity, lower bacterial population with anti-inflammatory properties and/or an increase in the proportion of bacteria with pro-inflammatory properties, has been observed in CD patients [6]. Several bacterial species have been associated with CD, including members of the Proteobacteria [7] such as adherent-invasive Escherichia coli [8], Campylobacter concisus [9] [10] and enterohepatic Helicobacter [11] [12]. So far, no definitive proof for any specific etiological agent has been highlighted. Additionally, several 16S rRNA sequencing-based studies have reported that members of the Firmicutes phylum were reduced in CD patients [13]. In previous work, we analyzed the composition of the ileal mucosa-associated microbiota of CD patients at the time of surgical resection for active disease, and six months later. Low proportions of Firmicutes, and particularly of Faecalibacterium prausnitzii, were consistently associated with an increased risk of post-operative recurrence of ileal CD [14]. Although controversial [15], we hypothesized that treatment with F. prausnitzii could be an effective strategy to counterbalance dysbiosis and reduce inflammation in CD patients. Following this, we and others demonstrated that F. prausnitzii exhibits anti-inflammatory effects both in vitro (cellular models) and in vivo (TNBS colitis model), associated with secreted metabolites that block NF-κB activation and IL-8 production by intestinal epithelial cells [14] [16] [17-18]. Moreover, gnotobiotic rodent models were previously used to show beneficial effects of F. prausnitzii on intestinal homeostasis and during an acute colitis [19] [20]. Identification of the active molecule(s) involved in this protective effect is of particular interest, since F. prausnitzii remains hardly cultivable due to its extreme oxygen sensitivity. Thus, finding the secreted molecule(s) responsible for this anti-inflammatory effect seems not only a cognitive issue but will also open the field of new therapeutic approach in CD. The cellular and molecular effects of commensal and probiotic bacteria are now recognized. Some bacteria have been shown to reinforce the intestinal barrier through the production of a soluble factor either by normalizing intestinal permeability of inflamed tissues [21] or inducing the expression of defensins [22] and type 2 zona occludens proteins in tight junctions between epithelial cells [23]. In mice, some commensal bacteria such as segmented filamentous bacteria [24] [25], Bacteroides fragilis and Clostridia members are able to shape gut immune responses [26] [27] [28] [29]. Among the molecules secreted by non-pathogenic bacteria, which are responsible for cellular effects in the host, very few bioactive molecules have been identified [30] [31]. An example of previously identified bioactive molecules can be seen in Lactobacillus rhamnosus GG, which secretes two proteins, p75 and p40, which have been characterized as being able to inhibit epithelial cells apoptosis induced by pro-inflammatory cytokines [32]. Small molecules, less than 10 kDa, from Saccharomyces boulardii and Bacteroides thetaiotaomicron that interact with NF-κB pathway have been detected but remain unidentified [33] [34]. Searching for such bioactive molecules remains challenging as finding these molecules in a cultured supernatant containing thousands of molecules is comparable to searching for a needle in a haystack. Preliminary experiments indicated that various treatments of F. prausnitzii supernatant (heated > 70°c), enzyme digestion (trypsin, lipase, amylase) or filtration (MW<15kDa), do not suppress anti-inflammatory effects (Maubert M.A, unpublished data). This prompted us to develop a peptidomic analysis of the supernatant in order to identify the presence of potential peptides derived from a unique and original protein from F. prausnitzii, which are able to interact with the NF-κB pathway in epithelial cells and are responsible for the anti-inflammatory effects. Methods Faecalibacterium prausnitzii bacterial strain, culture conditions and supernatant preparation F. prausnitzii (strain A2-165) was grown overnight at 37°C in an anaerobic (90 % N2, 5% CO2 and 5% H2) workstation (Whitley A35 anaerobic workstation) in LyBHI broth, 37 g. L−1 of brain-heart infusion (Difco) and 5 g. L−1 of yeast extract (Conda) at pH 7. The culture supernatant of F. prausnitzii was obtained by centrifugation at 1700 g at 4°C for 20 min. Faecalibacterium prausnitzii supernatant analysis Fractionation of the F. prausnitzii culture medium A solid/liquid extraction of culture medium using Waters Oasis HLB® SPE cartridges was carried out. Fractions were obtained by eluting with 20%, 40% and 80% of acetonitrile (F1, F2, F3 for F. prausnitzii supernatant and F1’, F2’, F3’ for LyBHI culture medium). After freeze-drying, these fractions were tested on epithelial cells through a cellular assay for anti-inflammatory effect (see below). For further investigations in mass spectrometry, F2/F2’ fractions were purified by preparative HPLC using a Waters Symmetry® C8 column (7.8 × 300 mm) Comparative mass spectrometry analysis of F. prausnitzii culture medium The bioactive anti-inflammatory fractions from the culture supernatant and LyBHI fractions eluted by SPE were analyzed by MALDI-TOF mass spectrometry (Voyager® DE Pro -AB Sciex). Peptide identification and synthesis Identification of peptides by FT-ICR mass spectrometry The ions of interest were fragmented by a FT-ICR mass spectrometer, equipped with a 7T superconducting magnet (Apex Qe®, Bruker Daltonics). The peptides were then identified by complete de novo sequencing. Peptides and functionalized peptides synthesis All protected amino acids were commercially available from Iris Biotech GMBH or Bachem. Peptide syntheses were carried