article
stringlengths 4.26k
146k
| summary
stringlengths 31
3.31k
| section_headings
stringlengths 9
553
| keywords
stringlengths 0
1.24k
| year
stringclasses 13
values | title
stringlengths 20
281
| article_length
int64 1.05k
36.8k
| summary_length
int64 14
689
|
|---|---|---|---|---|---|---|---|
Kidney function depends on the nephron, which comprises a blood filter, a tubule that is subdivided into functionally distinct segments, and a collecting duct. How these regions arise during development is poorly understood. The zebrafish pronephros consists of two linear nephrons that develop from the intermediate mesoderm along the length of the trunk. Here we show that, contrary to current dogma, these nephrons possess multiple proximal and distal tubule domains that resemble the organization of the mammalian nephron. We examined whether pronephric segmentation is mediated by retinoic acid (RA) and the caudal (cdx) transcription factors, which are known regulators of segmental identity during development. Inhibition of RA signaling resulted in a loss of the proximal segments and an expansion of the distal segments, while exogenous RA treatment induced proximal segment fates at the expense of distal fates. Loss of cdx function caused abrogation of distal segments, a posterior shift in the position of the pronephros, and alterations in the expression boundaries of raldh2 and cyp26a1, which encode enzymes that synthesize and degrade RA, respectively. These results suggest that the cdx genes act to localize the activity of RA along the axis, thereby determining where the pronephros forms. Consistent with this, the pronephric-positioning defect and the loss of distal tubule fate were rescued in embryos doubly-deficient for cdx and RA. These findings reveal a novel link between the RA and cdx pathways and provide a model for how pronephric nephrons are segmented and positioned along the embryonic axis.
The kidney eliminates metabolic waste in the body using highly specialized structures called nephrons. Individual nephrons are composed of a blood filter (renal corpuscle), a tubule that recovers or secretes solutes, and a collecting duct [1]. The renal corpuscle contains epithelial cells called podocytes that form the slit-diaphragm filtration barrier and allow collection of substances from the blood [2]. In a number of vertebrate species, including some mammals, the renal corpuscle is connected to the tubule by a short stretch of ciliated epithelium called the neck segment that guides filtrate entry into the tubule [3–5]. The mammalian nephron tubule is subdivided into a series of proximal and distal segments connected to a collecting duct [1,6]. The polarized epithelial cells in the tubule segments have a unique ultrastructure and express a select cohort of solute transporters [1]. Thus, each segment is functionally distinct and performs the transport of particular solutes that are required for proper renal function. In higher vertebrates, three kidneys of increasing complexity arise sequentially from the intermediate mesoderm (IM): the pronephros, the mesonephros, and the metanephros [7]. The pronephros and mesonephros degenerate in succession, with the metanephros serving as the adult kidney. Lower vertebrates, such as fish and amphibians, develop a pronephros during embryonic stages, and then form a mesonephros that will be used throughout their adult life [8–10]. Each of these kidneys contains the nephron as its basic functional unit [8]. To date, much of our knowledge of kidney development has come from gene-targeting studies in the mouse [7,11,12]. These experiments have identified a number of genes that play essential roles in the early stages of metanephros development, but there is a limited understanding of the molecular pathways governing the later stages of kidney ontogeny, when individual nephrons form and become segmented [7]. The zebrafish is an ideal genetic and developmental model system for dissecting the molecular mechanisms of nephron formation because of the anatomical simplicity of the pronephros, which contains two nephrons as opposed to the thousands of nephrons in a mammalian metanephros [9]. During zebrafish development, bilateral stripes of IM lying on either side of the trunk undergo a mesenchymal-to-epithelial transition to form the pair of pronephric nephrons. The anteriormost renal progenitors differentiate into podocytes, which migrate medially and fuse at the midline to form a single renal corpuscle. The nephrons also fuse posteriorly at the cloaca to form a shared exitway. From a functional standpoint, these pronephric nephrons have been thought to consist of three parts: (1) the blood-filtering renal corpuscle, (2) a very short tubule region that transports solutes, and (3) long pronephric ducts that convey the resulting waste to the cloaca [9]. Contrary to this model, recent studies have suggested that the ‘duct' region possesses regional segmentation, based on the restricted expression boundaries of solute transporter orthologues known to be expressed in the tubule segments of metanephric nephrons. For example, a rostral stretch of the pronephric duct expresses the endocytic receptor megalin (lrp2) [13] and the sodium bicarbonate transporter NBC1 (slc4a4) [14], which are expressed in the proximal tubule in mammals. These reports raise the possibility that portions of the pronephros considered to be duct might in fact be tubule, thus suggesting that the organization of the zebrafish pronephros is more complex than previously appreciated. However, a complete model of the molecular anatomy of the zebrafish pronephros and whether there is a functional correlation to the segments of the mammalian nephron remain unclear. Furthermore, the pathway (s) directing segmentation of the pronephros along the embryonic axis are unknown. Numerous factors are known to control segmental patterning along the anterior-posterior (A-P) axis during vertebrate development and thus provide candidate pathways that might act to establish pronephros segmentation. Retinoic acid (RA) signaling is vital for directing the A-P regionalization of tissues deriving from all three germ layers, such as the hindbrain, paraxial mesoderm, and gut [15–19]. Control of RA production via retinaldehyde dehydrogenase (RALDH) synthesizing enzymes [20] and the degradation of RA via the CYP26 catabolizing enzymes establishes both the location and timing of RA signaling [21,22]. In addition to RA, the caudal (cdx) transcription factors (Cdx1, Cdx2, and Cdx4 in mammals and cdx1a and cdx4 in zebrafish) are responsible for determining vertebral identity and directing posterior body formation [23–31]. cdx genes are known to act as master regulators of the homeobox (hox) transcription factors [25], and in turn, overlapping domains of hox gene expression along the A-P axis are thought to confer segmental identities [32]. In mice, loss of Cdx function causes posterior shifts in Hox gene expression that are associated with abnormal vertebral patterning, and posterior truncations due to defects in the extension of the embryo axis [23,25–27,33]. Similarly, studies in zebrafish have shown that the loss of cdx4 function or deficiency of both cdx1a and cdx4 causes shifts in hox gene expression domains, a shortened body axis, and altered patterning of the blood, vascular, and neural tissues [24,28–31]. These lines of evidence indicate that the cdx genes play essential roles in controlling cell fates along the embryonic axis; however, the molecular mechanisms underlying these effects have not been elucidated [34]. In this study, we undertook a functional genomics approach to identify new markers of the zebrafish pronephros. From this analysis, we found that the pronephros is composed of at least eight regions, including two proximal and two distal tubule segments. We explored how segmental identity is controlled during nephrogenesis by testing the roles of RA signaling and the cdx genes. We found that RA is required to induce proximal segment fates and prevent the expansion of distal segment fates, whereas the cdx genes are necessary for positioning the pronephros along the embryonic axis. Embryos deficient in cdx1a and cdx4 displayed a posterior shift in the location of the pronephros and formed proximal but not distal nephron segments. The cdx genes were found to control the expression boundaries of raldh2 (aldh1a2) and cyp26a1, suggesting a model in which the cdx pathway influences where the pronephros forms along the body axis by localizing the source of RA, while subsequent RA signaling acts to direct the segmentation of the pronephros.
To gain insight into the molecular mechanisms that control vertebrate renal development, we undertook a functional genomics approach to identify genes expressed in the kidney. We mined two gene collections, one comprising developmentally expressed genes from embryonic zebrafish cDNA libraries [35] and another compiled from an adult zebrafish kidney library [36]. Gene expression patterns were analyzed by whole-mount in situ hybridization using wild-type zebrafish embryos between the 5 somite stage and 144 hours post fertilization (hpf). We identified a number of genes, including 15 solute transporters, that were expressed within specific subregions of the pronephros. In total, eight distinct regions could be visualized, with some genes expressed in more than one region (Figure 1A). Representative examples of region-specific genes include wt1b, slc20a1a, trpm7, slc12a1, stc1, slc12a3, and gata3, as compared to expression of cdh17, which is found in all tubule and duct progenitors [37] (Figure 1A). We investigated where the mouse or human orthologues of some of these genes are expressed in the mammalian metanephric kidney, and found that many corresponded to segment-specific domains within the nephron. For example, Slc9a3 is expressed in podocytes, the proximal convoluted segment (PCT) and proximal straight segment (PST) (Figure 1B) [38]. Slc20a1 is expressed throughout the entire nephron epithelium, although stronger expression was observed in proximal tubule segments (Figure 1B). Transcripts for Slc13a3 are found in the PST [39], while Slc12a1 is restricted to the thick ascending limb (TAL) and macula densa (MD) (Figure 1B) [40,41]. Slc12a3 is expressed in the distal convoluted tubule (DCT) (Figure 1B) [41]. Lastly, GATA-3 expression specifically marks the collecting ducts (CD) (Figure 1B) [42,43]. Based on this cross-species gene expression comparison, the following identities were assigned to the zebrafish pronephros segments we observed (going from proximal to distal): podocytes (pod), neck (N), PCT, PST, distal early (DE), corpuscle of Stannius (CS), distal late (DL), and the pronephric duct (PD) (Figure 1A and 1E). Our division of PCT and PST within the tubule is based on the observation that the slc20a1a-expressing PCT cells undergo morphogenesis from a linear tube into a coiled structure by 5 days post-fertilization (dpf), while the trpm7- and slc13a1-expressing PST segment maintains a linear structure (Figure 1C). Expression of trpm7 and slc13a1 is discontinuous within the PST, an observation that has been shown recently to reflect the presence of two cell types in this region: transporting epithelia and multiciliated cells [44,45]. The renal corpuscle connects to the PCT via a short segment of cells that express the transcription factor rfx2, and fails to express almost all of our PCT solute transporters (Figure 1D). As rfx2 marks ciliated cells and rfx genes are essential regulators of ciliogenesis [46,47], we hypothesize that this region corresponds to the ciliated neck segment found in other fish species as well as mammals [3–5]. However, a more detailed analysis is needed to confirm this hypothesis. In addition to the neck segment, rfx2 expression was also detected in presumptive ciliated cells along the length of the PST and DE segments, as described previously [44,45]. For the distal tubule, we adopted the DE/DL nomenclature used in Xenopus [48], although the zebrafish DE appears analogous to the TAL segment in mammals and the DL appears analogous to the mammalian DCT segment according to our gene expression comparison. We included the CS as a discrete segment, as it initially arises from the tubular progenitors within the pronephros, but by 48 hpf, it is located just dorsal to the DE/DL boundary [49,50] (unpublished data). The DL segment expresses slc12a3 and connects to the cloaca via a short segment that expresses gata3 and likely represents the PD. Our data are consistent with the notion that the zebrafish pronephric kidney resembles a ‘stretched-out”' mammalian nephron, and suggests that rather than being composed of mostly nephric duct (as currently believed), it is made up of extensive proximal and distal tubule epithelium (Figure 1E). Between 24 and 48 hpf (the start of blood filtration), the pronephros undergoes significant morphogenesis, including the midline migration of podocytes and the growth/extension of the tubules [9]. In order to better quantitate these morphological changes, as well as to precisely define the anatomical boundaries of each segment, we mapped the expression domains of segment-specific markers relative to the somites by performing double whole-mount in situ hybridization with myosin heavy chain (mhc) at 24 and 48 hpf (Figures 2 and S1–S4). At 24 hpf, podocyte and neck progenitors are arranged in a slight curve at the level of somite 3–4 with the anterior boundary of the PCT level with somite 5 (Figures 2 and S1). By 48 hpf, the podocyte progenitors have fused at the midline (level with somite 3) with the presumptive neck region forming a lateral extension that connects with the PCT also situated at the level of somite 3 (Figures 2 and S3). During this time, the length of the PCT, PST, and DE segments increased, possibly due to cell division within each segment (Figure 2 and S1–S4). This growth may provide the driving force that is responsible for the shift in the anterior boundary of the PCT from somite 5 to somite 3 between 24 and 48 hpf, and for the coiling morphogenesis of the PCT observed between 72 and 144 hpf (Figure 1C). However, the DL segment did not increase in length between 24 and 48 hpf, indicating that there is not a uniform expansion in all segments during development. During juvenile development (2–3 wk post-fertilization) the DL segment is proportionately larger than the other segments, suggesting that its expansion predominates at later stages of development (unpublished data). Interestingly, at 24 hpf, we observed an overlap of the DL and PD expression domains at the level of somite 17 (Figure 2 and Figure S2). This overlap may indicate the presence of an additional segment (such as a discrete CNT equivalent), though to date we have not discovered any genes expressed solely in this domain. In addition to mapping the morphological changes that occur between 24 and 48 hpf, we also noted segment specific changes in gene expression patterns during this time. For example, transcripts for the solute transporters slc13a1 (inorganic sulphate transporter), slc13a3 (sodium-dicarboxylate carrier), and slc22a6 (organic anion transporter) were all absent from the PST segment at 24 hpf but were found expressed at 48 hpf (Figures 2 and S3; and unpublished data). Similarly, transcripts for trpm7 (divalent cation-selective ion channel) and slc41a1 (Mg 2+ transporter) were not found in cells of the CS at 24 hpf but could be detected at 48 hpf (Figure 2 and unpublished data). The up-regulation of these genes likely reflects the maturation/differentiation of the segment epithelia before the onset of blood filtration, which begins around 40 hpf [51,52]. However, the glomerulus does not fully mature until 4 dpf, based on the size exclusion of different-sized fluorescent dextrans [53]. Further profiling of segment expression patterns at later stages is needed to investigate whether the maturation of the transporting epithelial is also an ongoing process. We next sought to characterize the developmental pathways that establish the segmentation pattern of the pronephros. Retinoic acid (RA) is essential for the development of numerous tissues during embryogenesis [25]. In vertebrates, a gradient of RA in the upper trunk is responsible for directing the A-P patterning of the hindbrain into segmental compartments [15,54], and gain or loss of RA also affects vertebral identity [25]. Interestingly, RA has been reported to regulate pronephros formation in Xenopus [55]. To explore whether RA was needed for pronephros segmentation, we injected wild-type zebrafish embryos with a morpholino to retinaldehyde dehydrogenase 2 (raldh2), which encodes an enzyme required to produce RA [20]. We then examined raldh2 morphants at 48 hpf with our panel of segment-specific markers, in combination with mhc expression to map segment length and location relative to the somites. Embryos deficient in raldh2 had fewer podocytes, as evidenced by punctate staining of the podocyte markers wt1b, wt1a, and mafb (Figure 3A and unpublished data). Both proximal tubule segments were slightly shortened, based on the reduced expression domains of slc20a1a (PCT) and trpm7 (PST) (Figure 3). Conversely, the distal tubule was expanded in length, with the clck transporter (marking both the DE and DL) expressed in a greater proportion of the cdh17-positive pronephric tubule (note that the overall length of raldh2 morphants is reduced compared to wild-type) (Figure 3). Each segment within the distal tubule was moderately expanded, with a lengthened DE shown by expanded slc12a1 expression, enlarged clusters of stc1-expressing cells that comprise the CS, and lengthened DL shown by slc12a3 expression (Figure 3). The expression domain of gata3 (PD) was also expanded (Figure 3). In contrast, the cloaca marker aqp3 was unchanged (Figure 3A). These data suggest that RA signaling is necessary for podocyte formation and/or survival, as well as for establishing the normal pattern of nephron segmentation. Multiple enzymes are capable of synthesizing RA and a recent analysis of the neckless (nls) (now called aldh1a2) mutant, which is defective in raldh2, demonstrated that there are additional sources of raldh-like enzyme activity in the zebrafish embryo [56–58]. Based on this, we hypothesized that the nephron phenotype in raldh2 morphants represented the effect of reducing RA production, rather than a complete inhibition. To more fully block RA signaling, we utilized a competitive, reversible inhibitor of raldh enzymes, 4-diethylaminobenzaldehyde (DEAB) [59], which has been used to effectively prevent de novo RA synthesis in zebrafish embryos [58,60]. Wild-type embryos were treated with DEAB starting at 60% epiboly (early gastrula) until the 15 somite stage, and nephron segmentation was assayed at 48 hpf by double whole-mount in situ hybridization using mhc expression to mark the somites. Expression of the podocyte markers wt1b, wt1a, and mafb was absent in DEAB-treated embryos, suggesting that podocyte development was completely blocked (Figure 3 and unpublished data). Expression of the PCT and PST markers slc20a1a and trpm7 was also absent, suggesting that these segments had failed to be specified (Figure 3). In contrast, clck expression (marking the DE and DL segments) was dramatically expanded, such that it was present throughout the entire tubule territory (Figure 3). These findings suggest that the pronephric tubule adopts a distal tubule identity when RA synthesis is inhibited. Within this ‘distal-only' pronephros, the DE marker slc12a1 was expressed from the anterior limit of the tubule to almost the middle of its length, the DL marker slc12a3 was expressed from the middle of the tubule to near its posterior limit, and the PD marker gata3 was expanded by an additional four somite lengths (Figure 3). A marked expansion of stc1-expressing cells was also detected, with multiple clusters of cells arranged in bilateral stripes, as opposed to the small groups of stc1-expressing cells seen in wild-type embryos (Figure 3). Expression of aqp3 was not altered, suggesting that progenitors of the cloaca are unaffected by the inhibition of RA production over this developmental interval (Figure 3). It is not known if the observed DEAB phenotype represents a full loss of RA signaling, as trace amounts of maternal RA have been detected in the yolk [61] and as zygotic raldh2 transcripts are expressed prior to 60% epiboly [56,57]. Nevertheless, our results demonstrate that RA, produced by raldh2 and possibly one or more unknown raldh enzymes, plays an essential role in the formation of proximal nephron fates (podocytes, PCT, PST) and in suppressing the expansion of distal fates (DE, CS, DL, PD). To determine more precisely when RA signaling was needed for the development of proximal segment fates, we performed a DEAB timecourse experiment (Figure 3). Blocking RA signaling from 90% epiboly (late gastrula) to the 5 somite stage caused a milder phenotype than the longer 60%- 15-somite exposure that was characterized by a loss of podocytes, reduced lengths of the PCT and PST segments, and small increases in the lengths of the distal segments (Figure 3). A slightly longer DEAB treatment, from 90% epiboly to the 10 somite stage, expanded distal segments further than the raldh2 morphants or DEAB 90% epiboly-5 somite treated embryos (Figure 3). In addition, examination of the slc20a1a and trpm7 expression patterns showed that the PST segment was ablated, while the PCT was only slightly shortened (Figure 3). Loss of PST identity was also confirmed by the absence of slc13a1 and slc22a6 transcripts, which are additional PST markers (Figure 2 and unpublished data). These findings suggest that, at least for this DEAB time window, the expansion of distal fates occurs at the expense of the PST segment. Finally, we tested whether DEAB treatment during somitogenesis would affect pronephros segmentation. DEAB treatment from 5–15 somites or 10–15 somites had no effect on the segmentation pattern of the pronephros (unpublished data), thus suggesting that an inhibition of RA signaling must be initiated prior to the 5 somite stage in order to affect nephron patterning. Taken together, our DEAB timecourse data indicate that RA signaling is required to induce proximal nephron fates and to prevent an expansion of distal fates, thereby establishing the normal pronephric segmentation pattern. Interestingly, our data suggest that different segments have different temporal requirements for RA signaling. RA is essential between 90% epiboly and the 5 somite stage to induce podocytes, between 90% epiboly and 10 somites to induce PST formation, and between 60% epiboly and 15 somites to form the PCT. The PCT segment is the most refractory to RA inhibition and is only lost following the longest DEAB treatment window (60% epiboly to 15 somites). Given these findings, we tested whether increasing the concentration of RA by exogenous treatment would promote proximal nephron fates at the expense of distal fates. We exposed wild-type zebrafish embryos to RA during similar developmental intervals used for our DEAB experiments, and then assayed segment marker expression at 24 hpf by double in situ hybridization with mhc to mark the somites. Wild-type embryos treated with 1 × 10−7 M RA between 90% epiboly and 5 somites developed a normal number of podocytes, as evidenced by wt1b expression, but displayed expanded proximal tubule domains, shown by expression of slc9a3 (marking both the PCT and PST), slc20a1a (PCT), and trpm7 (PST) (Figure 4). Conversely, the clck-expressing distal tubule domain was reduced, due to reductions in the length of the DL (marked by slc12a3) and PD (marked by gata3) (Figure 4). However, the DE segment (marked by slc12a1) was unaffected (Figure 4). The position of the stc1-expressing CS segment cells was shifted posteriorly, and the level of expression was slightly reduced (Figure 4). A longer treatment window, from 60% epiboly–15 somites resulted in a more severe ‘proximalized' phenotype with a longer expanse of proximal tubule and a greater reduction in each distal tubule domain (Figure 4). In these embryos, the position of the stc1-expressing CS population was shifted even more posteriorly, and was located at the distal edge of the yolk sac extension (Figure 4A). We next treated wild-type embryos with higher dose of RA (1 × 10−6 M) over this same 60% epiboly–15 somites time window, as well as two shorter time periods: 60% epiboly–5 somites, and 90% epiboly–5 somites. Embryos treated for any of these time windows displayed a completely ‘proximalized' phenotype with the tubule domain being comprised entirely of proximal segment identities (Figure 4 and unpublished data). In these embryos, the proximal marker slc9a3 was expressed throughout the cdh17-expressing tubule population (Figure 4 and unpublished data). Within this ‘proximal-only' pronephros, the PCT marker slc20a1a was expressed from the anterior limit of the tubule to somite 13, and the PST marker trpm7 was expressed from somite 14 to the posterior limit of the tubule, where the trpm7-expressing tubules fused at the prospective site of the cloaca (Figure 4 and unpublished data). Expression of all distal segment markers was absent, suggesting that the DE, CS, DL, and PD had failed to be specified (Figure 4 and unpublished data). These results show that exogenous RA treatment from gastrulation stages until the 5 somite stage is sufficient to ‘proximalize' the pronephros, suggesting that this time period is the critical window when RA signaling is required for proximo-distal patterning of pronephric progenitors. To further explore the notion that RA alters the patterning of renal progenitors prior to the 5 somite stage, we examined the expression patterns of the Notch ligands deltaC (dlc) and jagged2a (jag2a), and the renal transcription factors wt1a, pax2a, pax8, and evi1, as these genes are detected in the IM during early somitogenesis and have been implicated in early nephron patterning [44,62–64]. Wild-type embryos were treated with DMSO, DEAB, or 1 × 10−6 M RA from 60% epiboly until the 6 somite stage, and then IM gene expression was assayed by whole-mount double in situ hybridization together with myoD to mark the somites. Transcripts for pax2a and pax8, which label all pronephric progenitors, were found throughout the IM domain in a similar pattern in wild-type, DEAB-treated, and RA-treated embryos (Figure S5). Consistent with our previous results, wt1a expression was absent in DEAB-treated embryos, and wt1a was strongly up-regulated in RA-treated embryos (Figure S5). The expression domains of deltaC and jag2a, normally restricted to a proximal region of the IM adjacent to somites 2–5, were absent in DEAB-treated embryos, and expanded posteriorly in RA-treated embryos (Figure S5). Expression of evi1, found in the distal portion of the IM starting around somite 6, was shifted anteriorly to somite 3 following DEAB treatment and reduced to the most posterior group of IM cells following RA treatment (Figure S5). These findings demonstrate that changes in RA dosage during gastrulation to early somitogenesis are associated with gene expression changes in IM progenitors, far in advance of their mesenchymal-to-epithelial transition that creates the pronephric tubules. In addition to RA, multiple tissues along the embryonic axis are segmentally patterned by the cdx genes [25,34]. Loss of cdx gene function leads to an expansion of anterior trunk fates and axial elongation defects that result in a loss/truncation of the posterior trunk and tail [23–30,34]. In zebrafish, cdx4–/– mutant embryos display expanded wt1a expression at the 15-somite stage, suggesting that the podocyte lineage might be expanded and thus implicating cdx genes in pronephros patterning [24]. We therefore examined the formation of the pronephros in cdx4–/– and cdx1a/4-deficient (herein referred to as cdx-deficient) to assess development of the renal corpuscle, tubule, and pronephric duct. In addition, we used differential interference contrast (DIC) optics to visualize the somite boundaries and determine the size and position of each segment relative to the somites. It is important to note that the size of posterior somites in cdx4–/– embryos, and both the size and number of the somites in cdx-deficient embryos, is greatly reduced toward the posterior, due to the axial elongation defect [24,29–30]. A similar defect has also been observed in mouse Cdx mutants [34]. At 24 hpf, wild-type embryos expressed wt1a in presumptive podocytes (located adjacent to somite 3, marked by an arrow in Figure 5A) as well as a population of presumptive mesenchymal cells located next to somites 1–3 in a broad lateral domain (indicated by a bracket and ‘M' in Figure 5A). cdx4–/– embryos had expanded wt1a expression at 24 hpf that ranged from a position anterior to somite 1 to approximately somite 7, and cdx-deficient embryos showed an even more dramatic posterior expansion that reached somite 12 (Figure 5A). As the expansion of wt1a in cdx mutant embryos could indicate increased numbers of podocytes and/or the mesenchymal population, we examined wt1b expression, which specifically marks podocytes [65]. Equivalent numbers of wt1b-expressing podocytes were formed in cdx4–/– mutant and wild-type embryos, but podocytes in cdx4–/– embryos were located more posteriorly, at the level of somite 5 (Figure 5A). cdx-deficient embryos developed a normal number of wt1b-expressing podocytes, though they were arranged in a pair of somewhat irregular linear groupings (rather than forming bilateral spherical clusters), and they were also located more posteriorly, adjacent to somite 7 (Figure 5A). We conclude from these observations that the loss of cdx4 and cdx1a/4 function progressively expands wt1a expression without increasing the number of podocytes and leads to podocyte formation at more posterior locations along the embryonic axis. To assess tubule formation in cdx mutants, we examined the expression of cdh17 at 24 hpf. While cdx4–/– embryos formed complete tubules that fused at the cloaca, cdx-deficient embryos displayed cdh17 expression that was reduced and discontinuous, with the tubules failing to fuse (Figure 5A). Consistent with the posterior shifts in podocyte position, the A-P position of the tubule, marked by cdh17 transcripts, was shifted caudally in both cdx4–/– and cdx-deficient embryos. In wild-type embryos, the tubule spans the length of somites 4–18, but was located from somites 6–20 in cdx4–/– embryos, and from somites 8–13 in cdx-deficient embryos (Figure 5A). We next characterized the tubule segmentation pattern in cdx mutant embryos by examining the expression patterns of slc20a1a (PCT), trpm7 (PST), slc12a1 (DE), stc1 (CS), slc12a3 (DL), gata3 (PD), and aqp3 (cloaca). cdx4–/– embryos showed a slight expansion of the PCT, which spanned an additional two somites compared to wild-types (Figure 5A and unpublished data). The PST, DE, CS, and PD were all shorter than normal in cdx4–/– embryos, with the PD displaying the most severe reduction in length (Figure 5A and unpublished data). Transcripts for aqp3 were not detected in cdx4–/– embryos, suggesting a defect in cloaca development (Figure 5A). These results indicate that loss of cdx4 leads to a slight expansion in PCT fate with corresponding reductions in more distal fates. In contrast to cdx4–/– embryos, cdx-deficient embryos showed discontinuous slc20a1a expression and failed to express trpm7, slc12a1, stc1, slc12a3, gata3, or aqp3 (Figure 5A). These results suggest that the tubule territory in cdx-deficient embryos acquires a PCT identity, while the remaining nephron segments fail to develop. These defects were not the result of delayed development, as at later developmental stages, cdx-deficient embryos continued to possess tubules that only expressed PCT-markers, and the tubules never fused caudally (unpublished data). In addition, expression analysis of the podocyte marker mafB in cdx-deficient embryos at 48 hpf revealed that podocytes fail to fuse into a single renal corpuscle. Instead, the podocytes formed bilateral corpuscles that were dilated compared to wild-type and cdx4–/– embryos, presumably due to fluid accumulation (Figure 5B). Consistent with this, cdx-deficient embryos had developed glomerular cysts by 72 hpf, as well as severe pericardial edema, indicative of renal failure (Figure 5C). In contrast, cdx4–/– embryos never exhibited glomerular cyst formation or edema, suggesting that although various segment lengths were shortened, these embryos were able to maintain adequate kidney function (Figure 5C). In summary, our expression analyses show that cdx deficiency causes a posterior shift in the location of the pronephros along the embryonic axis (Figure 5D). While the podocytes and PCT populations formed relatively normally, the PST and distal tubule segments were reduced or absent in cdx4–/– and cdx-deficient embryos, respectively (Figure 5D). Thus the cdx genes, acting either directly or indirectly, are required for the formation of the distal nephron segments and establishing the normal segmentation pattern of the pronephros. Given the largely opposite effects of cdx-deficiency and loss of RA signaling on nephron patterning, as well as the recent report that cdx genes control how the hindbrain responds to RA during its patterning [30,31], we wondered if an interplay between these pathways was operative during pronephros segmentation. The location and level of RA within tissues is dependent on the expression of raldh-synthesizing enzymes and cyp26-degrading enzymes [15,54]. We therefore investigated the expression of these genes in cdx-deficient embryos during early somitogenesis, as our previous experiments suggested that the IM is being influenced by RA signaling at this time. At the 5 somite stage, expression of raldh2 in the paraxial mesoderm was expanded posteriorly in cdx4–/– embryos compared to wild-types (Figure 6A). An even greater posterior expansion was seen in cdx-deficient embryos, with raldh2 transcripts being detected throughout the entire unsegmented paraxial mesoderm and tailbud region (Figure 6A). Expression of the RA-catabolizing enzyme cyp26a1 in the upper trunk region was also expanded posteriorly in cdx4–/– embryos at the 5 somite stage, and more extensively expanded in cdx-deficient embryos (Figure 6A). To visualize how the combined changes in raldh2 and cyp26a1 expression altered the source of RA along the trunk, we generated digital overlays of these expression patterns. This analysis suggested that the anterior boundary of RA production (i. e. , the junction of the cyp26a1 and raldh2 expression domains) was located more posteriorly in cdx4–/– mutants compared to wild-types, and that this posterior shift was more pronounced in cdx-deficient embryos (arrows in Figure 6A). To better quantitate these posterior shifts, we examined expression of raldh2 and cyp26a1 at the 10 somite stage when the somites could be visualized by staining for myoD transcripts. We found that the raldh2 expression boundary in cdx4–/– and cdx-deficient embryos was shifted posteriorly by 1 and 2 somites, respectively, compared to wild-type embryos (Figure 6B). The cyp26a1 domain in wild-type embryos occupied the region of somites 2–3, just rostral and slightly overlapping with raldh2, as shown by double in situ hybridization (Figure 6B). In cdx4–/– embryos, cyp26a1 transcripts were detected in the region of somites 3–5, whereas in cdx-deficient embryos they extended from somites 3–7 (Figure 6B). These analyses reveal a striking correlation between the presumptive source of RA at the 10 somite stage and the axial position of the pronephros at 24 hpf in each genotype (i. e. , the source of RA and the position of the pronephros both start at somites 3,5, and 7 in wild-type, cdx4–/–, and cdx-deficient embryos, respectively). The combination of these data, together with the results from our DEAB experiments, suggest a model in which the cdx genes act upstream of raldh2 and cyp26a1 to localize the source of RA along the A-P axis, and that RA, in turn, acts on the IM to induce the proximal segments and prevent an expansion of the distal segment fates. If our model is correct, then inhibiting RA synthesis in cdx-deficient embryos should rescue pronephric positioning and the formation of the distal tubule segments. To test this, we treated cdx4–/– and cdx-deficient embryos with DEAB from 90% epiboly to the 5-somite stage and examined pronephros segmentation. In support of our model, we found that cdx4–/– and cdx-deficient embryos exhibited a one-somite anterior shift in the position of the pronephros, as shown by expression of cdh17 (Figure 7). In addition, the development of podocytes was abrogated, and the length of the PCT was reduced in DEAB-treated cdx4–/– and cdx-deficient embryos (Figure 7), consistent with our findings in wild-type embryos (Figure 3). We observed that the DE segment was increased in DEAB-treated cdx4–/– embryos, as shown by an expansion of the slc12a1 expression domain (Figure 7). DEAB treatment also increased the number of CS cells in cdx4–/– mutants, shown by the expression of stc1, and increased the DL segment length, evidenced by expansion of the slc12a3, romk2, and clck expression domains (Figure 7 and unpublished data). In cdx-deficient embryos treated with DEAB, formation of the DE, CS, and DL segments was rescued, shown by expression of slc12a1, stc1, and slc12a3, respectively (Figure 7). A similar rescue of the expression of the DE and DL markers romk2 and clck was also observed (unpublished data). These findings demonstrate that cdx gene function is not necessary to specify distal segment identity directly, but instead suggests that the abrogation of distal segment formation in cdx-deficient mutants is related to the level of RA that the renal progenitors are exposed to. Taken together with the above results, these finding provide good evidence that the pronephric positioning defect and failure to form the distal tubule identities in cdx-deficient embryos is caused by mis-localization of the RA source along the A-P axis.
Retinoid signaling plays essential roles in the A-P patterning of a number of diverse tissues in the embryo. During early development, a source of RA in the upper trunk (cervical) region is produced by the action of the RA synthetic enzyme, Raldh2, which is expressed in the anterior paraxial mesoderm [54]. The coordinate expression of RA-catabolizing Cyp26 enzymes in surrounding tissues creates a so-called ‘sink' for this RA source [54]. Collectively, these enzymes are thought to create a gradient of RA activity that diffuses into surrounding tissues [34,54]. The functions of this RA source have been extensively studied in the developing hindbrain, where the effects of graded RA signaling are thought to create nested expression domains of RA-responsive genes that drive A-P segmentation of the hindbrain into a series of rhombomeres [15,54]. In addition to regionalizing the overlying neurectoderm, RA produced in the upper trunk paraxial mesoderm has been implicated in the regionalization of the underlying endoderm. Studies in zebrafish have shown that RA acts directly on the endoderm to specify hepatopancreatic progenitors that give rise to the liver and pancreas [17,18]. RA also influences mesodermal cell fate decisions during zebrafish development, including the formation of the pectoral fin field—which arises from the lateral plate mesoderm adjacent to the upper trunk somites [56,57,66]—and the heart [67]. In the latter case, inhibition of RA synthesis leads to an expansion of precardiac mesoderm, resulting in an excessive number of myocardial progenitors [67]. These findings indicate that, in addition to acting as an inducer of cell fates such as in the hindbrain and endoderm, RA also plays an important role in restricting certain cell fates. Our study now adds the IM as another mesodermal derivative that is patterned by RA. Our results show that RA production is essential during gastrulation and early somitogenesis for the induction of proximal nephron fates as well as to restrict the expansion of distal nephron fates. Over this period of development, RA is produced by the anterior paraxial mesoderm (PM). The IM, which gives rise to the pronephros, is located lateral to the PM (Figure 8). Given the role of RA as a diffusible morphogen in other tissues, we hypothesize that RA diffuses from the PM and establishes a gradient along the IM, with high levels of RA inducing proximal fates and low RA levels being permissive for distal fates (Figure 8). Our time-course experiments with DEAB support this view, with the most severe reduction in proximal fates (and concomitant expansion in distal segments) corresponding to the longest treatment window. However, further work is needed to determine the nature of the RA gradient, as well as how dynamic fluctuations in retinoid availability [54] affect the dose and length of time that the renal progenitors are exposed to RA. A gradient-free model has recently been proposed for RA-dependent hindbrain patterning, based on the finding that sequential expression domains of the cyp26a1, cyp26b1, and cyp26c1 genes are essential for rhombomere boundary establishment [68,69]. It is unclear if a similar mechanism might operate during pronephros segmentation, as cyp26b1 and cyp26c1 do not show a nested pattern of expression in the IM. Overall, the effects of RA on the patterning of the IM can be regarded as ‘anteriorizing.' Our finding that exogenous RA treatment induces proximal tubule fates to form throughout the pronephros supports this conclusion. Classically, RA is known as a ‘posteriorizing' factor due to its effects on the central nervous system, where an inhibition of RA signaling causes an expansion of anterior neural fates in the hindbrain [15,54]. Thus we conclude that RA can actually have both anteriorizing and posteriorizing activities, depending on the tissue in question. A unified way to characterize these effects would be to consider the upper trunk RA source as an organizing center, akin to the dorsal organizer in the gastrula, that locally patterns cell types in all three germ layers. Previous studies have implicated RA as a regulator of renal development. Animal cap experiments in Xenopus showed that RA, together with Activin, is sufficient to induce the formation of pronephric tubules [70]. A more recent study in Xenopus reported that overexpressing various RA antagonists results in a complete loss of the pronephros (glomus, tubules, and duct) [55]. This phenotype is more severe than what we observed and may reflect differences in the efficacy of DEAB to completely block RA production compared with other RA antagonists. The pronephric tubules in Xenopus are segmented into proximal and distal segments [48], similar to the zebrafish pronephros, however a role for RA in Xenopus nephron segmentation has not been reported. In mammals, it has long been known that vitamin A deficiency causes severe renal malformations [71]. Targeted mutagenesis of the RAR genes in mouse, followed by elegant rescue experiments, established an important role for RA as a dose-dependent inducer of the GDNF receptor Ret [72,73]. GDNF is an essential regulator of ureteric bud branching morphogenesis, and loss of GDNF signaling results most frequently in renal agenesis [74]. Because ureteric bud branching is an essential prerequisite for nephrogenesis, RA serves a key role in stimulating nephron formation. At present it is not known whether RA is also involved in the proximodistal patterning of metanephric nephrons. Interestingly, Raldh2 transcripts are found in podocyte progenitors, whereas Cyp26a1 is expressed by the tubule anlagen during metanephros development, suggestive of a role for RA in mammalian nephron patterning [75]. Transplantation studies in frogs suggest that RA may act directly on pronephric precursors [55]. However, the downstream targets of RA in the IM are not known. In the hindbrain, the presumptive RA gradient is thought to regulate rhombomere segmentation by activating the expression of the anterior, 3' Hox genes, i. e. , those comprising the 1st through 5th paralog groups [15,54,76]. In zebrafish, transcripts for hoxb1a, hoxb1b, and hoxb5a are found in proximal portions of the IM, thus making them potential candidates for mediating the effects of RA during pronephros segmentation [34]. Future studies using single, double, and triple morpholino injections can test the importance of these hox genes for renal development. The effects of RA on pronephric segmentation may also be coordinated by the action of non-Hox pathways. Our results suggest the intriguing possibility that RA signaling targets may include renal transcription factors as well as members of the Notch signaling pathway. The gene evi1 encodes a zinc-finger transcription factor that has been implicated in patterning distal regions of the pronephros in Xenopus, and overexpression of evi1 was found to inhibit proximal segment formation [63]. These data are consistent with our results showing that expression of evi1 in the IM was expanded following DEAB treatment, and reduced following exposure to exogenous RA. Another renal transcription factor candidate is the odd-skipped related transcription factor 1 (osr1) encoding a zinc-finger repressor. Recent studies in Xenopus and zebrafish have shown that osr1 is expressed in the ventral mesoderm during gastrulation and later in an anterior domain of the IM [77]. Morpholino-mediated knock-down of osr1 leads to defects in the formation of podocytes and proximal tubule progenitors [77], consistent with Osr1 participating in a common pathway with RA. The Notch pathway may also interact with RA during nephron segmentation. Conditional knockout of Notch2 in the mouse metanephros results in a loss of podocytes and proximal tubule fates, whereas distal markers are relatively unaffected [78]. In zebrafish, Notch signaling has been shown to regulate the differentiation of multiciliated cells and principle cells in the pronephric tubules [44,45]. However, a role for Notch signaling in the formation of proximal nephron fates is also suggested by the expression pattern of the Notch ligands deltaC, jag1b, and jag2a, which are restricted to proximal portions of the intermediate mesoderm [62,64]. Simultaneous knockdown of jag1b/2a results in an abnormally small renal corpuscle and dysmorphic proximal tubules, consistent with a conserved role for Notch signaling in proximal nephron development [79]. Our finding that DEAB treatment abrogates deltaC and jag2a expression in the proximal IM, while exogenous RA expands their expression, supports a role for RA acting upstream of the Notch pathway. Our study provides evidence that cdx genes control the expression domains of raldh2 and cyp26a1 along the embryonic axis. The boundaries of both raldh2 and cyp26a1 are progressively shifted toward the posterior in cdx4 and cdx1a/4-deficient embryos, suggesting that the upper trunk source of RA is posteriorly shifted. We hypothesize that this posterior shift in RA production results in a posterior shift in the position of the pronephros (Figure 8). We propose that this effect, combined with the axial elongation defects, leads to reduced or absent distal segment fates. The ability to rescue distal segments by treating cdx mutants with a pulse of DEAB is consistent with this model, and also demonstrates that cdx function is not requisite for the induction of distal fates from the intermediate mesoderm. Thus additional, as yet unidentified pathways, are responsible for directing distal fates. While our data supports the notion that Cdx factors exert their effects by the regulation of RA signaling, it does not rule out the possibility that Cdx factors may also function to repress proximal fates independent of RA signaling. Given the mounting evidence that the upper trunk RA source is an important organizing center, we would predict that both the patterning and positioning of numerous organs would be affected in cdx mutants. Consistent with this, defects in several mesodermal fates that arise in the anterior trunk region have been observed in cdx-deficient embryos. Vascular precursors are progressively expanded when cdx activity is abrogated, and blood precursors are both reduced and shifted posteriorly in cdx mutants [24,29]. In addition to mesodermal defects, cdx mutants also display patterning defects in the neurectoderm that gives rise to the anterior spinal cord [30,31]. We hypothesize that many, if not all, of these defects in cdx mutants are caused by the abnormal localization of RA along the A-P axis. The loss of cdx gene function in both zebrafish and murine models has been shown to cause global shifts in hox gene expression in the mesoderm and neurectoderm [24,25,29–30,34]. Given the rostral shifts and expansions of both raldh2 and cyp26a1 expression observed in cdx4 and cdx1a/4-deficient embryos, Hox transcription factors are attractive molecules for regulating raldh2 and cyp26a1 expression. Defects in blood formation in cdx4-null zebrafish can be rescued by the overexpression of several hox genes [24], and the overexpression of hoxa9a also results in a partial rescue of the axis elongation defect in cdx4–/– embryos [29]. Future studies are needed to examine whether hox gene overexpression (s) can rescue pronephros positioning and formation of distal segments in cdx mutant embryos. In conclusion, our studies have revealed an important link between the cdx genes and localization of RA, and provide evidence that RA signaling is a central determinant of pronephros A-P segmentation. Our results establish the zebrafish embryo as a simplified model of vertebrate nephron segmentation that will further our understanding of mammalian nephron segmentation, and provide insights into the causes of kidney birth defects and renal disease in humans.
Zebrafish were maintained and staged as described [80,81]. Tübingen strain wild-type embryos were used for all experiments. DEAB and all-trans retinoic acid (Sigma-Aldrich) were dissolved in 100% dimethyl sulfoxide (DMSO) to make a 1 M stock and aliquots were stored at −80°C. For DEAB and RA treatments: embryos were incubated in 1. 6 × 10−5 M DEAB/DMSO in E3 embryo media, 1 × 10−6 M or 1 × 10−7 M RA/DMSO in E3 embryo media, or 1. 6 × 10−5 M DMSO (control) in E3 in the dark over particular developmental intervals, then washed five times with E3 and then fixed at 24 or 48 hpf. These experimental treatments were fully penetrant and produced consistent results at the doses and treatment windows that were examined. raldh2 morpholino (CAACTTCACTGGAGGTCATCGCGTC) was injected into 1-cell wild-type embryos. Incrosses of kggtv205 heterozygous adults (maintained on the Tübingen strain) were used to obtain cdx4–/– embryos and were injected at the 1-cell stage with cdx1a morpholino (CAGCAGATAGCTCACGGACATTTTC) as described [29] to obtain cdx-deficient embryos. Both raldh2 and cdx1a morpholinos produced fully penetrant effects. Embryos were raised to appropriate stages and fixed in 4% paraformaldehyde (PFA) /1×PBST for gene expression analysis. For all reported gene expressions, at least 20 embryos were examined. Whole-mount in situ hybridization of zebrafish embryos was performed as previously described [24]. The expression patterns of cdh17, clck, cyp26a1, evi1, gata3, mhc, myoD, nbc1, pax2a, pdzk1, raldh2, ret1, sall1, sglt1, slc4a2, slc20a1a, wt1a, and wt1b were previously reported [14,24,29,37,55,56,61,82–86]. For antisense probe production, we used the following IMAGE clone template plasmids, restriction enzymes for DNA linearization, and RNA enzymes: mafb: 7995399, pExpress-1, EcoR1, T7; rfx2, pBK-CMV, template was PCR amplified using primers GTGAATTGTAATACGACTCACTATAGGG and TTAACCCTCACTAAAGGGAACAAA, T7; slc9a3: 6996791, pExpress-1, EcoRI, T7; slc26a2: 4760214, pBK-CMV, EcoRI, T7; slc13a1: 6793065, EcoRV, t7; slc13a3: 4744276, pCMV-sport6. 1ccdb, EcoRI, T7; slc22a6: 4744276, pBK-CMV, SalI, T7; slc12a1: pBK-CMV, EcoRI, T7; slc12a3: 7037010, pExpress-1, EcoRI, T7; stc1 was amplified from 24 hpf embryo cDNA using primers ATGCTCCTGAAAAGCGGATTT and TTAAGGACTTCCCACGATGGA and cloned into pGemTEasy, NcoI, Sp6. Gene-specific primers spanning 700–1,000 bp of the coding sequence were used to amplify DNA fragments from E15. 5/P0 kidney cDNA pools, and the PCR products of the right size were cloned into the pCRII-Topo vector (primer sequences available upon request). DNA templates for riboprobe production were generated by PCR with T7 and Sp6 Ready Made primers (Integrated DNA Technologies) from PCRII-TOPO clones or T7 and T3 Ready Made primers from Bmap library clones. Digoxigenin-labeled anti-sense riboprobes were synthesized from the PCR product and purified with Micro Bio-spin columns P-30 Tris RNase-free (Bio-Rad). Probes were diluted with prehybridization buffer (50% formamide, 5×SSC, pH4. 5,50 μg/ml yeast tRNA, 1% SDS, 50 μg/ml heparin) to 10 μg/ml and stored at −80 °C. Neonatal kidneys were dissected free of surrounding tissues except the ureter and fixed with 4% PFA at 4 °C for 24 h. After PBS washes, they were incubated with 30% sucrose at 4 °C overnight. Kidneys were swirled in five dishes of OCT to remove sucrose and mounted in OCT in a dry ice/ethanol bath. The OCT blocks were stored at −80 °C. Sections were cut at 20 μm and air dried. Sections were post-fixed with 4% PFA for 10 min, treated with 10 μg/ml proteinase K for 10 min and post-fixed for 5 min. Slides were acetylated (1. 33% Triethanolamine, 0. 065% HCl, 0. 375% acetic anhydride) for 10 min and dehydrated with 70% ethanol and 95% ethanol for 5 min each. Slides were air dried then incubated with 500 ng/ml digoxigenin-labeled riboprobes at 68 °C overnight. Hybridized sections were washed with 50% formamide, 1×SSC, pH4. 5 for 30 min at 65 °C, treated with 2 μg/ml RNase for 15 min at 37 °C, and washed with 2×SSC, pH4. 5 for 30 min, and twice with 0. 2×SSC, pH4. 5 for 30 min at 65 °C. Slides were washed three times at room temperature with 1×MBST (0. 1 M maleic acid, 0. 15 M NaCl, 0. 1% Tween-20, pH7. 5) for 5 min each, and incubated with blocking solution (2% Boehringer Mannheim (BM) blocking reagent) in 1×MBST, 20% heat-inactivated sheep serum) for 1 h. After incubation with anti-digoxigenin antibody-AP (Roche, 1: 4000) at 4 °C overnight, sections were washed with 1×MBST at room temperature, 5 min for three times, then with NTMT (0. 1 M NaCl, 0. 1 M Tris-HCl, pH9. 5,50 mM Mg2Cl, 0. 1% Tween-20,2 mM Levimasole) for 10 min, and developed with BM purple (Roche). Color reactions were stopped with fixatives (4% PFA, 0. 2% glutaraldehyde) and sections mounted with glycergel mounting media (DAKO). Images were captured with a Nikon DXM1200 digital camera attached to a Leitz DMRB microscope. | In the kidney, structures known as nephrons are responsible for collecting metabolic waste. Nephrons are composed of a blood filter (glomerulus) followed by a series of specialized tubule regions, or segments, which recover solutes such as salts, and finally terminate with a collecting duct. The genetic mechanisms that establish nephron segmentation in mammals have been a challenge to study because of the kidney' s complex organogenesis. The zebrafish embryonic kidney (pronephros) contains two nephrons, previously thought to consist of a glomerulus, short tubule, and long stretch of duct. In this study, we have redefined the anatomy of the zebrafish pronephros and shown that the duct is actually subdivided into distinct tubule segments that are analogous to the proximal and distal segments found in mammalian nephrons. Next, we used the zebrafish pronephros to investigate how nephron segmentation occurs. We found that retinoic acid (RA) induces proximal pronephros segments and represses distal segment fates. Further, we found that the caudal (cdx) transcription factors direct the anteroposterior location of pronephric progenitors by regulating the site of RA production. Taken together, these results reveal that a cdx-RA pathway plays a key role in both establishing where the pronephros forms along the embryonic axis as well as its segmentation pattern. | Abstract
Introduction
Results
Discussion
Materials and Methods | developmental biology
danio (zebrafish)
vertebrates
teleost fishes
nephrology | 2007 | The cdx Genes and Retinoic Acid Control the Positioning and Segmentation of the Zebrafish Pronephros | 15,617 | 370 |
White-nose syndrome is one of the most lethal wildlife diseases, killing over 5 million North American bats since it was first reported in 2006. The causal agent of the disease is a psychrophilic filamentous fungus, Pseudogymnoascus destructans. The fungus is widely distributed in North America and Europe and has recently been found in some parts of Asia, but interestingly, no mass mortality is observed in European or Asian bats. Here we report a novel double-stranded RNA virus found in North American isolates of the fungus and show that the virus can be used as a tool to study the epidemiology of White-nose syndrome. The virus, termed Pseudogymnoascus destructans partitivirus-pa, contains 2 genomic segments, dsRNA 1 and dsRNA 2 of 1. 76 kbp and 1. 59 kbp respectively, each possessing a single open reading frame, and forms isometric particles approximately 30 nm in diameter, characteristic of the genus Gammapartitivirus in the family Partitiviridae. Phylogenetic analysis revealed that the virus is closely related to Penicillium stoloniferum virus S. We were able to cure P. destructans of the virus by treating fungal cultures with polyethylene glycol. Examination of 62 isolates of P. destructans including 35 from United States, 10 from Canada and 17 from Europe showed virus infection only in North American isolates of the fungus. Bayesian phylogenetic analysis using nucleotide sequences of the viral coat protein geographically clustered North American isolates indicating fungal spread followed by local adaptation of P. destructans in different regions of the United States and Canada. This is the first demonstration that a mycovirus potentially can be used to study fungal disease epidemiology.
Pseudogymnoascus destructans (Pd; previously named Geomyces destructans) is an emerging fungal pathogen responsible for a fatal disease, white-nose syndrome (WNS) in hibernating bats in North America [1–3]. Experts estimate over 5 millions bats died from WNS in North America since the disease was first noted in New York in 2006 [4–6]. Currently WNS has spread to at least 29 states in the United States (plus three additional states where Pd presence has been confirmed, but not WNS) and five provinces in Canada [4]. The fungus is widely distributed in Europe [6,7] and recently has been reported from the northeast of China and Siberia [8,9], but no mass mortality has been reported in European or Asian bats [6,8]. Pd’s lethal effect on North American bats coupled with its clonal genotype in North American isolates [10,11], its single mating type [12] and the absence of close relatives [13] led many researchers to hypothesize a recent introduction to North America [1,6, 14,15]. Pd is confirmed in seven North American [1,4] and 13 European species of bats [4,9]. The natural history of the genus Pseudogymnoascus and its allies indicate that they are commonly isolated from soils in colder regions of the world [16]. Unlike Pd many of its close relatives are cellulolytic saprobes and non-pathogenic [16,17]. Mycoviruses associated with fungi have drawn interest because of their potential roles in fungal biology and pathogenicity [18]. Mycoviruses are very frequent in fungi and generally maintain a persistent lifestyle [19]. Horizontal transmission is very rare, and is likely restricted to closely related strains, although phylogenetic studies indicate transmission among species has occurred [20]. Transmission has only been documented in a few cases outside the laboratory [21]. Most mycoviruses are cryptic with no known biological effects on their fungal hosts, although there is a lack of research in this area. However, there are significant examples where mycoviruses play important roles in fungal biology and ecology [22]. Here we used mycoviruses of Pd as a tool to study the epidemiology of WNS. We investigated mycoviruses in Pd and show that population variation of a Pd-mycovirus can be useful in tracing the spread of WNS.
We examined 62 isolates of Pd from North American and European bats for mycoviruses (Table 1). The isolates were cultured from four North American and one European species of bats and were collected from 2008 to 2015. North American isolates included 14 from Pennsylvania, seven from New York, six from West Virginia, three from North Carolina, three from Vermont, one from Ohio, one from Indiana and 10 from New Brunswick, Canada. We screened 16 isolates of Pd from the Czech Republic and one isolate from Slovakia in Europe. Double-stranded RNA (dsRNA) extracted from all North American isolates showed two bands—a larger band close to 1. 8 kb (RNA 1) and a smaller band close to 1. 6 kb (RNA 2) in electrophoretic analysis (Fig 1A). None of the European isolates contained these dsRNAs, although two, CCF-4127 and CCF-4128, had dsRNAs profiles different from that of the North American isolates (Fig 1B). We found no dsRNAs of viral origin in five isolates of Geomyces sp. from Antarctic soil or in six isolates of Pseudogymnoascus sp. from cave soils in Pennsylvania (S1 Table). The dsRNA enrichment method is based on the premise that uninfected plants or fungi normally do not contain detectable amounts of high molecular weight dsRNA, and, when present, dsRNA is an indicator of a viral genome [23]. Sanger sequencing of cDNA clones from RNAs 1 and 2 of the North American isolates of Pd obtained from random primed RT-PCR provided nearly complete genomic sequences; ends were determined by 5' - primer ligated RNA ligase mediated-rapid amplification of cDNA ends (RLM-RACE) [24] providing consensus genomic sequences for RNAs 1 and 2 of 1761 bp and 1590 bp. Northern-blots using cDNA clones from RNA 1 or RNA 2 as probes confirmed the identity of the dsRNA bands (Fig 1C). We named this new virus Pseudogymnoascus destructans partitivirus-pa (PdPV-pa; the pa indicates the sequenced isolate is from Pennsylvania). A BLASTx search of GenBank showed closest similarity of RNA 1 of PdPV-pa with RNA 1 of Penicillium stoloniferum virus S (PsV-S), with 76% amino acid (aa) identity. Similarly, RNA 2 of Pd showed closest similarity with the RNA 2 of PsV-S with 67% aa identity. PsV-S is the type species of the genus Gammapartitivirus in the family Partitiviridae [25]. Sequence analysis of RNA 1 of PdPV-pa predicted a single open reading frame (ORF) of 540 aa (60 kDa) that codes for a putative RNA-dependent RNA polymerase (RdRp) (Fig 2A). RNA 2 also contained a single ORF of 470 aa (52 kDa) that codes for a putative coat protein (CP) (Fig 2B). Amino acid level sequence identity of PdPV-pa RdRp and CP with the approved members of genus Gammapartitivirus in the family Partitiviridae ranges from 58% - 76% and 36% - 67% respectively, which are within the species demarcation criteria (RdRp ≤ 90%; CP ≤ 80%) of the genus [42]. Further, the 5' termini of PdPV-pa RNAs 1 and 2 coding strand share a conserved CGCAAAA… sequence, where G is followed by A, U, or C but not G in the next 5 to 6 nucleotide positions, characteristic of the genus Gammmapartitivirus [25] (Fig 2C). Similarly, the 3' terminal 50 nucleotides of RNAs 1 and 2 were adenosine (A) rich in the range (7–24 nt) typical of members of the Gammapartitivirus genus [25] (Fig 2D). PdPV-pa particles were purified from mycelia of Pd and negative-stain transmission electron microscopy showed isometric particles of approximately 30 nm diameter, characteristic of members of the Partitiviridae (Fig 3A). PdPV-pa dsRNAs were also extracted from the purified virus particles to reconfirm their presence as genomic RNAs (Fig 3B). Bayesian trees constructed using aa sequences from the RdRp and CP of PdPV-pa clustered PdPV-pa with other members of genus Gammapartitivirus in the Partitiviridae family (Fig 4A & 4B). In both RdRp and CP trees, PdPV-pa appeared as a sister branch to PsV-S with strong posterior probability support of 92% and 100% respectively suggesting PdPV-pa is evolutionary close to PsV-S. The genome structure of PdPV-pa, conserved features in its RNAs explained above, its particle morphology, its RdRp and CP amino acid sequence identity within species demarcation criteria, and phylogenetic analyses all confirmed that PdPV-pa is a novel partitivirus belonging to genus Gammapartitivirus in the family Partitiviridae. We attempted several methods including single spore isolation, hyphal tip culture, protoplast culture, heat therapy and nutritional and chemical stress that involved application of the antiviral drugs cycloheximide or ribavirin, to cure Pd of the PdPV-pa infection. However, only partial success was achieved with high concentrations of cycloheximide (25 μg/ml) and ribavirin (300 μM) treatments after three passages. PdPV-pa remained suppressed in the fungus treated with cycloheximide or ribavirin when grown in media with the drug but once the fungus was transferred to drug-free media the virus reappeared. Finally, our attempt to cure the fungus using polyethylene glycol (PEG) -induced matric potential in minimal nutrition media made PdPV-pa undetectable. PdPV-pa infection in Pd was checked under matric potential gradients starting from -2MPa, -3MPa to -4MPa. We did not observe visible germination of Pd conidia or mycelia growth at -5MPa and -6MPa. PdPV-pa was undetected in PEG treated Pd isolates when evaluated by dsRNA extraction and RT-PCR with RdRp specific primers for PdPV-pa (Fig 5A & 5B). The detection limit of PdPV-pa in Pd was determined to be approximately 380 copies per cell (S1 Appendix). We enriched the viral dsRNA from total nucleic acid extracted from a defined number of Pd conidia followed by measurement of dsRNA concentration, and serial dilutions to determine the end-point of detection. Pd isolates where PdPV-pa was undetected after PEG treatment lost the characteristic gray pigmentation of wild type Pd and appeared white (Fig 6A). The virus-free isolate also produced significantly less conidia in comparison to wild type isolate (Fig 6B). Although PEG treatments were successful in obtaining a PdPV-pa free isolate of Pd, PdPV-pa tolerance to many other stresses mentioned above indicate that PdPV-pa is tightly associated with the Pd isolates from North America. Genetic variability of the RdRp and CP regions was analyzed in 45 North American isolates of PdPV-pa by amplification using specific primers followed by sequence analysis (Fig 7A & 7B). Using a 930 bp region of RdRp amplicons after trimming and alignment, we found the average percentage identity ranged from 99. 7 to 99. 9 among the 45 isolates. The high level of conservation in the RdRp is also reflected by a total of only 15 segregating sites, including seven singletons among the isolates examined. For the CP, nucleotide variability was higher: in a 1088 bp of amplicon of the CP, the average percent identity ranged from 96. 8 to 98. 4 and included 127 segregating sites out of which 69 were singletons. The Bayesian tree based on the RdRp nucleotide sequences of 45 North American isolates of PdPV-pa produced a largely unresolved tree with no clusters with significant support. However, the Bayesian tree constructed from the nucleotide sequences of the CP clustered the 45 PdPV-pa isolates into two major clades based on their geographical distribution (Fig 8). One clade was comprised of Canadian isolates; the other clade included isolates from the USA, although the posterior probability of this separation was lower than for other branching in the tree. The USA clade further included well supported clusters of isolates from New York, Pennsylvania, West Virginia, North Carolina, Vermont, Indiana and Ohio. Indiana and Ohio had one isolate each and separated as sister branches. The separate topologies of USA and Canadian clusters indicate independent diversification of Pd isolates subsequent to movement to particular regions. Within each major clade there were examples of sub-branching topologies representing isolates based on their local distribution although the pattern was not consistent throughout. The phylogeny of the PdPV-pa isolates showed no structure based on the taxonomy of the bats indicating that Pd is a generalist pathogen that is transmitted readily across bat species.
In this study, we isolated and characterized a novel virus, PdPV-pa, from the pathogenic filamentous fungus causing WNS in North American bats. Based on the nucleotide sequence, sequence properties at the 5' and 3' termini, genome organization, morphology of the virus particle and phylogenetic analysis, PdPV-pa was confirmed as a new member of the genus Gammapartitivirus, family Partitiviridae. PdPV-pa shows closest similarity with PsV-S within Gammapartitivirus. The branch supports of over 90% in posterior probability in the RdRp and 100% in the CP Bayesian trees separating PdPV-pa from PsV-S (Fig 4A & 4B) and Gammapartitivirus species delimitation criteria (≤ 90% aa-sequence identity in RdRp and/or ≤ 80% aa-sequence identity in CP [26]) confirmed PdPV' s taxonomic placement into a distinct species [25]. The occurrence of PdPV-pa infection in Pd isolates from diverse geographical locations and time suggests PdPV-pa is widely spread in North America. We could not rule out the possibility of PdPV-pa incidence in Europe considering the sample size of 17 isolates that we examined in this study. Previously, Warneke et al. [14] reported a Pd isolate from Germany (MmyotGER2) showing similar mortality effects to North American isolates when inoculated onto North American little brown bat (M. lucifugus) under experimental conditions. Unfortunately, we were not able to obtain the German isolate to evaluate the presence of PdPV-pa. However the close association of PdPV-pa in a diverse subset of the North American population of Pd sampled (35 isolates from 7 states) may provide some indications of the roles of PdPV-pa in WNS. Many mycoviruses have been reported to elicit phenotypic changes, including both hypovirulence and hypervirulence in their fungal hosts [18]. For example, the presence of Helminthosporium victoriae 145S virus (chrysovirus) in the plant pathogenic fungus, Helminthosporium victoria increased virulence in oat plants. The viral dsRNAs up-regulated Hv-p68, an alcohol oxidase/RNA-binding protein in the fungus that is likely responsible for the disease development [27]. Similarly, a high level of virulence was reported in the presence of a six kbp mycoviral dsRNA in Nectria radicicola, the causal fungus of ginger root rot [28]. The opportunistic fungal pathogen, Aspergillus fumigatus causing lung disease in immunocompromised humans and animals also exhibited hypervirulence in the presence of the uncharacterized A78 mycovirus [29]. We have not explored the roles of PdPV-pa in WNS in the present study, but some indirect evidence, including the difficulties in curing the fungus of PdPV-pa, the stability of the virus after numerous generations of laboratory cultures, the changes in pigmentation and the significantly reduced production of conidia in the virus-free isolate indicate close biological relationships between the fungus and the virus; hence future investigation on potential biological effects of PdPV-pa will be important. In our attempts to cure PdPV-pa, PEG-induced stress on the matric potential was found effective. PEG being non-toxic and metabolically inert to fungi is an ideal compound to manipulate matric-induced water stress in media [30]. Matric potential influences water availability of substrates through capillary actions and particle adsorptive forces [31]. Raudabaugh & Miller [32] showed that Pd is sensitive to matric induced water stress beyond -5MPa, which is consistent with our results. In addition to the Pd growth response, normal growth at lower matric stress and significant growth inhibition as negative values of matric potential increases are characteristic of most soil fungi [32,33]. It is possible that Pd may have originated as a soil fungus and the adaptive pressure due to competition expanded its niche. The capacity of a human pathogenic fungus, Cryptococccus neoformans, to infect several animals including cats, dogs, dolphins, sheep and many birds was explained based on the environmental selective pressures imposed on it while surviving in its primary niche: soil [34]. The recent findings that Pd is capable of surviving on various substrates like harvestmen, fungus gnats, moss, and cave soils in addition to bat skin [32,35,36], support this argument. Whether or not Pd susceptibility to matric stress is related to its origin, the inhibitory effect of the matric stress on both Pd and PdPV-pa confirms parallel biological response of both the virus and the fungus. The genetic variation in the RdRp (<1%) and the CP (2–3%) of North American populations of PdPV-pa seems low, but in fact is quite high for partitiviruses. In studies with plant partitiviruses we find less than 1% divergence after extended periods of evolution (MR, personal observation). This higher level of variation implies a recent introduction of PdPV-pa. According to our results, only one species of this virus appears to occur in the North American isolates of Pd. The phylogenetic analysis based on a Bayesian algorithm of CP nucleotide sequences showed geographical clustering of 45 North American isolates into two main clades: USA and Canada. This indicates the diversification of PdPV-pa isolates is the outcome of geographical separation followed by sequence variation. No bat host specialization was observed. This finding is consistence with the clonal populations of Pd reported previously [10,11] with only one mating type [12] despite its infection in several species of bats in North America. The phylogenetic signatures of PdPV-pa isolates relating to geography provide valuable insights on the spread of WNS. The phylogeny supports two major clusters and many sub-clusters corresponding to US States of PdPV-pa isolation, suggesting connections among North American isolates, which is valuable in tracing WNS. Additionally, clustering of Pd isolates based on location was observed in several occasions within the USA clades followed by divergence, most likely for local adaptation. This analysis can be successfully expanded incorporating CP sequences of PdPV-pa from wider geographical locations to study the spread of WNS.
Pseudogymnoascus destructans (Pd) was isolated from diseased bat wing tissue, live bat wing punches (2-5mm diameter) or wing swabs, cultured on 0. 5X (7. 5 g/L) Sabouraud dextrose agar (SDA) plates with 20 μg/ml of ampicillin, streptomycin and tetracycline at 10° C for 3 weeks in the dark. Identification of Pd was confirmed based on the species morphological characters i. e. , the presence of curved conidia [1] and DNA sequences from conserved regions: internal transcribed spacer1 (ITS1), elongation factor 1α (EF-1α) and glyceraldehyde 3-phosphate dehydrogenase (gdp) genes. The pure cultures of Pd were obtained either by single spore isolation or hyphal tip cultures. For single spore cultures, actively growing Pd plates (100 mm X 15 mm) of over three weeks old were flooded with 2 ml of sterile water and gently swirled to release the spores (conidia). The spore suspension was vortexed for one minute to avoid clumping of spores. The spore suspension was then picked using an inoculating loop and spread over water agar plate (19 g/L). About 1 ml of sterile water was added in the process to help to spread the spores uniformly. The plate was viewed under a dissecting microscope and concentration of the spore suspension was adjusted so that each plate had 20–30 spores. The plate was then cultured at 7°-10°C in the dark and checked for germination every alternate day. Once the spores germinated, an agar plug was cut containing hyphae from the single germinating spore without damaging growing hyphae and then plated on a regular SDA plate to culture. For hyphal tip culture, we used the protocols described by Kanematsu et. al. [37] with some modification. We plated spore suspension on regular SDA plates as described above but when spores geminated and mycelia mats were formed they were gently overlaid with sterile Whatman cellulose filter paper soaked in SDB. The plates were then cultured for an additional two weeks until the fungal hyphae penetrated the filter paper and started growing on the upper surface. At that point the filter paper was removed and its upper surface was scraped gently and hyphal segments were suspended in sterile water. The method produced hyphal segments ranging from 4–8 cells in length that were appropriate for the hyphal tip culture. The hyphal segment suspension was then plated on SDA plates adjusting the concentration so that each plate had uniform distribution of 20–30 hyphal segments. Finally agar plugs grown from individual hyphal segments were cultured in separate plates to obtain a pure culture. The fungal isolates were stored in SDA plates for short-term storage at 4°C and at -80°C in the form of mycelia in 50% glycerol for long-term storage. All Pd isolates from Pennsylvania, one from Vermont and one from Indiana used in this study were isolated and cultured in our laboratory. The substrates (bat wings, wing punches, swabs) for these cultures were obtained from routine surveys of the Pennsylvania Game Commission (http: //www. pgc. pa. gov/Wildlife/Wildlife-RelatedDiseases/WhiteNoseSyndrome). The isolates from New York, West Virginia, North Carolina, Ohio, the remaining two isolates from Vermont and all European isolates were obtained as sub-cultures from the Center for Forest Mycology Research, United States Forest Service, Madison, WI (http: //www. fpl. fs. fed. us/research/centers/mycology/culture-collection. shtml). The Canadian isolates were obtained as sub-cultures from New Brunswick Museum collections, New Brunswick, Canada (http: //www. nbm-mnb. ca). In addition, we obtained five isolates of Geomyces sp. collected from Antarctic soil from Dr. Robert A. Blanchette’s collection at the University of Minnesota and we used six isolates of Pseudogymnoascus sp. from cave soil in Pennsylvania for this study. We extracted dsRNAs from lyophilized mycelia of Pd with a minor modification in the protocol described by Márquez et. al. [38], specifically Pd was cultured using mycelial plugs or spores in 150 ml of 0. 5X Sabouraud dextrose broth (SDB) supplemented with 20 μg/ml of ampicillin, streptomycin and tetracycline in a shaker at 10°C under dark conditions for three weeks prior to lyophilization. In addition to binding to CF11 cellulose (Whatman) in the presence of ethanol, the chemical nature of the dsRNA was confirmed by its resistance to DNase and RNase with NaCl concentration > 0. 3M. Approximately 2 μg of dsRNA were mixed with 2 μM of random primer-dN10 with a linker sequence (5' CCTTCGGATCCTCCN103' ), 0. 5 mM of Tris-EDTA (pH 8. 0) and nuclease-free water to a final volume of 12 μl, and boiled for 2 min. The mixture was incubated on ice, and 8 μl of Reverse Transcriptase (RT) mix (SuperScriptTM III RT 200U, 5X First-Strand buffer 4 μl, 0. 1M DTT 1 μl and dNTP 0. 5 mM as recommended by the manufacturer) were added and incubation continued at 50°C for 1. 5 hours. The newly synthesized cDNA mixture was then incubated with 10 μg of boiled RNase A (Sigma) for 15 min. at room temperature and cleaned with E. Z. N. A Cycle Pure Kit (Omega Bio-tech) according to the manufactures instruction. About 0. 5 μg of cleaned cDNA was used as a template for a 25 μl polymerase chain reaction (PCR) with Taq DNA Polymerase (ThermoFisher Scientific), buffers, dNTPs supplied with 1μM concentration of the primer (5' CCTTCGGATCCTCC 3' ). The PCR was completed in a Idaho Technologies Rapid Cycler with a slope setting of 5, using the following cycles: 1 cycle of 94°C, 1 min. ; 25 cycles of 94°C, 0 sec. , 45°C, 0 sec. , and 72°C, 15 sec. ; 1cycle of 72°C, 5 min. ; 1 cycle of 37°C, 5 min. The PCR product was cleaned and cloned into the pGEM-T Easy Vector System (Promega) according to the manufacturers instructions. Sequence analysis of the cDNA plasmid clones were done by the Genomic Core Facility of Pennsylvania State University, University Park, PA. The sequences obtained were trimmed for plasmid and primer sequences and assembled using de novo assembly in Geneious version 8. 0. 2 [39]. All cloning and sequence analysis was based on the dsRNA from the LB-01 isolate cultured from a little brown bat from Pennsylvania. RNA ligase mediated-rapid amplification of cDNA ends (RLM-RACE) was performed to determine the terminal sequences of the PdPV-pa dsRNA segments. A 5' -phosphorylated oligodeoxynucletide (5' -PO4-GGAGGATCCGAATTCAGG 3' ) was ligated to the dsRNA termini as an adaptor before synthesizing cDNAs using a complementary primer (5' CCTGAATTCGGATCCTCC3' ) in combination with the internal primers designed for PdPV-pa RNA1 and RNA2 (RNA 1: 5' TTCAAGTTCGCCCTGTACC3' F, 5' TGAGCGAATGGAAGGTTG3' R; RNA 2: 5' CGCGTAATCATGACGACC3' F, 5' CCGAGGAGCACACACTATC3' R) in RLM-RACE. Ligation reactions were done in 50% PEG with 2 U of T4 RNA ligase 2 (New England BioLabs) mixed with approximately 2 μg of dsRNA along with the primers mentioned above and buffer supplied according to the manufacturers instructions, and incubated at 4°C overnight. RT-PCR of the primer-ligated dsRNA was performed exactly like described in the cDNA synthesis above except the enzyme used was Avian Myeloblastosis Virus (AMV) RT (New England BioLabs). The amplicons were cloned followed by sequence determination using Sanger sequencing. The complete nucleotide sequences of PdPV-pa RNA 1 and PdPV-pa RNA 2 have been deposited in GenBank with accession numbers KY20754 and KY207544, respectively. Consensus sequences for PdPV-pa RNA 1 and RNA 2 were analyzed for the open reading frames (ORFs) using ORF finding operation in Geneious version 8. 0. 2. A sequence similarity search was conducted with BLASTn and BLASTx available online from the National Center for Biotechnology Information (NCBI). Northern blotting was performed using non-radioactive isotopes probes, digoxigenin (DIG) -11-dUTP-labeled DNA fragments according to the manufacturers instructions (Roche Diagnostics). Representative clones of PdPV-pa RNA 1 and RNA 2 in the range of 500–700 bp were selected and the labeling was done in a PCR with DIG-11-dUTP and dNTPs mix (DIG-11-dUTP: dTTP = 1: 3; with equimolar amount of dATP, dCTP and dGTP), Taq DNA Polymerase (ThermoFisher Scientific), specific primers and buffer in Idaho Technologies Rapid Cycler as described above. About 2 μg of PdPV-pa dsRNA was electrophoresed in 1. 2% agarose gels and subsequently denatured by saturating with freshly prepared 50mM NaOH for 30 min followed by neutralization in 50mM sodium borate for 5 min. The cycle was repeated three times before dsRNA was transferred to a nylon membrane (Hybond N+ Amersham) by capillary action overnight. The membranes were UV-cross-linked in a Stratalinker at 200 J. Hybridization and washings were carried out as described by Li et al. [40] except we performed prehybridization and hybridization at 52°C instead of 42°C. The blots were incubated in antibody solution, anti-DIG-AP Conjugate (Roche) and CDP-STAR (Roche) for chemiluminescence detection. Virus particles were purified following methods described by Sanderlin and Ghabrial [41] with some modifications. Eight g of lyophilized mycelia of Pd isolate BB-06 was ground to powder in the presence of liquid nitrogen. The homogenates were mixed with extraction buffer (0. 1 M sodium phosphate. pH 7. 6 containing 0. 5% (v/v) thioglycolic acid) and mixed with chloroform followed by low speed centrifugation at 7000 rpm for 15 min at 4°C. The virus containing supernatant was then subjected to two cycles of differential centrifugations (low speed at 7000 rpm for 15 min and ultracentrifuge at 35,000 for 1. 5 hours). During the ultracentrifuge cycle, the virus containing supernatant was underlaid with a 10% sucrose cushion. The final pellets were suspended in 1 ml of 0. 03 M sodium phosphate buffer pH 7. 6. The virus preparation was examined under JEOL 1400 transmission electron microscope after negatively staining with uranyl formate in the Microscopy and Imaging Facility at Penn State College of Medicine, Hersey, PA. For the heat stress, actively growing Pd plates in three replicates were exposed to room temperature (22–23°C), 37°C and 42°C for 2,6, 12 and 24 hours before culturing the mycelia plugs from each treatments in liquid medium (SDB) under normal laboratory culture conditions for Pd described above. During the treatments, Pd plates in three replicates were also grown under normal culture condition as controls. The fungal mycelia were then harvested after three weeks to extract dsRNAs. However, only samples treated at room temperature and 37°C for 2 hours grew. Single spore isolation and hyphal tip cultures were done as described under the section, fungal isolation and culture. The protoplast isolation from Pd was performed on mycelia (~ 1. 7 g) harvested from SDB culture after two weeks at 10°C in a shaker (200 rpm) in the dark. The fungal mycelia were collected by centrifugation at 90 × g for 5 min followed by washing with KCl buffer (0. 6 M, pH 5. 8) as an osmotic stabilizing agent. The mycelia was treated with lysing enzyme mixture (Lysing enzyme from Trichoderma harzianum 20 mg/ml and driselase 20 mg/ml from Sigma) prepared in KCl buffer and incubated at 10°C at 70 rpm in the dark. Protoplast production was checked every half an hour until 35–40 protoplasts were observed under a 40X field with 10 μl of the mixture. The mixture was then passed through double-layered miracloth (VWR) soaked in STC buffer (1. 2 M Sorbitol; 10 mM Tris-HCl, pH 7. 5; 20 mM CaCl2) to filter out the cell debris. The filtrate was centrifuged at 90 × g for 5 min to collect the protoplasts which were resuspended in regeneration media (0. 5% yeast extract, 2% glucose, 0. 6 M Sorbitol and 25 mM CaCl2) followed by incubation at 10°C at 70 rpm in the dark. Once the protoplasts recovered completely with cell wall growth, they were transferred to agar supplemented regeneration media (0. 5% yeast extract, 2% glucose, 20% sucrose and 1% agar) and the concentration adjusted so that each plate had 25–30 uniformly distributed cells. The plates were then incubated under normal culture condition for Pd until hyphae developed uniformly around each protoplast without touching each other. Individual colonies were then picked and cultured in SDA. We also treated Pd with the antiviral drugs cycloheximide and ribavirin at different concentrations in SDA media. Cycloheximide was used at 2 μg/ml, 5 μg/ml, 10 μg/ml, 15 μg/ml and 25 μg/ml concentrations. Ribavirin treatment was at 80 μM, 100 μM, 150 μM, 200 μM and 300 μM concentrations. Three passages with both cycloheximide and ribavirin were also performed with higher concentrations. For PEG induced matric stress on water availability we used PEG 8000 (Fisher BioRegents) in a modified Spezieller Nährstoffarmer liquid media (SN: 0. 02 g sucrose, 0. 02 g glucose, 0. 08 g KNo3,0. 08 g KH2Po4,0. 04 g MgSo4. 7H2O and 0. 04 g NaCl/L) to make media with water potential gradients of -1 MPa, -2 MPa, -3MPa, -4 MPa, -5 MPa and -6 MPa. The amount of PEG 8000 in gram/gram of water was calculated based on Michel [42] equation: Ψ (water potential) = 1. 29 [PEG]2T – 140[PEG]2–4 [PEG] and the value was adjusted to the Pd culture temperature of 10°C. An agar plug containing actively growing Pd was placed in 50 ml autoclaved modified SN liquid media with a targeted amount of PEG 8000 (-1 MPa: ~ 0. 075 PEG g/g of water, -2 MPa: ~ 0. 11 PEG g/g of water, -3 MPa: ~ 0. 14 PEG g/g of water, -4 MPa: ~ 0. 16 PEG g/g of water, -5 MPa: ~ 0. 19 PEG g/g of water and -6 MPa: ~ 0. 21 PEG g/g of water) and grown as described above. After three weeks, pieces of newly growing mycelia of Pd were transferred to normal SBD routinely used to culture Pd and the fungus was harvested after a normal culture period. The fungi from different treatments were examined for PdPV-pa both by dsRNAs gel electrophoresis and RT-PCR with PdPV-pa specific primers. In all methods Pd isolate LB-01 was used. Genetic variation in North American PdPV-pa isolates were determined by sequence analysis of RdRp and CP segments amplified in RT-PCR using specific primers. The primer pairs specific to RdRp (5' ATGGAAGTATCTCCTTTCG3' F, 5' GTATAGAAGATTGAGTGCC3' R) and CP (5' ACTCTGTGTTAACGGAGG3' F, 5' CTGTAGTTGACACCTGTACC3' R) were designed from the consensus sequences of RNA 1 and RNA 2 assembled from LB-01 isolate cloned sequences. PCR products using RdRp and CP specific primers from 45 North American PdPV isolates were sequenced and aligned with MUSCLE default settings in the program Geneious 8. 0 [39]. The RdRp sequences have been deposited in GenBank under accession numbers KY207498 to KY207552 and the CP sequences have been deposited in GenBank under accession numbers KY207453 to KY207497. The alignment was visually corrected as necessary before recording segregating and singleton sites. The average percentage identity for each sequence was calculated by taking the average from a pairwise percentage identity matrix generated from the sequence alignment. Phylogenetic analysis was performed using MrBayes [43] implemented via a plug-in in Geneious. The amino acid sequences were used in studying the evolutionary relationships of PdPV-pa within the genus Gammapartitivirus. The tree was constructed using amino acid sequence (RdRp and CP) of 10 approved species of Gammapartitivirus available in the GenBank. The sequences of Pepper cryptic virus 1, type member of genus Deltapartitivirus, which is the closest group to Gammapartitivirus in Partitiviridae family was used as outgroup. We used nucleotide sequences (CP) to study phylogenetic relationships of PdPV-pa in North American population. The nucleotide sequence of PsV-S CP was used as outgroup in the analysis. In Bayesian trees construction using amino acid sequence of the RdRp and CP ORFs, Jukes-Cantor substitution model was applied and for nucleotide sequences of CP General time-reversible (GTR) model with gamma rate variation was used based on the best model tested out of 28 models. | Many species of bats in North America have been severely impacted by a fungal disease, white-nose syndrome, that has killed over 5 million bats since it was first identified in 2006. The fungus is believed to have been introduced into a cave in New York where bats hibernate, and has now spread to 29 states and 4 Canadian provinces. The fungus is nearly identical from all sites where it has been isolated; however, we discovered that the fungus harbors a virus, and the virus varies enough to be able to use it to understand how the fungus has been spreading. This study used samples from infected bats throughout Pennsylvania and New York, and New Brunswick, Canada, as well a few isolates from other northeastern states. The evolution of the virus recapitulates the spread of the virus across these geographical areas, and should be useful for studying the further spread of the fungus. | Abstract
Introduction
Results
Discussion
Materials and Methods | sequencing techniques
fungal spores
vertebrates
animals
mammals
fungi
phylogenetic analysis
rna isolation
molecular biology techniques
rna sequencing
fungal reproduction
research and analysis methods
sequence analysis
mycology
rna sequence analysis
bioinformatics
molecular biology
molecular biology assays and analysis techniques
biomolecular isolation
dna sequence analysis
database and informatics methods
bats
biology and life sciences
amniotes
organisms | 2016 | Using a Novel Partitivirus in Pseudogymnoascus destructans to Understand the Epidemiology of White-Nose Syndrome | 9,689 | 203 |
Sustainable dengue intervention requires the participation of communities. Therefore, understanding the health beliefs, knowledge and perceptions of dengue among the local people can help to design locally appropriate strategies for effective interventions. A combination of qualitative semi-structured in-depth interviews (SDIs) and quantitative household questionnaire surveys (HHSs) was used to investigate the beliefs, knowledge and perceptions of dengue among the Shan people in Eastern Shan Special Region IV (ESSR4), Myanmar. The SDI was administered to 18 key informants, and the HHS was administered to 259 respondents. Only 14. 7% (95% CI: 10. 6–19. 6%) of the HHS respondents could confirm that mosquitoes transmit dengue; 14. 3% (95% CI: 10. 3–19. 1%) knew that piebald or Aedes mosquitoes transmit dengue; and 24. 3% (95% CI: 19. 2–30. 0%) believed that dengue-transmitting mosquitoes mainly lived in small ponds. Merely ten (0. 4%) of the 259 respondents of the HHS thought that dengue-transmitting mosquitoes bite in the day time. The people in the villages where there were outbreaks of dengue had more knowledge about dengue. This study demonstrates that the health beliefs of the Shan people were closely associated with their lifestyles, social and natural environments. To stay healthy, the Shan people clean their houses and surroundings regularly. However, their knowledge about dengue was not adequate for effective dengue control because it was mostly learned from previous dengue experiences and in a context that lacks systematic health education. Thus, in this setting, with a weak public health structure, more international support should be provided to promote the knowledge of the Shan people about dengue and to increase their sensitive awareness to dengue, which might be beneficial for social mobilization and community participation during future dengue prevention.
Dengue fever (DF) is a mosquito-borne disease caused by the dengue virus. It has become a substantially increasing threat to public health and represents a challenge to health services and a burden to economies. It was recently estimated that there are 390 million dengue infections every year (95% confidence interval 284–528 million), of which 96 million (67–136 million) infections manifest clinically [1,2]. The World Health Organization (WHO) estimated that 3. 9 billion people in 128 countries are at risk for infection with dengue viruses [3] and that approximately 75% of the global population who are at risk for contracting dengue is distributed across the Asia-Pacific region [4]. However, dengue is still one of the most neglected tropical diseases, which are caused and maintained by the social and environmental determinants of health [5]. The social determinants of infectious disease control are difficult to address [6]. For that reason, the Special Programme for Tropical Diseases Research and Training from the World Health Organization has called for more research on the interplay between the demographic, social, and environmental factors in infectious disease occurrence [7]. In recent years, outbreaks of DF have occurred along the China-Myanmar-Laos border, and most of these outbreaks affected Dai or Shan ethnic communities [8,9]. Dengue virus is primarily transmitted by Aedes mosquitoes, particularly Aedes aegypti. Water containers and discarded tires are the main productive habitats for Aedes [10]. The population dynamics of Aedes aegypti are influenced by human behavior and by the weather [11]. Outbreaks of DF may also be attributed to the social and environmental characteristics of the residents. Human behavior change communication (BCC) is one of the currently adopted strategies to reduce the Aedes population and dengue virus transmission [4,11–13]. Understanding people’s health beliefs, knowledge and perception about dengue may help to generate better strategies for dengue control [14,15]. Data about people’s health beliefs, knowledge and perceptions about dengue are still lacking at the China-Myanmar-Laos border. This kind of data may help to explore locally adaptive strategies and solutions for the next steps in dengue prevention and control. This study investigated the health beliefs, knowledge and perception of the Shan people about dengue in Eastern Shan Special Region IV (ESSR4), Myanmar. The purpose of using both qualitative and quantitative methods is to triangulate the study outcomes, to try to gain in-depth understanding and to compare any potential differences between the general community and the key informants who are supposed to have more chances to receive health information. Here, we report our findings and provide a discussion on their implications.
This investigation was a cross-sectional study that used mixed methods, including a quantitative HHS and qualitative SDI. Based on the number of reported dengue fever cases in 2017, three types of villages (high, low and no dengue incidence) were chosen as sample sites in Mongla Township in ESSR4, Myanmar. The SDI was administered to six key informants in each type of sampled village, so a total of 18 SDIs were conducted. The targeted sample size of the HHS, i. e. , 250 household heads for the questionnaire survey, was determined based on a 95% confidence interval of the standard value normal distribution, a 5% precision and an estimated 20% of adult people who know that mosquitoes transmit dengue virus [16]. Generally, people’s health beliefs are supposed to affect people’s knowledge, perception and preventive behaviors regarding diseases. In this paper, health beliefs are the general perceptions of elements related to health and diseases, not just special beliefs about dengue, i. e. , people’s perception of the effect of religious, socioeconomic and natural elements on their health and their perception of disease causes. More specifically, in this article, the knowledge and perception were specifically about dengue fever, including their knowledge of the clinical symptoms, vectors, dengue virus transmission, preventive methods and their perceived risk of contracting dengue fever. First, two of the main researchers discussed the study themes, including common diseases and health problems in the community, causes of poor health, knowledge of dengue, perceived risks of dengue and preventive methods, and then drafted the guidelines for the qualitative SDI. Second, the guidelines were pretested through discussions with village leaders and health workers and were then revised and finalized according to the results of the pretests. Third, based on the qualitative SDI guidelines, a questionnaire was developed, pretested by interviewing five heads of households, and then revised and finalized. The questionnaire consisted of 55 questions that covered demographics, education, family economics, beliefs on health and disease causes, the knowledge of dengue (including symptoms, transmission, vectors and prevention), perceived risk and attitudes. ESSR4 of Myanmar adjoins the Xishuangbanna Prefecture of China [17] and has three administrative areas, i. e. , Mongla Township, Nanban and Selei County. The total population of ESSR4 is approximately 110,000, and the majority of residents are Shan people (known as Dai in China, Thai in Thailand and Lao in Lao PDR), which is one of the main ethnicities in the Greater Mekong Subregion (GMS). The hot climate and adequate precipitation in ESSR4 provide a suitable environment for the growth and reproduction of mosquitoes, thereby increasing dengue transmission. Outbreaks of DF have recently occurred in Mongla Township each year [8]. The hospital of ESSR4 is the sole health facility that can perform the laboratory-based diagnosis and treatment of DF, and this hospital reported a total of 114 dengue fever cases in 2017. To compare the responses of communities with different DF experiences and to explore the potential reasons behind these responses, this study was deliberately designed to be carried out in three types of villages (V1 = no DF patients, V2 = low DF incidence, and V3 = high DF incidence). After discussion with the hospital of ESSR4, Mangjingpa and Wangnali (V1), Wangmaidao (V2) and Wangdong (V3) were selected for the study, as a common decision between us and the hospital. In 2017, V1 had 867 people and no reported DF cases; V2 had 389 people and nine reported DF cases; and V3 had 487 people and 36 reported DF cases. The qualitative data were collected by interviewing 18 key informants, with six from each type of village (i. e. , V1, V2 and V3). Participants included one health worker, three village leaders and two villager representatives who were chosen by the other villagers. The interview was conducted by following the SDI guidelines. One researcher interviewed the key informants in the Shan language, and another took notes in Chinese. The contents that were discussed included local health problems, villagers’ beliefs, including religion, the causes of diseases, local dengue situations, the name for DF in the Shan ethnic language and its meaning, the linkage between mosquitoes and dengue, the time that mosquitoes bite humans, the breeding sites of dengue-transmitting mosquito larvae and people’s perceptions of the prevention of DF [18,19]. The unit of sampling of this study was a household, which was defined as all those eating from the same cooking pot. The family wealth index (FWI) was determined by the household’s physical assets, such as housing, walls, roofs, and bicycles, and was classified into five groups (Table 1). Households were selected by a simple computer randomization in V1, V2 and V3. The 50% ±10% of the targeted sample size of the HHS (250 household heads) was decided for villages with reported dengue cases and without, respectively. Chinese is the official language in the ESSR4, so the questionnaire was written in Chinese. Researchers first visited each selected household and introduced the purpose of this study, topics and related questions that would be asked. The questionnaire was subsequently administered to each household head only after oral informed consent was obtained. Investigators who understood both the Shan language and Chinese from the Hospital of ESSR4 asked every question in the Shan language and then filled out the questionnaire in Chinese [20,21]. Records of the qualitative SDI were coded according to the contents of the questions and were then entered into cells in Microsoft Office Excel 2007. The same content records were combined by code sequencing. Two researchers independently analyzed the records of every content to generate themes. Finally, the two researchers discussed and compared their findings to finalize the findings. Quantitative data were entered in Excel 2007 and were analyzed in Epi Info 7. 2. The percentage and its 95% confidence interval (CI) were calculated for each health belief, piece of knowledge and perception. A chi-squared test was used to compare the percentage of each health belief, piece of knowledge and perception across the three village categories (V1, V2 and V3) [20,21]. The study was approved by the Bureau of Health of ESSR4, Myanmar. Ethical approval for this study was also granted by the Ethics Committees of the Yunnan Institute of Parasitic Diseases in China. Verbal consent was approved by the Ethics Committee as an acceptable form, as the study was interview-based and did not include any human specimens. All studied participants were adults (over 18-years-old). According to the World Medical Association Declaration of Helsinki, the purpose and procedures of the study were explained and disclosed to the participants before obtaining their consent. Participation was entirely voluntary, and participants could pass on a question, take a break or withdraw their consent from the study without providing any explanation at any time. Their consent was assumed if they did not refuse to answer questions.
The SDI was administered to 18 key informants, including nine males and nine females, ranging from 32- to 54-years-old. The questionnaire was administered to 259 household heads, and all completed questionnaires were considered valid. Among the 259 households, 211 (81. 5%, 95% CI: 76. 2–86%) were less poor (FWI 4–5) (Table 1). The respondents were made up of 149 females and 110 males. The mean age of the respondents was 42. 3-years-old (median: 49. 0, range: 18–99). A total of 225 (86. 9%, 95% CI: 82. 1–90. 7%) respondents said that they had not received any formal health education. Most of the participants and their family members were Buddhists. Only two of the HHS respondents were Papists, and the other three were not religious. All 18 key informants were Buddhist and agreed that, “One good turn deserves another. Evil will be recompensed with evil. The Buddha will bless and protect good people. If anyone does bad things, he or she would be punished due to their evil behaviors”. Half of 18 key informants thought that the natural environment in which they were living affected their health. In V1, three of the key informants agreed that, “In hotter weather, more people get ill”. In V2, two key informants agreed that, “Extremely hot or cold weather can make people ill”. In V3, four of the key informants agreed that, “When there are more trees, there are more mosquitoes and therefore more diseases”. Nine of 18 key informants believed that a certain link existed between household economic income and diseases. One of the key informants explained, “In poorer economic conditions, it is easier to get diseases because the poor may more easily suffer from malnutrition and be less healthy”. Fourteen of the key informants thought that clean living and working environments benefited their health. In V1, the key informants agreed that, “Poor hygiene is one of the disease causes”. In V2, the key informants agreed that, “Dirty water leads to diseases”. In V3, the key informants agreed that, “When hygiene is poor, it is easier to contract diseases” (Table 2). The results of the HHS showed that 79. 9% (95% CI: 74. 5–84. 6%) of the respondents believed that Buddha would bless good people; 54. 1% (95% CI: 48. 2–60. 6%) thought that anyone who did evil would be punished; 47. 5% (95% CI: 41. 3–53. 8%) mentioned that poverty was one of the causes of disease; 62. 9% (95% CI: 56. 7–68. 8%) believed that all natural factors, including climate, weather, water and forests, influenced human health; 87. 0% (95% CI: 92. 2–90. 9%) believed that cleanliness and sound hygiene benefited human health; and 21. 7% (95% CI: 16. 7–27. 2%) thought that unpolluted water was healthier than polluted water (Table 3). DF is named “Paya Yong” in the Shan ethnic language, in which “Paya” represents illness and “Yong” represents mosquitoes. Shan people have connected DF to mosquitoes from the literal meaning of this ethnic language phrase. However, most participants did not know the clinical symptoms of DF. Six of 18 key informants regarded fever as one of the symptoms of DF, and one of the key informants also mentioned rash, pantalgia and joint pain (Table 2). The results of the HHS showed that there was a higher proportion of respondents who did not know the DF symptoms or who did not respond to the question in V1 than in V2 and V3 (P = 0. 0453); only 17. 8% (95% CI: 13. 3–23. 0%) of the respondents mentioned fever; 14. 3% (95% CI: 10. 3–19. 1%) mentioned headache; 4. 2% (95% CI: 2. 1–7. 5%) mentioned orbital pain; 9. 7% (95% CI: 6. 3–13. 9%) mentioned pantalgia; and 3. 5% (95% CI: 1. 6–6. 5%) mentioned rash (Table 3). Seven people perceived DF as a serious or deadly disease among the 18 key informants. In V2 and V3, all 12 key informants thought that local people could easily become infected with DF; however, in V1, only one out of the six key informants thought that local people could easily become infected with DF (Table 2). The HHS results revealed that 68. 7% (95% CI: 62. 7–74. 3%) of the respondents either did not know about the risk or did not respond to the question regarding risk; only 27. 8% (95% CI: 22. 4–33. 7%) of the respondents thought of DF as a serious disease; and 24. 7% (95% CI 19. 6–30. 4%) of the respondents considered DF to be deadly (Table 3). Both the quantitative and qualitative data showed similar results. Ten of the 18 key informants regarded DF as a contagious disease, but they thought that DF could be transmitted from person to person directly by breathing, speaking and physical contact, etc. (Table 2). Furthermore, the HHS results revealed that 52. 9% (95% CI: 46. 6–59. 1%) of the HHS respondents did not answer the transmission question; only 25. 0% (95% CI: 19. 4–31%) the respondents thought that DF was transmittable; 20. 0% (95% CI: 14. 0–24. 9%) thought that DF could be transmitted from person to person directly; and only 14. 7% (95% CI: 10. 6–19. 6%) of the HHS respondents confirmed that DF was transmitted by mosquitoes. Some respondents thought that bacteria, viruses, animals and improper or dirty food as one of the causes of DF (Table 3). Six of the 18 key informants knew that dengue-transmitting mosquitoes were piebald or Aedes; they also knew that dengue-transmitting mosquitoes could bite during all 24 hours of the day. Only one key informant asserted that dengue-transmitting mosquitoes bit in the daytime, compared with four informants who thought that dengue-transmitting mosquitoes only bit at night. Thirteen key informants confirmed that dengue-transmitting larvae live in water; an additional eight of the 13 key informants confirmed that containers or small-scale pond water is the principal breeding sites for the larvae. The other five key informants did not know the breeding habitats of the mosquito larvae or they did not respond to the question (Table 2). The HHS results indicated that 79. 2% (95% CI: 73. 7–83. 0%) of the respondents were unable to answer the vector question; only 14. 3% (95% CI: 10. 3–19. 1%) of the respondents knew that piebald or Aedes were dengue-transmitting mosquitoes, and there were more HHS respondents who knew about dengue-transmitting mosquitoes in V2 and V3 than in V1 (P = 0. 0293). Only 21 out of the 259 respondents knew that dengue-transmitting mosquitoes bit people during the day and night, and ten further confirmed that biting could occur during the daytime. More respondents knew that mosquitoes bite in the daytime in V2 and V3 than in V1 (P = 0. 0467). A total of 33. 2% (95% CI: 27. 5–39. 3%) of the HHS respondents thought that all types of water sites were the habitats of dengue-transmitting mosquitoes, and only 24. 3% (95% CI: 19. 2–30. 0%) confirmed that the dengue-transmitting mosquito larvae were only living in containers or in the water of small ponds. More of the HHS respondents in V1 than in V2 and V3 (P = 0. 0018) could not answer this question or did not respond to the habitat question (Table 3). All participants showed a keen interest in this topic and thus actively discussed this issue with the investigators. Thirteen key informants knew that clearing mosquito breeding sites helps to prevent dengue transmission. Ten key informants said that good hygiene was beneficial for their health and expressed their willingness to participate in the cleaning of the community environment (Table 2). In the HHS, only 20 participants did not answer the questions regarding the control of mosquito larvae, and only five had no response to the questions. The HHS results indicated that 43. 6% (95% CI: 37. 5–49. 9%) of the respondents knew that maintaining good hygiene was important for eliminating the habitats of the dengue-transmitting mosquito larvae; 67. 2% (95% CI: 61. 1–72. 9%) knew that turning containers upside down was important for eliminating the habitats of the dengue-transmitting mosquito larvae; and 20. 5% (95% CI 15. 7–25. 9%) knew that draining small-scale pond waters was important for eliminating the habitats of the dengue-transmitting mosquito larvae. More participants in V2 and V3 than in V1 understood that good hygiene helps to control DF. Regarding adult mosquito control, 20. 1% (95% CI: 15. 4–25. 5) of the HHS respondents knew about the use of door and window screens; 79. 9% (95% CI: 74. 5–84. 6) knew about mosquito coils; 27. 8% (22. 4–33. 7%) knew about spraying or fogging with insecticides; and 69. 9% (63. 9–75. 4%) knew about the use of bed nets. More HHS respondents in V2 and V3 than in V1 thought that using window and door screens (P = 0. 02) and spraying with insecticides is helpful (P = 0. 5891). However, more respondents in V1 than in V2 and V3 thought that using mosquito coils (P = 0. 03) and bed nets (P = 0. 1235) helped with dengue prevention. Additionally, more participants in V1 than in V2 and V3 (Table 3) expressed their willingness to eliminate bamboo and tree stump holes (P = 0. 3302), to clean up dumps and to turn containers upside down (P<0. 0001) (Table 3).
The dengue burden is growing and has become large enough currently. More than half of the global population lives in territories that are at risk of becoming infected with dengue [2,4]. However, ability to contain epidemics of the dengue virus is still limited. In addition to supportive treatments, effective antiviral therapies continue to be lacking. The only licensed dengue vaccine is only partially protective [22]. Thus, intensified vector control could still be the most important strategy for DF prevention for a long time to come. Adopting protective behaviors is a multifactorial process that depends on both sociocultural and cognitive factors [23]. The Shan people have beliefs that have originated from their primary social living and culture. They believe that health is associated with all natural and social environmental factors, and that regularly cleaning their houses and surroundings to maintain sound hygiene would benefit human health. The name for DF in the Shan ethnic language (Paya Yong) connects dengue with mosquitoes, however, their knowledge and awareness of DF remain at low level. These low levels might be attributed to the fact that the local Shan people are unable to effectively benefit from the services that are provided by both the national health and education systems. Only nine of the participating household heads had completed more than six years of formal education. Thus, the participant’s low levels of knowledge and perceptions about DF might be explained by the lack of health services that they benefitted from and the inadequate formal education that they received. Meanwhile, few actions involving information, education and communication on dengue have been taken, despite the fact that the burden of DF is increasing [8], shortly after the malaria burden was effectively reduced [24]. In V2 and V3, there was a higher response rate and level of knowledge than in V1 about DF symptoms, transmission, vectors, time of day when mosquitoes bite and the breeding sites of dengue-transmitting mosquitoes. This difference might be attributed to the fact that the V2 and V3 communities have suffered from DF and thus have more experience and knowledge regarding dengue. Over half of the participants could not answer questions about the causes of DF. Only a quarter of the participants understood that dengue is transmittable, but most of the participants thought that dengue could be directly transmitted from person to person by breathing, speaking, physical contact, etc. (Tables 2 and 3). What is worse, almost all key informants did not have enough knowledge to communicate with their fellow villagers to prevent future outbreaks. Most of the participants were unsure if they were at risk of being infected by dengue. Both the qualitative and quantitative data show that the participants had difficulty with answering the questions on the risks and seriousness of DF. Only 15. 8% (95% CI: 11. 6–20. 9%) of the participants perceived the risk of becoming infected. Furthermore, less than one-third of the participants (72/259) considered DF to be a serious disease and lethal. People in V2 and V3 were more susceptible to DF than those in V1. This indicates that their perceived risk may have correlated with their previous experience of DF outbreaks too. This Inappropriate preparedness due to the shortage of necessary resources might be one of the reasons for the outbreaks of DF in ESSR4, Myanmar. More of the participants in V1 than in V2 and V3 showed a willingness to eliminate bamboo and tree stump holes, to clean up dumps and to turn containers upside down. Based on some of the SDI data, local governments have frequently required villagers to clean their houses, public spaces in the community and the surrounding environment during DF outbreaks. In response to DF outbreaks, three ten-person teams in V3 would alternate to conduct spraying and fogging with insecticide once a week. During our communications, we felt that people in V2 and V3 became bored while hearing about these requirements. One villager said, “The campaign just stopped a few days ago”. This ennui might lower their willingness to perform environmental management and may become a new challenge for the sustainability of intensive vector control. Because of the malaria control program, the residents’ knowledge of adult mosquito control was improved to a certain level across ESSR4 [24]. In the HHS, 79. 9% (95% CI: 74. 5–84. 6%) and 69. 9% (95% CI: 63. 9–75. 4%) of the participants mentioned that they had used coils and bed nets, respectively. More of the participants in V1 than in V2 and V3 knew how to use mosquito coils and bed nets. However, only a few people knew that dengue-transmitting mosquitoes bite during the day and that small-scale pond water is the main breeding site of dengue-transmitting mosquito larvae. Most participants knew to protect themselves from the bites of adult mosquitoes, but they did not know that eliminating the habitats of dengue-transmitting mosquito larva is beneficial for the control of dengue vectors (Tables 2 and 3). Combining effective interventions against multiple arboviral diseases has been suggested by some scholars as one of the most cost-effective and sustainable strategies for the reduction of vector-borne diseases [25,26]. However, another study that conducted a meta-analysis documented that indoor residual spraying (IRS) did not significantly impact DF infection risk [27]. Both this literature and our study results illustrate that knowledge regarding the control of adult mosquitoes is not sufficient for controlling the dengue vector. Unavoidably, this study was constrained by two obvious limitations. First, one of the objectives of the qualitative SDI was to triangulate the study outcomes of the quantitative methods. Much more attention was paid to this objective in the qualitative SDI, and not enough nuanced data were collected. This limited our ability to delineate the outcome differences between the qualitative SDI and the quantitative HHS despite the qualitative and quantitative outcomes also showed that the key informants had better knowledge than overall community people. Further qualitative study should been needed. Second, the response rates of the HHS participants to certain questions were not high enough, and thereby might have caused information bias. However, all investigators felt that the Shan People were very friendly, and we very much enjoyed collaborating with them during the field research. This means that the Shan people may have been unlikely to reject answering questions that they could answer. When the individuals had no response to a specific question it was mainly because they did not know the answer. To add to this disclosure, we combined the answers of “I don’t know” and no response into a single category during analysis. In conclusion, the Shan people believe that health is associated with all natural and social environmental factors, and that regularly cleaning their houses and surroundings to maintain sound hygiene would benefit human health. However, their knowledge and aware sensitivity of DF remains at low level, and most of their knowledge and awareness was learned from previous experiences in controlling malaria and dengue. Thus, in this setting with a weak public health structure, more international support should be provided to promote the knowledge of the Shan people about dengue and to increase their aware sensitivity to dengue. With proper guidance, social mobilization and community participation might help increase the perception of DF and the involvement of the Shan people for dengue control. | The burden of dengue has been increasing over the last five decades, and dengue fever (DF) has become one of the most rapidly spreading mosquito-borne diseases. DF has become another disease that threatens public health after malaria has become successfully controlled along the China-Myanmar border. However, it is currently not easy to contain epidemics of the dengue virus. As part of an integrated vector management approach, a community-based method is effective in the prevention of DF by tailoring the approach in a local context. Consequently, mixed methods comprising qualitative semi-structured in-depth interviews (SDIs) and quantitative household questionnaire surveys (HHSs) were used to study the health beliefs, knowledge and perceptions about dengue among the Shan people in Eastern Shan Special Region IV, Myanmar. This study found that the Shan people believed that their health was closely associated with their lifestyle and the social and physical environment in which they lived. Their beliefs originated from their primary social activities and cultural heritage. Most of their knowledge about DF was learned from previous outbreaks and interventions for the disease. The Shan people had a relatively higher level of knowledge about adult mosquito control, which they learned from previous malaria control programs, but they lacked knowledge on DF symptoms, transmission, vectors and Aedes larval breeding sites. Their knowledge about the methods of adult mosquito control could not effectively control DF. More sound health education is urgently needed to increase the local people’s knowledge of dengue and to rouse community awareness and participation in cleaning vector breeding sites. In the context of a lack of the necessary technical and financial resources, these interventions might rely more on international aid and help from neighboring countries, such as China. | Abstract
Introduction
Methods
Results
Discussion | invertebrates
medicine and health sciences
behavioral and social aspects of health
tropical diseases
social sciences
ponds
neuroscience
animals
health care
developmental biology
cognitive psychology
bodies of water
neglected tropical diseases
insect vectors
language
public and occupational health
infectious diseases
dengue fever
marine and aquatic sciences
life cycles
hygiene
disease vectors
insects
arthropoda
socioeconomic aspects of health
mosquitoes
psychology
eukaryota
earth sciences
biology and life sciences
species interactions
viral diseases
cognitive science
larvae
organisms | 2019 | The Shan people’s health beliefs, knowledge and perceptions of dengue in Eastern Shan Special Region IV, Myanmar | 6,512 | 364 |
Despite the central role of estrogen exposure in breast and endometrial cancer development and numerous studies of genes in the estrogen metabolic pathway, polymorphisms within the pathway have not been consistently associated with these cancers. We posit that this is due to the complexity of multiple weak genetic effects within the metabolic pathway that can only be effectively detected through multi-variant analysis. We conducted a comprehensive association analysis of the estrogen metabolic pathway by interrogating 239 tagSNPs within 35 genes of the pathway in three tumor samples. The discovery sample consisted of 1,596 breast cancer cases, 719 endometrial cancer cases, and 1,730 controls from Sweden; and the validation sample included 2,245 breast cancer cases and 1,287 controls from Finland. We performed admixture maximum likelihood (AML) –based global tests to evaluate the cumulative effect from multiple SNPs within the whole metabolic pathway and three sub-pathways for androgen synthesis, androgen-to-estrogen conversion, and estrogen removal. In the discovery sample, although no single polymorphism was significant after correction for multiple testing, the pathway-based AML global test suggested association with both breast (pglobal = 0. 034) and endometrial (pglobal = 0. 052) cancers. Further testing revealed the association to be focused on polymorphisms within the androgen-to-estrogen conversion sub-pathway, for both breast (pglobal = 0. 008) and endometrial cancer (pglobal = 0. 014). The sub-pathway association was validated in the Finnish sample of breast cancer (pglobal = 0. 015). Further tumor subtype analysis demonstrated that the association of the androgen-to-estrogen conversion sub-pathway was confined to postmenopausal women with sporadic estrogen receptor positive tumors (pglobal = 0. 0003). Gene-based AML analysis suggested CYP19A1 and UGT2B4 to be the major players within the sub-pathway. Our study indicates that the composite genetic determinants related to the androgen–estrogen conversion are important for the induction of two hormone-associated cancers, particularly for the hormone-driven breast tumour subtypes.
Estrogen exposure is critical for the development of both breast and endometrial cancers and represents the most well-established risk factors for both diseases. Estrogen is a metabolic product whose circulating level is determined by de novo synthesis, conversion from other steroid hormones, and mechanisms of estrogen elimination. These metabolic processes are regulated by a network of enzymes encoded by different genes, suggesting that genetic variation within these metabolic genes may impact on breast and endometrial cancer risk. Genetic variation within the estrogen metabolic pathway has been intensively investigated, mostly by analyzing single variant effects in a limited number of candidate genes, SNPs and study subjects. The inadequacies of study design and analytical methodology have caused these studies to be underpowered for detecting moderate genetic effects which, not surprisingly, has led to inconsistent results [1]–[8]. We surmised that strategies for assessing the synergistic effect of multiple genetic variants within the estrogen metabolic pathway may provide a more realistic determination of genetic effect than a single gene, single SNP approach. Herein, we present a comprehensive analysis of genetic variation in the estrogen metabolism pathway and its association with breast and endometrial cancer risk using a pathway-based approach.
We performed single SNP association analysis in 1596 breast cancer cases, 719 endometrial cancer cases and 1730 population controls from Sweden. Of the 239 tagSNPs analyzed, 17 SNPs (7. 1%) had p-values less than 0. 05 for breast cancer, and 18 SNPs (7. 5%) had p-values less than 0. 05 for endometrial cancer (Table S4 and Table S5). For breast cancer, the smallest p-value was 0. 00034 at rs7167936 within CYP19A1, and for endometrial cancer, the smallest p-value was 0. 00017 at rs12595627 in CYP19A1. The single-SNP associations were all moderate. Only rs12595627 (for endometrial cancer) survived the conservative Bonferroni correction for multiple testing at α = 0. 05. Overall, however, the single-SNP p values appeared to deviate from their null distribution of no association (formally tested below). The single-SNP associations were suggestive, but instead of any single variant having a strong effect, there appeared to be multiple weak associations within the metabolic pathway. To evaluate the cumulative effect from multiple variants we employed the AML method [9] that assesses the experiment-wide significance of association by analyzing multiple SNPs through a single global test. The whole metabolic pathway can be sub-divided into three a priori defined sub-pathways, each performing specific metabolic function (Figure 1). Sub-pathway 1 is involved in the synthesis of androgen, sub-pathway 2 is involved in the conversion of androgens to estrogens, and sub-pathway 3 is responsible for removing estrogens. To investigate whether there is multi-SNP association for the whole pathway and whether any of the three sub-pathways is particularly important in influencing disease risk, we performed the progressive pathway-based global test on the whole metabolic pathway as well as the three sub-pathways using the AML method. The global test yielded marginally significant association for the whole metabolic pathway in both breast (pglobal = 0. 034) and endometrial (pglobal = 0. 052) cancers (Table 1). Dividing the metabolic pathway into three functional sub-pathways for the global test revealed strong association between the androgen-to-estrogen conversion sub-pathway and both breast (pglobal = 0. 008) and endometrial (pglobal = 0. 014) cancer (Table 1). The association evidence survived correction for performing 4 pathway-based tests in each cancer (pglobal corrected = 0. 032 for breast and 0. 056 for endometrial). In contrast, the other two sub-pathways showed no association with either form of cancer. For approximately half of the Swedish subjects in the breast cancer study (797 cases and 764 controls) we have genome wide association study (GWAS) data available. We used this to assess the possible influence of population stratification on our results. For the GWAS dataset, the genomic inflation factor, λgc, was1. 015. Assuming an equal level of population stratification (in terms of the fixation index FST) in the current study and the GWAS sub-study, we estimated the genomic inflation factor, λgc, to be 1. 030 in the current study, using the relationship between FST, sample size and λgc described in [10]. Using the λgc value of 1. 030 for genomic control-based correction of population stratification, the corrected global AML p-values for breast cancer are 0. 052 for the entire pathway and 0. 011 for the androgen-estrogen conversion sub-pathway, leaving our results largely unchanged. Even if λgc was as large as 1. 05 in the current study, the global test p-value for the androgen-estrogen conversion sub-pathway would still be as low as 0. 014. To further ensure that the observed associations could not be due to the employment of 319 paraffin-embedded tissue samples in the analysis, we re-ran analyses excluding 319 paraffin-embedded tissue samples, and (at the same time) excluding 33 SNPs with call rates of less than 95%. Results were very similar. For example, for breast cancer, p-values were 0. 028 and 0. 009 for the entire pathway and for the androgen-estrogen conversion sub-pathway, respectively. To validate the association in the androgen-to-estrogen conversion sub-pathway, we genotyped the 120 SNPs of this sub-pathway in an additional 2245 breast cancer cases and 1287 controls from Finland and performed the same AML analysis by using the 118 successfully genotyped SNPs. The validation analysis in the Finnish sample revealed similar evidence of association between the androgen-to-estrogen conversion sub-pathway and breast cancer (pglobal = 0. 015) (Table 1). The non-centrality parameter from the AML analysis of the androgen-to-estrogen conversion sub-pathway, which represents the size of the common effect of the associated SNPs, was estimated as 2. 90 for the Swedish sample and 2. 94 for the Finnish sample. The similar values indicate a consistent size of the genetic effect in the two samples. A joint analysis of the Swedish and Finnish samples further yielded a global p-value of 0. 001 (Table 1). The SNPs with the lowest p-values in the Finnish sample are listed in Table S6. Hormone-related risk factors may play a differential role in breast cancer subtypes. In particular, estrogens appear to drive the development of ER positive tumors. This prompted us to investigate the association in the androgen-to-estrogen conversion sub-pathway in hormone-related breast tumor subtypes. As surrogate markers for hormone driven tumour subtypes we constructed variables as combinations of menopausal status, family history and estrogen receptor (ER) status and divided all the patients into subgroups. We then compared subgroups of patients, defined on values of these variables, with controls, to evaluate the role of the androgen-to-estrogen conversion sub-pathway in different patient subgroups. First, we compared patient subgroups against all the controls in the combined Swedish and Finnish samples. The subgroup results showed that in the combined samples, significant association was observed in postmenopausal patients (pglobal = 0. 009 and 0. 018 respectively), postmenopausal patients without family history (pglobal = 0. 001 and 0. 04 respectively), and postmenopausal patients with estrogen receptor positive (ER+) tumors (pglobal = 0. 0006 and 0. 05 respectively) (Table 2). No significant association was observed in either premenopausal patients or postmenopausal patients with family history or estrogen receptor negative (ER−) tumors. Then, to rule out the possibility that the above subgroup results were caused by the mismatch between the patient subgroups and the controls in terms of the variables which defined patient subgroups, we performed the second subgroup analysis where the controls were also divided into subgroups according to family history and menopausal status (Table 3). The second subgroup analysis was only performed in the Swedish sample, because the Finnish controls lack information on family history and menopausal status. This yielded similar evidence for the association of the sub-pathway with the hormone-driven subtypes of breast cancer as in Table 2. We further investigated the impact of reproductive risk factors on the genetic association of the androgen-to-estrogen conversion sub-pathway with breast cancer. Because the risk factor information is not available for the Finnish controls, the analysis of the reproductive risk factors was performed in the Swedish samples where information on such factors is available. We performed the AML analysis of the androgen-to-estrogen conversion sub-pathway with adjustment for the reproductive risk factors (parity, age at the first birth, age at menarche and age of menopause) and HRT use. We investigated this primarily to assess whether any of the reproductive risk factors could be in the causal pathway. Since p-values remained almost unchanged in all analyses (Table 4), it appears that none of the reproductive risk factors are likely to be in the causal pathway. Attempting to refine the association within the androgen-to-estrogen conversion sub-pathway, we performed a gene-based AML analysis in the combined Swedish/Finnish breast cancer sample and the Swedish endometrial cancer sample. Among the 15 genes tested (Table 5), strong association was observed for CYP19A1 with both breast (pglobal = 0. 003) and endometrial (pglobal = 0. 006) cancer and UGT2B4 (pglobal = 0. 002) with breast cancer only. The associations in breast cancer survived correction for multiple testing of 15 genes (pglobal corrected = 0. 045 for CYP19A1 and 0. 03 for UGT2B4). We also observed suggestive association for UGT2B11 in breast and endometrial cancer as well as for HSD11B1, SULT2A1 and SULT2B1 in breast cancer. Consistent with the pathway-based associations, the gene-based associations are generally more significant in sporadic postmenopausal patient samples than in the whole breast cancer sample (except SULT2B1). Furthermore, the importance of CYP19A1 and UGT2B4 in breast cancer risk is supported by the fact that excluding either gene from the global test of the sub-pathway reduced the global significance of association for the sub-pathway, from 0. 0015 to 0. 011 for CYP19A1, and to 0. 010 for UGT2B4. However, the fact that the association for the sub-pathway remained significant, after excluding either gene, suggests that, although CYP19A1 and UGT2B4 are the major players, genetic variation within other genes also contributes to the association within the sub-pathway.
Our pathway-based multi-SNP association analysis revealed a significant association between genetic variants in the androgen-to-estrogen conversion sub-pathway and the risk of two hormone dependent cancers. The association was particularly strong for ER+, sporadic breast cancer. Single SNP analysis did not reveal a similar association. We used the AML-based multi-SNP analysis, which has been shown to be more powerful than single SNP tests to yield significant and consistent association, when genetic risk is carried by multiple risk alleles each with moderate effect [11]. Pathway-based approaches are just beginning to be applied in association analysis [12]. Recently, an association study of 9 candidate gene groups (involving 120 candidate genes) was performed in breast cancer by using the AML approach, and interestingly, only the group of 8 genes involved in the steroid hormone signalling were significantly associated [13]. Our study has moved one step further and highlights the fact that the power of the pathway-based association analysis can be increased when analysis is guided by well-defined biological information. We believe that pathway approaches have potential to move genome-wide association studies beyond their initial success of identifying some ‘low-hanging fruits’ to revealing many weak genetic risk alleles that have been missed by single SNP analysis. Unless one enzyme is the rate limiting step for the entire metabolic pathway, it is not likely that small functional perturbations of individual variants would have a major impact on the overall effect of the metabolic pathway. To test the hypothesis that several genetic variants, each conferring weak to moderate effects, contribute to genetic risk, we adopted a systematic pathway-based approach for association analysis by testing the joint effect of multiple genetic variants in a progressive fashion from the whole metabolic pathway to biochemical sub-pathways and further down to individual genes. Such a progressive approach allows us to not only establish consistent association in three cancer samples from two different populations but also to refine the association of the androgen-to-estrogen conversion component of the metabolic pathway. Our study may therefore have advanced our understanding of the role of estrongen metabolism in breast and endometrial cancers by 1) accounting for the ambiguity surrounding the genetic association results and 2) indicating the androgen-to-estrogen conversion to be the important component of the metabolic pathway in modulating the risk and therefore to be a worthy focus for future studies. After menopause, ovarian estrogen production dramatically declines and conversion of adrenal androgens to estrogens in peripheral tissues becomes the major source of circulating estrogens. The final step of this conversion is catalyzed by aromatase, encoded by CYP19A1 [11]. Thus, there is biological plausibility in the association between CYP19A1 polymorphisms and postmenopausal breast cancer. Moreover, pharmacological inhibition of aromatase prevents recurrences in postmenopausal women with estrogen-receptor-positive breast cancer and new contralateral primaries [14], which has challenged the previous routine of a 5-year course of tamoxifen alone [15]. Our study has advanced our understanding of CYP19A1 by suggesting that the modulation of aromatase activity by either germ-line variation or pharmacological agents can influence the development of ER+ tumour in postmenopausal women. Furthermore, the convergence of genetic and pharmacological effects of CYP19A1 also raises therapeutic possibilities. For example, other genes implicated by our genetic study, such as UGT2B4, might also be pharmacological targets for treating breast cancer. Hormone exposure is a common risk factor for breast and endometrial cancer. Our employment of the three samples of two different hormone-related cancers from two different populations allowed us to apply a very stringent criterion for declaring an association. Furthermore, results of our breast cancer patient subgroup analysis indicate that the genetic determinants within the androgen-to-estrogen conversion sub-pathway may play a more prominent role in postmenopausal women with sporadic ER+ tumors, further suggesting that the modulation of hormone exposure by genetic variation may have a differential impact on breast tumor subtypes. Endogenous sex hormone level appears to be associated with breast cancer risk in postmenopausal women [16], and particularly with the risk of ER+/PR+ breast tumors [17]. The effect of hormone-related factors on breast cancer risk apparently differs by ER status [18] and menopause status [19], [20]. It could also differ by the status of family history of the disease, as suggested by a recent study showing that most cases of hereditary breast cancer are probably not related to cumulative hormone exposure [21]. Our findings may have therefore advanced the development of a general model for breast cancer risk: hormonal factors, both genetic and reproductive, can play a key role in the genesis of post-menopausal and “sporadic” breast cancer, whereas genes involved in DNA repair, checkpoints, and genetic stability (such as BRCA1, BRCA2, p53, ATM, CHK2) appear to be more involved in predominantly breast cancers associated with family history of disease. It is worth noting that the contribution of genetic polymorphisms to risk is a function of both their prevalence and penetrance and thus the relative importance of individual SNPs may vary from population to population. More studies in different populations are needed to fully understand the role of the androgen-to-estrogen conversion sub-pathway in breast cancer. We also want to highlight that our results are of genetic association in nature, and further studies are needed to confirm the findings and to identify functional variants causally linked to cancer risk.
Swedish subjects were from a population-based case control study of breast and endometrial cancer as described [22], [23]. Briefly, the study included all incident primary invasive breast and endometrial cancers among Swedish-born postmenopausal women between 50 and 74 years of age at diagnosis, diagnosed with breast cancer between October 1993 and March 1995 and endometrial cancer between January 1994 and December 1995. All cases were identified through six regional cancer registries in Sweden, and all controls were randomly selected from the Swedish Registry of Total Population and frequency matched to the expected age distribution of the cases. Finnish breast cancer cases consist of two series of unselected breast cancer patients and additional familial cases ascertained at the Helsinki University Central Hospital. The first series of 884 patients was collected in 1997–1998 and 2000 and covers 79% of all consecutive, newly diagnosed cases during the collection periods [24], [25]. The second series, containing 986 consecutive newly diagnosed patients, was collected in 2001–2004 and covers 87% of all such patients treated at the hospital during the collection period [26]. An additional 538 familial breast cancer cases were collected at the same hospital as described [27]–[30]. 1287 anonymous, healthy female population controls were collected from the same geographical regions in Southern Finland as the cases and have been used in several studies previously [31]–[33]. Risk factor information and tumour characteristics were available for all the Swedish samples and the Finnish cases, but were missing for the Finnish controls. The Finnish samples (mean age = 56 for the cases and 41 for the controls) were younger than the Swedish samples (mean age = 63 for both the cases and controls). All the risk factor and tumour characteristics information of the subjects are summarized in Table S1 and Table S2. Written informed consent was obtained from all participating subjects, and the study was approved by the Institutional Review Boards in Sweden, Finland and at the National University of Singapore. DNA was extracted from 4 ml of whole blood using the QIAamp DNA Blood Maxi Kit (Qiagen) and non-malignant cells in paraffin-embedded tissue using a standard phenol/chloroform/isoamyl alcohol protocol [34]. We selected 35 genes involved in estradiol or estrone metabolism and expressed in the breast (based on published literatures). We selected 1007 single nucleotide polymorphisms (SNPs) in these genes and their 30kb flanking sequences from the dbSNP (build 124) and Celera databases, aiming for a marker density of at least one SNP per 5kb (Table S3). These SNPs were genotyped in 92 Swedish control samples to assess linkage disequilibrium pattern and coverage. Haplotypes were reconstructed using the PLEM algorithm [35] implemented in the tagSNPs program [36]. A subset of SNPs, tagSNPs, were selected based on the R2 coefficient, which quantifies how well the tagSNP haplotypes predict the genotype or haplotypes an individual carries. We chose tagSNPs so that common SNP genotypes and haplotypes (frequency ≥0. 03) were predicted with R2≥0. 8 [37]. To evaluate our tagSNPs' performance in capturing unobserved SNPs within the genes, we performed a SNP-dropping analysis [38], [39]. In brief, each of the genotyped SNPs was dropped in turn and tagSNPs were selected from the remaining SNPs so that their haplotypes predicted the remaining SNPs with an R2 value of 0. 85. We then estimated how well the tagSNP haplotypes of the remaining SNPs predicted the dropped SNP, an evaluation that can provide an unbiased and accurate estimate of tagSNP performance [38], [39]. Overall, we selected and genotyped 302 tagSNPs from the 35 genes in all the Swedish cases and controls. Genotyping was performed using the Sequenom system (San Diego, California). All genotyping results were generated with positive and negative controls and checked by laboratory staff unaware of case-control status. Of the 302 tagSNPs, 42 SNPs failed in the development stage of Sequenom genotyping assays. SNPs with a call rate <85% (8 SNPs), minor allele frequency <1% (9 SNPs) or out of Hardy-Weinberg Equilibrium (p<0. 05/252,4 SNPs) were excluded from further analysis. Overall, 239 tagSNPs from the 35 genes were successfully genotyped (Table S3). The genotype concordance was >99%, suggesting high genotyping accuracy. The Cochran-Armitage trend test was performed for each of the 239 SNPs. One approach for assessing the departure of the distribution of the (Cochran-Armitage) test statistics from the (global) null distribution (no SNPs associated) has been described by Tyrer et. al. [9]. The approach is based upon fitting a mixture model to the distribution of the test statistics, with two components, one representing SNPs which are independent of the case-control status, the other representing SNPs associated with case-control status. The Cochran-Armitage test statistics for the associated SNPs are assumed to all have the same (chi-squared) non-centrality parameter value. The distributed software for the “admixture maximum likelihood” (AML) test of Tyrer et. al. [9] calculates empirical p-values based on a “pseudo-likelihood ratio” test, comparing the ratio of values of the optimized likelihoods under the null and alternative hypotheses for the observed data, with the corresponding values obtained from a large number of data sets with case-control status permuted randomly. It also provides an estimate of the non-centrality parameter which is a measure of the common effect size of the associated SNPs within the pathway. We performed the AML-based global test of association for the full metabolic pathway as well as for 3 sub-pathways (see results section). In addition, we performed gene-specific analyses, using the AML-based global test on SNPs within genes, within the androgen-estrogen conversion sub-pathway. We also carried out AML tests adjusted for a non-genetic risk factor using software provided by the authors of Tyrer et al. [9]. | Estrogen exposure is the most important risk factor for breast and endometrial cancers. Genetic variation of the genes involved in estrogen metabolism has, however, not been consistently associated with these two cancers. We posited that the genetic risk associated with the estrogen metabolic genes is likely to be carried by multiple variants and is therefore most effectively detected by multi-variant analysis. We carried out a comprehensive association analysis of the estrogen metabolic pathway by interrogating SNPs within 35 genes of the pathway in three tumor samples from Sweden and Finland. Through pathway-based multi-variant association analysis, we showed that the genetic variation within the estrogen metabolic pathway is associated with risk for breast and endometrial cancers and that the genetic variation within the genes involved in androgen-to-estrogen conversion is particularly important for the development of ER–positive and sporadic breast tumors in postmenopausal women. Our study has demonstrated that the influence of genetic variation on hormone exposure has an impact on breast cancer development, especially on the development of hormone-driven breast tumor subtypes. Our study has also highlighted that future genetic studies of the estrogen metabolic genes should focus on the androgen-to-estrogen conversion process. | Abstract
Introduction
Results
Discussion
Materials and Methods | oncology/breast cancer
oncology/gynecological cancers
genetics and genomics/genetics of disease
genetics and genomics/cancer genetics
genetics and genomics/population genetics | 2010 | Multi-Variant Pathway Association Analysis Reveals the Importance of Genetic Determinants of Estrogen Metabolism in Breast and Endometrial Cancer Susceptibility | 5,791 | 263 |
Macrophage migration inhibitory factor (MIF) has emerged as a pivotal mediator of innate immunity and has been shown to be an important effector molecule in severe sepsis. Melioidosis, caused by Burkholderia pseudomallei, is an important cause of community-acquired sepsis in Southeast-Asia. We aimed to characterize the expression and function of MIF in melioidosis. MIF expression was determined in leukocytes and plasma from 34 melioidosis patients and 32 controls, and in mice infected with B. pseudomallei. MIF function was investigated in experimental murine melioidosis using anti-MIF antibodies and recombinant MIF. Patients demonstrated markedly increased MIF mRNA leukocyte and MIF plasma concentrations. Elevated MIF concentrations were associated with mortality. Mice inoculated intranasally with B. pseudomallei displayed a robust increase in pulmonary and systemic MIF expression. Anti-MIF treated mice showed lower bacterial loads in their lungs upon infection with a low inoculum. Conversely, mice treated with recombinant MIF displayed a modestly impaired clearance of B. pseudomallei. MIF exerted no direct effects on bacterial outgrowth or phagocytosis of B. pseudomallei. MIF concentrations are markedly elevated during clinical melioidosis and correlate with patients' outcomes. In experimental melioidosis MIF impaired antibacterial defense.
Macrophage migration inhibitory factor (MIF) was one of the first cytokines to be discovered almost half a century ago [1]–[4]. Since then MIF has emerged as a pivotal mediator of innate immunity in various inflammatory diseases such as rheumatoid arthritis and atherosclerosis [5], [6] and is considered to be an integral component of the host antimicrobial alarm system [4], [7]. MIF, a classical proinflammatory cytokine, is constitutively expressed by many tissues with environmental contact such as the lung and the gastrointestinal tract, and by numerous cell types, among others T- and B-lymphocytes, monocytes and macrophages [4]. MIF-deficient macrophages are hyporesponsive to lipopolysaccharide (LPS) due to a down-regulation of Toll-like receptor (TLR) -4 [8], [9]. In line, MIF knockout mice were resistant to LPS induced toxic shock [8]–[10]. Recently it was shown that blood concentrations of MIF are elevated in patients with sepsis and able to predict early mortality [11]–[14]. Similarly, MIF is increased in patients with meningococcal disease and highest in the presence of shock [15]. Excitingly, treatment with anti-MIF antibodies protected mice from lethal peritonitis induced by Escherichia coli or cecal ligation and puncture (CLP) [16]. Furthermore, ISO-1 and OXIM-11, new small molecule inhibitors of MIF, offered significant protection to mice from CLP-induced sepsis [17], [18]. These data identified MIF as a potential mediator of lethality following abdominal sepsis. In Southeast-Asia and Northern-Australia the gram-negative bacillus Burkholderia pseudomallei is an important cause of community-acquired sepsis [19], [20]. More than half of these cases of melioidosis, as this severe infection is named, habitually presents with pneumonia, frequently associated with bacterial dissemination to distant sites [19]–[21]. In the present study we aimed to characterize the expression and function of MIF in melioidosis. For this we analysed MIF expression patterns in patients with melioidosis and in a mouse model of B. pseudomallei infection. MIF function was investigated in experimental murine melioidosis using anti-MIF antibodies and recombinant MIF.
The patient study was approved by both the Ministry of Public Health, Royal Government of Thailand and the Oxford Tropical Research Ethics Committee, University of Oxford, England. We obtained written informed consent from all subjects before the study. The Animal Care and Use of Committee of the University of Amsterdam approved all murine experiments. We included 34 individuals with sepsis caused by B. pseudomallei and 32 healthy controls in this study. Individuals were recruited prospectively at Sapprasithiprasong Hospital, Ubon Ratchathani, Thailand in 2004. Sepsis due to melioidosis was defined as culture positivity for B. pseudomallei from any clinical sample plus a systemic inflammatory response syndrome (SIRS) [22]. Study design and subjects have been described in detail [23]. Human MIF was measured by ELISA, as described elsewhere [24]. In addition, MIF mRNA levels were measured as follows. Heparin blood samples were drawn from an antecubital vein and immediately put on ice. Leukocytes were isolated using erylysis buffer, dissolved in Trizol and stored at –80°C. Thereafter, RNA was isolated and analyzed by multiplex ligation-dependent probe amplification (MLPA) as described [25], [26] (MRC-Holland, Amsterdam, the Netherlands). Levels of mRNA were expressed as a normalized ratio of the peak area divided by the peak area of the β2 microglobulin (B2M) gene [25]. Male C57BL/6 mice (age 8–10 weeks) were purchased from Harlan Sprague Dawley Inc. (Horst, The Netherlands). Age-matched animals were used in each experiment. For the inoculum, B. pseudomallei strain 1026b, kindly provided by Dr. Don Woods [27], [28], was used and prepared as described [23], [29]–[31]. Pneumonia was induced by intranasal inoculation of a 50 µl (5×101,2. 5×102 or 7. 5×102 colony forming units (CFU) /50 µl) bacterial suspension. 48 hours after infection, mice were anesthetized and sacrificed by bleeding from the vena cava inferior [23], [30], [32]. CFUs were determined from serial dilutions of organ homogenates as described [23], [29]–[31]. In some experiments mice were injected intraperitoneally with 2 mg of anti-MIF or non-immune IgG 2 hours before bacterial inoculation or with 50 µg recombinant mouse MIF or control buffer at the onset of infection as described previously [16], [33], [34]. Rabbit polyclonal anti-MIF and recombinant MIF were generated as described [16], [34]. The ELISA for mMIF developed according to the 4-span approach was used as described in detail [35]. Tumor necrosis factor (TNF) -α, interferon (IFN) -γ, interleukin (IL) -6, IL-10 and IL-12p70 were determined using a cytometric bead array (CBA) multiplex assay in accordance with the manufacturer' s instructions (BD Biosciences, San Jose, CA). Four-µm thick lung tissue sections were sampled 48 hours after infection and mounted on aminopropylmethoxysilane-coated glass slides, deparaffinized in xylol, taken through to absolute alcohol and blocked for endogenous peroxidase with 0. 1% hydrogen peroxide in methanol. They were boiled in 10 mM citrate buffer in a microwave oven and rinsed in Tris-buffered saline (TBS). To reduce non-specific binding, sections were incubated in normal goat serum (Pel-Freez Biologicals, Rogers, AK) 1∶30 in TBS. After 40-minutes incubation with polyclonal rabbit anti-MIF purified IgG diluted 1∶200 in TBS containing 2% bovine serum albumin (final immunoglobulin concentration: 25 mg/l), the sections were incubated with biotinylated goat anti-rabbit IgG (Vector, Burlingame, CA) diluted 1∶400 and then with ABC-peroxidase complex solution (Vector). Peroxidase activity was revealed with 5-5′-diaminobenzidine as chromogen and the sections were counterstained in Meyer' s acid-free hematoxylin. As a negative control, the primary antibody was replaced by pre-immune rabbit purified IgG. Furthermore, to score inflammation, lung and livers from infected mice were harvested 48 hours after infection, fixed in 10%-formalin and embedded in paraffin. Four µm sections were stained with hematoxylin and eosin and analyzed by a pathologist blinded for groups exactly as described previously [23]. B. pseudomallei strain 1026b was used and prepared as described above. In short, B. pseudomallei at concentrations from 3×103–3×106 CFU/ml was grown in the presence of recombinant MIF (dose range from 5 to 50 µg/ml) diluted in LB-growth medium. Phagocytosis was evaluated as described [36], [37]. Heat-killed B. pseudomallei was labeled with carboxyfluorescein-diacetate-succinimidyl-ester (CFSE dye, Invitrogen, Breda, The Netherlands). Peritoneal macrophages (derived from 5 different mice per group) were incubated with CFSE-labeled B. pseudomallei (2. 5×107 CFU/ml) for 0,60 and 120 minutes. Phagocytosis was stopped by placing cells on ice; thereafter cells were washed in PBS and suspended in Quenching solution (Orpegen, Heidelberg, Germany). To determine the neutrophil phagocytosis capacity, 50 µl of whole blood was incubated with bacteria after which cells were suspended in Quenching solution, incubated in FACS lysis/fix solution (BecktonDickinson) and neutrophils were labeled using anti-Gr-1-PE (Pharmingen). Phagocytosis was determined using FACS. Values are expressed as means ± standard error of the mean (SEM). Differences between groups were analyzed by Mann-Whitney U test or Kruskal-Wallis analysis with Dunn' s posthoc test where appropriate. For survival analysis, Kaplan-Meier analysis followed by log rank test was performed. These analyses were performed using GraphPad Prism version 4. 00, GraphPad Software (San Diego, CA). Values of P<0. 05 were considered statistically significant.
To obtain an insight into MIF expression during melioidosis, we first measured MIF in plasma from 34 patients with culture proven B. pseudomallei infection and in plasma from 32 local healthy controls. The mortality rate in this cohort of patients was 44%. MIF was markedly elevated in melioidosis patients with mean plasma concentrations that were approximately 2-fold higher than in those of healthy subjects (Figure 1A, P<0. 01). Plasma concentrations of MIF were associated with an adverse outcome: on admission patients who went on to had higher MIF concentrations than those who survived (Figure 1B, P<0. 01). In line, MIF mRNA levels were significantly higher in peripheral blood leukocytes from patients than in leukocytes from healthy controls (Figure 1C, P<0. 001). Since the majority of severe melioidosis cases presents with pneumonia with bacterial dissemination to distant body sites [19]–[21] and considering the fact that it is not feasible to study MIF expression at tissue level in patients with melioidosis, we used a well-established murine model of pneumonia-derived melioidosis in which mice are intranasally infected with B. pseudomallei [23], [29], [31]. In agreement with the data obtained in patients with melioidosis, infected mice showed an abundant upregulation of MIF expression, both in the pulmonary and systemic compartment (Figure 2, both P<0. 01). Immunohistochemical staining of lung tissue was performed to further identify the distribution of MIF expression during melioidosis. Positive immunostaining for MIF was observed in untreated control animals in alveolar macrophages and within the bronchial epithelium (Figure 3A). Granulocytes did not stain positive for MIF. After infection with B. pseudomallei there was a marked increase in immunostaining of the epithelial submucosa, bronchial epithelial cells and inflammatory cells, most notably of alveolar macrophages (Figure 3B). To obtain a first insight into the function of MIF during experimental melioidosis, we treated mice infected with 2. 5×102 CFU B. pseudomallei mice with 50 µg recombinant MIF using a dose similar to that used previously in an experimental septic shock model [16], [33]. Treatment of mice with recombinant MIF at the time of infection resulted in increased MIF concentrations in lung homogenates 48 hours later (from 33±1. 3 to 729±56. 6 ng/ml; P<0. 001). Mice were sacrificed 48 hours after inoculation to determine bacterial loads in lungs (the primary site of the infection), liver and blood (to evaluate to which extent the infection disseminated to distant body sites) (Figure 4). Relative to infected but non-treated controls, mice treated with recombinant MIF displayed almost 10-fold higher bacterial loads in the liver (Figure 4, P<0. 01). In addition, a clear trend was seen towards higher bacterial loads in the pulmonary and systemic compartments of recombinant MIF treated mice, although the differences with control mice did not reach statistical significance (Figure 4). Having found that administration of supra physiological doses of MIF results in a partially impaired bacterial clearance during experimental melioidosis, we next hypothesized that treatment with anti-MIF antibodies would result in decreased bacterial outgrowth and performed the reverse experiment by examining the effect of anti-MIF treatment. Therefore, before inoculating mice with B. pseudomallei, we injected mice with anti-MIF antibodies using a dosing schedule previously found be protective in a mouse model of E. coli or CLP-induced peritonitis [16], [33]. To evaluate whether anti-MIF treatment interferes with bacterial clearance, we first determined bacterial loads 48 hours after infection with an inoculum of 2. 5×102 CFU B. pseudomallei (Figure 5). At this dose no significant differences in bacterial outgrowth in either lungs, liver or blood were observed. To determine whether the effect of anti-MIF therapy is dependent on the size of the infectious dose, we next infected mice with a higher (5×102 CFU B. pseudomallei) and lower (5×101 CFU B. pseudomallei) inoculum (Figure 6). At the highest dose no effect of anti-MIF treatment was seen on the bacterial outgrowth in the lungs, liver or blood (Figure 6B). However, at the lowest inoculum mice treated with anti-MIF had almost 10-fold less B. pseudomallei CFU in their lungs compared to control mice (Figure 6A, P<0. 05). With this low inoculum, none of the mice showed positive Burkholderia cultures in liver or blood, suggesting that anti-MIF treatment inhibits the growth of B. pseudomallei in the lungs after infection with a relatively low bacterial dose. Lastly, we performed a survival experiment in which mice were injected intraperitoneally with 2 mg of anti-MIF or non-immune control IgG 2 hours before intranasal inoculation with B. pseudomallei. In accordance with the modest protective effect of anti-MIF treatment on bacterial outgrowth a limited survival advantage was seen in the anti-MIF treated group (Figure 7). Since cytokines are important regulators of the inflammatory response to acute lower respiratory tract infection [38] and given the observation that protective anti-MIF treatment reduced TNFα concentrations in mouse model of sepsis induced by E. coli or CLP [16], [33], we measured the concentrations of TNFα, IL-6, IL-10, IL-12 and IFNγ in lung homogenates and plasma obtained 48 after infection with 2. 5×102 CFU B. pseudomallei (Table 1). Anti-MIF treatment did not influence pulmonary cytokine concentrations in our model of experimental melioidosis (Table 1). Previously it was shown that plasma TNFα concentrations induced by LPS were lower in MIF-deficient mice compared to wild-type mice [10], [33]. However also in plasma no differences in TNFα or IL-6, IL-10, IL-12 and IFNγ concentrations were seen between anti-MIF treated and control mice after inoculation with B. pseudomallei (data not shown). In addition treatment with 50 µg recombinant MIF did not influence cytokine concentrations in either the pulmonary (Table 1) or systemic compartment (data not shown). Considering that MIF is regarded as an important proinflammatory mediator, we determined whether modulation of MIF concentrations could have an effect on organ inflammation during experimental melioidosis. Therefore, we performed histopathological analyses of lung and liver tissues in control mice and mice treated with anti-MIF or recombinant MIF and infected with B. pseudomallei. Although all mice showed evidence of inflammation as characterized by diffuse infiltrates, interstitial inflammation and bronchitis there were no differences in total organ histopathological scores between groups (data not shown). Having found that anti-MIF treated mice showed lower bacterial loads in their lungs upon infection with a low inoculum while mice treated with recombinant MIF displayed a modestly impaired clearance of B. pseudomallei, we next wished to determine whether MIF has a direct effect on bacterial outgrowth and/or phagocytosis. Therefore, B. pseudomallei (at concentrations from 3×103–3×106 CFU/ml) was grown in the presence of recombinant MIF diluted in the growth medium (dose range from 5 to 50 µg/ml). However, no effects of recombinant MIF on bacterial outgrowth could be observed at any time point (up to 2 hours; data not shown). Lastly, we studied whether MIF contributes to phagocytosis of B. pseudomallei. However no effects of anti-MIF treatment on phagocytosis of B. pseudomallei by peritoneal macrophages or whole blood neutrophils could be observed (data not shown).
In the present study we aimed to characterize the expression and role for MIF in melioidosis, linking observational studies in patients with culture-proven disease with functional studies in mice in which we modulated the concentrations of MIF during experimentally induced melioidosis. Our study shows that patients with severe melioidosis have strongly increased MIF plasma and MIF mRNA leukocyte levels. High plasma MIF concentrations were associated with mortality. Similarly, mice intranasally inoculated with B. pseudomallei displayed a strong increase in pulmonary and systemic MIF expression. The functional role of MIF in our model of experimental melioidosis however was modest given the fact that modulation of MIF levels only moderately influenced the innate immune response towards B. pseudomallei. Anti-MIF treatment resulted in a modest survival benefit. Anti-MIF treatment only decreased bacterial outgrowth when mice were inoculated with a low dose of B. pseudomallei whereas - conversely - mice treated with recombinant MIF displayed a modestly impaired clearance of B. pseudomallei. These data are the first to report on the expression and function of MIF during melioidosis. MIF expression is increased in a wide variety of infectious diseases, ranging from viral infections, such as Dengue, HIV and West Nile virus infection [39]–[41], malaria [42], tuberculosis [43] and various forms of sepsis [11], [12], [15]. Our study further extends these findings by demonstrating increased plasma and blood leukocyte mRNA levels of MIF in patients with severe melioidosis. Importantly, we demonstrated a strong association between elevated MIF levels and increased mortality. This is in line with a recent study among pediatric and adult patients with severe sepsis or septic shock caused predominantly by Neisseria meningitides and other gram-negative bacteria in which elevated MIF levels were shown to be predictive of early mortality [12]. MIF, however, is not always upregulated after acute infection or inflammation. For instance, in children with acute malaria circulating MIF levels were significantly lower compared with healthy, malaria-exposed children [44]. Furthermore, MIF release could not be detected in a human endotoxemia model and is not produced by whole blood cells incubated with LPS [15]. Also in HIV seropositive patients low serum MIF levels were associated with a high 1-month mortality [41]. This further highlights the potential diverse roles MIF can play in the host response against various invading pathogens. Melioidosis, which is the most common form of community-acquired sepsis in Northern-Australia and Eastern-Thailand, is associated with a mortality of up to 50% in endemic areas [19], [20]. Severe pneumonia with bacterial dissemination to distant body sites is a common presentation of melioidosis [20], [21]. Sepsis caused by B. pseudomallei is characterized by a markedly proinflammatory cytokine profile; in the current cohort of patients we have demonstrated increased plasma concentrations of IL-6, IL-8 and IL-18 when compared to controls [31], [45]. In addition, high throughput mRNA profiling in these patients suffering from severe melioidosis furthermore demonstrated increased transcription of a whole array of proinflammatory genes in whole blood leukocytes [26]. In light of the proinflammatory properties attributed to MIF in sepsis, we studied the expression and function of MIF in a well-established mouse model of melioidosis [23], [29], [31]. In line with our patient data and in line with various other murine models of sepsis induced by LPS, E. coli or CLP [4], [16], [46], we observed a strong upregulation of MIF expression in both the lungs and blood of mice inoculated with B. pseudomallei. However, MIF seems to play a less important role in the innate immune response in melioidosis, which is in contrast with previous studies pointing towards a central role of MIF in other forms of infection. With regard to bacterial infection, the role of MIF has been first studied in abdominal sepsis caused by either intraperitoneal injection of E. coli or CLP [9], [16]. In these models, anti-MIF from the same source and administered in the exact same dose protected mice from mortality, reduced TNFα concentrations and diminished bacterial growth. Very recently it was shown that polymorphisms associated with higher MIF expression may have a beneficial effect in community-acquired pneumonia [47]. In addition, modulation of MIF may have therapeutic advantages in treating acute lung injury in patients with acute pancreatitis complicated by bacterial infection [48]. The fact that anti-MIF only has a minor impact on the immune response to B. pseudomallei could be related to differences in the primary site of infection and/or differences in the pathogens involved [9], [16]. In this respect it is worthwhile noting that MIF regulates innate immune responses in gram-negative infections through modulation of Toll-like receptor 4 [8]. We obtained recent evidence - counter intuitively for a gram-negative infection - that TLR2 impacts on the immune response of the intact host in vivo, whereas TLR4 does not contribute to protective immunity in melioidosis [23]. As such, the minor role of TLR4 in the innate immune response towards B. pseudomallei could be an explanation for our present findings revealing an equally limited role for MIF in melioidosis. Interestingly, during murine Listeria monocytogenes infection, the elimination of bacteria from the spleen and liver was not affected by anti-MIF antibody although this treatment was able to rescue mice from lethal infection [49]. In reverse experiments we found that treatment of B. pseudomallei infected mice with recombinant MIF caused impairment of the bacterial clearance capability. Earlier studies showed that recombinant MIF increased mortality during E. coli sepsis when co-injected with bacteria in mice [16], [33], [50]. In these investigations the effect of recombinant MIF on bacterial loads was not reported. These findings imply that increased concentrations of MIF can be harmful in the acute host response against invading bacteria. In this respect it is of interest that during the immune suppressed state which occurs in the late phase of the septic response and which is characterized by a reduced capacity of immune cells to produce proinflammatory cytokines such as TNFα, it was shown that treatment with recombinant MIF could protect animals from bacterial superinfection in a mouse model of CLP-induced peritonitis [51]. This further highlights the potential diverse nature of MIF function during the course of sepsis. Our study has several limitations. Our observations were done in patients with sepsis caused by B. pseudomallei and caution is required when extending these findings to less severe or chronic melioidosis, since we focused on the early acute phase of melioidosis. Furthermore, although our in vivo model of melioidosis has been important in elucidating the role of other inflammatory mediators in melioidosis [23], [29], [31], data obtained from a mouse model by definition should be extrapolated to patients with melioidosis with great caution. In addition, it would be of interest to confirm our results in MIF knockout mice, although we consider it less likely that the use of these mice will yield strongly different data in light of the modest differences observed in the different treatment groups. Lastly, obtaining new biological insights from studies using antibodies and recombinant proteins of interest remains a challenge, limited by the notion of considerable cooperation between inflammatory factors involved and extensive redundancy in the host response against invading pathogens [52]. In conclusion, MIF concentrations are markedly increased during melioidosis, and elevated levels correlate with mortality. Although mice with experimentally induced melioidosis showed strongly upregulated expression of MIF in lungs and blood, inhibition of MIF with a specific antibody only modestly influenced the host response. Similarly, administration of recombinant MIF did not strongly impact on the immune response to B. pseudomallei infection. These data argue against an important role for MIF in the pathogenesis of melioidosis. | Melioidosis is a severe tropical infection caused by the bacterium Burkholderia pseudomallei. B. pseudomallei is the major cause of community-acquired septicemia in northeast Thailand with a mortality rate in severe cases of around 40% Little is known, however, about the mechanisms of the host defense to B. pseudomallei infection. Macrophage migration inhibitory factor (MIF) is a proinflammatory cytokine that has emerged as an important mediator of the host defense in severe bacterial infections. In this article, we studied the expression and function of MIF both in patients with melioidosis and in mice during experimental melioidosis. We found that MIF concentrations were elevated in patients with melioidosis. Furthermore, high MIF concentrations are associated with poor outcome in patients with melioidosis. Also, in mice with experimentally induced melioidosis, we observed an upregulation of MIF concentrations. Furthermore, mice with melioidosis that were treated with a MIF blocking treatment showed lower bacterial counts in their lungs during infection. In conclusion, MIF seems to impair host defense mechanisms during melioidosis. | Abstract
Introduction
Methods
Results
Discussion | immunology/cellular microbiology and pathogenesis
respiratory medicine/respiratory infections
infectious diseases/neglected tropical diseases
microbiology/innate immunity
pathology/histopathology
immunology/innate immunity
critical care and emergency medicine/sepsis and multiple organ failure
pathology/immunology
immunology/immunity to infections | 2010 | Expression and Function of Macrophage Migration Inhibitory Factor (MIF) in Melioidosis | 6,418 | 275 |
At present, screening of the population at risk for gambiense human African trypanosomiasis (HAT) is based on detection of antibodies against native variant surface glycoproteins (VSGs) of Trypanosoma brucei (T. b.) gambiense. Drawbacks of these native VSGs include culture of infective T. b. gambiense trypanosomes in laboratory rodents, necessary for production, and the exposure of non-specific epitopes that may cause cross-reactions. We therefore aimed at identifying peptides that mimic epitopes, hence called “mimotopes, ” specific to T. b. gambiense VSGs and that may replace the native proteins in antibody detection tests. A Ph. D. -12 peptide phage display library was screened with polyclonal antibodies from patient sera, previously affinity purified on VSG LiTat 1. 3 or LiTat 1. 5. The peptide sequences were derived from the DNA sequence of the selected phages and synthesised as biotinylated peptides. Respectively, eighteen and twenty different mimotopes were identified for VSG LiTat 1. 3 and LiTat 1. 5, of which six and five were retained for assessment of their diagnostic performance. Based on alignment of the peptide sequences on the original protein sequence of VSG LiTat 1. 3 and 1. 5, three additional peptides were synthesised. We evaluated the diagnostic performance of the synthetic peptides in indirect ELISA with 102 sera from HAT patients and 102 endemic negative controls. All mimotopes had areas under the curve (AUCs) of ≥0. 85, indicating their diagnostic potential. One peptide corresponding to the VSG LiTat 1. 3 protein sequence also had an AUC of ≥0. 85, while the peptide based on the sequence of VSG LiTat 1. 5 had an AUC of only 0. 79. We delivered the proof of principle that mimotopes for T. b. gambiense VSGs, with diagnostic potential, can be selected by phage display using polyclonal human antibodies.
The chronic form of sleeping sickness or human African trypanosomiasis (HAT) in West and Central Africa is caused by the protozoan parasite Trypanosoma brucei (T. b.) gambiense while T. b. rhodesiense causes a more fulminant, acute form in East and Southern Africa. Both subspecies of T. brucei are cyclically transmitted by tsetse flies of the genus Glossina and mainly affect poor, rural populations. The true burden of this disease is unknown as many cases remain undiagnosed or unreported [1], [2]. Since untreated HAT is almost always fatal and no inexpensive, safe and easily administered drugs are available, accurate case detection is crucial. Parasite detection is laborious and insensitive, and remains therefore limited to disease suspects. In the absence of reliable clinical symptoms or antigen detection tests, HAT suspects are identified through screening of the population at risk for presence of trypanosome specific antibodies. The commonly used antibody detection tests, card agglutination test for trypanosomiasis (CATT) [3], LATEX/T. b. gambiense and ELISA/T. b. gambiense [4], [5] detect antibodies against the highly immunogenic variant surface glycoproteins (VSGs) of T. b. gambiense. Even though the genome of T. brucei contains >1000 VSG genes, only one variable antigen type (VAT) is expressed at a time. Stochastic switching of VSG allows the trypanosome to evade the specific antibody responses that were raised against earlier VATs [6]–[10]. Some VATs, such as LiTat 1. 3 and 1. 5, are recognised by almost all gambiense HAT patients and therefore called predominant. The dense VSG monolayer on the living trypanosome shields all non-specific epitopes. The hypervariable N-terminal VSG domain (300–400 residues) is exposed to the immune system and comprises the VAT-specific epitopes, while the relatively conserved C-terminal domain (40–80 residues) is hidden by the intact VSG coat [6], [9], [11], [12]. Disadvantages of the present antibody detection tests include the occurrence of non-specific reactions. This might be explained by exposure of non-HAT-specific epitopes that are normally shielded on the living trypanosome [12], [13]. In addition, diagnostic test production actually requires culture of infective T. b. gambiense in large numbers of laboratory rodents and poses an important risk of infection to the manufacturing staff [14]. These drawbacks can be circumvented through the use of synthetic peptides that mimic HAT-specific VSG epitopes (mimotopes) and can be produced in a standardised way [15]. One way to identify such mimotopes is by peptide phage display. This technique is based on DNA recombination resulting in foreign peptides with random sequences that are displayed fused to the pIII surface protein of the M13 phage. After an in vitro selection process based on binding affinity and several rounds of enrichment (panning), the encoded peptide insert sequence of the selected phage is deduced from the phage DNA. We previously reported successful identification of mimotopes for VSG LiTat 1. 3 and LiTat 1. 5 by performing phage display with three monoclonal antibodies [16]. However, by the use of only three monoclonal antibodies, representing only a fraction of the VSG-specific antibody response, some mimotopes with diagnostic potential might have been missed. Additionally, the mouse and human immune system may recognise different B cell epitopes. The use of polyclonal human antibodies might therefore increase chances of selecting diagnostic mimotopes [17]. Polyclonal antibodies from human sera have been previously used for selection of mimotopes with diagnostic potential for e. g. hepatitis C [15], typhoid fever [18] and Epstein Barr virus [17]. Some mimotopes have been patented for incorporation in commercially available tests, e. g. for neurocysticercosis [19]. In this manuscript we describe the identification of mimotopes for VSG LiTat 1. 3 and LiTat 1. 5 through phage display, using sera from HAT patients and endemic negative persons.
Sera from HAT patients and endemic controls were collected within different diagnostic studies [5], [20]. All individuals gave their written informed consent before providing blood. Permission for these studies was obtained from the national ethical committee of the Democratic Republic of the Congo (DR Congo) and from the Institute of Tropical Medicine Antwerp (ITMA) ethical committee, reference number 03 07 1 413 and 04 44 1 472. Forty additional endemic negative control specimens were obtained from the archived specimen bank of the Parasite Diagnostics Unit at ITMA. All specimens were anonymised. Variant surface glycoproteins were purified from cloned populations of T. b. gambiense Variable Antigen Type (VAT) LiTat 1. 3 and 1. 5 [4]. VSG LiTat 1. 3 or LiTat 1. 5 were coated onto magnetic particles (MP, Estapor, 10% suspension, 1. 04 µm, 9 µeq/g COOH). A volume of 250 µL of MP suspension was washed twice with 1 mL of buffer A (10 mmol/L NaH2PO4, pH 6. 0). The MP were activated with 2. 5 ml of buffer A containing 25 mg of 1-ethyl-3- (3-dimethylaminopropyl) carbodiimide (Pierce) and 15 mg of N-hydroxysuccinimide (Sigma). The MP were rotated for 15 minutes at room temperature (rT) and washed with 1 mL of buffer B (2 mmol/L HCl) where after 350 µg of VSG (LiTat 1. 3 or LiTat 1. 5) in 1. 5 mL of buffer C (20 mmol/L NaH2PO4/Na2HPO4, pH 7. 5) was added to the pellet of MP. After rotating for 2 h at rT, the MP were washed three times with buffer C and resuspended to a final concentration of 8% in buffer C containing 100 mmol/L glycin, 1% bovine serum albumin (BSA) and 0. 1% NaN3. Successful coating of the MP was evaluated by agglutination with a HAT positive serum diluted 1/4 in phosphate buffered saline (PBS, 0. 01 mol/L phosphate, 0. 14 mol/L NaCl, pH 7. 4) containing 0. 02% w/v NaN3. Antibodies specific to VSG LiTat 1. 3 or LiTat 1. 5 were purified from nine HAT positive sera originating from the DR Congo [20]. One mL of LiTat 1. 3 or LiTat 1. 5 coated MP was rotated for 2 h at rT with 125 µL of HAT positive serum. After five washes with 800 µL of PBS, the specific antibodies were eluted from the MP by adding 700 µL of 0. 2 mol/L glycine/HCl (pH 2. 2) followed by magnetic separation after five minutes. The eluates, corresponding to the affinity purified antibody fractions, were neutralised with 100 µL of 1 mol/L Tris/HCl pH 9. 1. Indirect ELISA was used to screen the affinity purified antibody fractions and all human serum samples on reactivity with VSG LiTat 1. 3 and LiTat 1. 5. ELISA plates (Nunc MaxiSorp™) were coated overnight (ON) at 4°C with 100 µL/well of 2 µg/mL of each VSG separately in phosphate buffer (PB, 0. 01 mol/L phosphate, pH 6. 5) or with 1. 7 1011 particle/mL of wild type phage (WTP) in PBS. One plate was left empty as antigen negative control (Ag0). The plates were tapped dry, saturated with 350 µL/well of PBS-Blotto (0. 01 mol/L phosphate, 0. 2 mol/L NaCl, 1% w/v skimmed milk powder, 0. 05% w/v NaN3) during 1 h at rT and washed three times with 0. 05% v/v Tween-20 in PBS (PBST) (ELx50, Bio-Tek ELISA washer). The purified antibody fractions were diluted 1/25 and human serum samples 1/150 in PBS-Blotto. One hundred µL/well of each dilution was applied in duplicate and incubated for 1 h at rT. After three washes with PBST 100 µL/well of horse radish peroxidase (PO) -conjugated goat anti-human IgG (H+L) (Jackson), 1/40000 diluted in PBST, was added. An hour and five washes later, wells were incubated for 1 h at rT with 100 µL/well of 2. 2′-azino-bis- (3-ethylbenzthiazoline-6-sulfonic acid) (ABTS) chromogen/substrate solution (50 mg tablet/100 mL of ABTS buffer, Roche). The plate was shaken for ten seconds and the optical density (OD) was read at 414 nm (Labsystems Multiskan RC 351). The measured OD was corrected (ODc) with the corresponding OD in the Ag0 wells. Three LiTat 1. 3 positive pools, each consisting of three different VSG LiTat 1. 3-specific antibody fractions, three LiTat 1. 5 positive pools, each consisting of three different VSG LiTat 1. 5-specific antibody fractions and one negative pool of four endemic negative sera were prepared. For each pool the antibodies were coated onto anti-human IgG (H+L) functionalised magnetic particles (MP) (1% w/v, 1. 05 µm, Estapor/Merck) according to the guidelines of the manufacturer. The panning was performed with the Ph. D. -12 (12-mer) phage display library (New England Biolabs, NEB) [21] through two rounds consisting of 1) a positive selection with anti-VSG (LiTat 1. 3 or 1. 5, respectively) antibodies coated on MP, 2) a negative selection with endemic negative serum antibodies coated on MP and 3) phage amplification [22]. Each positive selection was followed by phage titration and sandwich ELISA. After these two rounds a third positive selection was performed. Positive selection was performed as previously described [16]. Bound phages were eluted for ten minutes with 600 µL of 0. 2 mol/L glycine-HCl containing 1 mg/mL BSA (pH 2. 2) and neutralised with 90 µL of Tris-HCl (1 mol/L, pH 9. 1). Six hundred µL of the elution from the positive selection was rotated ON at 4°C with 1 mg of MP coated with endemic negative serum antibodies, in a total volume of 1 mL of PBSG. The unbound phages in 900 µL of the supernatant of the negative selection were amplified, in a culture of Escherichia (E.) coli (strain ER2738, NEB) at early log (0. 01–0. 05 A600), and purified with PEG-NaCl as previously described [16], [21]. Phages from the first, second and third positive selection were diluted in PBS 101 to 104,102 to 105,104 to 107, respectively. Ten µL of these dilutions were incubated for five minutes with 200 µL of an E. coli culture in mid-log (0. 4–0. 5 A600). The mixture was then pipetted into 4 mL of Top-Agar (50°C) and poured onto agar plates containing 1 mL/L IPTG/X-gal (1. 25 g isopropyl β-D-thiogalactoside, 1 g 5-bromo-4-chloro-3-indolyl-β-D-galactoside, 25 mL dimethylformamide); ninety-four blue clones were picked and each clone was inoculated in 200 µL of lysogeny broth (LB) in a sterile culture plate (BD Falcon™ Clear 96-well Microtest™ Plate) [21]. This plate was shaken overnight at 30°C, and then the bacteria were pelleted by 5 min centrifugation at 1312 g. The supernatant was tested in a sandwich ELISA. ELISA plates were coated ON at 4°C with 100 µL/well of VSG LiTat 1. 3- or LiTat 1. 5-specific positive antibody pools (5 µg/mL in PBS) or a 1/10000 dilution in PBS of the negative serum pool. The ELISA was performed as previously described [16]. Briefly, the wells were incubated for 1 h at rT with 100 µL of phage dilution in PBS-Blotto (1/3 for culture plate supernatant or 1/20 for PEG-NaCl purified phage). PO-anti-M13 pVIII mAb (GE Healthcare), diluted 1/2000 in PBST was added to the wells for 1 h at rT. The wells were then incubated for 1 h at rT with ABTS and read at 414 nm. Phage clones were withheld after the first two positive selections if 1) the OD with the corresponding positive pool (ODpos) >average ODpos+2*standard deviation (sdpos) and 2) the OD with the negative pool (ODneg) <average ODneg. After the third positive selection, phages were sequenced if 1) ODpos>average ODpos+1* sdpos with at least one of the positive pools, 2) ODpos with the 3rd positive pool >0. 150 or 0. 200 for phages selected for VSG LiTat 1. 3 or LiTat 1. 5 respectively and 3) ODneg<average ODneg. Withheld phage clones were sequenced and tested in a similar sandwich ELISA with as capture antibody the nine individual affinity purified antibody fractions, diluted 1/70 in PBS. Purification of phage DNA was performed according to the NEB manual [21]. Sequence determination was performed as described before [16]. The obtained sequence chromatograms were read with Chromas 2. 33 (Technelysium Pty Ltd). Sequence alignment was performed manually and with RELIC software [23]. A protein data base (pdb) model of the N-terminal domain of VSG LiTat 1. 5, was created using SWISS-MODEL [24], [25]. Modelling was based on the known structure of VSG MITat 1. 2 (pdb 1vsgA), previously derived by X-ray crystallography [26]. For VSG LiTat 1. 3 however the server could not find a template with sufficient sequence homology, hence the pdb was created by Thomas Juetteman from the PyMol helpdesk (PyMOL Molecular Graphics System, Schrödinger, LLC). In order to identify possible conformational epitopes, the 3D-Epitope-Explorer (3DEX) [27] was used to find structural homology between the mimotope sequences and the respective VSG protein sequence. Molecular graphics images were produced using the UCSF Chimera package from the Resource for Biocomputing, Visualization, and Informatics at the University of California, San Francisco (supported by NIH P41 RR001081) (http: //www. cgl. ucsf. edu/chimera). The peptides were synthesised at >85% purity (Peptide 2. 0, Chantilly, VA, U. S.). The GGGS-spacer, separating the library insert and the pIII phage protein, was added to the C-terminus of the peptides that were selected by phage display [16], [21]. All peptides were C-terminally elongated with an additional lysine-biotin and amidated (-CONH2), mimicking the uncharged peptide bond in a protein. All synthetic peptides were reconstituted in sterile deionised H2O to a concentration of 2 mg/mL. First, the reactivity of all biotinylated synthetic peptides was evaluated with the nine sera used for affinity antibody purification and nine endemic negative controls. Second, the diagnostic performance of the synthetic peptides was evaluated with human serum samples that were previously screened (indirect ELISA on VSG, serum dilution 1/100) on reactivity with VSG LiTat 1. 3 and 1. 5. All 102 serum samples from gambiense HAT patients originated from DR Congo [20]. Of the 102 endemic gambiense HAT negative serum samples, 71 originated from the DR Congo and 31 from Benin. The indirect ELISA on biotinylated peptides was performed as the indirect ELISA on VSG but 150 µL/well was applied in all but the saturation and washing steps. ELISA plates were coated with 10 µg/mL streptavidin (NEB) in carbonate buffer (0. 1 mol/L, pH 9. 2) or with 2 µg/mL VSG LiTat 1. 3 and LiTat 1. 5 in PB, or wells were left empty (Ag0). After saturation with PBS-Blotto, the peptides were added at a concentration of 2 µg/mL in PBS to the wells containing streptavidin. The peptide-free wells received only PBS. To the VSG-containing and Ag0 wells PBS-5% w/v sucrose was added. After incubation of 1 h at rT the plates were tapped dry, sealed and frozen at −80°C. The serum samples were centrifuged for 5 min at 15700 g and diluted 1/100 in PBS-Blotto. After thawing of the plates and three washes with PBST, the serum dilutions were applied in duplicate. After one hour, we added PO-conjugated goat anti-human IgG (H+L), 1/40000 diluted in PBST. ABTS was used as chromogen/substrate solution and the OD was read as described above. The measured OD was corrected by subtracting the corresponding OD in the peptide-free or Ag0 wells and the average of the duplicate corrected ODs was taken (ODc). The accuracy of the synthetic peptides to detect VSG-specific antibodies for diagnosis of sleeping sickness was assessed by the area under the receiver operator characteristics (ROC) curve (AUC) [28]. Confidence intervals were calculated according to DeLong [29]. For the whole range of cut-offs the Youden index was determined (Youden index = sensitivity+specificity−1) [30] and the cut-off with maximal Youden index was retained.
In indirect ELISA, all affinity purified antibody fractions reacted specifically with their corresponding VSG and not with WTP. The antibody fractions that were purified with VSG LiTat 1. 3, had an average ODc of 0. 533±0. 319 with VSG LiTat 1. 3, and average ODcs of only 0. 032±0. 032 with VSG LiTat 1. 5 and −0. 008±0. 015 with WTP. The antibody fractions purified with VSG LiTat 1. 5 had an average ODc of 1. 406±0. 487 with VSG LiTat 1. 5, and average ODcs of only 0. 037±0. 064 with VSG LiTat 1. 3 and −0. 017±0. 018 with WTP. The negative serum samples did not react with VSG LiTat 1. 3 (ODc 0. 034±0. 078), nor with VSG LiTat 1. 5 (ODc 0. 028±0. 069), nor with WTP (ODc 0. 013±0. 029). During the selection process, none of 94 phage clones of the first positive selection, eight of 188 phage clones of the second positive selection and 11 of 188 phage clones of the third positive selection reacted in the sandwich ELISA and were sequenced, resulting in 18 sequences (table 1). The alignment results of VSG LiTat 1. 3 [GenBank AJ304413] and the eighteen peptide sequences displayed by the phage clones are presented in figure 1. All peptides could be aligned within amino acid stretch (AA) 72 to 116 of the N-terminal domain of VSG LiTat 1. 3 (alignment 1). The common motive (F/W) ExDxK (A/V/L) x (A/V/L) was repeated twice in this VSG AA stretch. Therefore twelve sequences could be aligned twice within this region. The peptide displayed by phage 3-3-F6, ETDNMKPLHLRQ, could even be aligned three times within this region of VSG LiTat 1. 3, having ETD, DNxKP and ExD identical within amino acids 78 to 80,87 to 91 and 102 to 104 of the protein sequence. The peptide sequence displayed by phage 3-2-D10 had the highest identity within AA 72 to 116 of the VSG LiTat 1. 3 sequence (7/16 AA, 44%). The reverse sequence of the peptides displayed by phage clones 3-2-C5 and 3-3-E3 showed respectively 31 and 13% identity within AA 180 to 196 (alignment 2). Within the C-terminal domain (alignment 3), the peptide expressed by phage clone 3-2-C5 and 3-3-E3 had 25% identity (4/16 AA) within AA 404 to 443 of VSG LiTat 1. 3. The peptide expressed by phage clones 3-2-G10,3-2-G5 and 3-2-B12 were respectively 31%, 19% and 19% identical within AA 404 to 443 of VSG LiTat 1. 3. All selected phage clones were tested in a sandwich ELISA with the individual purified antibody fractions. The peptides displayed by the seven phage clones with the highest average ODc were withheld for synthesis as biotinylated peptides (table 1). The peptide displayed by phage clone 3-2-B12 was not withheld, since it was a homologue of 3-2-G10 and 3-2-G5 but had a lower average ODc. Based on the alignment results, AA stretch 78 to 110 and AA stretch 424 to 439 of the protein sequence of VSG LiTat 1. 3 were also synthesised as biotinylated peptides (respectively peptide 1. 3/78-110 and peptide 1. 3/424-439). The reactivity of all nine biotinylated synthetic peptides was evaluated in indirect ELISA with the nine HAT positive sera used for affinity antibody purification, and with nine endemic negative controls. Peptide 1. 3/78-110 was the best performing peptide with ODc 1. 469. Peptide 1. 3/424-439 gave a lower average ODc (0. 246) than peptides 3-2-G10 and 3-2-G5 (0. 564 and 0. 920), sharing the same common motive, and was not withheld for further testing. Peptide 3-2-E2, a homologue of peptide 3-2-D10, also gave a lower average ODc (0. 541 versus 0. 763) and was also not withheld for testing on diagnostic performance. During the selection process, one of 94 phage clones of the first positive selection, two of 94 phage clones of the second positive selection and 20 of 188 phage clones of the third positive selection reacted in the sandwich ELISA and were sequenced, resulting in 20 sequences (table 2). The alignment results of VSG LiTat 1. 5 [GenBank HQ662603] and the 20 peptide sequences displayed by the phage clones are presented in figure 2. The peptide expressed by phage clone 5-1-F9 (19% identity) could be aligned within AA 33 to 47 (alignment 1). Within the N-terminal domain, 18 phage peptides could be aligned with minimum 6% identity within AA 81 to 119 (alignment 2), by analogy with alignment 1 for VSG LiTat 1. 3. The peptides expressed by clones 5-3-C7,5-3-A8 and 5-3-B5 respectively had 0,0 and 6% identity within this AA stretch, but shared the common motive (W/F) Y with the peptide expressed by phage 5-3-B8 (19% identity). The peptide of phage clone 5-3-G6 had only 1/16 AA (6%) identity with VSG LiTat 1. 5 but two more AA were homologous within this region. The reverse sequences of phage clone peptides 5-3-F7 (13% identity), 5-1-F9 (19% identity) and 5-3-D5 (13% identity) could also be aligned within this VSG LiTat 1. 5 region. The peptide expressed by clone 5-3-B9 had the highest identity within the AA 81 to 119 stretch (5/16 AA, 31% identity, if a gap of 1 AA was allowed). Peptides expressed by phage clones 5-3-C1,5-3-A4 and 5-3-A6, with common motive “KLANP”, could also be aligned between AA 145 to 166 of the VSG LiTat 1. 5 protein sequence (alignment 3) with respectively 25,13 and 13% identity. Within the VSG LiTat 1. 5 AA stretch 245 to 281, the peptide expressed by phage clone 5-3-A6, showed 19% identity and the reverse peptide sequence of phages 5-3-B9,5-3-A3 and 5-3-A4 showed respectively 19,19 and 13% identity (alignment 4). Within the boundary with the C-terminal domain of VSG LiTat 1. 5, showed the peptides expressed by phage clones 5-3-A8,5-3-C1 and 5-3-B6 respectively 19,31 and 31% identity within AA stretch 341 to 368, if a gap of three AA was allowed for peptide 5-3-C1 (alignment 5). Within the C-terminal domain of VSG LiTat 1. 5, phage clone peptide 5-3-B9 and 5-3-A4 had respectively 31 and 13% identity between AA 468 to 489 (alignment 6). All selected phage clones were tested in a sandwich ELISA with the individual purified antibody fractions (table 2). The peptides displayed by the seven phage clones with the highest average ODc, were chosen for synthesis as biotinylated peptides, except for the peptide displayed by phage clone 5-3-A6, which was similar to 5-3-C1 but had a lower average ODc. Based on the alignment results and by analogy with VSG LiTat 1. 3, the AA stretch 81 to 109 of the protein sequence of VSG LiTat 1. 5 was synthesised as biotinylated peptide (peptide 1. 5/81-109). The reactivity of all eight biotinylated synthetic peptides was evaluated in indirect ELISA with the nine sera used for affinity antibody purification and with nine endemic negative controls. Peptide 5-3-A4 and 5-3-G6 had the lowest average ODcs (0. 145 and 0. 109) and shared a common motive with peptide 5-3-C1 with a higher average ODc (0. 289) and were therefore not withheld for testing on diagnostic performance. Peptide 1. 5/81-109 had an ODc of 0. 382 and was withheld. The accuracy of the biotinylated peptides to detect VSG-specific antibodies was assessed with sera from 102 gambiense HAT patients and 102 endemic negative controls (table 3). Among the mimotopes for VSG LiTat 1. 3, the highest AUC was obtained with peptide 3-2-G5 (0. 93) and peptide 3-2-G10 (0. 95). Sensitivities and specificities at the cut-off with the highest Youden index were respectively 0. 85 and 0. 94 for peptide 3-2-G5, and 0. 90 and 0. 93 for peptide 3-2-G10. Of the mimotopes for VSG LiTat 1. 5 the highest AUC was obtained with peptide 5-1-F9 (0. 95) and 5-2-D3 (0. 94) with respective sensitivities and specificities of 0. 94 and 0. 95 for peptide 5-1-F9 and 0. 92 and 0. 89 for peptide 5-2-D3. With peptide 1. 3/78-110, an AUC of 0. 95 was observed, with a sensitivity of 0. 96 and a specificity of 0. 85. With peptide 1. 5/81-109, an AUC of 0. 79, a sensitivity of 0. 81 and a specificity of 0. 75 were obtained. VSG LiTat 1. 3 and 1. 5 obtained an area under the curve of respectively 1. 000 and 0. 997. The sensitivity and specificity were both 1. 000 at cut-off 1. 318 for VSG LiTat 1. 3 and 1. 000 and 0. 990 at cut-off 1. 182 for VSG LiTat 1. 5. By using 3DEX software and setting the number of hits at a minimum of 5 AA, none of the VSG LiTat 1. 3 mimotopes with AUC>0. 90 could be mapped as a conformational epitope on the protein model of the VSG LiTat 1. 3. In contrast, among the VSG LiTat 1. 5 mimotopes with AUC>0. 90, peptide 5-2-D3 could be mapped with 8/12 AA (E 168|N 164|D 152|G 153|T 150|K 146|L 144|A 141) on the three-dimensional VSG LiTat 1. 5 protein model (figure 3).
In this manuscript we describe how mimotopes and regions that take part in epitope formation for VSGs LiTat 1. 3 and LiTat 1. 5 of T. b. gambiense were identified by screening of a Ph. D. -12 phage display library with polyclonal antibodies that were purified from sera of sleeping sickness patients. As sera from sleeping sickness patients contain an important fraction of trypanosome unrelated antibodies [31], the risk of selecting mimotopes that are unrelated to sleeping sickness by using human sera for the screening was considerable. We identified a linear region between amino acid 72 and 114 of the protein sequence of both VSG LiTat 1. 3 and LiTat 1. 5 wherein most of the peptide sequences could be aligned with the VSG protein sequence. This region is localised in the hypervariable N-terminal domain of the VSG and was for both VSGs synthesised as a linear biotinylated peptide and tested in indirect ELISA with a panel of 102 HAT positive and 102 endemic negative sera. Peptide 1. 3/78-110, corresponding to AA stretch 78 to 110 of VSG LiTat 1. 3, had an AUC of 0. 95, indicating diagnostic potential. The epitope of VSG LiTat 1. 3, recognised by the human serum antibodies used for screening of the peptide library, therefore seems to be linear and located within AA stretch 78 to 110. The peptide sequences that were selected for VSG LiTat 1. 3 had in average 3/16 amino acids in common within AA 72 to 114 with a maximum of 7/16 (44%) identical amino acids. Interestingly, a common motive of the peptide sequences was repeated twice within AA 72 to 114 of VSG LiTat 1. 3: (F/W) ExDxK (A/L/V) x (A/L/V), from AA 77 to 85 and 101 to 109. The two mimotopes of VSG LiTat 1. 3 with the highest AUC, peptide 3-2-G5 and 3-2-G10, seemed to share a common epitope (correlation coefficient of ODcs with human sera in ELISA 0. 71, data not shown), their motive WxxDxK reoccurred twice within AA 72 to 114. Their motive, (I/V/A) (T/S) DSK, could also be aligned within the C-terminal domain (AA 424 to 439). This AA stretch was synthesised as a biotinylated peptide as well, but had a low average ODc upon a first screening with nine HAT sera and was discarded for further evaluation of diagnostic performance. We think it unlikely that peptide 3-2-G5 and 3-2-G10 are mimotopes for a linear epitope localised in the C-terminal domain of VSG LiTat 1. 3. Additionally, epitopes localised in the relatively conserved C-terminal domain are more likely to react with non-VSG-specific antibodies. As for VSG LiTat 1. 3, a repetitive motive was present within AA 81 to 114 of VSG LiTat 1. 5: (Y/F/W) (x or xx) (A/L/I/V) A (A/I/L) (D or K) (A/L) xxxxE, from AA 83 to 94 and AA 99 to 111. The peptide sequences selected for VSG LiTat 1. 5 had in average 2/16 AA in common within AA 82 to 114 of the protein sequence, with a maximum of 5/16 (31%) identical AA. Contrary to the results for peptide 1. 3/78-110, peptide 1. 5/81-109, corresponding to AA stretch 81 to 109 of VSG LiTat 1. 5, had an AUC of only 0. 79, while the AUC of all of the individual peptides aligned in this region was >0. 85. Motive AYSxxxIKL of peptide 5-3-C1 (AUC 0. 87), corresponded to LYSxxxAKL (AA 99 to 106) of the VSG LiTat 1. 5 protein sequence. Peptide 5-3-C1 seems therefore to mimic an epitope that is, albeit partly, localised in this region. The similar peptide 5-2-D3, with motive F (x) xxxxKL, performed better in ELISA (AUC 0. 94). It is possible that peptide 5-2-D3 and 5-3-C1 are mimotopes for a discontinuous epitope as they share the common motive (A/I/L) (Y/F) xxxxxKLANPG with four other peptides, while ANPG was not found in the VSG protein sequence. We therefore suspect the epitope of VSG LiTat 1. 5, recognised by the human serum antibodies used for screening, to be discontinuous and to be at least partly localised within this region. This might explain the weaker performance of the linear peptide 1. 5/81-109 compared to the mimotope peptides. This finding was supported by the results of the 3DEX analysis of the mimotopes that had an AUC>90, locating peptide 5-2-D3 with 8/12 AA on the three-dimensional VSG LiTat 1. 5 protein model (E 168|N 164|D 152|G 153|T 150|K 146|L 144|A 141). In a previous study [16] we were able to identify mimotope peptides for the native trypanosomal variant surface glycoproteins by screening of peptide phage display libraries with monoclonal antibodies. Through phage display with polyclonal human antibodies we now identified different mimotopes and regions taking part in epitope formation. Because the three monoclonal antibodies used in the first study represent only a fraction of the VSG-specific antibody response, some mimotopes with diagnostic potential might have been missed. Additionally, the mouse and human immune system may recognise different B cell epitopes. Other factors may have contributed to finding different motives using the two approaches. As a result of a short infection period of two weeks, the mouse monoclonals do not recognise all VSG-epitopes. They were selected for strict VAT-specificity with purified VSGs and identified mimotopes, not necessarily dominant, that were located near the surface of the VSG N-terminal domain. The polyclonal human antibodies result from a long infection and recognise also less exposed VSG-epitopes. It may be that by affinity purification on purified VSG an antibody fraction that recognises non-surface epitopes was mainly retained, as the mimotopes of VSG LiTat 1. 3 seem to be located in this region. Another explanation may lie in the presence of self-reactive VSG-specific antibodies in sera from uninfected individuals, as has been demonstrated by Müller et al. [32]. Thus the negative selection with human antibodies from control sera may have eliminated the phages expressing the mimotopes for the VSG-specific epitopes also recognised by the mAbs. Both panning strategies thus seem complementary, in contrast to what has been described by Tang et al. [18], who selected a greater number of different 12-mer sequences with polyclonal serum for Salmonella enterica, but some of the common motives were also selected by panning of a monoclonal antibody. In the manuscript of Casey et al. [17] the mimotopes for Epstein-Barr (EBV) virus, selected with polyclonal EBV immune rabbit and patient sera were also not recognised by the monoclonal antibodies used for mimotope selection in a previous study. Diagnostic evaluation of individual mimotopes and combinations [patent application GB1202460. 0] indicates that screening of peptide phage display libraries with patient' s antibodies resulted in a more efficient selection of diagnostic peptides than with monoclonal antibodies. We therefore prefer screening with patient' s antibodies. As an alternative approach to phage display linear epitopes may be replaced by synthetic peptides identified by scanning of overlapping peptides spanning the native protein sequence [33]. Furthermore, there are alternative in vitro methods for phage display such as yeast cell or bacterial display or, non-cellular, ribosome or mRNA display [34]. Our study has nevertheless some limitations. First, considering the broad antibody spectrum in HAT sera due to polyclonal B cell activation [35], we opted to use antibody fractions that were affinity purified for VSG LiTat 1. 3 and 1. 5. Thus, mimotopes for other predominant VSGs or other trypanosome antigens with diagnostic potential have not been selected. An alternative approach to identify additional diagnostic mimotopes may consist of screening peptide phage libraries with patient antibodies against other candidate diagnostic proteins [36]. Examples are the T. b. gambiense-specific glycoprotein TgsGP [37], the T. b. rhodesiense-specific serum resistance associated (SRA) protein [38] and Trypanozoon-specific trypanosome antigens such as invariant surface glycoprotein (ISG) 65 and ISG 75 [39], microtubule associated repetitive protein 1 (MARP1) and GM6 [40]. Some of them have already been tested for their diagnostic potential in the form of recombinant fusion proteins expressed in E. coli but none are yet used in diagnostic tests for HAT. Second, even with affinity purified antibodies there is a risk that non-specific mimotopes are selected with antibodies against VSG epitopes that are normally hidden in the intact VSG coat. Usually, most of the phage particles display a consensus binding sequence after two or three rounds of enrichment [21]. We therefore performed three rounds of positive selection and two selections with negative sera. Remarkably, the mimotopes with the highest AUC for VSG LiTat 1. 3 and 1. 5 were selected after only two or even one round of panning. Third, no affinity measurements e. g. via surface plasmon resonance, have been performed. Considering the polyclonal character of antibodies in patients' sera and the inherent differences in antibody response between individual patients, we opted to assess only the diagnostic potential of the selected peptides by means of ELISA. Before the native T. b. gambiense VSGs LiTat 1. 3 and LiTat 1. 5 in the currently existing diagnostic formats can be replaced by synthetic peptides, further improvements should be considered. It is possible to define critical residues, essential for binding with the antibody, by e. g. alanine scanning mutagenesis [41]. Thus the epitope of the human serum antibodies might be recreated as has recently been done for a linear epitope on the VP1 protein of foot-and mouth disease virus [42]. Other, non-essential, parts of the peptides can then be eliminated in order to increase specificity. Phage clones that express peptides with a higher binding affinity might be selected by increasing the number of selection rounds and/or the stringency of the washing steps. In conclusion, with this study we demonstrate that mimotopes of T. b. gambiense VSG LiTat 1. 3 and 1. 5 can be selected from a phage display library and that these mimotopes and corresponding amino acid stretches within the VSGs have diagnostic potential. | Control of the chronic form of sleeping sickness or gambiense human African trypanosomiasis (HAT) consists of accurate diagnosis followed by treatment. We aim to replace the native variant surface glycoprotein (VSG) parasite antigens that are presently used in most antibody detection tests with peptides that can be synthesised in vitro. Antibodies recognising VSG were purified from HAT patient sera and were used to select phage-expressed peptides that mimic VSG epitopes from a Ph. D. -12 phage display library. The diagnostic potential of the corresponding synthetic peptides was demonstrated in indirect ELISA with sera from HAT patients and endemic negative controls. We proved that diagnostic mimotopes for T. b. gambiense VSGs can be selected by phage display technology, using polyclonal human antibodies. | Abstract
Introduction
Materials and Methods
Results
Discussion | medicine
infectious diseases
test evaluation
diagnostic medicine
african trypanosomiasis
neglected tropical diseases
parasitic diseases | 2012 | Identification of Mimotopes with Diagnostic Potential for Trypanosoma brucei gambiense Variant Surface Glycoproteins Using Human Antibody Fractions | 10,563 | 207 |
Accurate forecasts of influenza incidence can be used to inform medical and public health decision-making and response efforts. However, forecasting systems are uncommon in most countries, with a few notable exceptions. Here we use publicly available data from the World Health Organization to generate retrospective forecasts of influenza peak timing and peak intensity for 64 countries, including 18 tropical and subtropical countries. We find that accurate and well-calibrated forecasts can be generated for countries in temperate regions, with peak timing and intensity accuracy exceeding 50% at four and two weeks prior to the predicted epidemic peak, respectively. Forecasts are significantly less accurate in the tropics and subtropics for both peak timing and intensity. This work indicates that, in temperate regions around the world, forecasts can be generated with sufficient lead time to prepare for upcoming outbreak peak incidence.
Forecasting is an important tool in a number of fields, including weather and climate [1–3], agriculture [4,5], air quality [6,7], and consumer activity [8–10]. When operationalized for use in real time, predictions from probabilistic forecasts can be used in decision-making to inform, for example, emergency food aid allocation [4] or profit maximization [8]. Recently, forecasting systems have also been developed for a range of infectious diseases of high public health concern, including influenza [11–20], norovirus [21], dengue [22–25], Ebola [26–29], and, most recently, Zika [30,31]. The ability to generate accurate, real-time forecasts of infectious disease activity has important implications for public health. Currently, response to infectious disease outbreaks is primarily reactive: medical and public health professionals attempt to deal with unexpected spikes of disease incidence as they occur. By providing information on when an outbreak is expected to peak and how many cases are expected at that peak, forecasts have the potential to create a paradigm shift in infectious disease control and public health decision-making. For example, hospitals expecting a patient surge might ensure that adequate resources are available, avoiding bed and staff shortages. Seasonal influenza produces annual wintertime outbreaks in temperate regions, as well as sporadic outbreaks throughout the year in the tropics and subtropics [32,33]. The World Health Organization (WHO) estimates that influenza causes about 300,000–650,000 deaths and 3–5 million cases of severe illness each year [34]. To date, forecasts of influenza activity in the United States have been generated and operationalized [11,15]. However, while influenza forecasts have been generated for countries outside the US [13,14,17,18,35,36], these efforts are less numerous, and many countries have been ignored entirely. The tropics and subtropics are particularly neglected, with forecasts attempted for only Hong Kong [18] and Singapore [17]. This is true despite evidence suggesting that influenza burden in the tropics is similar to that in temperate regions [33]. The WHO collects influenza data year-round from several member states around the world. To our knowledge, no influenza forecasts have yet been generated using these data. Given differences in data collection procedures by country, and the importance of high data quality for generating accurate forecasts, whether these data can be used to generate accurate forecasts remains an open question. Here, we explore the following research questions: 1) Can the WHO data be used to generate accurate and well-calibrated retrospective forecasts at the country level? ; 2) Does forecast accuracy significantly differ between temperate and tropical regions? ; and 3) What factors are associated with substantial changes in forecast accuracy within both temperate and tropical regions? Based on past work, we expect that forecasting will be feasible in all regions, but that forecast accuracy will be substantially higher in temperate regions.
Influenza syndromic and virologic data were obtained from WHO’s FluID [37] and FluNet [38] web-tools, respectively. Briefly, these systems contain aggregated influenza data from WHO member states, which are either submitted by member states directly or downloaded by the WHO from existing regional databases. Good quality (see S1 Text) syndromic and virologic data were available for at least one season from 64 countries, primarily in Europe and North America (see Figs 1 and S1). Countries were classified as temperate or tropical based on both their latitude and whether they demonstrated seasonal or sporadic influenza dynamics (see S1 Text). Overall, eighteen countries were classified as tropical, and three (Australia, New Zealand, and Chile) were located in the southern temperate region. FluID data include diagnostic counts of influenza-like illness (ILI), acute respiratory infection (ARI), severe acute respiratory infection (SARI), and pneumonia, with different countries preferentially reporting different data types (see S1 Text for additional information). Because these data contain no information on laboratory testing, counts include both patients infected with influenza and patients infected by other pathogens that lead to similar signs and symptoms. To adjust for this lack of specificity, we use FluNet data, which includes the total number of tests performed for influenza and the number positive for influenza. Specifically, we multiply the syndromic case counts from the FluID tool by the proportion of tests positive for influenza in that same country during a given week. This calculation eliminates out-of-season syndromic cases that are unlikely to be due to influenza. Further, as the model used in this study (described below) simulates the transmission of a single pathogen, the removal of incidence due to non-influenza illness increases agreement between model input (data) and output. We refer to the resulting measures as ILI+, ARI+, SARI+, or pneumonia+, or, more broadly, syndromic+. For this study, we focused specifically on seasonal influenza outbreaks, and excluded the 2009 pandemic from the main analysis. While pandemic outbreaks often produce a strong incidence signal that is forecastable [17], they typically appear out-of-season in temperate regions. Seasonal influenza outbreaks, on the other hand, occur with enough frequency that, even in the tropics, where outbreak timing is less regular, future epidemics are almost certain to occur within the year. To maintain a consistent forecasting approach, we therefore focus on seasonal influenza. We present results from forecasting the 2009 pandemic alone, as well as associated methods, in S1 Text and S19 Fig. In addition, individual seasons were removed from the final dataset if: a) five or more consecutive weeks of data were missing near the outbreak peak (n = 2); b) the season consisted of fewer than 5 non-NA and non-zero data points (n = 2); c) the total attack rate of the season was less than 5% that of the largest outbreak (in other words, if case counts were unrealistically low; n = 5); d) data collection began or ended at the outbreak peak (n = 4); or e) no consecutive weeks of data were available (in other words, data were only available every other week; n = 1). We also removed data from 2010–11 in Mexico due to the continued disruption of typical seasonal patterns by the 2009 pandemic. Individual data points were removed if they occurred outside of the influenza season (as defined below under “Delineation of Influenza Seasons”) and were greater than 50% of the maximum value for the country over all seasons (n = 1). In total, 15 individual seasons were removed from the dataset, and 64 countries remained. In temperate regions, data were available for between one and seven seasons for each country for a total of 289 seasons. A complete list of countries and seasons used for forecasting can be found in S1 and S2 Tables, and the cleaned influenza data are available as S1 Dataset. Note that, for the seasonal forecasts, we began fitting tropical data at week 40 of 2010. Data on absolute humidity were obtained from NASA’s Global Land Data Assimilation System (GLDAS), which uses both observed data and modeling techniques to produce high-resolution surface meteorological data [39]. Data were available every three hours at a spatial resolution of 1°x1° for the years 1989–2008. Data from each grid cell were aggregated to the daily level, and anomalous records were identified by visual inspection and removed. Then, climatologies for each grid cell were generated by averaging daily specific humidity on each of 365 days across twelve to twenty years, depending on the amount of anomalous data removed. Finally, climatologies were aggregated to the country level by averaging the climatologies for all grid cells lying within a country, weighted by the proportion of the grid cell situated within the country in question. A more detailed description of how the humidity data were processed can be found in the S1 Text, and the processed data are available as S2 Dataset. The influenza season in temperate regions of the northern hemisphere is modeled as beginning in week 40 and ending in week 20 of the following year [40]. We shift these values by one half-year for temperate regions of the southern hemisphere; thus, the influenza season begins in week 14 and continues until week 46. For tropical regions, where consistent seasonality in influenza infection patterns is not observed, the above methods are not sufficient. Individual outbreaks are instead identified using methodology previously described in [18]. Briefly, outbreak onsets are defined as the first of three consecutive weeks where ILI+ rates exceeded the 33rd percentile of non-zero ILI+ values across all available data for a country. The end of an outbreak is defined as the first of two consecutive weeks below this threshold. To ensure that sporadic spikes in influenza are not counted, we remove any outbreaks where ILI+ counts never exceeded three times its respective onset threshold value. Country-level retrospective forecasts are developed using a model-data assimilation system consisting of: (1) influenza observations, as described above, (2) a model of influenza transmission, and (3) a filter to assimilate observations and optimize model simulation and ensemble forecast. The final two components are described here. These components differed slightly for temperate and tropical regions, and are therefore described separately. As described above, model output represents true influenza incidence per 100,000 population. Our data, on the other hand, are obtained by multiplying nonspecific syndromic data by influenza positivity rates among those who actively seek medical care. Furthermore, the majority of countries included in the WHO data provide no information on the total number of patients seen or the size of the catchment areas from which data were obtained. Thus, our data represent counts, not rates. In order to properly use the EAKF as described above, we must therefore first scale the data such that they are compatible with the model-simulated state space. In effect, the scaling factors map the observed syndromic+ data to the model state space. Scaled data, thus, represent the estimated number of syndromic+ cases per 100,000 population, and can be used for data assimilation. Model output—the simulations and forecasts—can then be scaled back to their original units (e. g. ARI+) for use by individual country public health departments. Our previous work has shown that SIRS simulations perform optimally when 15–50% of a model population of 100,000 is infected over the course of a modeled epidemic. Therefore, scaling values, γ, for each country were determined by first calculating the range of scaling values yielding a total attack rate between 15% and 50% for each season, i, ([γ15, i, γ 50, i]), then choosing a single country-specific scaling value based on the following rule: γ={if∃γ∈ℝ: γ15, i<γ<γ50, i∀i: maxi=1n (γ15, i) else: mini=1n (γ50, i) } (4) Although forecasts in the tropics were run continuously rather than by season, scaling factors for tropical countries were determined similarly using influenza outbreaks as identified under “Delineation of Influenza Seasons” above. Scaling values were allowed to vary by country, but not by season: that is, for each country, a single scaling value was chosen and used in retrospective forecasts of all available seasons. As scaling factors are controlling for differences in rates of seeking medical attention, size of the catchment area from which influenza data are collected, and overall population size by country, they vary substantially, from 0. 004 in Mexico to 374 in Peru. Forecasts were evaluated based on their ability to accurately predict outbreak peak timing (the week with the highest number of influenza cases), peak intensity (the number of influenza cases at the peak), and onset timing (the first of three consecutive weeks with influenza activity over some baseline value). Onset baseline values were chosen as 500 simulated cases for temperate countries, and 300 cases for tropical countries (see S1 Text). A forecast was considered accurate for peak timing and onset timing if the predicted value was within one week of the observed, and for peak intensity if the predicted influenza case count was within 25% of the observed. These thresholds, particularly the 1 week cutoff for peak timing accuracy, have been routinely used both in our past work [13,14,19,43,45,48,49] and in evaluating forecasts submitted to the Centers for Disease Control and Prevention’s (CDC) Predict the Influenza Season Challenge [15], allowing for comparison between the results of this work and past work. If the mode predicted onset timing is NA (no outbreak), predicted peak timing, peak intensity, and onset timing were set to NA, and the forecast was removed from consideration. Forecast accuracy was compared for temperate vs. tropical regions, as well as within temperate regions by hemisphere, region, data type, season, and scaling, and within the tropics by region, data type, and scaling. Because, in real time, the actual time to peak is unknown, we evaluated forecast accuracy by predicted lead time (i. e. the difference between the week at which a forecast is initiated and predicted peak timing). For most analyses, forecast accuracy was assessed at predicted lead weeks -6 to 4 (i. e. six weeks before the predicted peak through four weeks after). Comparisons were made for each individual variable using generalized estimating equations (see S1 Text for more details). To assess whether the effects of the explanatory factors change over time, GEE models were also run restricting the data to either before or after the predicted peak. Seasons with no identified onset (in other words, where no outbreak occurred) were removed before analyzing forecast accuracy. Additionally, because individual outbreaks within tropical countries are identified during the forecasting process, and therefore were not checked for quality previously, outbreaks where a) five or more consecutive weeks of data were missing; or b) data collection for an outbreak began at the outbreak peak were removed from tropical countries’ results before GEEs were run. To assess the impact of including humidity forcing in the temperate models, we generated an additional set of forecasts for the temperate regions, this time without including humidity forcing in the model structure (see S1 Text). This resulted in two distinct forecasts for each country, season, start week, and run: one incorporating humidity data and one not. In order to fully take advantage of this paired design, forecast accuracy was compared by observed lead week using the exact binomial test. Because individual comparisons were made for each lead week, we applied a Bonferroni correction and considered differences to be statistically significant when p-values were less than 0. 0045 (p = 0. 05 / 11). Unlike in previous analyses, rather than removing forecasts predicting no onset, we considered these forecasts to be “inaccurate. ” This was done to avoid ignoring pairs of forecasts where one failed to recognize an oncoming outbreak but the other accurately predicted peak timing or intensity. Sensitivity analyses were performed to test how forecast accuracy changes as a function of EAKF observational error variance, onset baseline value, scaling, and accuracy metric. Findings from these sensitivity analyses broadly agree with the results presented here (results in S1 Text).
Retrospective forecasts were performed using syndromic+ data from 64 countries, of which 18 were classified as tropical. In the temperate regions, data were available for between 2 and 7 seasons, with each country contributing an average of 6 seasons of data (data in S2 Table). In the tropics, data were available for between 29 and 345 weeks (mean = 166 weeks; median = 140 weeks). In the northern temperate region, onset timing occurred between weeks 45 and 64, and peak timing occurred between weeks 48 and 67. In the southern temperate region, these values were weeks 23 and 33 for onset timing and 29 and 38 for peak timing. Overall, we found that accurate forecasts of both peak timing and peak intensity for influenza outbreaks are possible using publicly available WHO data. In temperate regions, we were able to develop country-level, retrospective forecasts that exceeded 50% accuracy for peak timing (i. e. , 50% of forecasts predicted peak timing within one week of the observed value) up to four weeks before the predicted peak, and for peak intensity (within 25% of the observed value) two weeks before the predicted peak. Forecasts exceeded 75% accuracy for peak timing one week before the predicted peak, and for peak intensity at the predicted peak week (Fig 2A). Forecast accuracy was lower in the tropics, never exceeding 50% for either peak timing or peak intensity (Fig 2B). As expected [11,12,14,17,18], forecast accuracy varied as a function of lead time, with forecasts near and after the forecasted peak typically performing better than forecasts generated several weeks before the peak. Similar patterns were seen by observed lead time, although tropical forecast accuracy was much higher after the observed peak, exceeding 70% (results in S3 Fig). Broadly, these results remained consistent after altering the cutoff point at which forecasts were considered accurate (S11 Fig), and when correlation coefficients and symmetric mean absolute percentage error (sMAPE) over the entire forecast period were assessed (S12 Fig), although forecast accuracy assessed using sMAPE was comparable between temperate and tropical regions. For both temperate and tropical regions, forecasts of outbreak onset timing showed high accuracy post-onset, but forecasts were rarely generated in advance of the predicted onset week (Table 1). Specifically, no temperate forecasts predicted that onset would occur with more than a one week advanced lead, and very few forecasts in the tropics accurately predicted onset with more than a one-week lead. In temperate regions, onset timing accuracy (onset predicted within one week of the observed value) quickly increased and remained above 95% as soon as the predicted onset was in the past. In the tropics, accuracy reached almost 50% at the predicted onset, and remained around 65–70% for all later lead weeks. For the tropics only, we also evaluated how often forecasts correctly recognized an existing or upcoming outbreak, without mistakenly predicting outbreaks during periods in which no outbreaks occurred. Specifically, we calculated sensitivity, specificity, positive predictive value, and negative predictive value. We found that both sensitivity (98. 56%) and the negative predictive value (98. 10%) were high, but that specificity (56. 22%) and the positive predictive value (63. 12%) were much lower. Thus, while forecasts are unlikely to predict dormancy before or during an outbreak, forecasts suggesting a current or upcoming outbreak were inaccurate more often than accurate. It is important to consider not only how accurate forecasts are, but also forecast uncertainty. This is especially true in the case of real-time forecasting: different medical and public health responses might be affected given forecast of an 80% chance of a particular outcome rather than a 20% chance. Because each forecast is based on 300 individual ensemble members, we could assess forecast certainty through the spread of the ensemble variance, where narrower ensemble spread ideally indicated greater certainty. Fig 3A and 3B show average peak timing and intensity forecast accuracy, respectively, for temperate regions plotted against ensemble variance (separated into 10 quantiles). For peak timing, we generally saw a slight decrease in forecast accuracy as ensemble variance increases at all predicted lead weeks, indicating that we can infer expected forecast accuracy from ensemble spread. For peak intensity, this pattern only held prior to the predicted peak. Corresponding plots for the tropics are shown in Fig 3C and 3D. For peak timing, no clear relationship existed between ensemble variance and forecast accuracy, indicating that no information about expected forecast accuracy can be inferred from ensemble spread. For forecasts of peak intensity, on the other hand, increases in ensemble variance corresponded to substantial decreases in forecast accuracy. We also explored how often the observed peak timing and intensity fall within certain prediction intervals of ensemble spread prior to the predicted peak (Fig 4). In a well-calibrated forecast, we expect that the observed intensity will fall within the nth% prediction interval n% of the time. Overall, forecasts appeared to be well calibrated for both peak timing and intensity in temperate regions at all lead times, although prediction intervals tended to be too wide for peak timing, especially several weeks before the peak. In the tropics, peak intensity forecasts appeared well calibrated, while peak timing forecasts rarely included the observed peak timing. Further exploration of forecast calibration can be found in S10 Fig.