out on a 0.1-mmol scale using an ABI Model 431A peptide synthesizer (Applied BioSystems), starting from the appropriate Wang Tentagel resin, with 10 equiv. of the protected Fmoc-amino acid and HBTU/DIEA for the activation. The crude peptides were purified by HPLC to obtain purity over 97%. Production of MAM protein in a bacterial heterologous system The gene encoding the MAM protein was PCR amplified from genomic DNA. PCR product was digested, purified, and cloned in pSTABY plasmid and introduced in Escherichia coli. Protein extraction was carried out and verified by Coomassie blue staining and Western Blot. Cellular assays for anti-inflammatory effect Intestinal epithelial cell lines (Caco-2, HT29 and mucus-secreting HT29-MTX), from the European Collection of Cell Cultures (Wiltshire, United Kingdom) and INSERM U505 (Institut des Cordeliers, France), were cultured in supplemented DMEM (PAA) at 37°C in a 5% CO2 incubator. Cells were incubated with different F. prausnitzii supernatant fractions (1/100, 1/500 and 1/1000 dilutions in DMEM medium) or molecules of interest (2 to 10 μM) and, according to the cell lines used, stimulated with various pro-inflammatory cytokines (IL-1β 15 ng. mL−1 or TNFα 10 ng. mL−1). After 6h incubation, cell supernatants were removed for IL-8 analysis. Protein concentrations were determined in cell lysates using Bicinchoninic acid protein assay (Pierce, Rockford, IL) according to the manufacturers instructions. The IL-8 level was determined in duplicate in cell supernatants using ELISA kit DuoSet (R&D systems, Minneapolis, MN). Effect on NF-κB pathway (phosphorylated and total JNK, P38, ERK1/2, IκB, NF-κB) of F. prausnitzii supernatant was tested on Caco-2 cell line using multiplex assay (Bio-Plex®Bio-Rad). Transfected cell lines and NFκB reporter assay The cDNA encoding the protein ZP05614546.1 was cloned in 3xFlag (C-term) pCMV vector. Transfection was performed using transfectin™ Lipid reagent (Bio-Rad) in opti-MEM® medium (Gibco, Life Technologies) in different epithelial cells, namely HEK293T, HT29, and TLR4/MD2/CD14 stably-transfected HEK293T (Invivogen). Different concentrations of cDNA, depending on the considered plasmid, were used: 0.8 ng.μL−1 of the plasmid MAM, the plasmid containing cDNA of Carma1 and their empty equivalents, 0.08 ng.μL−1 of the NF-κB reporter plasmid and 0.2 pg.μL−1 of the Renilla luciferase control reporter vector. For HEK293T-TLR4/MD2/CD14, activation of the NF-κB pathway was performed by administration of LPS 100 ng.mL−1 (Sigma-Aldrich). Positive control of inhibition of NF-κB pathway was obtained with SN50 50μM (Enzo Life Sciences). NF-κB reporter assay was carried out, after 24h incubation, using Dual-Luciferase® Reporter Assay system (Promega). In the same manner, MAM activity on STAT3 pathway, activated by colivelin 0.1 nM (Santa Cruz Biotechnology), was evaluated. Cellular imaging IF in Hela transfected cells HeLa or HEK293T (both from American Type Culture Collection) cells were grown in DMEM (Invitrogen) with 10% FCS (HyClone) at 37°C and 5% CO2. Twenty-four hours following transfection, cells were washed and fixed. Cells were stained using appropriate antibodies for 1h. Cells were imaged using a Leica SP5 confocal microscope. In vivo assays for anti-inflammatory effect of MAM protein Plasmid construction MAM encoding plasmid (pILMAM) was created by a fusion between pIL253 [35] cut with PstI (Fermentas) and pCMV including DNA of MAM (see above) cut with Sbf1. Empty equivalent (pILEMPTY) was created using the same method, but with fusion of pIL253 empty pCMV. pILMAM and pILEMPTY were transformed in L. lactis MG1363 as described by Langella et al. (1993) [36]. L. lactis strains were thereafter grown on M17 medium. DNBS-Induced Colitis Mice were fed with L. lactis harboring pILMAM or pILEMPTY plasmid. After acclimatization, bacterial suspensions of L. lactis MG1363 (5.109 CFU in 200 μL), were administered to mice (C57BL/6, 6 weeks, male) daily by intragastric gavage from day 7 before until day 3 after induction of colitis. DNBS solution in 30% ethanol was administered intrarectally (at a dose of 100 mg. kg−1 body weight). Inflammation was monitored 72 h after DNBS administration. Mice were weighed before DNBS administration and at killing. Two groups of 8 mice were investigated, fed with L. lactis harboring pILMAM or pILEMPTY. Experiments were performed 3 times. Inflammation Score Assessment of DNBS Colitis The colon was removed, dissected free of fat and mesentery, carefully opened, and cleaned. Colon length was measured. Colonic damage and inflammation were assessed blindly according to the Wallace criteria. For histological assessment, a colon sample located in the most inflamed area was fixed in 4% paraformaldehyde acid (Sigma) and embedded in paraffin. Four micrometer sections were stained with hematoxylin/eosin and examined blindly according to the method described by Dieleman et al [37]. Histological scores were carried out according to Ameho criteria. MAM detection from mice purified intestinal epithelial cells Small and large intestine were dissected, washed, cut in 1cm pieces and incubated for 20 min at 37°C, 250 rpm in extraction buffer (EDTA 5 mM, DTT 0.145 mg. mL−1, 25 mM HEPES, 50 μM 2-mercaptoethanol (Sigma), RPMI (Lonza)). After incubation remaining solid tissues were withdrawn. Incubation mix containing epithelial cells was centrifuged and pellets lyzed by ultrasound in cold PBS (Lonza). After centrifugation supernatant was kept at −80°C in RNAlater (Qiagen) and RNA was extracted with RNeasy Mini Kit (Qiagen). Retro Transcription was performed using SuperScript® II Reverse Transcriptase (Sigma) with random primers. RT-PCR for MAM mRNA detection was performed using Phusion High-Fidelity DNA Polymerase (Thermo Scientific) with