We have shown that, in temperate regions, accurate and well-calibrated retrospective forecasts of seasonal influenza activity are feasible. Work is currently being conducted to determine whether real-time forecasts are similarly feasible, and future work will incorporate travel between countries with the goal of improving forecast accuracy, particularly onset timing accuracy. Although this work is at an early stage, we note the importance of eventually incorporating forecasts into medical and public health decision-making. Accurate real-time probabilistic forecasts have the potential to inform decisions such as antiviral stockpiling by governments or staff and bed management by hospitals, preventing morbidity and mortality. Therefore, it is critical that these forecasts not be produced solely as an academic exercise. | Influenza is responsible for an estimated 3–5 million cases and 300–650,000 deaths each year worldwide. If produced early enough, accurate forecasts of influenza activity could guide public health practitioners and medical professionals in preparing for an outbreak, reducing the subsequent morbidity and mortality. For example, hospitals could use these forecasts to determine how many beds will be needed when an outbreak is most intense. Despite this potential impact, influenza forecasts are primarily generated for the United States, with forecasts for other countries being comparatively rare. Here, we use publically available influenza data to forecast influenza activity in 64 countries. We find that accurate forecasts can be produced several weeks before the outbreak’s peak in temperate countries, where influenza outbreaks occur regularly during the winter. Forecast accuracy is lower in the tropics and subtropics, where outbreaks occur more sporadically. Overall, our results suggest that forecasts have potential as an important public health tool in many countries, not only in the US. | Abstract
Introduction
Materials and methods
Results
Discussion | medicine and health sciences
statistics
influenza
geographical regions
atmospheric science
temperate regions
simulation and modeling
seasons
mathematics
forecasting
regional geography
humidity
research and analysis methods
public and occupational health
infectious diseases
geography
mathematical and statistical techniques
tropical regions
meteorology
earth sciences
viral diseases
physical sciences
statistical methods | 2019 | Development and validation of influenza forecasting for 64 temperate and tropical countries | 4,897 | 210 |
The cell wall of Gram-positive bacteria is a complex network of surface proteins, capsular polysaccharides and wall teichoic acids (WTA) covalently linked to Peptidoglycan (PG). The absence of WTA has been associated with a reduced pathogenicity of Staphylococcus aureus (S. aureus). Here, we assessed whether this was due to increased detection of PG, an important target of innate immune receptors. Antibiotic-mediated or genetic inhibition of WTA production in S. aureus led to increased binding of the non-lytic PG Recognition Protein-SA (PGRP-SA), and this was associated with a reduction in host susceptibility to infection. Moreover, PGRP-SD, another innate sensor required to control wild type S. aureus infection, became redundant. Our data imply that by using WTA to limit access of innate immune receptors to PG, under-detected bacteria are able to establish an infection and ultimately overwhelm the host. We propose that different PGRPs work in concert to counter this strategy.
The complex cell surface of bacteria has been directly or indirectly associated with different strategies that bacterial pathogens use to interact with the host. These include acquisition of specific adhesion factors, formation of biofilms, adaptation to an intracellular environment, production of a protective capsular polysaccharide or evasion of innate immune defences (e. g. lysozyme) [1]. The host counters these strategies by targeting conserved molecules (pathogen associated molecular patterns or PAMPs), unique in bacteria, that are either present at the bacterial surface or are released by bacteria as they attempt to establish infection. Bacterial PAMPs include Peptidoglycan (PG), a heterogeneous polymer of glycan chains cross-linked by short peptides of variable length and amino acid composition [2]. Although PG recognition is essential to trigger an inflammatory response, this macromolecule may not be easily accessible for recognition at the surface of bacteria. In Gram-positive bacteria, PG is buried within a complex cell surface consisting of different molecules [3]–[5]. Such molecules include surface proteins, covalently linked or tightly associated with PG, capsular polysaccharides, usually required for the ability of different bacteria to cause disease [6] and wall teichoic acids (WTA), phosphate-rich glycopolymers involved in the resistance of bacteria to environmental stress and regulation of bacterial division [7]. It is not clear therefore, how the host would be able to sense bacterial PG buried within such complex structures. One hypothesis is that the innate immune system recognises soluble PG fragments that are released from the bacterial cell surface through the activity of enzymes produced by bacteria (such as autolysins) or by the host (such as lysozyme) [2], [8]. However, certain bacteria have the ability to modify their PG, turning it more resistant to the action of such enzymes [9], thus preventing the release of small soluble fragments capable of triggering an innate immune response in the host. This may be the case for Listeria monocytogenes that has the ability to de-N-acetylate its PG allowing them to survive the action of lysozyme and evade the host innate immune system [10]. Another hypothesis is that the components of the host innate immune system are able to bind directly to PG present within the bacterial cell surface. As discussed earlier, PG is decorated with a variety of large molecules that may sterically block access of host receptors to the underlying PG. In Gram-positive bacteria, cell wall glycopolymers, including WTA may play this role [1]. The role of WTA protecting the PG from recognition would have important implications regarding the onset of infection by major human pathogens such as Staphylococcus aureus (S. aureus) [1]. Recently, it has been shown that different components, present at the cell wall of S. aureus bacteria, may determine the survival of infected Drosophila. Specifically, S. aureus strains impaired in the expression of enzymes involved with the metabolism of cell wall components were unable to kill flies [11]. Moreover, it has been proposed that D-alanylation of the WTA produced by S. aureus may inhibit the recognition of PG by host receptors. This inhibitory effect was observed in vitro not only when WTA was covalently attached to polymeric PG but, surprisingly, also when WTA was covalently attached to monomeric PG [12]. The fruit fly Drosophila melanogaster recognises Gram-positive bacteria by either direct binding to PG or its smallest components [13]. Based on in vitro data [14] and infection studies of mutants [14], [15], the current working hypothesis is that a flexible system of pattern recognition receptors (PRRs) can be deployed by the host immune system to detect Lysine-type PG from different Gram-positive bacterial pathogens. Two Peptidoglycan Recognition Proteins (PGRPs), namely PGRP-SA and PGRP-SD are major components of this system [15], [16]. Depending on the bacterium, each, or both of these PGRPs – along with Gram-Negative Binding Protein1 (GNBP1) [17] – interacts with PG and activate a downstream proteolytic cascade, which culminates in Toll receptor signalling. The signal reaches the cytoplasmic NF-κB/I-κB complex via a receptor/adaptor complex comprising dMyD88, Tube and the IRAK homologue Pelle. At that point the I-κB homologue Cactus is phosphorylated and targeted for degradation while the NF-κB homologue Dif is free to enter the nucleus of host cells and regulate target genes [18]. Prominent among these genes, is a group of potent antimicrobial peptides (AMPs), which are synthesised by the fat body and secreted into the haemolymph. An AMP frequently used as a read-out for the Toll pathway is Drosomycin (Drs). AMPs and local melanization, along with the phagocytic activity of haemocytes constitute respectively the humoral and cellular arm of the fruit fly response to infection [18]. Here, we report for the first time that Drosophila PGRP-SA, a non-lytic PGRP was able to bind intact live bacteria in vivo. Access to PG was limited by the presence of WTA: binding of PGRP-SA to various live Gram-positive bacteria was minimal, but binding to purified PG, stripped of covalent modifications (including WTA) was far greater. Through inhibiting WTA synthesis, either by the addition of an antibiotic or genetically, we were able to potentiate detection of these bacteria by PGRP-SA. For S. aureus, this correlated with a reduced ability of the bacteria to proliferate within the host, and a reduced susceptibility of the host to infection in a PGRP-SA/GNBP1 dependent manner. We also observed that PGRP-SD, essential for sensing wild type S. aureus, became redundant as WTA levels were reduced. Overall, our results suggest that WTA may be part of a general mechanism used by Gram-positive bacteria, which limits the access of innate receptors to PG, thereby enabling bacteria to evade detection and establish infection.
To address the question of whether Gram-positive bacteria counter host recognition by limiting access of innate sensors to PG, we constructed a fluorescent derivative of the fruit fly Lys-type PG receptor, PGRP-SA (mCherry-PGRP-SA). This construct and an untagged version (rPGRP-SA) were expressed in Escherichia coli and the resulting proteins were purified. As shown in the supplementary material (Figure S1A), injection of mCherry-PGRP-SA, or rPGRP-SA, into PGRP-SA deficient flies restored Drs-GFP production induced by infection with Micrococcus luteus (M. luteus). Endogenous Drs expression was also restored as confirmed by qPCR (Figure S1B). These observations were consistent with our previous results when using a recombinant PGRP-SA expressed in the lepidopteran cell line Sf9 [19]. Taken together, these results showed that the fluorescently tagged PGRP-SA and the untagged versions are functional and capable of restoring an innate immune response in PGRP-SA deficient flies. Initially, we used rPGRP-SA and mCherry-PGRP-SA in co-precipitation experiments in order to study binding to PG from different Gram-positive bacteria. Both bound with similar affinity to PG purified from M. luteus, Enterococcus faecalis (E. faecalis), and S. aureus (data not shown and Figure 1A, respectively). For details of PG composition of these bacteria see Figure S2. Importantly, this indicated that the mCherry-tag appeared not to interfere with PGRP-SA binding, and thus, demonstrated that both proteins were able to bind Lys-type PG of different composition. We therefore assessed in vitro, the binding of mCherry-PGRP-SA to the surface of live bacteria harvested during exponential growth phase. Notably, the binding of the recombinant protein to live bacteria exhibited a range of different affinities in contrast to their respective purified PG. Binding to live E. faecalis and S. aureus was significantly reduced, when compared to binding to M. luteus (Figure 1B). However, the binding levels of PGRP-SA to the purified PG from these bacteria were similar (Figure 1A). We also noticed that while mCherry-PGRP-SA was capable of binding the entire surface of M. luteus cells, it bound at specific sites at the surface of S. aureus cells, similar to what has been described recently for mammalian bactericidal PGRPs [20]. These results suggested that although the three types of bacterial PG were similarly recognized by PGRP-SA, the presence of other components found at the surface of live bacteria might have prevented PGRP-SA from finding its PG ligand. The cell surface of a Gram-positive bacterium is a complex structure consisting of a thick layer of PG, surface proteins and glycopolymers such as capsular polysaccharides and WTA. As previous studies had shown that certain PG-binding proteins, such as bacterial autolysins, have a higher affinity for the surface of bacterial strains lacking WTA [21]–[23], it was decided to investigate whether presence of WTA could be preventing PGRP-SA from binding to the surface of live bacteria. Further support for the choice of WTA came from the fact that different Gram-positive bacteria can produce WTA with a variable composition [24]–[26]. M. luteus, for which mCherry-PGRP-SA displayed the highest affinity, does not produce WTA [24], [27], (Figure S2C). To test whether WTA mediated the differential binding of PGRP-SA, we cultured bacteria in the presence of tunicamycin, thereby inhibiting their ability to synthesize WTA. At lower, sub-inhibitory concentrations as those used in this study, tunicamycin specifically inhibits TagO [28]: a glycosyltransferase that specifically localizes to the division septum of S. aureus [29] and is required for the initial step of WTA biosynthesis, namely, the transfer of GlcNAc to the C55-P lipid anchor bactoprenol. We observed higher levels of mCherry-PGRP-SA binding to the newly synthesized cell material, located at the division septum, when Gram-positive bacteria cells were treated with tunicamycin (Figure 2A). S. aureus and S. saprophyticus exhibited a similar increase in binding, 63× and 84× respectively, whilst E. faecalis binding increased 8×. It should be noted that the effect of tunicamycin in these bacteria was not the same. While addition of the antibiotic resulted in binding of mCherry-PGRP-SA to the entire cell surface of S. aureus, binding was observed predominantly at the division septum in S. saprophyticus and exclusively at this region in E. faecalis. We attribute these differences to how and where the new cell wall synthesis occurs in these bacteria. Nevertheless, the results described above suggested that WTA in different bacteria might protect PG from exposure to host receptors. To confirm that WTA were indeed required to reduce access of PGRP-SA at the cell surface, we quantified the binding of mCherry-PGRP-SA to S. aureus mutants that produced varying amounts of WTA due to mutations in the tagO gene [29]. We chose S. aureus because it is a major human pathogen with a well-characterised WTA synthetic pathway [30], [31]. A complete absence of WTA, which occurs when tagO is entirely deleted (RNΔtagO pMAD), or when two highly conserved residues have been mutated (RNΔtagO ptagOD87A/D88A), resulted in equivalently enhanced levels of mCherry-PGRP-SA binding, when compared to the wild type strain (∼2×103 and ∼3. 3×103-fold respectively, Figure 2B). To verify that the observed result was indeed due to the loss of WTA, we expressed wild type tagO in the RNΔtagO background (RNΔtagO ptagO): this rescued the loss of WTA (WTA levels restored to 90% of wild type levels) [29], and reduced mCherry-PGRP-SA binding to levels close to those observed for the wild type strain (Figure 2B). A tagO mutant that could only support production of a reduced amount of WTA (RNΔtagO ptagOG152A; 24% levels of WTA compared to wild type) exhibited an intermediate level of mCherry-PGRP-SA binding relative to all strains (6×102-fold increase relative to the wild type strain, Figure 2B). Overall, our data indicated that WTA found in the cell wall of different live Gram-positive bacteria restricted PGRP-SA from binding their PG, and in S. aureus this occurs in a dose dependent manner. We next wanted to examine whether increased PGRP-SA binding – due to a lack of WTA – affected the ability of bacteria to survive in an in vivo system. We chose D. melanogaster because it is a well-established model for dissecting pattern recognition in innate immunity [18]. We know for example that in vitro, three PRRs – PGRP-SD/PGRP-SA/GNBP1 – form a ternary complex for binding to the PG of S. aureus [14]. As a first approach wild type and mutant S. aureus strains were injected into wild type flies and also into flies defective for PGRP-SD or PGRP-SA. We then determined the number of CFUs 6 and 17 hours post-infection; the latter time point being when the first flies succumb to infection (Figure 3 and S5). All flies were inoculated with low and statistically identical numbers of bacteria (∼102 CFUs per fly; Figure 3, Time 0). Our rationale was to induce infections that were comparable and that could evolve over time. For example, flies generally succumb to bacterial infection when their numbers increase beyond 106 CFUs per fly [18], [32], and therefore, high initial loads (e. g. 104–105 CFUs per fly) may overwhelm the host and consequently may not be informative regarding the course of an infection. We observed that wild type S. aureus (NCTC8325-4) CFUs increased in all fly backgrounds over the period of infection to numbers that were statistically separable, with PGRP-SA deficient flies carrying the heaviest load (Figure 3). In contrast, the numbers of the S. aureus mutant, which lacked WTA (NCTCΔtagO) [29], did not significantly increase in the wild type or PGRP-SD mutant background. However, the number of NCTCΔtagO bacteria in the PGRP-SA mutant was significantly higher at both the 6 and 17 hours time points (Figure 3). Two-way ANOVA revealed a significant interaction between the bacteria and fly strains, which was due to the large increase of NCTCΔtagO bacteria in the PGRP-SA mutant. Together, these data indicated that WTA were fundamental for S. aureus to counter recognition by PGRP-SA, and consequently, the bacteria were able to increase their number during the initial course of infection. We have previously observed that PG produced by NCTCΔtagO bacteria has reduced levels of cross-linking relative to the wild type strain [29]. To evaluate whether this contributed to the inability of NCTCΔtagO bacteria to increase their number in wild type or PGRP-SD mutant flies, we assessed mCherry-PGRP-SA binding to NCTCΔpbpD and determined CFUs at 6 and 17 hours. NCTCΔpbpD is a derivative of NCTC8325-4 in which pbpD (the gene encoding to penicillin binding protein 4, PBP4) has been deleted. Deletion of pbpD results in a strain that produces PG with a similar level of cross-linking to that found in NCTCΔtagO [29], but which still produces WTA. The inability of NCTCΔpbpD and NCTCΔtagO to produce a highly crosslinked PG did not interfere with bacteria growth in culture, as its duplication time at 30°C was very similar to the parental NCTC8325-4 strain (Figure S3B). In both experiments, NCTCΔpbpD behaved as the wild type bacteria. Firstly, binding mcherry-PGRP-SA similarly (Figure S3C) and secondly, for each fly background attaining numbers that were statistically inseparable from those for NCTC8325-4 (Figure 3, Time +17 hours). To assess whether the developing trend in bacterial numbers at 17 hours post-infection resolved into differences in how flies survive, we monitored the number of flies alive at 24 hour intervals over 3 days. In addition, we infected GNBP1 mutant flies, because GNBP1 has been postulated to work as part of a complex with PGRP-SA [14], [17]. Survival curves for a particular fly background when infected with either NCTC8325-4 or NCTCΔpbpD were statistically inseparable, except for those obtained for the wild type background, where flies succumbed more to NCTCΔpbpD (Figure 4; 62% and 38% survival at 72 hours post-infection, respectively). Nearly all PGRP-SA and GNBP1 mutant flies had died by 24 hours, whereas ∼40% of PGRP-SD mutant flies survived beyond this time point, succumbing to infection around 48 hours (∼5% of flies surviving). In contrast, ∼95% of wild type and PGRP-SD mutant flies survived the NCTCΔtagO infection up to 72 hours (furthermore, taking CFUs at this time-point revealed that NCTCΔtagO had been eliminated from these flies, 0 CFUs per fly). The majority of PGPR-SA and GNBP1 flies had succumbed to infection by 48 hours (3% of flies surviving). A similar trend in survival outcome was observed with NCTC8325-4 after treatment with tunicamycin (Figure 4). These data confirmed that WTA were indeed required to counter host immunity, because without them, infection could be controlled in a PGRP-SA/GNBP1 dependent manner. Differences in CFUs were apparent 6 hours post-infection suggesting that recognition and reduction of propagation or killing of bacteria, occurs rapidly following infection. Interestingly, these results also showed that a requirement for PGRP-SD was bypassed when WTA are removed and PGRP-SA has far greater access to PG. To further demonstrate the necessity for WTA to protect PG from host recognition, we monitored survival of flies infected with the aforementioned TagO point mutations (Figure 2B and Figure 5). In these experiments, we wanted to rule out unknown causes that may occur due to the absence of the TagO protein per se, and also, lessen adverse effects that may occur due to a complete lack of WTA. The survival trend for flies infected with RNΔtagO pMAD, that lacks tagO and carries an empty pMAD plasmid vector (vector control), was similar to that for NCTCΔtagO: the PGRP-SA mutant succumbed rapidly, whereas the PGRP-SD mutant and wild type flies generally survived, their curves being statistically inseparable (Figure 5). The injection of the complemented strain (RNΔtagO ptagO) resulted in survival outcomes that were characteristic of NCTC8325-4, with PGRP-SD mutant and wild type flies succumbing to the infection, with their curves being statistically separated (Figure 5). Notably, wild type and PGRP-SD mutant flies infected with RNΔtagO ptagOG152A (which produces ∼24% WTA relative to RNΔtagO ptagO but produces similar levels of the TagO protein) [29] survived to intermediary levels (Figure 5). Overall, survival of wild type flies decreased as WTA levels increased (with a concomitant decrease in PGRP-SA binding, Figure 2B), and likewise for the PGRP-SD mutant; with the difference between wild type and PGRP-SD mutant survival successively increasing. In contrast, survival of PGRP-SA mutant flies was independent of WTA levels, with flies succumbing strongly for all infections in a statistically inseparable manner (Figure 5). These data confirmed that it was indeed in vivo protection of PG by WTA against the consequences of PGRP-SA binding, and furthermore, suggested that a requirement for PGRP-SD gradually became redundant as WTA levels decreased. It has been reported previously that D-alanylation of WTA is also required for the pathogenicity of S. aureus [11]; D-alanylation is a process that incorporates D-alanine residues into the glycerol-/ribitol-phosphate backbone of WTA, thereby reducing the negative charge of the polymer [33]. We examined therefore, whether a S. aureus mutant that lacks the D-alanylation pathway (RNΔdltABCD) bound mCherry-PGRP-SA equivalently to RNΔtagO. Binding of, mCherry-PGRP-SA to RNΔdltABCD was similar to the binding to the wild type bacteria (Figure S3). This prompted us to assess how RNΔdltABCD affected survival of the wild type, PGRP-SD and PGRP-SA mutant flies. In contrast to RNΔtagO, PGRP-SA mutant flies did not succumb strongly to RNΔdltABCD infection, with 83% surviving at 72 hours post-infection (Figure S4); furthermore, survival was statistically inseparable for the different fly backgrounds (Figure S4). These data demonstrated that D-alanylation is not necessary for WTA to limit the access of PGRP-SA, that neither PGRP-SD nor PGRP-SA were required to control the RNΔdltABCD infection and consequently, the reduced killing effect of RNΔdltABCD had nothing to do with recognition.
The results shown here indicate that in respect to Gram-positive bacteria, where the cell wall is not concealed by outer membrane (e. g. staphylococci), pathogen recognition, via recognition of PG, is tightly linked to host survival. Our studies bring forward the notion that one of the strategies used by pathogens to reduce recognition is to restrict accessibility to inflammatory non-self components of the cell wall. Specifically, the results here show that presence of WTA in a range of Gram-positive bacteria impaired PGRP-SA binding. The use of tunicamycin to abolish WTA synthesis dramatically improved receptor recognition of bacteria as well as host survival of flies infected antibiotic treated S. aureus. Genetically deleting a major component of the WTA synthesis (TagO) in S. aureus also increased PGRP-SA binding leading to increased host survival. It should also be noted that, rPGRP-SA was capable of binding in vitro significantly better to WTA-free PG than to WTA-linked PG that were purified from wild type S. aureus bacteria, treated with trypsin to remove any attached surface proteins and adjusted to the same concentration of PG (Figure S2). This observation confirmed the results obtained with live bacteria and allowed us to eliminate the notion that deletion of tagO gene may influence the amount of protein present at the cell surface and that this change in protein levels was influencing the binding of PGRP-SA. Effectively during the course of this work we have removed WTA from PG by treatment with antibiotic, by deletion of the tagO gene and finally we have chemically removed them from PG. In all the cases binding of PGRP-SA to PG has increased. S. aureus produces WTA composed of about 40 ribitol phosphate-repeating units modified with N-acetylglucosamine (GlcNAc) and D-alanine [7]. The latter modification is mediated by the D-Alanine ligase DltA and partially neutralizes the negative charge of the cell surface thus reducing attraction of cationic AMPs [33]. ΔdltA mutants are more susceptible to killing by cationic AMPs and neutrophils in vitro and have markedly reduced virulence in several animal infection models including Drosophila [11], [34]. In one of these studies [11], Tabuchi and colleagues showed that S. aureus producing WTA without D-alanylation were impaired in their ability to kill Drosophila. Surprisingly, the ΔdltA mutant was more impaired in the ability to kill flies than an independently generated tagO mutant [11]; the latter according to the authors had the same pathogenicity as wild type S. aureus [11], contrary to our findings. There is a crucial point to be made in reference to this however, which is at the heart of our experimental design and gives physiological relevance to our results. We propose that WTA are important to reduce S. aureus recognition by the host and thus help the pathogen increase its numbers inside the fly. The host uses PGRP-SA to control bacterial numbers and the more PGRP-SA binds to the cell wall (see Figure 2B) the more the bacterial load is controlled (as seen by comparing CFUs between wild type NCTC8325-4 S. aureus and NCTCΔtagO in Figure 3A). In PGRP-SA mutants the control mechanism is absent and NCTCΔtagO was able to proliferate and kill the host (Figure 3B). We were able to observe this because we started from a low bacterial load (102 cells/initial infection/fly) and followed the progress of pathogen load inside the host. Tabuchi et al. injected 104–106 cells per fly for all bacterial strains used [11]. In our hands this concentration overwhelmed the host from the beginning and it is not surprising that these authors were unable to resolve statistical differences in host survival. In order to rule out possible pleiotropic effects produced by the inactivation of the tagO due to the insertion of non-replicative plasmids or reversion of the mutation by elimination of the plasmid from the chromosome, we have specifically deleted the tagO gene in a manner that left no resistance marker in the bacterial chromosome and thus minimized possible alterations on the transcription of neighbouring genes. Finally, in order to increase the confidence of our results, we have complemented the tagO null mutant with plasmids that allowed the expression of a partially active (TagOG152A), TagO protein and have statistically analyzed the estimated host survival probability curves obtained. Finally we should emphasize that deletion of the tagO gene in NCTC8325-4 strain (an agr positive strain) and in RN4220 (an agr negative strains) resulted in similar outcomes (Figure S3) - a reduced pathogenicity in the Drosophila infection model and the production of a bacterial cell surface that was better recognized by mCherry-PGRP-SA. In parallel experiments we have also generated a ΔdltA deletion mutant (this study) as well as a deletion of the ΔdltA operon (ΔdltABCD) [29] and found that both were indeed less pathogenic than wild type S. aureus (Figure S4), similar to what was previously reported [11]. However, this reduced pathogenicity was also observed in PGRP-SA and PGRP-SD single mutant flies (in contrast to ΔtagO). This indicated that the non-pathogenicity of ΔdltA was not linked to recognition by PGRP-SA or PGRP-SD. We propose that increased “visibility” of PG to PGRP-SA when WTA were removed, dramatically improved survival of the host. However, alternative interpretations of our results may exist. In the following section we will attempt to challenge and rule them out: Our results underline an important aspect of pathogen recognition by the host, which remains relatively unexplored. Namely, how does the host recognition machinery respond to changes in the surface of bacteria? Here we manipulated the amount of WTA on the cell surface of S. aureus. Previously, two host PGRPs, PGRP-SA and PGRP-SD were found to be involved in recognition of wild type S. aureus [14], [15]. We found here that when WTA were genetically removed, the requirement for PGRP-SD was abolished. Flies deficient for PGRP-SD had estimated survival probabilities comparable to wild type flies following infection by S. aureus ΔtagO or ΔtagOptagOD87/D88A. When a small amount of WTA was left on the surface through the residual activity of the S. aureus ΔtagOptagOG152A then PGRP-SD mutants were less able to survive infection. However this sensitivity was not as pronounced as when infected with S. aureus ΔtagOptagO, the strain with reconstituted wild type levels of WTA. Previous studies have established that PGRP-SD does not bind Gram-positive Lys-type PG [14], [39]. However, in its presence, PGRP-SA was able to bind substantially better to cell wall from S. aureus and S. saprophyticus [14]. Our results, combined with the latter observation, support a role for PGRP-SD in neutralizing the effect of WTA obstructing access to PG. The alternate hypothesis that PGRP-SD may directly recognize WTA, and is therefore not necessary when flies are infected with bacteria that lack teichoic acids, is also a possibility. The role of teichoic acids in concealing PG at the surface of Gram-positive bacteria may be also effective in preventing recognition by innate immune sensors of other organisms. It is now established that insect PGRPs have mammalian homologues and mice and humans express four genes encoding members of this family [35]. Our results correlate with data, which attributed a significantly reduced virulence of tagO mutants in cotton rat nasal colonisation model [40] as well as a mouse endophthalmitis model [41] and suggest a mechanism for how this may happen: absence of teichoic acids may render PG at the bacteria surface more exposed to the host immune system.
Isogenic wild type flies (Bloomington #25174) were used as the wild type control. For the survival and bacterial Colony Forming Unit (CFU) experiments, and DD1 flies for assaying Drs levels visually or via qPCR; the latter carries a Drs-GFP and a Diptericin-lacz reporter [42]. The PGRP-SA and PGRP-SD mutant backgrounds are, respectively: flies with the semmelweis mutation in PGRP-SA [16] and a 1499 bp deletion in PGRP-SD (PGRP-SDΔ3) [15]. The spzrm7 [43] and spz1 [44] Toll pathway mutant backgrounds, and the Dif1-key1 [45] Toll-IMD pathways double mutant background, were used to assess survival of flies deficient for AMPs. All fly stocks were reared at 25°C. Bacterial strains are listed below. S. aureus strains were grown in tryptic soy broth medium (TSB; Difco) supplemented with antibiotic (erythromycin 10 µg/ml; Sigma-Aldrich) when required. E. faecalis was grown in brain heart infusion medium (BHI; Fluka). M. luteus was grown in Luria-Bertani medium (LB; Difco). Bacteria were plated from -80°C stocks every 7 days Growth of all bacteria cultures were done at 30°C as S. aureus mutants impaired in the synthesis of teichoic acids are thermosensitive [46]. NCTC8325-4 (S. aureus reference strain from R. Novick); NCTCΔtagO (NCTC8325-4 tagO null mutant [29]); RN4220 (Restriction deficient derivative of S. aureus NCTC8325-4 that can be electroporated); RNΔtagO (RN4220 tagO null mutant [29]); RNΔtagOpMAD (RNΔtagO transformed with pMAD [29]– shuttle vector with a thermosensitive origin of replication for Gram-positive bacteria); RNΔtagO ptagO (RnΔtagO transformed with ptagO [29]); RNΔtagO ptagOD87A/D88A (RNΔtagO transformed ptagOD87A/D88A [29]); RNΔtagO ptagOG152A; RNΔtagO transformed with ptagOG152A, [29]); RNΔdltABCD (RN4220 dltABCD null mutant [29]); RNΔdltABCD (RN4220 dltABCD null mutant [29]); RNΔdltA (RN4220 dltA null mutant, this study). M. luteus strain: DMS20030 [47]; E. faecalis strain: JH2-2 [48]; B. subtilis strain MB24 [49]. To delete the dltA gene from the chromosome of S. aureus RN4220 we started by amplifying two 0. 55 Kb DNA fragments from the genome of S. aureus NCTC 8325-4 strain, corresponding to the upstream (primers 5′-AGATCTgaatgtatatatttgcgctgatg-3′ and 5′-gtaaaatcaccatatggaatcatattaagtctccctcattagaactc-3′) and downstream (primers 5′- gagttctaatgagggagacttaatatgattccatatggtgattttac-3′ and 5′-GAATTCcgaaacgtttgtaacgatcg-3′) regions of the dltA gene. The two fragments were joined by overlap PCR using primers P33 and P36 and the resulting PCR product was digested with BglII and EcoRI and cloned into the pMAD vector, producing the plasmid pΔdltA. This plasmid was sequenced and electroporated into S. aureus RN4220 strain. Insertion and excision of pΔdltA into the chromosome of RN4220 was performed as previously described [29] with the exception of the incubation temperature after excision of the plasmid, which was 30°C (instead of 43°C) due to the thermosensitive nature of the cells lacking D-alanylation. Deletion of dltA was confirmed by PCR, and the resulting strain was named RNΔdltA. Overnight 10 ml cultures of bacteria were washed and resuspended in an equal volume of sterile phosphate buffered saline (PBS), and further diluted 1/1000. Healthy looking adult flies from uncrowded bottles, 2–4 days old, were injected in the thorax with 32 nl of a bacterial cell suspension or PBS using a nanoinjector (Nanoject II, Drummond Scientific). For determination of CFUs, injected flies (6 females) were crushed immediately in media appropriate for the bacteria injected and the homogenates were diluted and plated on tryptic soy agar-media (TSA). The plates were incubated at 30°C for 20–30 hours and the colony forming units (CFUs) per fly were measured by counting the number of colonies on each plate, the CFUs per fly were used to adjust the initial dose of bacteria injected to approximately 100 CFUs per fly. For the time course (0,6, 17 hours) determination of CFUs, each value represents an arithmetic average derived from three biological repeat experiments (n = 3). Flies for survival and PGRP-SA mutant rescue assays were inoculated concurrently with those for determining CFUs, with ten or fifteen flies of each sex injected per bacteria-fly strain combination (or PBS-fly strain); each combination being repeated independently three times (n = 3). Following injection, flies were transferred to 30°C and survival assessed every 24 hours over a period of 3 days. Since the trends in survival were the same (i. e. survival curves were positioned similarly relative to one another) for each independent biological repeat, the data for each bacteria-fly strain combination was added (n = 60 or n = 90) and estimates of survival curves constructed. Flies injected with PBS were mostly unaffected for all fly backgrounds. A truncated version of PGRP-SA (in which the N-terminal sorting sequence was replaced with a T7 tag, and a poly-histidine tag was added to the C-terminus) was expressed in E. coli and purified using cobalt affinity resin (Talon; BD Biosciences) under denaturing conditions. A mCherry tagged derivative, mCherry-PGRP-SA was produced using the same procedure. Proteins were stored in 20 mM Tris-HCl pH 8. 0 and 150 mM NaCl. Functionality assays of the rPGRP-SA and mCherry-PGRP-SA proteins were performed as previously described [14]. Drs-GFP expression was monitored after 24 hours of the M. luteus infection through the production of fluorescent signal produced by the infected flies; and by qPCR using as template RNA extracted from 6 infected female flies, similar to what was previously described [50]. Peptidoglycan was prepared from exponentially growing cultures of S. aureus, B. subtilis, M. luteus, and E. faecalis as previously described [13]. 50 µg of recombinant PGRP-SA was incubated with 0. 2 mg of peptidoglycan and 17 µg of BSA (New England Biolabs) in 20 mM Tris-HCL pH 8. 0 and 300 mM NaCl in a final volume of 300 µl. Incubation was at 25°C with agitation for 30 minutes. Peptidoglycan and co-precipitated proteins were harvested by centrifugation, washed twice with 20 mM Tris-HCl pH 8. 0,300 mM NaCl and then resuspended in 1× SDS loading buffer, boiled for 5 minutes and run on 12% SDS PAGE mini gels. An aliquot of the supernatant, representing unbound protein, was also run. Gels were stained with Coommasie stain, destained and imaged using an ImageScanner (Amersham Biosciences/GE Healthcare). Quantifications of bands performed using ImageJ software [51]; each value represents an arithmetic average derived from three biological repeat experiments (n = 3). Bacteria were grown to mid-exponential phase. Washed cell cultures in PBS (500 µl) were incubated with 50 µl of mCherry-PGRP-SA (2 mg/ml in 150 mM NaCl, 20 mM Tris pH 8. 0) for 5 minutes on ice. The cells were washed twice with PBS and harvested at 4°C (3000 rpm, 10 minutes). Finally the bacteria were resuspended in 20 µl PBS. A drop of this culture was placed on a PBS, 1% agarose slide and visualised. Images were obtained using a Zeiss Axio ObserverZ1 microscope equipped with a Photometrics CoolSNAP HQ2 camera (Roper Scientific using Metamorph software, Meta Imaging series 7. 5) and analyzed using ImageJ software. WTA were extracted by alkaline hydrolysis from overnight cultures were analyzed by native polyacrylamide gel electrophoresis and visualized by combined alcian blue silver staining, as previously described [52]. ImageJ software [51] was used to quantify the percentage of WTA produced by each strain as previously described [29]. The signal intensity of each lane was quantified and normalized against the corresponding value for the wild type (considered as 100%). Tunicamycin minimum inhibitory concentration (MIC) assays were performed as previously described [28]. Overnight cultures of bacteria were grown in antibiotic free medium or in the presence of a subinhibitory concentration of tunicamycin (0. 8 ug/ml for E. faecalis – 17× less than the MIC - and 0. 4 ug/ml for S. aureus and S. saprophyticus – 32× less than MIC), that doesn' t interfere with the bacterial growth rate. For mCherry-PGRP-SA binding assays, overnight cultures were diluted 1∶100 into fresh medium, with or without tunicamycin at the appropriate concentration, and were grown until mid-exponential phase. For survival experiments, we used S. aureus overnight culture grown with tunicamycin as above described. As nonparametric tests lack statistical power with small samples, when required, data sets with three biological repeats (n = 3) were transformed to give a normal distribution (Lilliefors test, P>0. 05) and then checked for equal variance (Levene' s test, P>0. 05); subsequently, data was analysed using parametric tests. Data for the PGRP-SA-peptidoglycan co-precipitation assay was normal with equal variance, thus not transformed; One-way ANOVA was applied to the data. For the mCherry-PGRP-SA binding to bacteria assays data (n = 50) was non-normal but with equal variance, therefore nonparametric Kruskal-Wallis test followed by Dunn' s multiple comparison was applied. The complete CFU data set exhibited neither normality nor equal variance, and attempts to rectify this by transforming the data failed. Therefore, the data was separated into 6 groups, which were independently transformed via a Box-Cox transformation (Box-Cox returns a λ number, where a transformed data-point = data-pointλ – 1/λ) to give a normal distribution with equal variance, and statistical analysis performed as described. Firstly, for each bacterial strain (groups 1–3, graphical representations not shown), Repeated Measures Two-way ANOVA was used to look for differences over time and between the fly backgrounds. However, due to interactions between these two factors, Repeated Measures One-way ANOVA with 95% Tukey' s HSD Intervals was used to look for differences over time for each particular bacteria strain and fly background combination (i. e. 9 separate tests, data for each was normally distributed with equal variance). Secondly, at each time point (groups 4–6, Figure 3), Two-Way ANOVA was used to look for differences between the bacterial strains and between the fly backgrounds; where there was an interaction between these two factors, One-way ANOVA with 95% Tukey' s HSD Intervals was used to look for differences between the fly backgrounds for a particular bacterial strain. Estimated survival curves were constructed from the raw data sets and the Log-rank (Mantel-Cox) test used to determine statistical significance between the curves. For clarity in display, 95% confidence intervals have been omitted from the graphs. All data was plotted and analyzed using GraphPad Prism 5 (GraphPad Software, Inc.) or MATLAB R2009a. | Gram-positive bacteria such as the opportunistic pathogen Staphylococcus aureus have their cell wall exposed to the environment found within a host. Following an infection these bacteria need to find ways to evade or reduce recognition by the host in order to survive and potentially proliferate. The cell wall of Gram-positive bacterium consists of an intricate network of glycan chains cross-linked by short peptides called peptidoglycan (PG; a major target for host recognition in a variety of animals) covalently linked to surface proteins and glycopolymers including Wall Teichoic Acids (WTA). It has been proposed that lack of WTA reduce the pathogenicity of S. aureus. We asked whether this was due to better recognition of PG. We found that both bacterial recognition and survival of fruit flies (our model host) infected with bacteria lacking WTA was markedly increased compared to those infected with wild type S. aureus. This result was quantifiable: a reduction in the amount of WTA resulted in greater binding by host receptors and a higher host survival. We propose that the presence of WTA limit access to PG and therefore reduce the recognition ability of the host. Bacteria are thus able to increase in numbers and eventually overwhelm the host. | Abstract
Introduction
Results
Discussion
Materials and Methods | medicine
infectious diseases
model organisms
immunology
biology
microbiology
molecular cell biology | 2011 | Wall Teichoic Acids of Staphylococcus aureus Limit Recognition by the Drosophila Peptidoglycan Recognition Protein-SA to Promote Pathogenicity | 10,765 | 302 |
In convergent-extension (CE), a planar-polarized epithelial tissue elongates (extends) in-plane in one direction while shortening (converging) in the perpendicular in-plane direction, with the cells both elongating and intercalating along the converging axis. CE occurs during the development of most multicellular organisms. Current CE models assume cell or tissue asymmetry, but neglect the preferential filopodial activity along the convergent axis observed in many tissues. We propose a cell-based CE model based on asymmetric filopodial tension forces between cells and investigate how cell-level filopodial interactions drive tissue-level CE. The final tissue geometry depends on the balance between external rounding forces and cell-intercalation traction. Filopodial-tension CE is robust to relatively high levels of planar cell polarity misalignment and to the presence of non-active cells. Addition of a simple mechanical feedback between cells fully rescues and even improves CE of tissues with high levels of polarity misalignments. Our model extends easily to three dimensions, with either one converging and two extending axes, or two converging and one extending axes, producing distinct tissue morphologies, as observed in vivo.
Embryonic development requires numerous changes in tissue morphology. Convergent-extension (CE) is a basic tissue shape change [1–9], during which cells in an epithelial sheet rearrange to narrow (converge) the tissue along one planar axis while lengthening (extending) it along the perpendicular planar axis (Fig 1). Although CE has been observed in the development of many organisms [1–8], the specific cellular mechanisms that drive such movements are still subject of investigation [10]. Both asymmetric external forces on a tissue (passive CE) and asymmetric forces generated by the cells within a tissue (active CE) can lead to CE (Fig 1) [10]. Hypothesized mechanisms for CE include anisotropic cell edge/actin contraction [11,12], anisotropic cell adhesion and elongation [13,14], cell shape extension/retraction [11,15], combinations of a constraining boundary with undirected cell elongation [16] or with directed leading edge protrusion [17], and increased cell adhesion within tissue segments [18] (see Supplemental Material for a more detailed discussion of previous models). Existing models of CE, however, neglect the experimentally observed prevalence of filopodial extension parallel to the direction of tissue convergence [3,9, 19–24], which could produce anisotropic traction forces between cells or between cells and the extracellular matrix [25–28]. The observed asymmetry of filopodial protrusion led us to propose a filopodial-tension mechanism for CE based on anisotropic filopodial pulling forces between cells. We explicitly model the number of cell-cell connections, their range, angular distribution, strength, and frequency of formation and breakage. We define an appropriate set of metrics to quantify both the effects of model parameters and planar-polarization defects (such as misalignments and the passive cells) on the dynamics of tissue-level CE. Since our filopodial-tension model extends naturally to three dimensional tissues, we discuss the two types of 3D CE and their corresponding tissue morphologies.
Experiments show that long filopodia continuously form and retract during CE in epithelial sheets and that these filopodia preferentially form in-plane along angles near the axis of tissue contraction. Each model cells therefore extends and retracts filopodia (which we represent using the model concept of a link) distributed within a range of angles around the directions perpendicular to the cell’s planar-polarity axis. To simulate the observed binding of filopodial tips to other cells and the roughly length-independent pulling forces which retracting filopodia generate, in our model, an extending link binds to the cell it contacts, then generates a constant (length independent) tension force between the cells it connects [20,25,29]. We then test whether this tension force is sufficient to explain observed local cell intercalation and global tissue CE. In the filopodial-tension model (Fig 2, S1 Movie) cells form and eliminate links representing filopodia with a defined set of neighboring cells (terms in boldface identify model objects). Each cell carries a polarization vector (perpendicular to its planar-polarity axis) (Fig 2, red arrow) that defines its preferred direction of filopodial protrusion (Fig 2, blue horizontal line). We simplify the model by having the links connect the centers-of-mass of cells rather than connecting the actin cortex of one cell to the actin cortex of the contacted cell, as do real filopodia. Because filopodia typically form in a pair of growth cones roughly along the convergence axis and with a typical maximal length, we allow a cell to form links within a range of angles ±ϑmax around this axis on either side of the cell with a maximum length of approximately rmax. Specifically, a cell can form a link only with those cells whose centers-of-mass lie within a distance rmax from its center of mass and within an angle ±ϑmax of its polarization axis (Fig 2, blue horizontal line). A cell can have at most nmax links to other cells at any time (including links formed and received) and only one link is allowed between any pair of cells. The actual number of links a cell forms may be less than nmax. Each link between a pair of cells exerts a tension force of magnitude λforce along the line connecting the cells’ centers-of-mass. To model the finite lifetimes of filopodia, we define a relaxation time, tinterval, after which we remove the links of all cells and create new ones. In a simulation in which the links form and then persist indefinitely, the cells only move a few microns (lattice sites) from their original locations and the tissue does not converge or extend. We implement the filopodial tension model using the Cellular Potts model (CPM, also known as the Glazier-Graner Hogeweg model, GGH), where each cell is represented as a collection of lattice sites with the same cell index. An effective-energy cost function, H, specifies the cell’s properties (see supplemental material). The tension force along a link between a pair of cells is independent of its length and acts along the vector between their centers-of-mass. In the GGH/CPM formalism, the tension has the form: H=H0+∑σ, σ′λforce (σ, σ′) lσ, σ′ (1) where the sum is over all pairs of linked cells, λforce is the strength of the pulling force between cells σ and σ’, lσ, σ’ is the current distance between the cells, and the term H0 aggregates all the other GGH/CPM cost function terms. The GGH/CPM simulations evolve stochastically from random lattice-site updates subjected to the effective-energy cost function, H. The time unit is the Monte Carlo Step (MCS), defined as the rate of lattice-site updates (see supplemental material for more details on the GGH/CPM formalism). The filopodial-tension model has five intensive parameters (λforce, tinterval, rmax, nmax, ϑmax) and one extensive parameter (N, the number of cells), making a complete sensitivity analysis computationally costly. We therefore fixed all parameters to reference values that are within the ranges observed in vivo and produced biological plausible convergent-extension (Table 1), then studied the effects of varying each intensive parameter one-at-a-time. The biological parameters proposed by the model can be directly measured experimentally, but since the concept of a filopodial-based CE is new and applies more readily to CE of deep tissues, which are not as easily visualized as epithelial sheets, appropriate experimentally-derived values are harder to find. The most studied cases are chicken limb-bud mesenchymal intercalation [30] (tinterval = 2. 2 hours; rmax = 3 cell diameters; nmax = 11; ϑmax = 45°), Xenopus gastrulation and notochord formation [31–33] (tinterval = 2. 0–2. 7 min; rmax = 1. 5 cell diameters; nmax = 8–9; ϑmax = 60°), and Xenopus Keller explants [23,34] (tinterval = 0. 5–1. 0 hour; rmax = 1. 5 cell diameters; nmax = 8–9; ϑmax = 30°). All simulations start with a mass of identical cells uniformly distributed inside a rough circle. Each cell has the same planar-polarization vector (V). To quantify the degree of tissue deformation we calculate the inverse aspect ratio between the length of the minor (L-) and major (L+) axes of the tissue (Fig 3B, green line). Initially the aspect ratio is close to 1 and decreases in time to a final value κ (Fig 3B, dashed red line) that depends on the filopodial tension parameters (λforce, tinterval, rmax, nmax, ϑmax), the number of cells in the tissue (N) and the surface tension of the tissue γ (defined below). The final inverse aspect ratio quantifies the maximum elongation of the tissue, but does not convey how fast the tissue elongates. To quantify the elongation rate, we define the elongation time (τ) the time an initially isotropic tissue takes for its major axis (L+) to double the length of its minor axis (L-), which is equivalent to the time when the inverse aspect ratio (L-/L+) first decreases to 0. 5 (Fig 3B, dashed blue lines). We consider CE to fail if L-/L+ never reaches 0. 5. Since both the filopodial-tension model and the GGH/CPM are stochastic, we average the value of the elongation time (τ) over 10 simulation replicas. Because the tissue inverse elongation ratio converges to the same value independent of the simulation seed or initial conditions (S1 Fig), unless specified otherwise, we calculate the final inverse aspect ratio κ for a single simulation replica, with the standard deviation indicating the fluctuations in κ around its final value for that replica.
Successful CE depends on the ability of intercalating cells to generate forces stronger than the internal and external forces that oppose tissue deformation. Here, the opposing forces come from the superficial tension (γ) between the cells and the external medium, defined as [35]: γ=Jc, M−Jc, c2, (2) where Jc, c is the contact energy between cells and Jc, M is the contact energy between cells and medium (see supplemental material). When the filopodial tension is weak compared to the surface tension (λforce < 2γ), cells do not intercalate and CE fails. For larger filopodial tensions, the elongation time (τ) decreases as a power of λforce (τ ∝ λforce-1. 25±0. 03) (Fig 4A, red line). The final inverse aspect ratio (κ) decreases monotonically with increasing λforce (Fig 4B). Increasing γ shifts the κ vs. λforce curve to the right and decreasing γ shifts the κ vs. λforce curve to the left (Fig 4B, inset). Normalizing the filopodial tension by the surface tension (λforce/γ) collapses the κ vs. λforce curves (Fig 4B), showing the linear relationship between λforce and γ. The surface tension (γ), however, has little effect on the elongation time (τ), which depends on λforce, but is relatively insensitive to surface tension (Fig 4A). The κ vs. λforce/γ curve is sigmoidal on a log-log scale (Fig 4B), because the shape of the tissue changes little for weak filopodial tensions and because the total number of cells limits κ for strong filopodial tensions (see Fig S3A). At the inflection point of κ vs. λforce/γ, the tensions of the links (λforce) balances the external surface-tension forces that oppose tissue elongation (λforce/γ ~ 6). Near this inflection point κ varies as an approximate power law of λforce (κ ∝ λforce-1. 51±0. 08). Next we studied how the remaining filopodial tension parameters affect CE, specifically, the mean lifetime of the filopodia, modeled as the time interval between link formation and breakage (tinterval); the maximum length of the filopodia, modeled as the maximum distance of interaction between the cells’ centers-of-mass (rmax); the maximum number of filopodial interactions per cell (nmax); and the maximum angle between the filopodial direction and the cells’ convergence axis (ϑmax). Fig 5A shows that, for the reference parameter values (Table 1) the lifetime of filopodia, tinterval, has no effect on τ or κ for tinterval ≲ 200 MCS. For the reference parameter values, 200 MCS corresponds to the typical time the cells require to rearrange their positions in response to a given set of neighbors interactions. Increasing filopodial lifetimes above 200 MCS slows cell intercalation (increasing the elongation time) and increases the tissue’s final inverse aspect ratio (corresponding to less deformation). The maximum range (rmax) of filopodia interaction has different effects on the final inverse aspect ratio (κ) and elongation time (τ). For rmax < 2 cell diameters, κ decreases as a power law in rmax (κ ∝ rmax-3. 5±0. 2), then saturates for rmax ≥ 2 cell diameters, while the elongation time (τ) decreases monotonically with increasing rmax (Fig 5B). The same effect is seen with respect to the maximum number of links (nmax): κ decreases as a power law in nmax for nmax < 4 (κ ∝ nmax-1. 5±0. 03) and saturates for nmax > 4 (this saturation makes sense since the cell typically has 4 neighbors within the range of its filopodia for rmax = 2 and ϑmax = 45°) while τ decreases monotonically with increasing nmax (Fig 5C). Thus rmax and nmax have affect the rate of cell intercalation more than the final inverse aspect ratio, while the tissue’s surface tension affects only the final inverse aspect ratio and not the rate of cell intercalation (Fig 4). Both κ and τ are concave with respect to the maximum angle of filopodial protrusion (ϑmax), since for small ϑmax the number of cell center-of-mass within the cones defined by ϑmax and rmax is very small, while for ϑmax = 90° the forces on the cell are symmetric since it extends filopodia uniformly in all directions. In both limits CE fails (Fig 5D). Since the net intercalation force is the difference between the tension forces parallel and perpendicular to the convergence axis (roughly ∫0θmax (cos (θ) -sin (θ) ) dθ), we might expect the force to be greatest (and thus κ and τ to be smallest) when ϑmax = 45° and for their values to increase symmetrically away from ϑmax = 45°. The curves, however, have different minima and are not symmetric: the smallest final inverse aspect ratio (κ) is around ϑmax = 40° (Fig 5D, red dots) and the smallest elongation time (τ) is around ϑmax = 30° (Fig 5D, blue squares). This asymmetry is caused by the limited number of neighbors with which a cell can form a link. Both the maximum number of links per cell (nmax) and the number of cells within the link interaction range (rmax) can limit the actual number of links a cell forms. If the maximum number of links per cell is lower than the number of cell neighbors within a cone of range rmax (e. g. nmax = 3) and angle ϑ < ϑmax, increasing ϑmax leads to more links with cells at larger ϑ and thus reduces the net tension force applied along the direction of the convergence axis. In effect, large ϑmax causes the cell to waste its limited number of filopodia. For large nmax, links form to all cells within the cone of range rmax and small θ regardless of the value of ϑmax. Thus, for large nmax (e. g. nmax = 7), the κ and τ curves are roughly symmetrical around their minima at ϑmax ~ 45° (blue and red lines in Fig 5D). The filopodial tension model assumes that cells can extend filopodia, contact and pull other cells that lie within a given distance, even if they do not touch each other before filopodial extension. An example would be the formation of adhesion junctions between cells which coupled to a contractile stress fiber in both cells. To model these cases, we defined a contact-mediated cell tension model, which is identical to the filopodial tension model except that the maximum link length rmax in the filopodial tension model is replaced with the condition that cells must be in touch before pulling on each other (Fig 6A). The qualitative results for the contact-mediated cell tension model do not differ much from the filopodial tension model. The κ x λforce curve is sigmoidal on a log scale, τ decreases with a power law (κ ∝ λforce -1. 18±0. 06) and CE fails for λforce < 20 (Fig 6B). The dependence of κ on the number of filopodial interactions (nmax) is still a power law (κ ∝ nmax-1. 5±0. 03) and saturates when nmax = 4. The elongation time (τ), however, does not keep decreasing as it does for the filopodial tension model, but also saturates around nmax = 4 links (Fig 6C), as few cells have more than 4 neighbors with centers near the convergence plane. The (κ, τ) x ϑmax curves have minima at ϑmax = 40° and ϑmax = 35°, respectively, but are less skewed than in the filopodial tension model (compare Figs 6D and 5D). CE fails for ϑmax < 10° and ϑmax > 70°. Convergent-extension requires cells to have consistent planar polarity throughout an extensive region of tissue. This correlated orientation might result from a long-range bias from a morphogen gradient, cellular or intercellular differences in protein expression [36], or from a boundary-relay mechanism [37,38]. In our previous simulations we assumed that all cells had perfectly aligned polarization vectors (Fig 2, red arrows), i. e. , they all pointed in the same direction with the same magnitude, and they maintained their internal orientation throughout the simulation. To study the effect of polarization misalignment on CE we added a zero-mean Gaussian distributed displacement angle to the cells’ polarization vectors and varied the standard deviation of the distribution (σ) while keeping the mean direction (here, the vertical axis) constant. Since the final elongation ratio is sensitive to the distribution of polarization vectors, the values of κ were averaged over 5 simulations. The filopodial tension model tolerates small polarization misalignments, with a tissue with a displacement angle of σ = 10° reaching the same final inverse aspect ratio as in the perfectly aligned case with little decrease in elongation rate (an 11% increase in τ). The tissue remained aligned with the mean direction of cell polarization (the vertical axis) for small misalignments (σ < 40°, Fig 7B), but bent at around σ = 50° (Fig 7C). For polarization misalignments with σ > 60°, CE fails and the tissue breaks its symmetry, acquiring more complex shapes such as the caltrop (see Fig 7D). Both metrics are exponential functions of the variance σ2 (Fig 7A). So far we have assumed that the polarization vector of the cells remains constant throughout the entire process. In reality, however, cells are constantly communicating with their neighbors either through signaling or through mechanical interactions. During CE cells establish and maintain their polarity through the planar cell polarity (PCP) pathway, however there is growing evidence that mechanical feedback may also play a role in the maintenance of global tissue polarity during development [39–42]. The presence of a mechanical feedback mechanism may rescue CE in tissues with high polarization alignment defects. In order to investigate this we developed a simple model of mechanical feedback and applied it to the misalignment polarization cases that have been described in the last section (Fig 7). The feedback model assumes that the pulling forces on a cell due to filopodial interactions affect its polarization vector. We implement a simple phenomenological version of such an interaction by calculating the line of tension from the sum of all filopodial interactions of the cell with its neighbors (Fig 8A). From this line of tension we extract an orthogonal vector T which is averaged with the previous cell polarization vector V in the following way: Vt+Δt=Vt* (1−w) +Tt*w, (3) where Vt+Δt is the polarization vector of a cell at time t+Δt, Vt* is the normalized polarization of the same cell at time t, Tt* is the normalized tension vector of the cell at time t, and w is a feedback weighting factor ranging from 0 (no feedback) to 1 (no memory) (Fig 8A). This iterative processes repeated at discrete time intervals set equal to filopodia lifetime (Δt = tinterval). For simplicity, we do not distinguish between the pulling forces generated by the cell from the pulling forces that their neighbors exert on it. The normalized tension vector of a cell is calculated from the vector that maximizes the sum of all projections of the normalized lines of force from all the cells’ neighbors: Σi cos (∅i− ∅T) = 0, (4) where the sum is over all the cell’s neighbors that pull on it, ϕi is the angle of the line of force between the cell and the pulling neighbor i, and ϕT is the angle that defines the tension vector T* = (cos ϕT, sin ϕT). For tissues with a high starting level of polarization misalignment (σ ≥ 40°), addition of this mechanical feedback mechanism usually leads to a lower final elongation ratio κ, as long as the feedback factor w is below 0. 1. For these cases, tissue elongation times (τ) decrease with higher feedback levels (Fig 8C, green and red lines), while k usually decreases with lower feedback levels. For tissues with a low starting level of polarization misalignment (σ ≤ 20°), addition of this mechanism leads to lower final elongation ratios only for small levels of feedback (w ≤ 0. 001) (Fig 8B, blue and black lines), while the time of tissue elongation (τ) remains relatively unchanged with respect to the case with no feedback (Fig 8C, blue and black lines). In all simulations where the addition of mechanical feedback rescues CE, the cells established a global polarization axis emergently. We chose to implement the feedback update iteratively, which leads to fast destabilization of the tissue for high levels of feedback, as is expected for a case with no memory. Although a continuous model would be relatively more robust, we expect the same destabilization effect when the weighting factor w approaches 1. Next we varied the number of intercalating cells in the tissue to check if there is a minimum number of active cells needed to drive CE and how this change tissue dynamics. We defined two types of cells without filopodia: passive cells, which lack filopodia but can be pulled by the filopodia of other cells; and non-responsive, or refractory cells, which cannot be pulled by the filopodia of other cells. The former would correspond to cells whose surface adhesion molecules were compatible with those of the cells extending filopodia and the latter to cells with incompatible adhesion molecules. The parameters for cells which produced filopodia were the same as in Table 1. Since the final elongation ratio is sensitive to the distribution of active/non-active cells, the values of κ were averaged over 5 simulations. For tissues with a mixture of active and passive cells, both κ and τ decrease monotonically with the percentage of active cells in the tissue (Fig 9, red dots). However, even a fraction of active cells can drive CE. For 40% or more active cells (S2 Movie), the tissue deforms almost as much as a tissue composed entirely of active cells (Fig 9B, red dots), though the elongation time increases with the percentage of passive cells up to twice that for a tissue of all active cells (Fig 9A, red dots). For higher fractions of passive cells the final inverse aspect ratio increases significantly with the fraction of passive cells (Fig 9B). E. g. , for 90% passive and 10% active cells (Fig 9C and S3 Movie), the tissue’s final inverse aspect ratio never drops below 0. 3 (Fig 9B) and the elongation time τ is more than ten times that for a tissue of all active cells (Fig 9A). In all simulations, the active cells migrate towards the midline of the elongating tissue, leaving the passive cells at the lateral margins (Fig 9D). The presence of relatively high CE despite the presence of only 10% of active cells can be explained by the relative tissue area covered by the filopodia. Every active cell can pull on neighbors that lie up to a distance rmax from its center of mass and within an angle ϑmax on each side of its convergence plane (see Fig 2), thus covering an area of 2ϑmaxrmax2. For the reference parameters (ϑmax = π/2 and rmax = 2 cd, see Table 1) and a population of 10% of active cells (N/10, where N is the number of cells in the tissue) this amounts to ~1. 2N (cd) 2, which more than covers the whole area of the tissue (N (cd) 2). Refractory cells have a stronger effect on CE than passive cells. CE fails when the percentage of refractory cells is above 60% (Fig 9B, blue squares), while with passive cells it only fails for percentages higher than 90% (Fig 9B, red squares). For higher fractions of active cells, the two populations sort out, with the active cells extending normally and the refractory cells displaced to both sides of the elongating tissue (Fig 9F). Surface tension between the cells and the surrounding medium causes the refractory cells to form droplet-like clusters which bend the extending active-cell tissue into a wavy bar (Fig 9F and S4 Movie). The 2D filopodial tension model is a reasonable description of cells within epithelial sheets, where cell movement is confined to a plane. However, in many situations cell intercalation occurs in 3D. That is the case in radial intercalation during epiboly of the developing Xenopus Laevis embryo, where cells in a multilayered epithelium intercalate and converge perpendicular to the plane of the sheet [43]. The filopodial tension model can be easily extended to three dimensions, but due to the extra degree of freedom, it breaks in two versions, depending on which axis is rotated: In equatorial or extensional intercalation, obtained by rotating the 2D model around the polarization vector (the red arrow in Fig 2), the cells pull on all neighbors that lie in a convergence plane (Fig 10A). At the tissue level, equatorial intercalation results in the convergence of the tissue along the two directions perpendicular to the polarization vector and its extension along the polarization vector (Fig 10A’ and 10A”). In bipolar or convergent intercalation, obtained by rotating the 2D model around the convergence line (the blue line in Fig 2), the cells pull on all neighbors that lie along a convergence axis (Fig 10B). At the tissue level, the bipolar intercalation results in the convergence of the tissue along the axis of convergence and its expansion in the other two directions (Fig 10B’ and 10B”). Beginning with a spherical tissue with all the cells polarized in the same vertical direction, the 3D equatorial model produces a tissue resembling a prolate spheroid (cigar shaped, Fig 10A”, S5 Movie), while the bipolar model produces a tissue resembling an oblate spheroid (lentil shaped, Fig 10B”). The bipolar model has more biological correspondence than the equatorial model: cells with unipolar or bipolar protrusive activity are much more common during development than cells with equatorial protrusive activity, and the resulting tissue shape from the 3D bipolar model corresponds to the thinning and expansion associated with radial intercalation. For both versions of the 3D model, the dependence of the parameters κ and τ with λforce, rmax, nmax and tinterval are qualitatively the same as in the 2D model. The results only differ qualitatively with respect to ϑmax. For the same values of rmax and nmax, the 3D convergence model is slightly less skewed than the 2D version, with the best value for κ around ϑmax = 45° and the best value for τ around ϑmax = 35° (Fig 11B). The 3D extension model, however, presents a more drastic change in the (κ and τ) vs. ϑmax curve when compared to the 2D. While the 3D convergence model was slightly more symmetrical around ϑmax = 90°, the 3D extension model is very skewed towards small angles, with the best values for κ around ϑmax = 30° and the best value for τ around ϑmax = 15° (Fig 11A). In the extensional model CE fails for ϑmax < 3° and ϑmax > 60°, while in the bipolar model CE fails for ϑmax < 10° and ϑmax > 75°. The reasons for the asymmetry is that the final shape of the tissue in the 3D extension model—a two cell diameter tube orthogonal to the convergence plane—is more sensitive to perturbations than the lentil shape tissue obtained by the 3D bipolar model. Two cells pulling each other along a convergence axis leads them to be aligned in a plane perpendicular to the direction of the pulling. This plane fully coincides with the 3D bipolar model extension plane (Fig 10B”), but only partially with the extension plane of the 3D extension model (Fig 10A”). In the simulations results shown in Fig 11A, the value of ϑmax = 30° represents the optimal maximum angle value where the pulling forces in the 3D extension model are still able to align the tissue without destabilizing it.
Here we developed a new 2D model of for active cell intercalation. The model drastically differs from previous existing models by its explicit use of pulling forces between cells rather than anisotropy on adhesion energies or surface tensions on the cell surface. A recent experimental work by Pfister et al. [28] supports our force-driven model. The use of forces makes the model easily adaptable to three dimensions, but the extra degree of freedom gives two ways in which this can be achieved: either by rotating the model around the polarization vector or around the convergence line (see Fig 10A and 10B). The model was implemented in CompuCell3D using the Cellular Potts formalism. The core of the model, however, is independent of the mathematical formalism and can be easily implemented on other types of agent-based formalisms such as cell-center or vertex models, as long as they provide a volume exclusion mechanism. Although some quantitative results might differ, we expect the same qualitative results. The advantages of implementing and simulating our model using the Cellular Potts formalism include the ease to manipulate and study the effects of the surface tension on the tissue dynamics (Fig 4) and the addition of the model as an integrated part of the CompuCell3D simulation package [44] that allows for it to be immediately reusable by others. Model validation can be done at two different levels quantitative and semi-quantitative: when both microscopic and macroscopic measurements are available for a specific tissue, using model parameters that agree with those measured for the cells in the tissue should result in tissue-level model output (here, the rate of convergence and final aspect ratio) agreeing quantitatively with that of the experimental tissue. This agreement should persist under different experimental conditions. This form of validation shows that the hypothesized mechanisms included in the model are sufficient to reproduce the experiment quantitatively. Note that a model can only show the sufficiency of modeled mechanisms, not their necessity, since a different set of mechanistic hypotheses might yield the same results. In our case, since we are not modelling a specific tissue, model validation can only be semi-quantitative. We show that changes in the properties of the modeled cells (such as average number of filopodia, length and angular distributions) and tissue properties (such ratio of active and non-active cells) predict relative changes in the rate of CE and final tissue aspect ratios in real tissues that agree with those in experiments when the corresponding parameter and conditions are similarly modified. In this case, agreement demonstrates the plausibility of the hypothesized mechanisms, but detailed quantitative validation requires additional experimental measurements. We hope that our semi-quantitative analysis of the filopodial tension model of CE will inspire the additional experimental measurements that a more detailed quantitative validation requires. Our model predicts that external forces, such as surface tension and pressure, can only affect the final degree of tissue elongation (κ) (Fig 4), whereas the internal parameters that regulate cell-intercalation can affect both tissue dynamics (as measured by τ) and the final tissue shape (Figs 4 and 5). This can be easily tested experimentally by either changing the properties of the external environment (the surrounding cells/matrix) of the intercalating tissue or culturing it ex vivo. Of the five cell intercalation parameters, the time interval between link formation/breakage (tinterval) had negligible effects as long as it is below the typical time that the cells take to rearrange positions and/or shapes in response to a given set of external forces. This might be different if a refractory time interval between pulls is added to the model. We expect that in the presence of such refractory time, an increased frequency in link formation/breakage would slow down the speed of intercalation and reduce the final elongation ratio. The model also predicts that the maximum range of cell interaction (rmax) and the maximum number of links per cell (nmax) had no effect on the final tissue elongation after rmax = 2 and nmax = 3, but the time of elongation kept decreasing for higher values of rmax and nmax (Fig 5B and 5C). It was not possible to increase those parameters indefinitely to check if the speed of intercalation would also saturate because the simulated cells start to fragment past rmax ≥ 6 or nmax ≥ 7. The current implementation of the model, however, does not allow for more than one active link between cell pairs, which would likely decrease elongation time. All cell intercalation models assume some type of increased cell activity along the convergence axis, which is often translated, as is the case here, into the assumption that the cells are bipolar (one exception being our 3D equatorial/extensional filopodial tension model). This however is not necessarily true and we expect the model to also work in cases where the simulated cells are either monopolar in opposite directions of the same convergence axis or randomly alternate being monopolar in each direction of the convergence axis. Our model also suggests that CE can be successfully achieved even in the presence of relatively high degrees of polarization defects. We predict that tissues containing polarization misalignments of up to ±10° will be practically indistinguishable to the optimally aligned case. Even severe misalignments (about ±50°) would still lead to some CE, although to a much lesser degree of final elongation ratio and with longer tissue elongation times (Fig 8). Experimental disruptions of the PCP pathway that alter the global alignment of cells in a dose-dependent manner would provide a way to test some of these predictions. Addition of a simple mechanical feedback mechanism by which the cells readjust their polarization in response to the pulling forces from the neighbors does not have major effects on the speed of tissue elongation (Fig 8C), but can fully rescue and even improve on the final elongation ratio of tissues with low or even severe polarization misalignments (Fig 8B). We choose to implement a minimal phenomenological feedback mechanism to explore the general response and self-organization of the tissue, but the formulation of the model allows the replacement of this generic mechanism with more detailed feedback models that reflect a specific tissue. In cases where some cells fail to polarize, the severity of the effects on CE will depend on the type of interaction between the polarized (intercalating) cells and the unpolarized (non-intercalating cells). If the polarized (or active) cells can still pull on the unpolarized cells, then CE still happens even in a situation where the vast majority of cells (95%) are not active (Fig 9), although at a great reduction in both speed and final elongation ratio. If, on the other hand, the unpolarized cells are non-responsive and cannot be pulled, then the reductions in speed and final elongation ratio are much more sensitive to the presence of unpolarized cells (Fig 9) and CE completely fails when the population of active cells falls below 25% (Fig 9C). Another prediction of the model is the separation between the intercalating cells and the non-responsive/refractory cells (Fig 9D). Such defects could be induced experimentally by randomly distributed knock-out of intercalating cells, e. g. using electroporation of tissues with a dominant negative or RNAi, and would provide a way to further test model predictions. Finally, the model reduces to the more common implementations when the maximum range of interaction is replaced by the common contact area condition. In this case, instead of contracting (or increasing the tension of) the cell’s surfaces that are aligned with the polarization vector or the global direction of convergence, we pull the neighbors that are not aligned with it (Fig 6D). In both cases active CE is achieved by the same principle, promoted cell-cell activity along one axis and inhibition along the other. | The development of an embryo from a fertilized egg to an adult organism requires not only cell proliferation and differentiation, but also numerous types of tissue restructuring. The development of a relatively round initial embryo into one elongated along its rostral-caudal axis involves coordinated tissue elongation and cell reorganization in one or more groups of cells or tissues. Counterintuitively, in many organisms, cells in elongating tissues elongate and increase their protrusive activity in the direction perpendicular to the axis of elongation (convergent extension). Experimental and theoretical studies have not determined how this cell-level oriented protrusive activity leads to observed tissue-level changes in morphology. We propose a filopodial-tension model that shows how tension from oriented cell protrusions leads to observed patterns of tissue CE. | Abstract
Introduction
Methods
Results
Discussion | cell physiology
classical mechanics
fluid mechanics
condensed matter physics
anisotropy
surface tension
geometry
aspect ratio
simulation and modeling
cell polarity
developmental biology
mathematics
materials science
damage mechanics
embryos
morphogenesis
research and analysis methods
embryology
deformation
continuum mechanics
physics
cell biology
planar cell polarity
biology and life sciences
physical sciences
material properties | 2016 | Filopodial-Tension Model of Convergent-Extension of Tissues | 9,185 | 196 |
Recent outbreaks of locally transmitted dengue and Zika viruses in Florida have placed more emphasis on integrated vector management plans for Aedes aegypti (L.) and Aedes albopictus Skuse. Adulticiding, primarily with pyrethroids, is often employed for the immediate control of potentially arbovirus-infected mosquitoes during outbreak situations. While pyrethroid resistance is common in Ae. aegypti worldwide and testing is recommended by CDC and WHO, resistance to this class of products has not been widely examined or quantified in Florida. To address this information gap, we performed the first study to quantify both pyrethroid resistance and genetic markers of pyrethroid resistance in Ae. aegypti and Ae. albopictus strains in Florida. Using direct topical application to measure intrinsic toxicity, we examined 21 Ae. aegypti strains from 9 counties and found permethrin resistance (resistance ratio (RR) = 6-61-fold) in all strains when compared to the susceptible ORL1952 control strain. Permethrin resistance in five strains of Ae. albopictus was very low (RR<1. 6) even when collected from the same containers producing resistant Ae. aegypti. Characterization of two sodium channel kdr alleles associated with pyrethroid-resistance showed widespread distribution in 62 strains of Ae. aegypti. The 1534 phenylalanine to cysteine (F1534C) single nucleotide polymorphism SNP was fixed or nearly fixed in all strains regardless of RR. We observed much more variation in the 1016 valine to isoleucine (V1016I) allele and observed that an increasing frequency of the homozygous V1016I allele correlates strongly with increased RR (Pearson corr = 0. 905). In agreement with previous studies, we observed a very low frequency of three kdr genotypes, IIFF, VIFF, and IIFC. In this study, we provide a statewide examination of pyrethroid resistance, and demonstrate that permethrin resistance and the genetic markers for resistance are widely present in FL Ae. aegypti. Resistance testing should be included in an effective management program.
Local vector control programs play a part in public health. In many countries, these programs serve as the primary defense against the spread of several mosquito-borne diseases. Effective Integrated Vector Management (IVM) programs rely on surveillance information coupled with multiple vector control strategies such as chemical adulticiding to reduce vector populations and arbovirus transmission. Limited recent transmission of locally acquired dengue and Zika viruses in the southeastern continental US, primarily Florida, has brought renewed attention to the importance of IVM programs where the potential vectors, Aedes aegypti (L.) and Ae. albopictus Skuse, have long been present. However, for IVM programs in Florida to effectively control Aedes vectors and to reduce dengue and Zika virus transmission during an outbreak, it is essential to know which adulticide products are effective against local Ae. aegypti and Ae. albopictus strains. Organophosphates and pyrethroids are the only two classes of insecticides available to public health agencies for control of adult mosquitoes in the US. Compared to organophosphate insecticides, pyrethroids have higher public acceptance, rapid knockdown, relatively low costs, and are generally the product class of choice when adulticiding is required [1]. Unfortunately, years of pyrethroid insecticide use and previous DDT usage has increased the frequency of genetic resistance and enhanced enzymatic detoxification in insects like mosquitoes. Genetic target site changes of the sodium channel, known as knockdown resistance (kdr) mutations, are a relatively common insect response to selective pressure by pyrethroids [2,3]. Although a variety of other single nucleotide polymorphisms (SNPs) have been noted in Ae. aegypti, the primary two SNPs assessed to determine pyrethroid resistance are at codons 1016 and 1534 (positions according to standard M. domestica notation) [4,5]. One allele of the 1016 mutation is geographically distinct to the Western hemisphere and results in the replacement of the normal valine with an isoleucine (V1016I), while the 1534 mutation results in a phenylalanine to cysteine change (F1534C). There is currently debate about the individual toxicological effect of these two mutations, but they are consistently present in resistant strains [4]. Heterozygous and homozygous combinations of these two SNPs could result in nine possible genotypes, but strong linkage between the SNPs has been noted with only six of the genotypes observed in a study of Mexican Ae. aegypti [6]. More recent work indicates that the kdr genotype in Ae. aegypti may be more complicated than the combination of 1016I and 1534C SNPs. An additional SNP at position 410 has been identified and this valine to leucine (V410L) change, in addition to the other two SNPs, may be the strongly resistant phenotype [7]. The noted strong linkage between 410L and 1016I indicates that assessment of 1016 may still be adequate to assess the strongly resistant phenotype [7]. As expected based on the dilocus genotype distribution, several of the trilocus genotypes were also not frequently present in the wild [8]. Pyrethroid resistance and the distribution of kdr alleles have been well documented in Ae. aegypti strains from Central America, South America and the Caribbean [9–13]. Recent testing as part of the Zika emergency response in Puerto Rico has shown that isolated, early reports of resistance were indicative of widespread resistance on the island [13–16]. Studies have also shown the pyrethroid resistance is not specific to a particular pyrethroid but is generally class-wide to type I, type II, and non-ester pyrethroids [13,17,18]. Kdr alleles and pyrethroid resistance are widely distributed in Mexico, including Nuevo Laredo which lies just south of the US-Mexico border [11]. In contrast, little has been published about pyrethroid resistance in Aedes strains within the continental US. In the Garcia et al. [11] study mentioned above, the authors did not find kdr alleles in a Houston, TX collection from 1999. Recent reviews of the Aedes resistance literature listed no reports of resistance in continental US Ae. aegypti [4,19], but Cornel et al. [20] recently demonstrated toxicological resistance and sodium channel mutations in invasive populations of Ae. aegypti in California. Two recent studies using CDC bottle bioassays do indicate resistance in strains from the southern US, but these studies do not provide any quantification of the strength of the resistance nor did they examine the presence of kdr alleles [21,22]. In contrast, Ae. albopictus does not appear to frequently develop kdr resistance, and most US strains tested thus far have only shown very minimal pyrethroid resistance [4,23,24,25]. Resistance surveillance is recommended by the CDC and statewide initiatives to map pyrethroid resistance have begun in Florida and California [22,26]. Resistance information is critically important for operational decision making as part of an effective IVM program. In this study, a collaborative group of government, academic, industry and vector control district stakeholders collected Ae. aegypti and Ae. albopictus adults, eggs or larvae from more than 200 locations throughout Florida to assess the extent and intensity of pyrethroid resistance. The goal of this study was to improve vector control operations by producing a resistance map and apply this information to make more effective control decisions. We determined permethrin resistance ratios (RRs) relative to the susceptible ORL1952 strain for 21 wild type strains of Ae. aegypti and 5 strains of Ae. albopictus using direct topical application, the WHO gold standard assay for determination of intrinsic toxicity [27]. While the CDC bottle bioassay is a more commonly used assay, it is a threshold assay and does not quantitate resistance levels which was one of our primary goals. Nearly 5,000 Ae. aegypti from numerous locations were genotyped by allele-specific PCR to assess the frequency of V1016I and F1534C alleles and rapidly visualize the pattern of resistance throughout the state of Florida.
The toxicological profiles and rearing procedures for the susceptible (ORL1952) and resistant (Puerto Rico–BEIResources, NR-48830) strains of Ae. aegypti used in this study have been described previously [17]. Strain information, including specific location information, collectors, and dates of collection are noted in Table 1. Field strains for toxicology testing were often collected as mixed (Ae. albopictus and Ae. aegypti) eggs laid on seed germination paper (Anchor Paper, St. Paul, MN) placed in a variety of oviposition containers including black plastic stadium cups, plastic cemetery vases, and glass containers. Field collected eggs were hatched by soaking papers for 48–72 hours at 27 oC in rearing trays of untreated well water or deionized water (diH2O). Papers were carefully removed, and larvae were reared through adult emergence with a reduced feeding regimen compared to the standard rearing protocol for the ORL1952 susceptible strain [28] due to sensitivity to overfeeding. Strains from Monroe, Seminole, Orange, Hernando and Sarasota counties were collected as larvae from a variety of manmade and natural containers (tires, plant pots, bottles, buckets, etc.), rinsed with diH2O and then placed into the standard larval rearing procedure described above. Pupae were collected from rearing trays and placed in emergence chambers or 12” x 12” screen cages (BioQuip Models 1425 & 1450B, Rancho Dominquez, CA). After emergence, wildtype mosquitoes were briefly chilled to 4 oC and sorted to species. Strains for toxicology testing were produced from locations that had more than 30 wildtype founders. If multiple nearby locations were combined to produce a strain, the GPS location in Table 1 represents the GPS centroid of the sites that contributed the mosquitoes. Individual oviposition cup locations and life stages tested are provided in S1 File. Eggs were produced using standard rearing methods [17]. Colony strains were provided 2–10 bloodmeals on weekly intervals to collect F1 eggs. Warmed bovine blood was provided to produce 2,000 or more eggs per strain. If feeding was poor, warmed blood was spiked with 1 mM ATP as a phagostimulant. Bovine blood for mosquito feeding was purchased by the USDA under contract from a local, licensed abattoir. Blood was collected during normal operations of the abattoir from the waste stream after animal slaughter. Under CFR9, Parts 1–3, tissues, including blood, collected from dead livestock intended as food are exempt from IACUC regulation. Aedes aegypti strains from St. Augustine, Clearwater, and Vero Beach, FL were provided as F1 eggs produced at Anastasia Mosquito Control District, Pinellas County Mosquito Control, and the Florida Medical Entomology Laboratory, respectively. To produce mosquitoes for bioassay testing, eggs from the F1 or F2 generation of field collected strains, the lab susceptible strains (ORL1952) and the pyrethroid-resistant strain were hatched and reared as described above in covered trays at a density of approximately 1,000 mosquito larvae/tray. Adult mosquitoes emerged into 12” x 12” screen cages and provided with cotton saturated with 10% sucrose in diH2O. Females used for bioassays were 3–7 days post-emergence. The adult topical bioassay has been described previously in detail [28,29]. For these studies, the permethrin stock solution in DMSO (Product #N-12848-250MG, Chemservice, Westchester, PA) and all dilutions in acetone were prepared gravimetrically. An initial 10-fold dilution series was prepared over a range of relevant concentrations [17]. Sub-dilutions were prepared from the 10-fold dilution series as necessary to determine the critical region of the dose curve. Three assays (n = 3) were performed for all strains with listed LD50s unless limited numbers of F1 test mosquitoes allowed only two replicates. Average mass/female was calculated for each strain by weighing a cohort of 50–100 females before each replicate. LD50s, standard errors, and goodness of fit were determined from dose-mortality curves using SigmaPlot v13 (Systat software Inc. , San Jose, CA) with data fit to a four-parameter logistic model. To provide a comparative metric between strains that may have different body sizes, doses applied were divided by the average mass of the mosquitoes before curve fitting. This results in an LD50 of ng of active ingredient per mg of mosquito. Resistance ratios for strains were calculated using the LD50 of the field strain divided by the LD50 of the susceptible ORL1952 strain included in the same assay. In this study, we use the WHO scale to define the levels of resistance [30]. When RR is less than 5, the field population is considered susceptible. When the RR is between 5 and10, the mosquitoes are considered to have moderate resistance. A RR greater than 10 indicates that mosquitoes are highly resistant [31]. Genotypes for individual mosquitoes or eggs were determined using melt curve analysis with previously described allele-specific primers for the 1016 and 1534 SNPs [32,33]. Preliminary assays were conducted using eggs and adults from the same generation to verify that the resulting frequencies were similar (within 10%) between life stages rather than showing agenotype bias due to varied hatch rates or differing larval mortality (S2 File). Assays were performed in 96 well plates on a StepOnePlus (Applied Biosystems) or QuantStudio5 (Applied Biosystems, Thermo Fisher Scientific, Waltham, MA) using SYBR green chemistry. Plates were loaded with 8 μl of PCR master mix containing SYBR Select Master Mix (Applied Biosystems, Thermo Fisher Scientific, Waltham, MA), nuclease free water (NFW), and three primers (Table 2). Due to differences in efficiency from the added GC tails, initial primer titering was necessary to ensure that the melt curves of homozygous controls would be accurately called by the analysis software. Individual adult females were homogenized for 60 seconds at max speed in 100 μl of NFW on a bead beater (BioSpec, Bartlesville, OK). Individual eggs, randomly sampled from all available egg papers of the same generation, were similarly treated but homogenized in 50 μl of NFW. Immediately after homogenization, samples were centrifuged at 10,000 relative centrifugal force (rcf) for 60 seconds to pellet solids. Two microliters of each supernatant were added to 8 μl of PCR mix for each primer set and then subjected to standard cycling conditions (3 min @ 95 °C; 40 cycles @ 95 °C for 3 sec, 60 °C for 15 sec). Melt curve analysis followed cycling with acquisition of fluorescence data every 0. 3 °C as the temperature was ramped from 60 °C to 95 °C. Characteristic melting temperature (Tm) peaks in the derivative fluorescence data indicate the presence of specific alleles [32,33]. For the 1016 mutation, a codon for the susceptible valine has a Tm peak at 86±0. 3 °C while the isoleucine codon has a Tm of 77. 3±0. 3 °C. The 1534 phenylalanine has a Tm of 79. 8±0. 3 °C while the mutant cysteine produces a peak at 84. 7±0. 3 °C. Homozygotes for either allele produce one peak while heterozygotes produce peaks at both Tms. All assay plates included two susceptible ORL1952 strain and two resistant PR strain samples as negative and positive controls, respectively. Most plates also contained artificial heterozygotes created by including ORL1952 and PR homogenates in the same sample. Well positions of individual mosquitoes were maintained in both plates (one plate for each locus) to allow genotyping of an individual for both SNPs. Frequencies for each of the nine genotypes (VVFF, VVFC, VVCC, VIFF, VIFC, VICC, IIFF, IIFC, IICC) were calculated by dividing the specific genotype by the total number tested from each area. Base maps were created using ArcGIS software by Esri. ArcGIS® is the intellectual property of ESRI and is used herein under license. Copyright © ESRI. GIS data sources were ESRI and Tele Atlas. All rights reserved. (For more information about Esri® software, please visit www. esri. com). Permission to publish this content was verified from the ESRI Redistribution Rights Matrix at https: //www. esri. com/~/media/Files/Pdfs/legal/pdfs/redist_rights_106. pdf? la=en Maps were exported to GIMP 2. 8 and additional layers with the kdr or resistance ratio data was added. Pie chart representations of kdr allele frequencies were created with Microsoft Excel and exported to layers added to the basemaps.
All Florida strains of Ae. aegypti were resistant to permethrin when compared to the ORL1952 strain, which has been in a continuous laboratory colony for nearly seventy years (Fig 1 and Table 3). The field strains showed varied levels of resistance, from 6-fold to 61-fold compared to the ORL1952 strain. The two least resistant Ae. aegypti strains were collected from flowerpots on Big Coppitt Key (Monroe County) and from oviposition cups in Cortez (Manatee County) at 6. 0 and 6. 8-fold, respectively. Ae. aegypti mosquitoes collected from a tire facility in Orlando (Orange County) were about 17. 2-fold more resistant than Ae. aegypti collected from the same area generations earlier and used to produce the ORL1952 strain. While most of the FL strains were 15 to 35-fold resistant compared to the lab strains, several strains with higher RR were identified. The strain from New Port Richey in Pasco County had the highest RR at 61. 3-fold, which is like the RR of the Puerto Rico-resistant reference strain [17]. Strains from Miami Beach and Fort Myers had RRs above 50. The variability we observed in RR throughout the state was also seen at finer resolution within several, but not all, counties. In Miami-Dade County, the Miami Beach and East Wynwood strains were relatively resistant (57-fold and 33-fold), while nearby locations like South Wynwood and central Little River were much less resistant (Fig 1). Manatee County also had a range of RRs. The Anna Maria Island strain, collected from a densely populated barrier island, had a higher RR than nearby strains from Palmetto or Cortez (Fig 1 and Table 3). We also observed this same trend in Lee County but did not observe universal variation in RR. Aedes albopictus competes with Ae. aegypti for oviposition sites, and in several locations Ae. albopictus eggs were collected in conjunction with Ae. aegypti eggs. Ae. albopictus strains were also subjected to topical application along with the Ae. aegypti from the same locations (Fig 1, in blue). In the Ae. albopictus strains we tested, only slight resistance to permethrin (<2-fold) was observed when compared to the ORL1952 strain. In Miami-Dade County, we observed large differences in RR between Ae. albopictus and Ae. aegypti even when collected from the same sites in areas such as south Wynwood and central Little River. St. Johns (St. Augustine), Orange (Orlando), and Lee (Alva) counties also had resistant Ae. aegypti and low RR Ae. albopictus. Examination of kdr alleles in Ae. aegypti strains from 62 locations showed a range of genotypes (Figs 2 and 3). We observed that most strains were fixed or nearly fixed for the F1534C SNP. For many strains, more than 95% of the tested mosquitoes were homozygous for the 1534C (1534CC) and the remainder of the strain was made up of a few 1534 heterozygotes (1534FC). Only strains from Big Coppitt Key (13% FC), Cortez in Manatee County (38% FC), Central Wynwood (22% FC) and 2 strains from Little River (20% and 14% FC) had more than 10% of mosquitoes that were heterozygous at position 1534. Most strains had no mosquitoes without at least one copy of the 1534C allele but we did find two strains from central Wynwood (8% FF) and Cortez (11% FF) that still had appreciable numbers of susceptible alleles. There was much more variation throughout the state at position 1016. Strains from Big Coppitt Key, Longboat Key, Orlando and Clearwater had the lowest percentages of the homozygous 1016II at 11. 0,14. 1,14. 9, and 17. 8%, respectively. We did observe strains with high levels of 1016II. Six of eight strains examined from Pasco County had 1016II frequencies above 75%, including two strains from New Port Richey at 100% 1016II. Due to the 2016 Zika outbreak and resulting massive public health response, we heavily sampled strains from southeast Florida for allele frequencies. Two strains from Broward and several from Miami-Dade County had 1016II frequencies above 75% including a Miami Beach strain at 91% 1016II (Fig 3). Select strains from Lee and Collier counties in southwest Florida were also above 75% 1016II (Fig 2). As with RR, in several counties we observed large differences in allele frequency from one part of the county to another. In Lee County (Fig 2), an inland strain from Alva had a relatively low level 1016II frequency (20%) while four strains from near the Gulf Coast all had 1016II frequencies of greater than 55%. This variation was also observed at much closer scale between strains in neighboring cities. The Manatee County Anna Maria Island strain had a much higher 1016II allele frequency than nearby strains from Palmetto or Longboat Key (Fig 2). While overall IICC frequencies were high in Pasco County, we did find variation at the neighborhood level in south Pasco County, where a strain with 75% 1016II was found just blocks from a strain with 30% 1016II (Fig 2). In Miami-Dade, we were able to test four strains from Wynwood and three from Little River collected during the 2016 Zika response (Fig 3). Again, we observed variations in the levels of 1016II within each neighborhood. North Wynwood was slightly more than 25% IICC while south Wynwood was nearly 50%. In Little River, the four strains ranged from 20 to 60%. Just across the water from Wynwood, the Miami Beach strain was greater than 90% IICC. We regularly observed only six genotypes in the field (Table 4). Only a combined seven mosquitoes of the 4,810 analyzed in this study had genotypes of IIFF, IIFC, or VIFF, which would require an allele coding for a resistant isoleucine at 1016 (1016I) and a susceptible phenylalanine at 1534 (1534F) to be contributed from at least one parent. We did commonly observe the reverse where homozygous or heterozygous susceptible 1016V alleles paired with the homozygous resistant 1534C allele (VVCC or VICC). Plotting the combined dataset for strains with both kdr genotype frequencies and resistance ratios from topical application of permethrin showed a strong correlation between increasing RR and increasing 1016II frequency (Fig 4, Pearson correlation coefficient = 0. 905, P<0. 00001, n = 20). Strains from Miami Beach, Fort Myers, and New Port Richey, with high frequencies of 1016II, which, based on the rarity of IIFF and IIFC mosquitoes implies the genotype IICC, were the strains with the highest resistance ratios. In contrast, the strain from Big Coppitt Key had the lowest 1016II frequency and the lowest RR even though the 1534CC frequency was relatively high.
Recent local transmission of Zika virus during 2016 and small outbreaks of dengue virus in 2010 and 2011 demonstrate that an effective IVM plan that attacks multiple life stages to reduce mosquito numbers is a necessity. However, when disease transmission is active, chemical adulticiding can be the only means to immediately reduce the population of potentially infected mosquitoes. While there is some debate as to overall efficacy of adulticiding against Ae. aegypti and Ae. albopictus, there is little question that it is a necessary part of the response that allows other slower methods of control like larviciding and source reduction to gain a foothold. Pyrethroids are the major class of chemical adulticides that Florida mosquito control programs use operationally. Thus, the efficacy of pyrethroids on Aedes vectors is a critically important part of a control program. Direct topical application of permethrin clearly demonstrated that resistance was widespread in Ae. aegypti strains throughout Florida. Resistance ratios ranged from about 6-fold in strains from Manatee County and Big Coppitt Key to approximately 60-fold in Miami Beach, Cape Coral, and New Port Richey. Genotyping thousands of individuals indicated that common kdr alleles were also widely distributed in the state. While these findings are noteworthy as they represent the first published report of widespread pyrethroid resistance and kdr alleles in the southeast US, this result is not surprising and could be predicted based on the results from other locations including Puerto Rico, Mexico and, more recently, in invasive CA Ae. aegypti strains [11,13,20]. The dataset described in this study does reveal wide variations in both RR and kdr alleles within small geographic areas. Examining Miami-Dade County, the strain collected from Miami Beach was highly resistant and had high levels of kdr. However, strains from inland Wynwood and Little River were much less resistant than the Miami Beach strain. Even strains from the opposite ends of neighborhoods differed. We saw this disparate pattern in south Miami-Dade, Manatee, and Lee counties. In Pasco County, we saw very different genotypes in strains geographically close to one another, separated only by a major highway. This wide variation in resistance alleles has been observed in Mexico and has been proposed to predominate over natural gene flow [34]. Considering this variability, along with the relatively short flight ranges and limited immigration in Ae. aegypti [35], the very different allele frequencies we observed in geographically close strains would support performing testing in numerous areas of a control district to get an accurate resistance picture. In contrast to the resistance in the Ae. aegypti strains, we observed very little permethrin resistance in Ae. albopictus statewide. This was true whether Ae. aegypti were present or absent from the same collections. Waits et al. [24] showed very low levels of resistance in St. Johns County strains collected from areas without Ae. aegypti. Miami-Dade, Orange, and Lee counties all had the same pattern of resistant Ae. aegypti and much less resistant Ae. albopictus. It has been proposed that the development of pyrethroid resistance is much more difficult in Ae. albopictus due to sequence differences in the voltage-gated sodium channel (NaV) that make kdr much less likely, although recent reports indicate it may be possible [36,37]. Our laboratory efforts to induce permethrin resistance in wildtype Florida Ae. albopictus by gentle pressuring have failed. At this time, pyrethroid resistance in Florida Ae. albopictus does not appear to be an issue that could lead to adulticide failure. An important observation made due to this work is the correlation between increasing RR and the frequency of the IICC genotype in Ae. aegypti, which has been anecdotally observed in other studies using the CDC bottle bioassay [6,11,13,38]. The linkage between kdr mutations and a strong correlation with resistance ratios has also been observed in other dipterans [39,40]. While more research must be done to validate the correlation, this dataset adds another 20 strains with both resistance and kdr data that support using kdr genotype as a surrogate to estimate pyrethroid resistance levels in mosquitoes [9,41]. The use of allele frequencies has several potential benefits. Genotype data are relatively easy to collect, the results are produced in hours, and dead mosquitoes collected during standard surveillance activities can be used to provide information on resistance. With limited budgets and personnel at the operational vector control district level, predictive estimation of resistance levels could produce useful operational data from activities already being done without requiring additional efforts to collect or produce mosquitoes for bioassay testing. Nearly a decade after Donnelly et al. [41] asked whether kdr genotypes are predictive in mosquitoes there is, to our knowledge, no published study that shows a pyrethroid-resistant strain of Ae. aegypti without also showing the presence of kdr alleles. We suggest that the major benefits to be gained from use of allele frequencies as an estimator of resistance would be improvements in coverage area (this study, 62 strains with allele frequencies vs. 26 by direct topical application) and better operational decision-making as vector control programs could access more timely information on area specific resistance levels. These challenges of getting wide coverage and providing this information on an operationally useful timeline have been present in recent response efforts to Zika virus. The efforts of CDC and vector control units in Puerto Rico and the efforts of the authors and others in Florida to develop wide-area pyrethroid resistance maps by relying strictly on bioassay data show that it is currently a slow, labor intensive process. In Florida at least, the epidemic had passed by the time more than a few strains had been tested by bioassay. Until there is reliable, published evidence to argue against the use of kdr frequencies as a predictor, it is at present the only rapid way to assess strains across a large area. This study shows that permethrin resistance is widely present in variable intensity in Ae. aegypti throughout Florida. This variability points to the need to include resistance testing as part of an IVM plan as well as examine resistance in more than one location. But how do we use this resistance information to improve vector control? Clearly, the strain of Ae. aegypti in Miami Beach is very different from nearby Wynwood and would likely call for different treatment strategies. A treatment with permethrin would likely have much less effect in Miami Beach in comparison to other less resistant locations. Our study also points to the value of a collaborative approach from motivated stakeholders to develop resistance information. The state of Florida began regular conferences to bring together vector control districts, researchers, and public health resources months before the first locally transmitted case of Zika was reported in 2016. Like the CDC response in Puerto Rico, development of resistance information was an early and ongoing part of this process. The dataset in this study represents the result of thousands of hours of effort from vector control districts, vector control contractors, the state of Florida, and the federal government to produce operationally useful resistance information to protect public health and improve the efficacy of control operations. However, even with these efforts, this study is very limited in scope. We examined only permethrin resistance and although the literature shows the patterns we observed would likely be applicable to other pyrethroids [13,17], work to define statewide patterns of resistance to synergized products or organophosphates still needs to be addressed. | Aedes aegypti (Yellow-fever mosquito) and Aedes albopictus (Asian Tiger mosquito) can vector a variety of arboviruses that cause diseases and are thus a public health concern. Pyrethroid insecticide resistance is common in Ae. aegypti in many locations worldwide and can adversely affect vector control operations. However, the resistance status of these vectors in Florida is largely unreported and recent local transmission of dengue and Zika viruses has made this information critical for effective control operations. In this study, we showed that permethrin resistance and two common SNPs of the voltage gated sodium channel (V1016I and F1534C) previously associated with pyrethroid resistance were widely present in Florida Ae. aegypti strains. We also observed a strong correlation between the dilocus knock down response (kdr) genotype and resistance ratio (RR) as determined by topical application, which suggests, as have others, that kdr frequency may be a useful indicator of resistance in Aedes aegypti. | Abstract
Introduction
Methods
Results
Discussion | united states
invertebrates
dengue virus
medicine and health sciences
pathology and laboratory medicine
pathogens
variant genotypes
geographical locations
microbiology
animals
genetic mapping
north america
viruses
rna viruses
infectious disease control
insect vectors
florida
public and occupational health
infectious diseases
aedes aegypti
medical microbiology
microbial pathogens
disease vectors
insects
arthropoda
people and places
mosquitoes
eukaryota
flaviviruses
heredity
viral pathogens
genetics
biology and life sciences
species interactions
organisms
zika virus | 2018 | Quantification of permethrin resistance and kdr alleles in Florida strains of Aedes aegypti (L.) and Aedes albopictus (Skuse) | 7,836 | 265 |
During the Ebola virus disease (EVD) epidemic in Liberia, contact tracing was implemented to rapidly detect new cases and prevent further transmission. We describe the scope and characteristics of contact tracing in Liberia and assess its performance during the 2014–2015 EVD epidemic. We performed a retrospective descriptive analysis of data collection forms for contact tracing conducted in six counties during June 2014–July 2015. EVD case counts from situation reports in the same counties were used to assess contact tracing coverage and sensitivity. Contacts who presented with symptoms and/or died, and monitoring was stopped, were classified as “potential cases”. Positive predictive value (PPV) was defined as the proportion of traced contacts who were identified as potential cases. Bivariate and multivariate logistic regression models were used to identify characteristics among potential cases. We analyzed 25,830 contact tracing records for contacts who had monitoring initiated or were last exposed between June 4,2014 and July 13,2015. Contact tracing was initiated for 26. 7% of total EVD cases and detected 3. 6% of all new cases during this period. Eighty-eight percent of contacts completed monitoring, and 334 contacts were identified as potential cases (PPV = 1. 4%). Potential cases were more likely to be detected early in the outbreak; hail from rural areas; report multiple exposures and symptoms; have household contact or direct bodily or fluid contact; and report nausea, fever, or weakness compared to contacts who completed monitoring. Contact tracing was a critical intervention in Liberia and represented one of the largest contact tracing efforts during an epidemic in history. While there were notable improvements in implementation over time, these data suggest there were limitations to its performance—particularly in urban districts and during peak transmission. Recommendations for improving performance include integrated surveillance, decentralized management of multidisciplinary teams, comprehensive protocols, and community-led strategies.
In March 2014, Liberia detected its first cases of Ebola virus disease (EVD) in Lofa, a northern county bordering Guinea and Sierra Leone [1]. The Liberian Ministry of Health (MOH) (formerly Ministry of Health and Social Welfare) established a national task force and initiated control efforts, including contact tracing [1,2]. As the epidemic grew, the task force developed into an Incident Management System, which oversaw contact tracing in all 15 counties with support from international partners including World Health Organization (WHO), U. S. Centers for Disease Control and Prevention (CDC), and Action Contre la Faim [3,4, 5]. Continuous, widespread transmission continued until February 2015 [6], and 42 days after the last confirmed case had two negative samples in March 2015 [7], Liberia was declared free of Ebola on May 9,2015—marking an end to the epidemic [8]. Contact tracing is comprised of three main steps: identifying, listing, and monitoring persons who have been exposed to infected individuals, with the goal of rapidly diagnosing and treating new cases and preventing further spread of infection. This approach has been used to control transmission of infectious diseases including smallpox, tuberculosis, HIV, and syphilis [9,10,11,12]. Although contact tracing has been used in prior outbreaks of hemorrhagic fever, these outbreaks were small in scale [13,14]. Contact tracing is most efficient for diseases with low incidence, limited transmissibility [15,16], tight networks, and an incubation period long enough to allow intervention. Conversely, the effectiveness and optimal levels of investment for contact tracing, particularly for emerging diseases and for acute epidemics, are subjects of ongoing research and debate [15,16,17,18,19]. The 2014 EVD epidemic was the largest EVD outbreak in history and the first known EVD outbreak in West Africa [20]. The scale of the control efforts including contact tracing was unprecedented, particularly within both rural and crowded urban settings, which burdened existing surveillance capabilities and required immense commitment and cooperation on the part of government and the affected communities themselves. Furthermore, the strategies and implementation of contact tracing in Liberia evolved—from establishing operations to scaling them up—in order to respond to the changing phases of the epidemic. These aspects warrant the need to further examine contact tracing within this unique context. Here, we describe the scope and characteristics of contact tracing in Liberia and explore its performance during the 2014–2015 EVD epidemic in order to inform future contact tracing strategies in large-scale epidemics.
We performed a retrospective descriptive analysis of data collection forms for contact tracing that was conducted for the EVD epidemic in six of Liberia’s 15 counties during June 2014–July 2015. The six counties consisted of both rural and urban areas and represented 72% of the population of Liberia [21]. Three of the counties (Lofa, Bong, and Nimba) are at the border with Cote d’Ivoire, Guinea, or Sierra Leone, while the other three (Montserrado, Margibi, and Sinoe) extend from central areas to the coast. Additionally, both formal and informal sources of information regarding contact tracing organizational structures and implementation within these counties were reviewed to help provide context for the data analysis. A contact was defined as a person who had direct or indirect exposure to any confirmed, probable, or suspect EVD case, or bodily fluids of a case, within the past 21 days [22,23]. This definition also included any persons who had been discharged from an Ebola Treatment Unit (ETU) as not a case, due to their potential exposure to the virus while in the ETU. National contact tracing guidelines and forms, which were initially adapted from existing WHO and CDC materials and finalized during the waning days of the epidemic, were used as the foundation for implementing the three steps of contact tracing: contact identification, listing, and monitoring. Once a case was detected, contact identification and listing were conducted by interviewing the case and/or family members to gather an initial list of potentially exposed persons. In most instances, this process was conducted by case investigation teams, which were distinct from contact tracing teams, and any of the following six types of exposure were added: (1) sleeping or eating in the same household; (2) direct physical contact with the body; (3) touching bodily fluids; (4) manipulating clothes or other objects; (5) through breastfeeding; and (6) attending a case’s funeral. Contact tracers, chosen from within the community, located the listed contacts and identified any additional contacts missed in the initial investigation. Contact tracers transferred the information collected by case investigation teams to paper forms, including the contact’s name, county, district, town, and exposure (s). The name, age, location, and unique case identifier of the case for which contacts were listed, i. e. the “source case”, were also recorded. During contact monitoring, contact tracers were expected to visit contacts twice daily (morning and afternoon) for 21 days post-exposure in order to identify and record whether the contact had EVD symptoms. This was determined initially through self-reports and physical observation, and eventually temperature readings were added for more objective monitoring. Contacts were monitored for nine symptoms: joint pain, fever (>38° Celsius), weakness, nausea, diarrhea, headache, throat pain, red eyes, and mucosal bleeding. Following the outbreak, paper contact tracing forms were requested from all County Health Teams. Forms were received from six counties and the data were entered into a Microsoft Access database. Data were analyzed using Microsoft Access and Epi Info. Each form was considered a unique contact record and unit of analysis, though it was possible for individual contacts to be monitored more than once if re-exposed. Source cases were identified using unique case identifiers, name, age, county, and district. To assess the coverage of contact tracing, or the percentage of cases for which contacts were monitored, we calculated the ratio of source cases in the database to the total number of suspected, probable, and confirmed EVD cases in MOH situation reports for the same counties using the closest approximate dates [24,25,26,27,28]. The mean number of contacts per source case was presented as contact-to-case ratios. We analyzed records by county and district. Urban or rural classifications were assigned based on districts; districts that hold the county headquarters or that have settlements with a population of 5,000 or more persons were classified as urban [21]. Two districts, one each in Nimba and Montserrado counties, were divided into urban and rural sub-districts. Each of the six exposure categories and nine symptoms were analyzed, and medians and interquartile ranges (IQR) were calculated. We divided the timeframe into four phases based on the observed epidemic trends of cases within Liberia [5], per epidemiologic week (EW): “Phase 1”: the initial increase of cases, from June to mid-August 2014 (EW 22–33); “Phase 2”: the peak, from mid-August to mid-November 2014 (EW 34–46); “Phase 3”: a decline in the epidemic, from mid-November 2014 through February 2015 (EW 47–9); and “Phase 4”: sporadic clusters, from March through July 2015 (EW 10–31). The first date of contact monitoring or last date of exposure was used to categorize records by phase. Medians and IQRs for timeliness, determined by the difference between the last date of exposure and first date of follow-up, were calculated and stratified by urban-rural and phases. Each record was assigned one of seven outcomes of monitoring, either designated on the form or imputed using supplemental information: (1) “completed” the monitoring period of 21 days post-exposure; (2) “dropped” if the source was determined to be not a case; (3) “lost to follow-up” if the contact could not be located after three consecutive days; (4) “potential cases” if the contact presented with symptoms and/or died and monitoring was stopped; (5) “restarted” if monitoring was reinitiated due to a new exposure; (6) “transferred” if the contact moved to another jurisdiction; or (7) “unknown” for all remaining contacts with no outcome information. Contacts who presented with symptoms could be referred for medical evaluation without meeting EVD case definitions; hence, we use the terminology “potential cases”. We calculated the positive predictive value (PPV) defined as the proportion of traced contacts—excluding those with dropped and unknown outcomes—who were potential cases. Sensitivity was defined as the ratio of potential cases identified during monitoring to the number of new cases in situation reports in the same counties [24,25,26,27,28]. This analysis assumes all potential cases were infected with EVD, and that all source cases and potential cases in the database were included in the total counts from situation reports. Therefore, to the extent that these assumptions are overstated, the calculations serve as upper limit estimates. PPVs were stratified by urban-rural and epidemic phases, whereas sensitivity and coverage were stratified by phases. We used odds ratios and 95% confidence intervals to examine exposure types, symptom types, phases, and urban-rural amongst potential cases compared with contacts who completed monitoring; for ordinal variables, the lowest category was used as a reference group. Chi-square tests with p-values <0. 05 were statistically significant. Two multivariate logistic regression models were used: (1) urban-rural, phase, and exposure type covariates, limited to records with ≥1 exposures, and (2) urban-rural, phase, and symptom type covariates, limited to records with ≥1 symptoms. Only statistically significant variables in bivariate analysis were included in the models. Nonparametric tests were used for continuous variables. This assessment is included under Johns Hopkins School of Public Health Institutional Review Board no. 6296 with DHP as principal investigator. A letter of agreement was signed with the Liberia MOH concerning the publication of contact tracing analyses. This assessment used retrospective data collected for public health surveillance purposes so informed consent was deemed unnecessary according to the U. S. Common Rule. We followed the Declaration of Helsinki, aiming to provide assurance that the rights, integrity, and confidentiality of participants were protected.
We analyzed 25,830 records for contacts who had monitoring initiated or were last exposed between June 4,2014 and July 13,2015 in the six counties. Of these, 25,651 contacts were listed for 2,465 source cases; an additional 179 contacts had no source case provided. The overall contact-to-case ratio was 10: 1 (median = 7, range 1–424). The contact-to-case ratio increased with each subsequent phase and was higher in urban than rural districts. There were 9,241 EVD cases in situation reports in the six counties. The upper limit estimate of coverage, or the maximum percentage of cases for which contacts were monitored, was 26. 7%, and was lowest during Phase 1. (Table 1) In the six counties providing data, 89. 0% of the records were identified in Montserrado County, 8. 6% in Margibi, 1. 6% in Bong, 0. 4% in Lofa, 0. 4% in Sinoe, and 0. 1% in Nimba (Fig 1) (Table 2). Records pertained to 22 of Liberia’s 136 districts (Table 2); data from the remaining districts was unavailable due to no contact tracing records or no reported EVD cases. In total, 21,500 (83. 2%) contacts were in seven urban districts/sub-districts, mainly in the Monrovia capital district in Montserrado, while 4,327 (16. 8%) contacts were in 17 rural districts/sub-districts. Potential cases were less likely to be from urban districts (Table 3). Temporal trends for contact tracing aligned with disease transmission trends (Fig 2). For 25,690 records grouped by phase, 61. 7% were monitored during Phase 2 and 32. 9% during Phase 3. Only contacts in Montserrado were monitored during Phase 4. No contacts were monitored during May–June 2015, corresponding to the Ebola-free period. Based on 25,300 records, contact tracing was timelier in rural districts; overall, the median difference was 1 day (IQR 0–4) (Table 1). Of 25,830 total contacts, 17,876 (69. 2%) contacts reported 34,284 exposure types, and 7,954 (30. 8%) had zero exposures recorded. Among the 17,876 contacts reporting any exposure, direct physical contact with the body was the most common (73. 1%), while funeral attendance (2. 1%) and breastfeeding (0. 2%) were the least common (Table 3). Two or more exposure types were reported in 54. 9% of 17,876 contacts; the median was 2 (IQR 1–3). Multivariate analysis showed the odds of sleeping or eating in the same household, direct physical contact, or touching bodily fluids were higher amongst potential cases than contacts completing monitoring (Table 3). Of 25,569 contacts with an assigned outcome, 22,680 (87. 8%) completed monitoring, 1,768 (6. 8%) were dropped, 637 (2. 5%) restarted, 334 (1. 3%) were potential cases, 136 (0. 5%) were lost to follow-up, and 14 (0. 1%) were transferred. Most contacts completed monitoring during each phase except during Phase 4, when 53. 6% of contacts were dropped (Fig 2). More contacts restarted during phases 2 and 3 than other phases. Potential cases were less likely to be monitored during phases 2 or 3 compared to Phase 1 (Table 3). Twenty-two contacts were not located prior to monitoring. Of 46 recorded contact deaths, 56. 5% were in urban districts and 33 occurred during monitoring (15 after taken to an ETU). The PPV was 1. 4% overall, and was higher in rural (3. 0%) than urban (1. 1%) districts and highest during Phase 1 (4. 7%), after which it decreased for subsequent phases. The sensitivity of monitoring, or the maximum proportion of new cases detected, was 3. 6%, and was highest during phases 1 and 2. (Table 1) Table 4 shows the distribution of reported symptoms. Overall, 326 contacts reported 1,299 symptom types and 3,732 symptom-days; the median symptom types per contact was 4 (IQR 2–5). Contacts of all outcomes reported symptoms except transferred contacts; 218 (66. 9%) of 326 contacts reporting symptoms were potential cases, 92 (28. 2%) completed monitoring, 6 (1. 8%) restarted, 6 (1. 8%) were unknown, 3 (0. 9%) were dropped, and 1 (0. 3%) was lost to follow-up. In multivariate analysis, potential cases were more likely to report fever, nausea, or weakness compared with contacts who completed monitoring.
During 2014–2015, more than 25,000 persons in six of Liberia’s 15 counties were identified, listed, and monitored for EVD, representing one of the largest contact tracing efforts during an epidemic in history. Nationwide, these efforts were even more substantial and required the dedication of responders, including the Government of Liberia, counties, and contact tracing teams. As a result, 334 contacts were identified as potential cases with the intention of providing earlier treatment and preventing hundreds of new infections. Relative to the scale of these efforts, however, these data suggest there were limitations to the performance of contact tracing within Liberia. Overall, there was a small proportion of monitored contacts that were identified as potential cases, and more than 97% of reported EVD cases from the six counties were not detected through contact monitoring. This is greater than expected, especially compared to other examples in West Africa where approximately 69% to 78% of cases were not being traced prior to case identification [29,30]. This measure is dependent upon the level of contact tracing coverage, and based on our database, though admittedly not comprehensive, coverage only accounted for a maximum of one-quarter of all EVD cases reported for these six counties. While this ratio is aligned with similar findings in two Guinea prefectures (32% and 39%) and in Sierra Leone (19%) [29,30], it is possible that contact tracing was not initiated for up to three-quarters of the remaining EVD cases in Liberia, potentially due to a combination of factors discussed below. Potential cases were more likely to be identified in rural districts and early in the epidemic, despite intensified efforts as the epidemic progressed. Possible explanations for why contact tracing was less effective in urban areas could include the following: higher population density and complex social networks making it more difficult to identify all contacts; less cooperation within urban settings; higher burden and strained resources; or a combination of these factors. These results support the concept that contact tracing is most successful when transmission is low, and models have shown that expanding implementation of contact tracing yields diminishing reductions in disease prevalence [15,16]. Therefore, it is critical to conduct contact tracing rigorously and comprehensively as soon as an outbreak is identified, and to achieve higher sensitivity and coverage during this phase. There were, however, notable improvements in implementation over time; specifically, greater coverage, fewer contacts lost to follow-up, and higher contact-to-case ratios. During Phase 4, Liberia was able to focus more resources on eliminating the last transmission chains, including expanding the inclusion criteria to ensure no new cases went undetected [6]. This would have resulted in a larger contact-to-case ratio during Phase 4 compared to all other phases. The dynamics of contact tracing are complex, and its success is related to characteristics of the disease and etiologic agent, resources, and socio-political factors that influence its acceptability and implementation. Additionally, the approaches to contact tracing may differ depending on whether there is a vaccine or therapy available. Given that contact tracing remains one of the critical public health tools during outbreaks involving person-person transmission, optimizing its performance is paramount. While not exhaustive, we focus on four key challenges that may have limited the performance of contact tracing for EVD within Liberia, and propose recommendations for future efforts. First, an integrated surveillance and data management system was lacking and had to be established for reporting between the national laboratory, healthcare facilities and ETUs, and contact tracing and case investigation field teams [5]. Consequently, contact tracing was less functional at the beginning of the epidemic when it could have been most effective in slowing the epidemic. Initially, an insufficiently integrated system resulted in missed source cases and contacts, and led to delays in monitoring; this is reflected in that 25% of contacts with available information started monitoring four days after their last exposure. Additionally, contacts were listed and needlessly traced because of delays in receiving negative laboratory results, thereby lowering the PPV. Although mobile applications had the potential to improve reporting and data management, these were not piloted until after the peak of the outbreak. In contrast, contact tracing in urban Nigeria successfully and rapidly contained EVD transmission, largely thanks to robust surveillance systems and leveraging mobile applications for real-time monitoring [31,32]. Strengthening integrated surveillance and electronic data systems, and the early adoption of mobile technology, could help improve timely reporting for listing and monitoring contacts. Secondly, the organizational structure for contact tracing likely led to inefficiencies in its implementation and management, particularly in urban districts. For instance, case investigation teams, who conducted contact listing, were often distinct from contact tracing teams who conducted contact monitoring. In some rural areas, teams responded in tandem thereby reducing gaps, yet this was more difficult in dense urban areas such as in Montserrado County. Additionally, the county level coordinated all aspects of the response—not just contact tracing. In January 2015, Montserrado created decentralized sub-county sectors to oversee and synchronize all operations—a change previously recognized as a critical step for halting transmission [6]. Particularly in urban areas and in the absence of a robust surveillance system, using a decentralized management approach and multidisciplinary teams may improve contact tracing performance. Thirdly, there were challenges with adapting and implementing contact tracing protocols, which had to be used by novice teams during the epidemic. For instance, the number of contacts per source case ranged widely in our analysis, from 1 to 424, and nearly one-third of contacts had no exposure documented, indicating that some contacts may not have met the inclusion criteria, thereby straining resources. Also, written guidance for identifying potential cases during monitoring did not specify how contact tracers should determine when to refer a contact for medical evaluation [23]. Eighteen contacts, who presumably would have shown symptoms prior to death, died during monitoring without being referred for medical evaluation. Among contacts who reported symptoms, including multiple symptoms and symptom-days, 33. 1% continued under monitoring without being referred for further evaluation, indicating that triggers for identifying potential cases was subjective. During future outbreaks, clear and comprehensive protocols need to be initiated early in the epidemic and reinforced throughout implementation. Furthermore, if resources are limited, inclusion criteria could prioritize contacts with multiple exposures, and/or those with household contact or direct contact with the body or bodily fluids. Triggers for identifying potential cases could include contacts reporting multiple symptoms types, fever, nausea, and weakness. Finally, community perceptions, stigma, and mistrust reportedly led to challenges in obtaining complete and reliable information, to delays or an inability to trace contacts due to evasion, and even to violence [5,33]. Underreporting of symptoms due to fear or due to fever-reducing drugs may explain why relatively few symptoms were captured in our database. Also, contacts were instructed to self-isolate within their home, which disrupted normal routines and the ability to maintain jobs; without adequate support from the community or organizations, contacts are less likely to cooperate. These aspects stress the importance of community cooperation, trust, and engagement. Overall, less than 1% of contacts were lost to follow-up, and this improved during each phase along with more contacts listed per source case, suggesting that this cooperation probably improved as the outbreak progressed. For future outbreaks, community-led strategies for contact tracing should be an early priority to foster cooperation, trust, and ownership of the control efforts. This analysis represented both urban and rural settings, and Montserrado specifically, where the response was most intense. However, we were unable to collect forms from all 15 counties nor all forms from the six inclusive counties; for example, no forms were available for the EVD cluster that occurred in Margibi in July 2015. Despite commendable efforts, counties reported that paper forms were lost or destroyed due to perceived contamination risks. Using paper forms also led to variability in data quality, including illegible writing, misspellings, inversed source case and contact information, and difficulty in interpreting marks for visits and the presence of symptoms. Falsifying information on forms was a concern [33], such as documenting visits when the contact had not been seen, and this was an issue early in the epidemic. These factors, combined with the lack of information to ascertain the final status of EVD infection amongst source cases and potential cases, constrained our analysis. Likewise, we could not conclude whether symptoms reported amongst potential cases evidenced EVD infection. Our data primarily represented contact monitoring, as we did not have a comprehensive contact listing. Finally, EVD case counts from situation reports were unavailable to stratify coverage and sensitivity by urban-rural districts and/or phase, and these aggregated totals could not be linked to our individual-level database. Our findings suggest that despite the unprecedented scale of contact tracing for EVD in Liberia, there were limitations in its ability to detect new cases, especially in urban areas and during the peak case load. Since contact tracing remains a critical intervention for controlling outbreaks, we suggest rigorous implementation early in the outbreak and focusing on four key areas to optimize its performance within similar contexts: (1) strengthening integrated surveillance and electronic data systems, (2) decentralizing management of multidisciplinary teams for improved coordination and oversight, (3) instituting and reinforcing clear and comprehensive protocols, and (4) adapting community-led strategies to foster cooperation, trust, and ownership. | Contact tracing is comprised of three main steps: identifying, listing, and monitoring persons who have been exposed to infected individuals, with the goal of rapidly diagnosing and treating new cases and preventing further spread of infection. This approach has been used to control transmission of infectious diseases including smallpox, tuberculosis, HIV, and syphilis, and while contact tracing has been used in prior outbreaks of hemorrhagic fever, these outbreaks were small in scale. During the 2014–2015 Ebola virus disease (EVD) epidemic in Liberia, contact tracing was implemented in all 15 counties on a scale that was unprecedented, particularly within both rural and crowded urban settings. This work provides insight into the magnitude that which contact tracing was implemented, its characteristics, as well as an assessment on its performance. Given that contract tracing is a critical tool for controlling disease spread, these findings aid in informing future planning and decision making for its implementation. | Abstract
Introduction
Methods
Results
Discussion | medicine and health sciences
body fluids
pathology and laboratory medicine
infectious disease epidemiology
geographical locations
signs and symptoms
infectious disease control
africa
infectious diseases
geography
epidemiology
people and places
nausea
diagnostic medicine
anatomy
urban areas
fevers
physiology
earth sciences
liberia
geographic areas
biology and life sciences | 2018 | Contact tracing performance during the Ebola epidemic in Liberia, 2014-2015 | 5,857 | 221 |
Culex species are widespread across Cameroon and responsible for high burden of nuisance in most urban settings. However, despite their high nuisance, they remain less studied compared to anophelines. The present study aimed to assess Culex species distribution, susceptibility to insecticide, bionomics and role in Lymphatic Filariasis (LF) transmission in the city of Yaoundé. Mosquito collections were conducted from March to December 2017 using Centre for Disease Control light traps (CDC-LT), human landing catches (HLC) and larval collections. Mosquitoes were identified using morphological identification keys. Mosquitoes from the Culex pipiens complex were further identified using Polymerase Chain Reaction (PCR) to assess the presence of sibling species. Bioassays were conducted with 2–5 day-old unfed females to assess mosquito susceptibility to DDT, permethrin, deltamethrin and bendiocarb following WHO guidelines. Dead, control and surviving mosquitoes from bioassays were screened by PCR to detect the presence of knockdown resistance (kdr) alleles. Pools of mosquitoes were examined by PCR to detect the presence of Wuchereria bancrofti. A total of 197,956 mosquitoes belonging to thirteen species were collected. The density of mosquito collected varied according to the collection methods, districts and seasons. Culex quinquefasciatus emerged as the most abundant and the only species of the Culex pipiens complex in Yaoundé. Culex species were found breeding in different types of breeding sites including polluted and unpolluted sites. All Culex species including Cx antennatus, Cx duttoni, Cx perfuscus and Cx tigripes were found to be highly resistant to permethrin, deltamethrin and DDT. Culex quinquefasciatus was also found to be resistant to bendiocarb. A high frequency of the West Africa kdr allele was recorded in resistant Cx. quinquefasciatus. Out of the 247 pooled samples of 25 Culex spp. examined for the presence of Wuchereria bancrofti, none was found infected. The study confirms the high adaptation of Culex species particularly Culex quinquefasciatus to the urban environment and no implication of this species in the transmission of LF in Yaoundé Cameroon. Culex species predominance in urban settings highlight potential transmission risk of West Nile and rift valley fever in Yaoundé.
Study site- The study was conducted in Yaoundé (03°52’N; 11°31’E), the capital city of Cameroon from March to December 2017. The city has a population estimated at 2. 8 million inhabitants. Yaoundé belongs to Guinean subequatorial climate type, characterized by four distinct seasons: the short rainy season (Mars-June), the short dry season (June-July), the long rainy season (August-November) and the long dry season (November-February). The city receives annually over 1600 mm of rainfall and the annual average temperature is 24°C. Yaoundé is located about 750 m above sea level and surrounded by many hills. Although occurring at very low endemicity, human infection by Wuchereria bancrofti was estimated at 2. 3% during surveys conducted between 2009–2010 in Yaoundé and it surroundings [19]. The study was conducted under the ethical clearance N° 2016/11/832/CE/CNERSH/SP delivered by the Cameroon National Ethics Committee for Research on Human Health (CNERSH) Ref N°D30-172/L/MINSANTE/SG/DROS/TMC of 4 April 2017. For human landing catches all adult men who took part in the collection signed a written informed consent form before being enrolled in the study as recommended by the validated protocol and were given free malaria prophylaxis. Mosquito’s collection and breeding sites characterization- Adult and immature stages of Culicine mosquitoes were sampled in 32 districts of Yaoundé. Culicine collections were undertaken in the context of a big survey intended to assess mosquito distribution and malaria transmission pattern in the city of Yaoundé before a larval control trial and will allow in the future additional analysis with more data. Adult mosquitoes were collected using CDC light traps (CDC-LTs) and Human Landing Catches (HLCs) from 7pm to 6am. All potential larval breeding sites were inspected and positive sites (with at least one Culicine larvae or pupae) recorded. Three dips were undertaken for small breeding sites of less than 1 m2; and 5 to 10 dips were undertaken in breeding sites of more than 1m2. The average larval density (N) was estimated. Once collected larvae were classified according to their stages: early instars larvae (L1&L2) and late instars (L3&L4 and pupa). Other parameters measured included the type of breeding sites sampled (stagnant water pools, gutters, well, tyre print, footprint, pit latrine….), depth, the status organically polluted or not, the distance to the nearest house, the presence/absence of predators, the proportion of water surface covered by vegetation or algae. Larvae collected were kept in plastic containers and brought to the insectary for rearing. After emergence, adult mosquitoes were identified to species level under a binocular magnifying glass using morphological identification keys [31–33]. For mosquitoes collected using either CDC-LTs or HLC, a subsample of 50 culicine specimens per district was randomly selected for identification during each collection month. All mosquitoes collected were stored at -20°C for further molecular analyses. Susceptibility tests to insecticides-Bioassays were performed with 2–5 days old females emerging from larval collection. Mosquitoes were tested against permethrin 0. 75%, DDT 4%, bendiocarb 1% and deltamethrin 0. 05% following WHO guidelines [34]. For each test, batches of 25 mosquitoes per tube were exposed to impregnated papers for 1 hour. The number of mosquitoes knocked down by the insecticide was recorded every 10 minutes during exposure. After exposure, mosquitoes were fed with a 10% glucose solution and the number of dead mosquitoes was recorded 24 hours post-exposure. Mosquitoes used as controls were exposed to untreated papers. The mortality rates were corrected using the Abbot formula [35] whenever the mortality rate of the controls was between 5 and 20%. Susceptibility and resistance levels were assessed according to WHO criteria [34]. At the end of the assay, mosquitoes were classified into three different groups: 98%-100% mortality indicates susceptibility, 80%-97% mortality suggests possible resistance that needs to be confirmed, <80% mortality suggests resistance. The study objective was to assess culicine species distribution, bionomic and potential role in W. bancrofti transmission in the city of Yaoundé. High Culicine species diversity was recorded with up to 13 species collected. Culex species were the most prevalent and this was consistent with previous studies conducted in Cameroon and across Africa indicating the high adaptation capacity of species of this genus particularly Cx. quinquefasciatus to the urban environment [27,41–44]. The diversity of culicine species recorded could result from the presence of different landscapes across the city of Yaoundé made up of an alternation of highland and marshland covered with vegetation and exploited for agriculture, lakes invaded by vegetation, and rural environment. It is still unknown whether there is an intense competition between culicine species sharing similar habitats. Species such as Cx. tigripes larvae are known to be predators for early instars of different species. Culex quinquefasciatus emerged after molecular analysis, as the sole member of the Cx. pipiens complex in Yaoundé; its presence was consistent with the known distribution of members of the complex [37]. Species diversity and abundance were all found to vary according to collection methods and seasons. High species diversity was recorded using CDC-LT compared to HLC or larval collection and reflects the high efficiency of CDC-LT method for collecting culicines. The use of CDC-LT has now become common for sampling mosquito populations across the world and has been shown to be particularly effective for sampling Culex mosquitoes [27,45]. This tool was rather found to underestimate anophelines densities [27,45,46]. Both HLC and CDC-LT techniques were used because there was so far no available data on the efficiency of CDC-LT for collecting Culex species from Yaoundé. Seasonal variations in species composition was detected for mosquitoes collected from breeding habitats, however, no similar trend was detected for mosquitoes collected using CDC-LT or HLC. This likely suggest different breeding habitats preference for culicine species at different periods of the year or the influence of physico-chemical parameters [47,48] or xenobiotics selection [49] on Culex species distribution. Cx. quinquefasciatus larvae were found to be highly prevalent in polluted sites. It is likely that females of Culex species are more attracted by oviposition cues released by the microbial fauna in this type of habitats. In addition, these habitats are rich in nutrients and could thus reduce competition for resources between species. This could also be because mosquitoes in polluted sites are also frequently exposed to intensive selective pressure induced by pollutants and xenobiotics [27,50–52], different strategies were reported to promote Culex species adaptation to different ecological constraints. This include the development of resistance or detoxification mechanisms to a large set of insecticides and xenobiotics [53–55], the capacity for eggs to resist desiccation [56] and development of cuticle resistance in larvae [3,57,58]. Several Culex species including Cx. quinquefasciatus, Cx. antennatus, Cx. duttoni were found to display resistance to DDT, permethrin and deltamethrin. This is the first time that insecticide resistance in different Culex species is documented in Cameroon. The level of pyrethroid resistance was similar to data recorded for An. gambiae populations in the city of Yaoundé [59,60]. In addition to the fact that Culex species are known to breed in polluted environment and could thus be affected by xenobiotics selection, the high level of resistance recorded could also result from increased use of LLINs for malaria vector control and pesticides use in agriculture in the city of Yaoundé [27,61]. Our study also suggested the presence of kdr allele in Cx. quinquefasciatus populations. It is likely that resistance in Culex species is sustained by both kdr mutations and other mechanisms such as the metabolic detoxification machinery [62]. The present study also permitted to evaluate the role of Culex species in LF transmission after mass drug administration (MDA) scale up in Cameroon. Culex quinquefasciatus is the predominant vector of LF in both urban and rural settings in East Africa [3,23] but less so in Central and West Africa. However, with potential gene flow and changing climate, one cannot rule out that Cx. quinquefasciatius in Central Africa such as in Cameroon may also emerge as LF vector. Furthermore, because of the rapid expansion and predominance of this species in Cameroon cities, it’s potential implication in LF transmission in Yaoundé was examined. Analysis conducted with pool samples of Culex mosquitoes recorded no infection. In Cameroon LF is considered to be endemic with prevalence rates varying from 1 to 8% [19,63]. It is likely that the prevalence of parasite may have decreased over years due to the implementation of mass drug administration of ivermectin and abendazole to the population since 2009 [19]. So far, five to six rounds of MDA have been successfully conducted in endemic settings across the country and interruptions of LF transmission have been documented in some parts of the country [64]. The fact that only Culex species were screened during this study could have limited the capacity of detecting any ongoing transmission since mosquito species such as An. gambiae and An. funestus are also good vectors of LF [3,23]. Another important dimension which could explain the absence of W. bancrofti infection in Culex is that the area may have not been endemic for W. bancrofti before the introduction of MDA. Recent studies conducted in Cameroon and DRC suggested that the perceived endemicity of LF established by ICT test in the central African region could result from the presence of Loa filariasis which cross react to the ICT tests which was used to detect W. bancrofti in Central Africa, leading to false positivity [64–66]. During the last decade, several arboviral diseases such as chikungunya, dengue, yellow fever, West Nile, Sindbis, Tahyna, O’nyong-nyong and spondweni virus have been reported in circulation in human adults in both urban and rural settings [67–70]. With the rapid distribution of Culex species in the urban environment, the potential role that these species could play in spreading of these arboviral diseases deserves further consideration. The present study confirms high abundance of Cx. quinquefasciatus in the city of Yaoundé and high insecticide resistance in most Culex species populations. The study also suggests no transmission of W. bancrofti by Culex species in Yaoundé. In Cameroon, apart from malaria vectors, surveillance activities are not regularly conducted on other vectors of diseases because of lack of funding or technical capacities for these activities. In this context, combining surveillance activities of malaria vectors with other culicine species and strengthening capacities of medical entomologists on taxonomy, sampling, processing and calculation of key entomological indicators for endemic vector borne diseases could be cost effective and will enable better understanding of the distribution and epidemiology of various diseases. This could lead to the establishment of sustainable surveillance systems. | Culex species are highly prevalent in both urban and rural settings in Cameroon and are responsible for high nuisance and transmission of pathogens such as Wuchereria bancrofti and arbovirus. Despite the important epidemiological role, that Culex could play, they are still less studied. The current study was conducted to assess Culex species distribution, susceptibility to insecticide and role in W. bancrofti transmission in the city of Yaoundé. Mosquito collection was conducted using three collection methods human landing catches, CDC light traps and larval collection. Once collected, mosquitoes were identified using morphological identification keys and PCR diagnostic tools. They were later processed to determine their infection status. Bioassays with Culex females of 2 to 5 days old were conducted to determine their susceptibility level to different insecticide families. Culex quinquefasciatus emerged as the most abundant species. Up to 13 different culicine species were recorded. Culex species were recorded to be highly resistant to DDT, permethrin and deltamethrin. A high frequency of the West Africa kdr allele was recorded. No mosquito was detected to be infected by LF. The study confirms the need for further xenomonitoring activities in order to control the risk of outbreaks due to Culex mosquitoes in the city of Yaoundé. | Abstract
Introduction | death rates
invertebrates
medicine and health sciences
geographical locations
animals
developmental biology
molecular biology techniques
population biology
insect vectors
africa
research and analysis methods
infectious diseases
cameroon
agrochemicals
wuchereria bancrofti
artificial gene amplification and extension
wuchereria
culex quinquefasciatus
life cycles
molecular biology
disease vectors
insects
agriculture
arthropoda
people and places
insecticides
population metrics
mosquitoes
eukaryota
polymerase chain reaction
nematoda
biology and life sciences
species interactions
larvae
organisms | 2019 | Culex species diversity, susceptibility to insecticides and role as potential vector of Lymphatic filariasis in the city of Yaoundé, Cameroon | 3,332 | 316 |
A constellation of metabolic disorders, including obesity, dysregulated lipids, and elevations in blood glucose levels, has been associated with cardiovascular disease and diabetes. Analysis of data from recently published genome-wide association studies (GWAS) demonstrated that reduced-function polymorphisms in the organic cation transporter, OCT1 (SLC22A1), are significantly associated with higher total cholesterol, low-density lipoprotein (LDL) cholesterol, and triglyceride (TG) levels and an increased risk for type 2 diabetes mellitus, yet the mechanism linking OCT1 to these metabolic traits remains puzzling. Here, we show that OCT1, widely characterized as a drug transporter, plays a key role in modulating hepatic glucose and lipid metabolism, potentially by mediating thiamine (vitamin B1) uptake and hence its levels in the liver. Deletion of Oct1 in mice resulted in reduced activity of thiamine-dependent enzymes, including pyruvate dehydrogenase (PDH), which disrupted the hepatic glucose–fatty acid cycle and shifted the source of energy production from glucose to fatty acids, leading to a reduction in glucose utilization, increased gluconeogenesis, and altered lipid metabolism. In turn, these effects resulted in increased total body adiposity and systemic levels of glucose and lipids. Importantly, wild-type mice on thiamine deficient diets (TDs) exhibited impaired glucose metabolism that phenocopied Oct1 deficient mice. Collectively, our study reveals a critical role of hepatic thiamine deficiency through OCT1 deficiency in promoting the metabolic inflexibility that leads to the pathogenesis of cardiometabolic disease.
Hepatic energy metabolism is a major determinant of systemic glucose and lipid levels as well as total body adiposity, which in turn are key risk factors for cardiovascular and metabolic diseases [1,2]. Genome-wide association studies (GWAS) have provided a wealth of information on the genes and pathways involved in hepatic energy metabolism, including apolipoprotein E (APOE), proprotein convertase subtilisin/kexin type 9 (PCSK9), and low-density lipoprotein receptor (LDLR) [3–5]. In follow-up studies in cells and in preclinical animal models, most of these genes have been linked mechanistically to lipid metabolism [6]. In contrast, the mechanisms responsible for the genome-wide–level significant association of SLC22A1 (encoding the organic cation transporter, OCT1) with total and low-density lipoprotein (LDL) cholesterol [3] remains unexplored. In humans, the OCT1 gene is highly polymorphic. A number of reduced-function variants with high prevalence in European populations have been characterized [7–9]. In particular, 40% of Caucasians carry one and 9% carry two reduced-function OCT1 variants [7,8]. OCT1, which is highly expressed in the liver, has been widely characterized as a drug uptake transporter. Reduced-function polymorphisms of OCT1 have been associated with changes in the pharmacokinetics and pharmacodynamics of several drugs, including the opiate receptor agonist, morphine, and the anti-diabetic drug, metformin [10–12]. Recently, GWAS and fine mapping analysis showed that OCT1 functional variants are associated with acylcarnitine levels through efflux mechanism [13]. Previously, through metabolomic studies in Oct1-/- mice and in cells overexpressing human OCT1, our laboratory identified thiamine, vitamin B1, as a major endogenous substrate for OCT1, and Oct1 knockout mice were shown to exhibit hepatic thiamine deficiency [14]. Although systemic thiamine deficiency is well known to cause nerve damage and lead to beriberi and Wernicke-Korsakoff syndrome [15,16], the pathophysiologic effects of thiamine deficiency in the liver are not understood. Thiamine pyrophosphate (TPP), the active metabolite of thiamine, is an essential cofactor for several metabolic enzymes, including pyruvate dehydrogenase (PDH), α-ketoglutarate dehydrogenase (α-KGDH), and transketolase (TK), which have fundamental roles in regulating cellular energy metabolism [15]. In particular, in 1963 Randle proposed that PDH acts as a key metabolic switch in the glucose–fatty acid cycle, which underlies the metabolic disturbance of diabetes. Under the theory of substrate competition between glucose and fatty acids, an increase in fatty acid oxidation and a reduction in glycolytic flux result in a critical imbalance in energy metabolism in tissues. As noted by Randle, regulation of PDH activity greatly influences selection of fuel source [17,18]. Failure to flexibly adjust the choice of fuel (e. g. , fatty acids or glucose) for metabolic energy production has recently been proposed to underlie metabolic inflexibility and lead to the pathogenesis associated with metabolic disorders [19]. Metabolic inflexibility and indeed metabolic syndrome have been linked to an excess of macronutrients (e. g. , carbohydrates or fat); however, the role of micronutrients such as thiamine in metabolic syndrome has been largely ignored. Although many reports have identified a high prevalence of thiamine deficiency in patients with diabetes or obesity [20–23] and a beneficial effect of thiamine supplementation in these patient populations [24–26], the molecular mechanisms contributing to thiamine-associated metabolic disturbance are unknown. Here, we hypothesize that reduced OCT1 function or reduced dietary thiamine intake leading to decreases in hepatic thiamine levels modulates the activity of multiple enzymes and the levels of key metabolites involved in glucose and lipid metabolism. These effects result in dyslipidemias, increases in circulating glucose levels, and peripheral adiposity. Through extensive experiments in Oct1-/- mice, our data show that Oct1 deficiency results in substantial changes in hepatic energy metabolism, i. e. , reduction in glucose utilization, increased gluconeogenesis, and alterations in lipid metabolism. Similarly, feeding wild-type mice a thiamine deficient diet (TD) results in comparable effects on hepatic energy metabolism. Taken together, our studies suggest that hepatic thiamine deficiency, through deletion of Oct1 in mice, results in the development of metabolic inflexibility. Our studies provide a mechanistic explanation for the striking metabolic findings in large-scale human genetic studies, demonstrating that common OCT1 reduced-function polymorphisms are associated with dyslipidemias, obesity, and increased risk for type 2 diabetes.
The GWAS Catalog, database of Genotypes and Phenotypes (dbGAP) Association Results Browser, and Genome-Wide Repository of Associations Between SNPs and Phenotypes (GRASP) identified two major phenotypes (total cholesterol and LDL cholesterol levels) that were associated with genetic variants in SLC22A1 (OCT1) (Fig 1 and S1A and S1B Fig). In particular, rs1564348 and rs11753995 were associated with LDL cholesterol (p = 2. 8 × 10−21) and total cholesterol (p = 1. 8 × 10−23), respectively (Fig 1 and Table 1). Using HaploReg v4. 1 to obtain linkage disequilibrium information from 1000 Genomes Project, we noted that these two SNPs are in linkage with the OCT1 with methinone420 deletion (420Del), a common genetic variant in OCT1 that shows reduced uptake and altered kinetics of its substrates. Thus, the results suggest that reduced OCT1 function is significantly associated with higher total cholesterol and higher LDL levels. The GRASP database identified other phenotypes with significant, but weaker, p-values, relevant to glucose traits and coronary artery disease. Recent results from the UK Biobank cohort (http: //geneatlas. roslin. ed. ac. uk/), available in the Gene ATLAS database and from the Global Lipids Genetic Consortium, are also included in Table 1. As shown, the reduced-function OCT1 nonsynonymous variants, OCT1-R61C, OCT1-G401S, OCT1-420Del, and OCT1-G465R, were significantly associated with high total cholesterol, LDL cholesterol, and/or TG levels in at least one study (Table 1). In addition, two of the missense OCT1 variants, OCT1-P341L and OCT1-V408M, which are associated with lower SLC22A1 expression levels in several tissues [13,27,28], were also associated with higher cholesterol levels in at least one study. The OCT1 nonsynonymous variants in Table 1, except OCT1-P341L, are not in linkage disequilibrium (r2 < 0. 1) with SNPs in lipoprotein (a) (LPA) and lipoprotein (a) like 2 (LPAL2) genes (a known locus for plasma lipoprotein levels) [29–31] (S1C Fig), indicating that OCT1 constitutes an independent locus for association with plasma lipids, which was also recently shown in other studies [32,33]. Notably, the effect size of the OCT1 variants for associations with lipids traits are small; thus, larger sample sizes are needed for genome-wide level significance (p < 5 × 10−8) (Table 1). In the Type 2 Diabetes Knowledge Portal, weaker but significant associations (p < 0. 05) between OCT1 reduced-function variants and higher 2-hour glucose levels, higher fasting insulin levels, increased risk for type 2 diabetes, increased risk for coronary artery disease, and higher BMI were cataloged (Table 1). We performed burden test analysis using the data available in the portal. Interestingly, in the analysis, in which we included possibly or probably deleterious missense or protein truncating variants of OCT1, we observed strong associations of the reduced-function OCT1 variants with increased body weight (p = 0. 0002–0. 0005, beta = 0. 23–0. 3). When we performed a similar burden test analysis with type 2 diabetes, the significance was weaker and the results were only significant when we included only protein truncating variants of OCT1 (p = 0. 015, odds ratio = 2. 10). Consistent with our previous studies, deletion of Oct1 protected the mice from hepatic steatosis [14] (Fig 2A, S2A Fig). In this study, we observed that glycogen content was 3. 3-fold greater in livers from Oct1-/- mice compared to livers from Oct1+/+ mice after an overnight fast (Fig 2A and S2B Fig). Consistent with these results, hepatic glucose levels were 5. 9-fold higher (p = 0. 0006) in Oct1-/- mice compared to Oct1+/+ mice (S2C Fig). Significantly greater body weights were observed for Oct1-/- mice compared to their wild-type counterparts, starting at the age of 6 weeks (Fig 2B and S2D Fig). Body composition also differed, with dual-energy X-ray absorptiometry (DEXA) scans showing a higher percent of body fat in Oct1-/- compared to Oct1+/+ mice (p = 0. 001) (Fig 2C). Consistent with the greater proportion of body fat, Oct1-/- mice had greater epididymal fat pad weights and reduced liver weight compared to Oct1+/+ mice (p < 0. 0001) (Fig 2D and S2E Fig). To further assess the potential mechanism leading to increased weight gain in Oct1-/- mice, we analyzed energy expenditure, food intake, and activity by the comprehensive laboratory animal monitoring system (CLAMS). Before placing the mice into the CLAMS, the body composition of all mice was measured by EchoMRI. As shown in Fig 2E, Oct1-/- mice had greater fat and lower lean mass in comparison to Oct1+/+ mice (p < 0. 0001). When normalized to total body weight, Oct1-/- mice had significantly lower respiratory oxygen (O2) consumption and energy expenditure (Fig 2F and 2G), indicating lower metabolic rates of Oct1-/- mice in comparison to Oct1+/+ mice. These data are consistent with the lower lean mass of the Oct1-/- mice compared to Oct1+/+ mice, because lean mass contributes more to energy expenditure than more inert tissue, such as adipose tissue [39,40]. In fact, no differences in respiratory O2 consumption or energy expenditure normalized to lean mass were observed between Oct1+/+ and Oct1-/- mice (S2F and S2G Fig). Thus, the differences in metabolic rate between Oct1+/+ and Oct1-/- mice appear to be due to significant differences in body composition. Additionally, our Oct1-/- mice had no difference in activity but had slightly lower food intake and respiratory exchange ratio (RER) during the dark cycle compared to Oct1+/+ mice (S2H–S2J Fig). There were no deleterious effects of Oct1 deficiency on hepatic function and, in fact, some of the liver function tests improved in the Oct1 knockout mice in comparison to wild-type mice (S2K Fig). There were no major differences in the expression levels of thiamine transporters (Slc19a2 and Slc19a3) in the liver. In contrast, levels of organic cation transporter, Oct2, which also transports thiamine, were increased, albeit the expression levels of Oct2 in the liver were extremely low relative to Oct1 and Slc19a2 (S2L Fig). Collectively, our data suggest that Oct1 deletion had a significant effect on hepatic and peripheral energy homeostasis. We hypothesized that the systemic plasma levels of thiamine are higher in Oct1-/- mice as a result of reduced hepatic extraction of dietary thiamine (Fig 3A). As expected, Oct1-/- mice had significantly higher plasma levels of thiamine (Fig 3B and S3A Fig) compared to Oct1+/+ mice on thiamine-controlled and thiamine-enriched diets. In addition, Oct1 deletion preserved plasma thiamine levels in mice on TDs (Fig 3B). Thiamine deficiency is associated with life-threatening diseases, such as beriberi and Wernicke-Korsakoff syndrome [15,41]. We hypothesized that preserved circulating thiamine levels would delay the development of severe thiamine deficiency syndromes and increase the rate of survival when mice were challenged with a TD. As shown in Fig 3C, there was a significant improvement in the overall survival of Oct1-/- mice (p = 0. 012, Gehan-Breslow-Wilcoxon test; p = 0. 018, log-rank test) compared to Oct1+/+ mice. Modulation of Oct1 expression levels provides a means of studying the effect of hepatic thiamine levels per se as opposed to systemic thiamine levels or thiamine levels in other tissues. Manipulation of dietary thiamine may have additional effects, for example, in the central nervous system. Notably, Liu and colleagues determined that reduced levels of thiamine in the systemic circulation in mice resulted in neurological effects in the hypothalamus, with anorexia and resultant reduction in peripheral adiposity [42]. In human populations, the OCT1 gene is highly polymorphic [7–9,43]. Many loss-of-function polymorphisms of OCT1 have been characterized and found to affect hepatic uptake of drugs, leading to altered treatment response [43]. Here, in the uptake studies, cells expressing human OCT1 genetic variants (420Del or 420Del+G465R) had significantly reduced uptake of thiamine compared to the reference allele (Fig 3D), although they have comparable levels of OCT1 transcript (S3C Fig). In kinetic studies performed at 4 minutes, the maximum velocity (Vmax) of thiamine in cells expressing human OCT1 with methinone420 deletion (hOCT1-420Del) was 70% lower than in cells expressing the human OCT1 reference (hOCT1-Ref) (1. 80 ± 0. 09 nmol/mg protein/minute versus 5. 36 ± 0. 30 nmol/mg protein/minute) (Fig 3D). In contrast to humans, who express OCT1 primarily in the liver, mice express Oct1 in both the liver and the kidney; therefore, deletion of Oct1 in the kidney could potentially affect systemic levels of thiamine in mice. To address this limitation of the Oct1 knockout mice as a model for humans, we used hydrodynamic tail vein injection of mouse Oct1 short hairpin RNA (shRNA) lentiviral particle (or empty vector shRNA lentiviral particle as control) to specifically knock down Oct1 in the liver in both Oct1+/+ and Oct1-/- mice. Following a single intraperitoneal injection of 2 mg/kg thiamine (with 4% 3H-thiamine), we observed that the area under the plasma concentration-time curve (AUC) of thiamine was significantly greater in wild-type mice treated with Oct1 shRNA lentiviral particles compared to wild-type mice treated with vector control shRNA lentiviral particles (Fig 3E). Although not significant, similar trends were observed in the maximum concentration (Cmax) values (S3D Fig). Notably, the Oct1 shRNA did not affect Oct1 expression levels in the kidney (S3D Fig). Compared to wild-type mice with Oct1 shRNA lentiviral particle knockdown, higher systemic levels of thiamine were observed in Oct1-/- mice (Fig 3E), potentially reflecting an incomplete Oct1 knockdown (50% liver Oct1 expression reduction, S3D Fig) or an additive effect of renal Oct1 deletion in Oct1-/- mice. The data provide strong evidence that reduction of OCT1 expression in the liver alone can result in increased systemic thiamine exposure. Although the liver plays a role in pre-systemic thiamine metabolism, it should be noted that thiamine is metabolized in most tissues in the body; therefore, other tissues, such as the intestine, may contribute to pre-systemic metabolism of the vitamin. Collectively, alterations in OCT1 function through genetic polymorphisms affect thiamine uptake and disposition. Our previous studies indicated that Oct1 deletion resulted in reduced hepatic thiamine levels and levels of TPP [14], the cofactor of PDH. It is shown that reduced TPP levels directly affect the activity of PDH [44,45]. As PDH plays a key role in energy metabolism linking glycolysis to the tricarboxylic acid (TCA) cycle and fatty acid metabolism [17], we hypothesized that the activity of hepatic PDH was impaired in Oct1-/- mice. Because phosphorylation of PDH results in inactive forms of the enzyme [46], we measured levels of phosphorylated PDH (at two phosphorylation sites, Ser232 and Ser300) and mRNA levels of pyruvate dehydrogenase kinase 4 (PDK4). Both phosphorylated PDHs and PDK4 transcripts were significantly higher in livers from Oct1-/- mice (Fig 4A and S4A Fig). In addition, in Oct1-/- mice, glycogen synthase (GS), and glucose transporter 2 (Glut2) were present at significantly higher levels (Fig 4A). Although glycogen phosphorylase (PYGL), which plays a key role in breakdown of hepatic glycogen, was also expressed at higher levels, the ratio of GS to PYGL was significantly higher in livers from Oct1-/- mice (Fig 4A and S4B and S4C Fig). These data suggest that Oct1-/- mice had higher rates of glycogen synthesis, which could explain the higher hepatic glycogen content in Oct1-/- mice. Our data suggested that livers from Oct1-/- mice would have less activity of PDH, which in turn would result in a lower rate of conversion of pyruvate to acetyl-CoA entering the TCA cycle [47,48] and thus an overall reduction in oxidative phosphorylation of glucose. We hypothesized that the reduction of oxidative phosphorylation of glucose would increase the accumulation of the intermediates of gluconeogenic substrates. These intermediates would lead to increased gluconeogenesis as glycolysis and gluconeogenesis are reciprocally regulated and highly depend on the availability of gluconeogenic substrates [1,49]. The levels of glucose-6-phosphate (G6P), a strong allosteric activator of GS [50], were 2. 3-fold (p < 0. 0001) higher in the livers of Oct1-/- mice (Fig 4B). In addition, the ratio of phosphorylated GS to total GS was significantly lower in Oct1-/- mice (Fig 4C), consistent with a higher activity of GS in Oct1-/- mice. To further investigate the role of OCT1 in hepatic glucose metabolism, we performed three standard tests related to glucose homeostasis [51]. In the glucose tolerance test (GTT), the blood glucose rose following oral glucose dosing and fell back to normal in both Oct1+/+ and Oct1-/- mice (S4D Fig), although the Oct1-/- mice had higher blood glucose levels at baseline. After adjusting for baseline, there was a trend toward higher blood glucose levels and an overall greater glucose AUC after a bolus dose of glucose in Oct1-/- mice (Fig 4D). The GTT indicated that both Oct1+/+ and Oct1-/- mice could produce insulin in response to rising glucose. In contrast, pyruvate tolerance tests (PTTs) were different between Oct1+/+ and Oct1-/- mice (Fig 4E and S4E Fig). In particular, blood glucose was significantly higher at each time point after pyruvate injection in Oct1-/- mice, which suggested that Oct1-/- mice had higher rates of hepatic gluconeogenesis. In the insulin tolerance test (ITT), there was a trend toward higher blood glucose levels after insulin injection in the Oct1-/- mice and an overall greater glucose AUC (Fig 4F). Blood glucose levels are maintained by glucose uptake mainly in peripheral tissues and glucose output primarily from the liver [52]. Data from the PTT suggested that the knockout mice had significantly higher hepatic gluconeogenesis, which may have contributed to the higher glucose exposure in Oct1-/- mice following the ITT. To understand the role of thiamine in regulating glucose metabolism, age-matched mice were placed on dietary chow containing three different doses of added thiamine, following the experimental design shown in Fig 5A. Wild-type mice fed a TD for 10 days had higher levels of hepatic glycogen, hepatic glucose, and plasma glucose compared to mice fed control diets (Fig 5B–5D). In contrast, varying thiamine content in the diet resulted in no significant differences in hepatic glycogen, hepatic glucose, or plasma glucose levels among Oct1-/- mice (Fig 5B–5D). Furthermore, wild-type mice fed TDs had similar levels of hepatic glycogen, hepatic glucose, and plasma glucose as Oct1-/- mice irrespective of the thiamine content in their diets, consistent with the idea that Oct1 deficiency mimics thiamine deficiency in wild-type mice. Levels of G6P, an activator of GS, were significantly higher in livers from wild-type mice fed a TD diet and were comparable to liver levels of G6P in Oct1-/- mice irrespective of thiamine content in the diet (Fig 5E). As shown by western blotting (Fig 5F), livers from Oct1-/- mice in the control thiamine diet group and from both Oct1+/+ mice and Oct1-/- mice in the TD group had higher GS and Glut2 protein levels compared to Oct1+/+ mice in the thiamine control group. Taken together, our data suggest that thiamine deficiency impairs glucose metabolism in wild-type mice and that Oct1 deficiency phenocopies thiamine deficiency in wild-type mice. Oct1-/- mice exhibited increased adiposity (Fig 2C and 2D), and examination of fat cells through staining revealed significantly larger adipose cells in the epididymal fat pad (epididymal white adipose tissue [eWAT], p = 0. 004) and a trend toward larger adipose cells in retroperitoneal adipose tissue (rpWAT) from Oct1-/- mice (Fig 6A). To probe the mechanism of increasing adiposity and adipose cell size in the Oct1-/- mice, we measured the mRNA expression levels of genes related to adipose metabolism. Fat gain may be due to imbalances between rates of TG synthesis and lipolysis. The mRNA expression of patatin-like phospholipase domain-containing protein 2 (Pnpla2) and lipase, hormone sensitive (Lipe) involved in adipose lipolysis was reduced in adipose tissue from Oct1-/- mice compared to adipose tissue from Oct1+/+ mice (Fig 6B). In contrast, levels of genes involved in TG synthesis were similar between the two strains of mice (S5A Fig). Pnpla2 (coding for adipose triglyceride lipase [ATGL]), Lipe (coding for hormone sensitive lipase [HSL]), and Mgll (coding for monoglyceride lipase [MGLL]) are responsible for three major steps in mobilizing fat through hydrolysis of TGs to release free fatty acids from the adipocytes [53]. Lower expression levels of these genes are consistent with lower rates of lipolysis in adipose tissue from Oct1-/- mice. Insulin has antilipolytic effects in adipose tissue, regulating ATGL expression and promoting lipid synthesis, and chronic insulin treatment results in increased adipose mass [54,55]. Corresponding to the higher levels of glucose (Fig 6C), we observed higher circulating levels of insulin in the Oct1-/- mice (Fig 6D and S5B Fig), which suppressed lipolysis. Furthermore, fasting free fatty acid levels were lower in the plasma of Oct1-/- mice (Fig 6E), which may reflect the lower rates of lipolysis in adipose tissue [56]. Data in Oct1 knockout mice were corroborated by data from inbred strains of mice. In particular, Oct1 mRNA levels in the liver inversely associated with percent fat growth and fat mass among various strains of mice (S1 Table). In addition, down-regulation of mitochondrial uncoupling protein 2 (Ucp2) was observed in brown adipose in Oct1-/- mice (S5G Fig), which may associate with the reduced energy expenditure. Examination of total cholesterol, HDL cholesterol, LDL cholesterol, and TG in plasma samples revealed significant differences in the two strains of mice. Notably, Oct1-/- mice had higher plasma levels of total cholesterol and LDL cholesterol compared to Oct1+/+ mice, without significant differences in TG and HDL cholesterol (Fig 6F). The increase in LDL was due primarily to smaller LDL particles (Fig 6G). We observed no differences in the transcript levels of lipoprotein lipase (Lpl) and Ldlr in livers from Oct1+/+ mice and Oct1-/- mice. However, livers from Oct1-/- mice had higher transcript levels of 3-hydroxy-3-methylglutaryl-CoA reductase (Hmgcr), and Acyl-CoA: cholesterol acyltranferase 2 (Acat2) (S5E Fig). Consistent with lower activity of PDH, pyruvate levels were significantly higher in the livers from Oct1-/- mice, as less pyruvate was converted to acetyl-CoA. Interestingly, contrary to our expectation, Oct1-/- mice had higher levels of acetyl-CoA in their livers (Fig 6H and 6I). The higher accumulated acetyl-CoA may have resulted from higher fatty acid β-oxidation in Oct1-/- mice [17,48]. Our data suggest that up-regulation of enzymes involved in cholesterol synthesis and higher levels of the substrate precursor, acetyl-CoA, in the liver of Oct1-/- mice result in alterations in hepatic cholesterol metabolism, leading to increased production of LDL particles. Furthermore, lower thiamine levels were correlated with higher levels of cholesterol in plasma and liver in male mice from various inbred strains of mice (S2 Table and S6 Fig).
Through extensive characterization of Oct1 knockout mice, our data provide compelling evidence that Oct1 deficiency leads to a constellation of diverse effects on energy metabolism that are consistent with GWAS demonstrating strong associations between OCT1 polymorphisms and a variety of metabolic traits in humans (Fig 1 and Table 1). Our data support the notion that hepatic thiamine deficiency is the underlying mechanism for the phenotypes associated with reduced OCT1 function. Five major effects of OCT1 deficiency emerge from the current study: (1) a shift in the pathway of energy production from glucose to fatty acid oxidation due to lower activity of key thiamine-dependent enzymes in the liver; (2) increased gluconeogenesis and hepatic glucose output, with associated increases in liver glycogen and glucose levels; (3) increased peripheral adiposity stemming from alterations in energy metabolism; (4) changes in hepatic cholesterol homeostasis and plasma lipids that may contribute to cardiovascular disease risk; and (5) beneficial effects on life-threatening thiamine deficiency syndromes. As the major energy-generating organ, the liver has high metabolic flexibility in selecting different substrates to use in energy production in response to various metabolic conditions. The glucose–fatty acid cycle, first proposed by Randle in 1963, plays a key role in regulating metabolic fuel selection, and impairment in metabolic flexibility contributes to insulin resistance and metabolic syndrome [17,18,47,57,58]. Many studies have shown that there is a failure to shift from fatty acid to glucose oxidation during the transition from fasting to feeding in individuals with obesity or diabetes [48,57]. We observed lower hepatic steatosis (Fig 2A), largely due to increases in fatty acid oxidation in the liver [1,59,60]. Importantly, the observation that the Oct1-/- mice had significantly lower RERs during the dark cycle than their wild-type counterparts suggests that overall, Oct1-/- mice have a greater reliance on energy production from fatty acids over glucose during feeding than wild-type mice [61,62] (S2J Fig). PDH is the key enzyme switch for the glucose–fatty acid cycle [17,63]. Thus, alterations in its activity by reduced levels of the cofactor TPP in Oct1-/- mice (Fig 4A) disrupt the hepatic glucose–fatty acid cycle, resulting in an impairment of hepatic energy homeostasis. In livers from Oct1 deficient mice, β-oxidation of fatty acids becomes a major source of energy production, leading to impaired homeostasis in both lipid and carbohydrate metabolism. In both the current study and our previous study [14], we observed increased levels of phosphorylated 5′ adenosine monophosphate-activated protein kinase (AMPK) and its downstream target, acetyl-CoA carboxylase (ACC), in livers from Oct1-/- mice compared to livers from wild-type mice, indicative of a lower hepatic energy status in the knockout mice. As a result, fatty acid β-oxidation was stimulated in the livers from Oct1-/- mice. However, as was evident by lower hepatic ATP content, the increased rates of fatty acid oxidation were not sufficient to compensate for normal rates of ATP production. Reduced glucose oxidation as well as a reduction in flux through the TCA cycle due to reduced activity of α-KGDH, another TPP-associated enzyme, may have contributed to the lower ATP production. Consistent with a lower flux through the TCA cycle as well as increases in β-oxidation of fatty acids, we observed higher levels of acetyl-CoA in livers from Oct1-/- mice. Studies have shown that acetyl-CoA allosterically inhibits PDH, which results in further inhibition of glucose utilization [48,58,63]. Thus, in the livers from Oct1-/- mice, this loop continued to stimulate fatty acid β-oxidation, which further suppressed hepatic glucose utilization, shifting the major energy source from glucose to fatty acids. Increases in hepatic glycogen and glucose content in the Oct1-/- mice (Fig 2A, Fig 5B and 5C) were likely due to changes in intermediary metabolites resulting directly from alterations in the activity of the TPP-dependent enzyme, PDH (Fig 4A). Reduced PDH activity resulted in higher levels of pyruvate in the liver of Oct1-/- mice (Fig 6H), which is consistent with results from previous studies [63,64]. Higher pyruvate levels can drive hepatic gluconeogenesis, resulting in increased hepatic glucose production and associated increases in hepatic glucose and glycogen levels [1,52]. Whereas our data suggested that the glycogen accumulation in the livers of Oct1-/- mice resulted from changes in key intermediate metabolites that are involved in hepatic energy metabolism, other regulatory paths such as hormonal, transcriptional, and neural regulation need to be further studied. Oct1-/- mice exhibited increased adiposity (Fig 2), which was more likely due to downstream effects of reduced transporter expression in the liver rather than in extra-hepatic tissues. Multiple factors contributed to the increased adiposity, such as hyperinsulinemia, hyperglycemia, increased hepatic glycogen, and reduced energy expenditure. Consistent with the increased adiposity observed in the Oct1-/- mice, hepatic expression levels of Oct1 are inversely correlated with fat growth and fat mass in inbred strains of mice (S1 Table). In addition, hyperinsulinemia in Oct1 knockout mice may further result in increasing storage of TGs and suppression of lipolysis in peripheral adipose tissue, which reduced flux of fatty acids to the liver. Chronic insulin treatment has been shown to result in increased adipose mass due to suppression of lipolysis and increased lipid storage [54,55], and in the current study, a high correlation between plasma insulin levels and fat mass in both wild-type and Oct1-/- mice (S5D Fig) was observed. Furthermore, high insulin levels have been associated with low expression levels of the lipolytic enzyme Pnpla2 in adipose tissue [55], consistent with results in the Oct1-/- mice (Fig 6B), which is in agreement with the increased adiposity in these mice. In addition, high hepatic glycogen levels may have contributed to the increased adiposity. In particular, hepatic glycogen levels regulate the activation of the liver–brain–adipose axis [65]. Glycogen shortage during fasting triggers liver–brain–adipose neurocircuitry that results in stimulation of fat utilization. In contrast, in mice with elevated liver glycogen resulting from overexpression of GS or knockdown of PYGL, the liver–brain–adipose axis action is turned off, which preserves fat mass [65]. The greater stores of glycogen in the livers of Oct1-/- mice may have shut off the liver–brain–adipose axis, contributing to the increased peripheral adiposity in the mice. Overall, our data in Oct1-/- mice suggest that OCT1 plays a key role in regulation of peripheral metabolism, likely because of its effects on circulating glucose and insulin as well as increased stores of hepatic glycogen triggering a feedback loop mechanism between the liver, the brain, and the adipose tissue. Parallels between phenotypes observed in GWAS in humans and those in the Oct1-/- mice were striking. High plasma LDL, total cholesterol, and TG levels were observed in individuals with reduced-function polymorphisms of OCT1 (R61C, F160L, G401S, V408M, 420del, and G465R) (Table 1) as well as in the Oct1-/- mice (Fig 6F and 6G). In particular, the Oct1-/- mice, and humans with reduced-function polymorphisms of OCT1 (S3 Table) [37], have increased levels of small dense LDL particles (Fig 6G), which in humans are predictive of increased risk of cardiovascular disease [66] and are a characteristic feature of the dyslipidemia associated with excess adiposity [67] and insulin resistance [68]. We speculate that the increased LDL levels in Oct1-/- mice may result from a relative deficiency of hepatic thiamine. Specifically, livers from rats with thiamine deficiency have been shown to have lower TG but higher cholesterol content [69], consistent with our data in inbred strains of mice, in which inverse correlations were observed between plasma thiamine and cholesterol levels in both plasma and liver (S2 Table and S6 Fig). Importantly, our in vivo studies in mice showed that reduction of liver Slc22a1 expression levels resulted in higher systemic thiamine levels (Fig 3B and 3E). Furthermore, individual OCT1 polymorphisms were nominally associated with systemic plasma levels of thiamine in humans (S4 Table) [70] as well as when combining six of the OCT1 nonsynonymous variants that were genotyped in the cohort by Rhee and colleagues (see S5 Table) [71]. Published studies in animals and humans suggest that thiamine supplementation may improve blood lipid profiles [24,25]. Higher levels of the precursor for cholesterol synthesis, acetyl-CoA, as well as higher expression levels of enzymes involved in cholesterol synthesis [59,72] (S5E Fig), potentially leading to greater rates of hepatic cholesterol production, could have contributed to the higher cholesterol and LDL levels in plasma. Although we were unable to detect differences in total hepatic cholesterol content between Oct1+/+ mice and Oct1-/- mice (S5F Fig) corresponding to the observed differences in plasma cholesterol levels between the mouse strains, many factors that can modulate hepatic cholesterol content [73,74], including perhaps increased export, need further investigation. Further studies are warranted to investigate the mechanisms underlying the effects of reduced OCT1 function as well as thiamine bioavailability on cholesterol and lipoprotein metabolism. In addition to the metabolic changes that were observed in Oct1-/- mice, the knockout mice were found to survive substantially longer on TDs. This may have been due to the higher systemic levels of thiamine (Fig 3B), which would spare essential organs such as the brain and heart from thiamine deficiency, as well as to the increased adiposity in the knockout mice (Fig 2C), which could protect the mice from the starvation that ensues from thiamine deficiency [42,75,76]. Nevertheless, the results have implications for human ancestors who harbored reduced-function genetic polymorphisms of OCT1. Because of differences in the tissue distribution of OCT1 between humans and mice, our study has limitations in directly extrapolating all the results obtained in mice to humans. In particular, because Oct1 is also abundantly expressed in the kidney of mice, OCT1-mediated renal secretion of thiamine is another important determinant of systemic thiamine levels in mice. In contrast, in humans, OCT1 plays a role in modulating thiamine disposition largely in the liver and not in the kidney. Deletion of Oct1, particularly in the kidney of the knockout mice, therefore, may have modulated systemic thiamine levels and, thus, survival during TDs as well as other phenotypes observed in the current study. Today, thiamine deficiency is associated with aging, diabetes, alcoholism, and poor nutritional status [23,77–79]. In the setting of thiamine deficiency, OCT1 reduced-function polymorphisms today would have mixed effects. On the one hand, individuals who harbored reduced-function variants would have higher systemic levels of thiamine, which may protect essential organs from thiamine depletion. On the other hand, the individuals would have low hepatic thiamine levels, which may predispose them to the deleterious effects of dysregulated plasma lipids and to obesity and diabetes (Table 1). In fact, lower thiamine levels have been reported in individuals with diabetes [23,78] and, as noted, some studies have shown that high-dose thiamine supplementation has beneficial effects on diabetes [25,80,81]. Overall, the current study shows that OCT1 deficiency triggers a constellation of effects on hepatic and overall energy homeostasis (Scheme in Fig 6J). That is, reduced OCT1-mediated thiamine uptake in the liver leads to reduced levels of TPP and a decreased activity of key TPP-dependent enzymes, notably PDH and α-KGDH. As a result, there is a shift from glucose to fatty acid oxidation, which leads to imbalances in key metabolic intermediates, notably, elevated levels of pyruvate, G6P, and acetyl-CoA. Because of these imbalances, metabolic flux pathways are altered, leading to increased gluconeogenesis and glycogen synthesis in the liver. In addition, the increased acetyl-CoA levels along with elevated expression levels of key enzymes involved in cholesterol synthesis likely contribute to increases in plasma levels of total and LDL cholesterol observed in mice with Oct1 deficiency and in humans with reduced-function genetic polymorphisms of OCT1. Although many of the details of the mechanisms have still to be worked out, our study provides critical insights into the role of thiamine in the liver in maintaining metabolic balance among energy metabolism pathways. Finally, our studies provide mechanistic insights into findings from GWAS implicating reduced-function variants in the SLC22A1 locus as risk factors for lipid disorders and diabetes.
Animal experiments were approved by the Institutional Animal Care and Use Committee (IACUC) of University of California, San Francisco (AN119364), in accordance with the requirements of the National Research Council Guide for the Care and Use of Laboratory Animals and the Public Health Service Policy on the Humane Care and Use of Laboratory Animals. Humane end points were determined by body condition score of 2 or less or 15% body weight loss. Animals were euthanized once the humane end points were reached during the treatment in accordance with IACUC approved protocol. To limit pain and stress, mice were anesthetized deeply by isoflurane vaporizer and intraperitoneal injection of ketamine/medetomidine cocktail (75/1 mg/kg) prior to the physical cervical dislocation of euthanasia. Various publicly available databases were used to determine whether there are significant genetic associations of SLC22A1 reduced-function variants with human diseases and traits. The following databases were used: GWAS Catalog, dbGAP Association Results Browser, GRASP: Genome-Wide Repository of Associations Between SNPs [82], GIANT Consortium, Type 2 Diabetes Knowledge Portal, and Phenotypes and Genome Wide Associations Studies for Lipid Genetics. The first three databases, GWAS Catalog, dbGAP Association Results Browser, and GRASP, provide an easy to use interface to allow first-step information gathering of the types of human diseases and traits that have been reported in all published GWAS. Based on the results from the three databases, other specific databases relevant to the findings were then used. This includes searching for specific databases that have the GWAS summary statistics (beta coefficient and p-values). These are GIANT Consortium for body weight, Type 2 Diabetes Knowledge Portal for glucose and insulin traits, and all GWAS for lipid traits. These databases allow investigators to download the association studies for obtaining the p-values and the beta coefficients for the associations. In this study, we focused our search on nonsynonymous variants of OCT1 (SLC22A1) with minor allele frequencies ≥1% in populations with European ancestries (1000 Genome Project): R61C (rs12208357), F160L (rs683369), P341L (rs2282143), G401S (rs34130495), M408V (rs628031), 420Del (rs202220802) (rs662138 and rs1564348, which are in linkage disequilibrium to 420Del with r2 ≥ 0. 77, D′ > 0. 95), and G465R (rs34059508). All experiments on mice were approved by the IACUC of UCSF. Oct1-/- mice were generated as previously described [83] and backcrossed more than 10 generations to FVB/N background. Mice were housed in a pathogen-free facility with a 12-hour light and 12-hour dark cycle and given free access to food and water. Five- or six-week-old experimental mice were fed with thiamine control diet Cat# TD. 09549 (thiamine 5 mg/kg) containing 17. 5% protein, 65. 8% carbohydrate, and 5. 0% fat by weight (Envigo, Madison, WI). Other thiamine diets contained the same composition as the thiamine control diet but differed in thiamine levels (TD Cat# TD. 81029,0 mg/kg; adjusted thiamine diet with different thiamine doses added, Cat# TD. 120472,25 mg/kg; Cat# TD. 140164,50 mg/kg). The periods of dietary treatments and time for mouse being humanely killed are indicated in the Results section and figure legends. The animal studies were conducted in male mice; however, overall body weight and liver weight were assessed in female mice. Mice treated with TD developed thiamine deficiency syndromes, resulting in reduction of food intake and body weight loss. During the treatment period, mice were closely monitored and weighed daily. Animals were euthanized once the humane end points (body condition score of 2 or less or 15% body weight loss) were reached during the treatment. To limit pain and stress, mice were anesthetized deeply by isoflurane vaporizer and intraperitoneal injection of ketamine/medetomidine cocktail (75/1 mg/kg) prior to the physical cervical dislocation of euthanasia. Hydrodynamic tail vein injection procedure was conducted as described previously [84], with minor modifications. Briefly, the body weights of the mice were used to calculate the total volume (mL) required for injection based on the formula: body weight (g) * (mL/10g) + 0. 1 mL (dead volume). The injection solution included 5*10^7 TU virus/mouse and saline to just the final volume. Instead of anesthetizing the mice, we used TransIT-QR kit MIR5210 (Mirus Bio LLC, US) and followed the online protocol (https: //www. mirusbio. com/delivery/tailvein/). Dosing of 3H-thiamine was performed 48 hours after hydrodynamic tail vein injection. Mouse Slc22A1 shRNA lentiviral particle (TRCN0000070156) and nonmammalian shRNA control pLKO. 1 (SHC002V) were purchased from Sigma. Viruses were verified in HEK-293 cells. Briefly, pcDNA5 containing mouse Slc22a1 was cotransfected with the Slc22a1 knockdown vector or pLKO. 1 control virus. mRNA was isolated after 48 hours transduction, and mRNA expression of Oct1 was measured. Before and during dietary treatment, body composition was determined by either quantitative magnetic resonance on the EchoMRI-3in1 body composition analyzer (EchoMRI, Houston, TX) or by DEXA. For DEXA, live animals were anesthetized with isoflurane and scanned on the Lunar PIXImus densitometer (Lunar PIXImus Corporation Headquarters, Madison, WI). After 8 weeks diet treatment, mice were placed in single housing cages for 3 days before initiating the CLAMS (Columbus Instruments, Columbus, OH) experiments. CLAMS was used to monitor food and water intake, oxygen consumption (VO2) and carbon dioxide production (VCO2), and locomotor activities for a period of 96 hours. All these experiments were performed in the Diabetes and Endocrinology Research Center Mouse Metabolism Core at UCSF. Blood glucose levels from mice were measured using the FreeStyle Freedom Lite blood glucose meter (Abbott Laboratories, Chicago, IL) in samples obtained by the tail milking method. For oral glucose tolerance tests (OGTTs), mice were fasted for 5 hours and dosed with glucose 2 g/kg (Sigma-Aldrich, St. Louis, MO) by oral gavage. For ITTs, mice were fasted for 5 hours and dosed with 0. 75 U/kg humulin R insulin 100 U/ml (Henry Schein Animal Health, Dublin, OH) by intraperitoneal injection. Blood was sampled at 0,15,30,60,90, and 120 minutes. For PTTs, mice were fasted overnight for 16 hours and dosed with pyruvate 2 g/kg (Sigma-Aldrich, St. Louis, MO) by intraperitoneal injection. Blood was sampled at 0,15,30,60,90,120,150, and 180 minutes. For adipose tissues and liver glycogen staining, mice were fasted for 16 hours and perfused with 20 mL 4% paraformaldehyde (PFA) in PBS. Epididymal fat pad, rpWATs, and liver were incubated in 4% PFA for 48 hours at 4°C and transferred to 70% ethanol. For Oil Red-O (ORO) staining in liver, mice were fasted for 16 hours and perfused with 10 mL PBS. Livers were fixed via sucrose infiltration steps prior to freezing. After incubating in 30% sucrose in PBS at 4°C for 24 hours, tissues were frozen in OCT molds. Fixed or frozen tissues were transferred to the histology and light microscopy core at Gladstone Institutes for staining, imaging, and analysis. For the cell-size analysis, hematoxylin and eosin-stained paraffin-embedded sections (https: //labs. gladstone. org/histology/pages/section-staining-haematoxylin-and-eosin-staining) of mouse adipose tissues were imaged using a Nikon Eclipse E600 upright microscope equipped with a Retiga camera (QImaging, Vancouver, BC, Canada) and a Plan Fluor 20×/0. 3NA objective. For each sample, four independent fields were imaged for analysis and adipocyte size was determined using ImageJ (v. 2. 0. 0-rc-3) software (US National Institutes of Health) and the Tissue Cell Geometry macro (http: //adm. irbbarcelona. org/image-j-fiji). For quantifying the lipid droplets, ORO stained frozen sections of mouse liver were imaged as above. For each sample, four independent fields were imaged for analysis. RGB images were then color thresholded to ORO, and the total area of ORO-positive pixels was summed for each image using the Analyze Particles function. For quantifying the glycogen levels in liver sections, Periodic-Acid Schiff’s stained mouse liver was imaged as above. For each sample, four independent fields were imaged and the mean intensity and integrated density values were averaged for each image using the Analyze Particles function. Mice were humanely killed and blood was collected via posterior vena cava to BD microtainer tubes with dipotassium EDTA (365974) or heparin (365985). Plasma was sent to the Clinical Laboratory of San Francisco General Hospital for measurement of total, LDL, and HDL cholesterol and TGs and liver panel. Plasma was send to Children' s Hospital Oakland Research Institute for the measurement of lipoprotein particles size, as described previously [85]. Glucose (GAGO20), glycogen (MAK016), free fatty acid (MAK044), acetyl-Coenzyme A (MAK039), and G6P (MAK014) quantification kits were purchase from Sigma-Aldrich (St. Louis, MO). Pyruvate (ab65342), cholesterol (ab102515), and TG (ab65336) quantification kits were purchase from Abcam (Cambridge, MA). Plasma insulin was measure by ELISA (EMINS) from Thermo Fisher Scientific (Waltham, MA). Plasma was sent to Molecular MS Diagnostics, Inc. (Warwick, RI), for thiamine quantification, as previously described [14]. Tissues were homogenized in CelLytic MT lysis buffer (Sigma-Aldrich, St. Louis, MO) with cOmplete ULTRA protease inhibitor tablet and PhosSTOP phosphatase inhibitor tablet freshly added (Roche). After normalization, equal protein amounts from each sample were loaded in to 4%–20% criterion Tris-HCl gel (Bio-Rad, Hercules, CA) and run at 110 V. Protein was transferred to EMD milllipore immobilon PVDF membranes at 100 V for 1 hour at 4°C using the criterion blotter. GS (15B1) (#3886; 1: 1,000 dilution), phospho-glycogen synthase (Ser641) (#3891,1: 1,000 dilution), phospho-acetyl-CoA carboxylase (Ser79) (#3661; 1: 2,000 dilution), AMPKα (23A3) (#2603,1: 1,000 dilution), and phospho-AMPKα (Thr172) (#2535; 1: 1,500 dilution) were purchased from Cell Signaling Technology (Danvers, MA). Anti-PDH-E1α (pSer232) (#AP1063; 1: 2,000 dilution) and Anti-PDH-E1α (pSer300) (#AP1064; 1: 2,000 dilution) were purchased from EMD Millipore (Billerica, MA). PDH-E1α (D-6) (#sc-377092; 1: 200 dilution) and β-actin (C4) (#sc-47778; 1: 4,000 dilution) were purchased from Santa Cruz Biotechnology, Inc. (Dallas, TX). Anti-PYGL (#ab198268; 1: 1,000 dilution) was purchased from Abcam (Cambridge, MA). Anti-Glut2 (#600-401-GN3; 1: 1,000 dilution) was purchased from Rockland (Limerick, PA). Primary antibodies were incubated overnight at 4°C. Secondary antibodies, goat anti-rabbit IgG-HRP (#sc-2030; 1: 5,000; Santa Cruz Biotechnology, Inc.), and anti-mouse IgG-HRP (#7076; 1: 5,000; Cell Signaling Technology) were incubated for 2 hours at room temperature. Either Amersham ECL western blotting detection reagents (RPN2106) or ECL Prime western blotting detection reagents (RPN2232) were used for detection. Membranes were imaged by ProteinSimple western blot imaging system (San Jose, CA). For the quantification of western blot bands, the ImageJ (US National Institutes of Health) method was used. Total RNA from mouse tissues or cell lines was isolated using RNeasy Mini kit (Qiagen, Valencia, CA). Total RNA (2 μg) from each sample was reverse transcribed into cDNA using SuperScript VILO cDNA Synthesis kit (Life Technologies, CA). Quantitative real-time PCR was carried out in 384-well reaction plates using 2X Taqman Fast Universal Master Mix (Applied Biosystems, Foster City, CA), 20X Taqman specific gene expression probes, and 10 ng of the cDNA template. The reactions were carried out on an Applied Biosystems 7900HT Fast Real-Time PCR System (Applied Biosystems, Foster City, CA). The relative expression level of each mRNA transcript was calculated by the comparative method (ΔΔCt method), normalized to the housekeeping gene, β-actin. The stably overexpressing pcDNA5 empty vector, mouse OCT1, human OCT1-reference, OCT1-420 del, and OCT1-420 del+G465R cell lines were maintained in Dulbecco’s Modified Eagle Medium (DMEM H-21) supplemented with hygromycin B (100 ug/mL) (Thermo Fisher Scientific, Waltham, MA), penicillin (100 U/mL), streptomycin (100 mg/mL), and 10% fetal bovine serum. Cell culture supplies were purchased from the Cell Culture Facility (UCSF, CA). Cells were cultured on poly-D-lysine coated 96-well plates for 24 hours to reach 95% confluence. Before the uptake experiments, the culture medium was removed and the cells were incubated in Hank’s balanced salt solution (HBSS) (Life Technology, CA) for 15 minutes at 37°C. Radiolabeled thiamine [3H (G) ] hydrochloride (20 Ci/mmol) was purchased from American Radiolabeled Chemicals Incorporation (St. Louis, MO). Thiamine hydrochloride was purchased from Sigma-Aldrich (St. Louis, MO). Chemicals and radiolabeled compounds were diluted in the HBSS for uptake experiments. The details for drug concentrations and uptake time are described in the Results section and figure legends. The uptake was performed at 37°C, and then the cells were washed three times with ice-cold HBSS. After that, the cells were lysed with buffer containing 0. 1 N NaOH and 0. 1% SDS, and the radioactivity in the lysate was determined by liquid scintillation counting. For the transporter study, the Km and Vmax were calculated by fitting the data to a Michaelis-Menten equation using GraphPad Prism software 6. 0 (La Jolla, CA). All mice were randomly assigned to the control or each treatment group. No statistical method was used to predetermine sample size, and sample size was determined on the basis of previous experiments. Numbers of mice for each experiment are indicated in figure legends. Mice that were dead or sick before the end of experiments were excluded from the final analysis. Investigators were not blinded during experiments. Data were expressed as mean ± SEM. Appropriate statistical analyses were applied, as specified in the figure legends. Data were analyzed using GraphPad Prism software 6. 0 (La Jolla, CA). Differences were considered statistically significant at p < 0. 05; *p < 0. 05, **p < 0. 01, and ***p < 0. 001. | The liver is the major organ for glucose and lipid metabolism; impairment in liver energy metabolism is often found in metabolic disorders. Traditionally, excesses in macronutrients (fat and glucose) are linked to the development of metabolic disorders. Our study provides evidence that imbalances in a micronutrient, vitamin B1 (thiamine), can serve as an etiological cause of lipid and glucose disorders and implicates the organic cation transporter, OCT1, in these disorders. OCT1 is a key determinant of thiamine levels in the liver. In humans, reduced-function polymorphisms of OCT1 significantly associate with high LDL cholesterol levels. Using Oct1 knockout mice, we show that reduced OCT1-mediated thiamine uptake in the liver leads to reduced levels of TPP—the active metabolite of thiamine—and decreased activity of key TPP-dependent enzymes. As a result, a shift from glucose to fatty acid oxidation occurs, leading to imbalances in key metabolic intermediates, alterations in metabolic flux pathways, and disruptions of various metabolic regulatory mechanisms. The extensive characterization of Oct1 knockout mice provides evidence for the molecular mechanisms responsible for various metabolic traits and indicates an important role for imbalances in micronutrients in cardiometabolic disorders. | Abstract
Introduction
Results
Discussion
Materials and methods | b vitamins
carbohydrate metabolism
genome-wide association studies
medicine and health sciences
body fluids
chemical compounds
carbohydrates
glucose metabolism
organic compounds
glucose
glycobiology
genome analysis
lipids
genomics
chemistry
vitamins
blood plasma
cholesterol
biochemistry
blood
organic chemistry
anatomy
glycogens
physiology
monosaccharides
genetics
biology and life sciences
physical sciences
computational biology
fatty acids
metabolism
thiamine
human genetics | 2018 | Organic cation transporter 1 (OCT1) modulates multiple cardiometabolic traits through effects on hepatic thiamine content | 14,281 | 292 |
The peptide repertoire that is presented by the set of HLA class I molecules of an individual is formed by the different players of the antigen processing pathway and the stringent binding environment of the HLA class I molecules. Peptide elution studies have shown that only a subset of the human proteome is sampled by the antigen processing machinery and represented on the cell surface. In our study, we quantified the role of each factor relevant in shaping the HLA class I peptide repertoire by combining peptide elution data, in silico predictions of antigen processing and presentation, and data on gene expression and protein abundance. Our results indicate that gene expression level, protein abundance, and rate of potential binding peptides per protein have a clear impact on sampling probability. Furthermore, once a protein is available for the antigen processing machinery in sufficient amounts, C-terminal processing efficiency and binding affinity to the HLA class I molecule determine the identity of the presented peptides. Having studied the impact of each of these factors separately, we subsequently combined all factors in a logistic regression model in order to quantify their relative impact. This model demonstrated the superiority of protein abundance over gene expression level in predicting sampling probability. Being able to discriminate between sampled and non-sampled proteins to a significant degree, our approach can potentially be used to predict the sampling probability of self proteins and of pathogen-derived proteins, which is of importance for the identification of autoimmune antigens and vaccination targets.
Major histocompatibility complex (MHC) class I molecules play a crucial role in the adaptive immune response of higher vertebrates. These molecules, in humans referred to as human leukocyte antigen (HLA) class I molecules, bind peptides derived from endogenous proteins of host or, in the case of infected cells, of pathogen origin and present them to circulating CD8+ T lymphocytes and natural killer (NK) cells. The presentation of self peptides by an individual' s HLA class I molecules has an impact on positive and negative selection of CD8+ T lymphocytes in the thymus [1], [2], maintenance of naive T cells in the periphery [3], [4], and inhibition of NK cells through recognition of self peptides in the context of HLA class I molecules by killer cell immunoglobulin-like receptors (KIR) [5]. Generally, HLA class I ligands are derived from intracellular proteins, which are degraded by the proteasome into peptide fragments. These peptides are then translocated by the transporter associated with antigen processing (TAP) into the lumen of the endoplasmic reticulum (ER), where they may be loaded onto an HLA molecule if the peptide sequence fits the HLA molecule' s binding preference. The C-terminus of an HLA ligand is assumed to be mainly determined by the proteasome (even though recently a carboxypeptidase has been found to contribute to C-terminal editing [6]), whereas the N-terminus may be trimmed by cytosolic and endoplasmic aminopeptidases after proteasomal cleavage [7]. Finally, the HLA-peptide complexes are transported to the cell surface for presentation to CD8+ T cells and NK cells. Several studies analyzed peptide data sets obtained by peptide elution from specific cell lines and peptide sequencing by mass spectrometry to characterize the HLA peptide repertoire [8], [9], [10], [11]. Most of these studies focused on characterizing the function and subcellular localization of source proteins and suggested that HLA class I presented peptides are sampled from functionally and compartmentally diverse proteins, with a functional bias towards RNA-binding proteins [8]. In human cells, a weak correlation has been found between the abundance of HLA class I ligands presented and the corresponding mRNA levels [10], [11], whereas peptides eluted from murine thymocytes were preferentially derived from highly abundant mRNAs [12]. Here, we take a different angle to the question of what fraction of the human proteome is represented on the cell surface. We studied two large HLA ligand data sets obtained by peptide elution [13], [14] with the aim to quantify the role of several factors shaping the peptide repertoire of HLA class I molecules. We show that the gene expression level, protein abundance, rate of potential binding peptides in a protein and the processing quality of these peptides all contribute to which proteins are sampled and which peptides are chosen to be presented on the cell surface. Having studied the impact of each of these factors separately, we subsequently combined all factors in a logistic regression model in order to quantify their relative impact. This model can potentially be used to predict the sampling probability of self proteins and of pathogen-derived proteins.
We studied two different peptide elution data sets. One large set, which we will call the Johnson data, comprises 4717 human peptides and 105 vaccinia peptides eluted from vaccinia infected cells [13]. A second set of eluted peptides, the Ben Dror data, comprises 569 human peptides eluted from soluble HLA-B*27: 05 (see Materials and Methods for details) [14]. By mapping each eluted peptide to the human proteome, we were able to uniquely identify the source protein for 90% (4243 of 4717) of the Johnson data and 81. 9% (466 of 569) of the Ben Dror data. Peptides that mapped to several human proteins (about 9%) or for which no source protein could be identified (only 0. 4–1. 1%) were excluded from further analysis. The cell line used for the generation of the Johnson data was homozygous for HLA-A*02: 01, B*15: 01 and C*03. In order to assess which of these three HLA molecules each of the reported peptides was eluted from, we employed NetMHC 3. 2 [15], [16], a tool for HLA-peptide binding prediction. This tool is applicable for peptides of 8 to 13 amino acids in length [15]. Of all eluted peptides of appropriate length and that could be mapped uniquely to a human source protein (4113 of 4243 peptides), we were able to assign 86. 4% (3552 of 4113) to either being eluted from A*02: 01 or B*15: 01. The remaining 561 peptides could potentially have been eluted from C*03. Binding predictions, however, suggested C*03 binding only for a minor fraction of these (28% for all peptide lengths and 37% for 9mers), and therefore we decided to exclude these peptides from further analysis. Surprisingly, we identified twice as many potential B*15: 01 binders as A*02: 01 binders, originating from a larger number of source proteins (Fig. 1). This observation is in agreement with the estimation given by the original study [13], in which the assignment to the restricting HLA molecule was solely determined based on the C-terminal residues of the eluted peptides. Likewise, among the vaccinia-derived peptides, a larger number of peptides (1. 5-fold) were eluted from B*15: 01 than from A*02: 01, mapping to a larger number of vaccinia proteins, even though in this case the difference was less pronounced (Fig. 1). Of the 569 eluted peptides in the Ben Dror data, 466 peptides (81. 9%) mapped uniquely to 396 human source proteins. Among these peptides, 420 (90. 1%) were predicted to bind to HLA-B*27: 05, the soluble HLA molecule expressed by the cells studied. For all three HLA alleles studied, the majority of sampled proteins were represented by only a single peptide: 86. 9% (457 of 526) of the proteins sampled by A*02: 01,76. 1% (981 of 1289) of the ones sampled by B*15: 01, and 86. 4% (342 of 396) of the proteins sampled by B*27: 05 gave rise to only one eluted peptide. In total, the two elution data sets had 160 source proteins in common. GO-term enrichment analysis (see Materials and Methods) revealed that for this set of proteins biological processes relating to the cell cycle and its regulation as well as nucleic acid metabolic processes were overrepresented. The observation that a protein is represented on the cell surface by one or more peptides allows the assumption that the protein must have been available in sufficient amounts or must have been present at an accessible subcellular location to be available for the antigen processing machinery. What factors then determine which of the potential HLA binders of a given protein will be found on the cell surface? In order to characterize the obtained peptide set, we employed prediction methods for HLA binding and antigen processing (see Materials and Methods). For the identified source proteins, we predicted all potential (9mer) binders to HLA-A*02: 01, B*15: 01, and B*27: 05 and compared the predicted binding affinity of the eluted peptides (which form a subset of all potential binders) with the predicted binders from the same source protein that were not found in the elution. To ensure an unbiased comparison, the set of eluted peptides was limited to 9mers that were predicted to bind to the respective HLA molecule. We found that eluted peptides bind their HLA molecule with a significantly higher (predicted) binding affinity than other potential binders (Fig. 2A). However, not all predicted high-affinity binders were found in the elution. In order to investigate whether this observation may be due to inefficient processing of these peptides, we predicted the probability of C-terminal processing (using NetChop [17]) for all eluted peptides and all predicted binders that were not found in the elution data set. Based on these predictions, the set of eluted peptides is indeed more likely to arise from antigen processing as compared to the set of predicted binders (Fig. 2B). Possibly due to co-evolution between HLA class I molecules and the immunoproteasome [18], predicted binding affinity and C-terminal processing probability show a weak (but significant) correlation (Kendall' s tau = −0. 065, p-value<0. 0001). Therefore, we investigated the effect of processing without the influence of HLA binding by comparing the eluted peptide set to an affinity-matched subset of predicted binders to ensure that eluted and predicted peptides show the same distribution of binding affinities (Fig. S1A). Also for this subset of predicted binders we observed a significantly lower C-terminal processing probability (Fig. S1B). NetChop was trained on the C-termini of known HLA ligands and therefore predicts the combined effect of proteasome and TAP. Investigating the impact of these two processes separately (by employing prediction methods that are part of mhc-pathway [19], [20]) suggested that the C-termini of eluted peptides are more likely to be produced by the immunoproteasome and that these peptides are more efficiently transported by TAP (Fig. S2). For non-self peptides, we observed the same characteristics. Eluted peptides that originated from vaccinia proteins showed a significantly higher binding affinity to the respective HLA allotype than other potential binders derived from the same set of proteins (Fig. S3A). A difference in predicted C-terminal processing probability between eluted and other peptides was, however, only found for A*02: 01-binding vaccinia peptides (Fig. S3B). Interestingly, we did not observe a difference in predicted HLA binding affinity between the eluted peptides that originate from human proteins and vaccinia proteins (results not shown). This is in line with an earlier study, which showed that the HLA alleles analyzed here do not show a preference for presentation of non-self over self peptides, while others, foremost HLA-A alleles, do [21], [22]. After having investigated what factors determine which peptides of a given protein are chosen to be presented, we turned to investigate which features of a protein impact protein sampling itself. In other words, why are some proteins sampled while others are not? Previous studies have shown that proteins giving rise to HLA ligands are foremost intracellular, distributed over various intracellular compartments with a slight bias towards the cytosol [8], [23]. Predicting the subcellular localization of each of the sampled proteins in the two data sets of our study, we found similar results: Overall, the distribution of cellular compartments for the sets of source proteins significantly differed from the distribution for the complete human proteome (Fig. S4), and specifically, extracellular proteins were significantly underrepresented in both elution data sets (p<1e-09, Chi-squared test), while proteins resident in the cytosol were overrepresented (p<2e-05). In addition, we tested several protein characteristics for their ability to discriminate source proteins from proteins that were not sampled by the antigen processing pathway. For all three HLA allotypes studied, source proteins are longer, more abundant, and the corresponding genes are more highly expressed (Fig. 3). These factors, however, are not independent. As expected, gene expression level and protein abundance are moderately correlated (Spearman' s rho = 0. 3, p-value<2e-16). Additionally, we noticed that protein length and abundance are inversely correlated to each other, with shorter proteins being more abundant (rho = −0. 41, p-value<2e-16). Since we found that sampled proteins were longer but at the same time more abundant, correcting for protein length (by choosing a random subset of non-sampled proteins with the same length distribution as the set of sampled proteins) enhanced the difference in protein abundance even (Fig. S5). In addition, proteins that were sampled in both elution studies (n = 160) were found to be more abundant than source proteins that emerged only in one of the data sets (median abundance = 17. 54 ppm (parts per million, see Materials and Methods) compared to 3. 15 ppm, p = 1e-10). Moreover, we found a significantly higher rate of predicted binders (in the following referred to as the “predicted hit rate”) in sampled proteins, most pronounced for A*02: 01-specific source proteins (median hit rate = 0. 3 for sampled proteins vs. 0. 025 for non-sampled proteins, p = 7e-14). Interestingly, within the same cell line, proteins that were sampled only by B*15: 01 show a significantly lower predicted hit rate for A*02: 01 than proteins that have been sampled by A*02: 01 (median hit rate = 0. 030 vs. 0. 026, p = 7e-13), further emphasizing that the relative number of potential binding peptides does have an influence on sampling probability. We do not have abundance data for vaccinia proteins available, but Assarsson et al. [24] measured vaccinia gene expression at several time points after infection. For each time point, we found a positive correlation between the gene expression level and the sampling frequency of the proteins when comparing single-sampled and multiple-sampled source proteins with those that did not give rise to eluted peptides (Fig. S6). Additionally, as for human proteins, sampled vaccinia proteins are longer than the remaining vaccinia proteins (p<0. 01, data not shown). Overall, all tested factors - protein length, gene expression level, protein abundance, and predicted hit rate - show differences between the set of sampled proteins and the proteins that were not sampled. In order to quantify the contribution of each factor and to determine which combination of factors best describes the data, we performed a multiple logistic regression analysis. Starting from a maximal model including all factors as explanatory variables, we obtained a minimal model by iterative exclusion of non-significant factors. Before running the regression, we first randomly picked a subset of non-sampled proteins to form a “negative” set of equal size as the positive set of sampled proteins. This balanced set of negative and positive data points was then used to perform a logistic regression and performance analysis (see Materials and Methods), which was repeated 100 times with different random negative subsets. The performance was measured as the Spearman correlation coefficient between the known sampling status (i. e. , a binary value) and the predicted sampling probability. For all three HLA allotypes, a regression model combining protein abundance, protein length and predicted hit rate showed the best performance (the best examples are given in Fig. 4A–C). Since we found that eluted peptides are more efficiently processed than other HLA-binding peptides, we tested whether we could improve the model by filtering the set of predicted binders for processing efficiency. For all three HLA allotypes, this filtering step improved the prediction performance only to a minor extent (results not shown). As gene expression and protein abundance are moderately correlated, we tested which of these two factors would carry more information for predicting protein sampling. We found that protein abundance clearly outcompetes gene expression (Fig. 5). Among the three HLA allotypes, the prediction performance of the B*27: 05 model was best (Fig. 4C), with an average AUC (area under the receiver operating characteristic curve [25]) value of 0. 74 compared to 0. 70 for A*02: 01 and 0. 68 for B*15: 01 (Fig. 4D). Overall, the resulting logistic regression models were able to discriminate between sampled and non-sampled proteins to a significant degree (Fig. 4A–C).
Only a small fraction of the human proteome is sampled by the class I antigen processing pathway and presented on HLA class I molecules. Previously it was suggested that the cellular localization of a protein and its function play a role in this sampling process [8], [23]. Here we show that other protein characteristics like protein length, abundance, and rate of predicted binders also largely influence the sampling probability of a protein and thereby shape the peptide repertoire of an HLA class I molecule. We analyzed two large peptide elution datasets; one derived from a vaccinia virus infected cell line, and one obtained from cells transfected with a gene encoding a soluble HLA class I molecule. Identification of the source protein of each peptide showed that, in spite of the huge difference in proteome size between human and vaccinia virus, a similar fraction of either proteome (10–12% of all proteins) was sampled by the antigen processing pathway. We characterized the set of eluted peptides in terms of antigen processing and presentation and observed a significantly higher binding affinity to the respective HLA molecule and more efficient processing for the eluted peptides than for other potential binders derived from the same set of source proteins. The predicted median affinity of eluted peptides was 14 nM IC50 for A*02: 01,65 nM IC50 for B*15: 01, and 107 nM IC50 for B*27: 05. These values are much lower than the 500 nM, which are often used as a threshold to discriminate HLA-binding peptides from non-binders [26]. This observation could reflect that high-affinity binders are preferentially loaded onto the HLA molecule among others with the help of the ER resident chaperone tapasin [27], [28], or that they have a longer “life span” on the cell surface because they form a more stable complex with the HLA molecule, which increases their chance of being eluted (even though this is rather related to the off-rate of a peptide than to its affinity). Especially for the elution studies involving soluble HLA molecules, it is not surprising to identify foremost high-affinity binders after the long affinity purification process [14]. A higher binding affinity of eluted peptides has also been found for mouse MHC class I molecules [12]. If high-affinity peptides are able to outcompete lower affinity-peptides in binding to the HLA, this may result in a higher copy number of these peptides on the cell surface which in turn increases their chance to be detected by mass spectrometry. The nature of the data sets we analyzed does not allow us to study this because we do not have abundance data on the peptides. Instead we merely know whether a peptide was present in the eluate or not. To our surprise, most proteins were represented by only a single peptide in the elution data sets we studied. This is in line with the observation by Hickman et al. [8] who found only 9 of 189 source proteins (4. 8%) to be represented by more than one peptide. For the prediction of C-terminal processing, we employed a method that has been trained on the C-termini of known HLA ligands [17]. Initially intended as a predictor of proteasomal cleavage, the method automatically accounts for the contribution of other potential peptidases that are able to further process the carboxy terminus of proteasome products, as for example the carboxypeptidase ACE [6]. It does, however, not account for the activity of aminopeptidases in the cytosol and ER, which may further trim the amino termini of peptides. There is some evidence for the existence of N-terminal processing motives, which differ in specificity between cytosol and ER [29], [30]. However, the lack of appropriate prediction methods prevented us from assessing the effect of N-terminal trimming of peptides in our analysis. For the data sets we studied, we observed that (i) source proteins are longer and more abundant than non-source proteins, (ii) the corresponding genes show higher expression levels, and (iii) source proteins show a higher rate of predicted binders than proteins that were not sampled. We combined these factors in a logistic regression model and conclude that prediction of protein sampling probability is possible to some degree. The best model made use of protein length, abundance, and predicted hit rate to predict the sampling probability of a protein. Fortier et al. [12] observed that MHC-presented peptides are preferentially derived from highly abundant mRNAs. Our analysis confirmed the impact of gene expression reported earlier by Fortier et al. , but in addition, our results suggest that protein abundance carries more information for the prediction of protein sampling than transcript levels do. It has been argued that antigen processing should be correlated with protein turnover rather than cellular abundance of proteins [10]. In addition, a recent study suggested that the pioneer round of mRNA translation, which serves as a “proof-reading” step during mRNA maturation, might be a major source of HLA ligands [31]. We believe that the model presented in this paper will improve considerably when more data is available describing the specificity and kinetics of peptide generation via these processes. Finally, another source of antigenic peptides are so-called defective ribosomal products (DRiPs), which are truncated and/or misfolded polypeptides that are directly targeted to proteasomal degradation [32], [33]. The DRiP hypothesis suggests that the set of MHC-presented peptides reflects recent protein synthesis rather than the protein content of the cell, which should manifest itself in our analysis as a higher correlation with gene expression level than with protein abundance. Even though this is not what we see, the fact that HLA ligands are preferentially derived from long proteins is in accordance with the DRiP hypothesis, because the chance of incorporating errors and of misfolding increases with protein length. A limitation of our analysis is the presumably high noise in the protein abundance and gene expression data. The abundance data was derived through meta-analysis from a multitude of different tissue types, even though there is considerable variation of protein abundance between cells and tissues. However, as Weiss et al. [34] report, the abundance data set consists mainly of house-keeping genes whose tissue-to-tissue expression variability is limited. Ideally, the analysis presented here should be repeated on a data set where mRNA levels, protein abundance, and HLA peptide presentation are measured simultaneously for a single cell type or tissue sample to minimize noise. All the more striking it is, however, that we see a clear signal for both gene expression and protein abundance in their impact on protein sampling in spite of the noise introduced by averaging over cell types. In conclusion, the results presented in this paper demonstrate that protein characteristics such as gene expression level, protein abundance, and the rate of HLA ligands determine which protein will be sampled for antigen presentation. Moreover, our results suggest that sampling prediction may be extended to the proteomes of pathogens, allowing us to identify promising targets for vaccination studies.