the following primers: 5′-ACTCTGGTTGGCAACACCTT-3′ and 5′-CGATCGGGTTGCCCTTAACA-3′. Immunoblotting targeting the FLAG region of the MAM expressed protein (pILMAM construct) was used to detect MAM in purified epithelial cells from gut mucosa of sacrificed mice (on day 3). Lymphocyte Isolation and Measurement of Cytokine Production Lymphocyte suspensions were prepared from the Mesenteric Lymph Nodes (MLN) by pressing cells through a 70-mmol/L Falcon nylon cell strainer (BD Biosciences, San Jose, CA). Lymphocytes were counted by flow cytometry (Accuri C6), resuspended in culture medium (RPMI, Lonza with 100 Unit of Streptomicin Penicilin, PAA Laboratories and 10% SVF, Lonza) and activated with coated anti-mouse antibody CD3e and CD28 (eBioscience). Cytokine production was assessed in supernatant by ELISA (IL-17A and INF-γ, Mabtech) after 48h incubation. MAM detection from dixenic mice Caecum content of gnotobiotic mice harboring F. prausnitzii (A2-165) and Escherichia coli (K-12 JM105) subjected to 2,4,6-trinitrobenzenesulfonic acid (TNBS)-induced acute colitis were gifted by Sylvie Miquel and Muriel Thomas. To obtain dixenic E. coli/F. prausnitzii-diassociated mice, BALB/c germfree mice (obtained from the germfree rodent breeding facilities of Anaxem-Micalis; INRA, Jouy-en-Josas, France) were orally inoculated with a fresh culture of E. coli JM105 (108 to 109 CFU /ml). Following this, the E. coli monoassociated mice were inoculated with F. prausnitzii A2-165 as previously described [19]. One month after the stable implantation of the two strains, mice were intrarectally inoculated with 2,4,6-trinitrobenzenesulfonic acid (TNBS) (50 mg/kg body weight) to induce acute colitis, or with the vehicule (0.9% NaCl–ethanol) to obtain control mice [20]. Inflammation was monitored 48h after TNBS administration. 250 μL of phenol pH4,8 / chloroforme-isoamylalcohol (5 :1) (Sigma) was added to caecum contents and the mix was agitated strongly. After shaking, 12.5 μL of SDS 20% and 25 μL of sodium acetate at 3M pH8 was added. Samples were mixed with fast prep 40 s at power 5 and 20 s at power 5, followed by centrifugation for 15 min at 13000g. Superior phases were mixed with 250 μL of chloroforme-isoamylalcohol and centrifuged. Superior phases were treated with High Pure RNA Isolation Kit (Roche) according the manufacturers instructions. Retro Transcription was performed using SuperScript® II Reverse Transcriptase (Sigma) with the following primers: 5′-ACTCTGGTTGGCAACACCTT-3′ and 5′-CGATCGGGTTGCCCTTAACA-3′. RT-PCR for MAM mRNA detection was performed using Phusion High-Fidelity DNA Polymerase (Thermo Scientific) with the following primers: 5′-ACTCTGGTTGGCAACACCTT-3′ and 5′-CGATCGGGTTGCCCTTAACA-3′. Statistical analysis Statistical analysis for significant differences was performed using Student’s t-test or Mann-Whitney test when appropriate using JMP® software (Abacus Concepts, Berkeley, CA). Animal experiments were performed three times. Statistical significance was considered when P < 0.05. Results Identification of MAM protein Peptidomic analysis of F. prausnitzii supernatant Two fractions (F2 and F3) from F. prausnitzii supernatant exerted inhibitory effects on IL1-β-induced IL-8 secretion in intestinal epithelial Caco2 cells (Figure 1). For comparison, the MALDI-TOF MS spectra generated from the culture medium fractions F2’ and F3’ and F. prausnitzii culture supernatant fractions F2 and F3 were also analysed. Analysis of F2’ and F3’ vs F2 and F3 using MALDI-TOF MS allowed us to detect seven ions ([M+H]+ m/z 1733.93, 1832.92, 1946.97, 2047.95, 2146.94, 2579.72 and 4601.06) in F2 and F3 fractions only (Figure 2 A-B and Supplementary files S1, S2, S3). The seven ions of interest were fragmented by an ESI FT-ICR mass spectrometer, and identified by de novo sequencing. To illustrate the identification of compounds, one ion of interest (here: m/z 2146.94) corresponding to [M+2H]2+ m/z 1073.61 was fragmented by collision induced dissociation (CID) (Figure 2 C). From various CID spectra (zoom, deconvoluted, product ion at [M+2H]2+ m/z 831.96), the complete sequence was finally obtained unambiguously: VT[I/L]VGNTF[I/L]QST[I/L]NRT[I/L]GV[I/L] (supplementary files S4, S5 and S6). The isomeric amino acids in square brackets [I or L] were undefined. The same procedure was used to determine the sequence of the 6 other peptides. Identification of a protein and its derived peptides from F. prausnitzii The in silico analysis of F. prausnitzii genome (Blast and ProteinInfo software with NCBInr) first allowed to complete the sequence of the peptide 4600.28 Da, removing any ambiguity between leucine and isoleucine, and then to demonstrate that the seven isolated peptides all derived from the F. prausnitzii protein ZP 05614546.1. Sequences of protein ZP 05614546.1, called MAM (Microbial Anti-inflammatory Molecule or named after its discoverer Marie-Anne Maubert), and its derived peptides are presented in table 1. To investigate structure/function relationships of MAM protein, a sequence search was performed using the program BLAST. The BLAST search recovered several sequences that are homologous to MAM, all belonging to F. prausnitzii strains (Figure 3), with statistically significant, low e-values. The sequence alignment of MAM to its other F. prausnitzii counterparts (Figure 3) revealed high sequence identity (> 34%), with strong sequence conservation in the 1–71 N-terminal part and in the 105–118 segment (MAM numbering). The characterized peptides Pep1-5 corresponded to a region of the protein, which is conserved in F. prausnitzii orthologs. A sequence analysis based on ExPASy bioinformatics tools indicated that MAM does not contain regions of low sequence complexity and intrinsic disorder, and should therefore adopt a compact, globular fold. Although the sequence exhibited two highly hydrophobic regions around residues 23–38 and 103–121, MAM was not predicted as a transmembrane protein using DAS-TM, HMMTOP or PHDhtm programs. In order to get further structural information on MAM protein, the results of the BLAST search were examined using the hits displaying higher e-values. Some sequence similarity was observed with proteins belonging to other Firmicutes (Roseburia intestinalis) having a putative GGDEF domain. A sequence alignment of MAM with this GGDEF domain yielded 15% sequence identity and 38% sequence similarity over the whole sequence, both proteins having comparable sizes (Figure 4A). Based on these results, a homology model was tentatively built using Modeller program [38] (Figure 4C). The inspection of the model indicated that the spatial distribution of polar and hydrophobic residues was compatible with such a globular fold. GGDEF domains share a α/β topology typically composed of a five-stranded β-sheet core flanked by five α-helices (Figure 4B). The alignment showed that the short C-terminal β-strand was absent in the MAM model, as also observed in some GGDEF domains [39]. Interestingly, the region encompassing the characterized peptides Pep1-5 corresponds to helix H3 and is exposed at the surface of the model (Figure 4C). Anti-inflammatory effect of the MAM protein Direct anti-inflammatory effect of MAM protein and its derived peptides MAM protein produced through heterologous expression system (e.g. E. coli) was systematically found in hydrophobic fractions, thus compromising any direct testing of anti-inflammatory effects. To circumvent this problem, MAM derived peptides were functionalized (pegylation, CPP coupling, modification of peptidic bond…), and/or mixed with amphiphilic partners to optimize their solubility. However, none of these compounds showed significant and reproducible anti-inflammatory effect in cellular assays. We thus decided to change our strategy. Previous experiments suggested that F. prausnitzii supernatant acts distally on NF-κB pathway in epithelial cells (data not shown). We reasoned that MAM or its derived peptides could act on intracellular targets. We thus assessed the effect of direct overexpression of MAM protein in epithelial cells. Anti-inflammatory effect of MAM protein after transfection in eukaryotic epithelial cells A full-length Flag-MAM cDNA was transfected in HEK293T and HT29 cells and the effect of the expressed protein was tested on NF-κB pathway, since preliminary experiments, using a multiplex assay (see Material and Methods), suggested an effect on the distal part of this pathway (data not shown). Expression of MAM was confirmed in HEK293T, MD2-TLR4-CD14 HEK293T and HT29 cells by Western blot analysis with anti-Flag antibody (supplementary file S7). Using a NF-κB reporter system in these human epithelial cells, we showed that expression of MAM protein was able to block NF-κB activation induced by Carma-1 or LPS in a dose-dependent manner (Figure 5 A-C). The specificity of the NF-κB activation in our system was validated by the inhibitor effect of SN50, a well known NF-κB inhibitor. Moreover, the specificity of the effect of MAM on the NF-κB pathway was highlighted by the lack of effect on STAT3 pathway in the same eukaryotic system (Figure 5 D). When NF-κB activation was performed using IκκB (distal part of the NF-κB pathway), MAM transfected cells still exhibited a reduced NF-κB activity (Figure 6). Subcellular location of MAM protein in HeLa and in HEK293T cells In agreement with this last result, MAM protein was localized around the cell nucleus in MAM-Flag transfected HeLa cells (Figure 7 A-D) and co-localized with IκκB in transfected HEK293T (Figure 7 E-H). MAM effect in DNBS-induced colitis in mice To mimic more adequately the in vivo situation, a food-grade bacterium, Lactoccocus lactis was modified to produce the cDNA coding for MAM, resulting in a kind of “in vivo transfection”. This allowed delivery of MAM cDNA directly by bacteria in mice subjected to DNBS treatment. Although no heterologous protein is produced by the bacterial strain, we verified that there is no difference in the growth between L. lactis pILEMPTY and pILMAM (supplementary file S8). To confirm that our strategy effectively leads to in vivo production of MAM, we performed western blot analysis on small bowel and colon epithelial cells of mice fed with L. lactis (Figure 8). Expression of MAM at the mRNA level was also confirmed (supplementary file S9). Weight loss was significantly lesser in mice fed with L. lactis pILMAM at day 1 and 2 (Figure 9B). After dissection of the gut, the Wallace score was also significantly lower in pILMAM (Figure 9A). We subsequently isolated lymphocytes from MLN of both mice groups and assessed their cytokine production following stimulation with anti-CD3 and anti-CD28 antibodies. The two pro-inflammatory cytokines IL-17A and INF-γ were produced in lower amounts in cells isolated from L. lactis pILMAM fed mice (Figure 9C-D). No histological differences were observed between mice fed with L. lactis pILMAM and L. lactis pILEMPTY (supplementary file S10). Taken together, these results show that L. lactis pILMAM exhibits significant protective effect in DNBS-induced colitis. MAM detection from dixenic mice MAM mRNA was detected in vivo for the first time in our model of dixenic mice E. coli / F. prausnitzii (Figure 10). Sequencing confirmed that the detected mRNA was MAM encoding mRNA. Discussion Herein, we were able to identify a unique and original 15 kDa protein (ZP05614546.1), called MAM, with anti-inflammatory properties produced by F. prausnitzii. This protein and/or derived peptides involved in the anti-inflammatory effect are able to inhibit the NF-κB