Johnson et al. [13] performed peptide elution and mass-spectrometry analysis of vaccinia virus infected Epstein-Barr virus-transformed B-cells, homozygous for HLA-A*02: 01, B*15: 01, and C*03 (for details see [13]). With a false positive rate (FPR) of 5%, they identified 4717 unique human-derived peptides and 119 vaccinia derived peptides. Ben Dror et al. [14] eluted peptides from cultured cartilage cells and HeLa cells transfected with a soluble form of HLA-B*27: 05. Based on several criteria to assess the confidence in identified peptides, they categorized eluted peptides into three subsets: certain (569 peptides), probable (582 peptides), and possible (116 peptides). As the certain peptide set corresponds to a FPR of 4. 7%, we limited our analysis to this data set. Of note, in the original publication, peptides were selected as correct only if they contained the amino acids arginine or glutamine at their second position [14], which according to the authors (personal communication) was necessary in order to filter out peptides that were eluted from other HLA allotypes expressed by the cell line (which may become soluble due to cellular stress). We obtained the human proteome from Ensembl Genomes (ftp. ensembl. org/pub/, release 56) and used this collection of proteins to identify the source protein for each peptide in our elution data sets. Source protein identification required identical mapping of a peptide to the source protein sequence. Peptides that could not be uniquely mapped to one single protein were omitted from further analysis. In the case of several splice variants of the same protein (i. e. , the peptide matched to several protein sequences which all originate from the same gene), the longest splice variant was chosen for sequence analysis. Likewise, the longest splice variant was chosen for the set of non-sampled proteins. We were able to map 105 of the vaccinia peptides to the Vaccinia Western Reserve proteome (GenBank identifier: AY243312). We had abundance data available for 12,021 human proteins [34]. The abundance is expressed in parts per million (ppm), relative to the molecule counts of all other proteins in the proteome. The measured abundance for different proteins spans several orders of magnitude. The protein abundance data covers 1986 (78. 4%) of the 2533 Johnson source proteins and 340 (85. 8%) of the 396 Ben Dror source proteins. We used gene expression data from Juncker et al. [23], who provide the median of normalized mRNA levels of haematopoietic tissues originally obtained from the GNF gene expression database [35]. Expression levels of vaccinia virus genes were obtained from Assarsson et al. [24]. Throughout the study we used the method NetMHC 3. 2 [15], [16] to predict peptide-binding to the HLA molecules A*02: 01, B*15: 01, and B*27: 05. Binding predictions for C*03 were done using NetMHCpan 2. 4 [36]. We define predicted binders as peptides that have a predicted binding affinity of <500 nM IC50 for a particular HLA molecule. For the Ben Dror data, where all identified peptides were eluted from a known HLA molecule (namely, HLA-B*27: 05), this proved to be a suitable threshold, predicting 89. 6% (510 of 569) of the eluted peptides as binders. The NetChop version Cterm-3. 0 [17], [37] was used for the prediction of C-terminal processing. Furthermore, we used weight matrices provided by the mhc-pathway package [19], [20] for the prediction of cleavage probability by the immunoproteasome and for TAP transport efficiency. We employed WoLF PSORT [38] to predict subcellular localization of proteins and confirmed our results by GO-term enrichment analysis using the Cytoscape [39] plug-in Bingo [40]. The predicted hit rate for a given protein is defined as the ratio of the number of predicted binders for a particular HLA allotype to the total number of unique 9mer peptides in this protein. Multiple occurrences of the same peptide within one protein were counted as a single occurrence, because they would also not be detected as separate peptides in the elution analysis. The hit rate is calculated per HLA allele and hit rates may differ between alleles, because we use a fixed affinity threshold of 500 nM IC50 to define binders (instead of assigning a fixed fraction of peptides, e. g. top 1%, as binders). For this reason we did not directly compare hit rates between alleles, but instead performed separate analyses per HLA allele. Two-sided Mann-Whitney tests, correlation tests, Chi-squared tests, and logistic regression analysis were performed using R (http: //www. R-project. org). We used a generalized linear model with a binomial response distribution and a logit function for data transformation to model the impact of various factors on sampling probability. All figures were produced using R. | HLA class I molecules are expressed on the cell surface of almost all cells of the human body in complex with short fragments (peptides) of cytosolic proteins, thereby providing a snapshot of the intracellular state of a cell to circulating CD8+ T cells. Several processes are involved in shaping the peptide ligand repertoire of an HLA class I molecule, which generally represents only a small fraction of the proteins available in the cytosol. In our work we addressed protein sampling by HLA class I molecules to answer two questions: 1) Which proteins are sampled by the antigen processing pathway and why, and 2) which peptides of a given protein are picked to represent the source protein on the cell surface? To this end we quantified the contribution of each process involved in peptide processing and presentation individually and combined them into a logistic regression model. This simple model enabled us to predict the sampling probability of self proteins and may aid in the identification of autoimmune antigens. | Abstract
Introduction
Results
Discussion
Materials and Methods | sequence analysis
major histocompatibility complex
immunology
biology
computational biology | 2012 | Proteome Sampling by the HLA Class I Antigen Processing Pathway | 7,279 | 217 |
A statistical thermodynamics approach is proposed to determine structurally and functionally important residues in native proteins that are involved in energy exchange with a ligand and other residues along an interaction pathway. The structure-function relationships, ligand binding and allosteric activities of ten structures of HLA Class I proteins of the immune system are studied by the Gaussian Network Model. Five of these models are associated with inflammatory rheumatic disease and the remaining five are properly functioning. In the Gaussian Network Model, the protein structures are modeled as an elastic network where the inter-residue interactions are harmonic. Important residues and the interaction pathways in the proteins are identified by focusing on the largest eigenvalue of the residue interaction matrix. Predicted important residues match those known from previous experimental and clinical work. Graph perturbation is used to determine the response of the important residues along the interaction pathway. Differences in response patterns of the two sets of proteins are identified and their relations to disease are discussed.
Transfer of information between two points in a protein is a fundamental problem relating to function [1], [2]. Fluctuations of residues in the native protein are the essential determinants of information transfer. The three dimensional native conformation or the topology of a protein determines the fluctuations of its residues. Relationships between topology and fluctuations offer important clues for the function of the protein. Structure-function relations can conveniently be understood by treating the protein as a graph of interacting residues [1]. Significant progress has been made in this direction over the past decade. Residue fluctuations, correlations, locations of conserved and important residues, stability of the native state, information transfer, energy fluctuations, and recently the protein-protein and protein-ligand binding have been studied by recourse to the graph-like state of the native topology [3]–[14]. The residue interaction graph contains important information in this respect that allows the determination of important interactions in a protein. The criticality of important interactions in a complex system is often the determining factor of stability of graphs [1], [15]. The lack of rearrangements in over determinate and highly constrained graphs result in decreased stability and robustness [15]. Our work here is centered on the determination of the subset of important interactions in proteins and the relationships between this set and function. We apply our analysis to a set of ten HLA class I proteins, HLA-B27, which are relevant examples of the relationship between critical interactions, robustness, and function. The specific purpose of the present paper is to present a statistical thermodynamics model that gives a consistent explanation of structure-fluctuation-function relations in terms of the graph-like features of native proteins. We use the widely adopted Gaussian Network Model (GNM) based on a harmonic potential of residue-residue interactions, and propose a model for determining structurally and functionally important residues in relation to ligand-protein interactions as well as the path that the protein uses in transferring information form one point to the other. Our treatment is essentially an extension of the three recent papers [9]–[11] which we briefly summarize in the method section below in order to reduce cross-referencing. In the cited papers we showed, using statistical thermodynamics arguments, that the mode corresponding to the largest eigenvalue of the connectivity graph obtained from the contact map indicates the structurally and functionally important residues and that these residues are the ones for which energy and residue fluctuations are strongly correlated. We show that a few residues belong to the set of energetically active residues that are at the surface of the protein and are most efficient in energy exchange with the surroundings. We call these the ‘energy gates’. We also show that the residues that connect any such two surface residues along an interaction path are the ‘hub residues’ over which information is transmitted. From statistical mechanical arguments, a surface residue that is efficient in energy exchange with the surroundings is expected to be active in binding of a ligand, as the ligand-binding problem is an energy exchange problem. We also show that changes in the binding/interaction capacity of an energy gate or a hub residue changes the binding/interaction capacity of the other energy gate or hub residues. This has significant consequences relating to allostery and cooperative binding. The harmonic approximation that we adopt here is a coarse graining approach. However, many of the features obtained by this coarse graining are also indicated by more accurate treatments of protein behavior [16]–[19]. The GNM approach allows for a faster and easier visualization of structure-function relations. We study the structure-function, ligand binding and allosteric activities of ten models of HLA-B27 Class I proteins of the immune system. Five of these models, which belong to the HLA-B*2705 allele of the HLA-B27 protein, are known to be strongly associated with a tendency to develop a chronic inflammatory rheumatic disease, known as ankylosing spondylitis, by causing yet unknown functional abnormalities. The remaining five are chosen from the HLA-B*2709 allele of the same protein. These are the corresponding properly functioning ones with almost no susceptibility for ankylosing spondylitis [20]–[31]. Each pair of the protein structures, one from the HLA-B*2705 and the other from the HLA-B*2709 allele, contains the same peptide in their antigen binding groove to present to immune cells, and therefore serves as an excellent benchmark to test the predictions of the GNM. The only difference between the B*2705 and B*2709 alleles is that residue 116 in the former is always an ASP, whereas it is HIS in the latter. This single residue difference between the two alleles causes structural differences in the two types, and therefore in their contact maps. We show that these differences in the contact map of the two types lead to significant and consistent changes in the fluctuation profile, making the members of the HLA-B*2705 allele respond too strongly to perturbation. Based on these changes, we propose a mechanism that is responsible in the functional differences of the two types.
The system consists of the protein and its environment. The latter may contain ligands that are capable of binding to the protein. The protein and the environment form a closed system with fixed energy and amount of molecules. The protein exchanges energy with the environment. Since the total energy of the protein and the surroundings is constant, we have (1) (2) where, Uprot and Usurr are the energies of the protein and the surroundings, respectively. In the statistical thermodynamics treatment of proteins that we propose here, the thermodynamic variables for the protein are S = Entropy, U = energy, V = Volume of protein, R = Position of the residues. In the remainder of the paper, the thermodynamic variables are used for the protein only, without the subscript prot. The thermodynamic variables are averages. The instantaneous values of the energy, volume and residue positions are shown by, , , respectively. The fluctuations, , , result from the deviations of the instantaneous extensive variables from their thermodynamic averages. In the GNM model, the emphasis has been on the fluctuations, visualized as resulting from coupled harmonic motions of the residues from their mean positions [4]. The present treatment is based on the extension of the mechanistic description of the GNM to include the role of energy fluctuations, , as well. As in previous treatments, we adopt a coarse-grained model and represent each residue in terms of its alpha carbon. Thus, for a protein of n residues, is defined as (3) Here, represents the Cartesian coordinates of the ith residue alpha carbon. and are similarly defined. The probability of the instantaneous values, , , and, of the energy, volume and residue positions, is determined from the interrelation of the thermodynamic functions given in the Text S1. Therein, this probability function is used to derive the correlations between the fluctuations of residue positions and energy, as well as the cross correlations between the fluctuations of the energy and residue positions. The statistical thermodynamics interpretation of the GNM was given in full detail by Yogurtcu et al. , [11], which was successfully applied to the prediction of binding sites in receptor-ligand complexes [10], of specific sites for binding [9]. In the present paper, we use the statistical thermodynamics approach to predict the important residues along an interaction pathway. The starting point of the model is the equation relating fluctuations to thermodynamic averages. The derivation of this equation is given in the Text S1. We reproduce the resulting expression here to reduce cross-referencing (4) Here, represents the position vector of the alpha carbon of the ith residue and the superscript T indicates the transpose. k is the Boltzmann constant, T is the temperature. is the force on the jth alpha carbon. The subscripts of the parenthesis of the right hand side indicate that the temperature, pressure and the force on each residue except the ith is kept constant. Angular brackets indicate an average over all possible values of the argument. The right hand side is a thermodynamic quantity that expresses the change in the position of residues by the application of a force. The left hand side, on the other hand denotes an average of fluctuations. Thus, this equation relates fluctuations to average quantities. If fluctuations are associated with function, as is done in several previous studies [3], [6], [32], [33]. The right hand side of Eq. 4 requires the knowledge of a force-displacement relation. The simplest of such relations is that for the linear spring (5) where, Γ is the spring constant matrix. Multiplying both sides of Eq. 5 with the inverse of Γ and performing the differentiation shown in Eq. 4 leads to (6) In the GNM model, the Γ matrix is obtained by inserting a constant to the ij' th position if residues i and j are in contact and zero otherwise. Two residues are assumed to be in contact if they are separated by less than 7 Å. This value of the cutoff is approximately equal to the radius of the first coordination shell for residues in a protein. Each diagonal element of Γ is the negative sum of its row. In this way the protein is visualized as a graph, and the off-diagonal terms of Γ is the connectivity graph of the protein. Although the ij' th elements of the Γ matrix is set to zero when residues i and j are separated by more than 7Å, the ij' th element of the inverse matrix Γ−1 is not zero. This implies correlations between residues that are not in contact. The correlation between the fluctuations of the ith and jth residues is determined by the full graph structure, since the ij' th element of the inverse of Γ contains contributions from all nodes of the graph, i. e. , from all other residues, and not just from those in the close neighborhood of the ith and jth residues. As will be discussed in detail below, the essential features of the correlations may be understood largely by considering the largest eigenvalue of Γ. Energy fluctuations of the protein are assumed to result from fluctuations of inter-residue interactions. The correlation of the energy fluctuations of the protein with fluctuations of residue positions is derived as (See Text S1) (7) The left hand side gives the correlations of energy fluctuations with the fluctuations of residues. The right hand side consists of correlations among residue fluctuations only. Writing the difference (8) and using the right hand side of Eq. 7 with the appropriate choice of the indices, we can write (9) For the case of harmonic fluctuations, i. e. , GNM, this relation is derived in the Text S1. If is assumed to represent the mean-square fluctuation in the ‘spring length’ connecting residues i and j, the left hand side of Eq. 4 becomes proportional to the fraction of the incoming energy from the surroundings absorbed by the spring. The right hand side of Eq. 9 is positive. The terms in the angular brackets on the left hand side may be positive or negative depending on the sign of ΔU. But, for the average to be positive, there must be a constraint on the elements of the left hand side: Positive fluctuations of the energy, which indicates energy transfer into the protein from its surroundings, should couple to large values of and negative fluctuations should couple to small values. Stated in another way, energy that is absorbed from the surroundings are stored in pairwise interactions between residues i and j according to Eq. 8. In the application of the model to the HLA proteins, we define the variable as the sum of the ith row of the correlation matrix (10) A finite value of indicates that residue i belongs to the subset of energetically active residues that are either energy gates or lie along an interaction pathway. It also is a measure of the energy absorbed from the surroundings as may be seen from the second equality in Eq 10. According to graph theory, important features of the graph, such as graph perturbation that relates to allostery for example, may be obtained by considering the largest eigenvalue and eigenvector of the graph [34]. The choice of the largest eigenvalue mode is specifically relevant, because (i) it corresponds to localized effects where only a few residues are excited [3] and (ii) the largest eigenvalue is the most sensitive to perturbation of the graph [34]. In the case that the residues identified by the highest mode are adjacent in space, then they interact and form a path that is active in long distance communication. Our calculations for a large number of ligand-protein systems show that the largest eigenvalue and the corresponding mode of the matrix is in general sufficient to point to the known functionally relevant residues. Within the present approximation, we adopt the maximum eigenvalue interpretation. A residue at the surface with a large value of is an ‘energy gate’ through which the protein executes its energy interactions with the surroundings. If the residue with high is not at the surface but inside the protein, then it is a ‘hub residue’ that has important function along the interaction pathway connecting to an energy gate. Although there is no proof, hubs are generally located between two energy gates in allosteric processes [35], [36]. Examples shown below are in support of this statement. For a residue i at the surface, is a measure of whether residue i will interact with the ligand. For the hub residues, is a measure of the importance of that residue within the network of information exchange. The quantity introduced above indicates the extent of correlation of the given residue i with the rest of the protein. Any change in the connectivity state of residue i will affect the behavior of the rest of the protein through the subset of energetically active residues. One way to apply this change would be to bind a ligand to i, and to the residues within the cutoff distance of i. This corresponds to perturbing the entries in the ith row and column of the Γ matrix. The relation of this to allosteric manipulation is obvious. In this section we discuss the changes in the interaction energies of residues j when the parameters of residue i are modified. Binding to a point i on the protein may increase or decrease the residue interaction energy of other points. The calculations and the analysis are carried out with respect to the heavy chain A, taking the structure from the complex structure of chains A, B (beta-2-microglobulin) and C (peptide bound on the antigen binding groove of chain A) in the alleles HLA-B*2705 and HLA-B*2709 alleles.
Ten HLA-B27 protein structures are analyzed here. Five of the structures belong to the HLA-B*2705 allele and the remaining five belong to the HLA-B*2709 allele, where the residue 116 is ASP in the former and HIS in the latter. Each pair has the same peptide sequence bound. The PDB codes of the proteins and their alleles are presented in the first and third columns of Table 1. In Figure 1A, the ribbon diagram of the heavy chain of 1OF2. PDB of the HLA-B*2709 allele is shown. The nine residue peptide is shown in indigo. It sits in the groove between the two helices shown in red. In the same figure, we show the positions of the energetically active and functionally important residues 6,7, 27,101 and 164 for 1OF2 in yellow, suggested by the GNM calculations. Figure 1B is an enlarged version of Figure 1A where each residue is shown with a different color and the rest of the protein is not shown. Figure 1B clearly shows the important residues that form an interaction pathway, with GLU161 at one end and TYR27 on the other end of this path. The prediction of these residues, and the detailed discussion of their role in the functioning and malfunctioning of the HLA-B*2709 and HLA-B*2705 will be given below. In all of the calculations presented in this paper, we concentrate only on the largest eigenvalue which suffices for presenting a proof-of-principle discussion. A more detailed discussion may need to involve eigenvalues other than the largest, which might be plausible for additional functionally and structurally residues. We use Eq. 7 (see Methods) to obtain the energetically active residues using the high frequency mode. In Figure 2, we show the plots of the two proteins 1OF2 and 1OGT, where the subscript i indicates the residue index. The plots are obtained as follows: First, the Γ matrix is constructed with a cutoff distance of 7 Å and the correlations are calculated using Eq. 3. The components of the correlation matrix corresponding to the largest eigenvalues of the Γ matrix are determined by reconstructing the correlation matrix keeping the largest eigenvalue only, and the columns of the resulting matrix are added according to Eq. 10 in order to obtain the values presented in the figures. Figure 2 displays that residue CYS101plays the most significant role in the interactions of the protein and has the strongest correlations with other residues such as ARG6, TYR27, LEU160 and CYS164 in both 1OF2 and 1OGT. This is also observed in the other HLA-B*2709 and HLA-B*2709 alleles. The residues with high values that are observed in the ten proteins are shown in Table 1. The other feature observed that is in common for all of the ten proteins is that the value for the 101st residue is always larger for the HLA-B*2709 alleles than for HLA-B*2705. Since these residues are calculated from the largest eigenvalue of the Γ matrix, we call them the important residues and show with the examples below that these play role in the stability and function of the protein. From the definition given by Eq. 10, is the sum of the distance fluctuations of the intermolecular bonds which the ith residue makes with others. Equation 4 shows that reflects the energetic interactions of residue i with other residues. In this sense, CYS101 acts as the central hub, which controls the system. There are two different types of terms on the right hand side of Eq. 8, the self terms, and the cross term. In order for to be large, both and should be large, and should be negative, i. e. , residues i and j should make anti-correlated motions. Only in this case can be large and energy can be transferred from one to the other via the spring that connects them. Figure 2 shows that there are essentially four groups of residues that are of significance: (i) residue 6, (ii) residue 27, (iii) residues 101–116, and (iv) residues 160–164. In the remaining sections, we elaborate on the characteristic features of these four groups of residues that are also observed in the other HLA-B*2709 and HLA-B*2709 alleles (Table 1). In this section, we study the differences of the response of the residues for a perturbation along the interaction pathway between the two families. These differences arise from the presence of the negatively charged ASP116 in HLA-B*2705 and positively charged HIS116 in HLA-B*2709 that induce energetic changes along the interaction pathway, resulting in functional differences. The differences will be outlined in the following sections. In this section, we present the results of our calculations based on Eq. 10. In the interest of observing the response of a protein to an external stimulus, we induce changes in the interaction strength of each important residue and observe the response of the remaining residues. This is done by increasing the interacting strength of contacts of the ith residue by 1%. This amounts to multiplying the off diagonal elements of the ith row and column of the Γ matrix by 1. 01, and recording the difference in the values of obtained after and before this perturbation for each residue j. In this notation Δ is the change in, and the subscript i indicates that the perturbation is applied on the ith residue. A perturbation of 1% was chosen to ensure that the system was in the linear response region. Trial choice of values above 10% resulted in nonlinear response. In the linear response regime, changing the perturbation from 1% to 2%, for example, doubled the output. Figures 3A–F, given for the case of 1OF2-1OGT, show that the B*2705' s respond to perturbations strongly compared to B*2709' s. The other B*2705' s and B*2709' s, also show the same difference. This difference has its roots in the differences of residue-residue interaction energies. The residue ASP116 in B*2705' s results in strong interactions with its surrounding residues, making the protein respond strongly to perturbations. An examination of Figure 3 shows that positive perturbation of ARG6 induces a decrease in the response of CYS101 which is stronger for all of the B*2705 alleles than for B*2709. It is worth noting that the residue that is directly involved in the binding of the ligand is TYR7. However, its perturbation does not result in any noticeable perturbation in the rest of the protein, suggesting that it does not directly lie on the interaction pathway. However, perturbation of its neighbors MET5 and ARG6 induces strong changes in the behavior of the protein. This is because in the native structure, the environment of TYR7 is less compact than that of ARG6. In this respect, ARG6 plays a special role in the pathway we identified. For example, mutating ARG6 into ALA6 in 1OGT (see the following section describing energy calculations) caused four times more energy increase then mutating TYR7 into ALA7. Based on this evidence we hypothesize that the direct interaction of the ligand with TYR7 induces a perturbation of ARG6 which affects the protein structure significantly. In order to see the differences in response of the two alleles, we subtracted the values of B*2709 from those of B*2705 for ARG6 for each allele pair, and presented the results in Figure 4. The figures show that although ARG6 is perturbed positively by the same amount for both alleles, i. e. , the related elements of the Γ matrix are perturbed by 1% in both cases, the response of B*2705' s is stronger at ARG6 and the CYS101 response to this perturbation is always negative, and stronger again in all B*2705' s. Energy calculations show that the interactions of ARG6 with its environment is 9 kcal/mole stronger in B*2705 than in B*2709, which means that ARG6 is more rigidly embedded in its surroundings in B*2705. This is the result of the differences of residue 116 that affect the two alleles in different ways and the effects are seen on residues ARG6 and CYS101. The only difference between the sequence of B*2705' s and B*2709' s is in the residue 116. This residue is located at the bottom of the B-pocket where the peptide binds. This single mutation is thought to cause differences in the stiffness of the structure around 116 that result in the differences observed and reported above for the two alleles [21]. In order to understand the energetic differences of the two alleles, we calculated the interaction energy of residue 116 with its surroundings in the two alleles. The knowledge of the interaction energies of specific residues in the system may be helpful for understanding and comparing the behavior of the two alleles under study. Here, we present approximate calculations of interaction energies obtained by static minimization of the energies, briefly described in Text S1. The energy minimization calculations are only for comparison of the B*1705 and B*2709, where we either compare two different systems, or compare two different situations on the same system. Thus, the relative values rather than the absolute values of the energies reported here are of interest here to have an estimate on the differences between the B*1705 and B*2709. In our calculations, we first minimized the energy of the system. To calculate the interaction of a given residue with the rest of the protein at its minimum energy conformation, the residue is chosen in the matrix of the remaining residues that are kept in their native states. The interaction energy of the chosen residue is then minimized around the given conformation. In this calculation, only the residue of interest is left flexible and the conformations of the remaining protein residues are kept fixed at their native values. As the energies are sensitively dependent on the value of the dielectric constant chosen, different values of the dielectric constant are used to see the effect on the calculated energies. In the absence of explicit water, the value range of 1–4 is usually considered in the calculations of biological systems. We minimized the energy of 1OGT. PDB and 1OF2. PDB and calculated the energy of residue 116 in each structure as described in the preceding paragraph. We then removed the rest of the protein and minimized the energy of the isolated residue 116. The energy calculated in this way is the intra-residue energy for 116 and contains bond, bond angle, electrostatic, hydrogen bonded and nonbonded energies of the atoms that all belong to 116 only. The difference between the energy in the presence and absence of the surroundings gives an idea on how strongly the residue interacts with its neighbors in the protein. For the dielectric constant equal to unity, the energy of residue 116 is 156 kcal/mol lower in 1OGT (B*2705) because of the charge differences of the two residues, where HIS is positively charged with a pK of 6. 5 while ASP is negatively charged with a pK of 3. 1, which is the ‘random coil’ or ‘model compound’ small peptide pKa value. In 1OGT, the carbonyl group of the negatively charged ASP116 is within 2. 9 Å of the positively charged amino end of LYS70, and within 5. 5 Å of the positively charged HIS114, whereas in 1OF2 (B*2709), the positively charged HIS116 is 6 Å to the nearest negatively charged ASP122. In 1OGT, the two residues of the peptide binding site, ARG6 and ASN97 have lower energies in 1OGT compared to 1OF2. This means, these two residues are embedded strongly in their environments in 1OGT, and perturbing their states results in strong responses in the protein. The unrealistically high energy values reported here are upper bounds that are obtained by taking the dielectric constant as unity. We calculated the binding energies by varying the dielectric constant over a wide range. The results are presented in Figure 5. The difference between the two proteins vanishes when the dielectric constant is around 20. The realistic values of ε used for biological systems vary in the range 1–4. Even with a value of ε = 4, the energy difference is as high as 30 kcal/mole. In Table 2, we present the differences in these energies for 1OGT (B*2705) and 1OF2 (B*2709) for residue 116 and for a few other residues with the values calculated for ε = 1. These values are the upper bounds. We see that GLU163 in 1OGT is bound to its neighborhood less strongly than the one in 1OF2 by an energy difference of 61. 0 kcal/mole. This difference comes from the presence of the negatively charged LYS3 of the peptide in close vicinity of GLU163 in 1OF2. In the energy minimized structure, the oxygen of the carbonyl group of GLU163 is 2. 3 Å from the hydrogen of the amino group of LYS3 of the peptide, whereas this distance is 5. 6 Å in 1OGT. This interaction indicates the specificity of binding of the peptide to 1OF2, which is lacking in 1OGT. In Table 3, differences in the interaction energy of the peptides of the two alleles and the energy of the residue GLU163 are presented. The approximations involved in these calculations are explained in the Text S1. In the second column, we compare the binding energy of the full peptide to the proteins. The values are the differences of the binding energies to alleles B*2705 and B*2709. Among the different residues, GLU163 exhibits a peculiar difference in that its energy is much lower in B*2709' s. For this reason we calculated the difference in the energy of GLU163 for the two alleles and presented the results in the third column of Table 3. Although the overall binding energy of the peptide to B*2705 is more favorable in all cases, the energy of GLU163 is more favorably in B*2709. It is worth noting that the computations by using the largest eigenvalue approach indicated LEU160 and CYS164 as the important residues and GLU163 does not appear as an important residue. A similar trend is observed for several other systems not reported here, where the maximum eigenvalue approach points to a close neighbor of an important residue as in the present study. The difference arises mostly from the presence of electrostatic interactions in the neighborhood of the important residue. Predictions by the maximum eigenvalue method do not directly consider the electrostatic interactions. However, their presence affects the topology that is reflected in the maximum eigenvalue method. In the energy calculations described above, in order to see why ARG6 appeared in the interaction path, we mutated ARG6 into ALA6 in 1OGT, which resulted in fourfold increase in the energy of the system (See Text S1). Same calculations are performed by mutating TYR7 into ALA7. These values, though approximate and relative, it consistently points to some important features of the system. Nevertheless, it is worth stating here that the appropriate and rigorous computational practice in computational biology is to perform an extensive molecular dynamics simulation of the protein and the ligand in aqueous medium and extract the required energies as thermodynamic averages, which may include the evaluation of the free energy as well. Our present energy minimization approach is only for exploratory purposes.
We identified the interaction paths of the B*2705 and B*2709 alleles. This path contains the residues TYR27 at one end and CYS164 at the other. Along the path lies CYS101 as the most interactive residue, which we termed as the hub residue. The important residues along this path are shown in Figure 6 for 1OF2, of the B*2709 allele. Differences on this path for the B*2705 alleles are summarized in Figures 7A–D for 1OGT. The roles of the residues shown in Figures 6 and 7 relating to the structural and functional features of this path are discussed in this section. (1) Substitution of ASP116 for HIS116 in the B*2705 allele results in stronger bonds both between the peptide and the protein and between the residues of the protein in the neighborhood of 116. In Figure 6 for B*2709, HIS116 exhibits no interactions with other residues along the path. For B*2705, on the other hand, ASP116 shown for 1OGT in Figure 7A makes three hydrogen bonds with ARG5 of the peptide and one hydrogen bond with ASN97. In other members of B*2705 and B*2709, interactions other than the ones shown in Figures 6 and 7 are also present. In molecular dynamics simulations by Starikov et al. [21], for example, a salt bridge between LEU9 of the peptide and ASP116 in B*2705 was observed to limit the relative motions of these two residues. This causes changes along the interaction path and makes the system B*2705 becomes more fragile against nontrivial rearrangements, in parallel with recent findings on graphs [1]. Apart from differences in the important residues on the pathway predicted by various works the major common finding relates to the effects of replacing HIS116 in B*2709 with ASP116 on B*2705. (2) Our calculations show that ARG6 exhibits strong response to external perturbation in B*2705' s compared to those in B*2709. This implies that ARG6 is more rigidly embedded in its surroundings in B*2705. In Figure 6, ARG6 of 1OF2 is observed to make a single hydrogen bond directly with TYR99. In B*2705, however, ARG6 makes more bonds to its neighbors. In Figure 7B for 1OGT, for example, ARG6 and its neighbor MET6 are hydrogen bonded to CYS101, which in turn is covalently bonded to CYS164. Similarly, in Figure 7C, ARG6 and its neighbor TYR7 are hydrogen bonded to TYR27. In Figure 7D, ARG6 is observed to be an element of a cycle that is a loop of hydrogen bonded elements. This loop contains the residues in clockwise order: ARG6, MET5, CYS101, CYS164, GLU163, bridged by ARG1, ARG2, LYS3 of the peptide, followed by TYR99, TYR 7, terminating with ARG6. All of the residues along this loop are identified by the GNM and the maximum eigenvalue method. Perturbation of B*2705 and B*2709 at ARG6 results in a strong change in the correlations of CYS101 with its neighbors. The response of CYS101 is stronger in B*2705. Comparative energy calculations reported above, show that the interactions of ARG6 with its environment is stronger in B*2705 than in B*2709. (3) Although ASN97 is not identified as a path member by the maximum eigenvalue method, this residue is shown in Figure 6 to make a hydrogen bond with TYR99. Although the latter does not show up as a path member, it makes a short loop of hydrogen bonding with ASN97 which is expected to reinforce the binding pocket. In Figure 7A, ASN97 is observed to make a hydrogen bond with HIS9 and with ARG5 of the peptide, thereby accentuating the tight binding of the peptide in the B*2705 allele. The pocket region that contains HIS9 is referred to as the B-pocket. According to comparative energy calculations reported above, ASN97 shows energetic differences for the two alleles, being more strongly bound to its environment in B*2705 than in B*2709. (4) There is a major difference between the binding modes of the peptides to B*2705 and B*2709. In B*2709, ARG1 and ARG2 of the peptide bind strongly to the B pocket and the rest of the peptide remains relatively flexible contributing to the entropic advantage. The residues of the peptide in the C terminal are subject to nonpolar interactions. These interactions allow for only a few residue types, thus restricting the number of different peptides to only a few. B*2705 on the other hand, is capable of forming bonds with a multitude of residues because of the presence of ASP116. Hence, several different peptides may bind to B*2705. Therefore binding is not specific to a few peptides. Furthermore, the stronger bonding in the B*2705' s presented in Table 3 results in an enhanced entropy penalty. More specifically, for the B*2709 allele shown in Figure 6, ARG1 of the peptide is hydrogen bonded to GLU163, and ARG2 of the peptide is hydrogen bonded to GLU45 and GLU63. For the B*2705 allele, more extensive hydrogen bonding is observed between the peptide and the protein: In Figure 7A, ARG5 of the peptide binds to ASN97 and ASP116; in Figure 7B, ARG1 of the peptide binds to GLU163; in Figure 7C, ARG2 of the peptide binds to THR24; in Figure 7D, ARG2 of the peptide is bonded to TYR7 and LYS3 of the peptide is bonded to TYR99. Peptide flexibility is observed only for the HLA-B*2709 [37]. This suggests an entropic control of peptide recognition. The constraints on the strongly bound peptides in B*2705 constitutes an entropy disadvantage, or an entropy penalty. This hypothesis is supported by thermodynamic data [37]. Figures 7A–D show that the peptide is capable of forming several hydrogen bonds with various residues of the protein. Among these, the interaction with ASP116 and ARG5 of the strong peptide binding capability of the B*2705 binding groove that we observed raises the possibility that B*2705 allele may be capable of binding various different peptides. On the contrary, B*2709 exhibits a limited peptide binding capacity. There is indeed significant amount of experimental work aimed at understanding the differences in binding capacities of the two alleles. B*2709 shows a high specific preference for ligands with nonpolar C-terminal residues. The reason for this is the lack of ASP116. B*2705 accepts other residues at this position [38]–[39]. This is one reason of the ligand specificity of the B*2709' s. (5) In Figure 6, GLU163 is hydrogen bonded to ARG1 of the peptide. Its neighbor CYS164, is also shown in Figure 6. The corresponding conformation for the HLA-B*2705 allele is presented in Figure 7B. In both cases, the crystal structures show that CYS164 is covalently bonded to CYS101. Thus, in both alleles, the gate residue GLU163 can transfer the effects of the peptide to the rest of the protein through the bridge over the CYS164-CYS101 pair. We found that CYS101 is the residue that is strongly correlated with several other residues of the protein. In this sense, we call it the hub residue that controls the function of the protein. The present analysis shows that perturbation at the peptide binding site affects the behavior of CYS101. As shown in Figure 4, this response is stronger in B*2705 when compared with B*2709. A decrease of correlations of CYS101 is expected to result in an important change in the behavior of the protein. Warburton et al. mutated the residue CYS101 in another HLA Class I protein by replacing the CYS with SER, denoted by C101S mutation [40]. Due to the loss of the disulfide bond between CYS101 and CYS164 located between the alpha-helix and beta-sheet portions of the alpha2 domain of the A protein (heavy chain), the proteins lost stability and function. (6) The stronger response of CYS101 to perturbations is a consequence of the strong inter-residue interactions around the binding region. The presence of ASP116 in B*2705 leads to strong inter-residue interaction. On the contrary, the presence of HIS116 in B*2709 makes the protein more flexible due to weaker interactions in the F pocket [21]. This socket is the region that contains the residues 114 and 116 and accommodates the carboxy terminus of the bound peptide. (7) In Figure 6, the CO group of ASN97 is seen to make a hydrogen bond with the backbone NH of TYR99. The importance of this hydrogen bond for the stability of the protein has been shown by Blanco-Gelaz et al. [22] In that work, ASN97 was mutated to ASP97 which prevented the protein from gaining a stable conformation. (8) As a general rule, if a residue is strongly coupled to its environment, then it leads to stronger response when its environment is perturbed, for example, when the residue is replaced with another amino acid of different size. It is therefore expected that when a protein exhibits strong inter-residue interaction energies at a given site, then it is less stable against external perturbations. This parallels the reasoning behind the stability of graphs [1], [8], [41]. (9) In Figure 7, we see that TYR27 makes a hydrogen bond with TYR63 of the beta-2-microglobulin. This is true for both the B*2705 and B*2709 alleles. Interaction of the heavy chain with the light chain is known to be a necessary determinant of stability, and any change in this interaction may be one reason for the misfolded or unfolded protein response [42]. However, although TYR27 appears as a significant residue on the interaction pathway, perturbation of the structure presented in Figures 4 and 5 does not induce a strong response in TYR27. The contribution of TYR27 to the unfolded protein response may not therefore be significant. However, although misfolding is associated with the activity of the peptide, the possible role of the B pocket of the heavy chain in unfolding has not been discarded [43]. The B pocket contains ARG6, TYR7, HIS9, THR24, GLU45, GLU63, and TYR99. The hub residue CYS101 makes two hydrogen bonds, one with MET5 and the other with ARG6, and ARG6 in turn makes two hydrogen bonds with TYR27. There are two different lines of thought or hypothesis from the patogenetic perspective in the association of ankylosing spondylitis disease with HLA-B27 alleles [44]–[50]: Our model shows a difference between the structural stability between B*2705 and B*2709 but also corroborates both hypothesis together by suggesting a role of binding peptides on the stability of structure. Due to strong interactions between the peptide and B*2705, specifically the presence of ARG at positions 2 and 5 in the ligand, this allele of HLA-B27 protein can bind a multitude of different peptides. Strong binding of these peptides influences the stability of the protein through interactions extending from residues ARG6 and TYR7 all the way to CYS101. The interaction between certain peptides and B*2705 heavy chain may result in the enhanced folding problems and an inflammatory reaction due to unfolded protein response. B*2709 on the other hand is highly selective for the peptides, having only a binding residue in the F-pocket, the B-pocket being rather floppy, leading to stable binding if the peptide is extremely suitable for this purpose. This selectivity may be an advantage for B*2709 allele by avoiding the binding of certain peptides which may increase the likelihood of structural instability. | We propose a statistical thermodynamics model for determining structurally and functionally important residues in ligand-protein interactions. Our method identifies the path that the protein uses in transferring information from one point to the other. We show that a few energetically active residues are most efficient in energy exchange with the surroundings acting as ‘energy gates’. The remaining important residues that we identify are situated along the interaction path. These are the hub residues. Strong correlations exist between energy gates and hub residues along the interaction path, thus relating to allostery and cooperative binding. We studied the structure-function, ligand binding and allosteric activities of ten models of HLA Class I proteins of the immune system. Five of these models belong to the HLA-B*2705 allele and are strongly associated with a chronic inflammatory rheumatic disease. The remaining five from the HLA-B*2709 allele of the same protein are the corresponding properly functioning ones. We show that differences in the contact maps of the two types lead to significant and consistent changes in the fluctuation profile, making the HLA-B*2705 alleles respond too strongly to perturbation. | Abstract
Introduction
Methods
Results
Discussion | biophysics
biophysics/theory and simulation
biophysics/biomacromolecule-ligand interactions | 2010 | Predicting Important Residues and Interaction Pathways in Proteins Using Gaussian Network Model: Binding and Stability of HLA Proteins | 10,036 | 261 |
The scaling relationship between the size of an appendage or organ and that of the body as a whole is tightly regulated during animal development. If a structure grows at a different rate than the rest of the body, this process is termed allometric growth. The zebrafish another longfin (alf) mutant shows allometric growth resulting in proportionally enlarged fins and barbels. We took advantage of this mutant to study the regulation of size in vertebrates. Here, we show that alf mutants carry gain-of-function mutations in kcnk5b, a gene encoding a two-pore domain potassium (K+) channel. Electrophysiological analysis in Xenopus oocytes reveals that these mutations cause an increase in K+ conductance of the channel and lead to hyperpolarization of the cell. Further, somatic transgenesis experiments indicate that kcnk5b acts locally within the mesenchyme of fins and barbels to specify appendage size. Finally, we show that the channel requires the ability to conduct K+ ions to increase the size of these structures. Our results provide evidence for a role of bioelectric signaling through K+ channels in the regulation of allometric scaling and coordination of growth in the zebrafish.
Organ growth is a complex process that requires attaining not only a certain shape but also an appropriate size. The maintenance of proper proportions between organs is tightly regulated [1]. The growth of a structure at a different rate with respect to the rest of the body results in changes in proportions during development. Such allometric growth accounts for the morphological differences between juvenile and adult stages in numerous organisms. This process also contributes to changes in shape and morphology during evolution [2], [3]. Growth is regulated by both organ-intrinsic signals as well as growth factors and hormones that originate outside the target organ. Their relative contribution can vary depending on the species or even between different structures within the same organism [4], [5]. Analysis of chimeras, obtained from transplantation experiments during embryonic stages, has shown that in many cases the final size of an organ is independent of extrinsic factors, such as nutrients or hormones, suggesting that determination of size and shape are organ-autonomous properties [6]. For instance, reciprocal xenografts of limb buds between salamander species of different sizes lead to limbs that attain the final size of the donor species [7]. Further, grafting experiments in avian models have shown that the mesenchyme harbors the instructive information that specifies the final size and shape of structures such as the limb and the beak [8]–[11]. The final size of an organ or appendage results from a combination of cell number and cell size. Perturbation of the Hippo pathway causes massive proliferation of Drosophila tissues and tumorigenesis in mouse [12], while hyperactivation of the TOR pathway stimulates cell growth and can trigger entry into the cell cycle [13]. Locally acting molecules such as insulin-like growth factors (IGFs) and fibroblast growth factors (FGFs) are essential regulators of growth [6]. Yet, how these components are integrated to establish proper patterning and size during development as well as during regeneration is still unclear. Two-pore domain potassium (K2P) channels are a family of potassium (K+) channels that play an important role in determining membrane potential and cell excitability [14]. These leak K+ channels conduct instantaneous currents that are independent of voltage and show open rectification, i. e. they mediate primarily outward currents under physiological conditions. K2P channel function is modulated by neurotransmitters and pharmacological compounds as well as physiological parameters such as temperature, oxygen, osmolarity and pH [15]. Due to their ability to respond to multiple biological stimuli and their wide expression across tissues, they are thought to control many physiological processes besides determining the membrane potential. Although these ion channels have not been implicated in organ size control so far, evidence has been accumulating that endogenous bioelectrical signals orchestrate patterning and growth [16]. Endogenous electrical currents are associated with limb development and regeneration in vertebrates [17], [18] and changes in voltage accompany cessation of regenerative growth in earthworms [19]. In Xenopus laevis, a species with limited regenerative capacity, artificial induction of currents can enhance the regeneration process [20], [21], while chemical, pharmacological or molecular inhibition of ionic currents can abrogate regeneration in this species [22]–[24]. Fins are structures that show an enormous diversity in shape and size in different fish species. They also possess a remarkable regenerative capacity [25]; they can easily be manipulated and unlike internal organs, fins do not have obvious limitations on growth. The skeleton of zebrafish fins consists of a proximal endochondral and a distal dermal skeletal component. The latter is formed by segmented, concave fin rays, the lepidotrichia, which originate from mesenchymal condensations [26]. Fins grow through sequential addition of lepidotrichial segments at their distal tip via migration of mesenchymal cells along the actinotrichia, clusters of collagenous fibers that emerge from the tip of each lepidotrichium [27], [28]. Segment length slightly decreases along the proximo-distal axis [26], but does not change once joints are formed and segment boundaries are established [29]. In zebrafish numerous fin mutants have been isolated over the years [30]–[33]. Most of these mutants have reduced fins [34]. For example, impairment of the ectodysplasin signaling causes loss of fin rays in finless and nackt mutants [35], while in short fin (sof) mutants defects in connexin 43 (cx43) lead to decreased fin size with shorter segments [36]. A few mutants exhibit increased allometric growth of the fin. Among these, longfin (lof) and rapunzel (rpz) mutants have an increased number of ray segments [32], [37], whereas another longfin (alf) mutants tend to have elongated segments [36]. So far, the genetic lesion has only been identified for rpz, which is mutated in a novel teleost-specific gene with unknown function [38]. Here, we report that the allometric fin overgrowth displayed by alf mutants is due to the altered function of Kcnk5b, a K2P channel. Our analysis indicates that mutant Kcnk5b acts locally within the mesenchyme of fins and barbels to increase appendage size. Furthermore, we demonstrate that K+ conductance is required to cause allometric growth during development. Genetic experiments suggest that kcnk5b may act independently of, or in parallel to, cx43. Taken together our results provide in vivo evidence for a role of K+ channels in the determination of appendage size and proportion in the zebrafish.
another longfin (alfdty86d) was identified in a large-scale mutagenesis screen as a mutant with elongated fins and irregular segmentation of the fin rays [30], [34]. In a subsequent mutagenesis screen we isolated a second mutation (alfd30mh) showing an identical phenotype and mapping to the same chromosomal region as the original alf allele (see below). Besides the longer fins, alf mutants show overgrowth of the barbels, (Figure 1A, arrows). Homozygous mutants have a stronger phenotype (Figure S1) and their fins tend to be particularly susceptible to breakage leading to accretion of bone around the lesions. Overgrown fins and barbels in alf mutants retain their general organization; however, the fins have an altered segmentation pattern, as joint formation is variable in the mutants. On average, the length of lepidotrichial segments is increased [36] (Figure 1B and 1C); however, structures appearing as very short segments are occasionally observed (arrows in Figure 1B). In contrast to other fin overgrowth mutants such as lof or rpz [32], [37], the number of segments is not increased in alf mutants (Figure 1C). Analysis of the caudal fins during development showed that the increase in size seen in the mutants is due to an increased growth rate (Figure 1D). Wild type (wt) fins exhibit only a slight increase in relative growth during development (k = 1. 29) as growth is essentially isometric [32]. alf heterozygotes showed positive allometric growth during development of the fin with an allometric coefficient k near 2 (Figure 1D). Histological analysis of fins from heterozygous fish does not reveal appreciable differences in the size of scleroblasts and epidermal cells over those seen in wild type sections (Figure 2A). However, increased staining of the proliferating cell nuclear antigen (PCNA) during fin regeneration suggests that proliferation is increased in the mutants (Figure 2B). In sof mutants defects in cx43 are known to cause a reduction in both fin size and segment length [36]. We therefore tested whether the alf overgrowth phenotype requires the function of cx43. Crosses between alfdt30mh and a dominant sof allele, sofdj7e2, showed no epistatic interaction between the two genes (Figure 1E), suggesting that the two mutations most likely affect independent processes that both contribute to the determination of final appendage size during fin development. We mapped the alf mutations to overlapping regions on chromosome 20 (Figure 3A). We further refined alfdty86d to a genomic interval of 125 kb coding for 4 genes (bpnt1, ylpm1, kcnk5b, and syt14). In both alf alleles, distinct missense mutations (W169L and F241Y) were identified in the coding sequence for kcnk5b (Figure 3B). This gene encodes a K2P channel. The affected residues are highly conserved in kcnk5b homologs of other vertebrate species (Figure 3C). Thus, the alf phenotype is due to allelic mutations in kcnk5b. To assess the nature of these alleles we generated a phenotypic revertant (j131x8) of the dominant alfdty86d mutant (Figure 3D). PCR analysis of genomic DNA showed the presence of a 384 bp deletion leading to a frameshift and a premature termination codon. The resulting protein is predicted to lack 3 of the 4 transmembrane (TM) domains (Figure 3E). This suggests that the revertant is a null mutation for kcnk5b. Homozygotes harboring the deletion are viable and fertile; thus, kcnk5b is not essential for zebrafish development. As kcnk5b has a close paralog in zebrafish, kcnk5a (Figure 4A), which is expressed in similar tissues (Figure 4B), the lack of a loss-of-function phenotype in normal development may be due to functional redundancy between the paralogs. Together, these data endorse our finding that kcnk5b is the gene responsible for the alf overgrowth phenotype and demonstrate that these mutations are due to a gain of function rather than haploinsufficiency. We used the known structure of human KCNK4 (K2p4. 1) [39] as a template for modeling Kcnk5b and assessing the mutations. These models revealed that the affected amino acids are positioned in two distinct TM domains towards the cytoplasmic side of the protein (Figure 5A). To assess how the identified amino acid substitutions might affect Kcnk5b function, the channel properties were tested in a two-electrode voltage clamp experiment in Xenopus oocytes. This technique permits measurement of currents across the cell membrane when the membrane potential is clamped to a given value. Oocytes injected with kcnk5b (wt) cRNA react steadily to a change in voltage and do not exhibit a delay in current flow, as is expected for a K2P channel. A similar situation is also seen with kcnk5b (W169L) or kcnk5b (F241Y) cRNAs. However, oocytes injected with either of both mutant cRNAs show an almost two-fold increase in K+ conductance over that of oocytes injected with wild type cRNA (Figure 5B). The current-voltage relationship of the wild type channel shows the typical outward rectification of a K2P channel, i. e. current flows preferentially out of the cell, from the side of high K+ concentration to the side of low K+ concentration [40]. In contrast, the increase in K+ currents in the Kcnk5b mutant variants is accompanied by reduced outward rectification (Figure 5C) suggesting that the change in K+ conductance results from altered biophysical features of Kcnk5b rather than a simple increase in the number of channels at the plasma membrane. K2P channels are often referred to as leak channels since they account for the constant leaking current that sets the resting membrane potential observed in neurons. They are known to control both cell excitability and membrane potential [41], and the human homolog of kcnk5b, KCNK5 (TASK2), was shown to contribute significantly to the stabilization of the membrane potential in articular chondrocytes [42]. Therefore, we hypothesized that zebrafish Kcnk5b might also play a role in setting the membrane potential. Indeed, the membrane potential values of oocytes injected with wild type and mutant kcnk5b cRNAs are correlated with the amplitude of the ion current measured at a constant voltage of 50 mV (Figure 5D): the higher the conductance for K+ measured at 50 mV, the more negative the membrane potential of the oocyte. Consistently, the mutant channels lead to stronger hyperpolarization causing a shift in the membrane potential towards −90 to −100 mV, the equilibrium potential for K+ in Xenopus oocytes. To show where kcnk5b is expressed we performed in situ hybridization experiments on adult fins, however no specific signal above background was observed, indicating that expression levels might be below detection with this technique. Nevertheless, RT-PCR analysis showed that kcnk5b is expressed in fins of adult fish (Figure 4B). To assess whether kcnk5b acts locally within fins and barbels to control growth, we transplanted kcnk5bdt30mh/+ mutant cells into wild type hosts (Figure S2A). Local overgrowth of these structures was detected in 29 out of 120 chimeras raised to adulthood (Figure S2B–D), whereas global overgrowth of all fins and barbels was never observed. This suggests that the mutations act locally within the appendages to increase their size. We further attempted to induce the mutant phenotype by local overexpression of the channel within fins and barbels of wild type fish. Whereas the electrophysiological analysis indicated that the dominant kcnk5b mutations lead to an increase of channel conductance, the current of K+ ions through the plasma membrane depends not only on individual channel conductance, but also on the number of channels present in the membrane. Therefore, we argued that increasing the number of channels should also promote fin overgrowth. We generated a construct in which either kcnk5b (wt) or kcnk5b (W169L) expression is driven by the elongation factor 1 alpha (ef1a) promoter from Xenopus laevis; this promoter was recently shown to be active in all major fin tissues [43]. To mark the cells that express the transgene, DsRed expression was driven under a second ef1a promoter positioned in cis within the same plasmid (Figure 6A). This plasmid was injected into wild type one-cell stage zebrafish zygotes along with Tol2 transposase mRNA as described before [43]. Injected fish were raised to adulthood, screened for DsRed positive cells in the fins and the effects on growth were recorded. No overgrowth was observed in fish injected with a control plasmid expressing only DsRed under the ef1a promoter (0/240), despite the presence of DsRed-positive cells in various tissues within the fin (Figure S3). In about 40% of the fish injected with plasmids encoding wild type or mutant kcnk5b and showing DsRed positive cells in the fins we found a local overgrowth phenotype (Figure 6B and H). Analysis showed a strong correlation of overgrowth with DsRed positive mesenchymal tissue (89. 2%, N = 37, Figure 6C, D and I), whereas DsRed positive cells in other tissues were not associated with increases in size. The fin ray segments were enlarged in the overgrown fins similar to alf mutants (Figure 6D). The marked fibroblast-like cells typically occupied the intra-ray space and were excluded from the arteries (Figure 6E). These vessel-surrounding clones extended along the actinotrichia to the distal ends of the overgrown fins (Figure 6F). In the case of barbel overgrowth, DsRed positive cells were found in the mesenchymal tissue surrounding the central rod (Figure 6G), an acellular, non-cartilaginous, non-mineralized structure that supports this organ [44]. In a few cases no DsRed fluorescence signal could be detected within or next to overgrown fin tissue (kcnk5b (W169L), 2/26; kcnk5b (wt), 2/11), probably due to variegation of promoter activity [43]. In conclusion, these findings indicate that kcnk5b overexpression within fibroblasts of the mesenchyme is sufficient to induce fin outgrowth. To test whether kcnk5b-induced overgrowth requires conductance of K+ ions by the channel, we generated an overexpression construct encoding a non-conductive version by mutating the GFG motif of the selectivity filter to AAA, kcnk5b (GFGAAA). This modification has previously been shown to block ion conductance in K+ channels [45]. Electrophysiological measurements in Xenopus oocytes showed that this channel is unable to conduct K+ (Figure 6J). The plasmid was injected into wild type embryos along with Tol2 transposase mRNA and injected fish were reared to adulthood and assessed for overgrowth. No overgrowth was detected in these fish (Figure 6H), although fins containing DsRed positive tissue (n = 32), including fibroblasts (Figure 6J, inset), were found. These data indicate that the increase in conductance of the Kcnk5b channel is essential for the coordinated overgrowth of the fins and barbels in the mutants.
The size of an organ depends on cell size and cell number. The mammalian homolog of kcnk5b has been implicated in both, regulation of cell volume [50], [51] and cell proliferation [52], [53]. In alf mutants we could detect an increase in cell proliferation but not in cell size (Figure 2). Importantly, the mutant phenotype does not arise simply by dysregulation of cell proliferation, which would cause tumorous overgrowth; rather the overgrown structures in the mutants preserve tissue organization and patterning. It is unclear how K+ channels regulate proliferation. Studies have proposed that this might occur through regulation of the membrane potential [54]. In apparent contrast to some studies [55]–[58] but in agreement with others [59], [60], we found that hyperpolarization caused by mutations in a K+ channel can lead to tissue overgrowth. Although we observed a hyperpolarizing effect of the alf mutation in Xenopus oocytes, we cannot exclude that this, in turn, triggers a depolarization, either at cellular level or in the surrounding tissues during development of the fin. In fact, experiments employing depolarization-sensitive dyes, suggest that this might indeed be the case (Figure S4). The importance of hyperpolarization during growth is supported by regeneration studies in Xenopus [22], [24]. Regenerating tadpole tails are initially depolarized, but, unlike tails in the refractory state, subsequently undergo hyperpolarization. Notably, impairing hyperpolarization through inhibition of V-ATPase activity leads to a reduction of cell proliferation and failure to regenerate [22]. Transient hyperpolarization of the cell might lead to a cytosolic increase of the second messenger Ca2+, activate integrin-dependent or PTEN phosphatase-dependent cascades, or favor the uptake of mitogens such as serotonin through voltage-dependent transporters [61]. Recent reports suggest that in some cases K+ channels can induce cell proliferation independently of their effect on membrane potential [62], [63]. We show that, in the case of Kcnk5b, conductance is essential for the regulation of fin growth. Overexpression of a non-conducting version of the channel does not cause a phenotype, whereas wild type and alf variants induce local overgrowth. Our analysis of transgenic mosaics indicates that cells of the mesenchyme are sufficient to provide cues that alter the size of the fins. This is consistent with results of classic xenograft studies between chicken and quail where cells of the mesenchyme impart donor-specific characteristics to the limbs [8], [64]. During development tetrapod limbs are patterned by signaling interactions between mesenchymal cells and the overlying ectoderm. A prominent signaling center, the apical ectodermal ridge (AER), is active at the distal tip of the limb bud during this process. The AER and the mesenchyme of the progress zone continuously communicate with each other to direct limb outgrowth and development. Similar epithelial-mesenchymal interactions from the apical fin fold are likely to be required for the patterned overgrowth of fins in alf mutants. In support of this mechanism, we consistently find labeled mesenchymal cells in the distal-most regions of overgrown tissue in mosaic animals. AER signaling in amniotes requires connexin-mediated electrical connectivity between cells to coordinate pattern and growth of the vertebrate limb [65]–[67]. An analogous mechanism may be functioning in fish. We show here that altering ionic communication in the developing fin of the zebrafish is sufficient to induce growth. Our analysis of the genetic interactions between alf and sof indicate that Kcnk5b and Cx43 may act in parallel pathways to modulate final fin size. In both mutants segment length and fin size are correlated, however the role of segment patterning in size regulation of the fin is unclear. In contrast to alf [36], [68] (Figure 1) and sof [33], the overgrowth mutants lof and rpz have wild type sized lepidotrichial segments [32]. Moreover, the evx1 mutation, which leads to fins rays devoid of joints, does not affect final fin size in a wild type nor lof background [69]. Several experiments suggest that bioelectrical signaling is a shared common mechanism used across bilaterians to control organ growth and patterning [21], [70], [71] and indicate that ion flow may have an instructive role during development [72], [73], as well as regeneration [18], [74]. Here, we provide genetic evidence showing that changes in K+ channel activity result in allometric scaling of an organ, rather than causing uncontrolled proliferation. We favor a model for size regulation in which modulation of ionic current by K+ channels within the organ shifts positional information, thereby setting a different register of size during development. In fact, there is evidence for a rostro-caudal and medio-lateral gradient of voltage within the developing embryo suggesting that electric fields are a component of the positional information [75], [76]. External electrical currents have been shown to alter positional information in axial regenerates of planaria [77]. However, the underlying mechanism of signaling from electrical fields is largely unknown, and possibly depends on electrical coupling between cells [78]. This hypothesis is supported by studies in pigment pattern formation, where both K+ channels and connexins have been implicated in proper formation of the zebrafish stripes [49], [79]–[81]. Further studies will be needed to uncover the signaling mechanism from K+ channels to regulate size and pattern. However, our work, in concert with that of others, clearly shows that ion flow is not just an epiphenomenal event accompanying growth but one of the major factors specifying pattern and form during development and regeneration.
Zebrafish were bred and maintained as previously described [82]. alfdty86d was isolated in the 1996 Tübingen screen [30], [34] as a mutant affecting adult fin formation. The alfdt30mh (pfau) mutant was identified in F1 fish of a standard F3 screen (ZF Models) and isolated based on its fin and barbel phenotype. Fish were anesthetized in tricaine solution for measurements; fin length and standard length was measured using handheld calipers. Fish were imaged under a stereo microscope (Zeiss, SteREO Discovery) and measurements were performed using AxioVision software (Zeiss). p-values from unpaired Student' s t-test were obtained with Microsoft Excel. Fin regenerates were fixed at 4 dpa in 4% PFA overnight and decalcified with 0. 5 M EDTA for 24 h. Sample were embedded in paraffin and sectioned at 5 µm. Immunohistochemistry with anti-PCNA antibody (Sigma) was performed as described [83]. Percentage of PCNA positive nuclei over Hoechst positive nuclei was determined on three to four sections of four independent samples for each genotype. Adult zebrafish organs were dissected on ice and stored in RNALater (Invitrogen) at 4°C. Total RNA was isolated using RNeasy Mini kit (Qiagen). cDNA was synthesized from 200 ng RNA from each sample with SuperScript III and oligo (dT) primers (Invitrogen). PCR analysis was performed using Taq polymerase S (Genaxxon) with intron spanning primers (β-actin forward OSP-31, TGC GGA ATA TCA TCT GCT TG, β-actin reverse OSP-32: AGC ATC ATC TCC AGC GAA TC, kcnk5b forward OSP-390: CAT TCC TCT GTG CCT CAC CT; kcnk5b reverse OSP-324 AGG CCA TCC ACA GAC TCA TC, Tm = 61°C, 30 cycles). Mapping was performed as described [82]. The alfdty86d mutation mapped between z11841 (5 recombinants/96 meioses) and z21067 (2/96) and fine mapped using SNPs. alfdt30mh mapped between z7803 (1/48) and z21067 (1/48). Full length kcnk5b was cloned into pGEM-T Easy from cDNA of fin blastema amplified with LA Taq polymerase (TaKaRa) (forward primer OSP-379: TGG GAG TGT GGA GTG TGT GT, reverse OSP-382: TTT TTG GTC CAG CTT TGG TC, Tm = 60°C, 45 cycles). Sperm from alfdty86d homozygotes was irradiated with X-rays (1125 rads, Faxitron 43855D) and used to fertilize wild type eggs (AB strain). F1 progeny was reared to approximately three weeks of age (9433 fish) and screened for the alf phenotype. 11 fish showed wild type fins. 10 of these survived to adult stages. SSLP analysis revealed that 9 of these were deletions of some or all of the upper arm of chromosome 20. q-RT-PCR of candidate genes in the remaining revertant (j131x8) showed no change in transcript levels for bpnt, ylpm1 and syt14, but little or no transcript for kcnk5b. PCR analysis of genomic DNA showed that this revertant has a 384 bp deletion of the 3′ end of intron 2 and the 5′ end of exon 3. This deletion results in a frameshift and early truncation of the protein. The amino acid sequence of zebrafish Kcnk5b was retrieved from Ensembl (http: //www. ensembl. org) and used to search the PDB database with HHpred (http: //toolkit. tuebingen. mpg. de/) [84]. The first hit in the search (human KCNK4, PDB ID: 3um7 [39] identity 36%, similarity 0. 646; 22nd March 2012) was used to build the 3D model. The model was processed with MacPyMol (http: //pymol. org). kcnk5b was subcloned from pGEM-T Easy to pSGEM expression vector via SacII and SpeI sites. After linearization with NheI, cRNA was synthesized with Ambion mMessage mMachine (Invitrogen) and cleaned up with mRNeasy Mini Kit (Qiagen). X. laevis oocytes were injected as described previously [85] (kcnk5b single alleles: 4 ng wild type or mutant kcnk5b cRNA; co-injections of two kcnk5b alleles: 2 ng cRNA each, for a total of 4 ng per oocyte). Measurements were done from a holding potential of −80 mV with 0. 5 s long pulses from −100 to +60 mV with increments of 20 mV. Recorded currents (n = 5–26) were averaged and normalized to the mean value recorded for oocytes injected with the wild type channel at +60 mV. PCR mutagenesis was performed as described [86] using Pfu polymerase (Fermentas) (OSP-15 CCC TGA CGA CTG TCG CTG CAG CTG ACT ATG TGG CAG GGG C; OSP-16 CCT GCC ACA TAG TCA GCT GCA GCG ACA GTC GTC AGG GTG G, Tm = 70°C, 30 cycles) on pSGEM: kcnk5b (wt). ef1a: DsRed vector. A ef1a: DsRed cassette generated with KOD Hot Start DNA Polymerase (Toyobo) (primers: TAA TTT AAA TAG ATC TTC GAG CAG GGG GAT CAT CTA ATC A; CTA GAT GGC CAG ATC TGC CCG GGA CTT GAT TAG GGT GAT GGT TCA CGT AGT G, Tm = 59°C, 30 cycles) from plasmid Ale237 (kind gift of Alessandro Mongera) was inserted in plasmid 587jk (kind gift of Dr. Jana Krauß) using BglII restriction sites through In-Fusion Advantage (Clontech) cloning according to manufacturer' s protocol. ef1a: DsRed; ef1a: kcnk5b wild type and mutant vectors. The ef1a promoter was amplified from plasmid Ale237 (primers: ATT AAT TCG AGC TCG GTA CCC CTC GAG CAG GGG GAT CAT CT; GAA CAA GCA AGC TGG GTA CCC CGG CCG TCG AGG AAT TCT TTG, Tm = 59°C, 30 cycles) and inserted into the pSGEM vector at the KpnI restriction site using In-Fusion Advantage (Clontech) cloning. The ef1a: kcnk5b cassette was amplified from the resulting plasmid as above (primer: AAA CCT AGG TCG AGC AGG GGG ATC ATC T; AAA CCT AGG ATG ACC ATG ATT ACG CCA AGC TAT), digested with AvrII and inserted into ef1a: DsRed vector using the SpeI restriction site. Plasmids (5–20 ng/µl), Tol2 mRNA (25 ng/µl) and 20% (v/v) phenol red solution (Sigma- Aldrich, P0290-100ML) were injected into the zygote of 1-cell stage embryos under a dissecting microscope (Zeiss, Stemi 2000) using 275 Pa (40 psi) injecting pressure for 100 ms (World Precision Instruments, Pneumatic PicoPump PV820). Adults were analyzed with Zeiss, SteREO Discovery and Zeiss LSM 5 Live. Transplantations were performed as previously described [82]. At mid blastula stage (1000 cell stage), about 20–40 cells were transplanted from the pfaudt30mh/+ donors into the recipient close to the yolk cell and chimeras were raised to adulthood. Fluorescent dye experiments were performed by adapting described protocols [22], [87]. Briefly, wild type and mutant juvenile fish (STL = 16–18 mm) were incubated in fluorescent dye diluted 1∶2000 in fish water (stock solutions: DiSBAC2 (3) (Bis- (1,3-Diethylthiobarbituric Acid) Trimethine Oxonol, Life Technologies): 1 mg/ml in DMSO) for 30 min in the dark, anesthetized with tricaine solution and placed on a custom-made chamber for confocal imaging. The chamber was obtained by removing the bottom of a 55 mm plastic dish and by replacing it through a round cover slip fastened with silicone. Fish were held in place with a tissue soaked in dye and imaged upon excitation at 561 nm. Unstained animals were imaged as a negative control. p-values from unpaired Student' s t-test were obtained with Microsoft Excel. | The proportions of an animal can change during its lifetime. This often occurs through the phenomenon of relative growth, or allometry. In humans, for example, the head grows at a lower rate than the body resulting in a change in proportion between children and adults. The regulation of size and proportion is not well understood. We investigated fin growth in zebrafish as a model to understand this phenomenon. The mutant another longfin develops proportionally larger fins and barbels. Here, we show that another longfin mutants carry a mutation in kcnk5b, a gene coding for a potassium channel. Potassium channels control the electrical potential of cells and are known to regulate processes such as heart beat, neurotransmission and hormone secretion. We demonstrate that increased potassium channel activity can drive changes in growth in the zebrafish. Clonal analysis reveals that the channel acts directly in the fins and barbels to increase their size in a coordinated manner. Our work shows that potassium channels are involved in the determination of appendage size in zebrafish and suggests consistent with recent studies, an instructive role for bioelectrical signaling in development. | Abstract
Introduction
Results
Discussion
Materials and Methods | animal models
mutagenesis
developmental biology
zebrafish
animal genetics
genetic mutation
organism development
model organisms
genetic screens
genetics
biology
anatomy and physiology
electrophysiology
morphogenesis
pattern formation
limb development | 2014 | Bioelectric Signaling Regulates Size in Zebrafish Fins | 8,337 | 256 |
Chagas disease is a vector-borne disease of major importance in the Americas. Disease prevention is mostly limited to vector control. Integrated interventions targeting ecological, biological and social determinants of vector-borne diseases are increasingly used for improved control. We investigated key factors associated with transient house infestation by T. dimidiata in rural villages in Yucatan, Mexico, using a mixed modeling approach based on initial null-hypothesis testing followed by multimodel inference and averaging on data from 308 houses from three villages. We found that the presence of dogs, chickens and potential refuges, such as rock piles, in the peridomicile as well as the proximity of houses to vegetation at the periphery of the village and to public light sources are major risk factors for infestation. These factors explain most of the intra-village variations in infestation. These results underline a process of infestation distinct from that of domiciliated triatomines and may be used for risk stratification of houses for both vector surveillance and control. Combined integrated vector interventions, informed by an Ecohealth perspective, should aim at targeting several of these factors to effectively reduce infestation and provide sustainable vector control.
Chagas disease is a vector-borne disease of major importance in the Americas where it is endemic. It affects an estimated 8–10 million people, and nearly 25 millions are at risk of infection [1]. In terms of disability-adjusted life years (DALYs), and with a burden of 0. 7 million DALYs, it is the fourth most important disease in Latin America, following only hookworm, ascaris and trichuris infections [2]. The disease is caused by the protozoan parasite Trypanosoma cruzi and most transmission occurs by hematophagous triatomine vectors [3]. Disease control and prevention are mostly limited to vector control to reduce triatomine infestation of human dwellings and concomitant transmission of T. cruzi to humans [4]. Indoor residual spraying of pyrethroid insecticides and housing improvement are the main methods of interventions. Inter-governmental vector control initiatives during the 1990s are believed to have eliminated vectorial transmission to humans in several Latin American regions [4]. However, this success is mitigated by difficulties in sustaining vector control activities [5] and the emergence of insecticide resistance [6]. Additionally, some triatomine species are not well domiciliated but transiently invade house to feed on humans, representing an important source of infection. However, traditional control methods are less effective at preventing these triatomines from entering houses [3]. Integrated vector control based “on a rational decision-making process for the optimal use of resources in the management of vector populations” [7], targets ecological, biological and social determinants of vector-borne diseases to achieve improved control. It shares a common Ecohealth perspective, an ecosystem approach to health that promotes use of transdisciplinary participatory research to attain better health outcomes through environmental management [8]. Such multidisciplinary strategies are emerging as more rational, sustainable, and cost-effective than widespread empirical insecticide spraying [7], [8]. However, they require extensive knowledge of the eco-bio-social determinants leading to disease transmission. Most research on Chagas disease vectors has focused on domiciliated Triatoma infestans, the major vector species in South America. Risk factors for house infestation (and colonization) by T. infestans are typically associated with housing structure and quality; houses providing abundant hiding refuges for bug resting and reproduction (cracked adobe walls, dirt floors, thatched roofs, poor hygiene, darkness, etc. .) and easily accessible feeding sources (indoor dogs or chickens, large families) are more likely to be infested [9]–[11]. Similar results have been observed for other domiciliated vector species and populations, including T. dimidiata in Central America [12]–[16]. Accordingly, integrated vector control interventions targeting these risk factors such cement floors or roofs, or wall plastering are being evaluated [17]–[19] as alternatives to conventional insecticide spraying [20], [21]. Less is known about determinants of invasion by non-domiciliated triatomines, which have limited ability to establish domestic colonies, but transiently enter houses for blood feeding on animal and human hosts. In urban areas, infestation by T. pallidipennis or T. dimidiata has been found to be much less dependent on housing characteristics, but is instead associated with the availability of various peridomestic refuges (large peridomestic area, adjacent empty or abandoned lots) and feeding sources such as dogs, squirrels, and opossums [22], [23]. In the Yucatan peninsula, sibling species from the T. dimidiata complex typically infest houses on a seasonal basis during the months of March–July, with very limited ability to colonize houses [24]–[27]. This contrasts with its level of domiciliation in Central America and has been hypothesized to be associated with genetic differences within the T. dimidiata complex [20], [28]. This infestation is responsible for a seroprevalence of T. cruzi infection in humans of about 1–5% in the region [29], [30] and these invasive vectors cannot be fully controlled by conventional insecticide spraying [31], [32]. While we have now an accurate description of the dynamics of house infestation [33], [34], we still have limited understanding of the factors driving this process, limiting the design of vector control interventions. For example, houses located in the periphery of rural villages, in close proximity to the surrounding vegetation, were found twice as likely to be infested compared with houses located closer to village centers [35], suggesting that specific spatial targeting of vector control may be appropriate [36]. Proximity of public street lights was also found to be a significant contributor to infestation [37]. Importantly, housing quality seems irrelevant for this transient infestation [24], but more detailed ecosystemic and social studies are needed to fully identify and understand the interplay of factors contributing to this infestation pattern and to develop integrated vector control interventions. In the present study, we performed a detailed analysis of the eco-bio-social characteristics of rural villages in the Yucatan peninsula, Mexico, to identify the key determinants associated with transient house infestation by T. dimidiata.
The study was carried out from July 2010 to July 2011 in the rural villages of Bokoba (21. 01°N, 89. 07°W), Teya (21. 05°N, 89. 07°W) and Sudzal (20. 87°N, 88. 98°W), located about 15–20 km apart in the central part of the Yucatan state, Mexico. The regional climate is warm and humid, with an average annual temperature of 26°C and 1150 mm of rainfall, and the villages are surrounded by a mixture of secondary bush vegetation and agricultural/pasture land. There are a total of 570,702 and 416 houses in Bokoba, Teya and Sudzal, respectively, all of which have been georeferenced previously [35]. The respective populations are of about 2,000 inhabitants in both Bokoba and Teya and 1,600 in Sudzal, with about 40% of the population below 14 years of age. Most of the population (over 90%) is of Mayan descent and culture. Each of these communities has a health center run by the state public health system (Secretaria de Salud de Yucatán), as well as public primary and secondary schools and at least one church. The population is largely Catholic (over 95%). Previous work indicated that there was no significant difference in the overall housing characteristics and living conditions of these villages, all three of them being rather representative of the conditions in rural Yucatan [32]. There has been no systematic vector control program in these villages, but entomologic monitoring by community participation has been performed since 2006 and pilot interventions were implemented in a limited number of houses in 2007 [32]. An extensive survey was developed and validated to identify the key eco-bio-social dimensions of the households. The survey included a total of 127 variables describing in detail housing structure and characteristics (34 variables), peridomicile structure and characteristics (56 variables), socio-demographic characteristics and cultural practices (38 variables) (see Table S1 for the complete list of those variables). Housing structure variables included the number of rooms and construction materials of the different parts of the house (floor, wall, roof). Peridomicile variables included data on the size of the peridomicile, its vegetation, the presence of different structures (storage, corrals, others) and the presence of different species of domestic animals. Socio-demographic characteristics and cultural practices included a detailed description of the composition of the household, its socioeconomic status, and common practices related to the maintenance and care of the house and the peridomicile, including use of insecticides and other potential vector control measures (e. g. storage of grains or construction material, cleaning habits, etc…). Practices related to the care of domestic animals were also investigated. A total of 346 households randomly sampled in the three villages was selected for the survey (representing 20% of the total number of households). We used a random sampling scheme to avoid any bias in selecting specific households and ensure that all types of households would be included in the survey, irrespective of the housing type, structure, position in the village, infestation status. Teams of two trained field workers performed individual visits to each household to apply the survey following obtention of written informed consent. The protocol was approved by both the World Health Organization and the Autonomous University of Yucatan institutional bioethics committees. We were able to obtain data from 308 households, the remaining being abandoned houses, households that declined to participate, or households in which inhabitants were unavailable after three visits. In all three villages, house infestation was monitored from July 2010 to July 2011 by community participation, which we have found highly reliable and more sensitive than timed manual searches for entomologic surveys of houses with low and transient infestation [24], [25]. Infestation was defined as the catch and notification of at least one triatomine (adult or nymph) inside the domicile at any time of the year. Community members were asked to collect any triatomine-like bug observed in their houses, using plastic bags to avoid direct contact with the bugs, and take them to the health center where the bugs were registered together with basic information on the household. We later visited each household that had collected a bug to confirm the coordinates we had previously georeferenced [35]. Regular community meetings were held during the study to promote Chagas disease awareness and ensure community participation. Conventional indices were calculated to describe infestation: infestation index (percent of houses with indoor triatomine presence at any time of the year), colonization index (percent of infested houses with nymphal stages), density index (number of triatomines per infested house) [23], [24]. We used a mixed modeling approach by first performing univariate analyses of the 127 variables describing the eco-bio-social conditions of the 308 houses using logistic regression, to explore the potential association between the transient presence of T. dimidiata at any time of the year (house infestation) and these variables. A few data were missing in the database, corresponding to different houses for different variables, and thus missing data were considered randomly distributed (up to 8% of missing values for one variable retained in the model). We then used a multiple imputation method implemented in the R package ‘amelia II’ to estimate the missing values [38], [39]. Following established guidelines we constructed 5 datasets with imputed data [38], [40]. For each logistic regression, the value of each coefficient was calculated as its mean value across the 5 datasets, and standard errors were calculated taking into account the mean intra-dataset variance and the between-dataset variance [40]. We then selected for further multivariate analysis a sub-set of 29 variables that had P-values<0. 1 from a likelihood ratio test in the univariate analysis. Because of the redundancy and overlap of several of these variables, as assessed by correlation analysis, these were further reduced to a subset of 9 variables selected to incorporate the maximum independent information in our model while keeping covariation among variables at a minimum (see results and Table S2). We evaluated the goodness of fit of the complete model (i. e. the model with all nine variables) based on a generalised coefficient of determination [41], and the potential for over-dispersion in the data [42]. Multivariate analyses were performed using the framework of multimodel inference and selection followed by model averaging [42] to assess the support of each model in terms of Akaike' s Information Criterion (AIC). We also determined the odds ratio associated with each of the 9 variables, with their 95% confidence intervals (95%CI) constructed using multi-model estimations of each coefficient' s variance [42], [43]. All analysis were performed in Matlab (R2012b, The Mathwork). We performed 512 logistic regressions including different combination of zero up to nine of the selected variables and determined the likelihood () for each of the models. The AIC of each model (corrected for small sample size) was calculated as where k is the number of parameters estimated in the model. The final AIC was taken as the mean AIC from the 5 imputed datasets [44]. Models were then ranked from the best supported model (with the minimum value of AIC, AICmin) to the least supported one (maximum value of AIC). Akaike differences were calculated as and models with were considered to have a considerably lower support than the best supported model [43]. The Akaike weight (WAIC) for each model was defined as: and it provides the probability for each model to be the best model. The relative importance of a particular variable was then calculated as the sum of the Akaike weights of all models that contained this particular variable. Finally, to obtain a model including the most complete information and the best predictive ability, we performed model averaging, in which each parameter was weighted by the WAIC of each model and averaged for all 512 models [45]. Thus, for parameter linking infestation with a variable in the logistic regression, the average value over all R models is: A confidence interval was calculated for each parameter assuming a normal distribution with a total variance associated to the parameter. This variance accounted for (1) the uncertainty in the parameter value given a certain model, and for (2) the uncertainty associated with model selection, as follows: We then derived a confidence interval for the odds ratio associated with each parameter by taking the exponential of the lower and upper bounds of the parameter confidence interval. Finally, we evaluated a posteriori interactions among the six most supported variables of the best model, and constructed 15 models including one pairwise interaction. The statistical significance of each interaction was tested using a Student' s t test. For each household, the probability of infestation was calculated based on the model' s averaged parameters to assess the reliability of the model in terms of sensitivity and specificity. We thus calculated the generalized coefficient of determination (following [41]) to assess the fit of the model to observed data. To assess predictions at a coarser level, we also defined 15 groups of 20 houses according to their predicted probability of infestation. Houses from the first group had the 20 lowest predicted probabilities of infestation, and those from the last group had the 20 highest predicted probabilities (due to a sample size of 308, the first and last group actually consisted of 23 and 24 houses). For each group, we calculated the observed infestation probability defined as the proportion of infested houses in the group. The relationship between predicted and observed infestation was evaluated by correlation and regression analysis.
The eco-bio-social characteristics of a total of 308 households were obtained from the field survey. A typical household was composed of a family of 4–5 persons (4. 1±0. 1), led by a man in 77% of the cases. Most worked as subsistence farmers (38%), some in construction or manufacture (14%) and only 22% had a regular work contract. Sixty-three percent received social welfare benefits (“Oportunidades” program). Education level reached primary school for most men and women (63%). The houses had been built 20±1 years ago and consisted of 2. 1±0. 1 adjacent rooms. This included 1. 6±0. 1 bedrooms and rooms had an average of 1. 5±0. 1 windows (Fig. 1). Houses were of cement block construction (96%), often fully plastered walls (63%), with cement floors (93%). Similarly, roofs were made of cement/concrete; only 5% were thatched and 8% from tin. Fourteen percent of houses had no sanitation system. Houses were surrounded by a peridomestic area of on average 1300 m2, limited by a fence of piled rocks (84% of houses). Vegetation was scarce close to the house and trees were somewhat denser further from it (>10 m). Many families kept domestic animals all year round (58%), the most frequent being dogs (52%), chickens (49%), cats (34%), and songbirds (14%). Other animals such as rabbits, pigs, sheep, horses and cows were rather rare (<3%). Animals were usually kept close to the house (<10 m, 83% of cases). Songbirds were kept in suspended cages attached to the fronts of houses, chickens and turkeys were sometimes kept in a corrals/coops (22%), while other animals had free range in the peridomicile. Construction materials were sometimes stored in the peridomicile (25%), while corn or other grains (13%) and firewood (10%) were stored inside the house. The peridomicile area was cleaned at least once a week in most cases (66%), usually by men. About 55% of families used domestic insecticide products such as mosquito coils (45%), plug-in repellents (32%) or spray insecticides (55%) on a regular basis, but only a few (5%) resorted to professional insecticide spraying. Thirty-nine percent of households reported having seen triatomines in their house. Triatomine transient domestic infestation was detected in 46 of the 308 households (14. 9%), corresponding to 10/70 houses (14. 3%) in Sudzal, 9/106 houses (8. 5%) in Bokoba, and 27/132 houses (20. 5%) in Teya. The colonization index was of 5/46 (10. 8%) and the density index was of 2. 9 triatomines/house, ranging from 1 to a maximum of 43 triatomines/house. All these data were similar to what has been previously observed in these same villages [37], which are also very representative of other villages from the region [23], [24], [46]. The potential association of house transient infestation with eco-bio-social characteristics was first assessed by logistic regressions and 29/127 variables included in the survey were found correlated with house infestation at a P<0. 1 level (Table S2). Importantly, none of the variables related to the socio-economic status of the household, the education level or general cultural practices such as sleeping or cleaning habits of the house or peridomicile were found to be associated with infestation. The use of various domestic insecticide products or peridomestic pesticides had also no relationship with infestation. Similarly, most variables describing housing and peridomicile structure and organization, including floor, walls or roof type, number of rooms, number of inhabitants, peridomestic vegetation type and density, were not significantly associated with infestation. On the other hand, the 29 variables associated with transient domestic infestation (Table S2) were related to the presence of specific domestic animals such as dogs, chickens and perching birds, housing condition such as wall plastering, and the surroundings of the houses such as the presence of piles of rocks in the backyard/peridomicile, the proximity of a street light and the location of a house in the periphery of the village. We first further reduced the number of variables because of the redundancy and correlation of several of these, so that we could incorporate the maximum independent information in our model and limit covariation among explanatory variables (Table S3). For example, three variables described the presence of dogs: ‘Presence/absence of dogs’, ‘Number of dogs’ and whether dogs were ‘free ranging, enclosed in a corral or tied on a leash, or absent’. We thus eliminated the first variable because of its complete redundancy with the other two and further selected the variable with the lowest P-value in the univariate analysis, which led to retain the variable ‘Number of dogs’ for further modelling. This process resulted in the elimination of 20 variables and only 9 variables were kept for multivariate analysis, with pairwise correlations among them always lower than 0. 25 (Table S3). These variables included the number of dogs, presence of chickens in a corral or free-ranging, proximity to the periphery of the village, practice of removing trash from the peridomicile, presence of rock piles close to the house, presence of songbirds, the storage of firewood inside the house, complete wall plastering and the proximity of a public street light. We then evaluated the complete model (i. e. the model with all 9 variables) and found no obvious over-dispersion in the data (Pearson' s goodness of fit, , , or using the residual deviance:). Furthermore, this model accounted for 29% (R2) of the variation in infestation. Analysis of all 512 models including different combinations of the 9 variables indicated rapidly increasing AIC scores and, but 10 models presented a ΔAIC of less than 4. 5 and were thus considered to receive support from the data (Table 1). These 10 models included the complete model, but the best supported model was comprised of only 8 of the variables and had a coefficient of determination of. Since all 10 models contained at least 6 variables, infestation by triatomines may not be attributed to a single or few factors, but rather seemed to depend on a complex combination of conditions. We then proceeded to model averaging to identify the strongest determinants for house infestation and evaluate the predictive power of our model. The Akaike weight of each variable in the averaged model indicated that five variables could be considered of high importance in defining house infestation with WAIC>0. 9, two additional variables were of secondary importance with 0. 7<WAIC<0. 9, and the remaining had limited contributions (Table 2). As reported before [35], the location of a house at the periphery of a village increased the risk of infestation and had a very high weight. Keeping chickens in a coop or corral was a major determinant of infestation - it resulted in a 2. 4 fold higher risk of infestation by triatomines - while having free-ranging chickens had no effect (Table 2). Cleaning trash from the peridomicile area similarly doubled the risk of infestation, and the presence of more than two dogs also significantly increased this risk. The storage of firewood inside the house was a major protective factor in the averaged model, although the individual effect was not statistically significant. Risk factors of secondary importance consisted of the proximity of a public street light, which had been identified before [37], and the presence of rock piles in the peridomicile. Again, the individual contributions of these factors in terms of odds ratio was not significant. Finally, the complete plastering of walls and the presence of perching birds had very minor weights in the model and may thus be of limited relevance as determinants for infestation. Potential interactions between the six most supported variables were further evaluated a posteriori, but none of them reached statistical significance. We then tested if these determinants of infestation could be used to predict house infestation, and thus be used to target potential vector control interventions. The generalized coefficient of determination of the averaged model was, indicating that about a third of the variance in the observed pattern of infestation at the level of a single house could be predicted by the model. The model allowed the correct identification of 90% of non-infested houses (specificity), while 41% of infested houses could be correctly identified (sensitivity). We also grouped houses according to their predicted probability of infestation to assess the reliability of the model at a somewhat larger scale (groups of 20 houses) by comparing the predictions with the observed infestation index of each group of houses. In this case, the averaged model provided an excellent prediction of infestation probability compared with the observed infestation index (R2 = 0. 85, P<0. 0001, slope = 0. 878) and was thus able to explain most of the variance in the infestation pattern (Fig. 2).
Transient house infestation by non-domiciliated triatomine vectors remains a key challenge for the design of sustainable vector control interventions and further reduction of the burden of Chagas disease [47]. This infestation pattern makes conventional insecticide spraying poorly effective, and a better understanding of the determinants of infestation is needed to formulate novel vector control strategies [31], [32], [36], [48]. In fact, risk factors for house infestation by these non-domiciliated triatomines remain poorly understood, limiting the breadth of potential control interventions to be tested. We performed here the first detailed analysis aimed at identifying possible determinants of domestic infestation by T. dimidiata. From the initial univariate screening of 127 eco-bio-social variables describing the rural ecosystem, it was clear that variables associated with infestation by invasive triatomines are distinct from the determinants usually associated with infestation and colonization with domiciliated triatomines. Indeed, socio-economic status or housing quality were clearly not relevant for infestation [23], [24]. Similarly, even indoor use of a variety of domestic insecticide products was found irrelevant to prevent triatomine infestation. On the other hand, the strongest five determinants for infestation that were identified through our model selection and averaging approach included the number of dogs, having chickens in a corral, the practice of cleaning of trash from the peridomicile, and being located in the periphery of the village, which all favoured infestation, while the presence of firewood inside houses appeared protective. To a lesser extent, the proximity of public lights and the presence of rock piles in the peridomiciles were also associated with infestation according to our model. The relationship between the storage of firewood and the cleaning of the peridomicile with house infestation is difficult to interpret. Indeed, the presence of firewood has usually been associated with increased infestation risk [49] due to the passive transport of bugs and the potential refuge it provides. Alternatively, households may use smoke from firewood to repel insects from their house, as reported in the state of Chiapas, Mexico [50]. Peridomicile cleaning would also be expected to limit peridomestic infestation, and as a consequence the dispersal of peridomestic bugs towards houses. Studies on peridomicile management aimed at eliminating peridomestic bug colonies suggest it may indeed contribute to integrated vector control [32], [51]. Alternatively, peridomicile cleaning may reduce the availability of refuges and increase bug dispersal, and as a consequence favour domestic infestation. This potential effect of environmental management has not been considered yet, and may have contributed to the pattern observed in this study. On the other hand, all other determinants for transient infestation that we identified here are consistent with our previous hypothesis that poorly-fed bugs from both the peridomicile and surrounding sylvatic areas, foraging for blood sources, are infesting houses [26], [52]. Thus, houses located in the periphery would be more at risk of infestation as shown before [34], [35], [52]. The presence of dogs and chickens, the most common domestic animals, would be attractive food sources that may be effectively detected by foraging bugs [53]. Further studies of T. cruzi infection in dogs may provide additional information on their role as domestic reservoirs in the villages. Interestingly, only chickens held in a corral or coop seem to contribute to infestation, while free-ranging chickens do not play a role. Thus, a concentrated and captive chicken population may provide a stronger signal to attract bugs and an easier food source. The contribution of dogs and chickens in infestation by T. infestans has been observed previously [11], [54], [55]. Public street lights may then interfere with the dispersal process and attract bugs to nearby houses as suggested before [37] and rock piles may provide additional peridomestic refuges for bugs. The reliability of our model was further assessed by evaluating its ability to predict house transient infestation based on the determinants identified. At the level of an individual house, our averaged model was able to account for about 30% of the presence/absence of infestation by T. dimidiata. However, when houses were grouped according to their predicted infestation, the model then accounted for up to 85% of the variations in the observed infestation index. Unfortunately, the predictive value of models is rarely reported by authors attempting to identify risk factors, even though they often base subsequent vector control interventions on such studies, with little certainty that infestation will actually be affected [9], [12]–[15], [17]–[19]. The general practice is to use odds ratio statistics to identify key risk factors, but the actual capacity of the key factors to predict the level of risk is not assessed. This lack of assessment occurs with risk studies of other vector borne diseases as well, including malaria or dengue [56]–[60]. To our knowledge such evaluation has been attempted only once from a multimodel inference approach similar to the one adopted in this paper. This study identified a limited number of risk factors contributing to the infestation of domiciles or of chicken coops by T. infestans in the Grand Chaco, Argentina [54] and the authors genuinely tested the predictive capacity at the site level of their best statistical model. They reported levels of sensitivity and specificity of 49% and 82% in domicile infestation, and 65% and 71% in chicken coops infestation, which is similar to the sensitivity (41%) and specificity (90%) we reported. In our study, despite the fact that bug dispersal is a complex dynamic process with potentially many interacting factors, which can lead to very variable infestation outcomes, the few variables we identified account for most of the variability in terms of infestation at the level of groups of houses (R2 = 0. 85). These data suggest that we have identified the key determinants of infestation in this model and that additional determinants we may have missed will be minor contributors to infestation. While the model was not tested on additional villages, the representativeness of the studied villages suggests that our findings may be extrapolated to villages with similar characteristics. Additional studies should help explore further generalization of our results. Thus, infestation is actually associated with a rather limited set of risk factors, each having a modest contribution, and their particular combinations and synergism is what seems to be associated with infestation of specific houses. Importantly, our results suggest that targeting a single risk factor may be ineffective for vector control and that combined integrated interventions targeting multiple variables may be required to adequately reduce infestation by T. dimidiata. Nonetheless, the small number of factors to be modified suggests that such integrated control might be feasible. In conclusion, our search for eco-bio-social determinants for house transient infestation by non-domiciliated T. dimidiata clearly identified several factors, including the presence of dogs and chickens in the peridomicile, potential refuges such as rock piles, and the location of the house close to the vegetation at the periphery of the village, and proximity to public street lights. These factors allowed us to explain most of the variation in infestation within villages in rural Yucatan, Mexico. These results may be used for risk stratification within villages for both vector surveillance and control. They also suggest that combined integrated vector interventions, informed by an Ecohealth perspective, should target several of these factors to effectively reduce infestation and provide sustainable vector control. | Chagas disease is a parasitic disease of major importance in the Americas, transmitted by triatomine insects. Integrated control interventions targeting a combination of factors associated with the presence of the insect vectors are increasingly investigated for improved control. Here we identified the factors associated with the seasonal intrusion of triatomine vectors in houses from the Yucatan peninsula, Mexico, by studying the characteristics of 308 houses from 3 villages. The presence of triatomine vectors was associated with the presence of dogs, chickens and potential bug refuges, such as rock piles, and the proximity of houses to vegetation at the periphery of the villages and to public light sources. Thus, factors favoring seasonal intrusion of triatomines appear different from those favoring their domiciliation. Integrated control interventions based on this Ecohealth perspective should focus on several of the factors identified in this study to achieve effective and sustainable vector control. | Abstract
Introduction
Materials and Methods
Results
Discussion | 2013 | Eco-Bio-Social Determinants for House Infestation by Non-domiciliated Triatoma dimidiata in the Yucatan Peninsula, Mexico | 7,182 | 198 |
|
GABAergic interneurons (INs) in the dorsal lateral geniculate nucleus (dLGN) shape the information flow from retina to cortex, presumably by controlling the number of visually evoked spikes in geniculate thalamocortical (TC) neurons, and refining their receptive field. The INs exhibit a rich variety of firing patterns: Depolarizing current injections to the soma may induce tonic firing, periodic bursting or an initial burst followed by tonic spiking, sometimes with prominent spike-time adaptation. When released from hyperpolarization, some INs elicit rebound bursts, while others return more passively to the resting potential. A full mechanistic understanding that explains the function of the dLGN on the basis of neuronal morphology, physiology and circuitry is currently lacking. One way to approach such an understanding is by developing a detailed mathematical model of the involved cells and their interactions. Limitations of the previous models for the INs of the dLGN region prevent an accurate representation of the conceptual framework needed to understand the computational properties of this region. We here present a detailed compartmental model of INs using, for the first time, a morphological reconstruction and a set of active dendritic conductances constrained by experimental somatic recordings from INs under several different current-clamp conditions. The model makes a number of experimentally testable predictions about the role of specific mechanisms for the firing properties observed in these neurons. In addition to accounting for the significant features of all experimental traces, it quantitatively reproduces the experimental recordings of the action-potential- firing frequency as a function of injected current. We show how and why relative differences in conductance values, rather than differences in ion channel composition, could account for the distinct differences between the responses observed in two different neurons, suggesting that INs may be individually tuned to optimize network operation under different input conditions.
The dorsal lateral geniculate nucleus (dLGN) receives input from retinal ganglion cells and transmits processed information to visual cortex. About 75–80% of the neurons in the dLGN are thalamocortical (TC) neurons, also called relay neurons, as they relay information from the retina to the cortex. Local GABAergic interneurons (INs) constitute the remaining 20–25%, and are responsible for most of the intra-nuclear connections [1]. By providing feed-forward inhibition from retinal ganglion cells to TC neurons, INs control the number of visually evoked spikes in TC neurons, and refine the receptive fields of TC neurons (e. g. [2]–[5]). The INs are also important for synchronizing thalamic oscillations [6]–[7]. Comprehensive mathematical network models will likely be important for a comprehensive understanding of the key functional features of early sensory processing [8]–[11]. Due to the relatively high abundance of experimental data, the early visual system has attracted particular interest from theoretical neurobiologists. Several mechanistic network models aimed at mimicking responses of neurons in primary visual cortex, and neurons in the dLGN that provide the feed-forward input to visual cortex, have been developed [12]–[21]. Such network models require (i) detailed information about network connectivity and (ii) mathematical neuron models that capture the salient physiological properties of the individual neurons types. For dLGN, a host of physiological and anatomical experiments have provided detailed information about the neuronal connectivity, as well as the morphology and physiology of TC neurons and INs (see e. g. reviews in [22]–[24]). Different electrophysiological characteristic of TC neurons have been captured in a series of modeling works [25]–[31], and the accumulated insight has been incorporated in a high-resolution model which comprises a detailed description of the cell morphology, a set of different ion channels and their distribution over the somatodendritic membrane [32]. For the INs, the situation is more problematic: models are few and less detailed [33]–[36]. Until now, network models for the early visual system have either omitted INs entirely, or represented them in a very simplified manner (but see [37]). A satisfactory theoretical understanding of the computational properties of the dLGN circuit, and thus also the input to visual cortex, will likely require network models incorporating more detailed IN models. The development of such models is the topic of the present paper. INs may exhibit a rich variety of firing patterns, including (i) initial “sags” by hyperpolarizing current injections, (ii) rebound bursts when released from hyperpolarization, (iii) tonic firing of action potentials (APs) by depolarizing current injections, (iv) initial bursts by depolarizing current injections, (v) spike-time adaptation during depolarizing current injections, and (vi) periodic bursting during depolarizing current injections [36], [38]–[44]. Most previous models focus on aspects of passive signal propagation in INs [33]–[35]. To our knowledge, the only currently available model that includes a variety of the active mechanisms in INs was primarily developed in order to study the mechanisms behind the rebound bursts [36], whereas other properties such as dendritic conductances and the relationship between somatic current injections and action potentials frequency (hereby referred to as the I/O curve) were not taken into account. The morphology and distribution of dendritic ion channels are crucial for the integration of synaptic input (reviewed in [45]), and can even influence the neuron' s response to somatic current injections [46]–[49]. Several active conductances have been identified in the dendrites of INs [38], [50]–[54]. Dendritic ion channels are likely of particular importance in INs, as their dendrites have not only postsynaptic contacts for excitatory retinal and cortical input, but also presynaptic terminals for inhibitory output to TC dendrites [2]. An understanding of how INs provides feed forward inhibition to TCs thus requires models that incorporate the electrically active processes in the dendritic tree. No previous IN model includes such properties. We here propose a new and detailed multi-compartment model of the IN, which advances previous models in several aspects. Firstly, it includes a detailed 3D reconstruction of IN morphology, and a set of somatic and dendritic active conductances which could reproduce some of the key response patterns of INs (including (i) - (vi) listed above). In this way it lays the foundation for simulating active dendritic signaling on a fine spatiotemporal scale. Secondly, we present two different parameterizations (P1 and P2) of the model, which were constrained by current-clamp data from two example neurons (IN1 and IN2). This approach allowed us to do a comparative study, and relate differences in the parameterizations P1 and P2 to differences in the firing patterns between IN1 and IN2. The two parameterizations contain the same set of ion channels such that differences in the responses of the two neurons could be explained solely in terms of relative differences in the peak values of the conductances. Thereby, we demonstrate that relative differences in conductance values of the included ion channels may account for the substantial variations in firing patterns observed between INs. Finally, we were able to explain the experimentally observed responses in different neurons (IN1 or IN2) under 8 different input conditions, obtaining quantitative agreement with the experimental I/O-curves over the entire input range studied. This is a significant advance compared to previous models of similar cells. Since the overall input to any given cell in a network generally varies with time over a wide dynamic range, our model will likely allow a more accurate representation of the integration and computational properties of these neurons when included in a realistic dLGN network. Preliminary results from this model have previously been presented in abstract form [55]. The model will be publicly available on ModelDB (http: //senselab. med. yale. edu/modeldb).
Brain slices containing dLGN were prepared from GAD67-GFP (Δneo) knock-in mice [56] in accordance with the guidelines and approval of the Animal Care Committee in Norway. Mice, 29–33 days old, were deeply anaesthetized with halothane and sacrificed by rapid decapitation. A block of the brain was dissected out and 250–300 µm thick coronal slices were cut in 4°C oxygenated (5% CO2–95% O2) solution containing (mM): 75 glycerol, 87 NaCl, 25 NaHCO3,2. 5 KCl, 0. 5 CaCl2,1. 25 NaH2PO4,7 MgCl2, and 16 D-glucose, and kept submerged in oxygenated (5% CO2–95% O2) artificial cerebrospinal fluid (ACSF) containing (mM): 125 NaCl, 25 NaHCO3,2. 5 KCl, 2 CaCl2,1. 25 NaH2PO4,1 MgCl2 and10 D-glucose at 34°C for at least 30 min before the experiment. During experiments, slices were kept submerged in a small (∼1. 5 ml) chamber and perfused with ACSF at the rate of 5 ml min−1 heated to at 36°C through an inline heater. In some of the experiments, as indicated, 4-Ethylphenylamino-1,2-dimethyl-6-methylaminopyrimidinium chloride (ZD 7288; 20 µM; Tocris Bioscience, Bristol, UK) was included in the perfusion ACSF to block the hyperpolarization-activated cation current Ih. Whole-cell voltage- or current-clamp recordings were made from INs in dLGN. Neurons were visualized using DIC optics and infrared video microscopy. INs were identified by expression of GFP which was specifically expressed in GABAergic neurons under control of the endogenous GAD67 promoter in the GAD67-GFP knock-in mice [56] we used. Recordings were obtained with borosilicate glass electrodes (4–6 MΩ) filled with (mM): 115 potassium gluconate, 20 KCl, 10 HEPES, 2 MgCl2,2 MgATP, 2 Na2ATP, 0. 3 GTP (pH adjusted to 7. 3 with KOH). For morphology reconstruction, biocytin (0. 25%; Sigma-Aldrich, St Louis, USA) was included in the intracellular solution. Current traces were recorded and filtered at 3 kHz with a HEKA EPC 9 amplifier (HEKA Electronik, Lambrecht, Germany), while voltage traces were recorded and filtered at 10 kHz with an Axoclamp 2A amplifier (Molecular Devices, Palo Alto, CA, USA). After recordings, slices were fixed by 0. 1 M phosphate buffer containing 4% paraformaldehyde and kept there for at least 24 h. After biocytin histochemistry with avidin-biotin complex (Vectastain ABC kit, Vector Laboratories, Inc, USA) and diaminobenzidine (DAB; Sigma-Aldrich, St. Louis, USA), interneurons were drawn under a x100 objective using software for neuron reconstruction (Neurolucida, MicroBrightField, Inc. USA).