pathway in several intestinal epithelial cells lines. Interestingly, L. lactis delivering a MAM encoding plasmid was also able to prevent DNBS-colitis in mice. This discovery is in line with a series of work performed by our group. Indeed, for many years, we and others have described CD-associated dysbiosis and observed a strong restriction in the biodiversity of the Firmicutes (one of the two major bacterial phyla in the normal gut microbiota [40]) and a quantitative decrease in bacteria belonging to Clostridium leptum group [6] [14] [41]. Faecalibacterium prausnitzii is a dominant species of this group. We have gone a step further, looking at the impact of these changes in gut microbiome on CD patient outcomes and inflammation pathways. In this setting, we were able to show that low level of F. prausnitzii in ileal mucosa of CD patients was predictive of postoperative recurrence [14] and again, more recently, we observed that low level in F. prausnitzii in faeces was predictive of CD relapse in patients in remission [42]. Moreover, this bacterium exerts anti-inflammatory properties in vitro and in vivo [14]. In this previous work the anti-inflammatory effects of F. prausnitzii were mediated by secreted molecules in the culture supernatant. These bioactive molecules were highly suspected to be protein-derived peptides from F. prausnitzii. Thus, it remained a major task to isolate these molecules, to characterize and to analyze their immunomodulatory effects. We met several difficulties in achieving these goals. First, finding bioactive molecules in F. prausnitzii supernatant was challenging. We made it possible by using a bio-guided strategy generated from the culture medium fractions and differential MALDI-TOF MS analysis. Thanks to ultra-high resolution of the FT-ICR, the complete sequences of the peptides were finally obtained unambiguously by de novo sequencing. To note, the easy coupling of mass spectrometry with separation techniques, such as liquid chromatography, and its sensitivity makes it a method of choice for detection of various molecules present at trace levels in complex mixtures, such as a bacterial supernatant. Surprisingly, after in silico analysis, the seven peptides all originate from the same unknown protein ZP05614546.1 from F. prausnitzii. Some sequence similarities were observed with proteins belonging to other Firmicutes (Roseburia intestinalis) having a putative GGDEF domain. These proteins are described in the UnitProtKB database as putative since they have not yet been characterized and have unknown functions. Strikingly, MAM protein does not resemble any other known protein from Gram-positive or Gram-negative bacteria. These GGDEF domains are typically found in enzymes endowed with diguanylate cyclase activity and involved in the biosynthesis of cyclic di-GMP, a widespread signaling molecule in bacteria [43]. Accordingly, a BLAST search against the PDB database also revealed a hit with a protein from Methylococcus capsulatus containing a GGDEF domain. Although MAM protein might fold as a GGDEF domain, it lacks the critical catalytic residues (including the GGDEF sequence) that are involved in diguanylate cyclase activity. Furthermore, other residues involved in nucleotide binding such as the RXXD motif [39] are also missing. It is therefore very unlikely that MAM has any catalytic or regulatory functions involved in cyclic di-GMP signaling. Thus, no clear function could be assigned to this protein. High prevalence of nonpolar residues prevented us to provide direct characterization of the putative anti-inflammatory activity. To overcome this technical lock, we applied a molecular approach by transfecting MAM cDNA in a series of epithelial cell lines and showing a significant decrease in the activation of the NF-κB pathway with a dose-dependent effect. Finally, a L. lactis strain delivering a MAM encoding plasmid was able to prevent DNBS colitis in mice. These results demonstrate that MAM protein supports, at least partly, the anti-inflammatory effect exerted by F. prausnitzii. One can hypothesized that F. prausnitzii could exert its anti-inflammatory effect on host cells through many molecular patterns. In fact, when performing preliminary experiments on F. prausnitzii supernatant, we observed that anti-inflammatory effect was not abrogated by heat (above 70°c), enzyme digestion (trypsin, lipase, amylase) or MW filtration (below 15kDa) (Maubert M. A., unpublished data). This indicates that various metabolites other than MAM could contribute to this effect. Furthermore, it has been shown recently that F. prausnitzii is a major inducer of Clostridium-specific IL-10-secreting regulatory T cell subset present in the human colonic lamina propria and blood [44]. This indicates that several cell types are targeted through interactions between F. prausnitzii and the host to maintain and shape gut barrier immune function. In this setting, bacterial metabolites such as the abundant microbial-derived short-chain fatty acids (SCFA) have been identified to induce signaling effects regulating colonic regulatory T cell homeostasis [30]. To note human colonic butyrate producers are Gram-positive Firmicutes from which the two most abundant groups are related to Roseburia spp. and to F. prausnitzii. In fact, the intestinal tract of mammals is home to 1013 to 1014 commensal bacteria composed of hundreds of species of which certain are able to play specific roles in determining the immunological balance in host [27 45]. Atarashi and colleagues have recently demonstrated that Clostridial species induce regulatory T cells through SCFA pathway stimulating epithelial cells to produce TGFβ, contributing to regulatory T cells differentiation and expansion [46]. Other species-specific bacterial