The set of empirical data used for constraining the model is presented in Figure 2. The firing patterns in two different INs (IN1 and IN2) under 8 different experimental conditions are shown in Figure 2A–B. IN1 had a resting potential of -63 mV (Figure 2A1). The onset of a strong hyperpolarizing current injection (-150 pA) made the membrane potential drop rapidly to a hyperpolarized peak value, before it increased to a less hyperpolarized plateau value. This initial sag is a trademark of the Ih current. IN2 had a resting potential of -69 mV (Figure 2A2). The initial sag for hyperpolarizing current injections was less pronounced in IN2 than in IN1. Depolarizing stimuli gave rise to an initial, transient response (characterized by a high AP-firing frequency), followed by regular AP- firing at lower frequency. In IN2 the initial response tended to be distinct and burst-like (see e. g. Figure 2A2,40 pA), whereas IN1 tended to have a more gradual transition between the initial response and the regular AP-firing (see e. g. Figure 2A1,70 pA). We did not quantify these differences, but use the term initial burst for all initial responses that can be distinguished from the later, more regular AP firing. IN1 needed stronger depolarizing input (≥55 pA) to initiate AP-firing than IN2 (≥40 pA). However, IN1 had the steepest I/O-curve (see below), so that both neurons had about the same firing frequencies when stimulated at 70pA. In one experiment, small current injections were used to hold the membrane potential at a depolarized value (-57 mV in IN1 and -58 mV in IN2). In this case, strong hyperpolarizing current pulses (-150 pA) were followed by a rebound burst in IN1 (Figure 2B1), but not in IN2 (Figure 2B2). The Ih -current in rat INs has no calcium dependence [82] and has been measured and thoroughly described [41]. A comparison of the time course of the initial sag for recordings from rat INs [36], [41] with ours from mouse INs, especially IN1 (Figure 2A1), indicated that Ih may differ between the two species. We therefore recorded the Ih response to different hyperpolarizing voltage steps in our mouse INs (Figure 2C), and used these data to estimate the Ih kinetics. We derived Ih -activation curves for mouse INs using the following procedure [41]: With Imax denoting the maximum observed amplitude throughout the trial (e. g. , about -180 pA in the dataset shown in Figure 2C), and Iinf denoting the steady-state value of the response for a given command potential (V), the ratio Iinf/Imax was plotted against V for three data sets (Figure 3A). The data points were then fitted by a Boltzmann curve (Figure 3A; full line), given by: Iinf/Imax = (1+exp ( (V-shift) /stp) ) −1, where shift is the potential at half-inactivation, and stp determines the steepness of the activation curve. Using the inbuilt optimizer fminsearch in Matlab, we obtained the optimal parameters shift = -96 mV and stp = 10 mV. The corresponding values estimated for rat INs were -79 mV and 7. 4 mV [36], [41]. In order to determine the voltage dependence of the time constants for steady-state activation, the current traces (interval from 0–1000 ms after stimulus onset in Figure 2C) were fitted by exponential curves on the form Ih = (1-exp (-t/τh) ) ·Iinf. Only a single exponential term was used, as it yielded a good fit. In this way we estimated the time constant (τh) at each command potential as plotted in Figure 3B. The data points for τh were in turn fitted with a bell shaped curve (as in [41]) of the form: τh (V) = exp[ (V+a1) /a2]/ (1+exp[ (V+a3) /a4]), with V measured in mV and τh in ms. Using fminsearch, we obtained the optimal fit for [a1, a2, a3, a4] = [250,30. 7,78. 8,5. 78]. Note that the maximum value of τh is about 200 ms, while the corresponding value found for rat INs was about 1000 ms [36], [41]. The faster kinetics was in good agreement with the time course of the initial sag (see Figure 4). We sought two parameterizations of the model (P1 & P2), which separately should explain the characteristic features of the respective data sets (IN1 and IN2) under all 8 experimental conditions (Figure 2A–B). As an additional constraint we assumed that the two neurons contained the same types of ion channels and had the same axial resistivity (Ra). This means that we aimed to explain the differences between IN1 and IN2 in terms of differences in three passive parameters (Rm, Epas, Cm), seven parameters representing the density of ion channels (gh, gNa, gKdr, gCaT, gCaL, gAHP, gCAN), and two parameters (ShNa, ShKdr) representing shifts in the activation/inactivation curves of Na and Kdr along the voltage axis. Shifts in the kinetics of the AP-generating currents were allowed to be free parameters as we observed clear differences in spiking threshold between neurons IN1 and IN2. In order to limit the number of free parameters, the voltage and calcium dependence of the remaining ion channels were assumed to be identical in the two neurons. The experimental protocols for the two example neurons (Figure 2A–B) were replicated in the simulations shown in Figure 4, where we show that many significant features of the experimental results are reproduced by the model using the two sets of parameters (P1 and P2, Table 1). P1 had resting potential -63 mV, and P2 had resting potential -69 mV (as in the empirical data sets). Stimulus intensities between -150 pA and 20 pA gave rise to sub-threshold responses in both models. The higher response amplitudes in P2 were due to the significantly higher membrane resistance found for this neuron. Initial sags in P1 and P2 for strong hyperpolarizing current injections (-150 pA) were due to Ih. Simulated action potentials had a width (at half max response) of about 0. 4 ms in P1 as well as in P2 (Figure 5A), which is typical for INs [39], [40], and agreed well with the AP width of IN1. The APs elicited by IN2 were somewhat broader, and had a width of about 0. 6 ms. We did not change the kinetics of Na and KDR channels to account for the variability in AP shapes. The mechanisms behind other characteristics in the response patterns are discussed in the following subsections. As in the data sets (Figure 2A–B), depolarizing current injections to the model (55–70 pA to P1 and 40–70 pA to P2) gave rise to an initial burst (or a few APs with short intra-spike intervals) followed by regular activity with a lower AP firing frequency (Figure 4A). The slope of the I/O curve was mainly regulated by the interplay between calcium entering through CaL channels and a single, calcium activated potassium channel (IAHP). As the high-voltage activated CaL channels open during APs, the intracellular calcium concentration will accumulate during high firing frequencies, so that also IAHP increases with firing frequency. In this way the CaL/IAHP-mechanism flattens the I/O curves, as has been well described in earlier modeling studies (e. g. [83]). Without this regulatory mechanism, the firing frequency (as resulting from the Na and Kdr channels) was generally too high, and the I/O curves were too steep compared to the data in Figure 2A (results not shown). With parameters as in Table 1, the I/O curves of P1 and P2 agreed well with the data sets, being steeper in P1, as the conductance values for both these channels (gCaL and gAHP) were higher in P2 (Figure 5B). In order to investigate the impact of IAHP, we interchanged the conductance values (gAHP) between P1 and P2, leaving all other parameters at their original values. The high gAHP resulted in a lower spike frequency in P1 (Figure 6B1), and in a threefold reduction in the slope of the I/O-curve (from the original ∼0. 8 spikes/pA to ∼0. 25 spikes/pA (results not shown) ). Correspondingly, the low gAHP increased the spiking frequency in P2 (Figure 6B2), and made the I/O curve three times steeper (from the original ∼0. 3 spikes/pA to ∼0. 9 spikes/pA). The initial bursts for depolarizing current injections were well reproduced by the model (Figure 4A), and depended mainly on CaT. The Ih conductance had negligible impact on these initial bursts (results not shown), but the bursts vanished when gCaT was set to zero (Figure 5C), and became stronger when gCaT was increased (Figure 5D). Despite P1 having the highest gCaT, the initial bursts were most pronounced in P2. This may seem counterintuitive, but we found that this is mainly due to differences in resting membrane potential. The lower resting potential in P2 means less inactivation of CaT at rest, and compensates for the lower gCaT. Note that our CaT kinetics (Figure 1D) was shifted +8 mV compared to the empirical data set for rats [44]. Using the CaT activation from that empirical study would thus give even stronger bursts (Figure 2A). Due to the high intra-burst firing frequencies, initial bursts gave rise to strong IAHP activation, resulting in a period of afterhyperpolarization which distinctly separated initial bursts from the regular AP firing. The afterhyperpolarization was most pronounced in P2, as P2 had the higher gAHP (Figure 4A), but became most pronounced in P1 when gAHP was interchanged between the two parameterizations (Figure 6B). Previous experiments have shown that some INs respond to depolarizing input by periodic bursting [42], [52]. Although periodic bursting was not observed in our experiments (IN1 and IN2), we ran test simulations to see if the mechanism that explained the initial bursts and subsequent afterhyperpolarization in IN1 and IN2 also could give rise to periodic bursting. Using the parameter sets P1 and P2 as a starting point, an increase in gCaT (by a factor 2) increased the intensity of the initial burst, but also the overall AP firing frequency, suggesting a nonzero CaT-activity throughout the stimulus period. In P2, the increased gCaT also gave rise to a periodic firing of pairs of APs (Figure 6D2), indicating a periodic interplay between CaT-driven bursts and IAHP-driven afterhyperpolarizations. By also increasing gAHP (by a factor 2), both the afterhyperpolarization (triggering the bursts) and the bursts (triggering the afterhyperpolarization) became more intense, and periodic bursting was obtained in both P1 and P2 (Figure 6E). The model thus predicted that the interplay between CaT and IAHP could explain the periodic bursting observed in some INs [42], [52], and that periodically bursting neurons have high CaT and IAHP conductances relative to neurons that do not show this behavior (e. g. IN1 and IN2). Small current injections (22 pA in P1 and 12 pA in P2) were used to shift the membrane potential from rest to the holding potentials of -57 mV in P1, and -58 mV in P2. When held at these depolarized potentials, strong hyperpolarizing current injections (-150 pA) were followed by a rebound burst in P1, but not in P2 (Figure 4B). In order to elicit bursts, a preceding hyperpolarization of the membrane potential is often required [36], [53]. However, the somatic current injections used in vitro to study this effect do not occur in vivo. For example, the -150 pA current injections used in our experiments gave rise to a membrane potential much more hyperpolarized than the typical GABAergic reversal potential. We therefore simulated a more realistic, synaptic GABAergic input, to investigate whether our model could elicit rebound bursts under more realistic conditions. We found that 50 synapses of moderate strength (maximum conductances of 1 nS), activated with 10 ms intervals over a time period of 300 ms, reproduced the essential findings from the current-clamp experiments. When held at normal resting potentials, no rebound burst was elicited in either of the parameterizations (Figure 7A). However, when held at -57 mV, the series of synaptic activations provoked a rebound burst in P1, but not in P2 (Figure 7B). Rebound bursts in various neurons are normally mediated by CaT and/or Ih channels [39], [64], [84]–[86]. We used the simulation setup with synaptic input to explore the contributions from these two ion channels to the rebound bursts in INs. In the original parameterizations (Table 1), P1 had higher values than P2 for both gCaT and gh, which explains why P1 and not P2 elicited rebound bursts. When gh was interchanged between P1 and P2, leaving all other parameters as in Table 1, the rebound burst in P1 was reduced to include only a single AP, while the (sub-threshold) rebound response in P2 became more pronounced (Figure 7C). Similar results were obtained when gCaT was interchanged between the two parameterizations: The rebound response in P1 fell below the AP firing threshold, while the (sub-threshold) rebound response in P2 became stronger (Figure 7D). Finally, if both parameters (gh and gCaT) were interchanged between P1 and P2, also the rebound responses were entirely interchanged: P1 showed a small sub-threshold rebound response, while P2 elicited a pronounced rebound burst, resembling that originally seen in P1 (Figure 7E). This suggests that CaT and Ih are of comparable importance for rebound bursts in mouse INs. To investigate how sensitive our results are to the IN morphology, we ran test simulations using three additional IN morphologies with (a) a similar, (b) a smaller and (c) a larger total membrane area compared to the original morphology (o). We replaced the original morphology (o) with the new morphologies (a–c), but kept all other model parameters fixed at the values in Table 1. As the ion channel densities (i. e. conductances per µm2) were kept fixed, the new morphologies (a–c) corresponded to INs with (a) a similar, (b) a higher, and (c) lower input resistance compared to the original morphology (o). The AP shape was not strongly affected by the morphology changes, except from small variations in the afterdepolarization (Figure 5A). The I/O curve in cases (o) and (a) nearly coincided (Figure 5B). As expected from the differences in total input resistance, morphology (b) gave rise to a I/O curve that was steeper and shifted in the negative direction along the current axis, whereas morphology (c) gave rise to an I/O curve that was flatter and shifted in the positive direction along the current axis compared to the cases (o) and (a) (Figure 5B). Although occurring at different current input levels, the characteristic responses of P1 and P2 did not change qualitatively with morphology. Depolarizing current injections of sufficient intensity gave rise to initial bursts followed by regular AP firing, and with the initial bursts being most pronounced for the parameterization P2. Furthermore, strong hyperpolarizing current injections (-150 pA in case of morphology (o), (a) or (b) and -200 pA in case of morphology (c) ) were followed by rebound bursts when using parameter set P1 at a holding potential of -57 mV, but not when using the parameter set P2 at a holding potential of -58 mV. The essential firing patterns of P1 and P2 were thus preserved under changes of morphology, and the I/O curve was always steeper for parameterization P1 (Figure 5C).
The set of seven ionic conductances that we used to fit the experimental findings (Table 1) is the same as in the previous model by Zhu et al. [36], but with kinetics that were updated to account for recent findings for activation/inactivation kinetics and somatodendritic distributions, and with conductance values constrained by a broader range of I/O data. Although these seven channel types were successful in reproducing the observed spiking patterns, we cannot exclude the presence of additional mechanisms, either overlapping with an included conductance type in terms of their action on the firing properties, or with a minor or negligible effect on the somatic response observed in the current clamp experiments. Several of the included ion channels have well documented roles in INs, including the role of CaT in burst generation [36], [39], [44], the role of Ih in generating initial sags [41] and the role of CaL in increasing the intracellular calcium concentration [51], [53]. ICAN is involved in making the dendrites leakier in connection with cholinergic modulation [52]. Its influence on the somatic response pattern of INs is less clear, although a previous modeling study has suggested that ICAN may be involved in generating plateau potentials and prolong bursts [36]. Due to the high calcium sensitivity of this channel [36], [88], we found that even small depolarizations of the membrane gave rise to a tonically active ICAN. In our model, the main function of ICAN was to reduce the magnitude of the depolarizing current required for the neuron to reach AP firing threshold. In comparison to somatic voltage recordings from rat INs [39], [41], our recordings from mouse INs show more pronounced initial sags for strong hyperpolarizing current injections (see Figure 2A, -150 pA stimuli). This suggests that the kinetics of Ih may differ between INs in rats and mice, as is the case in CA1 pyramidal neurons [89]. This we confirmed by measuring the voltage dependence of Ih in three mouse INs (Figure 3). Simulations with Ih kinetics based on our own measurements not only agreed better with the sag-shape in the mouse INs, but also predicted that the impact of Ih on rebound burst generation was comparable to that of CaT (Figure 6). This differs from the situation in rat INs, where rebound responses are mainly due to CaT, with no measurable contribution from Ih [39]. Conversely, in CA1 pyramidal neurons it has been shown that rebound spiking can be generated by Ih alone [86]. However, a joint involvement of CaT and Ih in burst generation, were found to underlie intrinsic rhythmic bursting in subpopulations of TCs during inattentiveness in guinea pigs [64], [84]. Intrinsic rhythmic bursting was not observed in our neurons. The presence of CaL-conductances in IN dendrites [51], [53], [54] makes it likely that also inhibitory, calcium-dependent mechanisms (such as IAHP) are present. The role of IAHP in INs has not been previously documented. Our model predicted that the interplay between CaL and IAHP conductances was sufficient for explaining the modulation of the I/O curves, although additional mechanisms could be involved. For example, a slowly activating potassium channel (KM) with a high threshold and no inactivation gave similar results to our CaL and IAHP mechanism, but without any calcium dependence (simulations were made using KM kinetics taken from [90], results not shown). However, no clear functional role could be assigned to KM other than that covered by IAHP. We thus did not explore this issue further, since IAHP gave a better agreement than KM with the time course of the intra-spike membrane potential and, especially, with the afterhyperpolarization following the initial bursts in P2. However, the possibility that KM and IAHP have overlapping functions in regulating the spiking frequency cannot be excluded. This could be experimentally tested by blocking the respective channels. We presented two parameterizations of the model (P1 and P2), which reproduced the electrophysiological properties of two different INs (IN1 and IN2). Rebound bursts as those generally observed only in a small subset of INs [39] were elicited by P1 but not P2. On the other hand, P2 elicited more pronounced initial bursts than P1 when exposed to depolarizing stimuli. P2 also had a less steep I/O curve and required weaker depolarization than P1 in order to reach AP-firing threshold. Our simulations showed that differences between P1 and P2 in terms of response properties and preferred input conditions arose from relative differences in specific conductances. Our model thus supports the idea that conductances values (i. e. channel density) in different subgroups of INs may be tuned in such a way as to optimize network operation under different input conditions. Under in vivo conditions, changes in input conditions (e. g. in synaptic input and/or shifts in membrane potential) may be mediated by mGlu5-receptor activation, GABAergic input from the reticular nucleus, or cholinergic modulation [4], [53], [54], [91], [92]. Data on the somatic voltage responses to somatic current injections do not uniquely determine the distribution of passive and active properties in the dendrites (see e. g. [48], [93]). Assumptions on the somato-dendritic distributions of CaT and CaL in our model were therefore based on calcium imaging data [50], [51]. This may be very useful in future studies, as the specific sub-cellular localization of different types of calcium channels may be particularly important for their specific functional role [43]. The assumed distributions of the remaining ion-channels were based on what we judged to be the most relevant literature to date. With the assumption that dendritic Na and Kdr densities were 10% of those in the soma, simulations showed that APs got broader during propagation in the dendrites, whereas the amplitude did not get significantly attenuated (results not shown). This is, at least at a qualitative level, in good agreement with recent experiments using voltage-sensitive dye [70]. We cannot exclude the possibility that important aspects of dendritic signaling are not captured by our model at this stage. For instance, there is an uncertainty regarding the distribution of our dendritic CaT channels. We based our CaT distribution on electrophysiological data [50], which indicated that the CaT density increases with distance to soma, while other, anatomy-based studies, have indicated a uniform distribution of dendritic CaT channels [94]. In test simulations we showed that we could obtain essentially the same results as in Figure 4 and Figure 5 also with a uniform CaT distribution, simply by rescaling the total CaT conductances (results not shown). However, although it may not be crucial for an IN' s response to somatic current injections, the CaT distribution will likely influence aspects of dendritic signaling, such as the probability for dendritic GABA-release. A related source of uncertainty concerns the presence of dendritic A-type potassium channels (KA), which, suggested by early studies, counteracted the CaT-channels in the dendrites of INs, and suppressed bursting [38]. This mechanism has, however, not been observed consistently, and several experiments have reported bursting INs [36], [39], [40], [43]. For thalamic reticular neurons, the ability versus inability to burst was rather explained by a varying density of CaT-channels [95], as also fits well with our findings. KA channels are often most densely present in distal dendrites and might have an influence on backpropagating action potentials [96]–[98]. However, recent experiments on INs found that attenuation of dendritic Ca-signals was relatively small [53]. As the importance of KA is unclear, these channels were not included in our model. Test simulations, using a moderate KA density (channel kinetics from [97]) in the dendrites, did not affect the somatic response to somatic stimuli significantly (results not shown). Such channels could therefore be readily added to the model if future experiments identify a clear functional role of KA in IN dendrites. Parts of the distal dendrites of INs form so called triadic synapses with axons from retinal ganglion cells and dendrites from TC neurons [4], [99]. In these triads, the IN terminals are, at the same time, postsynaptic to retinal input, and presynaptic to TC neurons. The conditions for GABAergic release from IN dendrites are not fully known, but may depend on intracellular calcium levels which are elevated by (backpropagating) sodium spikes as well as signals evoked by local synapses [4], [53], [54], [99]. It is known that cholinergic modulation reduces the membrane resistance of INs, through the M2-receptor mediated activation of Ih, ICAN and a linear, unspecified potassium current [52]. During sleep, when the cholinergic tone tends to be low, interneurons are likely to be electronically compact. INs may then provide long range inhibition, as synaptic input at one location in the dendritic tree then result in GABA release throughout the dendritic and axonal arbors. During awake states, when the cholinergic tone is high, distal dendritic regions may become electronically isolated from each other and from the soma. During these conditions, the triads may function as independent units, being excited by the presynaptic retinal afferents and then directly inhibiting only the postsynaptic dendrites of TC neurons [33], [52], [100], [101]. In our model, somatic action potentials successfully invaded distal dendritic region, and it is thus likely that our parameterizations (P1 and P2), correspond to conditions with a low cholinergic tone. The present model includes dendritic sodium-, potassium-, and calcium channels, and at least some of the mechanisms that are affected by cholinergic modulators. It computes the time course of the intracellular calcium levels in each compartment, and contains essential mechanisms for addressing signaling in IN dendrites on a fine spatiotemporal scale. In a network model including the dLGN circuitry, this will be of paramount importance for simulating the interactions between TCs and INs, as well as inputs to these cells from retina, cortex, thalamic reticular nucleus and possibly modulatory input from the brain stem. | The dorsal lateral geniculate nucleus (dLGN) is a part of the visual thalamus. This region contains two types of neurons: thalamocortical neurons and local interneurons. Thalamocortical neurons receive information from the retina and transmit information to visual cortex. The interneurons regulate the activity of thalamocortical neurons through inhibitory connections. This regulation is not properly understood, but it is believed to promote contrast enhancement and other vital visual functions. A powerful tool for development of a mechanistic understanding of dLGN functions is computer models that include the involved neurons, their interconnections and their interactions. Quite sophisticated models are available for thalamocortical neurons, but previous interneuron models are too simple for adequate mechanistic understanding of the functional properties of interneurons. We here present a detailed compartmental interneuron-model based on experimental data. The typical response patterns vary between different interneurons, but also within a given neuron, depending on the stimulus it receives. The model identifies a set of ionic mechanisms that can explain this diversity of activity patterns. In addition to being a useful building block for future network simulations of the dLGN, the model gives useful insight into the operating principles of dLGN interneurons. | Abstract
Introduction
Methods
Results
Discussion | systems biology
visual system
computer science
theoretical biology
computational neuroscience
single neuron function
biology
computational biology
sensory systems
neuroscience
genetics and genomics | 2011 | A Multi-Compartment Model for Interneurons in the Dorsal Lateral Geniculate Nucleus | 9,429 | 298 |
The microbial communities that inhabit the distal gut of humans and other mammals exhibit large inter-individual variation. While host genetics is a known factor that influences gut microbiota composition, the mechanisms underlying this variation remain largely unknown. Bile acids (BAs) are hormones that are produced by the host and chemically modified by gut bacteria. BAs serve as environmental cues and nutrients to microbes, but they can also have antibacterial effects. We hypothesized that host genetic variation in BA metabolism and homeostasis influence gut microbiota composition. To address this, we used the Diversity Outbred (DO) stock, a population of genetically distinct mice derived from eight founder strains. We characterized the fecal microbiota composition and plasma and cecal BA profiles from 400 DO mice maintained on a high-fat high-sucrose diet for ~22 weeks. Using quantitative trait locus (QTL) analysis, we identified several genomic regions associated with variations in both bacterial and BA profiles. Notably, we found overlapping QTL for Turicibacter sp. and plasma cholic acid, which mapped to a locus containing the gene for the ileal bile acid transporter, Slc10a2. Mediation analysis and subsequent follow-up validation experiments suggest that differences in Slc10a2 gene expression associated with the different strains influences levels of both traits and revealed novel interactions between Turicibacter and BAs. This work illustrates how systems genetics can be utilized to generate testable hypotheses and provide insight into host-microbe interactions.
The intestinal microbiota has profound effects on host physiology and health [1–3]. The composition of the gut microbiota is governed by a combination of environmental factors, including diet, drugs, maternal seeding, cohabitation, and host genetics [4–7]. Together, these factors cause substantial inter-individual variation in microbiota composition and modulate disease risk [8,9]. Alterations in the composition of the microbiota are associated with a spectrum of cognitive, inflammatory and metabolic disorders [10–12], and a number of bacterial taxa have been causally linked with modulation of disease [13–15]. A major challenge in the field is deciphering how host genetics and environmental factors interact to shape the composition of the gut microbiota. This knowledge is key for designing strategies aimed at modifying gut microbiota composition to improve health outcomes. Several mouse and human studies have examined the role of host genetics in shaping the composition of the gut microbiota [16]. Mouse studies comparing gut bacterial communities from inbred mouse strains [17,18] and strains harboring mutations in immune-related genes [19–22] support this notion. Additionally, quantitative trait locus (QTL) analyses in mice have identified genetic regions associated with the abundance of several bacterial taxa and community structure [23–26]. Twin studies and genome-wide association studies (GWAS) in humans have identified heritable bacterial taxa and SNPs associated with specific gut microbes. While comparing these studies is often difficult due to differences in environmental variables among populations, some associations are consistently detected among geographically discrete populations, such as the association between Bifidobacterium abundance and the lactase (LCT) gene locus [27–29], indicating the abundance of specific taxa is influenced by host genetic variation. Gut microbes and the host communicate through the production and modification of metabolites, many of which impact host physiology [30–34]. Bile Acids (BAs) are host-derived and microbial-modified metabolites that regulate both the gut microbiome and host metabolism [35–37]. BAs are synthesized in the liver from cholesterol, stored in the gallbladder and are secreted in the proximal small intestine where they facilitate absorption of fat-soluble vitamins and lipids. Once in the intestine, BAs can be metabolized by gut bacteria through different reactions, including deconjugation, dehydroxylation, epimerization, and dehydrogenation, to produce secondary BAs with differential effects on the host [33,35]. In addition to their direct effects on the host, BAs shape the gut microbiota composition through antimicrobial activities [38,39]. The detergent properties of BAs cause plasma membrane damage. The bactericidal activity of a BA molecule corresponds to its hydrophobicity [40]. Additionally, the microbiota modulates primary BA synthesis through regulation of the nuclear factor FXR [41]. Thus, we hypothesized that host genetic variation associated with changes in BA homeostasis mediates alterations in gut microbiota composition. To investigate how genetic variation affects gut microbiota and BA profiles, we used the Diversity Outbred (DO) mouse population, which is a heterogenous population derived from eight founder strains: C57BL6/J (B6), A/J (A/J), 1291/SvImJ (129), NOD/ShiLtJ (NOD), NZO/HiLtJ (NZO), CAST/EiJ (CAST), PWK/PhJ (PWK), and WSB/EiJ (WSB) [42,43]. These eight strains capture a large breadth of the genetic diversity found in inbred mouse strains. Additionally, the founder strains harbor distinct gut microbial communities and exhibit disparate metabolic responses to diet-induced metabolic disease [18,44,45]. The DO population is maintained by an outbreeding strategy aimed at maximizing the heterozygosity of the outbred stock. The genetic diversity and large number of generations of outbreeding make it an ideal resource for high-resolution genetic mapping of microbial and metabolic traits [43]. We characterized the intestinal microbiota composition and plasma and cecal BA profiles in ~400 genetically distinct DO mice fed a high-fat/high-sucrose diet for ~22 weeks and performed quantitative trait loci (QTL) analysis to identify host genetic loci associated with these traits. Specifically, we focused our analysis on potentially pleiotropic loci, which we defined as a single genetic locus that associates with both bacterial and BA traits. Our analysis revealed several instances of bacterial and metabolite traits attributed to the same DO founder haplotypes mapping to the same position of the mouse genome, including a locus associated with plasma BA levels and the disease-modulating organism Akkermansia muciniphila. Additionally, we identified the ileal BA transporter Slc10a2 as a candidate gene that regulates both the abundance of Turicibacter sp. and plasma levels of cholic acid.
We investigated the impact of genetic variation on gut microbiota composition and bile acid (BA) profiles using a cohort of ~400 DO mice maintained on a high-fat high-sucrose diet (45% kcal from fat and 34% from sucrose) for ~22 weeks (range 21–25 weeks), starting at weaning. We previously showed that this diet elicits a wide range of metabolic responses in the eight founder strains that are associated with microbiome changes [18,46]. Furthermore, we incorporated in our analyses previously published clinical weight traits collected from the same DO mice [47]. All animals were individually housed throughout the duration of the study to measure food intake and minimize microbial exchange. We performed LC-MS/MS analyses of plasma and cecal contents to assess relative variation in the levels of 27 BAs. Both plasma and cecal bile acids were measured to provide a comprehensive picture of systemic BA homeostasis. There was substantial variation in the plasma and cecal BA profiles across the 384 mice (Fig 1A and 1B; S1 Table). Additionally, we examined gut microbiota composition (n = 399) using 16S rRNA gene amplicon sequencing of DNA extracted from fecal samples collected at the end of the experiment. Within the cohort, there were 907 unique Exact Sequence Variants (ESVs), (100% operational taxonomic units defined with dada2 [48]), which were agglomerated into 151 lower taxonomic rankings (genus, family, order, class, phyla). The microbial traits represented each of the major phyla found in the intestine and the relative abundance of these phyla was highly variable among the DO mice (Fig 1C). For instance, the abundance of taxa classified to the Bacteroidetes phylum ranged from 1. 17–89. 28%. For subsequent analysis, we identified a core measurable microbiota (CMM), which we defined as taxon found in at least 20% of the mice [24]. This was done to remove the effects of excessive variation in the data due to bacterial taxa that were low abundance and/or sparsely distributed. In total, the CMM was comprised of 86 ESVs and 42 agglomerated taxa (S2 Table). The CMM traits represent a small fraction of the total microbes detected, but account for 94. 5% of the rarefied sequence reads, and therefore constitute a significant portion of the identifiable microbiota. Since mice were received in cohorts (i. e. , waves) of 100, we examined whether animals in each wave were more similar to each other than mice in other waves. The fecal microbiota composition significantly clustered by wave (p < 0. 001, PERMANOVA) and sex (p < 0. 001, PERMANOVA) (S1 Fig). PCA analysis of plasma and cecal bile acids showed a significant effect of sex, but not wave, on both plasma (p < 0. 0001, Kruskal Wallis) and cecal BA profiles (p < 0. 05, Kruskal Wallis) (S2 Fig). There is substantial evidence implicating gut microbiota and BAs in metabolic disease development [36,37]. To identify potential relationships among these traits, we performed correlation analysis which yielded many significant associations after FDR correction (FDR < 0. 05) (S3 Table, discussed in S1 Data). To identify associations between regions of the mouse genome and the clinical and molecular traits discussed above, we performed QTL analysis using the R/qtl2 package [49]. We used sex, days on the diet, and experimental wave as covariates. We identified 13 significant QTL (LOD ≥ 7. 66; P ≤ 0. 05) and 50 suggestive QTL (LOD ≥ 6. 80; P ≤ 0. 2) for bacterial [36], bile acid [13], and body weight [1] traits (Fig 2, S4 Table). Of the microbial QTL, we found 23 QTL for 17 distinct bacterial ESVs from the Bacteroidetes and Firmicutes phyla that met the LOD ≥ 6. 80 threshold. ESVs with the strongest QTL (LOD > 8) are classified to the Clostridiales order and map on chr 12 at ~33 Mbp, the Lachnospiraceae family on chr 2 at 164 Mbp, and the S24-7 family on chr 2 at ~115 Mbp. We also identified 12 QTL for microbial taxa collapsed by taxonomic assignment (i. e. , genus to phylum). The genera Lactococcus and Oscillospira were also associated with host genetic variation, which is consistent with previous studies [23,24,50,51]. Similarly, BA QTL mapped to multiple loci spanning the mouse genome and most BA traits mapped to multiple positions. BA synthesis and metabolism are regulated by multiple host signaling pathways: there are >17 known host enzymes involved in the production of BAs [36], transporters, which play a critical role in maintaining the enterohepatic circulation and BA homeostasis, and receptors that respond to BA in a variety of host tissues [52–54]. Therefore, it is not surprising that our results indicate that BA levels are polygenic and shaped by multiple host factors. To identify instances of overlapping QTL, we applied a less stringent threshold of LOD ≥ 6. 1 (P < 0. 5). We observed multiple instances of related BA species associating to the same genetic locus, indicating the presence of pleiotropic loci. Interestingly, several of these loci associate with levels of related BA species in different stages of microbial modification. For example, cecal taurocholic acid (TCA) and plasma CA QTL overlap on chr 7 at 122 Mbp. Likewise, QTL for plasma TDCA and cecal DCA, overlap on chr 12 between ~99–104 Mbp. For the cecal DCA, the WSB founder haplotype was associated with higher levels of this BA, while the NOD founder haplotype was associated with lower levels. The opposite pattern was observed for plasma TDCA, where the NOD and WSB haplotype were associated with higher and lower levels, respectively (S3A and S3B Fig). We also identified overlapping QTLs on chr 11 at ~71 Mbp for cecal levels of the secondary BAs lithocholic acid (LCA) and isolithocholic acid (ILCA), the isomer of LCA produced by bacterial epimerization (S3C Fig). Higher levels of these cecal BAs are associated with the 129 founder haplotype and lower levels are associated with the A/J founder haplotype (S3D and S3E Fig). We identified the positional candidate gene Slc13a5 (S3F Fig), which is a sodium-dependent transporter that mediates cellular uptake of citrate, an important precursor in the biosynthesis of fatty acids and cholesterol [55]. Recent evidence indicates that Slc13a5 influences host metabolism and energy homeostasis [56–58]. Slc13a5 is a transcriptional target of pregnane X receptor (PXR) [59], which also regulates the expression of genes involved in the biosynthesis, transport, and metabolism of BAs [60]. We searched for regions of the chromosome that were associated with both BA and bacterial abundance, as this may provide evidence of interactions between the traits [61]. We identified 17 instances of overlapping microbial and BA QTL on 12 chromosomes (LOD ≥ 6. 1; P ≤ 0. 5). This QTL overlap indicates there might be QTL with pleiotropic effects on BAs and the microbiota, suggest that genetic variation influencing host BA profiles has an effect on compositional features of the gut microbiota, or genetic-driven variation in microbiota composition alters BAs. Examples of notable instances of overlapping bacterial and BA QTL, including Akkermansia muciniphila and Peptostreptococcaceae family are discussed in the Supporting Information (S1 Data). We focused our co-mapping analysis on chr 8 at ~ 5. 5 Mbp, where Turicibacter sp. QTL and plasma cholic acid (CA) QTL overlap (Fig 3A and 3B). These traits were particularly interesting because both have been shown to be influenced by host genetics by previous studies. Turicibacter has been identified as highly heritable in both mouse and human genetic studies [24,27,45,50], and multiple reports have found differences in CA levels as a function of host genotype [18,46]. Furthermore, CA levels are influenced by both host genetics and microbial metabolism since it is synthesized by host liver enzymes from cholesterol and subsequently modified by gut microbes in the intestine. Notably, these co-mapping traits also share the same allele effects pattern, where the A/J and WSB haplotypes have strong positive and negative associations, respectively (Fig 3C and 3D). To assess whether the trait patterns observed in the DO founder strains correspond to the observed allelic effects in the QTL mapping, we performed a separate characterization of the fecal microbiota composition and plasma bile acids in age-matched A/J and WSB animals fed the HF/HS diet. The founder strain allele patterns inferred from the QTL mapping closely resembled the observed levels of Turicibacter sp. (Fig 3E) and plasma CA in the founder strains (Fig 3F), where A/J animals had significantly higher levels of Turicibacter sp. and CA than WSB animals. However, Turicibacter levels in the founder strains do not completely mirror the estimated allele effects. This may be due to other genetic factors that also influence Turicibacter levels, as this taxa may be influenced by multiple host genes and levels of Turicibacter have previously been associated on chr 7 [24], 9 and 11 [50] in mice. Furthermore, Turicibacter and plasma CA were positively correlated in the DO mice (r = 0. 43, p = 3. 53e-10). This finding is consistent with a previous study that found positive correlations between Turicibacter and unconjugated cecal BAs [62]. Taken together, the overlap between the Turicibacter sp. QTL and plasma CA QTL, along with the similar allele effects pattern, which reflect the values observed in the founder strains, provide strong evidence that these traits are related and they are responding to the common genetic driver. We searched in the QTL confidence interval for candidate genes via high-resolution association mapping on chr 8 and identified SNPs associated with both microbial and BA traits. Among these we identified SNPs upstream of the gene Slc10a2, which encodes for the apical sodium-bile transporter (Fig 3G). Slc10a2 is responsible for ~95% of BA reabsorption in the distal ileum and plays a key role in BA homeostasis [63]. In humans, mutations in this gene are responsible for primary BA malabsorption, resulting in interruption of enterohepatic circulation of BAs and decreased plasma cholesterol levels [64]. Likewise, Slc10a2-/- mice have a reduced total BA pool size, increased fecal BA concentrations and reduced total plasma cholesterol in comparison to wild-type mice [63]. Additionally, a comparison between germ-free and conventionally-raised mice found that expression of Slc10a2 is downregulated in presence of the gut microbiota, suggesting microbes may influence the expression of the transporter [41]. Our analysis identified SNPs associated with levels of Turicibacter sp. and plasma CA at the QTL peak (Fig 3G). The SNPs with the strongest associations were attributed to the WSB and A/J haplotypes and fell on intergenic regions near Slc10a2. There is growing evidence that non-coding intergenic SNPs are often located in or closely linked to regulatory regions, suggesting that they may influence host regulatory elements and alter gene expression [65,66]. To assess if candidate gene expression patterns in the DO founders corresponds to the estimated allelic effects in the QTL mapping, we quantified Slc10a2 expression in distal ileum samples from A/J and WSB mice by quantitative reverse transcriptase PCR (qRT-PCR). A/J mice exhibited significantly higher expression of Slc10a2 compared to WSB mice (Fig 3H), which is consistent with estimated allele patterns for the overlapping Turicibacter and plasma CA QTLs on chr 8 (Fig 3A and 3B). Remarkably, several studies have noted concomitant changes in microbiota composition and Slc10a2 mRNA levels [67–69]. We mapped QTL for Turicibacter sp. and for plasma CA levels to a common locus on chr 8 at 5–7 Mbp. Since the LOD profiles and allelic effects are highly similar, the QTL may be due to a single shared locus (pleiotropy) or multiple closely linked loci. We examined this question using a likelihood ratio testing of the null hypothesis of pleiotropy versus the alternative of two independent genetic regulators of these traits [70]. Analysis of 1000 bootstrap samples resulted in a p-value of 0. 531, which is consistent with the presence of a single pleiotropic locus that affects both traits. We next sought to understand the causal relationships between the microbe and the BA. We asked whether the relationship between the microbe and BA was causal, reactive or independent. To establish the directionality of the relationship, we applied mediation analysis where we conditioned one trait on the other [71]. When we conditioned Turicibacter sp. on plasma CA (QTL → BA → Microbe), we observed a LOD drop of 3. 2 (Fig 4A and 4B). Likewise, when we conditioned the plasma cholic acid on the microbe (QTL → Microbe → BA) there was a LOD drop of 3. 32 (Fig 4C and 4D). The partial mediation seen in both models suggests that the relationship between the microbe and the BA could be bidirectional, where they exert an effect on one another. From this analysis, we can hypothesize this relationship can be explained by a pleiotropic model, where a single locus influences a microbial and a BA trait, and the microbial trait is also reactive to changes in the BA trait. It is important to note that statistical inference only partially explains the relationship between the traits and there may be other hidden variables that may further explain the relationship. The complex relationship depicted by the causal inference testing is consistent with the interplay between gut microbes and BAs in the intestine and their known ability to influence the other. Due to the strong correlative relationship between the QTL, we tested whether there was a direct interaction between bile acids and Turicibacter. Turicibacter inhabits the small intestine where BAs are secreted upon consumption of a meal [72,73]. We screened the human isolate Turicibacter sanguinis for deconjugation and transformation activity in vitro by HPLC/MS-MS. We found that T. sanguinis deconjugated ~96–100% of taurocholic acid and glycochenodeoxycholic acid (Fig 5A) within 24 hours. It also transformed ~6 and 8% of CA and CDCA to 7-dHCA and 7-ketolithocholic acid (7-KLCA), respectively (Fig 5B and 5C). Both of these transformations require the action of the bacterial 7α-hydroxysteroid dehydrogenase. Based on these results, we asked if conjugated and unconjugated bile acids differentially modulate T. sanguinis growth. BA concentrations range from ~1–10 mM along the small intestine [74] to ~0. 2–1 mM in the cecum [75]. Therefore, we grew T. sanguinis in the presence of either conjugated or unconjugated bile acids at physiologically relevant concentrations ranging from 0. 1–5 mM. T. sanguinis growth decreased with increasing concentrations of BAs and growth was completely inhibited at 1 mM for unconjugated BAs and 5 mM for conjugated BAs (Fig 5D and 5E). Growth rate was significantly slower in the presence of 1 mM conjugated and 0. 5mM unconjugated bile acids (Fig 5F). These results suggest that levels of BAs may affect abundance of Turicibacter in the gut. To compare T. sanguinis sensitivity to conjugated bile acids relative to other small intestine colonizers, we grew four taxa (Bacteroides thetaiotaomicron, Clostridium asparagiforme, Lactobacillus reuteri and Escherichia coli MS200-1) known to colonize this region of the intestine with or without 1 mM conjugated bile acids. Members of these genera are known to have bile salt hydrolase (BSH) activity to deconjugate bile acids [35]. Unlike T. sanguinis, the addition of high levels of conjugated bile acids had little to no effect on the growth of these four gut microbes (S4 Fig). Consistent with these findings, Turicibacter abundance was negatively correlated with cecal TCA levels in the DO mice (r = -0. 262, p = 0. 0035). Taken together, these data indicate that T. sanguinis is sensitive to higher concentrations of BA compared to other small intestine colonizers. These reciprocal effects between the BA and the bacterium provide biological evidence for the correlative relationship shown by the causal model testing. In summary, using a genetic approach, we identified and provide validation of a relationship between a genetic locus containing the BA transporter Slc10a2, and levels of Turicibacter and plasma cholic acid. Based on our findings, we hypothesize that the identified locus regulates expression of Slc10a2, altering active BA reabsorption in the ileum, leading to increased intestinal BA concentrations and alterations in the intestinal BA environment. Consequently, the resulting environmental change provides an unfavorable habitat for Turicibacter. In turn, lower levels of Turicibacter BA deconjugation activity leads to a decrease in circulating free plasma cholic acid levels. In this study, we performed the first known genetic mapping integration of gut microbiome and BA profiles. Using DO mice, we identified multiple QTL for gut microbes and bile acids spanning the host genome. These included loci that associated with individual microbial and BA traits, as well as loci with potential pleiotropic effects, where a single genetic region influenced both the abundance of a gut microbe and levels of a BA. While several studies suggest that host genetic variation has a minor impact on microbiota composition, there are overlapping findings among different studies in both human and mouse populations that indicate that specific bacterial taxa are influenced by host genetics. Our results in the DO population corroborate several of these key findings (discussed in S1 Data). Turicibacter sp. is among the microbes consistently associated with host genetics. This work plus data from previous reports suggest that alterations in the BA pool driven by Slc10a2 genetic variation and concomitant changes in expression/activity elicit an impact on gut microbiota community structure and influence the ability of Turicibacter to colonize and persist in the intestine. Although this microbe deconjugates primary BAs, we found that it is also sensitive to elevated concentrations of both conjugated and unconjugated BAs. Future experiments are needed to examine how a decrease in Slc10a2 expression changes intestinal BA profiles and the consequences on Turicibacter colonization. Additionally, this work identified multiple host-microbe-metabolite interactions that need to be validated with additional molecular studies. More broadly, our work demonstrates the power of genetics to identify novel interactions between microbial and metabolite traits and provides new testable hypotheses to further dissect factors that shape gut microbiota composition.
Animal care and study protocols were approved by the University of Wisconsin-Madison Animal Care and Use Committee (A005821) and were in compliance with all NIH animal welfare guidelines. Animal care and study protocols were approved by the University of Wisconsin-Madison Animal Care and Use Committee. DO mice were obtained from the Jackson Laboratories (Bar Harbor, ME, USA) at ~4 weeks of age and maintained in the Department of Biochemistry vivarium at the University of Wisconsin-Madison. Mice were housed on a 12-hour light: dark cycle under temperature- and humidity-controlled conditions. Five waves of 100 DO mice each from generations, 17,18,19,21, and 23 were obtained at intervals of 3–6 months. Each wave was composed of equal numbers of male and female mice. All mice were fed a high-fat high-sucrose diet (TD. 08811, Envigo Teklad, 44. 6% kcal fat, 34% carbohydrate, and 17. 3% protein) ad libitum upon arrival to the facility. Mice were kept in the same vivarium room and were individually housed to monitor food intake and prevent coprophagy between animals. DO mice were sacrificed at 22–25 weeks of age. The eight DO founder strains (C57BL/6J, A/J, 129S1/SvImJ, NOD/ShiLtJ, NZO/HILtJ, PWK/PhJ, WSB/EiJ and CAST/EiJ) were obtained from the Jackson Laboratories. Mice were bred at the University of Wisconsin-Madison Biochemistry Department. Mice were housed by strain and sex (2–5 mice/cage), with the exception of CAST that required individual housing. Inbred founder mice were housed under the same environmental conditions as the DO animals. Like the DO mice, the eight founder strains were maintained on the HF/HS diet and were sacrificed at 22 weeks of age, except for NZO males that were sacrificed at 14 weeks, due to high mortality attributable to severe disease. For both DO and founder mice, fecal samples for 16S rRNA sequencing were collected immediately before sacrifice after a 4 hour fast. Cecal contents, plasma, and additional tissues were harvested promptly after sacrifice and all samples were immediately flash frozen in liquid nitrogen and stored at -80°C until further processing. DNA was isolated from feces using a bead-beating protocol [18]. Mouse feces (~1 pellet per animal) were re-suspended in a solution containing 500μl of extraction buffer [200mM Tris (pH 8. 0), 200mM NaCl, 20mM EDTA], 210μl of 20% SDS, 500μl phenol: chloroform: isoamyl alcohol (pH 7. 9,25: 24: 1) and 500μl of 0. 1-mm diameter zirconia/silica beads. Cells were mechanically disrupted using a bead beater (BioSpec Products, Barlesville, OK; maximum setting for 3 min at room temperature), followed by extraction with phenol: chloroform: isoamyl alcohol and precipitation with isopropanol. Contaminants were removed using QIAquick 96-well PCR Purification Kit (Qiagen, Germantown, MD, USA). Isolated DNA was eluted in 5 mM Tris/HCL (pH 8. 5) and was stored at -80°C until further use. PCR was performed using universal primers flanking the variable 4 (V4) region of the bacterial 16S rRNA gene [76]. Genomic DNA samples were amplified in duplicate. Each reaction contained 10–30 ng genomic DNA, 10 μM each primer, 12. 5 μl 2x HiFi HotStart ReadyMix (KAPA Biosystems, Wilmington, MA, USA), and water to a final reaction volume of 25 μl. PCR was carried out under the following conditions: initial denaturation for 3 min at 95°C, followed by 25 cycles of denaturation for 30 s at 95°C, annealing for 30 s at 55°C and elongation for 30 s at 72°C, and a final elongation step for 5 min at 72°C. PCR products were purified with the QIAquick 96-well PCR Purification Kit (Qiagen, Germantown, MD, USA) and quantified using Qubit dsDNA HS Assay kit (Invitrogen, Oregon, USA). Samples were equimolar pooled and sequenced by the University of Wisconsin–Madison Biotechnology Center with the MiSeq 2x250 v2 kit (Illumina, San Diego, CA, USA) using custom sequencing primers. Demultiplexed paired end fastq files generated by CASAVA (Illumina) and a mapping file were used as input files. Sequences were processed, quality filtered and analyzed with QIIME2 (version 2018. 4) (https: //qiime2. org), a plugin-based microbiome analysis platform [77]. DADA2 [48] was used to denoise sequencing reads with the q2-dada2 plugin for quality filtering and identification of de novo exact sequence variants (ESVs) (i. e. 100% exact sequence match). This resulted in 20,831,573 total sequences with an average of 52,078 sequences per sample for the DO mice, and 2,128,796 total sequences with an average of 34,335. 4 sequences per sample for the eight DO founder strains. Sequence variants were aligned with mafft [78] with the q2-alignment plugin. The q2-phylogeny plugin was used for phylogenetic reconstruction via FastTree [79]. Taxonomic classification was assigned using classify-sklearn [80] against the Greengenes 13_8 99% reference sequences [81]. Alpha- and beta-diversity (weighted and unweighted UniFrac [82] analyses were performed using q2-diversity plugin at a rarefaction depth of 10000 sequences per sample. For the DO mice, one sample (DO071) was removed from subsequent analysis because it did not reach this sequencing depth. For analysis of the eight DO founder strains, one sample (NOD5) was removed because it did not reach this sequencing depth. Subsequent processing and analysis were performed in R (v. 3. 5. 1), and data generated in QIIME2 was imported into R using Phyloseq [83]. Sequencing data was normalized by cumulative sum scaling (CSS) using MetagenomeSeq [84]. Summaries of the taxonomic distributions were generated by collapsing normalized ESV counts into higher taxonomic levels (genus to phylum) by phylogeny. We defined a core measurable microbiota (CMM) [24] to include only microbial traits present in 20% of individuals in the QTL mapping. In total, 86 ESVs and 42 collapsed microbial taxonomies comprised the CMM. 40 μL of DO plasma collected at sacrifice (30 μL used for founder strains) were aliquoted into a tube with 10 μL SPLASH Lipidomix internal standard mixture (Avanti Polar Lipids, Inc.). Protein was precipitated by addition of 215 μL MeOH. After the mixture was vortexed for 10 s, 750 μL methyl tert-butyl ether (MTBE) were added as extraction solvent and the mixture was vortexed for 10 s and mixed on an orbital shaker for 6 min. Phase separation was induced by adding 187. 5 μL of water followed by 20 s of vortexing. All steps were performed at 4°C on ice. Finally, the mixture was centrifuged for 4 min at 14,000 x g at 4°C and stored at -80°C. For targeted bile acids analysis, samples were thawed on ice. 400 μL of ethanol were added to further precipitate protein, as well as 15 μL of isotope-labeled internal standard mix (12. 5 μM d4-TαMCA, 10 μM d4-CDCA). The samples were vortexed for 20 s and centrifuged for 4 min at 14,000 g at 4°C after which the supernatant (ca. 1000 μL) was taken out and dried down. Dried supernatants were resuspended in 60 μL mobile phase (50%B), vortexed for 20 s, centrifuged for 4 min at 14,000 g and then 50 μL were transferred to vials with glass inserts for MS analysis. 30 ± 7. 5 mg cecal contents along with 10 μL SPLASH Lipidomix internal standard mixture were aliquoted into a tube with a metal bead and 270 μL MeOH were added for protein precipitation. To each tube, 900 μL MTBE and 225 μL of water were added as extraction solvents. All steps were performed at 4°C on ice. The mixture was homogenized by bead beating for 8 min at 25 Hz. Finally, the mixture was centrifuged for 4–8 min at 11,000 x g at 4°C. Subsequent processing for the DO mice and eight DO founder strains differed due to other analyses performed on the samples that are not presented in this paper. For DO samples, 100 μL of the aqueous and 720 μL of organic layer were combined and stored at -80°C. For analysis, these were thawed on ice and 400 μL of ethanol were added to further precipitate protein, as well as 15 μL of isotope-labeled internal standard mix (12. 5 μM d4-TαMCA, 10 μM d4-CDCA). The samples were vortexed for 20 s and centrifuged for 4 min at 14,000 g at 4°C after which the supernatant (ca. 1000 μL) was taken out and dried down. Dried supernatants were resuspended in 100 μL mobile phase (50%B), vortexed for 20 s, centrifuged for 8 min at 14,000 g and then 50 μL were transferred to vials with glass inserts for MS analysis. For the eight DO founder strains, the mixture was dried down including all solid parts and stored dried at -80°C. For targeted bile acid analysis, these dried down samples were then thawed on ice and reconstituted in 270 μL of methanol, 900 μL of MTBE, and 225 μL of water. 400 μL of ethanol were added to further precipitate protein, as well as 15 μL of isotope-labeled internal standard mix (12. 5 μM d4-TαMCA, 10 μM d4-CDCA). The mixture was bead beat for 8 min at 25 Hz and centrifuged at 14,000 g for 8 minutes after which the supernatant (ca. 1500 μL) was taken out and dried down. Dried supernatants were resuspended in 100 μL mobile phase (50%B), vortexed for 20 s, centrifuged for 4 min at 14,000 g and then 90 μL were transferred to vials with glass inserts for MS analysis. LC-MS analysis was performed in randomized order using an Acquity CSH C18 column held at 50°C (100 mm × 2. 1 mm × 1. 7 μm particle size; Waters) connected to an Ultimate 3000 Binary Pump (400 μL/min flow rate; Thermo Scientific). Mobile phase A consisted of 10 mM ammonium acetate containing 1 mL/L ammonium hydroxide. Mobile phase B consisted of MeOH with the same additives [85]. Mobile phase B was initially held at 50% for 1. 5 min and then increased to 70% over 13. 5 min. Mobile phase B was further increased to 99% over 0. 5 min and held for 2. 5 min. The column was re-equilibrated for 5. 5 min before the next injection. Twenty microliters of plasma sample or ten microliters of cecum sample were injected by an Ultimate 3000 autosampler (Thermo Scientific). The LC system was coupled to a TSQ Quantiva Triple Quadrupole mass spectrometer (Thermo Scientific) by a heated ESI source kept at 325°C (Thermo Scientific). The inlet capillary was kept at 350°C, sheath gas was set to 15 units, auxiliary gas to 10 units, and the negative spray voltage was set to 2,500 V. For targeted analysis the MS was operated in negative single reaction monitoring (SRM) mode acquiring scheduled, targeted scans to quantify selected bile acid transitions, with two transitions for each species’ precursor and 3 min retention time windows. Collision energies were optimized for each species and ranging from 20–55 V. Due to insufficient fragmentation for unconjugated bile acids, the precursor was monitored as one transition with a CE of 20 V. MS acquisition parameters were 0. 7 FWHM resolution for Q1 and Q3,1 s cycle time, 1. 5 mTorr CID gas and 3 s Chrom filter. In total, 27 bile acids, including 14 unconjugated, 9 tauro- and 4 glycine-conjugated species, were measured. The resulting bile acid data were processed using Skyline 3. 6. 0. 10493 (University of Washington). For each species, one transition was picked for quantitation, while the other was used for retention time confirmation. Normalization of the quantitative data was performed to the internal standard d4-CDCA as indicated in Eq 1. Genotyping was performed on tail biopsies as previously described [42] using the Mouse Universal Genotyping Array (GigaMUGA; 143,259 markers) [86] at Neogen (Lincoln, NE). Genotypes were converted to founder strain-haplotype reconstructions using a hidden Markov model (HMM) implemented in the R/qtl2 package [49]. We interpolated the GigaMUGA markers onto an evenly spaced grid with 0. 02-cM spacing and added markers to fill in regions with sparse physical representation, which resulted in 69,005 pseudomarkers. We performed QTL mapping using the R package R/qtl2 [49]. QTL mapping was done through a regression of the phenotype on the founder haplotype probabilities estimated with an HMM designed for multi-parental populations. Genome scans were performed for each phenotype with sex, cohort (wave), and days on diet included as additive covariates. Genetic similarity between mice was accounted for using a kinship matrix based on the leave-one-chromosome-out (LOCO) methods [87]. For microbial QTL mapping, normalized gut microbiota abundance data transformed to normal quantiles. For bile acid QTL mapping, normalized plasma and cecal bile acid levels were log2 transformed. The mapping statistic reported is the log10 likelihood ratio (LOD score). The QTL support interval was defined using the 95% Bayesian confidence interval. Significant and suggestive QTL were determined at a genome-wide threshold of P ≤ 0. 05 (LOD ≥ 7. 66) and P ≤ 0. 2 (LOD ≥ 6. 80), respectively. We used a common significance threshold for all phenotypes, by pooling the permutation results for the individual phenotypes. No adjustment was made for the search across multiple phenotypes. To assess whether two co-mapping traits were caused by a pleiotropic locus, we used a likelihood ratio test implemented with the open source R package R/qtl2pleio [70]. Here, we compared the alternative hypothesis of two distinct loci with the null hypothesis of pleiotropy for two traits that map to the same genetic region. Parametric bootstrapping was used to determine statistical significance. Mediation analysis was applied to identify whether a microbe or bile acid were likely to be a causal mediator of the QTL as presented in Li et al. [88]. This analysis was adapted from a general approach previously described to differentiate target from mediator variables [89]. The effect of a mediator on a target was evaluated by performing an allele scan or SNP scan using the target adjusted by mediator. Only individuals with both values for both traits were considered for mediation analysis. Traits with a LOD drop >2 after controlling for the mediator were considered for further causality testing. To statistically assess causality between microbial and bile acid trait sets (causal, reactive, independent, undecided), a causal model selection test [90] was applied using the R packages R/intermediate and R/qtl2. Causal model selection tests were evaluated on both alleles and SNPs in peak region. Total RNA was extracted from flash-frozen distal ileum tissues by TRIzol extraction and further cleaned using the RNeasy Mini Kit (Qiagen, Germantown, MD, USA). DNA was removed by on-column DNase digestion (Qiagen). Purified RNA was quantified using a Nanodrop 2000 spectrophotometer. SuperScript II Reverse Transcriptase with oligo (dT) primer (all from Invitrogen, Carlsbad, CA, USA) was used to synthesize 20 μl cDNA templates from 1 μg purified RNA. cDNA was diluted 2X before use and qRT-PCR reactions were prepared in a 10 μl volume using SsoAdvanced Universal SYBR Green Supermix (Bio-Rad, Hercules, CA, USA) and 400 nM specific primers targeting the gene of interest (SLC10A2-F [5’- TGGGTTTCTTCCTGGCTAGACT-3’]; SLC10A2-R [5’- TGTTCTGCATTCCAGTTTCCAA-3’] [91]). All reactions were performed in triplicate. Reactions were run on a CFX96 Real-Time PCR System (Bio-Rad, Hercules, CA, USA). The 2-ΔΔCt method [92] was used to calculate relative changes in gene expression and all results were normalized to GAPDH. Bacterial strains were obtained from DSMZ and ATCC. All strains were cultured at 37°C under anaerobic conditions using an anaerobic chamber (Coy Laboratory Products) with a gas mix of 5% hydrogen, 20% carbon dioxide and 75% nitrogen. Strains were grown in rich medium (S5 Table) that was filter sterilized and stored in the anaerobic chamber at least 24 hours prior to use. L. reuteri was grown in medium supplemented with 20 mM glucose. For all in vitro assays, cultures used for inoculation were grown overnight at 37°C in 10 mL 14b medium in anaerobic Hungate tubes. Stock solutions of conjugated bile acids (TCA, GCDCA) and unconjugated bile acids (CA, CDCA, DCA) were prepared to a final concentration of 100 mM and used for all in vitro assays. All bile acids used were soluble in methanol. Stock solutions of conjugated and unconjugated bile acids (100 mM) were added to 3 ml 14b medium to obtain a final concentration of 100 μM total bile acid. Tubes were inoculated with a T. sanguinis cultured overnight, then incubated in the anaerobic chamber at 37°C for 48 hours. At the 24- and 48-hour timepoints, 1 mL of each culture was removed and the supernatant was collected after brief centrifugation. Each culture supernant was diluted 10x in initial running solvent (30: 70 MeOH: 10 mM ammonium acetate). Samples were spun at max speed for 3 minutes to remove suspended particles prior to loading on the uHPLC. Samples were analyzed using a uHPLC coupled with a high-resolution mass spectrometer. 10 μL aliquots of diluted supernatant samples were analyzed using a uHPLC-MS/MS system consisting of a Vanquish uHPLC coupled by electrospray ionization (ESI) (negative mode) to a hybrid quadrupole-high-resolution mass spectrometer (Q Exactive Orbitrap; Thermo Scientific). Liquid chromatography separation was achieved on an Acquity UPLC BEH C18 column (2. 1-by 100-mm column, 1. 7-μm particle size) heated to 50°C. Solvent A was 10 mM Ammonium acetate, pH 6; solvent B was 100% methanol. The total run time was 31. 5 minutes with the following gradient: 0 min, 30% B; 0. 5 min, 30% B; 24 min, 100% B; 29 min, 100% B; 29 min, 30% B; 31. 5 min, 30% B. Bile acid peaks were identified using the Metabolomics Analysis and Visualization Engine (MAVEN) [93]. Bacterial growth rate was measured in medium 14b supplemented with either 100 μM, 300 μM, 1 mM bile acids or methanol control. Medium was dispensed inside an anerobic chamber into Hungate tubes. Tubes containing 10 mL of medium were inoculated with 30 μL of an overnight culture and incubated at 37°C for 24 hours. T. sanguinis was grown with shaking to disrupt the formation of flocculent colonies. Growth was monitored as the increase in absorbance at 600 nm in a Spectronic 20D+ spectrophotometer (Thermo Scientific, Waltham, MA, USA). Growth rate was determined as μ = ln (X/Xo) /T, where X is the OD600 value during the linear portion of growth and T is time in hours. Values given are the mean μ values from two independent cultures done in triplicate. All statistical analyses were performed in R (v. 3. 5. 1) [94]. Unless otherwise indicated in the figure legends, differences between groups were evaluated using unpaired two-tailed Welch’s t-test. For multiple comparisons, Krustkal-Wallis test was used if ANOVA conditions were not met, followed by Mann-Whitney/Wilcoxon rank-sum for multiple comparisons and adjusted for multiple testing using the Benjamini-Hochberg FDR procedure. The correlation between the abundance of microbial taxa was performed using Spearman’s correlation in the “Hmisc” (v. 4. 1–1) R package [95]. The p-values were adjusted using the Benjamini and Hochberg method, and correlation coefficients were visualized using the “pheatmap” (v. 1. 0. 10) [96]. Multiple groups were compared by Kruskal-Wallis test and adjusted for multiple testing using the Benjamini-Hochberg FDR procedure. Significance was determined as p-value < 0. 05. To assess magnitude of variability of the CMMs, summary statistics were calculated on each CMM (taxa and ESVs). Non-parametric-based PERMANOVA statistical test [97] with 999 Monte Carlo permutations was used to compare microbiota compositions among groups using the Vegan R package [98]. | Inter-individual variation in the composition of the intestinal microbiota can in part be attributed to host genetics. However, the specific genes and genetic variants underlying differences in the microbiota remain largely unknown. To address this, we profiled the fecal microbiota composition of 400 genetically distinct mice, for which genotypic data is available. We identified many loci of the mouse genome associated with changes in abundance of bacterial taxa. One of these loci is also associated with changes in the abundance of plasma bile acids—metabolites generated by the host that influence both microbiota composition and host physiology. Follow up validation experiments provide mechanistic insights linking host genetic differences, with changes in ileum gene expression, bile acid-bacteria interactions and bile acid homeostasis. Together, this work demonstrates how genetic approaches can be used to generate testable hypothesis to yield novel insight into how host genetics shape gut microbiota composition. | Abstract
Introduction
Results and discussion
Materials and methods | medicine and health sciences
body fluids
gut bacteria
microbiome
quantitative trait loci
population genetics
bile
microbiology
genetic mapping
population biology
microbial genetics
bacteria
microbial genomics
genetic polymorphism
medical microbiology
genetic loci
haplotypes
anatomy
heredity
physiology
genetics
biology and life sciences
genomics
evolutionary biology
organisms | 2019 | Genetic determinants of gut microbiota composition and bile acid profiles in mice | 11,846 | 219 |
Monocytes and T-cells are critical to the host response to acute bacterial infection but monocytes are primarily viewed as amplifying the inflammatory signal. The mechanisms of cell death regulating T-cell numbers at sites of infection are incompletely characterized. T-cell death in cultures of peripheral blood mononuclear cells (PBMC) showed ‘classic’ features of apoptosis following exposure to pneumococci. Conversely, purified CD3+ T-cells cultured with pneumococci demonstrated necrosis with membrane permeabilization. The death of purified CD3+ T-cells was not inhibited by necrostatin, but required the bacterial toxin pneumolysin. Apoptosis of CD3+ T-cells in PBMC cultures required ‘classical’ CD14+ monocytes, which enhanced T-cell activation. CD3+ T-cell death was enhanced in HIV-seropositive individuals. Monocyte-mediated CD3+ T-cell apoptotic death was Fas-dependent both in vitro and in vivo. In the early stages of the T-cell dependent host response to pneumococci reduced Fas ligand mediated T-cell apoptosis was associated with decreased bacterial clearance in the lung and increased bacteremia. In summary monocytes converted pathogen-associated necrosis into Fas-dependent apoptosis and regulated levels of activated T-cells at sites of acute bacterial infection. These changes were associated with enhanced bacterial clearance in the lung and reduced levels of invasive pneumococcal disease.
Innate immunity is critical for the rapid recognition and response to pathogenic micro-organisms [1]. A complex relationship exists between innate immune responses and T-cells. Innate immune responses recruit and activate T-cells at sites of infection but T-cells in turn regulate phagocyte function and can therefore modify inflammatory responses. Monocytes are key effectors of the innate immune response to bacteria and contribute to recruitment of T-cells at sites of infection [2]. In contrast to differentiated macrophages, however, monocytes have not been viewed as having a major role in the downregulation of the inflammatory response [3]. Streptococcus pneumoniae is one of the leading causes of infection-related mortality globally [4]. T-cells are key to host defense against pneumococci, making this a useful model with which to study the regulation of T-cells during bacterial infection [5], [6]. CD4+ T-cells are found at sites of pneumococcal colonization in the upper airway [7] and T-cells migrate to sites of infection in the lung [8]. In murine models CD4+ T-cell Th17 responses facilitate clearance of colonizing bacteria [7], [9] while CD4+ T-cells enhance clearance of bacteria from the lungs [5]. Other studies have emphasised an important role for CD8+ T-cells during pneumococcal pneumonia by demonstrating CD8+ T-cells limit the extent of the inflammatory response [10]. Despite these observations, CD4+ T-cell inhibition may also limit inappropriate degrees of inflammation in some models of pneumococcal infection and improve disease outcome, emphasizing that numbers of T-cell populations must be carefully regulated to ensure effective clearance of bacteria while limiting lung pathology [10], [11]. There is limited information on how T-cell numbers are regulated during the immune response to pneumococci and in particular what role cell death plays in this process. Lymphocyte apoptosis has been observed in peripheral blood of patients with pneumococcal infection [12] and is a well-recognized feature of bacterial sepsis [13]. Nevertheless it remains unclear whether the lymphocyte apoptosis described during pneumococcal infection is part of a non-specific response, associated with microbial products and the altered cytokine responses that are a feature of infection, or whether it might be the result of a more specific host programme regulating the immune response. We therefore examined whether the interaction of T-cells with pneumococci results in cell death and have characterized features of this process. In particular we observed that the pneumococcal protein pneumolysin induces T-cell necrosis but that in the presence of monocytes T-cells undergo Fas-dependent apoptosis. Moreover we have found a requirement for Fas-signaling in the regulation of CD3+ T-cell death, during the early T-cell dependent phases of the host response to pneumococci.
We first examined whether peripheral blood mononuclear cells (PBMC) incubated with pneumococci demonstrated features of apoptosis, using a range of morphologic and biochemical assays of apoptosis. Almost all PBMC became Annexin V+ after 24 h of incubation with pneumococci (Figure 1A). PBMC also showed evidence of loss of inner mitochondrial transmembrane potential (Δψm) (Figure 1B) and of increased caspase activation (Figure 1C), cell shrinkage (Figure 1D) and DNA strand breaks (Figure 1E), following exposure to pneumococci. We confirmed cell death of purified lymphocytes 6 h after exposure to bacteria with evidence of both early apoptotic Annexin V+/TO-PRO-3− cells and late apoptotic/necrotic Annexin+/TO-PRO-3+ cells (data not shown), that was directly proportional to the MOI (Figure 1F). We also confirmed accumulation of hypodiploid DNA, a feature of apoptosis [14], in CD3+ T-cells (Figure 1G) and that the CD3+ T-cell death was caspase dependent (Figure 1H). The combination of features confirmed an apoptotic form of cell death. Apoptosis was apparent with an MOI as low as 0. 1 (Figure 1B). Cell death required cell contact between bacteria and PBMC and was not the consequence of utilization of growth factors in the media by bacteria since the insertion of transwells between bacteria and PBMC inhibited cell death at representative early and late time points after bacterial challenge (Figure S1). In PBMC cultures as expected T-cells predominated but cultures also contained 17. 5±1. 7% CD19+ B-cells and 8. 7±1. 1% CD14+ monocytes. CD19+ B-cells showed a trend towards an increased rate of cell death following 16 h pneumococcal challenge (Figure S2A). As we have previously reported during pneumococcal infection [15], CD14+ monocytes were highly susceptible to cell death (Figure S2B). We also confirmed that different subsets of T-cells were highly susceptible with early evidence of cell death (Figure S2C–E). These subsets included CD161+ cells, which are Th17 cells or cells with the potential to differentiate into this subset [16]. We did not find evidence for the selective involvement of any of these T-cell sub-sets as they were equally susceptible to cell death following bacterial challenge. Since monocytes/macrophages can induce apoptosis in T-cells in specific circumstances [17], [18] we addressed whether monocytes influenced CD3+ T-cell death in cultures of PBMC following exposure to pneumococci. The percentage of CD14+ monocytes in PBMC was 11. 8±1. 1%, while the number in the non-adherent fraction of plastic purified PBMC cultures was reduced to 1. 6±0. 4% (Figure S3) and the percentage of CD14+ monocytes in highly pure CD3+ T-cells cultures was negligible at only 0. 1±0. 05%. In these experiments the highly pure CD3+ T-cell cultures contained 95±0. 7% CD3+ T-cells. As shown in Figure 2A, the increase in total Annexin V+ cells, in PBMC cultures following challenge with pneumococci, was similar in magnitude despite varying numbers of monocytes. However, the PBMC cultures demonstrated significantly more Annexin V+/TO-PRO-3− cells and significantly fewer Annexin V+/TO-PRO-3+ cells than the highly purified CD3+ T-cell cultures (Figure 2B–C), while purified CD3+ T-cell cultures showed no increase in Annexin V+/TO-PRO-3− cells above baseline. Moreover few cells in the purified CD3+ T-cell culture showed accumulation of hypodiploid DNA or cytochrome c translocation into the cytoplasm, a specific feature of apoptosis indicating mitochondrial outer membrane permeabilization [19]. In contrast PBMC or plastic purified PBMC had significant numbers of cells with hypodiploid DNA and evidence of cytochrome c translocation (Figure 2D–F). To ensure cytochrome c translocation was measured only in CD3+ T-cells these cells were purified from PBMC cultures exposed to bacteria. We confirmed the lack of apoptotic features, such as hypodiploid DNA accumulation, in the purified CD3+ T-cell culture was not just the result of altered kinetics of cell death since the differences in accumulation of hypodiploid DNA between purified CD3+ T-cell cultures and PBMC were seen at all time points from 6–16 h after pneumococcal challenge, the latest time point being a time at which cell death was extensive in both cultures using less selective assays (Figure S4). The ability of monocytes to induce apoptosis in T-cells was not altered in the presence of neutrophils or apoptotic neutrophils, demonstrating this role of monocytes is not subverted during an acute inflammatory response or during the resolution of this response (Figure S5). To prove that monocytes were responsible for the enhanced levels of apoptotic T-cells we purified both CD3+ T-cells and CD14+ monocytes and confirmed that addition of monocytes to CD3+ T-cells was sufficient to induce accumulation of hypodiploid DNA and caspase 3 activation in CD3+ T-cells (Figure 3A–B). Apoptosis was not apparent in CD3+ T-cells cultured without CD14+ monocytes, even when cultured with high doses of bacteria for prolonged periods. Cultures containing CD14+ monocytes also showed evidence of a further apoptosis marker, loss of full length Bid indicative of Bid activation (Figure 3C–D). Bim expression remained constant in these experiments. In cultures containing only purified CD3+ T-cells there was no evidence of Bid activation. We were able to show that as the percentage of CD14+ monocytes increased the numbers of CD3+ T-cells with hypodiploid DNA increased (Figure 3E). Our isolation method initially only collected ‘classical’ CD14+ monocytes (Figure S6). ‘Non-classical’ CD14loCD16+ monocytes have emerged as important effectors of innate immunity [2]. We therefore modified our isolation protocol to include ‘non-classical’ and also ‘intermediate’ populations [20], as shown in Figure S6. Addition of these sub-populations to the ‘classical’ monocytes did not alter the level of CD3 T-cells with hypodiploid DNA in comparison to co-cultures with only CD14+ monocytes (Figure 3F). In the absence of monocytes CD3+ T-cells underwent a form of necrotic cell death. There was evidence of extensive loss of membrane integrity in the purified CD3+ T-cells and evidence of morphological features of necrosis such as prominent membrane disruption without features of cell shrinkage or of the nuclear condensation observed in the CD3+ T-cells isolated from PBMC (Figure 4A). Purified CD3+ T-cells showed prominent membrane disruption as evidenced by greater levels of Trypan blue staining (Figure 4B). The cell death in purified CD3+ T-cell cultures was not altered by necrostatin (Figure 4C) and death was not converted to apoptosis (Figure 4D) suggesting the death process did not involve necroptosis. Purified CD3+ T-cells did not demonstrate caspase 1 activation suggesting pyroptosis was unlikely to be the mechanism of cell death (Figure S7A). Cell death in PBMC required live bacteria (Figure S7B). The pneumococcal virulence factor pneumolysin induces not only cell lysis with membrane permeabilization but also apoptosis in mammalian cells [21]. Distinct regions of pneumolysin govern cytolytic and non-cytolytic activity such as complement activation. Overall cell death in CD3+ T-cells in PBMC cultures, which we show above is apoptotic, was partially reduced but not restored to baseline by a mutant (Δ6), which produces toxin that lacks cytolytic activity (Figure 4E and Figure S7C). Purified pneumolysin could induce cell death of lymphocytes (Figure 4F) but these cells did not have features of apoptosis such as accumulation of hypodiploid DNA (Figure 4G). These results demonstrate that the purified T-cells die by a primarily necrotic process with prominent membrane permeabilization and implicate pneumolysin in the cell death. We next addressed whether monocytes altered CD3+ T-cell activation in our model, since monocytes can enhance MHC-independent T-cell activation [22] and enhanced activation can alter susceptibility to apoptosis [23]. Monocytes/macrophages can also induce death receptor mediated cell death in activated T-cells [18]. We showed enhanced expression of the CD3+ T-cell activation markers CD69, CD25 and HLA-DR in PBMC cultures exposed to viable or heat-killed pneumococci (Figure 5A–B). CD3+ T-cells from HIV-1 seropositive individuals are known to have enhanced levels of activation and to be particularly susceptible to macrophage-mediated cell death [18]. In keeping with this we documented that cells isolated from HIV-seropositive individuals who were not receiving antiretroviral therapy (since HIV viral load influences activation state and susceptibility to apoptosis [24]), had increased levels of CD3+ T-cell death, as compared to controls, when exposed to pneumococci (Figure 5C). These results suggested T-cell activation influenced levels of cell death and activation was enhanced in PBMC co-cultures containing monocytes. To determine the apoptotic pathway inducing monocyte-dependent CD3+ T-cell apoptosis we next examined the potential role of Fas ligand (FasL). Fas signaling contributes to T-cell receptor mediated activation-induced cell death but has also been implicated in macrophage-induced cell death of activated T-cells [18], [23]. PBMC were challenged with pneumococci and then CD3+ T-cells were purified to prevent other mononuclear cells from confounding the results. This demonstrated increased levels of caspase 8 activation in CD3+ T-cells co-cultured with monocytes (Figure 6A), suggesting death receptor-mediated apoptosis [23]. We next examined involvement of Fas in the death process using an infectious dose at the low end of the range we had shown causes death (see Figure 1B) to limit competing effects of necrosis. A Fas blocking antibody prevented Bid activation and cytochrome c accumulation in the cytosol of CD3+ T-cells present in PBMC (Figure 6B–E). The numbers of CD3+ T-cells with accumulation of hypodiploid DNA was also reduced following treatment with anti-Fas antibody (Figure 6F). In addition, when we supplied a Fas-blocking antibody and measured the level of Annexin V+ cells in CD3+ T-cells isolated from PBMC following pneumococcal infection, there was also a reduction of cell death (Figure 6G). Similar experiments in purified CD3+ T-cells revealed Fas inhibition failed to reduce the levels of Annexin V+ CD3+ T-cells (data not shown). In addition, a granzyme B inhibitor (added since it has been suggested that pneumolysin can substitute for perforin and combine with granzyme B to induce lymphocyte apoptosis [25]) did not alter levels of T-cell apoptosis (Figure S8). TRAIL blocking antibody or PDL-1 blocking antibody also had no effect on levels of T-cell apoptosis (data not shown). These results confirmed that Fas mediated CD3+ T-cell apoptosis in PBMC cultures. To confirm the relevance of these findings in vivo we documented T-cell apoptosis in the spleen, an environment where macrophage-mediated T-cell death would be anticipated to occur in vivo [26], in mice with invasive pneumococcal disease (Figure 7A). Differences in levels of bacteremia can influence levels of T-cell apoptosis [27]. Therefore our investigation of apoptosis in vivo was carried out using a high dose of serotype 4 S. pneumoniae, a virulent strain of bacteria that establishes high levels of bacteremia [28], which we confirmed were similar between FasL deficient gld and control mice (data not shown). Under these conditions, which normalized blood bacterial colony counts, we documented a significant reduction in the levels of hypodiploid DNA in CD3+ T-cells from gld mice, as compared to TRAIL deficient or wild-type mice (Figure 7B). However, T-cells play an important role in the early stages of pneumonia [5], [6]. To test the relevance of FasL to pneumonia outcomes we confirmed that gld mice had increased levels of bacteria in the lung and blood in a model when there were very low levels of recruited neutrophils (Figure 7C–E). In association with this gld mice had increased levels of TNF-α (Figure 7F), though not IL-6 (Figure 7G). We also established that T-cells in this model play an important role in host defense by establishing that control mice treated with cyclosporine, a known T-cell immunosuppressant and cause of T-cell lymphopenia [11], had reduced control over bacterial replication and that this T-cell mediated control was lost in the gld deficient mice (Figure 7H–I). This suggested FasL was critical for T-cell dependent host defense in the early stages of pneumococcal infection when there are still relatively few recruited neutrophils.
Monocytes are usually regarded as contributing to the ramp up of inflammatory responses rather than triggering their cessation [2]. We now demonstrate an important role for monocytes in controlling the number of activated T-cells during bacterial infection. Although a variety of T-cell populations contribute to the immune response to bacteria such as pneumococci [5], [7], [10], the regulation of these cell populations and the role of cell death in this process are incompletely defined. We show that monocytes play an important and unrecognized role ensuring Fas-dependent apoptosis to ensure the safe disposal of activated T-cells. In the absence of monocytes pathogen driven T-cell necrosis predominates. Moreover we also demonstrate that in mice T-cell dependent early host responses to pneumococcal infection are attenuated in the absence of FasL, suggesting that monocyte-mediated Fas-dependent apoptosis represents a critical component of the early stages of an effective host response against pneumococci. We were also able to show this response could occur in the presence of viable or apoptotic neutrophils, suggesting it may also be important at later stages of infection. Cholesterol-dependent cytolysins, including listeriolysin, are recognized stimuli for T-cell death [27]. We found the related pneumococcal toxin pneumolysin triggered non-apoptotic death with early membrane permeabilization in purified T-cells. Pneumolysin enhanced monocyte-dependent T-cell death, though in this case via apoptosis without evidence of early membrane disruption, suggesting apoptosis did not require outer cell membrane pore formation. Cytolytic activity of pneumolysin has been traditionally viewed as a virulence determinant but recent evidence suggests that pneumolysin is also an important trigger for effective innate immune responses against the pneumococcus, including those mediated by the nucleotide-binding oligomerization domain-like receptor family pyrin domain containing 3 (NLRP3) inflammasome [29]. In our model of monocyte-mediated apoptosis it appears that the requirement for cytolytic toxin was as a trigger for the host response that induced maximal levels of apoptosis. Mutants that lack cytolytic activity such as ST306 serotype 1 and ST53 serotype 8 pneumococci have emerged as major causes of invasive pneumococcal disease [30]. On the basis of our findings we would predict these strains would induce less T-cell apoptosis and that persistence of activated T-cells could contribute to the invasive potential of these non-cytolytic strains. In contrast, once these strains have invaded the propensity to cause lower rates of apoptosis might explain the lower mortality associated with them since lymphocyte apoptosis may influence outcomes during sepsis [31]. Surprisingly PBMC needed to be in contact with the bacteria for the toxin to mediate its effect. Although the toxin is soluble when released from lysed bacteria it may also remain associated with the bacterial cell wall [32], which might explain this finding. Alternatively cell contact with PBMC might upregulate toxin production or enhance release by increasing bacterial lysis [33]. LPS can stimulate MHC-independent T-cell activation via a monocyte-dependent process involving CD80 [22]. Pneumolysin is also a TLR4 ligand that activates T-cells; therefore a similar mechanism may explain the enhanced T-cell activation we observed in PBMC cultures [5], [21], although the fact that heat killed bacteria also activated T-cells suggests active toxin is not essential. Other microbial factors could also stimulate monocytes to activate CD3+ T-cells. One consequence of increased T-cell activation is to enhance susceptibility to apoptosis after a period of sustained activation and this paradigm is exemplified by the apoptosis of uninfected CD3+ T-cells during HIV infection [34]. In keeping with this PBMC from HIV-seropositive individuals had elevated levels of CD3+ T-cell apoptosis following challenge with pneumococci. In addition to enhancing susceptibility to apoptosis T-cell activation could increase resistance to the competing toxin-related cytolytic cell death through cytokine production. For example, IFN-γ, produced by activated Th1 cells, reduces the susceptibility of epithelial and monocytic cell lines to pneumolysin associated cytolytic death [35]. Consistent with this possibility mice lacking IFNγ are more susceptible to listeriolysin dependent cell death [27]. In the presence of ‘classical’ monocytes we found that CD3+ T-cell death had classic features of apoptosis and this effect was maintained even in the presence of small numbers of monocytes, as evidenced by high levels of CD3+ T-cell apoptosis in plastic purified PBMC. ‘Non-classical’ monocytes played less of a role in this process though the numbers we isolated were relatively small so we cannot exclude a possible role when their numbers expand. The absence of early membrane permeabilization or of inhibition by necrostatin argued against the cell death being a form of programmed necrosis and the absence of cell swelling, early loss of membrane integrity or caspase 1 activation was inconsistent with pyroptosis, which has recently been described in T-cells [36]. In contrast cell death in purified CD3+ T-cells which were not co-cultured with monocytes was characterized by early membrane disruption without significant mitochondrial outer membrane permeabilization, Bid activation or accumulation of hypodiploid DNA, consistent with necrosis. These results define a novel role for monocytes in downregulating the inflammatory response during acute bacterial infection. A key role for monocytes in downregulating the inflammatory response has been previously ignored. Unlike macrophages, monocytes are not thought to clear apoptotic cells by efferocytosis [3]. Nevertheless, monocytes express FasL [37] and evidence that monocytes induce Fas-mediated apoptosis of activated T-cells has been provided during HIV infection [17] and chronic periodontitis [38]. Nevertheless this role of monocytes has not been demonstrated previously during acute bacterial infections, where downregulation of the inflammatory response is a critical component of inflammation resolution. In our model monocytes did not influence overall T-cell viability (levels of death were similar in the presence or absence of monocytes) but their role was to ensure the induction of an injury limiting apoptotic form of cell death. There are a number of potential benefits to the host of monocyte-dependent CD3+ T-cell apoptosis: ensuring the timely removal of activated T-cells downregulates pro-inflammatory cytokine expression by T-cells, induction of apoptosis rather than necrosis limits secondary tissue injury and the removal of apoptotic cells by efferocytosis further detunes pro-inflammatory cytokine expression by macrophages during pneumococcal pneumonia [39]. Consistent with these possibilities we noted that FasL deficient mice had reduced T-cell apoptosis (when we normalized the level of bacteraemia to exclude confounding effects of differences in bacterial number) and that during the early T-cell dependent stages of pneumococcal pneumonia FasL deficiency was associated with impaired bacterial clearance, in a model that had few recruited neutrophils and was therefore unlikely to be confounded by any direct effects of Fas-signalling on neutrophil function [40]. FasL deficiency was associated with upregulation of TNF-α which would be anticipated to in turn upregulate receptors involved in the translocation of bacteria into the blood such as the PAF receptor [41]. Activated T-cells enhance TNF-α production by monocytes, through release of IFN-γ, but also through membrane interactions with monocytes [42], [43]. Induction of T-cell apoptosis not only would be anticipated to reduce these stimuli but would also provide apoptotic cells which when ingested by macrophages would further reduce TNF-α expression [39]. IL-6 production, however, was not reduced even though production is also decreased in the presence of apoptotic cells [39]. The reason for this finding was not apparent but may reflect the pleiotropic role of IL-6 both as an inflammatory cytokine and as a cytokine which can downregulate T-cell responses [44]. It remains possible that its production was sustained since it is produced as part of a counter-regulatory response to the sustained presence of activated T-cells [44], although we were not able to formally test this possibility. Recently it has been shown that selective CD4+ T-cell depletion, non-selective depletion of T-cells or reduction of T-cell activation can improve survival in murine models of invasive pneumococcal disease, suggesting excessive T-cell activation can be harmful and must be tightly controlled [11]. Our observations provide a mechanism for this process. Prior observations of T-cell apoptosis and of increased FasL expression in a small cohort of patients with pneumococcal disease are also consistent with this [12]. We established a hierarchy of cell death programs such that monocytes induced apoptosis as the dominant mechanism of T-cell death during pneumococcal infection. In the absence of monocytes the combination of non-apoptotic cell death and the unopposed effects of pro-inflammatory microbial factors, such as pneumolysin, would be predicted to result in increased tissue injury and bacterial invasion [21]. In the context of pneumolysin-mediated T-cell activation [5], and in light of the potential for pneumococci to induce a predominantly cytolytic cell death in purified CD3+ T-cells, our observation that monocytes regulate T-cell death, ensuring Fas-mediated apoptosis predominates, is significant. These findings emphasise the important and unrecognised role of monocytes in downregulating the inflammatory response to bacterial infection by regulating numbers of activated T-cells during the immune response to infection.
Human PBMCs were isolated from whole blood donated by healthy volunteers as previously described with written informed consent as approved by the South Sheffield Regional Ethics Committee of Royal Hallamshire Hospital (Sheffield, United Kingdom). All animal experiments were performed in accordance with the UK Animals (Scientific procedures) Act, authorized under a UK Home Office License, and approved by the animal project review committee of the University of Sheffield. Experiments were performed with serotype 2 S. pneumoniae, strain D39 (NCTC7466) or its Δ6 mutant unless otherwise stated. In murine experiments serotype 1 (SSISP strain; Statens Seruminstitut) or serotype 4 S. pneumoniae (TIGR4 strain) were utilized. To test the contribution of the pneumococcal endonuclease EndA [45] in increasing the extent of cleavage of hypodiploid DNA we screened a panel of serotype 1 strains for EndA production using the method of Beiter et al [45] and used strain NCTC7465 that is EndA negative and INV104B that expresses EndA. All bacterial stocks were prepared as described previously [46]. In certain experiments bacteria were heat-killed by boiling for 10 min prior to incubation with cells and inhibition of growth was confirmed by documenting no growth following plating on blood agar. The Δ6 mutation refers to deletion of amino acid residues A146 and R147 from pneumolysin, introduced by site-directed mutagenesis [47]. The altered gene replaced the chromosomal gene [48]. The Janus cassette, which encodes markers for kanamycin resistance and streptomycin sensitivity, was linked to 300 bp of upstream and downstream flanking DNA from the pneumolysin gene by splice overlap PCR [48]. This construct was transformed into a streptomycin resistant version of D39. Homologous recombination in the flanking DNA introduced the Janus cassette in place of the pneumolysin gene giving an intermediate form of D39 that was resistant to kanamycin and sensitive to streptomycin. The intermediate form was isolated by growth on plates containing 150 mg/ml kanamycin and transformed with the Δ6 gene. Homologous recombination in the flanking DNA replaced the Janus cassette with the altered form of the pneumolysin gene giving D39 that was resistant to streptomycin and sensitive to kanamycin. Recombinants were selected by growth on plates containing 300 mg/ml streptomycin. The selected strain was checked for production of pneumolysin by western blot and the toxin produced was shown to be non-haemolytic using a standard hemolytic assay of cell extracts as described previously [49], see Figure S7C. Peripheral blood mononuclear cells (PBMC) plated at 1×106 cells/ml were isolated and cultured as previously described [50]. Plastic purified lymphocytes (PBL) were obtained by plating PBMC at 2×106 cells/ml in 25 cm2 flasks (Costar) for 1 h, transferring the non-adherent cells to a new flask for a further 1 h, removing, washing and counting the non-adherent cells and then resuspending at 1×106 cells/ml in a 24-well plate. Purified CD3+ T-cells and monocytes with and without CD16 depletion were isolated from PBMC by magnetic immunoselection using EasySep human T cell enrichment kit, EasySep human monocyte enrichment kit and EasySep human monocyte enrichment kit without CD16 depletion (Stemcell Techonologies). In the absence of CD16 depletion monocyte sub-populations corresponding to ‘non-classical’ and ‘intermediate’ monocytes, as described by Ziegler-Heitbrock and colleagues [20] were identified (Figure S6). Neutrophils were isolated by dextran sedimentation and plasma-Percoll (Sigma) density gradient from the peripheral blood of healthy volunteers and were approximately 98% pure. Neutrophils were made apoptotic as previously described [51]. In brief neutrophils were cultured for 20 h to ensure a population that was approximately 80% apoptotic and <5% necrotic, as defined by trypan blue. Neutrophils immediately after isolation or apoptotic neutrophils were added to PBMC cultures at a ratio of 10 neutrophils per PBMC for the indicated periods. Freshly isolated PBMC, plastic purified lymphocytes and CD3 enriched T-lymphocytes or cultures of purified CD3+ T-lymphocytes and purified monocytes were infected with S. pneumoniae or mock-infected for 4–24 h. In certain experiments transwells with 0. 4 µm pore size (BD Biosciences) were used to separate bacteria from lymphocytes [52]. To inhibit caspase activation 10 µM of N-benzyloxycarbonyl-Val-Ala-Asp (O-methyl) fluoromethyl ketone (Z-VAD-FMK) (R&D Systems Inc.) or DMSO (Sigma) as vehicle control was added for 30 min before infection. ZB4 neutralizing Fas antibody at 1 µg/ml (Enzo Life Sciences), or isotype control, was added to cultures 30 min prior to infection [18]. In additional blocking experiments, cells were pre-treated for 30 min with 2. 5 µg/ml of neutralizing TRAIL antibody 2E5 (Enzo Life Sciences), 10 µg/ml of mouse anti-human CD274 (PD-L1) (MIH1, eBioscience), 30 nM of necrostatin-1 (Sigma) to inhibit necroptosis, or 5 µM of granzyme B inhibitor (Calbiochem). In experiments involving antibodies monocyte Fcγ receptors were pre-blocked with 100 µg of human IgG1 (Sigma). Cytospin preparations were generated from non-adherent PBMC after infection (Cytospin 3; Thermo Shandon) and slides stained with TUNEL (terminal deoxynucleotidyl transferase dUTP nick end labelling) reagents per the manufacturer' s instructions (Oncor), counterstained with 4′6′-diamidino-2-phenylindole (DAPI (Vectashield) ) and mounted with cover slips containing Vectashield. TUNEL and DAPI-positive cells were viewed via fluorescence light microscopy (Leica DMRB 1000). PBMC or CD3 enriched T-cells were challenged with S. pneumoniae (serotype 2, D39) or mock-infected for 6 h then fixed in ice-cold 3% glutaraldehyde/0. 1 M phosphate buffer overnight at 4°C. The cell pellets were processed as previously described [15]. Sections (85 nm) were cut on a Reichert Ultracut E ultramicrotome and stained with 1% toluidine blue in 1% borax. Sections were examined using an FEI Tecnai transmission electron microscope at an accelerating voltage of 80 kV and micrographs were taken using a Gatan digital camera. Flow cytometric measurements were performed using a four-colour FACSCalibur flow cytometer (Becton Dickinson). Forward and side scatter light was used to identify cell populations by size and granularity. Fcγ receptor blockade was with 100 µg/ml human IgG1 (Sigma). Cell surface marker expression was with 1 µg/ml mouse anti-human anti- CD14 (61D3) phycoerythrin (PE), (eBioscience), anti-CD3 (SK7) fluorescein isothiocyanate (FITC), (BD Pharmingen), anti-CD161 (HP-3G10) allophycocyanin (APC), (eBioscience), anti-CD4 (CSK3) peridinin chlorophyll protein (PerCP) and anti-CD19 (HIB19) FITC (BD Pharmingen) with appropriate isotype controls. To determine percentages of monocyte sub-populations after each isolation procedure we used 1 µg/ml mouse anti-human anti-CD14 (TuK4) pacific blue (Invitrogen) and anti-CD16 (B73. 1) PE (eBioscience) or appropriate isotype controls. Activation marker expression was with 1 µg/ml of mouse anti-human anti-CD69 (FN50) APC (eBioscience); anti-human anti-CD25 (M-A251) APC-H7 (eBioscience) and anti-human anti-HLA-DR (L243) PE (BD Pharmingen). In murine experiments T-lymphocytes were identified using rat anti-mouse anti-CD3 (17A2) eFluor 450 (eBioscience). Annexin V-PE and TO-PRO-3 was used to assess cell death. Annexin V+/Topro3− cells were regarded as early apoptotic and Annexin V+/TO-PRO-3+ cells as late apoptotic/necrotic [46]. To measure loss of the inner mitochondrial transmembrane potential (Δψm), the cationic dye JC-1 was used. Loss of Δψm was confirmed by loss of red and gain of green fluorescence. DNA fragmentation was confirmed by the hypodiploid peak assay [14]. Apoptotic cells were visible as a distinct population with low fluorescence compared to healthy cells in growth phase 0/1. Since the events recorded were predominantly on the far left of the hypodiploid peak when depicted on a linear scale, we have presented the histograms on a log scale [53]. We tested the role of the pneumococcal endonuclease EndA in further degrading the DNA of the apoptotic cells [45]. Using pneumococcal strains that are sufficient in functional EndA, or lack significant EndA activity, we confirmed that EndA increased the degree of DNA degradation, pushing events to the left on a log scale and that this appearance could be reproduced by incubating apoptotic cells with 1 µg/ml of DNAse (Promega) (Figure S9). In all flow cytometry experiments 10,000 events were captured and analyzed with FlowJo software version 9. 3. 2 (Tree Star, Inc.). Caspase 3 and 8 activities were measured using the Caspase-Glo 3/7 Assay (Promega, USA) and Caspase-Glo 8 Assay (Promega, USA). After challenge with D39 (16 h, MOI 10), 4×104 PBMC (or in some experiments CD3+ T-cells isolated from the PBMC by magnetic immunoselection as above) were suspended in 50 µl of media and were combined with 50 µl of the Caspase-Glo 3/7 or Caspase-Glo 8 reagent for 1 h at room temperature. Luminescence was measured with a Packard Bioscience Fusion universal microplate analyser (Perkin Elmer, Beaconsfield, UK). Cells were lysed, quantified and separated by SDS-PAGE as described with 15–40 µg protein/lane [49]. Proteins were transferred to Immobilon-P membrane (Millipore), blocked in PBST/5% non-fat dry milk powder and incubated with primary and secondary antibodies. Antibodies used were: mouse anti-human cytochrome c (7H8. 2C12, BD Biosciences), mouse anti-human Cox-4 (Molecular Probes), rabbit anti-human Bid (Cell Signalling), rabbit anti-human Bim (Chemicon Ltd.), rabbit anti-human caspase 1 (Abcam) and actin (A2086, Sigma-Aldrich). Detection was with HRP-conjugated goat anti-rabbit and anti-mouse immunoglobulins (Dako), and enhanced chemiluminescence (Amersham Pharmacia). PBMC were isolated from 8 HIV-seropositive individuals who were not receiving antiretroviral therapy. Individuals had CD4 T-cell counts of 498±49/µl (Mean ± SEM) and HIV viral loads of 42379±15823 as assessed by Roche Cobas Ampliprep/Cobas TaqMan HIV-1 version 2. 0. PBMC were challenged with pneumococci and apoptosis in CD3+ T-cells was measured by estimating hypodiploid DNA accumulation as above. C57BL/6 mice (Harlan, Oxford, U. K.), mice homozygous for the FasLgld mutation (B6Smn. C3-Faslgld/J, gld) [54] and TRAIL deficient mice [55] both on a C57BL/6 background, received 5×105 or 1×107 cfu serotype 1 or 4 S. pneumoniae by intratracheal instillation for 24 h as described [46]. Excised spleens were homogenized and splenocytes separated from tissue using a 100 µm cell strainer (BD biosciences) before staining and analysis by flow cytometry. Bacterial colony counts were estimated in lung and blood as previously described [46]. Bronchoalveolar lavage (BAL) was performed by instilling 5×0. 8 ml of ice-cold heparinized-saline intratracheally (Leo laboratories) and total cell counts estimated by hemocytometer [46]. Cytospins were prepared (Cytospin 3, Thermo Shandon) and the percentage neutrophils in the BAL estimated as described [46]. TNF-α and IL-6 were measured by EIA as described [39]. Results were recorded as mean and standard error of the mean with the number of individual donors cells contributing to each data set shown as the ‘n’ value. Differences between groups of treatments were calculated by ANOVA (Bonferroni' s post-test), with Kruskal-Wallis test with Dunn' s Multiple Comparison Test for non-parametric comparisons or with paired t-test for comparison of means using GraphPad Prism 5 (Graphpad Software, Inc.). Significance was defined as p<0. 05. | T-cells are important contributors to the early host response to pneumonia, but their numbers must be tightly regulated to limit inflammatory lung injury. Cell death regulates T-cell numbers but the mechanism of execution influences the inflammatory cost with apoptosis viewed as predominantly anti-inflammatory and necrosis as pro-inflammatory. We show that monocytes determine the mechanism of T-cell death during acute bacterial infection. Monocytes triggered Fas-dependent T-cell apoptosis but in the absence of monocytes T-cells died by necrosis, which required the pneumococcal virulence factor pneumolysin. We also show that Fas ligand is required to regulate the early T-cell dependent host response to pneumococci during pneumonia. Although monocytes have previously been associated with enhancement of the inflammatory response our results imply that a key role of monocytes is to dampen the inflammatory response through induction of Fas-mediated apoptosis of activated T-cells during S. pneumoniae pneumonia. Our data identifies a critical and unrecognized regulatory role for monocytes during pneumonia. | Abstract
Introduction
Results
Discussion
Materials and Methods | medicine
cell death
infectious diseases
molecular cell biology
cell biology
biology
microbiology
host-pathogen interaction | 2012 | Monocytes Regulate the Mechanism of T-cell Death by Inducing Fas-Mediated Apoptosis during Bacterial Infection | 10,360 | 267 |
Zaire ebolavirus (ZEBOV), a highly pathogenic zoonotic virus, poses serious public health, ecological and potential bioterrorism threats. Currently no specific therapy or vaccine is available. Virus entry is an attractive target for therapeutic intervention. However, current knowledge of the ZEBOV entry mechanism is limited. While it is known that ZEBOV enters cells through endocytosis, which of the cellular endocytic mechanisms used remains unclear. Previous studies have produced differing outcomes, indicating potential involvement of multiple routes but many of these studies were performed using noninfectious surrogate systems such as pseudotyped retroviral particles, which may not accurately recapitulate the entry characteristics of the morphologically distinct wild type virus. Here we used replication-competent infectious ZEBOV as well as morphologically similar virus-like particles in specific infection and entry assays to demonstrate that in HEK293T and Vero cells internalization of ZEBOV is independent of clathrin, caveolae, and dynamin. Instead the uptake mechanism has features of macropinocytosis. The binding of virus to cells appears to directly stimulate fluid phase uptake as well as localized actin polymerization. Inhibition of key regulators of macropinocytosis including Pak1 and CtBP/BARS as well as treatment with the drug EIPA, which affects macropinosome formation, resulted in significant reduction in ZEBOV entry and infection. It is also shown that following internalization, the virus enters the endolysosomal pathway and is trafficked through early and late endosomes, but the exact site of membrane fusion and nucleocapsid penetration in the cytoplasm remains unclear. This study identifies the route for ZEBOV entry and identifies the key cellular factors required for the uptake of this filamentous virus. The findings greatly expand our understanding of the ZEBOV entry mechanism that can be applied to development of new therapeutics as well as provide potential insight into the trafficking and entry mechanism of other filoviruses.
Zaire ebolavirus (ZEBOV, Genbank: AF086833), a member of the family Filoviridae, genus Filovirus, causes a highly fatal hemorrhagic fever in humans and non-human primates. Over the past three decades numerous human outbreaks have occurred in Central Africa involving hundreds of cases with fatality rates ranging from 50–89% [1]. In addition, outbreaks of ZEBOV infection have been implicated in deaths of tens of thousands of gorillas, chimpanzees and duikers in Central and Western Africa posing a considerable threat to the wildlife and ecology in those areas [2]. Due to a very high case fatality rate in humans, significant transmissibility of the virus, lack of effective preventive or therapeutic measures against the disease, ZEBOV is considered a serious emerging viral pathogen. Currently no specific therapy or vaccine is approved for human or animal use against this pathogen. As for other members of Filoviridae, ZEBOV is morphologically distinct from other animal viruses. The virions are long and filamentous with an average length of 800–1000 nm and a diameter of about 80 nm but can form a variety of shapes ranging from straight rods to closed circles [3]. Virions are surrounded by a host cell-derived lipid envelope. The envelope contains virally-encoded glycoprotein (GP) spikes composed of homotrimers of two virally-encoded glycoproteins, GP1 and GP2. The approximately 19 kb single-stranded, negative-sense genomic RNA complexed with nucleocapsid, VP35, VP30 and L proteins form the nucleocapsid, while VP40 forms the matrix that underlies the viral membrane [4]. Like all viruses, ZEBOV largely relies on host cell factors and physiological processes for key steps of its replication cycle. Identification of these processes and factors will not only allow a better insight into pathogenic mechanism, but may identify novel targets for future therapeutic development. As the first step of replication, entry into the host cell is an attractive target for therapeutic intervention as infection can be stopped before virus replication disrupts cellular functions. However, the entry mechanism of ZEBOV, and that of other large enveloped viruses is very limited. Many enveloped viruses, including ZEBOV, require endocytosis to infect cells. The internalized virus is transported through successive endocytic vesicles to reach a vesicle/compartment where conditions are conducive (low pH and/or presence of proteolytic enzymes) for the GP to attain a suitable conformation needed for membrane fusion [5], [6], [7]. Upon fusion of viral and endocytic membranes, the capsid moves into the cell cytoplasm to begin genome replication. Several distinct endocytic mechanisms exist in mammalian cells. They are distinguished from each other on a number of criteria including the size and morphology of endocytic vesicles, the type of cargo they carry, the cellular factors involved in their control and their origins and destinations [8]. Different viruses employ different routes of endocytosis, and the route taken by a given virus largely depends on the receptor it interacts with. Clathrin-mediated endocytosis (CME) is the best understood endocytic pathway. A number of viruses including Influenza A (Genbank: M73524), Semliki forest (Genbank: X04129) and vesicular stomatitis viruses (VSV; Genbank: J02428) employ this pathway for entry [9], [10]. The internalization of the virus-receptor cargo occurs in specialized areas of cell membrane called clathrin-coated pits (CCPs). CCPs are formed on the cytoplasmic face of the plasma membrane through sequential assembly of proteins including clathrin that form a cage-like structure lining the cytoplasmic side of the pit. The pit then invaginates and buds from the plasma membrane forming a clathrin-coated vesicle approximately 120 nm in diameter containing the internalized cargo. Subsequently, the vesicle sheds its clathrin coat, a prerequisite for further trafficking and merging with other compartments. A number of accessory, adaptor and signaling molecules participate in this process, and provide a tight regulation of the pathway. Some, such as accessory protein-2 (AP2) and Eps15, are specifically associated with CCPs, while others such as dynamin, which is responsible for vesicle budding from the plasma membrane, are shared with other endocytic pathways [8]. Caveolin-mediated endocytosis (CavME), first observed for the cellular uptake of simian virus 40 (SV40) [11], differs from CME in terms of internalization mechanism and vesicular transport route. Caveolae are flask-shaped invaginations in the plasma membrane that are rich in caveolin protein, and are predominantly associated with cholesterol-rich plasma membrane microdomains termed lipid rafts. Therefore, extraction or perturbation of membrane cholesterol severely impedes entry of viruses that use CavME. Vesicles derived from CavME are indicated by the presence of caveolin and are termed caveosomes. Other cellular factors such as Eps15-related (Eps15R) protein are thought to be specific for CavME, but as with CME, dynamin is still required for severing of caveolae from the plasma membrane. Another distinguishing feature is that caveloae are smaller than CCPs and have an average diameter of approximately 60–80 nm [8]. Recently, the importance of macropinocytosis, as a distinct endocytic uptake mechanism for virus infection, has started to be realized for some viruses [9]. Macropinocytosis is associated with membrane ruffles such as those formed by filopodia and lamellipodia, which are outward extensions of the plasma membrane driven by actin polymerization underneath the membrane surface [12]. When a ruffle folds back upon itself a cavity can be formed. Subsequent fusion of the distal end of the loop with the plasma membrane results in formation of a large vesicle called a macropinosome. These can range in size from 200 to 10,000 nm across and take up cargoes of similar dimensions [9]. Morphological and regulatory characteristics that distinguish macropinocytosis from other endocytic processes have also begun to emerge [8], [9], [13], [14]. Macropinosomes are best characterized for uptake of fluid phase markers such as high molecular weight dextran and horse-radish peroxidase and is sensitive to inhibitors of Na+/H+ exchangers, such as amilorides [14]. As for CavME, they are dependent on cholesterol-rich lipid rafts, but dynamin is not required. Instead, scission of macropinosomes appears to require CtBP/BARS [9], [15], [16]. Other work indicates the involvement of cell signaling factors PI3K, Akt, PKC, PLCγ and PLC-A2 that act to promote membrane ruffling by stimulating actin remodeling through Rac and cdc42 [9]. All endocytic pathways used by viruses serve to deliver virus to vesicles and compartments conducive to virus membrane fusion and release of the core into the cell cytoplasm at a site where replication proceeds optimally. Many endocytic pathways share common features, such as acidification, yet each virus type appears to prefer one trafficking pathway over others and misdirection into alternative pathways can result in inhibition of infection. For most enveloped viruses, the point at which membrane fusion occurs appears to be at the early or late endosome stage. This evidence has been gathered by comparing pH-sensitivity of the GP to known pH of the endosome at different stages of maturation. More recently, the use of dominant negative GTPases, that are involved in endosomal maturation, have been used in determining virus exit points from the endosome [17], [18], [19]. In general, the two methods agree but provide little detail as to whether viruses have additional requirements in terms of site of release other than the early or late endosome. Currently, a detailed understanding of ZEBOV endocytosis and trafficking is lacking. Each of the previous studies on understanding ZEBOV entry pathway have indicated involvement of different pathways, including CME [20], [21], CavME [20], [22] and a Rho GTPase-dependent pathway that may suggest involvement of macropinocytosis [23]. These conflicting findings may be due to the use of surrogate models of ZEBOV such as pseudotyped retroviruses, which are morphologically and biochemically distinct from wild type filamentous ZEBOV and/or reliance on one analytical approach, such as use of pharmacological agents, which are likely to act on more than one cellular target [24]. Here we have used multiple independent approaches employing replication-competent, infectious ZEBOV and/or morphologically comparable virus-like particles (VLPs). We have examined the contribution of each endocytic pathway to ZEBOV entry and infection by quantitative analysis. The work involves measuring the impact of drugs, siRNA and/or expression of well characterized dominant negative (DN) mutants of cell trafficking proteins on virus entry and infection. We also use fluorescently-labeled virus-like particles (VLPs) to follow virus internalization and trafficking through different endocytic compartments. The product of combining all these approaches provides, for the first time, an accurate and detailed description of ZEBOV uptake mechanism. Our data clearly indicate that wild type ZEBOV enters and infects Vero and HEK293T cells independently of clathrin, caveolae and dynamin. Instead, virus entry required the presence of cholesterol in the cell membrane and was inhibited by the amiloride derivative, EIPA. A marked induction in fluid-phase uptake was also observed shortly after virus binding to cells and internalized virus particles showed significant colocalization with high molecular weight dextran. In addition, inhibition of p53-activated kinase (Pak1) or CtBP/BARS resulted in significant reduction in virus entry and infection. Importantly, the virus particles appear to stimulate uptake through this pathway directly by promoting localized actin polymerization and this is consistent with our previous work where the GP triggered the PI3 kinase signaling cascade and Rac1 activity [25]. No evidence for involvement of clathrin- or caveolae-dependent endocytosis was seen. Instead, the primary mechanism of virus uptake appears closely related to macropinocytosis. Subsequent to internalization, the virus utilizes the conventional endolysosomal pathway and is trafficked through early and late endosomes before membrane fusion takes place. This study provides novel information regarding ZEBOV entry, and is likely to be useful in understanding the entry mechanism of other filoviruses.
A recent study suggested that Ebola virus uses CME for cellular entry [21], while an earlier study had implicated both CME and CavME [20], [21]. However, these studies utilized either pseudotyped virus, which due to morphological and/or biochemical differences may not accurately depict the ZEBOV entry pathway, and/or relied solely on the use of pharmacological agents, which may alter multiple processes important for membrane trafficking. To more closely examine the role of the each endocytic pathway we used replication-competent infectious virus and morphologically comparable ZEBOV virus-like particles (VLPs) to determine the impact of specific dominant-negative (DN) forms of Eps15 (OMIM: 600051) and caveolin-1 (cav-1, OMIM: 601047) on infection and VLP uptake into cells. Eps15 and caveolin-1 (cav-1) are required for the formation and trafficking of CME and CavME vesicles respectively, and their DN forms inhibit the respective endocytosis with high specificity [26]. HEK293T cells were transfected with plasmid encoding GFP alone or GFP-tagged forms of DN-Eps15 or DN-Cav1, and subsequently infected with ZEBOV. Infected cells were detected by immunofluorescence staining for ZEBOV matrix protein (VP40) protein and the proportion of cells that co-expressed the transfected protein and ZEBOV VP40 (infection marker) was calculated. It was found that the proportion of ZEBOV-infected cells in cultures expressing DN-Eps15-GFP or DN-Cav1-GFP was not significantly different (P>0. 05) to that in cultures expressing GFP alone, indicating that neither DN protein had a significant impact on ZEBOV infection (Fig. 1A). This lack of effect of either DN protein on ZEBOV entry was confirmed using a sensitive contents-mixing entry assay that measures virus-endosomal membrane fusion by monitoring luciferase release from VLPs into the cell cytoplasm [25]. Expression of either DN protein failed to have any significant effect on entry of ZEBO-VLP (P>0. 05; Fig. 1B). As control, VLPs bearing VSV envelope glycoprotein (VSV-VLP) were used. VSV is known to use CME for cellular entry [27]. Entry of VSV-VLP was significantly inhibited only in cells expressing DN-Eps15 (Fig. 1B, P<0. 001). To confirm that DN-Cav1 expression impacted caveolar endocytosis, murine leukemia virus 10A1 (MLV-10A1) infection and cholera toxin B subunit (CTxB, Pubchem: 53787834) uptake were measured as both processes are known to require CavME [28], [29]. 10A1 infection of cells expressing DN-Cav-1 was reduced by half (Fig. 1C) and is consistent with previously reported observations [28]. Uptake of CTxB (Fig. 1D) was more strongly inhibited, as cells expressing DN-Cav1 had little CTxB inside the cytoplasm, indicating that DN-Cav1 was functional, blocking caveolar endocytosis. As a further test, colocalization of internalized GFP-labeled ZEBO-VLPs (gfpZEBO-VLP) with established markers of CME (clathrin light chain A; OMIM: 118960 and transferrin; OMIM: 190000) or CavME (caveolin-1) pathways was examined. Confocal microscopy revealed no significant colocalization of gfpZEBO-VLP with any of the markers used (Fig. 1E). Similar results were obtained when Vero cells were used (not shown). Taken together, the above findings indicated that neither CME nor CavME plays a major role in entry and infection of ZEBOV into HEK293T or Vero cells. Dynamin (OMIM: 602377) is a large GTPase and plays a critical role in numerous endocytic pathways including CME and CavME as well as some of the non-clathrin/non-caveolin-dependent (NC) pathways [8]. Dynamin acts by mediating the release of newly-formed endocytic vesicles from the plasma membrane. To determine ZEBOV dependence on dynamin, the effect of dynasore (Pubchem: 56437635), a potent and specific dynamin inhibitor [30] was tested. A recombinant infectious ZEBOV that encodes GFP (gfpZEBOV) was used. This virus is comparable to wild-type ZEBOV in terms of replication and cytopathic effects (CPE) in cultured cells but has been engineered to express GFP as an infection marker [31]. As control, a recombinant infectious VSV that encoded red fluorescent protein (rfpVSV) was used. Dynasore treatment of Vero cells greatly reduced rfpVSV infection but failed to have any significant effect on infection by gfpZEBOV even at the highest concentration tested (Fig. 2A, B). Similar results were obtained in HEK293T cells (not shown). This result was confirmed using the VLP-based entry assay. Just like the gfpZEBOV, entry of ZEBO-VLP was unaffected at any of the doses used while VSV-VLPs were strongly inhibited by dynasore in a dose-dependent manner (Fig. 2C). To ensure that dynasore inhibited dynamin-mediated endocytosis, its effect on internalization of transferrin (CME marker) or CTxB (CavME marker) was determined. Confocal microscopy revealed that treatment reduced internalization of both markers by >80% and 96% respectively in Vero cells (Fig. 2D). As a further test of the dynamin independence of ZEBOV infection cells were made to express a DN form of dynamin-2 (Dyn2-K44A) and VLP entry assays were performed. As with dynasore, there was a significant drop in the entry of VSV-VLP in cells transfected with Dyn2-K44A (P<0. 05). In contrast, the entry of ZEBO-VLPs actually increased significantly (P<0. 05), suggesting that the suppression of dynamin function may enhance entry of ZEBOV (Fig. 2E). Furthermore, the majority of GFP labeled ZEBO-VLPs (gfpZEBO-VLPs) did not colocalize with endogenous dynamin at any point up to 60 min after cell contact (Fig. 2F). These findings indicated that cell entry of ZEBOV is independent of dynamin and that dynamin activity may actually redirect virus to a non-productive pathway. Many of the cellular endocytic pathways including CavME, macropinocytosis and certain NC pathways occur in cholesterol-rich membrane microdomains such as lipid rafts. An earlier study suggested that lipid rafts may play a role in ZEBOV infection [32]. Consistent with this, we found significant co-localization of gfpZEBO-VLPs with lipid rafts during entry (Fig. 3A). Furthermore, methyl-β cyclodextrin (Pubchem: 3889506) or nystatin (Pubchem: 6433272), which disrupt lipid rafts by extracting or sequestering cholesterol out of the plasma membrane, respectively were able to block gfpZEBOV infection in a dose-dependent manner (Fig. 3B). Similar effects of both drugs were observed when tested using the ZEBO-VLP-based entry assay (not shown). These data indicated that cholesterol-rich lipid raft domains are the likely site of ZEBOV entry. The above findings clearly indicated that ZEBOV uptake occurs through a dynamin-independent, lipid raft-dependent, non-clathrin/non-caveolar endocytic mechanism but remains cholesterol dependent. Macropinocytosis is one pathway that is known to be cholesterol-dependent, but independent of clathrin, caveolin and dynamin and has been shown important for uptake of vaccinia virus into cells as well as bacteria [9]. Also, our previous work indicated involvement of PI3K and Rac1 (a rho family GTPase) in ZEBOV entry and infection [25]. Work by others had also indicated involvement of Rho GTPase in Ebola virus entry [23]. Each of these signaling proteins is also thought to be important for macropinocytosis [13], [33]. To assess the involvement of macropinocytosis, the effect of EIPA (5- (N-ethyl-N-isopropyl amiloride; Pubchem: 1795) on ZEBOV infection was determined. EIPA, an amiloride, is a potent and specific inhibitor of Na+/H+ exchanger activity important for macropinosome formation [34], [35], [36], [37]. Consistent with this activity, EIPA caused a significant reduction (>80%) in the uptake of high molecular weight dextran, a marker of macropinosomes (Fig. 4A and B). When tested in Vero cells, a dose-dependent inhibition of gfpZEBOV infection was observed in the presence of EIPA (Fig. 4C, top panels; Fig. 4D), while infection by VSV was not significantly affected (Fig. 4C, middle panels; Fig. 4D). As used, EIPA had no significant cytotoxic effect as assessed by cell monolayer integrity (Fig. 4C, bottom panels). Counting of infected cells revealed that VSV infection was inhibited by 30% but increasing the dosage of the drug did not further reduce infection, indicating a small portion of VSV infection may occur through an EIPA-sensitive pathway. In contrast, the majority of ZEBOV infection inhibition was dosage dependent, potent and indicative of inhibition of a single uptake pathway (Fig. 4D). Similar results were observed in HEK293T cells (data not shown). To rule out the possibility that EIPA blocked ZEBOV infection at a post-entry step, the VLP entry assay was used. Here, EIPA treatment had no significant effect on entry of VSV-VLP (P<0. 05) while the level of ZEBO-VLP entry was inhibited similarly to that seen for infectious virus (Fig. 4E). The impact of EIPA on virus binding to cells was also tested. Cells were pretreated with EIPA and then incubated with luciferase-containing ZEBO-VLPs for 10 min on ice. Unbound particles were washed away and then the amount of VLP associated with cells was measured by lysis in non-ionic detergent to release virus-encapsulated luciferase. Compared to DMSO-treated (control) cells, no significant difference was observed in luciferase activity in samples that were treated with EIPA, indicating that ZEBO-VLP binding to cells was unaffected (Fig. 4F). Finally, to directly visualize the effect of EIPA on virus uptake, Vero cells treated with DMSO or EIPA were incubated with gfpZEBOV-VLPs. Confocal microscopy revealed that there was a marked drop (3. 5-fold) in gfpZEBO-VLP uptake in cells treated with EIPA as compared to that in DMSO-treated cells (Figs. 4G and H). Vaccinia virus was shown to induce fluid phase uptake and exhibit colocalization with fluid phase markers such as dextran [36]. To see if a similar induction of fluid phase uptake was seen with ZEBOV, dextran and virus were incubated together on cells and examined by confocal microscopy. Starting within 10 min and continuing until at least 60 min post-binding, at any one time, approximately 20% of internalized VLPs overlapped with dextran (Fig. 4I). An additional 20–30% of the remaining VLPs were also found juxtaposed to vesicles containing dextran, indicating a close association with this compartment. Furthermore, while performing these experiments we observed that cells incubated with ZEBO-VLPs appeared to have more dextran containing vesicles than intact cells. Indeed, when studied in detail, a (2–3 fold) increase in the number of dextran-containing vesicles per cell was seen after incubation with ZEBOV as compared to cells incubated with VSV or medium alone (Fig. 4J). A similar outcome was seen for cells incubated with VLPs (not shown). The above data indicated that cellular uptake of ZEBOV primarily occurs through a pathway that has characteristics of macropinocytosis. Another hallmark of macropinocytosis is its dependence on the activity of Pak1 [9]. Therefore, the role of Pak1 in entry of ZEBOV was investigated. First, we measured the effect of siRNA-mediated suppression of endogenous Pak1 and found that the infection of gfpZEBOV was significantly reduced in cells transfected with Pak1 siRNA (Fig. 5A). To confirm, cells transfected with plasmid encoding wild-type Pak1 or DN Pak1 were challenged with gfpZEBOV. The infection of gfpZEBOV was reduced >95% in cells expressing the DN Pak1 protein than in cells expressing the wild-type Pak1 protein (Fig. 5B). A similar effect of DN Pak1 was observed when ZEBO-VLPs were tested in the entry assay (not shown). The protein CtBP/BARS, is also known as important for macropinocytosis [15], [16] and substitutes for dynamin in promoting vesicle scission from the plasma membrane. Again, siRNA were used to suppress expression and the impact on infection measured. Two independent siRNA were able to suppress expression of CtBP/BARS by >70% and 80% respectively (Fig. 5C, left panels). Infection was also reduced by 50% and 90% respectively (Fig 5C, right). This observation indicated that the suppression of CtBP/BARS expression must cross a threshold before becoming limiting to ZEBOV infection but plays an important role. Both sets of data support roles for Pak1 and CtBP/BARS in ZEBOV infection. Macropinocytosis is heavily actin-dependent. Actin is required for the formation of plasma membrane ruffles in macropinosome formation, as well as trafficking of macropinosomes into the cell [38]. Ligands that utilize macropinocytosis often promote changes in the cell actin dynamics by regulating various cellular proteins involved in controlling F-actin assembly and disassembly. Arp2 protein is an integral component of a multi-protein complex that serves as a nucleation site for de novo actin assembly. We observed a significant increase in the size of Arp2-containing complexes shortly after ZEBOV binding to cells (Fig. 5D, E). Analysis of the data indicated a 2-fold increase in the number of large (>0. 25 µm2) Arp2-containing complexes (Fig. 5D). A similar outcome was seen with cells incubated with VLPs (not shown) and a significant proportion of VLPs were associated with Arp2 complexes (Fig. 5F). Further support for a role of actin in ZEBOV entry came from the observation that gfpZEBO-VLPs were associated with F-actin foci within the interior of the cell but this was not seen for VSV-VLPs (Fig. 5G). Similarly, the gfpZEBO-VLPs were also seen associated with vasodilator-stimulated phosphoprotein (VASP), an actin-associated protein that promotes actin nucleation (Fig. 5H). In each of these cases, VLPs and staining for each marker often did not completely overlap. Instead VLP and actin or VASP often were closely juxtaposed and is consistent with nucleation occurring around vesicles containing the VLP. These observations suggested that the virus actively promotes actin assembly and associates with actin-based structures to facilitate its uptake and/or trafficking. The above findings indicated that ZEBOV is primarily internalized by a macropinocytosis-like pathway in Vero and HEK293 cells. However, the subsequent trafficking route to the site of penetration into the cytoplasm remained unknown. We found that fluorescently-labeled ZEBOV particles significantly co-localize with early endosomal antigen-1 (EEA1; OMIM: 605070) shortly after incubation with cells (Fig. 6A). At any time up to 60 min after the start of incubation, more than 30% of VLPs were associated with this marker (Fig. 6B). This confirmed a role for endocytic uptake into cells and suggested that following internalization, ZEBOV is delivered to an EEA1-positive compartment, likely sorting endosomes. Typically, the cargo from EEA1-positive compartments is delivered to early endosomes followed by trafficking to late endosomes. These vesicles are characterized by the presence of Rab5 (OMIM: 179512) and Rab7 (OMIM: 602298) GTPases on the cytoplasmic face of the vesicle, respectively, which play a key role in regulating their trafficking. Consistent with a role for early and late endosomes and in contrast to the lack of effect of DN Eps15 and Cav-1 expression, GFP-tagged DN Rab5 or DN-Rab7 resulted in significant reduction (P<0. 001 for each) in infection by gfpZEBOV (Fig. 6C). To determine if the effect was due inhibition of virus entry, VLP entry assays were performed. As compared to the negative control (GFP alone), wild-type Rab5 had no significant effect on entry of either ZEBO-VLP or VSV-VLPs, while there was >50% reduction in entry of both ZEBO-VLPs in cells transfected with either DN-Rab5 or DN-Rab7. The level of entry inhibition seen for ZEBO-VLP was similar to that of VSV-VLPs (Fig. 6C) and indicated that like VSV, ZEBOV is taken up by Rab5-dependent early and Rab7-dependent late endosomes. However, this does not mean that both virus types are present in the same vesicle population but that similar trafficking proteins are required at this stage of endocytosis. Currently, it is unknown whether ZEBOV envelope fusion occurs in late endosomes or further trafficking to a different compartment is needed.
Endocytosis offers an efficient way for viruses to cross the significant physical barrier imposed by the plasma membrane and to traverse the underlying cortical matrix. Viruses have also evolved to target distinct endocytic pathways that are capable of delivering the capsid into the cell cytoplasm at sites suitable to initiate replication and to avoid destructive compartments like the lysosome. Understanding the pathway of virus entry and deciphering the mechanism regulating it is important for understanding viral pathogenesis as virus entry into host cell is the first critical step in pathogenesis of infection. While there is ample evidence that ZEBOV enters cells through endocytosis in a pH-dependent manner [6], [7], [20], the specific endocytic and trafficking pathways have not been clearly defined. Previous studies to elucidate the ZEBOV entry pathway have produced conflicting findings. Most of these studies relied on the use of retrovirus-based pseudotypes in which the Ebola virus GP is coated onto the surface of a retrovirus capsid containing a recombinant genome. The use of this system overcomes the need for high bio-containment but suffers from not having native virus morphology, GP density, and other biochemical characteristics. One early study on ZEBOV uptake using pseudotyped virus indicated caveolae as important [22] but later work indicated that caveolin activity was not required [39]. In contrast, a recent study concluded that clathrin-mediated endocytosis was the major entry pathway for ZEBOV [21]. While multiple approaches were used, including dominant-negative mutant expression and siRNA to specifically disrupt clathrin-mediated endocytosis, the key data was obtained using lentivirus-based retroviral pseudotypes. In comparison, previous work using wild type virus [20] implicated both clathrin and caveolar endocytosis in entry of ZEBOV. However, only pharmacological inhibitors were used in this study and drug specificity was not examined, making interpretation difficult. Indeed, the only evidence of clathrin involvement in infection was provided using chlorpromazine. Chlorpromazine is a useful drug and there is ample evidence indicating that it disrupts clathrin-coated pits, but it has recently been demonstrated to also interfere with biogenesis of large intracellular vesicles such as phagosomes and macropinosomes [24]. Here, by combining distinct and independent approaches we have performed a detailed analysis of each major endocytic pathway and have obtained, a clear and accurate picture of how ZEBOV enters the cell and identified important cellular proteins that are required. Careful assessment of specificity and functionality of each pathway was performed and correlated to infection and virus uptake. Replication-competent infectious ZEBOV, as well as ZEBO-VLPs (which are morphologically similar to infectious ZEBOV and contain the native matrix protein in addition to GP) were used to study the virus entry mechanism. Drugs were used to inhibit pathways but issues of specificity and pleiotropy were assessed by testing the function of each pathway after treatment. This was done by using independent markers such as transferrin, CTxB and high molecular weight dextran for CME, CavME and macropinocytic uptake respectively. We also assessed the association of fluorescent VLPs with each marker as well as markers of each endocytic compartment being examined. Furthermore, highly specific dominant-negative mutants and/or siRNAs were also used to corroborate the data obtained by pharmacological inhibitors. Importantly, throughout this work a sensitive contents-mixing virus entry assay was used in discriminating against blocks in virus entry versus blocks in downstream steps in the infection cycle. This is particularly important to do when using drugs that often affect multiple cellular functions. It is noteworthy that in each case, virus infection with wild type or the GFP-expressing ZEBOV correlated exactly with the outcomes of the VLP-based assays. This approach gives a highly detailed view of the mechanism of ZEBOV uptake into cells. Unlike previous studies [20], [21], [22], we found no evidence for the involvement of either CME or CavME in ZEBOV entry and infection. However, there was strong association of fluorescently-labeled ZEBO-VLPs with lipid rafts, and a marked reduction of ZEBOV infection by MBCD or nystatin, as reported previously [32]. This signified that cholesterol-rich lipid raft domains are required for productive entry of the virus. However, cholesterol-rich membrane microdomains play important roles in many forms of endocytosis including caveolae-dependent, non-clathrin/non-caveolar pathways, and macropinocytosis [38], [40]. Our previous work indicated that entry of ZEBOV was dependent on signaling through PI3K and Rac1 [25], which are important regulators of macropinocytosis [38]. Work by others also showed that Rho GTPases play a role in ZEBOV uptake [23]. Each of these cellular signaling proteins are known to be important in macropinocytosis. Macropinocytosis is also distinguished from the other pathways principally by criteria that include actin-dependent structural changes in the plasma membrane, regulation by PI3K, PKC, Rho family GTPases [9], [13], [33], Na+/H+ exchangers, Pak1, actin, actin regulatory factors, involvement of CtBP/BARS [14], [38] as well as ligand-induced upregulation of fluid phase uptake and colocalization of the internalized ligand with fluid phase markers [14], [36], [37]. In our examination of ZEBOV entry mechanism, we found that EIPA, a potent and specific inhibitor of the Na+/H+ exchanger [34], [35], [36], [37] blocked ZEBOV infection and entry. Furthermore, ZEBOV caused significant induction of dextran uptake (a fluid phase marker) and the internalized virus particles colocalized with dextran. Pak1 regulates macropinocytosis by promoting actin remodeling and macropinosome closure through phosphorylation of proteins LIMK and CtBP/BARS, respectively [9], [16]. We found that suppression of both Pak1 and CtBP/BARS activity by siRNA or expression of a DN form of Pak1 reduced virus entry and infection. Actin plays a central role in formation and trafficking of macropinosomes. Actin remodeling is a key event during macropinocytosis and is often triggered by stimuli that promote macropinocytosis. Arp2, among other actin regulatory proteins, has been implicated in macropinocytosis [9]. Arp2 also plays an important role in actin remodeling. It is an integral component of a large multi-protein complex that forms in response to stimuli that trigger actin assembly, and serves as a nucleation site for assembly of actin monomers to form F-actin [41]. We observed a significant increase in the size of the Arp2-containing complexes shortly after ZEBOV binding to cells, indicating stimulation of actin nucleation by the virus. The increase in Arp2 nucleation paralleled an increase in large dextran containing vesicles inside cells corresponding to macropinosomes. This activity appears to be associated with the ZEBOV glycoprotein as VLPs were also capable of inducing a similar increase in Arp2 nucleation and dextran uptake (data not shown). Additionally, we found marked association of fluorescently-labeled ZEBO-VLPs with F-actin foci, as well as with the Arp2-containing complexes and actin-regulatory protein, VASP, that resides in membrane ruffles and promotes actin foci formation. Together, these data provide evidence for a role of actin in ZEBOV entry and suggest that the virus can actively promote localized actin remodeling to facilitate its uptake through macropinocytosis or a similar mechanism. Despite using multiple approaches, we found no evidence for a role of dynamin in ZEBOV entry. Dynamin is a large GTPase that is involved in scission of newly-formed endocytic vesicles at the plasma membrane [42], [43]. Dynamin-independent entry of ZEBOV further ruled out roles for clathrin or caveolae-mediated pathways as both require dynamin activity [38], [44], [45]. In contrast, the majority of studies suggest that macropinocytosis is independent of dynamin [9]. Recently a novel mechanism has been described for scission of shigatoxin-containing vesicles in which Arp2-dependent actin-triggered membrane reorganization directly leads to vesicle severance [46]. As indicated above, we observed a marked increase in the size of Arp2 complexes shortly after incubation with ZEBOV and a significant association of ZEBO-VLPs with these and F-actin foci but it is unclear if this resulted in membrane scission. In addition, several reports have indicated that C-terminal binding protein (CtBP/BARS), originally identified as a nuclear transcription factor, likely replaces dynamin in scission of nascent macropinosome from the plasma membrane [15], [16]. As discussed above, suppression of CtBP/BARS by siRNA reduced infection and is consistent with the requirement for macropinocytosis in ZEBOV infection. Interestingly, ZEBOV VP40-based VLPs bearing VSV envelope glycoprotein were found to enter cells through clathrin-mediated endocytosis, as has been reported for the wild-type virus [27]. This suggested that the choice of the internalization pathway is primarily determined by envelope glycoprotein specificity. This is in contrast to a study in influenza virus, where profound differences were seen in the entry characteristics of early passage filamentous virus compared to the laboratory grown spherical isolates that tend to use clathrin-mediated endocytosis [47]. These data indicated a more pronounced role of virion morphology on the choice of endocytic pathway. The apparent reason for this discrepancy is not clear but may relate to the differences in biological characteristics of the viruses and/or cell types used in the two studies. Overall, our data provide strong evidence that in HEK293T and Vero cells infection by ZEBOV occurs by a process that is closely related to macropinocytosis. We cannot say that entry occurs exclusively by this pathway, but that its disruption blocks the majority of infection and particle uptake. Our work also indicates that clathrin and/or caveolar endocytosis play at most, only a minor role in infection by wild type virus. A few other viruses as well as bacteria, require macropinocytosis to establish infection [14]. Each uses different mechanisms to induce macropinocytosis. Vaccinia virus has been shown to trigger macropinocytosis by mimicking apoptotic bodies [36]. In contrast, Coxsackie virus and adenovirus activate macropinocytosis by binding to the cell surface proteins occludin and integrin αV, respectively [48], [49]. The mechanism by which ZEBOV triggers macropinocytosis is currently unknown but likely involves GP interaction with cell receptors. Axl (a receptor tyrosine kinase) and integrin βI have been suggested to act as virus receptors [50], [51]. Although, the role of Axl or integrin βI has not been studied in the context of macropinocytosis, there is evidence that several other receptor tyrosine kinases and integrins can trigger macropinocytosis [52], [53], [54]. Therefore, it will be important to analyze the role of Axl and/or integrin βI in this context. After formation, macropinosomes traffic further into the cytoplasm and may acquire new markers and/or undergo heterotypic fusion with other vesicles of the classical endolysosomal pathway thereby successively transferring the cargo to more acidic compartments such as early and late endosomes [38], [55]. Consistent with this, we found that ZEBO-VLPs co-localized with EEA1-positive vesicles soon after binding [25]. Interestingly, the timing of colocalization of VLPs with EEA1 positive vesicles coincided with their appearance in dextran-containing macropinosomes (within 10 min after binding). Possible explanations may be that the macropinosomes acquire EEA1 shortly after formation or that they undergo prompt fusion with EEA1 positive vesicles. Our data also provided evidence that ZEBOV infection and entry was dependent on Rab5 and Rab7 function, indicating the involvement of early as well as late endosomes in ZEBOV uptake and infection. While a role of early endosomes in Ebola virus entry has not been previously reported, our finding that ZEBOV is trafficked to late endosomes is consistent with prior studies that showed inhibition of Ebola pseudovirion infection by dominant-negative Rab7 [56] and proteolytic processing of Ebola GP1 by late endosome-resident cathepsins [6], [7]. However, it is important to note that many distinct endocytic vesicles associate with Rab5 and Rab7 during maturation but differ by the ligands they carry [57]. This explains why transferrin, a marker of CME, was never seen associated with ZEBOV containing vesicles, even though both require Rab5 and Rab7 for endocytosis. The intracellular trafficking of the macropinosome is not well understood and existing data provide evidence both for and against the involvement of classical endolysosomal pathway [58]. However, little mechanistic information is available with respect to virus entry by macropinocytosis. Prior to our work only one study analyzed trafficking in any detail, using vaccinia virus and found that virus particles did not colocalize with markers of classical endolysosomal pathway [59]. This difference is likely due to the fact that ZEBOV requires transport to an acidic compartment for membrane fusion while vaccinia virus, which is a pH-independent virus, may undergo nucleocapsid release prior to fusion of macropinosomes with more acidic compartments of the endolysosomal pathway. Our findings now add novel and valuable information regarding macropinosome trafficking mechanism in general and in the context of virus entry. In conclusion, the evidence presented here demonstrates that ZEBOV utilizes a macropinocytosis-like pathway as the primary means of entry into HEK293T and Vero cells. Once taken up by endocytosis, virus trafficking occurs through early and then late endosomes; however, the exact site where envelope fusion and nucleocapsid release occur is unknown. We do not know if ZEBOV and other filoviruses follow the same pathway into other cell types, like macrophages, that are thought to be a primary target for infection. However, most cell types are capable of macropinocytosis and it is likely that the same or a similar pathway will be used. These findings are important as they not only provided a detailed understanding of ZEBOV entry mechanism, but also identified novel cellular factors that may provide new potential targets for therapies against this virus. It will be important to determine if other filoviruses share the same pathway. If so, it may be possible to develop broad-spectrum therapies that temporarily block this pathway in cells.
Human Embryonic Kidney HEK293T and Vero cells were maintained in Dulbecco' s modified Eagle' s (DMEM) medium supplemented with 10% fetal bovine serum (Gemini Bioproducts, GA), 1% non-essential amino acids (Sigma, MO) and 1% penicillin-streptomycin solution (Sigma, MO). All pharmacological inhibitors were purchased from Calbiochem (San Diego, CA) or Sigma (St. Louis, MO). Stock solutions were prepared either in water, DMSO, or methanol, as per manufacturer' s recommendation, and stored at −80°C in small aliquots. Alexafluor-labeled reagents including cholera toxin B subunit, transferrin, dextran (10,000 MW) and secondary antibodies were from Invitrogen (Eugene, OR). Specific antibodies against clathrin light chain, caveolin, dynamin, cholera toxin B, Arp2, CtBP/BARS, phospho-VASP and Pak1 were purchased from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA) or Cell Signaling Technology (Beverly, MA). siRNA were from Qiagen (Valencia, CA) and transfections were performed using Dharmafect transfection reagent 1 according to the manufacturer' s (Dharmacon, Lafayette, CO) instructions. 4–6 pmol of siRNA were used per transfection of cells in 0. 1 ml of medium per well of a 96-well plate. Assays were performed 48 h after transfection. All plasmids were prepared using Qiagen kits or by CsCl gradient centrifugation following standard procedures. The plasmid encoding VSV-G envelope glycoprotein (pLP-VSVG) was purchased from Invitrogen. Construction of the plasmid encoding Nef-luciferase fusion protein (pCDNA3-nef-luc) has been described previously [60]. Plasmids encoding ZEBOV matrix protein (VP40), ZEBOV envelope glycoproteins were kindly provided by Christopher Basler (Mount Sinai School of Medicine), Paul Bates (University of Pennsylvania), and Luis Mayorga (Universidad Nacional de Cuyo, Argentina) respectively. Plasmids expressing dominant-negative Eps15, caveolin-1, and dynamin-2 K44A have been described previously [19]. The Pak1 expression plasmids were obtained from Addgene (Cambridge, MA). ZEBOV-VLPs were produced by co-transfecting HEK293T cells with plasmids encoding ZEBOV matrix (VP40) protein, ZEBOV envelope glycoproteins, and Nef-luciferase fusion protein using the calcium phosphate method. For VSV-VLP, plasmid encoding ZEBOV glycoproteins was replaced with one encoding VSV-G. Cell culture supernatant was collected 48 h after transfection and cell debris was cleared by centrifugation (1,200 rpm for 10 min at 4°C). Subsequently, VLPs were purified by centrifugation (25,000 rpm in SW28 rotor for 3. 5 h at 4°C) through a 20% (w/v) sucrose cushion in PBS. The VLP pellet was resuspended in 0. 01 volume of DMEM, aliquoted and stored at 4°C. Assays were performed within 2–3 days after purification of VLPs. HEK293T cells were used for contents mixing assays to measure nucleocapsid release into the cell cytoplasm. The cells were removed from plates by trypsin treatment, pelleted by centrifugation and then resuspended in fresh medium. Cells (106 per assay point) were mixed with nef-luciferase containing VLPs in a volume of 0. 2 ml and incubated at 37°C on a rotating platform for 3 h. Subsequently, the cells were washed 2–3 times with DMEM to remove the unbound VLPs and the final cell pellet was resuspended in 0. 1 ml of luciferase assay buffer lacking detergent (Promega, WI). Luciferase activity was then measured using a Turner Design TD 20/20 luminometer and expressed as counts/sec. To study drug activity on virus entry, cells were pre-treated with drug for 1 h, followed by incubation with VLPs in the continued presence of the drug. Virus entry was then measured as described above. For measuring the effect of ectopic gene expression, cells were transfected with the control plasmid or one encoding the protein of interest. Cells were then used for entry assays 36 h after transfection. Typical transfection efficiency was 50–70%. Wild type ZEBOV (Mayinga strain) was provided by Michael Holbrook (UTMB, TX) and the recombinant virus encoding GFP (gfpZEBOV) was from Heinz Feldman (NIH, Rocky Mountain Laboratory, MT). The virus was cultivated on Vero-E6 cells by infection at an MOI of approximately 0. 1. All infected cells expressed GFP approximately 24 h post-infection. Culture supernatants were collected after 7 d and clarified by centrifugation at 2000 x g for 15 min. Virus titer was determined by serial dilution on Vero-E6 cells. Cells were incubated with virus for 1 h and then overlaid with 0. 8% tragacanth gum in culture medium. 10 d post-infection cells were fixed with formalin, and stained with crystal violet 10 d post-infection for plaque counting. All experiments with ZEBOV were performed under biosafety level 4 conditions in the Robert E. Shope BSL-4 Laboratory at UTMB. Cells were pre-treated with inhibitors for 1 h and then incubated with gfpZEBOV at 37°C for 2 h (except in the case of MBCD and nystatin, where cells were washed to remove the inhibitors prior to incubation with the virus). Subsequently, the unbound virus particles were removed by washing with PBS, and cells incubated in fresh growth medium. Twenty-four h later, cells were washed and fixed with 10% formalin for 48 h. Images were taken by epifluorescence microscopy and infected foci counted. Counting was performed using the Cell Profiler software package [61]. The processing pipeline used by the software is available upon request. HEK293T or Vero-E6 cells were cultivated overnight on chambered coverglass slides (Nunc, Rochester, NY) at a density of 50%. The following day, cells were incubated with GFP-tagged ZEBO-VLPs. Cells were then washed three times in DMEM and fixed in 4% fresh paraformaldehyde in PBS. After one wash in PBS residual paraformaldehyde was neutralized by addition of 0. 1 M glycine buffer, pH 7. 4 and cells were permeabilized using 0. 1% Triton X-100 for 1 min at room temperature. For immunofluorescence, cells were incubated with the appropriate primary antibody, typically diluted 1∶200 in PBS. After washing in PBS, the cells were then incubated with the indicated secondary Alexafluor conjugated secondary antibody. Cells were imaged using a Nikon TE Eclipse inverted microscope with a 100× oil immersion lens or a Zeiss LSM 510 confocal microscope in the UTMB optical imaging core. Cells were transfected with plasmids encoding GFP-tagged forms of the protein of interest. For work with Pak1, myc-tagged or GST-tagged expression constructs were also used. After 24 h, the cells were challenged with wild type ZEBOV. After an additional 48 h, the cells were fixed in formalin. Cells were then stained for ZEBOV VP40 using a rabbit polyclonal antiserum (Ricardo Carrion, Southwest Foundation for Biomedical Research, San Antonio, TX) followed by an Alexa633 secondary antibody. For Pak1 work, cells were also stained for myc-tag or GST-tag using the corresponding primary and a fluorescently labeled secondary antibody. Cell nuclei were also stained using DAPI (Invitrogen). Images were taken using an epifluorescence microscope and the intensity of GFP fluorescence and VP40 staining was evaluated on a cell per cell basis using the Cell profiler software package [61]. For this, cells were first identified by DAPI staining of the cell nuclei. Then cytoplasmic fluorescence intensity for GFP and VP40 staining was determined. The algorithm pipeline used for this part of the analysis is available from R. Davey upon request. The output data, which gives intensities on a scale of 0 to 1, was converted to a scale of 0–1024 using Excel. This data file was then converted to a text file and processed using A2FCS software, which is part of the MFI/FCS Verification Suite (Purdue University) and is available at http: //www. cyto. purdue. edu/flowcyt/software/Catalog. htm. This conversion makes the data accessible to conventional FACS analysis software. The data were then analyzed using FlowJo v7. 5 (http: //www. flowjo. com). Gates were set to exclude cells that were not infected and not expressing the tagged protein, as determined in control experiments. These were done for normal cells infected by virus but not stained, cells not infected by virus but stained with VP40 specific antibody and cells expressing the tagged protein and not infected (stained with antibody against the tagged protein when used). To quantitate the infection dependency of ZEBOV on expression of each construct, the proportion of cells that were expressing each tagged protein construct and infected by ZEBOV was calculated as a fraction of the total cell population. While not used here, this analytical approach can be extended further by setting gates for low, moderate and high levels of ectopic gene expression and then correlating the outcome on infection. To measure the effect of DN-Cav1 on 10A1 MLV infection, HEK293 cells were transfected with plasmid encoding GFP or GFP tagged DN-Cav1 protein. Thirty six hours after transfection cells were infected with 10A1 MLV pseudotype encoding truncated CD4 receptor (Miltenyi, Germany) as a marker for infection. 36 h after infection the cells were stained for CD4 expression with PE-labeled mouse anti-human CD4 antibody (BD Pharmingen Cat#555347). After 1 h the cells were washed in PBS and fixed in 4% paraformaldehyde. Cells were stained with DAPI to identify cell nuclei and were imaged by a Nikon TE eclipse microscope with an automated motorized stage. To analyze the effect of DN-Cav1 on infection, images were analyzed using Cell Profiler software (Broad Institute, Cambridge, MA) to detect total cells, cells expressing the expression construct and those infected by detection of CD4. Analysis was then performed as above. | Filoviruses, including Zaire ebolavirus (ZEBOV), are among the most pathogenic viruses known. Our understanding of how these viruses enter into host cells is very limited. A deeper understanding of this process would enable the design of better targeted antiviral therapies. This study defines in detail, key steps of ZEBOV cellular uptake and trafficking into cells using wild type virus as well as the host factors that are responsible for permitting virus entry into cells. Our data indicated that the primary mechanism of ZEBOV uptake is a macropinocytosis-like process that delivers the virus to early endosomes and subsequently to late endosomes. These findings aid in our understanding of how filoviruses infect cells and suggest that disruption of macropinocytosis may be useful in treatment of infection. | Abstract
Introduction
Results
Discussion
Materials and Methods | virology/emerging viral diseases
cell biology/membranes and sorting
virology
microbiology
microbiology/cellular microbiology and pathogenesis
virology/host invasion and cell entry | 2010 | Cellular Entry of Ebola Virus Involves Uptake by a Macropinocytosis-Like Mechanism and Subsequent Trafficking through Early and Late Endosomes | 13,981 | 182 |
GcvB is an archetypal multi-target small RNA regulator of genes involved in amino acid uptake or metabolism in enteric bacteria. Included in the GcvB regulon is the yifK locus, encoding a conserved putative amino acid transporter. GcvB inhibits yifK mRNA translation by pairing with a sequence immediately upstream from the Shine-Dalgarno motif. Surprisingly, we found that some target sequence mutations that disrupt pairing, and thus were expected to relieve repression, actually lower yifK expression and cause it not to respond to GcvB variants carrying the corresponding compensatory changes. Work prompted by these observations revealed that the GcvB target sequence in yifK mRNA includes elements that stimulate translation initiation. Replacing each base of an ACA trinucleotide near the center of the target sequence, by any other base, caused yifK expression to decrease. Effects were additive, with some triple replacements causing up to a 90% reduction. The enhancer activity did not require the ACA motif to be strictly positioned relative to the Shine-Dalgarno sequence, nor did it depend on a particular spacing between the latter and the initiating AUG. The dppA mRNA, another GcvB target, contains four ACA motifs at the target site. Quite strikingly, replacement of all four ACAs by random trinucleotide sequences yielded variants showing over 100-fold reduction in expression, virtually inactivating the gene. Altogether, these data identify the ACA motif as a translation-enhancing module and show that GcvB' s ability to antagonize the enhancer function in target mRNAs is quintessential to the regulatory effectiveness of this sRNA.
A relevant chapter in the expanding field of RNA-mediated gene regulation is devoted to the activities of multi-target trans-encoded small RNAs in bacteria. Acting in concert with chaperon protein Hfq, these RNA regulators function by base-pairing with short, often imperfectly complementary sequences in the 5′ untranslated regions (UTR) of target messenger RNAs. They can affect translation and turnover of several mRNAs simultaneously thus reprogramming gene expression of whole gene families in a coordinate manner in response to environmental cues (reviewed in [1]–[3]). Archetypal examples of this class of regulators are the RyhB small RNA (sRNA) which represses expression of mRNA for dispensable iron-sequestering proteins when iron is limiting [4]–[8]; RybB, which downregulates several outer membrane protein mRNAs under envelope stress conditions [9]–[13], Spot 42, which amplifies the regulatory range of catabolite repression by targeting several mRNAs involved in sugar uptake and consumption [14] and GcvB, which downregulates dozens of different mRNAs involved in amino acid uptake or metabolism in E. coli and Salmonella [15]–[18]. GcvB, a 200 nucleotide-long sRNA, was identified serendipitously during a study of gcvA, the gene for the main transcriptional regulator of the glycine cleavage operon gcvTHP [18]. The latter encodes the enzymes of the glycine cleavage system, the pathway generating one-carbon units from the oxidative cleavage of glycine [19]. The gcvB gene is located immediately adjacent to gcvA in the opposite orientation with its promoter partially overlapping the gcvA promoter. In the presence of excess glycine, the GcvA protein activates transcription of the gcvTHP operon as well as of gcvB [18]. Initial characterization of GcvB showed this sRNA to downregulate the synthesis of DppA and OppA proteins, main components of dipeptide- and oligopeptide-transport systems, respectively [16], [18]. Since then, the number of genes found to be regulated by GcvB has increased exponentially. A recent transcriptomic study in Salmonella enterica set this number to more than 40, making the GcvB regulon the largest of its kind [17]. The vast majority of these loci are linked directly or indirectly to amino acid metabolism and are negatively controlled by GcvB. Typically, regulation is exerted during exponential growth in nutrient rich environments and possibly aimed at coordinating the expression of interconnected metabolic pathways [16], [17]; however, its precise role remains incompletely understood. GcvB uses a specific sequence region to pair with most, although not all [20] of its mRNA targets. This pairing domain – named the R1 region [16] – is characterized by its high degree of sequence conservation, the lack of secondary structure and a typical GU-rich sequence bias. Hence, most sequences targeted by GcvB include CA-rich repeats. They are typically found inside, or immediately adjacent to, the ribosome binding sites (RBS) of target mRNAs. In one of these targets - the gltI mRNA for a glutamate-aspartate transport protein – the CA-rich element is located 45 nucleotides (nt) upstream from the translation initiation codon. Removal of this sequence (as part of a 27 nt deletion), besides causing the loss GcvB regulation, affected gltI translation, suggesting that the CA-rich element acts as a translational enhancer. Consistent with this interpretation, crafting the 27 nt segment at the corresponding position of an unrelated mRNA conferred simultaneously GcvB control and increased translational efficiency [16]. Some years ago, our laboratory performed a lac-based genetic screen aimed at identifying genes controlled by trans-encoded small RNAs in Salmonella. A random library of lacZ fusions to chromosomal genes was generated using a phage Mu-derived transposon (MudK) and screened for isolates whose LacZ levels changed (either increased or decreased) upon inactivating Hfq [21]. Among the candidates that were found, two independent isolates upregulated in the hfq mutant background, carried the lacZ insert translationally fused to the yifK gene [21]. Presumptive identification of this gene as an amino acid transporter suggested that yifK might be a GcvB target. We thus proceeded to test this hypothesis and characterize yifK regulation. While this work was underway, Sharma and coworkers identified yifK mRNA as a member of the gcvB regulon by microarray analysis; however, these authors could not confirm direct regulation by GcvB due to low reporter fluorescence of the yifK-gfp fusion used in the study [17]. Since this study also identified global regulator Lrp as a GcvB target [17], the possibility remained that the GcvB effects on yifK expression might be indirect. Here we present in vivo and in vitro evidence that GcvB downregulates yifK directly by pairing with a sequence immediately preceding the Shine-Dalgarno (SD) motif in yifK mRNA. A surprising observation in the course of this study was that some target sequence mutations that disrupted pairing did not cause yifK expression to increase – as expected for the relief of GcvB repression – but had the opposite effect. The drop in expression was not suppressed by deleting gcvB nor was it accentuated in a GcvB mutant carrying the appropriate compensatory changes. Closer analysis revealed that the GcvB target sequence includes elements that stimulate yifK mRNA translation. In the absence of such elements, the role of GcvB pairing in regulation becomes marginal.
Our original screen for Hfq-regulated genes yielded two isolates carrying the MudK (lac) transposon in the yifK gene; one predicted to produce a LacZ protein fusion to the 48th amino acid (aa) of the 461 aa YifK (yifK87: : MudK); the other with LacZ inserted after the 95th aa of YifK (yifK88: : MudK) [21]. Preliminary tests showed both fusions to be regulated in a closely similar manner; however, yifK87: : MudK produced significantly higher ß-galactosidase activity and was chosen for the present study. A survey of protein sequence databases showed YifK to be a highly conserved protein with the characteristic signature of amino acid transporters. The known role of GcvB in the regulation of some members of this family made this small RNA the likeliest candidate to control yifK expression. This was confirmed by deleting the gcvB gene and testing the effects of the deletion on the expression of the yifK87: : MudK fusion (hereafter referred to as yifK-lacZY). As shown in Figure 1, the gcvB deletion causes a nearly 5-fold increase of ß-galactosidase activity in exponentially growing cells, while effects decline in stationary phase. Somewhat surprisingly, LacZ levels in the gcvB-deleted strain are not as high as the levels measured in a strain deleted for hfq (Figure 1). This might reflect the existence of one or more additional sRNA (s) participating in yifK repression. Alternatively, Hfq could repress yifK directly [22]. The data in Figure 1 show that loss of Hfq is epistatic to the gcvB deletion. Primer extension experiments mapped the 5′ end of yifK mRNA to 64 nucleotides upstream from the initiating AUG (Figure 2). This 5′ untranslated region (UTR) includes a 14-nt stretch complementary to the 3′ half of GcvB' s R1 region. As an initial step to characterize GcvB involvement in yifK regulation, we tested whether point mutations in the gcvB gene or in the promoter-proximal portion of yifK relieved GcvB-mediated repression. For this, DNA fragments spanning either of these two regions were randomly mutagenized by the polymerase chain reaction (PCR) under error-prone conditions and introduced into the chromosome of a strain harboring the yifK-lacZY reporter fusion via lambda red recombination. Most of the isolates originating from the gcvB mutagenesis carried changes in the gcvB promoter or in the promoter of the adjacent gcvA gene (Figure S1). Thus, these mutations appeared to lower the levels rather than the activity of GcvB and were not further considered. Mutagenesis of yifK promoter-proximal segment yielded three mutants with elevated yifK-lacZY expression. One isolate carried a C: G to A: T change 33 base-pairs upstream from the 5′ end of yifK mRNA. The position and the nature of the change (producing a -35 promoter consensus match, TTGACA, Figure 2A), strongly suggest that the mutation increases the strength of the yifK promoter. The mutation leads to a sharp rise in the intensity of the primer extension product (lane “-33A” in Figure 2B) and a more than 10-fold increase in ß-galactosidase activity (data not shown). These findings confirmed that the 5′ end identified by primer extension corresponds to yifK transcription initiation site. The remaining two mutations affect residues within the 5′ UTR (Figure 2A). One allele, resulting in a U to C change at position +21, falls within a AU-rich segment (AUAACAAUAA) that might constitute a site for Hfq binding [3], [23]. Consistent with this interpretation, the mutation has no effect in Δhfq background (Figure 2C). Finally the third allele (G to A at +27) affects the CG-rich stem of a presumptive secondary structure immediately adjacent to the AU-rich segment. The change causes a generalized increase of yifK-lacZY expression by an unidentified mechanism. Northern blot analysis was used to assess the effects of GcvB regulation on yifK mRNA levels. This study critically benefited from the availability of the -33 promoter mutant (see above), yifK mRNA being otherwise undetectable when expressed from the wild-type promoter (data not shown). The analysis identified two yifK mRNA species, a 1. 4 kilobase (Kb) mRNA covering just the yifK coding portion and a longer, 2. 0 Kb RNA extending into the adjacent argX-hisR-leuT-proM tRNA operon. As shown in Figure 3A, both RNAs accumulate upon RNase E inactivation, whereas only the shorter species accumulates in cells lacking GcvB or Hfq. This suggested that derepression of yifK translation in ΔgcvB or Δhfq cells protects the 1. 4 Kb RNA against RNase E cleavage. To confirm this interpretation, the analysis was repeated with strains that, besides the promoter “up” mutation, carried a mutation in the Shine-Dalgarno sequence (G to Cat position +59; described in more detail below). As shown in Figure 3C, the SD mutation causes the intensity of 1. 4 Kb band to sharply decrease in the ΔgcvB or Δhfq strains but not in the rne ts mutant, consistent with the idea that reduced translation renders yifK mRNA susceptible to RNAse E degradation. Absence of any obvious transcription termination signals in the intercistronic region between yifK and the tRNA operon suggests that the 1. 4 Kb RNA originates from processing of the longer form. Likely, under normal conditions (i. e. , wt yifK promoter) yifK transcription contributes only to a small fraction of the four tRNAs, as most the tRNA operon transcription results from a strong promoter located immediately upstream from the argX gene [24]. The approximately 500 nt RNA accumulates in the RNase E mutant (Figure 3B). Previous work in E. coli, showed that this tRNA precursor is processed by the concerted actions of RNase E and RNase P in a pathway that, intriguingly, also sees the participation of Hfq [25]. Early on in this study, it became apparent that yifK expression was exquisitely sensitive to the growth medium and virtually silenced in minimal medium. As a result, a strain with the yifK-lacZY fusion is phenotypically Lac− when plated in minimal medium. We exploited this phenotype to positively select for spontaneous Lac+ mutants. The selection yielded two classes of mutations, one genetically linked to the yifK-lacZY locus, the second mapping elsewhere. All of the linked mutants that were analyzed were found to harbor the -33 C: G to A: T promoter change obtained previously (see above). The unlinked mutations mapped in a chromosomal interval encompassing the gene for leucine response regulator, Lrp. Prompted by this observation, we introduced an lrp insertion mutation into the yifK-lacZY-containing strain. The resulting strain acquired a Lac+ phenotype (Figure 4A), indicating that yifK silencing in minimal medium results from Lrp repression. Addition of leucine efficiently relieves repression (Figure 4). The data in Figure 4 also show that GcvB does not contribute to yifK repression to any significant extent in minimal medium. This is not surprising as GcvB is transcribed at very low level under these conditions [16] and inactivating Lrp does not reverse this pattern (Figure S2). The data in Figure S2 differ from those of Modi et al [26] who reported an approximate 30-fold increase in GcvB levels in an lrp deletion mutant of in E. coli. This discrepancy might reflect differences in the organisms used or in media composition. The above approach yielded no mutations affecting the presumptive pairing sequences of GcvB or yifK. Reasoning that single base changes might not disrupt regulation enough to be revealed by the MacConkey plate screen, we resorted to introducing multiple changes by site-directed mutagenesis. An initial test involved changing a UGUG quadruplet in the GcvB segment thought to pair with yifK mRNA. The alteration caused expression of the yifK-lacZY fusion to increase approximately threefold, thus corroborating the postulated role of this sequence in yifK repression. Unexpectedly, however, when the ACAC sequence at the corresponding position in yifK mRNA was changed, yifK-lacZ expression did not increase but actually declined (data not shown). Trying to clarify this observation, portions of the region of interest were mutagenized separately. As shown in Figure 5A, converting the AAA sequence in the middle of the target sequence to UGU, or making the opposite change (UGU to AAA) in GcvB, similarly relieves yifK-lacZY repression. Repression is restored upon combining the compensatory alleles. Thus, this portion of the target sequence behaves as expected, and the behavior of the compensatory mutant strongly suggests that GcvB represses yifK through a base-pair interaction. In vitro toeprint experiments, showing that GcvB inhibits the binding of ribosomal 30S subunit to yifK translation initiation site, specifically and in a dose-dependent manner (Figure 6), provided independent support to this conclusion. Again, however, changing the CA doublet on the 3′ side of the AAA sequence to UC produced an unusual pattern: like in the quadruplet mutant above, yifK-lacZY expression decreased rather than increase, becoming insensitive to a GcvB variant carrying the compensatory change (Figure 5B). To verify that the compensatory change did not hamper GcvB' s function in an unpredictable way, we took advantage of the fact that the replaced nucleotides do not participate in the pairing with dppA [16] and tested the mutant' s ability to repress a dppA-lacZ translational fusion. This analysis showed both GcvB variants to be as efficient as wild-type in repressing dppA, indicating that both alleles remain fully functional (Figure S3). Besides being insensitive to the compensatory GcvB allele, the yifK CA49,50 to UC49,50 mutant fails to respond to gcvB or hfq deletions (Figure 7A). We interpreted these findings to suggest that the CA to UC conversion lowers translation efficiency and under such conditions, GcvB action is no longer rate-limiting for yifK expression. The effects of the mutation on yifK translation were examined in vitro using a reconstituted system. Results in Figure 7B showed an epitope-tagged Cat protein to accumulate at significantly greater levels when made from a gene fusion to the wt yifK 5′ UTR than from an equivalent construct carrying the CA to UC change. These data confirmed that the CA49,50 doublet stimulates translation and suggested that GcvB effectiveness in regulation reflects the targeting of an activating element. To characterize the enhancer element, an 8-nt segment preceding the SD sequence was modified by systematically changing individual residues to each of the three alternative bases. Effects on expression of the yifK-lacZ fusion were measured in a strain deleted for the gcvB gene. As shown in Figure 8, any change in the ACA sequence at positions +48 to +50 lowers yifK-lacZY expression. Variations range between 23% and 62%, with G residues exerting the most adverse effects at any position. Significantly, the effects appear to be additive since a separate experiment in which all three bases in the ACA sequence were randomized yielded alleles undergoing as much as 92% reduction in yifK-lacZY expression (bottom portion of Figure 8). Alteration of a second ACA motif between +53 and +55 produced a somewhat different pattern. Changes in the central C were either neutral or stimulatory; in contrast, having a C at the first position was highly deleterious resulting in nearly 95% reduction of ß-galactosidase activity (Figure 8). Altogether, these data suggested that ACA motifs enhance the efficiency of yifK translation. Conservation of these motifs in distant members of the Enterobacteriaceae family (Figure S4) is consistent with their functional importance. A peculiarity of the yifK translation signal is the unusually short distance (four nucleotides) between the most conserved base of the SD motif [27] and the initiating AUG. We thus envisaged that the role of the enhancer could be to somehow compensate for such suboptimal arrangement. To test this possibility, we generated a 7-nt tandem direct duplication of the SD region and then inactivated either copy of the SD by changing the GAGGA motif to GACGA (Figure 9). Thus, the resulting constructs have the functional SD sequence positioned either 4 or 11 nt from the AUG. As shown in Figure 9, these two variants (n. 4 and n. 5) express ß-galactosidase levels that are similar to each other and to the strain in which both SD are functional (n. 1). However, when the upstream ACA motif is replaced by GGG, yifK-lacZ expression drops sharply in both constructs (compare n. 4 ton. 6, and n. 5 ton. 7). Similar effects were observed in a separate construct where the segment between SD sequence and the initiating AUG was replaced by the sequence found at the corresponding position in the chiP gene [28] where the spacing (9-nt) is optimal (n. 8 and n. 9). While construct n. 8 is tightly repressed by GcvB, its variant lacking the upstream ACA (n. 9) shows a weak response to this sRNA (Figure S5). In conclusion, these results indicate that the enhancer activity does not require strict positioning of ACA relative to the initiation site, nor it depends on the spacing between the SD and the initiation codon. Loss of the enhancer function causes yifK expression to be less sensitive to repression by GcvB. To assess the generality of the ACA effects, we turned to the dppA gene, a major GcvB target [16], [18]. The target sequence of GcvB in dppA mRNA includes four ACA motifs clustered within a 15 nt segment near the SD sequence (Figure 10). As an initial test, this 15-nt sequence was deleted in a strain carrying a dppA-lacZY translational fusion. The resulting mutant showed 96% lower of ß-galactosidase activity than the parental strain (data not shown). In the next experiment, we randomly mutagenized all four ACA repeats (in the same lacZ fusion background) and screened the mutants on MacConkey lactose indicator plates. Out of 43 mutants analyzed, 18 formed white colonies and had ß-galactosidase activities ranging between 0. 5 and 5% of the wild-type levels. Four representative isolates from this group are shown in Figure 10. They are essentially Lac− mutants. 4 of the initial 43 mutants formed red colonies and expressed significant levels of ß-galactosidase (three shown in Figure 10). Interestingly, in two of these strains, the mutagenic process regenerated an ACA sequence. The remaining 21 isolates had an intermediate phenotype (pink colonies) and were not analyzed. Examination of the mutant sequences by the Mfold algorithm [29] showed a complete lack of correlation between presence/absence of secondary structures (or free energy values) and lacZ expression. Hence the most likely conclusion from this analysis is that the variations in lacZ expression levels are solely dictated by primary sequence determinants. Although the possibility that decreased expression in some of the mutants could be due to reduced mRNA stability, independent of ribosome binding, cannot be formally ruled out, it seems most likely that the observed differences reflect variations in translation initiation rates. Thus, on one hand, the data in Figure 10 further corroborate the positive role of the ACA motif in translation initiation; on the other hand, they reiterate the notion that an SD motif and a properly spaced AUG are not sufficient to promote initiation if placed in unfavorable sequence contexts [30].
In the present work, we have characterized the regulation of Salmonella' s yifK locus encoding a putative amino acid transporter highly conserved in Enterobacteriaceae. Our analysis showed yifK to be negatively controlled at the transcriptional level by the leucine response regulator Lrp, and at the post-transcriptional level by GcvB sRNA. These findings place yifK at the intersection of two global regulatory networks devoted to amino acid management [17], [31]. The relative impacts of two systems on yifK expression vary as a function of growth conditions, with the Lrp control predominating in leucine-deprived poor media and the GcvB control operating when amino acids are plentiful, possibly in excess. The sole condition where yifK appears to escape negative control is leucine-supplemented minimal medium, where Lrp repression is relieved. This response closely parallels that of the oligopeptide permease operon, oppABCDF [31] whose transcript is also a target of GcvB repression [16], [18]. Likely, the overlap of Lrp and GcvB networks reflects the link between amino acid metabolism and one-carbon units production; however, the precise physiological role and the implications of the above responses remain incompletely understood. Genetic analysis of GcvB: yifK mRNA interactions revealed that the GcvB target sequence in yifK mRNA contains an enhancer element. Intriguingly, mutations that disrupt the enhancer - and lower yifK expression as a result - render yifK expression totally insensitive to GcvB repression. This suggests that the effectiveness of GcvB regulation is dependent on the enhancer function and that when this component is removed, GcvB-mediated repression no longer constitutes a rate-limiting step in yifK expression. Sharma and coworkers (2007) previously showed that GcvB' s target sequence in the gltI gene of Salmonella acts as transferable translation enhancer (see Introduction). Unlike in our study, the effects of GcvB as a translational repressor were much greater than the effects of removing the enhancer, leading the authors to conclude that GcvB did not simply block the enhancer effect [16]. It seems possible that the plasmid-borne nature of the gcvB gene in the study by Sharma and coworkers made the GcvB repression tighter than when the sRNA is expressed from the chromosome. Alternatively, the contribution of the enhancer to gltI expression might be less important than in yifK expression. The gltI enhancer, located 45 nt upstream from the initiation codon, was characterized as part of a 27 nt segment and not analyzed in any further detail [16]. Here we found that nucleotide replacement in either of two ACA triplets within GcvB target site in yifK can result in more than 90% reduction in yifK expression. Although our data do not allow defining the contours of the enhancer element, they unequivocally identify the ACA motif as a determinant of its activity. We also found that the enhancer activity is maintained following a 7 nt shift in the position of the initiation site, suggesting the absence of strict spatial requirements for the functioning of the element. This is consistent with the data from the gltI system and with a report showing CA repeats to stimulate translation even when placed downstream from the start codon [32]. Translation initiation efficiencies have been known to vary greatly as a function of the sequence context of the initiation region [30], [33]. Computational analysis of sequences surrounding translation initiation sites of E. coli genes showed that the spacing between the SD and the initiation codon affects SD sequence conservation and its pattern. This study did not reveal significant biases outside these main elements [27]; however, conserved patterns occurring at variable positions might have been difficult to identify by the statistical analysis. Indeed, separates lines of evidence point to the role of the ACA motif in translation initiation. The motif is found in other translation enhancer sequences [34], [35] and, as an ACAA repeat, was shown to promote translation initiation in the absence of a SD sequence [36]. ACA is also found in the loops of pseudoknots formed by RNA ligands to ribosomal protein S1, obtained through Systematic Evolution of Ligands by Exponential Enrichment (SELEX) [37] and is part of the SELEX-determined consensus sequence for binding of protein CsrA, a translational regulator [38]. Finally, ACA is the recognition sequence of the MazF endonuclease that inactivates E. coli mRNAs by preferentially cleaving near the translation initiation codon [39]. The lack of position requirements for the functioning of the enhancer suggests that its role is to provide an anchor point for the 30 S ribosomal subunit so as to facilitate subsequent recognition of the SD sequence. Some of the evidence reviewed above tentatively identifies protein S1 as the possible candidate for the interaction. In vitro S1-binding studies with some of the mutants constructed in the course of this work should allow testing of this idea. Combined with the mutational analysis of other GcvB-regulated mRNAs, this approach might provide further insight into how the ACA motif participates in the translation initiation step.
Strains used in this study were derivatives of Salmonella enterica serovar Typhimurium strain LT2 [40]. Strain SV4280 was a gift of J. Casadesús. Except for the latter strain and for strain MA7224, all other strains were derived from MA3409, an LT2 derivative cured for the Gifsy-1 prophage [41]. The genotypes of the relevant strains used are listed in Table S1. Bacteria were cultured at 37°C in liquid media or in media solidified by the addition of 1. 5% Difco agar. LB broth [42] was used as complex medium. Carbon-free medium (NCE) [43], supplemented with 0. 2% glycerol or 0. 2% lactose was used as minimal medium. Antibiotics (Sigma-Aldrich) were included at the following final concentrations: chloramphenicol, 10 µg ml−1; kanamycin monosulphate, 50 µg ml−1; sodium ampicillin 100 µg ml−1; spectinomycin dihydrochloride, 80 µg ml−1; tetracycline hydrochloride, 25 µg ml−1. MacConkey agar plates containing 1% lactose [44] were used to monitor lacZ expression in bacterial colonies. Liquid cultures were grown in New Brunswick gyratory shakers and growth was monitored by measuring the optical density at 600 nm with a Shimazu UV-mini 1240 spectrophotometer. T4 polynucleotide kinase and Taq DNA polymerase were from New England Biolabs, Pfu-Turbo DNA polymerase was from Stratagene, T4 DNA ligase was from New England Biolabs. DNA oligonucleotides were custom synthesized by Sigma Aldrich or Eurofins MWG/Operon. The complete list of DNA oligonucleotides used in this study is shown in Table S2. DNA sequencing was performed by GATC biotech. Acrylamide-bisacrylamide and other electrophoresis reagents were from BioRad. Agarose was from Invitrogen. Hybond-N+ membranes and hybridization buffer used for Northern blot analysis were from GE Healthcare and from Applied Biosystems-Ambion, respectively. The rNTPs were from Promega and the 32P-NTPs were from PerkinElmer or Hartmann Analytic. 32P-labeled nucleic acids were detected by phosphorimaging using ImageQuant software. Generalized transduction was performed using the high-frequency transducing mutant of phage P22, HT 105/1 int-201 [45] as described [46]. Chromosomal engineering (recombineering) was carried out by the λ red recombination method [47]–[49] implemented as in [47]. Donor DNA fragments were generated by PCR using plasmid DNA or chromosomal DNA or DNA oligonucleotides as templates. Amplified fragments were electroporated into appropriate strains harboring the conditionally replicating plasmid pKD46, which carries the λ red operon under the control of the PBAD promoter [47]. Bacteria carrying pKD46 were grown at 30°C in the presence of ampicillin and exposed to arabinose (10 mM) for 3 hours prior to preparation of electrocompetent cells. Electroporation was carried out using a Bio-Rad MicroPulser under the conditions specified by the manufacturer. Recombinant colonies were selected on LB plates containing the appropriate antibiotic. Constructs were verified by PCR and DNA sequence analysis (performed by GATC company). PCR amplification of DNA fragments under error-prone conditions was carried out as previously described [50]. Scarless modification of chromosomal DNA sequences at the single base-pair level was achieved with a two-step recombineering procedure as previously described [51]. Briefly, this involved: 1) inserting a tetAR module (produced by PCR) at the chromosomal site to be modified and: 2) replacing the tetAR module by a DNA fragment carrying the desired changed through positive selection tetracycline-sensitive recombinants [52]. Typically, the DNA fragment in the second step was also obtained by PCR using oligonucleotides with complementary sequences at their 3′ ends priming DNA synthesis on each other (“reciprocal priming”). In site-directed mutagenesis experiments, one of the two primers contained the desired nucleotide changes or randomized sequence stretches. All constructs were verified by DNA sequencing. Table S3 shows the list of alleles made by standard or scarless recombineering. RNA was prepared by the acid-hot-phenol method from exponentially growing cells (OD600 of 0. 35) as previously described [50]. Reverse transcriptase reactions (enzyme Superscript II from Invitrogen) were carried out using 5 µg of bulk RNA and 32P-labeled primer ppF49. The same DNA primer was used for the sequencing reactions. Reactions were performed with the fmol DNA Cycle Sequencing System from Promega, according to the manufacturer' s protocol. Reaction products were fractionated on a 10% polyacrylamide-8 M urea gel. For Northern blot analysis, RNA was fractionated on a 1% agarose-formaldehyde gel, blotted onto a nylon membrane, and hybridized to the appropriate 32P-labeled DNA oligonucleotide probes. In vitro coupled transcription/translation was performed using New England Biolabs' PURExpress In vitro Protein Synthesis kit (NEB #E6800) according to the manufacturer instructions. Genes to be analyzed were cloned under T7 promoter control in the DFRH plasmid provided with the kit. The hybrid genes carried yifK wt or mutant 5′ UTR sequences fused to the cat-3×FLAG coding sequence (chloramphenicol acetyl transferase in-frame fusion to the 3×FLAG epitope). Final volume of the transcription/translation reaction was 25 µl in all cases. In addition to kit solutions A and B, reaction mix contained, 10 U of RNase inhibitor SUPERase (Ambion) and template plasmid DNA added to either 0. 5 or 5 pM final concentration. Incubation times at 37°C varied from 15 to 90 min. Reactions were stopped by addition of equal volume of 2× Laemmli buffer and immediate freezing. Aliquots were loaded on 12. 5% Acrylamide gels and Western analysis performed as previously described [53]. Toeprinting reactions were carried out as described by Darfeuille et al [54] with minor modifications. RNA fragments spanning positions +1 to +135 of yifK mRNA were synthesized in vitro from T7 DNA templates generated by PCR amplification of chromosomal DNA (from strains MA8020 or MA11793) with primers ppI22 and ppI23. 2 pmol of RNA were annealed with 5′end-labeled primer ppI23 (1 pmol) in 10 mM Tris-acetate [pH 7. 6], 0. 1 M potassium acetate, and 1 mM DTT for 1 min at 90°C and chilled in ice for 5 min. Then, all dNTPs (final concentration 1 mM), Mg Acetate (10 mM final) were added; this was followed by preincubation with 2 pmol of 30S ribosomal subunit (a gift of Dominique Fourmy and Satoko Yoshizawa) at 37°C for 5 min. In experiments involving GcvB, 5,1 or 0. 5 pmol of sRNA were added prior to both, addition of the 30S ribosomal subunit and the preincubation step. After the 5-min period, 2 pmol of tRNAfMet were added and preincubation at 37°C continued for 15 additional min. Finally, Reverse Transcriptase (Superscript II, Invitrogen, 200U) was added and samples incubated for 15 min at 37°C. Following phenol chloroform extraction and ethanol precipitation, resuspended samples were loaded onto a 10% polyacrylamide-8 M urea gel along with the sequencing reaction samples generated with the same primer. β-galactosidase activity was assayed in toluene-permeabilized cells as described in [55] and is expressed in Miller units throughout this work. Typically, measurements were performed on duplicate or triplicate cultures grown in late exponential phase (OD600≈0. 7). All experiments included parental or reference strains as normalization controls. Standard deviations were generally less than 5% of the mean. | The majority of small RNA (sRNA) regulators in bacteria act by inhibiting translation initiation in target messenger RNAs. The study of this regulatory mechanism not only allows a better understanding of sRNA function but it can also provide new insight into aspects of the translation initiation process that remain incompletely characterized. This was the case in the work described here. Analyzing the mechanism by which GcvB, a multi-target sRNA, downregulates a putative amino acid transporter in Salmonella, we discovered that the sequence base-pairing with GcvB in the target mRNA functions as a translational enhancer. Replacing an ACA motif near the center of the sequence with unrelated trinucleotide sequences leads to a decrease in translational initiation efficiency that can be as severe as more than 90%. Interestingly, some of these replacements concomitantly render the mRNA insensitive to GcvB variants carrying the appropriate compensatory changes, suggesting that targeting the enhancer element is paramount for GcvB regulatory effectiveness. Overall the data presented in the paper unveil the role of the ACA motif in the translation initiation process and lay the grounds for further analysis of the mechanism involved. | Abstract
Introduction
Results
Discussion
Materials and Methods | 2014 | Translation Enhancing ACA Motifs and Their Silencing by a Bacterial Small Regulatory RNA | 9,322 | 275 |
|
Sjögren’s syndrome (SS) is a common, autoimmune exocrinopathy distinguished by keratoconjunctivitis sicca and xerostomia. Patients frequently develop serious complications including lymphoma, pulmonary dysfunction, neuropathy, vasculitis, and debilitating fatigue. Dysregulation of type I interferon (IFN) pathway is a prominent feature of SS and is correlated with increased autoantibody titers and disease severity. To identify genetic determinants of IFN pathway dysregulation in SS, we performed cis-expression quantitative trait locus (eQTL) analyses focusing on differentially expressed type I IFN-inducible transcripts identified through a transcriptome profiling study. Multiple cis-eQTLs were associated with transcript levels of 2' -5' -oligoadenylate synthetase 1 (OAS1) peaking at rs10774671 (PeQTL = 6. 05 × 10−14). Association of rs10774671 with SS susceptibility was identified and confirmed through meta-analysis of two independent cohorts (Pmeta = 2. 59 × 10−9; odds ratio = 0. 75; 95% confidence interval = 0. 66–0. 86). The risk allele of rs10774671 shifts splicing of OAS1 from production of the p46 isoform to multiple alternative transcripts, including p42, p48, and p44. We found that the isoforms were differentially expressed within each genotype in controls and patients with and without autoantibodies. Furthermore, our results showed that the three alternatively spliced isoforms lacked translational response to type I IFN stimulation. The p48 and p44 isoforms also had impaired protein expression governed by the 3' end of the transcripts. The SS risk allele of rs10774671 has been shown by others to be associated with reduced OAS1 enzymatic activity and ability to clear viral infections, as well as reduced responsiveness to IFN treatment. Our results establish OAS1 as a risk locus for SS and support a potential role for defective viral clearance due to altered IFN response as a genetic pathophysiological basis of this complex autoimmune disease.
Sjögren’s syndrome (SS) is a common systemic autoimmune disease with a prevalence rate (~0. 7% of European Americans) second only to rheumatoid arthritis (RA) [1]. SS is distinguished by immune cell infiltration, functional destruction, and irreversible dysfunction of exocrine glands, most notably salivary and lacrimal glands [2]. Secondary manifestations of exocrine gland dysfunction may include severe dental decay and corneal scarring. Approximately one-third of patients experience extra-glandular manifestations of disease, such as debilitating fatigue, a 16-fold increased risk of developing lymphoma, neuropathies, Raynaud’s phenomenon, arthralgia, and dermatologic symptoms [2–6]. Both glandular dysfunction and extra-glandular manifestations are associated with autoantibodies, a hallmark of autoimmunity [7–9]. Approximately 70% and 40% of SS patients exhibit autoantibodies targeting ribonucleoproteins, Ro/SSA (Ro52 and Ro60) and La/SSB, respectively [10]. These autoantibodies have the capacity to bind necrotic and apoptotic material, thus creating RNA-immune complexes that can activate cells of the immune system and aggravate autoinflammation [11]. Such RNA-containing immune complexes are taken up by the Fc gamma receptor IIa on plasmacytoid dendritic cells (pDCs) [12], which activates intracellular Toll-like receptors 7 and 9 and stimulates type I interferon (IFN) responsive loci [13]. The etiology of SS is still largely unknown, though it involves a complex interplay between both genetic and environmental factors [14–16]. Viral infections, such as Epstein-Barr virus (EBV) and cytomegalovirus [17–19], may initiate prolonged inflammation in glandular lesions and formation of germinal center-like structures commonly linked to autoantibody production in SS [14,19]. Autoantibodies can be detected up to 18–20 years prior to diagnosis in 81% of SS patients [20]. Indeed, cross-reactivity between antibodies against EBV and the Ro60 antigen has previously been reported [21], and possible subclinical reactivation of the virus has been associated with active joint involvement in SS [22]. Recently, the virus-like genomic repeat element L1 was identified as an endogenous trigger of the IFN pathway, and its expression correlates with type I IFN expression and L1 DNA demethylation [23]. Type I IFNs are key antiviral immune mediators of innate immune responses in infected cells, while at the same time enhancing antigen presentation and inducing production of pro-inflammatory cytokines and chemokines [24], thus initiating adaptive immunity [25]. Overexpression of type I IFN-inducible genes, known as “the IFN signature”, is a common feature of many autoimmune diseases [26,27], including RA patients with poor clinical outcome [28–30] and systemic lupus erythematosus [31,32], where the predominant IFN producing cells, pDCs, are reduced in number in the blood but are abundant in skin and lymph nodes [33]. In SS, the IFN signature is observed in both peripheral blood and salivary glands [12,34–37], and associates with systemic manifestations, greater disease severity, and autoantibody titers [8,9]. It has been proposed that viral infections contribute to perpetual activation of type I IFN signaling and the resulting dysregulation of innate immunity, ultimately resulting in activation of the adaptive immune response and autoantibody production in SS and other autoimmune diseases [34,38]. Genome-wide association [GWA] studies in autoimmune diseases have identified multiple genetic risk variants involved in type I IFN signaling pathways [39–41], including associations of IRF5 and STAT4 with SS susceptibility [15,42,43]. Suggestive associations of FCGR2A, PRDM1 (PR domain containing 1, regulated by IRF5) [44], and IRF8 with SS have also been reported [15], along with two genes within the NF-κB pathway (TNIP1 and TNFAIP3), which regulates early phase type I IFN production during viral infection [45]. Despite the evidence indicating an important role of the type I IFN pathway in SS, no direct functional mechanisms for SS-associated variants contributing to the substantial upregulation of IFN signature transcripts have been described. The vast majority of disease associated single-nucleotide polymorphisms (SNPs) identified in GWA studies are non-coding [46] and are not likely to impact protein function directly, thus requiring a combination of genetic studies and gene expression analyses to point towards mechanisms that link genetics with functional effects [47]. Specifically, the extensive linkage disequilibrium (LD) observed between associated polymorphisms renders it hard to identify causal variant (s) of disease. Systemic evaluation of genome-wide functional elements by the Encyclopedia of DNA Elements (ENCODE) project reveals that 80% of the human genome has at least one biochemical function, and many of the genetic variants are within cis- or trans- regulatory sites that impact gene expression [48]. Furthermore, genome-wide cis-expression quantitative trait locus (eQTL) mapping studies in different tissues have identified more than 3,000 genes associated with nearby genetic variants [49,50]. Through combining GWA and gene expression data from SS patients, we sought to identify and characterize SS-associated variants that influence the expression of genes within the IFN signature by utilizing a genomic convergence approach (Fig 1). Through cis-eQTL analyses we identified an association of a SNP rs10774671, located within the OAS1 gene locus, with SS. Functional studies were performed to assess biological consequence of the OAS1 variants.
To select candidate genes in the IFN signature, we first evaluated dysregulated transcripts in SS through a microarray-based gene expression profiling study. Whole blood transcriptome profiles from 115 anti-Ro/SSA positive SS cases and 56 healthy controls of European ancestry were compared, as the IFN signature is enriched in SS patients seropositive for anti-Ro/SSA [34]. After quality control (QC) and normalization, 13,893 probes (in 10,966 genes) remained, for which Welch’s t-tests, false discovery rate (FDR) -adjusted P values (q values), and fold changes (FC; the difference of the mean between log2-transformed values from cases and controls) were calculated (see Methods for details). Differentially expressed genes were selected by q < 0. 05 and FC > 2 or < -2. We found 73 differentially expressed genes in our dataset, among which 57 genes are regulated by type I IFNs (S1 Table). The majority of dysregulated genes (66 out of 73) were overexpressed in SS patients and formed the IFN signature in cases after unsupervised hierarchical clustering (Fig 2A). Of note, the IFN signature was observed in most but not all anti-Ro/SSA positive SS cases, in accordance with our previous work [34], and the intensity of this feature was heterogeneous among patients (Fig 2B). These results suggest that the expression of IFN signature genes might be influenced by genetic variants, which could be identified through cis-eQTL analysis. We hypothesized that variants near the differentially expressed IFN signature genes may potentially influence disease susceptibility through cis-regulatory mechanisms. We would, however, expect to identify many cis-eQTLs in these regions regardless of whether they associate with disease susceptibility or not. Therefore, instead of performing cis-eQTL analyses for all of the 73 dysregulated IFN signature genes, we first sought to identify variants that showed a disease association of Passoc<0. 05 for subsequent evaluation of their role in altering gene expression. Genetic associations with SS susceptibility for 2,163 SNPs in regions of the 73 differentially expressed genes were tested using a combined dataset (Dataset 1; Table 1) from genome-wide genotyping arrays consisting of 765 SS cases and 3,825 population controls of European ancestry. We identified suggestive associations (Passoc<1×10−4; this threshold was determined by Bonferroni correction for independent variants with r2<0. 2) of genetic variants within the OAS1 region (top association at rs10774671, Passoc = 8. 47×10−5), which is regulated by type I IFNs (S1 Table). Furthermore, we identified suggestive associations in ARGN (S1 Table) and observed nominal associations (1×10−4≤Passoc<0. 05) with SS susceptibility in 42 additional regions (S1 Table). To determine whether these disease-associated genetic variants (Passoc<0. 05) were related to the altered expression levels of their nearby differentially expressed IFN signature genes, we performed cis- and trans-eQTL analyses for all SNPs with Passoc<0. 05 (173 SNPs in 44 regions; S1 Table) using a linear model by integrating the transcriptome and genotype datasets in 178 European individuals (108 anti-Ro/SSA positive SS cases, 55 anti-Ro/SSA negative SS cases, and 15 healthy controls). Variants within and near OAS1 showed significant association with OAS1 transcript expression (Fig 3A; S1 Table). In particular, three microarray probes targeting OAS1 passed QC and were evaluated for cis-eQTLs (Fig 3B). The OAS1 transcript levels measured by all of the three probes were found to be associated with nearby genetic variants (Fig 3C). The most significant cis-eQTL for all the three probes targeting OAS1 was rs10774671 (PeQTL-Probe1 = 5. 14×10−4, PeQTL-Probe2 = 2. 86×10−6, and PeQTL-Probe3 = 6. 05×10−14; Fig 3C). No eQTL was detected in any other differentially expressed genes (S1 Table). We also determined that none of these eQTL variants were associated with the two nearby genes, OAS2 and OAS3. Additionally, no significant trans-eQTL was detected for OAS1. Therefore, we identified a variant associated with both SS susceptibility and gene expression in the IFN signature gene OAS1. To fine map this disease-associated region, imputation was then performed for the SS-associated OAS1 region to increase the informativeness of the genetic association and eQTL analyses results. After imputation, the most significant association with SS in the OAS1 region was at rs4767023 (Passoc = 3. 82×10−5; r2 = 0. 98 with the most significant genotyped SNP rs10774671; Fig 4A), whereas the top eQTL remains at rs10774671 (Fig 3A). All the variants with Passoc<1×10−4 in the OAS1 region were strongly correlated to each other (r2>0. 9; Fig 4B) and could explain the association of the whole region through conditional analyses (Fig 4C and 4D). The top SS-associated variants and cis-eQTLs in the OAS1 region, including rs10774671, were in strong LD (r2>0. 9; Fig 4B), thus challenging the selection of potentially functional variant (s) based on results from the association analyses. However, the top eQTL variant, rs10774671, is an A/G substitution within the consensus sequence of a splice acceptor site at the junction of the 5th intron and the 6th exon of OAS1 (Fig 3B), and is known to alter normal splicing and induce isoform switching of OAS1 [51]. In addition, all other SS-associated variants (Passoc<1×10−4) in the OAS1 locus were either intronic or outside of coding regions, lacking functional genomic elements mapped to the SNP as determined by the ENCODE project [48,52]. Also, we performed a co-localization analysis using eCAVIAR [53] to identify the potential causal variant in the OAS1 region. We estimated colocalization posterior probability (CLPP) scores for all the tested 453 variants, and rs10774671 has the highest CLPP score among all the variants (S2 Table). Therefore, we prioritized this SS-associated cis-eQTL variant, rs10774671, for further replication and functional studies. We replicated the genetic association of rs10774671 with SS susceptibility in an independent dataset (Dataset 2; Table 1) consisting of 514 European SS cases and 3,466 European population controls (genotyped using TaqMan assays, Prep = 5. 16×10−6; odds ratio = 0. 71; 95% confidence interval = 0. 63–0. 83). Meta-analysis was performed to combine the results between the initial genetic association study (Dataset 1) and the replication cohorts (Dataset 2) and established the association of rs10774671 with SS risk (Pmeta = 2. 59×10−9; odds ratio = 0. 75; 95% confidence interval = 0. 66–0. 86; risk allele [the A allele] frequency: case = 0. 70, control = 0. 64; with no heterogeneity between the two datasets as determined by I2 = 0). We also performed a stratified analysis and a permutation analysis using merged samples from Dataset 1 and Dataset 2 to determine whether the observed genetic association was restricted to anti-Ro/SSA positive or negative patients. We did not find any evidence to support the genetic effect to be specific to any sub-group of the patients (S1 Fig). In summary, we identified a potential causal variant, rs10774671, that was associated with SS susceptilibty, likely through its impact on the expression of a key IFN signature gene, OAS1. Following establishment of the association between rs10774671 and SS susceptibility, we further determined the influence of different genotypes on the alternative splicing of OAS1. Four isoforms of OAS1 are annotated in the NCBI Reference Sequence (RefSeq; http: //www. ncbi. nlm. nih. gov/refseq) database, of which we analyzed p46, p42, and p48, and p44, an un-annotated isoform previously reported in RNA-sequencing (RNA-seq) studies [54–56] (Fig 3B). The difference between these isoforms is confined to their 3' end where rs10774671 influences alternative splicing, yielding amino acid sequences of different lengths and composition. In the microarray experiments, one probe targeting OAS1 specifically recognized the 3' end of the p42 isoform (Fig 3B). The risk allele A of rs10774671 was correlated with higher expression levels of p42 (Fig 3C, right panel). However, we were not able to determine the influence of rs10774671 on the expression of other isoforms due to lack of isoform-specific probes on the microarray. In order to determine the influence of rs10774671 on the expression of each alternatively spliced isoform of OAS1 and compare OAS1 isoform composition, we performed RNA-seq on whole blood from 57 SS cases and 27 healthy controls. After QC, the reads were aligned to the human genome using TopHat [57] without gene annotation to facilitate the detection of potentially novel isoforms of OAS1. The transcript level of each isoform was compared across samples with different genotypes of rs10774671 based on the measurement of fragments per kilobase of transcript per million mapped reads (FPKM) using Cufflinks [58]. Consistent with our microarray results, the SS risk allele A of rs10774671 was correlated with higher expression levels of p42 (P = 1. 30×10−7; Fig 5A). Increased production of other alternatively spliced isoforms of OAS1, including p48 and p44, was also observed in subjects with the SS risk genotypes (GA and AA) of rs10774671 (Fig 5B and 5C). In contrast, transcript levels of the p46 isoform, was decreased in samples with the A allele (P = 3. 48×10−10; Fig 5D), consistent with previous reports that interruption of the splicing consensus sequence inhibit formation of the p46 isoform [56]. These results were further confirmed by quantitative real-time PCR using primer sets targeting the specific OAS1 isoforms (S2 Fig; S3 Table). Therefore, we found that the SS-associated variant rs10774671 is a functional variant that influences alternative splicing of OAS1. Since OAS1 is part of the IFN signature and its expression levels are correlated with the autoantibody status, we also performed a stratified eQTL analysis to investigate whether the eQTL effects are specific to any sub-group of the SS patients based on their anti-Ro/SSA positivity. We stratified the SS case samples into anti-Ro/SSA positive patients (n = 27) and anti-Ro/SSA negative patients (n = 30), and performed eQTL analyses on each of the OAS1 isoforms using linear regression while adjusting for sex. Despite reduced statistical power, we identified significant eQTL results for the p46, p42, and p48 isoforms in both subsets of samples. By using the Z-test as described in S1 Fig, we did not find any significant difference of the eQTL effects between the two sub-groups (S3 Fig). Comparing the total OAS1 transcript level from the microarray study within each genotype revealed significantly higher gene expression in SS patients as compared to control in the GA group (Fig 6A). There is a trend towards higher total OAS1 transcripts in the AA and GG groups of SS patients as well, though it is not statistically significant. Interestingly, the highest transcript levels are seen in the anti-Ro/SSA positive cases (Fig 6B), significantly higher than both anti-Ro/SSA negative cases and healthy controls. The same results were observed in the RNA-seq data (S4 Fig), indicating that the total OAS1 transcript levels regardless of isoform are also influenced by disease status or the presence of autoantibodies. To further dissect the functional mechanism of rs10774671 in predisposing disease risk, we utilized Western blots to evaluate the difference in protein levels of the normally spliced isoform p46 (formed by the protective allele G of rs10774671) and the alternatively spliced isoforms of OAS1 in EBV-immortalized B cells from SS patients. Consistent with the RNA-seq results, the protein expression of p46 was substantially lower in subjects carrying the A allele of rs10774671, whereas p42 was the dominant isoform in the GA and AA subjects without stimulation (Fig 7A). Interestingly, both protein and mRNA levels of the p46 isoform were upregulated after stimulation by type I IFN in the GG and GA subjects (P = 3×10−4 and P = 2×10−3, respectively; Fig 7A and 7B). However, protein expression of the p42 isoform remained unchanged upon IFN stimulation (Fig 7A), even though its transcript level significantly increased after stimulation (Fig 7B). The protein expressions of p48 and p44 were low in all of the samples, and not responsive to type I IFN stimulation (Fig 7A). We then cloned and transfected each isoform into human embryonic kidney (HEK) cell line 293T cells and observed similar protein expression results as in EBV cells: lower protein levels for p48 and p44 compared to p46, even though their transcript levels were equivalent (Fig 7C). These results suggest that the alternative isoforms of OAS1 that are associated with the disease risk variant of rs10774671 fail to generate proteins after transcription. The catalytic OAS1 domain is located at the N terminus, though the isoforms differ in their C terminus. It has been suggested that these differences affect affinity of OAS1 protein for different viruses [59]. However, our data suggested that the alternatively spliced 3' -terminus influenced the lack of post-transcriptional expression of the p48 and p44 isoforms. To test this hypothesis, we generated several truncated forms of p48 and p44 at the 3' -terminus and transfected them into HEK 293T cells. The truncation of both p48 and p44 transcripts at the 3' -end resulted in restoration of protein expression (Fig 7D). Our results demonstrated that the alternatively spliced 3' -end between 1,047 and 1,155 bp of the p48 isoform and the 3' -end between 1,083 and 1,137 bp of the p44 isoform were responsible for the impaired protein expression (Fig 7D). In addition, we recombined the green fluorescent protein (GFP) transcript with the 3' -end from different OAS1 isoforms and expressed them in HEK 293T cells. The alternatively spliced 3' -terminus of p48 and p44 resulted in reduced expression of GFP when linked to the 3' end of GFP transcript (Fig 7E; S5 Fig). These results further confirmed the impact of the alternatively spliced 3' -end of OAS1 on protein expression. Determining mechanisms for how the 3' -terminus from the alternatively spliced OAS1 isoforms influences protein expression and type I IFN responsiveness needs further study.
Overexpression of genes in the IFN pathway is a distinctive feature of multiple autoimmune diseases, though no evident mechanism has thus far been revealed. We identified and established rs10774671 as a risk locus for SS. The A allele of rs10774671 is correlated with reduced OAS1 enzymatic activity in human peripheral blood mononuclear cells [51], and is associated with increased susceptibility to West Nile virus [60] and chronic hepatitis C virus infections [61]. OAS1 is a member of the 2' -5' -oligoadenylate synthetase family, which is upregulated by type I IFNs during innate immune responses to viral infection and activates latent RNase L, leading to viral RNA degradation and clearance [62,63]. The SS risk allele A of rs10774671 causes alternative transcript splicing and consequently less functional isoforms are activated by type I IFNs. Failure to clear virus might lead to subclinical, chronic infection that drives the sustained overexpression of IFN, but viral proteins may also indirectly cause IFN production through adaptive immune responses. For example, antibodies generated towards EBV nuclear antigen-1 cross-react with Ro/SSA [21], and anti-Ro/SSA antibodies may in turn form immune complexes that stimulate type I IFNs [64]. As viruses evidently play a role in SS pathophysiology [18], genetic variants affecting the antiviral properties of OAS1 might be a contributing factor. A recent study showed that while the presence of antibodies to hepatitis D virus was equal in SS patients and otherwise healthy controls, the virus itself was present in significantly more patients [65], indeed suggesting that viral clearance is restrained. Epitope spreading [66,67], antibody cross-reaction [21], or molecular mimicry [68,69] are likely consequences of subclinical, chronic or recurrent infection. The basal activity of OAS1, which varies greatly among individuals, is thought to be under strong genetic control [51]. The enzyme activity in the GG genotype with predominantly the p46 isoform is higher than GA (intermediate) and AA (low) [51]. OAS1 isoforms p42 and p46 have been detected at the protein level in human cells, whereas the p44a/p44b, p48, and p52 isoforms have been detected at mRNA levels [54,70–72]. In addition to the RNase L activation properties, the tetramer forming p48 isozyme also exhibits proapoptotic activity [73], a property partly accredited to IFN-γ [74], and is shown to interact with Bcl-2 [75]. Bcl-2 is an anti-apoptotic protein negatively regulated by Ro52 [76], and in salivary gland epithelial cells Bcl-2 is essential in regulation of IFN-γ induced apoptosis [77]. It has been postulated that p46 is a more efficient synthetase than p48, explaining the increased basal activity of p46 [51]. Although there is no evidence showing that the differences in the C-terminus alter protein function [78], our truncation experiments indicated that the alternatively spliced C-terminus governs the post-transcriptional protein expression. Interestingly, we found lower protein expressions of p44 and p48 after type I IFN stimulation despite equivalent transcript levels compared to p46. This indicates that p44 and p48 production, which is governed by the A allele, is less responsive to IFN stimulation as compared to p46. Lack of response to IFNs has also been seen in multiple sclerosis (MS), in which patients carrying the homozygous rs10774671 GG genotype, a protective genotype in MS associated with less active disease, were more responsive to IFN-β treatment than AA and AG patients, as measured by time to first relapse [79]. We searched the Genotype-Tissue Expression (GTEx) database and confirmed the association between rs10774671 and OAS1 expression in whole blood [80]. We also searched eQTL for all our top variants in the 43 SS-associated genes (besides OAS1) as well as any variants in LD with those top variants (r2>0. 8). Out of the 614 variants we checked, two variants were also eQTLs for their corresponding genes: the top SS-associated variant in ANKRD22 (rs1147601, Passoc = 2. 38×10−3, PeQTL-GTEx = 6. 4×10−6) and the top SS-associated variant in EPSTI1 (rs7323736, Passoc = 1. 79×10−2, PeQTL-GTEx = 2. 6×10−6). However, both of these variants were only nominally associated with SS susceptibility and did not pass our suggestive significance threshold for disease association (Passoc<1×10−4). Nevertheless, these variants and genes could be plausible targets for future replication studies to assess their disease associations. The rs10774671 A/G variant is a common splice site variation, and there is a skewed distribution of genotypes in autoimmune diseases like type I diabetes (T1D) [81] and MS [79] despite ambiguous genetic association with disease: the alternative allele A renders risk to SS and MS, whereas the reference allele G increases susceptibility to T1D. We hypothesized that these opposite risk effects may be due to different functional isoform usages in different disease-relevant tissues. Through searching the GTEx database for rs10774671 eQTLs, we found rs10774671 is a significant eQTL of OAS1 in 5 tissues (S6A Fig). Interestingly, the eQTL effect in the Esophagus—Mucosa tissue is in the opposite direction compared to other tissues. In whole blood, the p46 isoform is predominant, thus the A allele caused reduced expression of OAS1 as a whole (S6B Fig); however, in Esophagus Mucosa, the isoform is p42 (S6C Fig) and results in an opposite effect of rs10774671 on the total OAS1 expression. We propose that the ambiguous genetic effects of rs10774671 on different diseases might be due to different functional isoforms in disease-relevant tissues (not necessarily Esophagus—Mucosa). While reduced expression of the functional isoform p46 in whole blood increases risk of SS and MS, it protects individuals from T1D. The downstream differences of various isoforms in protein levels, isoform expression, responsiveness to IFN, and basal activity between genotypes flag OAS1 as a highly relevant protein in autoimmune diseases, despite no direct effect on IFN expression. OAS1 is one of several genes relevant in overall IFN response found to be disease associated in SS. Others include IL-12A [15], which can induce both type I and type II IFNs [82]; STAT4 [15], which, although not explicitly overexpressed in the IFN signature, plays an important role in the cross-talk between type I and type II IFNs [83–85]; and IRF5 [15], a transcription factor in the IFN pathway [86]. The rs10774671 is a known cis-eQTL and splicing QTL, observed in whole blood [87] as in our study, in lymphoblastoid cells [88], and in monocytes, both naïve CD14 and in cells stimulated with LPS and IFN-γ [89]; but no trans-eQTLs are known. It is possible that other variant (s) in high LD with rs10774671 could contribute additional functional impact (s), such as the rs11352835 in exon 7 seen in MS [59]. Genomic editing approaches that introduce single point mutations or deletions in the OAS1 region will further advance the dissection of the causal SS-associated variant in this haplotype. Animal models that express the risk isoforms due to the risk allele can also be used to observe whether they spontaneously develop SS-like symptoms, and whether the chances for developing such symptoms increase after exposure to viral infections. Our study also highlights the importance of utilizing genomic convergence to identify and prioritize susceptibility genes for human complex disease. The complex mechanisms underlying the IFN signature in SS cannot be explained as a single eQTL driven overexpression. However, we have in this study established OAS1 as a risk locus with functional consequences affecting isoform composition, and that may play a fundamental role in dysregulation of both viral clearance and apoptosis.