molecules, such as B. fragilis-derived polysaccharide A, have previously been demonstrated to have immunomodulatory functions [47] [48] [49]. Another study demonstrated that a mixture of probiotic strains, including Lactobacillus and Bifidobacterium, enhanced the production of TGF-β and IDO from dendritic cells and consequently induced Treg cells [50]. Thus, our work uncovering a new anti-inflammatory protein from F. prausnitzii, a major microbiota species of gut microbiota, reinforces the role of a metabolic interface of promiscuous bacterial molecules on gut mucosa physiology. However, a metabolomic approach based on gnotobiotic model permitted to propose other hypothesis concerning F. prausnitzii anti-inflammatory properties [20]. Authors identified various metabolites, particularly salicylic acid, specifically associated with the presence of F. prausnitzii. This study, associated to our results, confirm the complex action mechanisms of F. prausnitzii to limit the inflammatory process. To conclude, our work opens new lines of evidence that the impact of CD-associated dysbiosis could, at term, change our practice and lead to novel strategies to prevent and treat inflammatory bowel diseases. Although discovering an anti-inflammatory molecule constitutes a first step towards new anti-inflammatory drugs in CD, MAM could represent a targeted biomarker for CD, regarding the value of F. prausnitzii for predicting CD relapse. In these perspectives, there is further need for deciphering the role of MAM in the gut ecosystem. In particular, emphasis should be placed on finding the function of MAM and the mechanism of its production within the bacteria before considering it a target for CD management. Supplementary Material Supplement Acknowledgments We thank Association François Aupetit (2009) and Agence Nationale de la Recherche (Mi2 2010) for funding this work. We also thank Lucette Groisard, Loïc Brot, Chantal Bridonneau and Isabelle Naas for technical assistance and Joëlle Masliah and Muriel Thomas for their scientific support. Marie-Anne Maubert thanks Jean Claude Tabet and the TGE High Field FT-ICR (CNRS) for providing the access to the FT-ICR mass spectrometer. Caecum content of Gnotobiotic mice harboring F. prausnitzii (A2-165) and Escherichia coli (K-12 JM105) subjected to 2,4,6-trinitrobenzenesulfonic acid (TNBS)-induced acute colitis have been gently given by Sylvie Miquel and Muriel Thomas. This last study was a part of the FPARIS collaborative project selected and supported by the Vitagora Competitive Cluster and funded by the French FUI (Fond Unique Interministériel; FUI no. F1010012D), the FEDER (Fonds Européen de Développement Régional; Bourgogne no. 34606), the Burgundy Region, the Conseil Général 21, and the Grand Dijon. This work was also supported by Merck Médication Familiale (Dijon, France) and Biovitis (Saint Étienne de Chomeil, France). Abbreviations IBD inflammatory bowel disease CD Crohn’s disease MALDI-TOF Matrix-Assisted Laser Desorption/Ionisation - Time Of Flight FT-ICR Fourier Transform - Ion Cyclotron Resonance HPLC High-performance liquid chromatography DNBS Dinitrobenzene Sulfonic Acid TNBS Trinitrobenzene Sulfonic acid Figure 1 IL-8 response of human Caco2 cells to stimulation in DMEM medium with IL-1β, F. prausnitzii supernatant (SN) or LyBHI medium, and F. prausnitzii supernatant fractions F1 (20% of acetonitrile), F2 (40% acetonitrile) and F3 (80% acetonitrile). The values are expressed as the mean ± SEM in pg IL-8/mg protein. * p < 0.05 (compared to DMEM control with IL-1β in three experiments). Figure 2 (A) MALDI TOF MS spectra (zoom scan for the m/z range 1630–1933) generated from F2′ (A) and F2 fractions (B) showing two ions at m/z 1733.93, 1833.92, only in (B) MS spectrum. (C) FTICR CID spectrum of the [M + 2H]2+ m/z 1073.61 precursor ion (corresponding to the ion of interest [M + H]+ m/z 2146.94). De novo sequencing generated a probable partial amino acid sequence from singly charged ions. The accuracy of mass determination made it possible to attribute this series unambiguously and to differentiate between the isobaric amino acids K and Q. Figure 3 Sequence alignments between MAM protein and 7 other homologous proteins of F. prausnitzii. Sequences were identified using a BLAST search and aligned using ClustalW2 program. The protein identifiers correspond to the following F. prausnitzii strains: C7H4X2, A2-165; R6QJG8 and R6Q1X1, sp. CAG:82; D4KBR2, SL3/3; A8SAI8, M21/2; E2ZMJ5, KLE1255; D4K191 and D4K193, L2-6. The region corresponding to identified peptides Pep1-5 is shown by an arrow. Figure 4 Homology model of MAM protein based on a GGDEF domain template. (A) Sequence alignment of MAM and GGDEF protein from M. capsulatus (Q60BX6); (B) X-ray structure of template protein (PDB entry 3ICL) showing the secondary structure elements; (C) three-dimensional model of MAM calculated with Modeller. The rms deviation on Cα positions of aligned residues is 1.1 Å. The region 49-68 corresponding to the identified peptides Pep1-5 is coloured in yellow. Figure 5 Decrease in activation of the NF-κB pathway after transfection of MAM protein in different epithelial cells: in HEK293T in a dose-dependent manner (A), in TLR4/MD2/CD14 stably transfected HEK293T stimulated by LPS 100 ng. mL−1 (B) and in intestinal cells HT29 (C). SN50 (50 μM) was used as positive control of NF-κB pathway inhibition in HEK293T. No activity of transfected MAM protein was observed on the STAT3 pathway activated by colivelin 0.1 nM (D). * p < 0.05 (compared to activation control). Figure 6 Decrease in activation of the NF-κB pathway after co-transfection of MAM protein and IκκB in HEK293T ( * p< 0.05) Figure 7 Subcellular location of MAM protein in HeLa cells (A, B, C, D) and co-localisation of MAM protein with Iκκβ in HEK293T cells (F, G, H, I). E and J labels correspond to staining controls in untreated cells. Figure 8 Western blot with anti-flag antibody for MAM protein detection in small intestine enterocytes (1) and large intestine enterocytes (2) of mice fed with pILEmpty L. lactis and in small intestine enterocytes (3) and large intestine enterocytes (4) of mice fed with pILMAM L. lactis Figure 9 Effects of intragastric administration of L. lactis bacterial suspension (pILMAM or Empty equivalent) on TNBS-induced colitis in C57BL/6 mice considering Wallace score (A), weight after induction of colitis (B) and quantification using ELISA of IL-17A (C) and INF-γ (D) in colons obtained 48 h after DNBS colitis induction (in pg/mL of total proteins). The values are expressed as the mean ± SEM (*p < 0.05 compared to L. lactis pILEmpty controls). Figure 10 PCR detection of MAM gene on RNA extract from feces of dixenic mice colonized with Faecalibacterium prausnitzii and Escherichia coli (1), on pILMAM plasmid (2, positive control) and on RNA extract of same sample (1) traited by RNAse before RT-PCR (3, negative control). Table 1 Sequences of the 7 peptides identified in F. prausnitzii supernatant and sequence of the MAM protein. Pep1 1732.95 Da GNTFLQSTINRTIGVL Pep2 1832.02 Da VGNTFLQSTINRTIGVL Pep3 1945.10 Da LVGNTFLQSTINRTIGVL Pep4 2046.15 Da TLVGNTFLQSTINRTIGVL Pep5 2145.26 Da VTLVGNTFLQSTINRTIGVL Pep6 4600.28 Da FSGNTTWKEVGNIGKNLFGTNVKGNPIEKNNFGDYAMNALGIA Pep7 2578.40 Da AAVYNLGVA PTKNTVKETE VKFTV Protein ZP 05614546.1  MMMPANYSVIAENEMTYVNGGANFIDAIGAVTAPIWTLDNVKTFNTNIVTLVGNTFLQSTINRTIG  VLFSGNTTWKEVGNIGKNLFGTNVKGNPIEKNNFGDYAMNALGIAAAVYNLGVA  PTKNTVKETEVKFTV Summary Box What is already known about this subject? Crohn’s disease (CD) associated dysbiosis is characterized by low proportion of Faecalibacterium prausnitzii in fecal and mucosa-associated microbiome. Loss of F. prausnitzii is predictive of CD relapse after surgery or in patients treated with immunosupressants. F. prausnitzii exhibits anti-inflammatory effects in vitro and in vivo by secreted metabolites that block NF-κB activation. What are the new findings? F. prausnitzii produces bioactive peptides derived from a single 15 kDa protein (ZP05614546.1) of unknown function named Microbial Anti-inflammatory Molecule (MAM). MAM expression in epithelial cells lines is able to block the NF-κB pathway. Lactococcus lactis harboring a MAM-cDNA encoding plasmid is able to alleviate DNBS-colitis in mice How might this impact on clinical practice in the foreseeable future? Discovery of anti-inflammatory molecules from a commensal bacterium such as F. prausnitzii constitutes the first step towards new anti-inflammatory drugs in CD. Knowing the predictive value of F. prausnitzii for CD relapse, MAM could represent a targeted biomarker for CD. Philippe Seksik declares consulting fees from Abbvie, MSD and Biocodex. Harry Sokol received consulting fee from Danone and Enterome and lecture fee from Abbvie and Biocodex. For the other authors, they declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. References 1 MacDermott RP Alterations of the mucosal immune system in inflammatory bowel disease J Gastroenterol 1996 31 907 16 9027661 2 Hugot JP Chamaillard M Zouali H Association of NOD2 leucine-rich repeat variants with susceptibility to Crohn's disease Nature 2001 411 599 603 11385576 3 Allez M Mayer L Regulatory T cells: peace keepers in the gut Inflamm Bowel Dis 2004 10 666 76 15472533 4 Sartor RB Therapeutic manipulation of the enteric microflora in inflammatory bowel diseases: antibiotics, probiotics, and prebiotics Gastroenterology 2004 126 1620 33 15168372 5 Khor B Gardet A Xavier RJ Genetics and pathogenesis of inflammatory bowel disease Nature 2011 474 307 17 21677747 6 Sokol H Seksik P Rigottier-Gois L Specificities of the fecal microbiota in inflammatory bowel disease Inflamm Bowel Dis 2006 12 106 11 16432374 7 Mukhopadhya I Hansen R El-Omar EM IBD-what role do Proteobacteria play? Nat Rev Gastroenterol Hepatol 2012 9 219 30 22349170 8 Darfeuille-Michaud A Boudeau J Bulois P High prevalence of adherent-invasive Escherichia coli associated with ileal mucosa in Crohn's disease Gastroenterology 2004 127 412 21 15300573 9 Man SM Zhang L Day AS Campylobacter concisus and other Campylobacter species in children with newly diagnosed Crohn's disease Inflamm Bowel Dis 2010 16 1008 16 19885905 10 Kaakoush NO Mitchell HM Campylobacter concisus - A new player in intestinal disease Front Cell Infect Microbiol 2012 2 4 22919596 11 Bohr UR Glasbrenner B Primus A Identification of enterohepatic Helicobacter species in patients suffering from inflammatory bowel disease J Clin Microbiol 2004 42 2766 8 15184464 12 Man SM Zhang L Day AS Detection of enterohepatic and gastric helicobacter species in fecal specimens of children with Crohn's disease Helicobacter 2008 13 234 8 18665930 13 Sokol H Seksik P Furet JP Low counts of Faecalibacterium prausnitzii in colitis microbiota Inflamm Bowel Dis 2009 15 1183 9 19235886 14 Sokol H Pigneur B Watterlot L Faecalibacterium prausnitzii is an anti-inflammatory commensal bacterium identified by gut microbiota analysis of Crohn disease patients Proc Natl Acad Sci U S A 2008 105 16731 6 18936492 15 Hansen R Russell RK Reiff C Microbiota of de-novo pediatric IBD: increased Faecalibacterium prausnitzii and reduced bacterial diversity in Crohn's but not in ulcerative colitis Am J Gastroenterol 2012 107 1913 22 23044767 16 Zhang M Qiu X Zhang H Faecalibacterium prausnitzii inhibits interleukin-17 to ameliorate colorectal colitis in rats PLoS One 2014 9 e109146 25275569 17 Miquel S Martin R Rossi O Faecalibacterium prausnitzii and human intestinal health Curr Opin Microbiol 2013 16 255 61 23831042 18 Martin R Chain F Miquel S The commensal bacterium Faecalibacterium prausnitzii is protective in DNBS-induced chronic moderate and severe colitis models Inflamm Bowel Dis 2014 20 417 30 24418903 19 Wrzosek L Miquel S Noordine ML Bacteroides thetaiotaomicron and Faecalibacterium prausnitzii influence the production of mucus glycans and the development of goblet cells in the colonic epithelium of a gnotobiotic model rodent BMC Biol 2013 11 61 23692866 20 Miquel S Leclerc M Martin R Identification of Metabolic Signatures Linked to Anti-Inflammatory Effects of Faecalibacterium prausnitzii MBio 2015 6 21 Mennigen R Nolte K Rijcken E Probiotic mixture VSL#3 protects the epithelial barrier by maintaining tight junction protein expression and preventing apoptosis in a murine model of colitis Am J Physiol-Gastr L 2009 296 G1140 G49 22 Vora P Youdim A Thomas LS beta-Defensin-2 expression is regulated by TLR signaling in intestinal epithelial cells Journal of Immunology 2004 173 5398 405 23 Zyrek AA Cichon C Helms S Molecular mechanisms underlying the probiotic effects of Escherichia coli Nissle 1917 involve ZO-2 and PKCzeta redistribution resulting in tight junction and epithelial barrier repair Cell Microbiol 2007 9 804 16 17087734 24 Lecuyer E Rakotobe S Lengline-Garnier H Segmented filamentous bacterium uses secondary and tertiary lymphoid tissues to induce gut IgA and specific T helper 17 cell responses Immunity 2014 40 608 20 24745335 25 Gaboriau-Routhiau V Rakotobe S Lecuyer E The key role of segmented filamentous bacteria in the coordinated maturation of gut helper T cell responses Immunity 2009 31 677 89 19833089 26 Ivanov II Atarashi K Manel N Induction of intestinal Th17 cells by segmented filamentous bacteria Cell 2009 139 485 98 19836068 27 Atarashi K Tanoue T Shima T Induction of colonic regulatory T cells by indigenous Clostridium species Science 2011 331 337 41 21205640 28 Round JL Mazmanian SK The gut microbiota shapes intestinal immune responses during health and disease Nat Rev Immunol 2009 9 313 23 19343057 29 Schnupf P Gaboriau-Routhiau V Cerf-Bensussan N Host interactions with Segmented Filamentous Bacteria: an unusual trade-off that drives the post-natal maturation of the gut immune system Semin Immunol 2013 25 342 51 24184014 30 Smith PM Howitt MR Panikov N The microbial metabolites, short-chain fatty acids, regulate colonic Treg cell homeostasis Science 2013 341 569 73 23828891 31 Bouskra D Brezillon C Berard M Lymphoid tissue genesis induced by commensals through NOD1 regulates intestinal homeostasis Nature 2008 456 507 10 18987631 32 Yan F Cao H Cover TL Soluble proteins produced by probiotic bacteria regulate intestinal epithelial cell survival and growth Gastroenterology 2007 132 562 75 17258729 33 Sougioultzis S Simeonidis S Bhaskar KR Saccharomyces boulardii produces a soluble anti-inflammatory factor that inhibits NF-kappaB-mediated IL-8 gene expression Biochem Biophys Res Commun 2006 343 69 76 16529714 34 Kelly D Campbell JI King TP Commensal anaerobic gut bacteria attenuate inflammation by regulating nuclear-cytoplasmic shuttling of PPAR-gamma and RelA Nat Immunol 2004 5 104 12 14691478 35 Holo H Nes IF High-Frequency Transformation, by Electroporation, of Lactococcus lactis subsp. cremoris Grown with Glycine in Osmotically Stabilized Media Appl Environ Microbiol 1989 55 3119 23 16348073 36 Langella P Le Loir Y Ehrlich SD Efficient plasmid mobilization by pIP501 in Lactococcus lactis subsp. lactis J Bacteriol 1993 175 5806 13 8376328 37 Dieleman LA Palmen MJHJ Akol H Chronic experimental colitis induced by dextran sulphate sodium (DSS) is characterized by Th1 and Th2 cytokines Clinical and Experimental Immunology 1998 114 385 91 9844047 38 Eswar N Webb B Marti-Renom MA Comparative protein structure modeling using MODELLER Curr Protoc Protein Sci 2007 Chapter 2:Unit 2 9 39 Whitney JC Colvin KM Marmont LS Structure of the cytoplasmic region of PelD, a degenerate diguanylate cyclase receptor that regulates exopolysaccharide production in Pseudomonas aeruginosa J Biol Chem 2012 287 23582 93 22605337 40 Eckburg PB Bik EM Bernstein CN Diversity of the human intestinal microbial flora Science 2005 308 1635 8 15831718 41 Manichanh C Rigottier-Gois L Bonnaud E Reduced diversity of faecal microbiota in Crohn's disease revealed by a metagenomic approach Gut 2006 55 205 11 16188921 42 Rajca S Grondin V Louis E Alterations in the intestinal microbiome (dysbiosis) as a predictor of relapse after infliximab withdrawal in Crohn's disease Inflamm Bowel Dis 2014 20 978 86 24788220 43 Ryan RP Cyclic di-GMP signalling and the regulation of bacterial virulence Microbiology 2013 159 1286 97 23704785 44 Sarrabayrouse G Bossard C Chauvin JM CD4CD8alphaalpha Lymphocytes, A Novel Human Regulatory T Cell Subset Induced by Colonic Bacteria and Deficient in Patients with Inflammatory Bowel Disease PLoS Biol 2014 12 e1001833 24714093 45 Hill DA Artis D Intestinal bacteria and the regulation of immune cell homeostasis Annu Rev Immunol 2010 28 623 67 20192812 46 Atarashi K Tanoue T Oshima K Treg induction by a rationally selected mixture of Clostridia strains from the human microbiota Nature 2013 500 232 6 23842501 47 Mazmanian SK Round JL Kasper DL A microbial symbiosis factor prevents intestinal inflammatory disease Nature 2008 453 620 5 18509436 48 Round JL Mazmanian SK Inducible Foxp3+ regulatory T-cell development by a commensal bacterium of the intestinal microbiota Proc Natl Acad Sci U S A 2010 107 12204 9 20566854 49 Geuking MB Cahenzli J Lawson MA Intestinal bacterial colonization induces mutualistic regulatory T cell responses Immunity 2011 34 794 806 21596591 50 Kwon HK Lee CG So JS Generation of regulatory dendritic cells and CD4(+)Foxp3(+) T cells by probiotics administration suppresses immune disorders P Natl Acad Sci USA 2010 107 2159 64