All patients in this study fulfilled the 2002 American-European Consensus Group (AECG) criteria for primary SS [7]. Seropositivity of anti-Ro/SSA autoantibodies was determined by the antibody index ≥1 using the Bio-Plex assay (Bio-Rad) following the manufacturer’s protocol. The present study was approved by the Oklahoma Medical Research Foundation Institutional Review Board (IRB#1—Biomedical), operation under Federalwide Assurance (FWA) # 00001389 and IRB # 00000114 under IORG 0000079 approved by the Office for Human Research protection (OHRP), Department of Health and Human Services (DHHS). The OMRF IRB is in compliance with local regulations and the regulations of the United States Food and Drug Administration as described in 21 CFR Parts 50,56 and 11, the International Conference on Harmonization (ICH) E6, and the United States Department of Health and Human Services at 45 CFR 46. The current study was approved under IRB#07–12 and all patients provided written informed consent. Quantitative levels of the differentially expressed transcripts from the microarray analysis were used as phenotypic traits in 178 European subjects described above. Variants showing nominal association with SS (Passoc<0. 05) in the genetic association analysis were selected to test for cis-eQTLs, defined by variant-transcript pairs within 50kb of the target genes or trans-eQTLs for variants at least 1Mb away. Association of genotype with transcript expression was evaluated using both linear regression (adjusted for gender and disease status) and analysis of variance (ANOVA) in Matrix-eQTL [107]. FDR-adjusted P values were calculated to determine the significance of the eQTL. The results of the cis-eQTL analyses were plotted in Prism 6. We also used a tool, PEER, based on a Bayesian framework to adjust for unknown non-genetic factors in gene expression [108]. We transformed our expression values using all the genes that passed QC by running PEER for 15 factors. We then used the PEER residuals from the 44 SS-associated (Passoc<0. 05) IFN signature genes as quantitative traits to determine eQTL while adjusting for other known potentially confounding factors: sex, disease status, anti-Ro/SSA status, and age. In addition to additive genetic models, we also performed eQTL analyses using linear regression by three other models: recessive (recode genotype from 0,1, 2 [where 2 equals to AA] to 0,0, 1), dominant (0,1, 1), and overdominant (0,1, 0). We used the coefficient of determination (R2) to evaluate the goodness-of-fit in each of these models. As shown in S4 Table, both the p42 and p48 isoforms fit the additive model best (highest R2), whereas the recessive model outperformed in the p46 isoform regression (R2rec = 0. 57 vs. R2add = 0. 54). However, the difference of the R2 between the dominant and additive models in the p46 eQTL analysis is subtle. Also, the outperformance of the recessive model cannot be confirmed by qPCR (where the additive model has the highest R2). Therefore, we only reported the additive results in the main text. However, the alternative genetic models for the eQTL effect observed in different isoforms may reflect distinct disease mechanisms rendered by these isoforms, and thus detailed contribution of different isoforms on disease susceptibility warrant further functional study. The co-localization analysis between genetic association and cis-eQTL results in the OAS1 region was performed using eCAVIAR [53]. We used the z-scores (calculated by β/standard error) and the LD matrix (calculated using PLINK—r) from both the genetic association and cis-eQTL results as input and assumed one causal variant to obtain colocalization posterior probability (CLPP) scores for all the tested 453 variants in the OAS1 region. Peripheral blood mRNA transcripts from 27 anti-Ro/SSA positive SS cases, 33 anti-Ro/SSA negative SS cases, and 30 healthy controls were isolated and measured as described above. RNA-seq was performed using the Illumina HiSeq 2000 employing standard procedures. Multiplexing of 6 samples per lane was utilized. Post sequence data were processed with Illumina Pipeline software v. 1. 7. Quality of raw sequence data was assessed using FASTQC. We assessed the quality of each sample using AQM [92] as described above. A total of 6 samples were removed from analysis due to significantly different expression patterns revealed by PC analysis. Raw FASTQ files were aligned to the human reference genome (hg19) using TopHat [57] that aligns the reads across splicing junctions independent of gene annotations, which benefits de novo detection of alternative splicing events. The total gene transcript level was determined by normalized read counts (raw read counts divided by estimated size factor) in DESeq [109]. To determine alternative splicing events, the reference-independent construction of the transcripts was performed using Cufflinks [58] to identify transcripts >1% of the most abundant isoform in each sample. We only kept the transcripts that were detected in more than 10% of the samples for further analysis. The previously annotated isoforms (p46, p42 and p48) and an un-annotated isoform identified across multiple samples (p44) were used as reference to reconstruct the isoforms of OAS1. The novel identified isoforms of OAS1 were also checked manually in the Integrative Genomics Viewer (IGV) [110] to confirm the transcripts and cross-exon reads. The FPKM values calculated by Cufflinks were used to determine the expression levels of each isoform of OAS1. Total RNA was extracted using TRIzol reagents (Life Technologies) from EBV-immortalized B cells pre-selected for the presence of target OAS1 isoforms based on the RNA-seq results from whole blood. Following DNase treatment (Life Technologies) and cDNA synthesis (iScript kit from Bio-Rad), full-length and truncated OAS1 transcripts were amplified from cDNA using primer sets specific for the different OAS1 isoforms and truncated forms (S3 Table). Each OAS1 isoform transcript was individually cloned into pcDNA3. 1 (Invitrogen) with an Xpress epitope tag at the 5' -terminus to facilitate the detection of transfected protein using Western-blot with anti-Xpress antibody. The plasmid was transfected into the HEK 293T cells using FuGENE transfection reagents (Promega) following manufacturer’s protocols. The protein expression of OAS1 isoforms was evaluated in EBV-immortalised B cells from SS patients, four independent samples from each genotype group GG, GA and AA, treated or not treated with type I interferon (universal type I IFN, 1500 U/mL, for 24 hours). The cells were lysed in RIPA buffer and cell lysate protein concentration determined using the Qubit Protein Assay kit (Thermo Fisher Scientific). A total of 30 μg protein from each cell extract was separated on a 10% Bis-Tris gel (10% Criterion™ XT Bis-Tris Gel, BioRad, Cat #: 3450112) following the manufacturer’s instructions, the gels cut according to the weight of the OAS1 protein, and simultaneously transferred to a single PVDF membrane, thus ensuring the comparability of Western blot bands from all gels. The OAS1 isoforms were visualized using an anti-OAS1 antibody targeting the shared epitope (Rabbit polyclonal anti-human OAS1, Abcam, Cat #: ab86343) and ECL Prime Western Blotting Detection Reagents (Amersham, Cat #: RPN2232). | Sjögren’s syndrome (SS) is a common autoimmune condition where immune cells infiltrate moisture-producing glands, leading to dryness typically in the eyes and mouth. SS patients also manifest debilitating fatigue as well as other diseases in liver, lung, kidney, and skin. The cause of this complex disease is still not fully understood; however, an environmental trigger, such as viral infections, in individuals with genetic risk factor (s) is thought to contribute to the development of SS. Type 1 interferons (IFNs) are one of the first defenders after viral infection and induce the expression of various virus-responding genes. Perpetual elevation of type 1 IFN signaling has been observed in SS patients. Here, we first replicated previously identified RNA transcripts contributing to the abnormal type 1 IFN signaling in SS patients. We then identified a disease-associated genetic variant in an IFN-inducible gene, OAS1. This variant governs splicing of OAS1, altering the transcript into multiple isoforms that lack protein expression and responsiveness to IFNs. The results of this study may provide insight into the genetic basis of SS, as well as other autoimmune disease with similar dysregulation in the type 1 IFN system. | Abstract
Introduction
Results
Discussion
Materials and methods | medicine and health sciences
immune physiology
body fluids
immunology
autoantibodies
alternative splicing
protein expression
bioassays and physiological analysis
molecular biology techniques
antibodies
research and analysis methods
immune system proteins
proteins
gene expression
molecular biology
molecular biology assays and analysis techniques
microarrays
biochemistry
gene expression and vector techniques
rna
rna processing
anatomy
nucleic acids
blood
physiology
interferons
genetics
biology and life sciences
genetics of disease | 2017 | Identification of a Sjögren's syndrome susceptibility locus at OAS1 that influences isoform switching, protein expression, and responsiveness to type I interferons | 9,752 | 282 |
Enteroviruses, members of the family of picornaviruses, are the most common viral infectious agents in humans causing a broad spectrum of diseases ranging from mild respiratory illnesses to life-threatening infections. To efficiently replicate within the host cell, enteroviruses hijack several host factors, such as ACBD3. ACBD3 facilitates replication of various enterovirus species, however, structural determinants of ACBD3 recruitment to the viral replication sites are poorly understood. Here, we present a structural characterization of the interaction between ACBD3 and the non-structural 3A proteins of four representative enteroviruses (poliovirus, enterovirus A71, enterovirus D68, and rhinovirus B14). In addition, we describe the details of the 3A-3A interaction causing the assembly of the ACBD3-3A heterotetramers and the interaction between the ACBD3-3A complex and the lipid bilayer. Using structure-guided identification of the point mutations disrupting these interactions, we demonstrate their roles in the intracellular localization of these proteins, recruitment of downstream effectors of ACBD3, and facilitation of enterovirus replication. These structures uncovered a striking convergence in the mechanisms of how enteroviruses and kobuviruses, members of a distinct group of picornaviruses that also rely on ACBD3, recruit ACBD3 and its downstream effectors to the sites of viral replication.
Enteroviruses are small RNA viruses that belong to the Enterovirus genus of the Picornaviridae family. They are non-enveloped positive-sense single-stranded RNA viruses with icosahedral capsids, currently consisting of 15 species. Seven enterovirus species (Enterovirus A-D and Rhinovirus A-C) contain human pathogens, such as polioviruses, numbered enteroviruses, echoviruses, coxsackieviruses, and rhinoviruses. They cause a variety of diseases ranging from common cold to acute hemorrhagic conjunctivitis, meningitis, myocarditis, encephalitis, or poliomyelitis [1]. The genome of the enteroviruses encodes the capsid proteins and seven non-structural proteins (named 2A-2C and 3A-3D). The latter carry out many essential processes including genome replication, polyprotein processing, host membrane reorganization, and manipulation of intracellular trafficking. To facilitate these functions, several host factors are recruited to the sites of enterovirus replication through direct or indirect interactions with viral proteins. For instance, the enterovirus non-structural 3A proteins directly bind to the Golgi-specific brefeldin A-resistant guanine nucleotide exchange factor-1 (GBF1) [2] and acyl-CoA-binding domain-containing protein-3 (ACBD3, also known as GCP60) [3]. ACBD3 is a Golgi resident protein involved in the maintenance of the Golgi structure [4] and regulation of intracellular trafficking between the endoplasmic reticulum and the Golgi [5]. ACBD3 is a multidomain protein composed of several domains connected by flexible linkers. Its central glutamine rich domain (Q domain) interacts with the lipid kinase phosphatidylinositol 4-kinase beta (PI4KB) and with the Rab GTPase-activating proteins TBC1D22A and TBC1D22B [6]. The interaction of ACBD3 and PI4KB causes membrane recruitment of PI4KB and enhances its enzymatic activity [7]. The C-terminal Golgi-dynamics domain (GOLD) of ACBD3 has been reported to interact with the Golgi integral protein giantin/golgin B1, which results in the Golgi localization of ACBD3 [5]. However, in enterovirus-infected cells, the ACBD3 GOLD domain interacts preferentially with viral non-structural 3A proteins, which causes re-localization of ACBD3 to the sites of virus replication [8]. The role of ACBD3 in enterovirus replication is not yet fully understood. It has been proposed that recruitment of ACBD3 to the sites of viral replication can lead to the indirect recruitment of its interactors and downstream effectors such as PI4KB, a well-known host factor essential for generation of PI4P-enriched membranes suitable for enterovirus replication [9,10]. The 3A-ACBD3-PI4KB route represents one of the major described mechanisms of PI4KB recruitment to the sites of enterovirus replication [3,11], although some other mechanisms employing the viral proteins 2BC [12] or 3CD [13] might be involved as well. Moreover, the formation of the 3A-ACBD3-PI4KB complex represents the major described mechanism of PI4KB recruitment to the replication sites of kobuviruses, members of a distinct group of picornaviruses [3,14–16]. Previously, it has been suggested that PI4P directly recruits the viral RNA-dependent RNA polymerase (3Dpol) [9]. Further studies, however, revealed that the affinity of PI4P to 3Dpol is too weak to attract 3Dpol to target membranes by itself, suggesting that other factors may be involved [17]. Notably, PI4P gradients between various membranes can be used for the transport of other cellular lipids against their concentration gradient [18,19]. The PI4P/cholesterol exchange machinery was implicated in replication of several enteroviruses [12,20], suggesting that PI4P can be used by the viral machinery as a mediator to prepare membranes with a specific lipid composition suitable for viral replication. ACBD3 is an important host factor of various enterovirus species [21], however, the structural determinants of its recruitment to the viral replication sites are poorly understood. To date, the structural information about any picornavirus 3A proteins is limited to a solution NMR structure of the uncomplexed poliovirus 3A protein [22] (pdb code 1NG7) and our previously published crystal structure of the aichivirus 3A protein in complex with the ACBD3 GOLD domain [23] (pdb code 5LZ3). Unfortunately, the latter cannot be used for homology modeling of the enterovirus 3A proteins, given the unrelated primary sequences of the enterovirus and kobuvirus 3A proteins, which indicates distinct mechanisms of hijacking ACBD3 by these two groups of viral pathogens. In this study, we present a structural, biochemical, and biological characterization of the complexes composed of human ACBD3 and the 3A proteins of four representative enteroviruses. The crystal structures revealed the details of the ACBD3-3A interaction, the 3A-3A interaction causing the assembly of the ACBD3-3A heterotetramers, the interaction between the ACBD3-3A complex and the lipid bilayer, and the roles of these interactions in facilitation of enterovirus replication. The comparison of the structures of the ACBD3: enterovirus 3A complexes and the previously known structures of the ACBD3: kobuvirus 3A complexes [23] uncovered a striking convergence in the mechanisms of how the two distinct groups of picornaviruses recruit ACBD3 and its downstream effectors to the sites of virus replication.
For the structural characterization of the enterovirus 3A proteins in complex with the host ACBD3 GOLD domain, we selected 3A proteins of six human-infecting enteroviruses each representing different species as follows: enterovirus A71 (EVA71), coxsackievirus B3 (CVB3), poliovirus 1 (PV1), enterovirus D68 (EVD68), rhinovirus A2 (RVA2), and rhinovirus B14 (RVB14) (Fig 1a). The recombinant cytoplasmic domains of all the 3A proteins were poorly soluble and tended to aggregate and precipitate at the required concentrations. Therefore, we used 3A proteins N-terminally fused to a GB1 solubility tag. For the crystallographic analysis of the complexes composed of the ACBD3 GOLD domain and the viral 3A proteins (hereafter referred to as GOLD: 3A complexes), the GB1-fused cytoplasmic domains of the 3A proteins were directly co-expressed with the ACBD3 GOLD domain in bacteria. The GOLD: 3A complexes were then purified, and the GB1 tag was cleaved off. The GOLD: 3A complexes exhibited better protein solubility than the uncomplexed 3A proteins, sufficient for the subsequent crystallographic analysis. Of the six GOLD: enterovirus 3A complexes, only GOLD: 3A/EVD68 and GOLD: 3A/RVB14 formed crystals that diffracted to a resolution suitable for subsequent structure determination (i. e. 2. 3 Å and 2. 9 Å, respectively). Both structures were solved by molecular replacement using a previously published structure of the unliganded ACBD3 GOLD domain (accession number 5LZ1 [23]) as a search model (Fig 1b, Table 1). To improve the crystallization properties of the other four GOLD: 3A complexes, we used two different strategies. The first strategy was based on mutagenesis of selected surface-exposed hydrophobic residues of the 3A proteins to improve the solubility of the respective GOLD: 3A complexes and their capability to be crystallized at higher protein concentrations. This approach led to a successful crystallization of the GOLD: 3A/PV1 complex with an L24A point mutation within the PV1 3A protein. Its structure was then solved at a resolution of 2. 8 Å (Fig 1b, Table 1). The second strategy took advantage of the fact that in all three solved GOLD: 3A structures the C terminus of the ACBD3 GOLD domain was located in the vicinity of the N terminus of the ordered part of the 3A protein. This allowed us to design GOLD-3A fusion proteins with the last residue of ACBD3 (R528ACBD3) fused through a short peptide linker (GSGSG) to the first predicted ordered residues of the respective 3A proteins (e. g. K153A/EVA71). This approach led to a successful crystallization of the GOLD-3A/EVA71 fusion protein and its structure solution at a resolution of 2. 8 Å (Fig 1b, Table 1). The GOLD: 3A/CVB3 and GOLD: 3A/RVA2 complexes, however, failed to form diffracting crystals even after extensive optimization using both the mutagenesis and fusion-protein strategies. The overall structures of all solved GOLD: 3A complexes are highly similar to each other. This suggests that neither the L24A mutation in the GOLD: 3A/PV1 complex nor the fusion-protein strategy used for the GOLD: 3A/EVA71 complex affected the overall fold of the complexes (Fig 1b). No electron density was observed for the N termini of the 3A proteins (approximately the first 15 residues) and we, therefore, assume that this region is intrinsically disordered. This part of the 3A proteins has been previously reported to be involved in the interaction with another host factor GBF1 [2] or it is largely absent (e. g. RVA2) (Fig 1a). In order to determine the strength of the interaction between the ACBD3 GOLD domain and multiple enterovirus 3A proteins in vitro, we used microscale thermophoresis (Fig 1c). The dissociation constants of the GOLD: enterovirus 3A complexes ranged approximately from 1 μM (EVD68 and RVB14) to 15 μM (EVA71). In summary, our experiments confirmed that the enterovirus 3A proteins interact with the host ACBD3 protein through the GOLD domain of ACBD3 and the cytoplasmic domains of the 3A proteins. These proteins interact directly with dissociation constants within the low micromolar range. Using several approaches, four GOLD: enterovirus 3A complexes were crystallized and their structures were solved. Taken together, these structures document a conserved mechanism how diverse enterovirus species recruit the host ACBD3 protein. We performed an analysis of the GOLD: 3A interface to identify amino acid residues important for the ACBD3: 3A interaction, co-localization, stimulation of PI4KB recruitment, and facilitation of virus replication in human cells. For this analysis, we chose the GOLD: EVD68 3A complex because we resolved its structure at the highest resolution. Given the high similarity of the various GOLD: enterovirus 3A structures, we assume that the conclusions drawn from the ACBD3: EVD68 3A complex can be applied to the other ACBD3: enterovirus 3A complexes as well. In the GOLD: EVD68 3A crystal structure, we could trace the polypeptide chain of the 3A protein from T163A to I583A. It contains four secondary elements: two alpha helices P193A-V293A (α13A, Fig 2a) and Q323A-K413A (α23A, Fig 2b), and two beta strands I443A-I463A (β13A, Fig 2c) and V533A-I583A (β23A, Fig 2d). All these segments contribute to the GOLD: 3A interaction mediated through multiple hydrophobic interactions and hydrogen bonds (Fig 2a–2d). The helices α13A and α23A bind to a mild cavity of the GOLD domain that is formed by four antiparallel beta strands of ACBD3. The strand β13A interacts with the strand K518ACBD3-R528ACBD3 of the ACBD3 GOLD domain, while the strand β23A binds to the strand V402ACBD3-P408ACBD3, both in the antiparallel orientation. The conformation of all these secondary elements is highly conserved among various GOLD: enterovirus 3A complexes. The lowest homology of the tertiary structures of these complexes within short linkers between the β13A and β23A strands of the 3A proteins corresponds to the lowest homology of the primary sequences of these proteins within this region (Fig 1a and 1b). Calculations [24] of the changes of the interaction energies of various to-alanine mutants of these complexes based on their crystal structures uncovered that multiple amino acid residues of both 3A proteins and ACBD3 are involved in the interaction (S1 Fig). To evaluate the relative importance of various segments of the 3A protein on the complex formation, we designed the following EVD68 3A mutants: NLD (N23A/L26A/D30A), QRD (Q32A/R35A/D36A), IVH (I44A/V45A/H47A), and LVK (L52A/V54A/K56A) (Fig 3a; S1 Fig, panel a). For all the mutants, the ACBD3: 3A interaction was significantly attenuated both in the mammalian-two-hybrid assay (Fig 3b) and in the co-immunoprecipitation assay (Fig 3c), confirming that all four segments of the 3A protein are important for the ACBD3: 3A interaction. Nevertheless, some residual affinity of the 3A mutants to ACBD3 was still observed. All the 3A mutants co-localized with endogenous ACBD3 in the Golgi as did the wild-type 3A protein. The lipid kinase PI4KB, however, was recruited to the Golgi significantly more effectively in the cells expressing wild-type 3A compared to the cells expressing the 3A mutants (Fig 3d and 3e). Under physiological conditions, PI4KB cycles between the cytoplasm and Golgi, where it is recruited by a direct interaction with ACBD3 [7]. In enterovirus-infected cells, the viral 3A protein has been proposed to promote the ACBD3: PI4KB interaction [11]. Thus, considering that no direct interaction between the enterovirus 3A proteins and PI4KB has ever been observed, our data indicate that the stimulation of the ACBD3: PI4KB interaction by the 3A protein and the subsequent increase of the PI4KB recruitment to target membranes in infected cells depends on the ACBD3: 3A interaction. The Golgi-localized PI4P lipid was redistributed in the 3A-expressing cells possibly due to the Golgi disintegration caused by 3A overexpression, nevertheless, no significant change in the PI4P levels was observed in the wild-type 3A-expressing cells compared to the mock-transfected or mutant 3A-expressing cells (S2 Fig). Thus, a cooperation with some other viral proteins can be required to increase the PI4KB activity during viral infection. To analyze the impact of these 3A mutations on enterovirus replication, we established a reporter subgenomic replicon assay for EVD68. To determine the background reporter expression directly from the transfected RNA, we used a viral polymerase-lacking mutant (Δ3Dpol). Unexpectedly, no significant replication of the wild-type replicon RNA compared to the Δ3Dpol mutant was observed in HeLa cells. However, screening of several human cell lines uncovered the U-87 MG glioblastoma cells and HaCaT keratinocytes in which the wild-type replicon RNA significantly replicated. For all analyzed mutants, the viral RNA replication was attenuated in both cell lines (Fig 3f; S3 Fig). We observed no replication of the NLD, QRD, and IVH mutants, and a significantly reduced replication of the LVK mutant. Notably, this mutant was the weakest ACBD3 interactor in both co-immunoprecipitation and mammalian-two-hybrid assays, indicating additional unknown important effects distinct from the strength of the ACBD3-3A interaction affecting virus replication. We tested whether this mutant gained resistance to the PI4KB inhibition, nevertheless, we found that this mutant was still sensitive to a highly specific PI4KB inhibitor (compound 10 in Mejdrova et al. [10]) (Fig 3g). To address the effect of mutagenesis of selected residues within ACBD3, we designed the following ACBD3 mutants: WR (W375A/R377A), VTVRV (V403A/T404A/V405A/R406A/V407A), SYLF (S414A/Y415A/L416A/F417A), and RVYYT (R523A/V524A/Y525A/Y526A/T527A) (Fig 4a; S1 Fig, panels b-c). In the mammalian-two-hybrid assay (Fig 4b) and in the co-immunoprecipitation assay (Fig 4c), all these ACBD3 mutants displayed a significantly reduced ability to interact with the 3A protein. A weak yet significant effect was observed for the SYLF and RVYYT mutants, while a strong effect resulting in no detectable interaction in both assays was achieved for the WR and VTVRV mutants. Proper intracellular localization of these ACBD3 mutants was verified by their ectopic expression in ACBD3 knock-out cells derived from HeLa cells by CRISPR/Cas9 technology [21]. All these ACBD3 mutants co-localized with giantin, an integral Golgi protein, which has been proposed to directly recruit ACBD3 to the Golgi [5] (Fig 4d). Finally, we tested the ability of these ACBD3 mutants to rescue enterovirus replication in ACBD3 knock-out cells. The ACBD3 F258A/Q259A mutant, which does not interact with the lipid kinase PI4KB and cannot rescue virus replication [21], was used as a control. The ACBD3 WR and VTVRV mutants failed to rescue virus replication as expected. However, virus replication was still sufficiently restored by the other tested ACBD3 mutants SYLF and RVYYT (Fig 4e). These data document that the remaining affinity of these ACBD3 mutants to the viral 3A protein is still sufficient to fully facilitate enterovirus replication. Surprisingly, the SYLF mutant supports viral replication significantly better than wild-type ACBD3. It is possible that this mutation affects some other ACBD3 properties, such as its ability to interact with some other (known or unknown) proteins involved in enterovirus replication, nevertheless, the exact mechanism of the enhanced enterovirus replication in the ACBD3 SYLF mutant-expressing cells remains unclear. Compared to the ACBD3 WR mutant, the single mutants W375A and R377A still could rescue virus replication (S4 Fig, panels a-b), indicating that both mutations at the ACBD3: 3A interface are required to sufficiently disrupt the ACBD3: 3A interaction to affect virus replication. Several other tested ACBD3 mutants (such as V403A/V405A/ V407A, Y415A/F417A, and R523A/Y525A/Y526A) displayed a reduced affinity to the enterovirus 3A protein and still were able to restore enterovirus replication (S4 Fig, panels c-d). Alternatively, virus replication can be inhibited by single mutations interfering with a proper intracellular localization of ACBD3 (through ACBD3 misfolding and/or loss of the interaction with giantin) as documented by the E419A mutant (S4 Fig, panels e-f). In conclusion, our data document that the ACBD3: 3A interaction is essential for enterovirus replication. The viral replication, however, can be facilitated by weakly interacting ACBD3 mutants, provided that they are correctly folded and localized in the Golgi in non-infected cells. The enterovirus 3A proteins have been proposed to form homodimers [22,25]. Analysis of the crystal structures of the GOLD: 3A complexes revealed that the 3A proteins formed either one of the crystal-packing contacts (as in the case of EVD68 and PV1) or contacts with the second 3A molecule when two GOLD: 3A complexes per asymmetric unit were present (as in the case of EVA71 and RVB14). This putative dimerization interface is formed by the two central alpha helices of the 3A proteins, which are bent 180° to form a helical hairpin (Fig 5a). These helices are amphipathic with several hydrophobic residues oriented towards the hydrophobic residues of the other 3A monomer. Surprisingly, the C termini of the 3A proteins, which in the cellular environment are anchored to the membranes, are located on the opposite sides of the GOLD: 3A heterotetramers. Therefore, we were interested whether the plasticity and flexibility of the 3A dimerization interface together with the plasticity of the lipid bilayer allows to form the GOLD: 3A heterotetramers at the intracellular membranes. To identify amino acid residues of the 3A proteins involved in the dimerization of the GOLD: 3A complexes, we calculated [24] the changes of the dimerization energies of various to-alanine mutants of these complexes based on the crystal structures (S5 Fig, panel a). The dimerization interface of the GOLD: 3A/EVD68 complex consists of the hydrophobic core formed by the residues L25, V29, V34, and Y37, and an additional intermolecular salt bridge between the residues D24 and K41 (Fig 5a). To analyze the dimerization of the GOLD: 3A complexes in more detail, we generated a mutant EVD68 3A protein (hereafter referred to as LVVY mutant) with the following four mutations at the putative dimerization interface: L25A, V29A, V34A, and Y37A. As expected, retention volumes of the recombinant wild-type 3A and its LVVY mutant in size exclusion chromatography were significantly shifted corresponding to the dimeric and monomeric state of the wild-type 3A and its LVVY mutant, respectively (Fig 5b). At the request of a reviewer of our manuscript, we analyzed the dimerization of the L25V, V29Y, L25V/V34L, and V29Y/Y37V mutants (S6 Fig). Both L25V and V29Y mutations attenuated the 3A dimerization. The dimerization of the L25V mutant was restored by the V34L mutation, likely due to the compensation of weakening the L25-L25 interaction by strengthening the V34-V34 and V34-V29 interactions. On the other hand, the potential" rescue" Y37V mutation had a negative impact on the 3A dimerization, likely due to the attenuation of the Y37-L25 interaction and a loss of the hydrogen bond between Y37 and D24 (S6 Fig). Next, we investigated the stoichiometry of the GOLD: EVD68 3A complexes. To ensure that the 3A protein is fully complexed with the ACBD3 GOLD domain and to avoid the formation of partial complexes with 1: 2 stoichiometry, we designed a GOLD-EVD68 3A fusion protein using a similar approach as for the GOLD-EVA71 3A fusion protein used for the crystallographic analysis as described earlier. Taking advantage of the vicinity of the C terminus of the ACBD3 GOLD domain and the N terminus of the ordered part of the EVD68 3A protein, we connected the last residue of ACBD3 (R528ACBD3) through a short peptide linker (GSGSG) to the first ordered residue of the EVD68 3A protein (T163A/EVD68) (S5 Fig, panel b). Both GOLD-3A wild-type and LVVY mutant fusion proteins formed crystals, which diffracted to a resolution suitable for further structure determination (S5 Fig, panel c). The crystal structures of the GOLD: 3A complex consisting of two individual proteins, the GOLD-3A fusion protein, and its LVVY mutant were almost identical with low root-median-square deviations (S5 Fig, panel b), confirming that neither the fusion-protein approach nor the LVVY mutation affected the correct folding of these proteins. Three lines of evidence document the dimeric state of the wild-type GOLD-3A fusion protein and the monomeric state of its LVVY mutant in vitro. First, the retention volumes of these proteins in size exclusion chromatography are significantly shifted (Fig 5c). Secondly, the small-angle X-ray scattering (SAXS) profiles of these proteins correspond to the calculated scattering curves of a dimer of the wild-type GOLD-3A fusion protein (Fig 5d; S7 Fig, panel a) and of a monomer of its LVVY mutant (Fig 5e). Thirdly, crystal contacts corresponding to the 3A dimerization interface are not preserved in the crystal structure of the GOLD-3A LVVY mutant (S7 Fig, panels b-c), indicating that this mutant cannot dimerize through this interface even at very high protein concentrations (of approximately 20 mM) present within the protein crystal. Next, we investigated the stoichiometry of the GOLD: 3A complexes in cells. For this purpose, we ectopically co-expressed either wild-type GOLD-3A fusion protein or its LVVY mutant N-terminally fused to mAmetrine and mPlum fluorescent proteins in HeLa cells and determined the Förster resonance energy transfer (FRET) efficiency by flow cytometry. Owing to the crowding effect, the energy transfer was observed in the case of both proteins. Nevertheless, we observed a significant difference in FRET efficiency between the wild-type GOLD-3A fusion protein and its LVVY mutant (Fig 5f and 5g). These results confirm that the GOLD: 3A complexes are flexible enough to allow the formation of the heterotetramers consisting of two molecules of the viral 3A protein and two molecules of host ACBD3 even in cells at the respective intracellular membranes (Fig 5h). Finally, we analyzed the impact of the LVVY mutation on enterovirus replication. Using a reporter subgenomic replicon assay for EVD68 established earlier, we found replication of this mutant significantly attenuated in both U-87 MG and HaCaT cells (Fig 5i; S3 Fig). These findings document that the intact dimerization interface of the viral 3A proteins is required for enterovirus replication. In a previous study [23], we identified a novel ACBD3 membrane binding site (MBS) consisting of the residues R399, L514, W515, and R516. The hydrophobic residues L514 and W515 can be inserted directly into the lipid bilayer, while the positively charged residues R399 and R516 can interact with the negatively charged phospholipid head groups (Fig 6a). A vicinity of ACBD3 MBS and the expected position of the transmembrane domain of the enterovirus 3A protein within the ACBD3: 3A complexes suggests that ACBD3 MBS may be involved in the stabilization of the ACBD3: 3A complexes at the membrane as well. To experimentally evaluate this hypothesis, we designed the following ACBD3 mutants with several point mutations within MBS: LWR514AAA and, to increase repulsion between ACBD3 MBS and the lipid bilayer, LWR514DDD. Then, we ectopically expressed wild type ACBD3 or its MBS mutants N-terminally fused to EGFP in HeLa ACBD3 knock-out cells. We found that wild-type ACBD3 co-localized with the Golgi marker giantin, while both ACBD3 MBS mutants LWR514AAA and LWR514DDD were mostly released to the cytoplasm, although minor yet significant fractions of their pools were still preserved at the Golgi (Fig 6b). Remarkably, when the ACBD3 MBS mutants were co-expressed with the enterovirus 3A protein, they were completely (LWR514AAA mutant) or partially (LWR514DDD mutant) re-localized back to the Golgi (Fig 6c and 6d). Thus, an intact MBS is required for ACBD3 recruitment to the Golgi by the action of giantin or other cellular factors, however, it is dispensable for ACBD3 stabilization at target membranes through its interaction with enterovirus 3A proteins. Finally, we tested the capacity of wild-type ACBD3 and its MBS mutants to rescue virus replication in ACBD3 knock-out cells. Both ACBD3 wild type and the LWR514AAA mutant, but not the LWR514DDD and FQ258AA (used as a control [21]) mutants, effectively restored virus replication (Fig 6e). Thus, it seems that not ACBD3 MBS itself but rather the orientation of the ACBD3: 3A complex with respect to the membrane plays a role in facilitation of enterovirus replication. To-alanine mutations of ACBD3 MBS still allow the ACBD3: 3A complex at the membrane to adopt a conformation suitable for viral replication. On the contrary, to-aspartate mutations of ACBD3 MBS, which repel the negatively charged phospholipids of the lipid bilayer, result in an orientation of the ACBD3: 3A complex with respect to the membrane that does not support enterovirus replication. In summary, ACBD3 MBS is not required for ACBD3 recruitment to target membranes by the action of the enterovirus 3A proteins, however, the proper conformation of the ACBD3: 3A complexes at the membrane mediated by ACBD3 MBS is essential for enterovirus replication.
Considering the commonness of enterovirus-mediated infections within human population, it is surprising that no antiviral therapy for enteroviruses has been approved yet. Targeting essential host factors instead of rapidly mutating viral enzymes represents a promising strategy. Several host factors essential for enterovirus replication are recruited to the sites of viral replication by a direct protein-protein interaction between the host factor and a viral protein. A detailed knowledge of the structures of such complexes can open up prospects for a structure-guided development of small chemical compounds targeting these interactions, yielding a novel class of antivirals to combat infections caused by these pathogens. In this study, we present a series of crystal structures of complexes composed of the non-structural 3A proteins of four enterovirus species and the 3A-binding GOLD domain of the host factor ACBD3. Previously, the genetic inhibition of ACBD3 mediated by siRNA has yielded conflicting results on the importance of ACBD3 for virus replication [26,27]. This conflict in the literature has been recently addressed using CRISPR/Cas9-generated ACBD3 knock-out cells, in which enterovirus replication was severely impeded [8,21]. This confirmed that ACBD3 is, indeed, an essential host factor for enterovirus replication. However, it seems that a very low concentration of ACBD3 within the cells is still fully capable of facilitating enterovirus replication. This hypothesis is in agreement with our observations that enterovirus replication in ACBD3 knock-out cells can be restored by several ACBD3 mutants with a very low affinity to the viral 3A proteins even at the detection limit of conventional methods assessing the protein-protein interactions, such as protein co-immunoprecipitation. Among picornaviruses, the interaction between the viral 3A protein and host ACBD3 is not unique for enteroviruses. ACBD3 has been proposed to interact also with the 3A proteins of kobuvirus (e. g. aichivirus), hepatovirus, salivirus (klassevirus), and parechovirus, but not with those of cardiovirus (e. g. Saffold virus) or aphthovirus (foot-and-mouth disease virus, FMDV) [6]. To our best knowledge, all picornaviruses sensitive to PI4KB specific inhibitors (such as enteroviruses and kobuviruses) are able to hijack ACBD3, arguing for ACBD3 as a main mediator of PI4KB recruitment by these viruses. Notably, hepatovirus replicates independently of PI4KB [28], indicating either functionally irrelevant interaction with ACBD3 or another, PI4KB-independent, role of ACBD3 in hepatovirus replication. Picornaviruses that cannot hijack ACBD3 through their 3A proteins are either PI4P-independent (such as FMDV [29]) or their replication depends on another PI4P-producing lipid kinase PI4KA (e. g. cardioviruses [30]). During the past decades, multiple enterovirus mutants resistant to the inhibitors of PI4KB and the oxysterol binding protein (OSBP), which acts downstream of PI4KB, were isolated and characterized. Most of the resistance-conferring mutations were localized to the 3A-encoding regions of these viruses, e. g. PV1 N45Y, R54W, N57D, A70T, and A71S, CVB3 V45A, I54F, and H57Y, and RVB14 E30D/V/Q, I42V, and M54I (S8 Fig) [31–34]. Although many of the PI4KB/OSBP-inhibition resistance-conferring mutations are localized within the ACBD3-interacting regions of the 3A proteins, they seem unlikely to act through modulation of the ACBD3-3A interaction. On the other hand, it is possible that the mutations clustered within the β2 strands of the 3A proteins, such as CVB3 H57Y or PV1 R54W, can act through modulation of the interaction of the ACBD3-3A complex (or the uncomplexed 3A protein) with the membrane. The loss of the positive charge of the mutated residues can possibly compensate for the loss of the negative charge of the PI4P head groups upon PI4KB inhibition. Nevertheless, at least the mechanism of action of the mutations located distally with respect to the membrane, mostly clustered within the β1 strands of the 3A proteins, remains unclear. Apart from the crystal structures of the enterovirus 3A: GOLD complexes, to date only the structures of the kobuvirus 3A: GOLD complexes are known [23]. The enterovirus (e. g. poliovirus) and kobuvirus (e. g. aichivirus) 3A proteins share a common overall architecture, i. e. a similar size of approximately 10 kDa, large N-terminal soluble cytoplasmic domains followed by hydrophobic membrane-anchoring regions and small C-terminal domains (Fig 7a). Despite this common architecture, primary and predicted secondary structures of the enterovirus and kobuvirus 3A proteins are unrelated and cannot be aligned. Furthermore, positions of the ACBD3 binding regions of the enterovirus and kobuvirus 3A proteins are distinct. The ACBD3 binding region of the enterovirus 3A proteins forms the C-terminal segments of the cytoplasmic domain (and is preceded by the N-terminal GBF1 binding region), while the ACBD3 binding region of the kobuvirus 3A proteins is located at the N terminus (and a GBF1 binding region is completely missing). Superposition of the crystal structures of the enterovirus and kobuvirus 3A: GOLD complexes reveals that the enterovirus and kobuvirus 3A proteins bind to the same regions of the ACBD3 GOLD domain, nevertheless, the polypeptide chains of the enterovirus and kobuvirus 3A proteins have opposite polarities (Fig 7b and 7c). For instance, the poliovirus 3A strand β23A/PV1 binds in the antiparallel orientation to the strand V402ACBD3-P408ACBD3 of ACBD3, while the aichivirus 3A strand β13A/AiV1 binds at the same position to the same strand of ACBD3, but in the parallel orientation (Fig 7b and 7c). The reverse orientation of the enterovirus 3A proteins compared to the kobuvirus 3A proteins may be caused by the specific need of the enterovirus 3A proteins to bind GBF1. A notable difference between the enterovirus and kobuvirus 3A proteins is represented in the way they are anchored to the membrane. In addition to the C-terminal hydrophobic membrane binding regions, kobuvirus 3A proteins are membrane-anchored by the myristoylated N-terminal glycines, which is very unusual among picornaviruses [3]. Among enteroviruses, the N-terminal myristoylation is not present. To gain more insight into the membrane binding mode of the enterovirus 3A: GOLD protein complex, we performed all-atom molecular dynamics simulation of this complex at the membrane and compared it with our previously published [23] simulation of the kobuvirus 3A: GOLD protein complex at the surface of the lipid bilayer (Fig 7d). These simulations uncovered a similar conformation of the ACBD3 GOLD domain recruited to the membrane by the poliovirus or aichivirus 3A protein, including the insertion of the ACBD3 membrane binding site residues into the lipid bilayer. The position of the N-terminal myristoylation of the aichivirus 3A protein functionally substitutes the position of the C-terminal transmembrane domain of the poliovirus 3A protein, whereas the C-terminal transmembrane domain of the aichivirus 3A protein has no equivalent in the case of the poliovirus 3A protein. In summary, our findings reveal structural details of how two groups of viral pathogens, enteroviruses and kobuviruses, developed a similar mechanism of hijacking the same host factor (ACBD3) and its downstream effectors (such as PI4KB). These viruses use their 3A proteins with a common architecture yet totally unrelated primary sequences to bind to the same regions of the host ACBD3 protein in the opposite orientations, representing a striking case of convergence in picornavirus evolution. Our results are in agreement with a pioneering work by Greninger and colleagues [6,35], which has forecast such convergent evolutionary strategies of kobuviruses and enteroviruses based on extensive mutagenesis of the ACBD3-3A interface. There are still several other picornavirus genera proposed to recruit ACBD3 through the 3A: GOLD interaction (such as salivirus, hepatovirus, or parechovirus [6]) whose 3A: GOLD complexes remain structurally unexplored. Structural details of their ACBD3 recruitment uncovering whether they utilize the mechanism described for enteroviruses, kobuviruses, or another mechanism distinct from those two, remain to be further elucidated.
For expression in E. coli, full-length human ACBD3 and various enterovirus 3A proteins and their deletion mutants were cloned into pRSFD vector (Novagen) with an N-terminal 6xHis tag followed by a GB1 solubility tag and a TEV protease cleavage site using PCR and restriction cloning. For bacterial expression of the EGFP-fusion proteins, the EGFP encoding sequence was inserted between the TEV cleavage site and the target gene encoding regions. For expression of the EGFP-fusion proteins in human cells, target genes encoding regions were recloned into pEGFP-C1 vector (Clontech) with an N-terminal EGFP tag. For expression of the GST-, mAmetrine-, and mPlum-fusion proteins in human cells, the EGFP encoding region was replaced by GST or corresponding fluorescent protein encoding sequence by PCR and restriction cloning. The pRib-EVD68/mCherry plasmid for viral subgenomic replicon assays was generated by subcloning of the EVD68 cDNA of a prototypical Fermon strain under T7 promoter and replacing the capsid proteins-encoding region with the mCherry fluorescent protein-encoding gene by Gibson assembly. Mutations were generated using the Q5 Site-Directed Mutagenesis Kit (New England BioLabs). All DNA constructs were verified by sequencing. The pEGFP-ACBD3 and pBIND-ACBD3 plasmids were kindly provided by Carolyn Machamer and Jun Sasaki, respectively. The mAmetrine-C1 and mPlum-C1 plasmids were gifts from Robert Campbell and Michael Davidson (Addgene plasmids #54660 [36] and #54839 [37]). The pEGFP-GalT plasmid was a gift from Jennifer Lippincott-Schwartz (Addgene plasmid #11929 [38]). All recombinant proteins used in this study were bacterially expressed as fusion proteins with an N-terminal 6xhistidine (His6) tag followed by a GB1 solubility tag and a TEV protease cleavage site. For the crystallographic analysis of the GOLD: 3A complexes, the N-terminally His6-GB1-TEV site-fused cytoplasmic domains of the 3A proteins were directly co-expressed with the untagged ACBD3 GOLD domain. The proteins were expressed in E. coli BL21 DE3 NiCo cells (New England Biolabs) using the autoinduction ZY medium. Bacterial cells were harvested and lysed in the lysis buffer (50 mM Tris pH 8,300 mM NaCl, 3 mM β-mercaptoethanol, 30 mM imidazole, 10% glycerol), the lysate was incubated with the HisPur Ni-NTA Superflow agarose (Thermo Fisher Scientific), and the bound proteins were extensively washed with the wash buffer (50 mM Tris pH 8,300 mM NaCl, 1 mM β-mercaptoethanol, 20 mM imidazole). The protein was eluted with the elution buffer (50 mM Tris pH 8,200 mM NaCl, 3 mM β-mercaptoethanol, 300 mM imidazole). For the biochemical analysis of the 3A proteins by microscale thermophoresis or SAXS, the N-terminal His6-GB1 tags were preserved uncleaved to increase protein solubility and to avoid aggregation at required concentrations. For the crystallographic analysis of the GOLD: 3A complexes and for the biochemical analysis of the GOLD-3A fusion proteins by SAXS, the N-terminal His6-GB1 tags were removed with home-made TEV protease. Next, the proteins were purified using the size exclusion chromatography at HiLoad 16/60 Superdex 75 or Superdex 200 prep grade columns (GE Healthcare) in the storage buffer (10 mM Tris pH 8,200 mM NaCl, 3 mM β-mercaptoethanol). In addition, the GOLD: 3A complexes used for the crystallographic analysis were further purified by reverse immobilized metal affinity chromatography using the HisTrap HP column (GE Healthcare), while the EGFP-fused ACBD3 GOLD domain used for microscale thermophoresis was further purified using the ion exchange chromatography at a MonoQ 10/100 GL column (GE Healthcare) and then dialyzed back into the storage buffer. The molecular weight and purity of all proteins was verified by SDS-PAGE (S9 Fig) and Matrix-Assisted Laser Desorption/Ionisation (MALDI). Purified proteins were concentrated to 1–10 mg/ml, aliquoted, flash frozen in the liquid nitrogen, and stored at -80 °C until needed. Crystals grew at 291 K in sitting drops by the vapor diffusion method. They were cryoprotected, flash frozen in liquid nitrogen, and analyzed. Measurements were carried out at the MX14. 1 beamline of the synchrotron BESSY II at Helmholtz-Zentrum Berlin [39]. The crystallographic datasets were collected from single frozen crystals. Data were integrated and scaled using XDS [40] and XDSAPP [41]. Structures were solved by molecular replacement using the uncomplexed ACBD3 GOLD domain structure (pdb code 5LZ1) as a search model. The initial models were obtained with Phaser [42] from the Phenix package [43]. The models were further improved using automatic model building with Buccaneer [44] from the CCP4 suite [45], automatic model refinement with Phenix. refine [46] from the Phenix package [43], and manual model building with Coot [47]. Statistics for data collection and processing, structure solution and refinement are summarized in Table 1. Structural figures were generated with PyMol [48]. The atomic coordinates and structural factors were deposited in the Protein Data Bank (www. pdb. org). MST measurements were carried out using the Monolith NT. 115 instrument (NanoTemper Technologies) according to the manufacturer' s instructions. The Monolith NT. 115 standard treated capillaries were loaded with a mixture of a recombinant EGFP-fused protein at a constant concentration of 150 nM in the MST buffer (30 mM Tris pH 7. 4,150 mM NaCl, 3 mM β-mercaptoethanol) and its binding partner in the appropriate series of concentrations. The thermophoretic motion of the fluorescent protein and its temperature-dependent changes of fluorescence were analyzed with the Monolith NT Analysis Software. Human cervical-carcinoma cells HeLa (American Type Culture Collection / ATCC), embryonic kidney cells HEK293T (ATCC), and keratinocytes HaCaT (AddexBio) were maintained in Dulbecco' s modified Eagle' s medium (Sigma) supplemented with 10% fetal calf serum (Gibco). Human glioblastoma cells U-87 MG (ATCC) were maintained in Minimum Essential Medium Eagle (Sigma) supplemented with 10% fetal calf serum (Gibco), GlutaMAX Supplement (Thermo Fisher Scientific), and non-essential amino acids (Biowest). HeLa cells were transfected using Lipofectamine2000 reagent (Thermo Fisher Scientific) or X-tremeGENE HP DNA Transfection reagent (Sigma/Roche) according to manufacturer' s instructions. Transfections of HEK293T cells were performed using polyethylenimine (Sigma) or Fugene6 (Promega). HEK293T cells were transfected with the appropriate mutants of the EGFP-fused EVD68 3A protein and GST-fused ACBD3. The next day, cells were harvested, washed twice with phosphate-buffered saline (PBS) and lysed in the ice-cold lysis buffer (20 mM Tris pH 7. 4,100 mM NaCl, 50 mM NaF, 10 mM EDTA, 10% glycerol, 1% NP-40), supplemented with protease inhibitors (Complete protease inhibitor cocktail, Sigma/Roche). After solubilizing for 15 min on ice, the lysate was pre-cleared by centrifugation at 16,000g for 15 min. The resulting supernatant was incubated with sepharose beads coupled either to GFP nanobody (GFP-Trap, ChromoTek) or glutathione (Protino Glutathione Agarose, Macherey-Nagel) for 1h at 4 °C. After three washes with 10 volumes of the lysis buffer, the bound proteins were directly eluted with the Laemmli sample buffer, subjected to SDS-PAGE, and analyzed by immunoblotting. The whole cell lysates and eluted proteins were stained with mouse monoclonal antibodies to ACBD3 (Santa Cruz Biotechnology, sc-101277) and GFP (Santa Cruz Biotechnology, sc-9996). The images were acquired using the LI-COR Odyssey Infrared Imaging System. HeLa cells were co-transfected with plasmids encoding target proteins fused to the FRET pair of fluorescent proteins mAmetrine (" donor" ) and mPlum (" acceptor" ). The next day, cells were harvested, washed twice with PBS, and analyzed by flow cytometry using BD LSR Fortessa (BD Biosciences). The donor and acceptor fluorescence as well as the energy transfer was determined using the optical configurations as follows: mAmetrine—405 nm laser—525/50 nm bandpass filter; mPlum—561 nm laser—670/30 nm bandpass filter; FRET—405 nm laser—655/8 nm bandpass filter. Acquired data were analyzed with the FlowJo software. The acquired fluorescence intensities were compensated and the same gate corresponding to the live transfected cells with the approximately 1: 1 donor: acceptor ratio was applied to all samples. The acquired events were binned on the basis of the acceptor fluorescence intensity, and the average FRET fluorescence intensities of each bin were plotted against the respective acceptor fluorescence intensity. HEK293T cells grown in 96-well plates were co-transfected with 50 ng of each pACT, pBIND, and pG5Luc plasmids using Fugene6 (Promega). At 24 hours post transfection, the cells were lysed, and both firefly and Renilla luciferase activities were measured using the Dual-Luciferase assay kit (Promega) and Centro LB 960 luminometer (Berthold Technologies) according to the manufacturer' s instructions. The firefly luciferase activity was normalized to the Renilla luciferase activity (used as an internal control of the transfection efficiency) and then to the activity determined in cells co-expressing wild-type ACBD3 and 3A (which was set to 100%). HeLa cells grown on coverslips in 24-well plates were transfected with 400 ng of the plasmid DNA using Lipofectamine2000 (Thermo Fisher Scientific). At 16 hours post transfection, the cells were fixed with 4% paraformaldehyde for 15 min at room temperature, permeabilized with 0. 1% Triton X-100 in PBS for 5 min, and immunostained with the appropriate primary and secondary antibodies diluted in 2% normal goat serum in PBS. Sources of the antibodies were as follows: anti-ACBD3 (Sigma, WH0064746M1), anti-PI4KB (Merck, 06–578), anti-GM130 (BD Biosciences, 610822), anti-giantin (Enzo Life Science, ALX-804-600-C100), anti-myc (Thermo Fisher Scientific, PA1-981), anti-PI4P (Echelon, Z-P004), and goat-anti-mouse and goat-anti-rabbit secondary antibodies conjugated to AlexaFluor 488,596, or 647 (Molecular Probes). Nuclei were stained with DAPI. Coverslips were mounted with FluorSave (Calbiochem), and confocal imaging was performed with a Leica SpeII confocal microscope. The pRib-EVD68/mCherry wild-type and mutant plasmids were linearized by cleavage with SalI-HF (Thermo Fisher Scientific) and purified using the mini spin columns (Epoch Life Science). Viral subgenomic replicon RNA was generated with TranscriptAid T7 High Yield Transcription Kit (Thermo Fisher Scientific) and purified using the RNeasy mini spin columns (Qiagen). For replicon assays, U-87 MG or HaCaT cells grown in 12-well plates were transfected with T7-amplified RNA using the TransIT mRNA transfection kit (Mirus Bio). At 12 hours post transfection, the reporter mCherry fluorescence was determined by flow cytometry using BD LSR Fortessa (BD Biosciences) and the following optical configuration: 561 nm laser, 670/30 nm bandpass filter. Acquired data were analyzed with the FlowJo software. The level of RNA replication was expressed as a transfection efficiency-normalized percentage of cells with the mCherry signal above the threshold determined using the viral polymerase-lacking mutant Δ3Dpol. To test the effect of PI4KB inhibition on virus replication, a PI4KB-specific inhibitor (compound 10 from [10] kindly provided by Radim Nencka) was added to the medium at a final concentration of 1 μM 30 min prior transfection of the viral subgenomic replicon RNA. Wild-type or ACBD3 knock-out HeLa cells grown in 96-well plates were transfected with plasmids encoding wild-type or mutant ACBD3 or another Golgi-resident protein GalT as a control. At 24 hours post transfection, the cells were infected with the Renilla luciferase-expressing CVB3 virus (RLucCVB3) [49]. At 8 hours post infection, the intracellular Renilla luciferase activity was determined using the Renilla luciferase assay system (Promega) and a Centro LB 960 luminometer (Berthold Technologies). The Renilla luciferase activity was normalized to the activity determined in wild-type ACBD3-expressing cells (which was set to 100%). Proteins were dialyzed against the SAXS buffer (30 mM Tris pH 7. 4,150 mM NaCl, 1 mM TCEP) and concentrated as follows: c1 = 1. 03 mg/ml, c2 = 1. 4 mg/ml and c3 = 1. 54 mg/ml for the wild-type GOLD-3A fusion protein, and c1 = 0. 96 mg/ml, c2 = 2. 36 mg/ml and c3 = 3. 26 mg/ml for the GOLD-3A LVVY mutant. The SAXS data were collected using the beamlines BioSAXS Beamline BM29 (ESRF, Grenoble) and EMBL SAXS beamline P12 (Petra III DESY, Hamburg) that are both equipped with the 2M Pilatus detector (Dectris). The three datasets overlay after rescaling, indicating no protein aggregation in the samples. To structurally interpret the SAXS data, we incorporated the missing loop (D437-K473) into the structures of the wild-type GOLD-3A dimer and GOLD-3A LVVY mutant monomer. Next, we performed the coarse-grained molecular simulations [50] in which only the conformations of the D437-K473 loop were sampled while the crystallized portion of the protein was kept rigid, yielding 10,000 structural models of both proteins. For all structural models, we computed the SAXS intensity profiles using the previously developed algorithms [51], compared them individually to the experimental SAXS data, and selected the models best fitting the SAXS data collected on the samples with the highest protein concentrations. The best models fit the SAXS data with χ = 1. 5 for the wild-type GOLD-3A dimer and χ = 1. 4 for the GOLD-3A LVVY mutant monomer. The intrinsically disordered region of the ACBD3 GOLD domain, which was missing in the crystal structure (D437-K473), was modeled as described previously [23]. The C-terminal segment of the poliovirus 3A protein (N57-Q87) was modeled as a loop (N57-M60) followed by a transmembrane alpha helix (T61-F83) and a short C-terminal tail (A84-Q87). This segment was positioned in a planar segment of a lipid bilayer using the PPM server [52]. MD simulations of the ACBD3 GOLD domain in complex with the poliovirus 3A protein in the environment of a lipid bilayer were performed following the procedures recently used to study the ACBD3 GOLD domain in complex with the aichivirus 3A protein [23]. The initial system for MD simulations was prepared using VMD version 1. 9. 2 [53]. Namely, a POPC bilayer segment with the lateral dimensions of 10 nm by 10 nm was formed using the Membrane Plugin version 1. 1 in VMD. The GOLD: 3A complex was placed on top of the resulting lipid patch. The lipids overlapping with the transmembrane alpha helix of the 3A protein were removed. The system was solvated using the Solvate Plugin version 1. 5 in VMD. Sodium and chloride ions were added to neutralize the simulated system and to reach a physiological concentration of 150 mM. The MD simulations were performed using the NAMD package [54] version 2. 9. The CHARMM22 force field [55,56] with the CMAP correction [57] and the TIP3P water model were used. The simulations were carried out in the NPT ensemble. Temperature was kept at 298K through a Langevin thermostat with a damping coefficient of 1/ps. Pressure was maintained at 1 atm using the Langevin piston Nose-Hoover method with a damping timescale of 50 fs and an oscillation timescale of 100 fs. Short-range non-bonded interactions were cut off smoothly between 1 and 1. 2 nm. Long-range electrostatic interactions were computed using the particle-mash Ewald method with a grid spacing of 0. 1 nm. Simulations were performed with an integration time step of 2 fs. After initial energy minimization with a conjugate gradient method, a 10 ns simulation was performed with constraints on the protein backbone atoms in order to equilibrate the lipids, ions and water molecules. Namely, a harmonic potential with the spring constant of 5 kcal/ (mol Å2) was applied to all backbone atoms of the GOLD: 3A complex. After the equilibration, the system was simulated with no constrains for 200 ns. The resulting MD trajectory was visualized and analyzed using VMD. To localize the PI4KB/OSBP-inhibition resistance-conferring mutations within the structure of the GOLD: CVB3 3A complex, a homology model of this complex was generated by the I-TASSER server [58] using the crystal structure of the GOLD: EVD68 3A complex as a template. In the graphs, data are presented as mean values ± standard errors of the means (SEMs) based on three independent experiments, unless stated otherwise. For statistical analyses, two-tailed two-sample Student' s t-tests were used. P-values below 0. 05 were considered significant. The crystal structures of the ACBD3 GOLD domain in complex with the 3A proteins from EVD68, RVB14, and PV1 (L24A mutant), and the ACBD3 GOLD domain fused to the 3A proteins from EVA71, EVD68 (wild type), and EVD68 (LVVY mutant) from this publication have been submitted to the Protein Data Bank (www. pdb. org) and assigned the identifiers 6HLN, 6HLT, 6HLV, 6HLW, 6HM8, and 6HMV, respectively. | Enteroviruses are the most common viruses infecting humans. They cause a broad spectrum of diseases ranging from common cold to life-threatening diseases, such as poliomyelitis. To date, no effective antiviral therapy for enteroviruses has been approved yet. To ensure efficient replication, enteroviruses hijack several host factors, recruit them to the sites of virus replication, and use their physiological functions for their own purposes. Here, we characterize the complexes composed of the host protein ACBD3 and the ACBD3-binding viral proteins (called 3A) of four representative enteroviruses. Our study reveals the atomic details of these complexes and identifies the amino acid residues important for the interaction. We found out that the 3A proteins of enteroviruses bind to the same regions of ACBD3 as the 3A proteins of kobuviruses, a distinct group of viruses that also rely on ACBD3, but are oriented in the opposite directions. This observation reveals a striking case of convergent evolutionary pathways that have evolved to allow enteroviruses and kobuviruses (which are two distinct groups of the Picornaviridae family) to recruit a common host target, ACBD3, and its downstream effectors to the sites of viral replication. | Abstract
Introduction
Results
Discussion
Materials and methods | medicine and health sciences
crystal structure
protein interactions
pathology and laboratory medicine
pathogens
condensed matter physics
microbiology
viruses
rna viruses
crystallography
physical chemistry
chemical properties
infectious diseases
solid state physics
dimerization
proteins
medical microbiology
enterovirus infection
microbial pathogens
chemistry
viral replication
physics
enteroviruses
biochemistry
virology
poliovirus
viral pathogens
protein domains
biology and life sciences
physical sciences
viral diseases
organisms | 2019 | Convergent evolution in the mechanisms of ACBD3 recruitment to picornavirus replication sites | 14,951 | 300 |
Coarse-grained (CG) simulations have become an essential tool to study a large variety of biomolecular processes, exploring temporal and spatial scales inaccessible to traditional models of atomistic resolution. One of the major simplifications of CG models is the representation of the solvent, which is either implicit or modeled explicitly as a van der Waals particle. The effect of polarization, and thus a proper screening of interactions depending on the local environment, is absent. Given the important role of water as a ubiquitous solvent in biological systems, its treatment is crucial to the properties derived from simulation studies. Here, we parameterize a polarizable coarse-grained water model to be used in combination with the CG MARTINI force field. Using a three-bead model to represent four water molecules, we show that the orientational polarizability of real water can be effectively accounted for. This has the consequence that the dielectric screening of bulk water is reproduced. At the same time, we parameterized our new water model such that bulk water density and oil/water partitioning data remain at the same level of accuracy as for the standard MARTINI force field. We apply the new model to two cases for which current CG force fields are inadequate. First, we address the transport of ions across a lipid membrane. The computed potential of mean force shows that the ions now naturally feel the change in dielectric medium when moving from the high dielectric aqueous phase toward the low dielectric membrane interior. In the second application we consider the electroporation process of both an oil slab and a lipid bilayer. The electrostatic field drives the formation of water filled pores in both cases, following a similar mechanism as seen with atomistically detailed models.
Since the first introduction of physics-based coarse-grained (CG) models in computational biology [1], CG models have become increasingly popular in the simulation of complex biological systems [2]. They significantly reduce the computational complexity in comparison to all-atom (AA) models and allow sampling over much longer time scales and of larger system sizes. One of the most widely applied CG models is the MARTINI force field [3]. The MARTINI model was initially developed for lipid systems [4] and has recently been extended for proteins [5] and carbohydrates [6]. In general a four-to-one mapping is used in MARTINI, which means that on average four atoms and associated hydrogens are represented by one CG bead. The CG particles interact with the other CG particles in the system by means of Lennard-Jones (LJ) interactions; in addition charged groups (e. g. ions, lipid head groups, charged amino acid side chains) interact via a Coulombic energy function. Water is treated explicitly, at the same level of coarse-graining as all other molecules implying that four water molecules are combined into a single coarse-grained bead. MARTINI water beads, just as many other CG water models, do not bear charges and, consequently, are blind to electrostatic fields and polarization effects. To compensate for the neglect of explicit polarization, screening of electrostatic interactions is done implicitly, assuming a uniform relative dielectric constant. While this is a reasonable approximation for bulk water, problems arise at the interfaces between water and other phases and in the vicinity of charged particles. Because of the implicit screening, the interaction strength of polar substances is underestimated in non-polarizable solvents. Correct modeling of the partitioning of polar and charged compounds into a low dielectric medium, e. g. a lipid bilayer, has proven a big challenge for CG models in general [7]. Applications involving the formation of polar/charged complexes in a non-polar environment are especially prone to be affected. A potential solution is to make the interaction potentials dependent on the local environment (see e. g. [8]), especially useful in solvent free approaches. With explicit solvent particles present, more flexibility is achieved with a polarizable water model. Attempts to include the effect of polarization in simplified water models date already back to the early days of biomolecular modeling. Notably the development of the soft sphere dipole model is worth mentioning [9], [10]. In this model, water molecules are represented by point dipoles that can reorient in response to the electrostatic field of an embedded (macro) molecule. Recently, induced dipoles were also added to a CG solvent model, and made compatible with a CG protein force field [11]. The polarizability challenge also stands in all-atom (AA) force fields at a more fine-grained level. The AA force fields lack electronic polarizability, which has proven to be a significant drawback in simulations of ions and highly polarizable systems [12], [13]. Several approaches to develop polarizable AA force fields, such as the inducible point dipole model [14], the model with Drude oscillators [15], [16], the fluctuating charge model [17] and the multipole expansion model [18] exist. The general idea of all these methods is to introduce a fluctuating dipole to each polarizable particle, which responds to the local electric field in the vicinity of this particle. In this work, we introduce orientational polarizability to the water beads of the MARTINI force field using an approach similar to that of the Drude oscillator [15], [16]. The resulting polarizable CG water model, in combination with the MARTINI force field, allows modeling the interaction of water with charged particles in a more realistic way. In the parameterization of the polarizable water model the following three criteria were used: i) The dielectric constant of bulk polarizable water should be sufficiently close to the value in real water; ii) The particle density of the polarizable water should be close to the particle density of the water in standard MARTINI; iii) The reproduction of partitioning free energies between water and organic solvents for a large variety of small compounds, one of the corner stones of the MARTINI model, should remain unaffected. The rest of this paper is organized as follows. In the next section, we first describe the details of the model, and the way we set out to parameterize it. This is followed by the Results section in which we explore the parameter space and arrive at the optimal parameter set, based on reproduction of the density and dielectric constant of bulk water, and the water/oil partitioning behavior of the MARTINI building blocks. We then test a number of properties of the new model, including the dynamical behavior of bulk water, the surface tension of the water/vapor interface, and structural properties of ionic solutions and of a lipid bilayer. We also look at the effect of long-range electrostatic interactions. Finally, two applications are shown which would not have been feasible with the standard MARTINI model, nor with most other CG models. The applications are a realistic description of the free energy of ion transport across a lipid bilayer, and the electroporation process of both an octane slab and a lipid bilayer. A discussion section about the limitations and prospects of the model concludes this paper.
The polarizable CG water consists of three particles instead of one in the standard MARTINI force field (Fig. 1). The central particle W is neutral and interacts with other particles in the system by means of the Lennard-Jones interactions, just like the standard water particle. The additional particles WP and WM are bound to the central particle and carry a positive and negative charge of +q and −q, respectively. They interact with other particles via a Coulomb function only, and lack any LJ interactions. The bonds W-WP and W-WM are constrained to a distance l. The interactions between WP and WM particles inside the same CG water bead are excluded, thus these particles are “transparent” toward each other. As a result the charged particles can rotate around the W particle. The dipole momentum of the water bead depends on the position of the charged particles and can vary from zero (charged particles coincide) to 2lq (charged particles are at the maximal distance). A harmonic angle potential with equilibrium angle θ and force constant Kθ is furthermore added to control the rotation of WP and WM particles and thus to adjust the distribution of the dipole momentum. The average dipole momentum of the water bead will depend on the charge distribution and is expected to be on average zero in an apolar environment, such as the interior of the lipid bilayer. In contrast, some non-zero average dipole will be observed in bulk water or in some other polar environment. The masses of the charged particles as well as of the central particle are set to 24 amu, totaling 72 amu (the mass of four real water molecules). There are five adjustable parameters in the polarizable water particle: The charge q, the distance l, the angle parameters θ and Kθ, and the atom type of the central particle W. The accessible range of the dipole momentum of the water bead is determined by both l and q; to restrict our parameter space we used q as the only adjustable parameter and fixed l at the value 0. 14 nm. This distance is small enough to prevent the overlap of the charged particles of adjacent water beads (which could result in very large forces) and large enough to represent the instantaneous dipole of a cluster of four water molecules. Similarly, only Kθ was varied. The equilibrium angle was fixed at θ = 0 to ensure that the water bead in an apolar solvent has a vanishing dipole moment (recalling that one CG water bead effectively represents a cluster of four real water molecules). It is clear that the polarizable water beads attract each other stronger than the standard CG water beads because of additional electrostatic interactions between their charged particles WP and WM. This additional attraction should be counter-balanced by a reduced LJ self-interaction of W particles. Thus we tested less attractive interaction levels II, III and IV (the standard MARTINI water has the atom type P4, which has a self-interaction strength level I. Note that, for each of these levels, the LJ parameter σLJ = 0. 47 nm. The LJ well depth εLJ = 5. 0,4. 5,4. 0,3. 5 kJ mol−1 for levels I–IV, respectively). Concerning the LJ interactions between the W particles and other particles in the MARTINI force field, our expectation was that these could stay unaffected, i. e. correspond to those for a P4 particle (for the full interaction matrix, see [3]). However, as we will show below, the cross-interaction strength has to be reduced slightly in order to reproduce the correct partitioning behavior. Since the polarization of water is treated explicitly in our polarizable model, the global dielectric constant εr = 15 used in the standard MARTINI should be adjusted accordingly. This value of εr compromises between large ε in water and small ε in the hydrophobic regions like the core of the lipid membrane. In the polarizable model, the global dielectric constant is reduced to εr = 2. 5 to ensure a realistic dielectric behavior in the hydrophobic regions. Other force field parameters are the same as in standard MARTINI [3]. All simulations were performed with the GROMACS suite of programs, versions 3. 3. 1 [19], 4. 0. 2 and 4. 0. 5 [20]. Standard simulation parameters associated with the MARTINI force field [3] were used unless stated otherwise. A time step of 20 fs was used in all simulations. We have repeated some of the simulations using 10 fs and 30 fs time steps; the results were virtually identical to the ones reported below. Temperature and pressure were kept constant by using weak coupling schemes [21], with time constants of τT = 0. 3 ps and τp = 3. 0 ps, respectively. The distance l between the central W particle and the charged WP/WM particles was constrained using the LINCS algorithm [22]. Visualization of the results was done with VMD [23]. Error estimates were obtained using a block-averaging procedure [24]. Details of the system composition and set-up are given alongside the presentation of the results. Times are reported as actual simulation time, except when explicitly stated as effective time in order to compare the kinetics to either all-atom simulations or experiment. The effective time accounts for the speed-up in coarse-grained dynamics (see [4]) and equals four times the actual simulation time. The parameter files are available in Dataset S1. They can also be downloaded from http: //cgmartini. nl, together with some example applications.
The neglect of orientational polarizability in many water models associated with CG lipid force fields [54], [55], [56], [57], [58] is arguably one of the crudest approximations made. Water in those force fields is represented by spherically symmetric interaction sites either based on analytic potentials or effective potentials derived from atomistic simulations. None of these water models include electrostatic interactions, implying they are non-polarizable. The MARTINI model suffers from the same approximation. The inability to form a transmembrane water pore upon dragging a lipid across the membrane [3], or upon binding of antimicrobial peptides [41], [59], [60] are examples pointing at the shortcoming of the standard MARTINI water model. To improve the behavior of the water model, inclusion of electrostatic interactions is needed; to account for the orientational polarizability, the minimum requirement is a point dipole, as in the models of e. g. Warshel and coworkers [9], [10], [13] and Orsi et al [61]. Our new water model is a three-bead model, consisting of a central particle with two charges-on-a-spring embedded, and was chosen as it combines simplicity with versatility. It is similar to the classical Drude model used in polarizable all-atom (AA) force fields to mimic electronic polarization [15], [16]. In contrast to the AA case, were the charged particles are massless and their position is solved in an expensive, self-consistent way, in our CG model the particles carry mass and follow the normal equation of motions. The model has only few adjustable parameters, yet enough of them to reproduce the dielectric properties of bulk water on the one hand and keeping at par with the standard MARTINI philosophy on the other. Despite the limited amount of free parameters in the model, a full exploration of parameter space is practically impossible; guided partly by intuition and partly through extensive testing we eventually settled on a combination of parameters which, overall, perform very well. Compared to the standard MARTINI water model, the polarizable model has improved properties, not only with respect to its dielectric behavior, but also for instance in the somewhat reduced freezing point. It can not be excluded that other combinations of parameters might perform even better, and we anticipate that further optimization of the model will take place in the future alongside with extending the range of applications of the model. The main reason for having included polarizability into the model is the expectation that processes involving interactions between charged and polar groups in a low-dielectric medium are more realistically described. As an example we presented two applications for which standard CG models, including MARTINI, are less well suited, namely the translocation of ions across a lipid membrane and the electroporation of an octane slab and a lipid bilayer. Both processes involve the movement of charges from a high dielectric environment (water) to a low dielectric medium (membrane interior). A realistic description of such processes requires a model capable of performing local electrostatic screening. The two applications presented show that, despite being coarse-grained, our polarizable water model can do this at a level comparable to that of atomistic simulations. This opens the way to explore a number of important (bio) physical processes using the MARTINI model, including membrane poration by antimicrobial and cell penetrating peptides, DNA transfection, salt-induced membrane fusion, functioning of the voltage gated membrane channels, electroporation, and electrokinetic phenomena in general. Finally, it is important to point out a few limitations of the polarizable water model: First, it is slightly more expensive from a computational point of view (for a pure water system the simulations are slowed down by a factor of approximately three). Second, the current parameterization of the model is not as thoroughly tested yet in comparison to the standard MARTINI model. For example lipid phase behavior, or the effect on proteins and peptides is largely unexplored. Third, despite an overall improved performance, some properties are still not at par with experimental measurements or data from atomistic simulations. These include the air/water surface tension, which is significantly too low, and also the sign of the membrane dipole potential which is opposite to that observed with more detailed force fields. Further improvement could be obtained by changing the analytical form of the non-bonded potential (i. e. moving away from the LJ 12-6 form), and by adding polarizability to other beads in the force field. The latter idea may also lead to a more realistic description of the protein backbone, allowing secondary structure formation to be described with MARTINI, an option we are currently exploring. We finally note that the polarizable MARTINI water model is not meant to replace the standard MARTINI water model, but should be viewed as an alternative with improved properties in some, but similar behavior at reduced efficiency in other applications. | Many biomolecular processes involve charged species moving between regions of high polarity, such as the water phase, and regions of lower polarity, such as the lipid membrane. Due to the change in electrostatic screening between these two environments, the strength of the interactions between the moving charge and the surrounding molecules also changes. This has important consequences for the way biological activity is controlled. To help understand the forces driving the movement of biomolecules, we developed a computational model which is capable of describing these processes at near-atomic detail. To do so efficiently, we use a coarse-grained description of the molecules, in which some of the atomistic detail is averaged out. To capture the inhomogeneous nature of the dielectric response, we re-introduce some detail in the water model; the new model effectively mimics the orientational polarizability of real water molecules, and screens electrostatic interactions realistically. This enables the study of a number of important biological processes that were hitherto considered challenging for coarse-grained models, such as the permeation of ions across a lipid membrane and the rupture of membranes due to an electrostatic field, at relatively low computational cost. | Abstract
Introduction
Methods
Discussion | biophysics/theory and simulation
biochemistry/theory and simulation | 2010 | Polarizable Water Model for the Coarse-Grained MARTINI Force Field | 4,114 | 263 |
Schistosomiasis japonica is a serious debilitating and sometimes fatal disease. Accurate diagnostic tests play a key role in patient management and control of the disease. However, currently available diagnostic methods are not ideal, and the detection of the parasite DNA in blood samples has turned out to be one of the most promising tools for the diagnosis of schistosomiasis. In our previous investigations, a 230-bp sequence from the highly repetitive retrotransposon SjR2 was identified and it showed high sensitivity and specificity for detecting Schistosoma japonicum DNA in the sera of rabbit model and patients. Recently, 29 retrotransposons were found in S. japonicum genome by our group. The present study highlighted the key factors for selecting a new perspective sensitive target DNA sequence for the diagnosis of schistosomiasis, which can serve as example for other parasitic pathogens. In this study, we demonstrated that the key factors based on the bioinformatic analysis for selecting target sequence are the higher genome proportion, repetitive complete copies and partial copies, and active ESTs than the others in the chromosome genome. New primers based on 25 novel retrotransposons and SjR2 were designed and their sensitivity and specificity for detecting S. japonicum DNA were compared. The results showed that a new 303-bp sequence from non-long terminal repeat (LTR) retrotransposon (SjCHGCS19) had high sensitivity and specificity. The 303-bp target sequence was amplified from the sera of rabbit model at 3 d post-infection by nested-PCR and it became negative at 17 weeks post-treatment. Furthermore, the percentage sensitivity of the nested-PCR was 97. 67% in 43 serum samples of S. japonicum-infected patients. Our findings highlighted the key factors based on the bioinformatic analysis for selecting target sequence from S. japonicum genome, which provide basis for establishing powerful molecular diagnostic techniques that can be used for monitoring early infection and therapy efficacy to support schistosomiasis control programs.
Schistosomiasis is caused by Schistosoma haematobium, S. mansoni, S. japonicum, and less frequently, S. mekongi and S. intercalatum. It occurs in the tropics and subtropics and it is among the most important parasitic diseases worldwide, with a significant socio-economic impact [1]. The disease affects people living in 74 endemic countries, with approximately 120 million individuals being symptomatically infected and 20 million being severely affected [2], [3]. Moreover, schistosomiasis represents an increasing problem in non-endemic areas, due to the growing number of immigrants and tourists [4], [5]. As schistosomiasis control programs are chiefly based on treatment of infected populations, adequate case-finding is important for the effective consecution of the control programs. Herein, diagnosis plays a crucial role in the monitoring of early infection and therapy efficacy. However, the currently available diagnostic assays are not ideal, since the examination of eggs in stools, such as Kato-Katz assay, and detection of circulating antigens lack sensitivity due to low disease prevalence, post-treatment situations where chemotherapeutic agents mask the presence of existing disease and methodologically, antibody detection lacks specificity [6]. In addition, current ELISA methods cannot be used to evaluate the efficacy of chemical treatment as IgG antibody levels remained elevated despite the fact that the disease was cured, indicating that there may be false positive results [7]. In the last several years, several research groups have developed specific and sensitive PCR-based methods for detecting S. japonicum, S. mansoni and S. haematobium DNA from humans and the intermediate hosts [7]–[15]. These PCR assays have proven useful alternatives for the accurate diagnosis of human schistosomiasis. Numerous factors influence the sensitivity and specificity of PCR assays for the diagnosis of schistosomiasis, in particular, the target sequences selected for PCR amplification. Several research groups have developed PCR methods using various target sequences and these PCR assays showed different sensitivities for detection of Schistosoma genomic DNA. An 121-bp rDNA sequence was the major target sequence for detecting S. mansoni [8], [9], [11], [13]. Mitochondrial NADH I gene (nad1) has been used as genetic marker for detection of S. mansoni and S. japonicum DNA [15]–[17]. SjR2, a new RTE-like, non-long terminal repeat retrotransposon from S. japonicum, is firstly described by Laha et al. [18], and its 230-bp sequence was first used as target sequence in PCR and LAMP assays for detecting S. japonicum DNA [7], [12]. Recently, 29 retrotransposons were identified in the genome sequence of S. japonicum [19], including known Gulliver, SjR1, SjR2 and Sj-pido elements as well as 25 novel elements. In the present study, primers targeting these 25 novel retrotransposons were designed and used in PCR assays for detecting S. japonicum DNA, and their sensitivities and specificities were compared.
This study was carried out in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. The protocol was approved by the Committ-ee on the Ethics of Animal Experiments of the Soochow University (Permit Number: 2007–13). All surgeries were performed under sodium pentobarbital anesthesia, and all efforts were made to mini-mize suffering of animals. Serum samples of healthy individuals were obtained from Suzhou, Jiangsu Province, China. The protocol was approved by the Scientific Committee of the Soochow University. Serum samples of patients were obtained from the endemic area of Hunan Province, China. Written informed consent was obtained from all donors. Ethical clearance for the project was obtained from the Scientific Committee at the Hunan Institute of Parasitic Diseases, which is responsible for schistosomiasis control within the Hunan Province where targeted villages are located. Schistosome-infected snails were obtained from Jiangsu Institute of Parasitic Diseases, China. The institute provided live Oncomelania hupensis snails exposed to the Chinese strains of S. japonicum. All living snails were putting into a tray which was filled with 4/5 volume of water and exposed to a light source to induce shedding of live S. japonicum cercariae. Clonochis sinensis adult worms were provided by Prof. Kuiyang Zheng, Xuzhou Medical College, China, Trichinella spiralis adult worms were provided by Prof. Zhongquan Wang, Medical College of Zhengzhou University, China, and S. mansoni adult worms were obtained from Dr. Donato Cioli, Institute of Cell Biology, Monterotondo, Italy. Six female New Zealand rabbits, weighing 1. 8–2. 2 kg, were randomly divided into three groups of two rabbits each. Blood collected before infection from all rabbits was served as negative control. Group I and Group II were percutaneously infected with 200 mixed sexual cercariae of S. japonicum (light infection), and Group III was percutaneously infected with 500 mixed sexual cercariae of S. japonicum (medium infection). On the seventh week, the eggs were detected in the feces. The EPG was 96 (Group I), 92 (Group II) and 397 (Group III), respectively. In Group I, between 1 and 7 weeks post-infection, blood were collected weekly. On the seventh week, the rabbits were anaesthetized with sodium pentobarbital and the adult worms and liver samples were collected. Between 1 and 24 weeks post-infection, blood was collected from rabbits in Group II. Serum of each rabbit was separated by centrifugation (1500 g for 15 min) and kept at −80°C until use. The rabbits infected with S. japonicum in Group III were treated with two doses of 150 mg/kg praziquantel on the seventh and eighth week post-infection. In this group, blood was collected weekly between 1 and 30 weeks post-infection (23 weeks post-treatment) and all of the rabbits were sacrificed on the last week, and no adult worms were found in portal system. Serum of each rabbits was separated by centrifugation (1500 g for 15 min) and kept at −80°C until use. Forty-three S. japonicum-positive serum samples of patients with positive Kato-Katz stool examination results were obtained from the endemic areas in Hunan Province, China, and the egg counts of these patients ranged from 8 to 1160 eggs per gram (EPG). 51 serum samples from healthy individuals were collected from healthy donors at the Center of Health Examination, the First Affiliated Hospital of Soochow University, Suzhou, Jiangsu Province, China, and were used as negative control to evaluate the specificity of the PCR assay. DNA from all of the collected samples was extracted using the method described by Steiner et al [20]. as modified by Xia et al. [7]. Briefly, 0. 2 g liver homogenate samples was mixed with 200 µl distilled water, grinded and dissolved in 200 µl Triton-X-100 (10 mMol/L Tris-HCL; 0. 45% Triton-X-100; 0. 45% NP-40; 300 µg/mL protease K, pH 7. 4) and were incubated at 60°C for 2 h with vigorous agitation for 10 min on a shaker, and then centrifuged at 12,000 g for 10 min. 300 µl supernatant of the liver homogenate was collected in a separate tube and incubated at 100°C for 10 min, centrifuged at 12,000 g for 10 min and the supernatant was collected. DNA from adult worms was also extracted using the method modified by Xia et al [7]. Five adult worms of each S. japonicum, S. mansoni, C. sinensis and T. spiralis were homogenized in 300 µl physiological saline and then digested with equal volume of extraction buffer containing 3 mg/ml proteinase K, 0. 1 mol/L Tris-HCl, pH 8. 5,0. 05 mol/L EDTA, and 1% SDS. Then the mixture was incubated at 60°C for 1 h. 200 µl serum of infected rabbits or humans were diluted in 400 µl serum extraction buffer containing 150 mol/L NaCl, 10 mol/L EDTA, 10 mol/L Tris-HCl, pH 7. 6,2% SDS, 5 µg/ml salmon sperm DNA, 4 µg 25 mg/ml proteinase K, and were incubated at 37°C overnight. All of the extraction mixtures (liver homogenate, adult samples and serum) were extracted twice with phenol, chloroform, isoamyl alcohol (25∶24∶1), once with chloroform, isoamyl alcohol (24∶1), and then precipitated with dehydrated alcohol and 3 M sodium acetate. The supernatant was discarded and the pellet was washed twice with 1 mL of 75% ethanol, finally the pellet was left to dry at 37°C for 30 min and then re-suspended with 100 µl of TE (10 mmol/L Tris-HCl, 1 mmol/L EDTA, pH 8. 0) for DNA of liver homogenates and adult worms, and 20 µl of TE (10 mM Tris-HCl, 1 mM EDTA, pH 8. 0) for DNA of serum samples, respectively. The DNA extracted from adult worms, liver and serum were used as the template. Primers were designed targeting the 25 novel retrotransposons repeat DNA sequences of S. japonicum (Table S1). PCR reaction (25 µl) contained 2. 5 µl of buffer, 1. 5 µl of 25 mmol/L MgCl2,2 µl of 2. 5 mmol/L dNTP, 0. 5 µl of each 20 pmol/L primer, 0. 4 µl of 5 U/L Taq polymerase (Takara) and 4 µl of template. The conditions for PCR were as follows (with the exception of SjCHGCS4 and SjCHGCS6 using 2-step PCR: 94°C for 4 min, followed by 30 cycles of 94°C for 30 s, 68°C for 60 s, and a final extension of 72°C for 7 min): 94°C for 3 min, followed by 35 cycles of 94°C for 60 s, Tm/ (list in Table S1) for 60 s, 72°C for 60 s and a final extension of 72°C for 7 min. The PCR was performed using GeneAmp PCR System (Eppendorf, Hamburg, Germany). Finally, a 5 µl aliquot of the PCR product was run on a 2. 0% agarose gel along with DNA ladder marker (Takara, Dalian) in TBE buffer containing 0. 5 µg/ml ethidium bromide, and the bands were visualized under UV light on a transilluminator. The nested-PCR reaction system was similar to the normal PCR as described above. The first-round amplification was carried out in the same manner except that the degenerate temperature was 55°C. Primers were designed on the basis of the SjCHGCS19 retrotransposons repreat DNA sequence of S. japonicum, the primers employed were P3 (5′-CCAAATCGCAACACTACGC-3′ (forward) and P4 (5′-ATCGGATTCTCCTTGTTCAT-3′) (reverse). DNA samples extracted from serum of rabbits and humans were used as the template. The expected length of the amplification product was 607 bp. Second-round amplification (nested-PCR) was carried out in the same manner as the first-round except that the DNA sample was a 1∶10000 dilution of the first-round PCR product and the degenerate temperature was 65°C. The sequences of these primers are listed in Table S1. The expected length of the amplification product was 303 bp. The first-round PCR product targeting SjCHGCS19 was cloned into plasmid by means of a pMD20-T II cloning reagent Kit (Tiangen, Beijing, China). Plasmid purification was done with a TIAN pure Mini Plasmid Kit (Tiangen, Beijing, China). Plasmids were quantified by spectrophotometry. Sequencing of the cloned amplification product confirmed that it was identical to part of the S. japonicum retrotransposon SjCHGCS19. The standard plasmid was tested in 10-dilution series by nested-PCR.
Primers were designed from the 25 novel retrotransposons of S. japonicum and a series of diluted genomic DNA of S. japonicum adults were used as the template. The target fragments were amplified by nested-PCR assay and the minimum amounts detectable were different among 25 new retrotransposons and SjR2. In addition to the 230-bp fragment from SjR2, a new 303-bp fragment from non-LTR retrotransposon (SjCHGCS19) displayed high sensitivity and specificity. Bioinformatic analysis showed that both the SjCHGCS19 (303-bp sequence) and SjR2 (230-bp sequence) have higher genome proportions (4. 09% and 4. 43%), higher repetitive complete copies (793 and 400) and partial copies (17,373 and 23,755), and higher EST numbers (39 and 79) than the other retrotranposons in the genome of S. japonicum (Table 1). Furthermore, to determine the limit of the 303-bp DNA fragment for detecting S. japonicum DNA, 10-fold serial dilutions of the standard plasmid clones of SjCHGCS19 and SjR2 were amplified by nested-PCR assay. The minimum detectable amount of standard plasmid from SjCHGCS19 was only 2. 02 copies, whereas the 230-bp fragment from SjR2 was 10. 2 copies. In addition, the 303-bp DNA fragment was amplified by nested-PCR assay from male and female adults of S. japonicum, from liver homogenate and sera of infected rabbits (Figure 1). Interestingly, the target DNA was amplifiable from both S. japonicum and S. mansoni, but no cross-reaction was detected with DNA samples representing C. sinensis and T. spirals (Figure 2). To clarify the potential diagnostic value of the 303-bp DNA fragment for early infection and chemotherapy evaluation in the rabbit model, a series of rabbit sera in Group II (light infection) and Group III (praziquantel treatment) were examined by nested-PCR assay. As shown in Figure 3, the 303-bp DNA fragment was amplified at 3 d sera post-infection in Group II and could be amplified until 24 weeks post-infection by nested-PCR assay. The 303-bp DNA fragment was amplifiable between 1 and 24 weeks post-infection and it became negative from 25 weeks post-infection (17 weeks after praziquantel treatment) in Group III (Figure 4). To illuminate the clinical utility of the 303-bp DNA from SjCHGCS19 as the target sequence for the diagnosis of human schistomomiasis, 43 human serum samples from patients with S. japonicum infection confirmed by stool examination and 51 serum samples from normal healthy individuals were examined. Of the 43 patient serum samples, 42 (97. 67%) were positive by the nested-PCR assay. Of the 51 serum samples from healthy individuals, 49 (96. 07%) were detected as negative by the nested-PCR assay. the percentages of sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) for nested-PCR assay were 97. 67%, 96. 07%, 97. 67% and 96. 07%%, respectively. For patients with heavy infection with S. japonicum (EPGs≥400) and medium infection (100≤EPG<400), the percentage of sensitivity was 100%. For patients with light infection (EPG<100), the percentage of sensitivity was 96. 87% (Table 2, Figure S1).
Because schistosomiasis control programs are chiefly based on treatment of infected populations, adequate case-finding is important for the effective consecution of the control programs. Herein, accurate diagnosis plays a crucial role in the monitoring of early infection and therapy evaluation. Nucleic acid-based diagnosis has been used in the clinical testing of a wide variety of pathogenic infections, such as human immunodeficiency virus [21], Mycobacterium tuberculosis [22], Plasmodium falciparum [23], Trypanosoma cruzi [24] and Leishmania braziliensis [25]. In parasitic diseases such as schistosomiasis, it was possible to detect cell-free parasite DNA circulating in plasma, and this could be used to diagnose schistosomiasis. Importantly, nucleic acid-based diagnostic methods have the same diagnostic value as parasitological diagnostic methods. Recently, several PCR-based methods for detecting Schistosoma DNA from various samples have been developed [7]–[15]. Numerous factors may influence the sensitivity and specificity of PCR assays for the diagnosis of schistosomiasis, in particular the target sequences selected. A 121-bp tandem repeat rDNA sequence was the major target sequence for detecting S. mansoni DNA in mouse serum samples and in human plasma by PCR [11], [13]. Although a real-time PCR assay using nad1 as target sequence for detecting S. japonicum DNA has proven highly sensitive, even for samples containing less than 10 EPG, it was negative for examining serum and urine samples from the infected pigs [14]. While previous efforts of PCR diagnosis of S. mansoni and S. japonicum infections have been relied on mitochondrial or rDNA sequences, we have focused on the sequence of the highly repetitive retrotransposon SjR2 in S. japonicum. In our previous studies, the 230-bp sequence from the highly repetitive retrotransposon SjR2 of S. japonicum was used as target sequence and it showed high sensitivity and specificity in detecting S. japonicum DNA, and the 230-bp fragment was amplified with DNA equivalent of 1. 1 egg from feces [7]. In particular, the 230-bp sequence was able to be amplified from the sera of the rabbit model at 1 week post-infection, which is one week earlier than that of the 121-bp sequence in the mouse-S. mansoni model [7], [11], [12]. SjR2 was 3. 9 kb in length and was constituted of a single open reading frame encoding a polyprotein with apurinic/apyrimidinic endonuclease and reverse transcriptase domains [18]. Phylogenetic analyses based on conserved domains of reverse transcriptase or endonuclease revealed that SjR2 belonged to the RTE clade of non-long terminal repeat retrotransposons and SR2 elements are members of a non-LTR retrotransposons typified by the RTE-1 non-LTR retrotransposon of Caenorhabditis elegans[26]. According to the phylogenetic tree, both SjCHGCS19 and SjR2 belonged to RTE clade of non-LTR retrontransposon (Figure S2). SjCHGCS19 showed the closest relationship with Perere-3 of S. mansoni, while SjR2 located in the same branch with SR2. At amino acids level, SjCHGCS19 showed 74% identity with Perere-3, while only 30% identity with SR2 and SjR2. Between SR2 and SjR2, the polyprotein showed 55% identity. Furthermore, hybridization analyses indicated that 10,000 copies of SjR2 were dispersed throughout the S. japonicum genome, accounting for up to 14% of the nuclear genome [18]. In our previous study, 29 retrotransposons were identified, including the known Gulliver, SjR1, SjR2 and Sj-pido elements as well as 25 novel elements, together constituting 19. 8% of the genome. Of the 25 novel retrotransposons, 18 were LTR forms, four were non-LTR forms and three were Penelope-like elements—enigmatic retroelements that retain introns. Each type of retrotransposons was represented by 1 to 793 intact copies or hundreds to thousands of partial copies [19]. The non-LTR retrotransposons such as SjR1, SjR2, Sj-pido, SjCHGCS19, SjCHGCS20, SjCHGCS21 and SjCHGCS22 have significantly higher copy numbers, constituting 12. 6% of the genome. We then wanted to determine whether there are any associations between the numbers of copies of target sequences and the sensitivity of the PCR detection. In the present study, primers were designed to amplify the 25 novel S. japonicum retrotransposons. The target fragments were amplified by nested-PCR assay for comparing sensitivity and specificity of detecting S. japonicum DNA of the target sequences. The results showed that a new 303-bp fragment from the highly repetitive retrotransposon, SjCHGCS19, had high sensitivity for detecting S. japonicum DNA in addition to the 230-bp fragment from SjR2. Importantly, bioinformatic analysis of 26 S. japonicum retrotransposons showed that both the SjCHGCS19 (303-bp sequence) and SjR2 (230-bp sequence) have higher genome proportions, repetitive complete copies and partial copies, and active ESTs than the others in the chromosome genome (Table 1). The minimum amount of the standard plasmid detectable using nested-PCR assay was 2. 02 copies per reaction using the 303-bp target sequence from SjCHGCS19, whereas 10. 2 copies per reaction of S. japonicum DNA was detected targeting the 230-bp sequence from SjR2 by the nested-PCR assay. This indicated that the 303-bp fragment from SjCHGCS19, as target sequence for detecting S. japonicum DNA, was more sensitive than the 230-bp fragment from SjR2 previously identified. Additionally, as shown in Figure 1, we evaluated the specificity of the 303-bp target sequence. The expected product was amplified by the nested-PCR assay from S. japonicum male and female adults, from liver homogenate and from sera of infected rabbits. We did not find any false positive bands in non-infected rabbit sera. Interestingly, the target DNA was amplified from both S. japonicum and S. mansoni, but no cross-reaction was detected in DNA samples representing C. sinensis and T. spirals (Figure 2). Furthermore, in the present study, the 303-bp sequence was amplified in rabbit sera at 3 d day post-infection, which is 4 d earlier than that of the 230-bp sequence in the rabbit-S. japonicum model by nested-PCR (Figure 3). The results showed that higher sensitivity is achieved using the 303-bp sequence for the detection of S. japonicum DNA in serum samples. In addition to the obvious advantage of an early diagnosis, our findings showed that the 303-bp target sequence might be valuable for the evaluation of chemotherapy efficacy, and it became negative at 25th week post-infection (17 weeks after praziquantel treatment) by nested-PCR, and was 7 weeks longer than detection by using the 230-bp sequence from SjR2, indicating a higher sensitivity of the 303-bp sequence than that of the 230-bp sequence (Figure 4). The effectiveness of the 303-bp target sequence was validated by examining serum samples from patients infected with S. japonicum (Figure S1). The findings showed that the sensitivity was 97. 67%, and the specificity was 96. 07% (Table 2). In particular, the PPV was 100% in patients with heavy infection (EPGs≧400) and medium infection (100≦EPG<400), indicating its high sensitivity. For patients with light infection (EPG<100), the percentage of sensitivity was 96. 87%. Only 1 serum sample was detected as false negative, and this sample was from a patient with particularly low infection (EPG = 56). This could be due to little S. japonicum DNA in human blood circulation and/or DNA loss during template extraction procedures. There appeared to be a correlation between PCR results and the EPGs of the patients (Table 2). However, in this study, the examined number of patient serum samples was small. It is imperative to carry out further investigation using a large number of patient serum samples. Since the prevalence of human infection with S. japonicum has been decreasing year by year [27], the rapid and reliable diagnosis of Schistosoma infection is central to the control as well as to the environmental monitoring and disease surveillance, especially for evaluation of treatment efficacy. DNA amplification assays provide alternative approaches for sensitive and specific diagnosis of Schistosoma infection, provided that reliable genetic markers are employed in the tests. Our results demonstrated that this new 303-bp sequence from non-LTR retrotransposon (SjCHGCS19) had high sensitivity and specificity for the detection of S. japonicum DNA, which may provide the new target sequence useful for the early diagnosis and for the evaluation of chemotherapy efficacy of schistosomiasis. More importantly, although many factors may affect the sensitivity and specificity in the target sequence selection, such as the length of target sequence, the specificity of the conserved sequence, the present study highlighted the key factors based on bioinformatic analysis for selecting a new perspective sensitive target sequence from genome sequences, which provides new insights into selecting suitable target sequence which may play a key role for the sensitive and specific detection of Schistosoma DNA. These findings would provide basis for establishing powerful molecular diagnostic techniques that can be used in clinical settings and as laboratory tools for surveillance and for environmental monitoring to support schistosomiasis control programs. | Schistosomiasis remains a serious parasitic disease worldwide. Adequate case-finding is important and crucial for the effective control programs as schistosomiasis control programs are chiefly based on treatment of infected populations. However, the currently available diagnostic assays are not ideal, and DNA detection can provide useful alternative approaches for the sensitive and specific diagnosis of schistosomiasis, provided that reliable genetic markers are employed in the tests. However, different target sequences used for DNA detection of samples showed different sensitivity. In previous studies, we identified a 230-bp sequence of high sensitivity and specificity from the retrotransposon SjR2 of Schistosoma japonicum. Here, new primers based on 25 novel retrotransposons were designed and their sensitivities and specificities for detecting S. japonicum DNA were compared. Of these, a new 303-bp DNA sequence from the non-LTR retrotransposon (SjCHGCS19) had high sensitivity and specificity in detecting S. japonicum DNA. More importantly, we also found that both the SjCHGCS19 (303-bp sequence) and SjR2 (230-bp sequence) have high repetitive copies, higher genome proportions and active ESTs in the genome of S. japonicum. Our findings provide new insights into selecting suitable target sequences which may play a key role for the sensitive and specific detection of S. japonicum DNA. | Abstract
Introduction
Materials and Methods
Results
Discussion | medicine
biochemistry
infectious diseases
nucleic acids
neglected tropical diseases
dna
biology | 2012 | Sensitive and Specific Target Sequences Selected from Retrotransposons of Schistosoma japonicum for the Diagnosis of Schistosomiasis | 6,875 | 342 |
The C. elegans germline is pluripotent and mitotic, similar to self-renewing mammalian tissues. Apoptosis is triggered as part of the normal oogenesis program, and is increased in response to various stresses. Here, we examined the effect of endoplasmic reticulum (ER) stress on apoptosis in the C. elegans germline. We demonstrate that pharmacological or genetic induction of ER stress enhances germline apoptosis. This process is mediated by the ER stress response sensor IRE-1, but is independent of its canonical downstream target XBP-1. We further demonstrate that ire-1-dependent apoptosis in the germline requires both CEP-1/p53 and the same canonical apoptotic genes as DNA damage-induced germline apoptosis. Strikingly, we find that activation of ire-1, specifically in the ASI neurons, but not in germ cells, is sufficient to induce apoptosis in the germline. This implies that ER stress related germline apoptosis can be determined at the organism level, and is a result of active IRE-1 signaling in neurons. Altogether, our findings uncover ire-1 as a novel cell non-autonomous regulator of germ cell apoptosis, linking ER homeostasis in sensory neurons and germ cell fate.
Apoptosis, also known as programed cell death (PCD), is a highly conserved fundamental cellular process that provides a self-elimination mechanism for the removal of unwanted cells. PCD is critical for organ development, tissue remodeling, cellular homeostasis and elimination of abnormal and damaged cells [1], [2]. The apoptotic machinery that actually executes cell death is intrinsic to all cells and can be activated in response to extracellular or intracellular cues. These are thought to be mediated by cell death receptors or by cytotoxic stress respectively [3]. In C. elegans, 131 somatic cells invariably undergo apoptosis during hermaphrodite development [4], [5]. In contrast, in the adult C. elegans, only germ cells undergo apoptotic cell death. These cell deaths can be either physiological or stress-induced [6], [7]. So far, stress-induced germ cell apoptosis has been associated with DNA damage, pathogens, oxidative stress, osmotic stress, heat shock and starvation [7]–[9]. These apoptotic events are restricted to germ cells at the pachytene stage which are located in the loop region of the gonad [6], where oogenesis transition normally occurs [10]. The physiological germ cell apoptosis pathway acts during oogenesis and is thought to act either as a part of a quality control process, preferentially removing unfit germ cells from the gonad, or as a resource re-allocation factor important for maintaining oocyte quality or simply as a gonad homeostatic pathway removing excess germ cells [6], [11], [12]. Both somatic and germ cell apoptosis rely on the highly conserved core apoptotic machinery comprised of the Caspase-3 homolog ced-3, the Apaf-1 homolog ced-4 and the anti-apoptotic Bcl-2 homolog ced-9 [6], [13]–[16]. All germ cell apoptosis, physiological and stress-induced, relies on the core apoptotic machinery [7]–[9]. However, different upstream genes activate the core apoptotic machinery in the germline in response to different stresses. For example, DNA damage-induced germ cell apoptosis involves the proteins EGL-1, CED-13, and the DNA damage response protein p53 homolog CEP-1 [9], [17]–[19]. In contrast, oxidative, osmotic, heat shock and starvation stresses induce germ cell apoptosis through a CEP-1 and EGL-1 independent pathway and rely on the MEK-1 and SEK-1 MAPKs instead [7]. The endoplasmic reticulum (ER) fulfills many essential cellular functions, including a role in the secretory pathway, in lipid metabolism and in calcium sequestration. Accordingly, ER homeostasis is essential for proper cellular function [20]. A specialized, conserved cellular stress response, called the unfolded protein response (UPR), is in charge of detecting ER stress and adjusting the capacity of the ER to restore ER homeostasis. In C. elegans, as in humans, three proteins located at the ER membrane sense ER stress and activate the UPR: the ribonuclease inositol-requiring protein-1 (IRE-1), the PERK kinase homolog PEK-1 and the activating transcription factor-6 (ATF-6) [21]. Of the three, IRE-1 is the major and most highly conserved ER stress sensor. In response to ER stress, IRE-1 activates the ER stress-related transcription factor XBP-1, which induces the transcription of genes that help restore ER homeostasis [22]–[24]. Accordingly, ire-1 and xbp-1 deficiencies perturb ER homeostasis [25], [26]. Although IRE-1 typically protects cells, upon excessive and prolonged ER stress, IRE-1 can also trigger cell death, usually in the form of apoptosis [27], [28]. For example, IRE-1 can lead to activation of the cell death machinery via JNK and caspase activation [29], [30] or by mediating decay of critical ER-localized mRNAs through the RIDD pathway, tipping the balance in favor of apoptosis [31]. These functions of IRE-1 are independent of XBP-1 [29]–[33]. Highly proliferating cells with a high protein and lipid biosynthetic load are thought to rely on ER function to a greater extent than other cells. This together with the general sensitivity of the germline to cellular stresses prompted us to investigate the effects of ER stress on germ cell fate. Strikingly, we discovered that ER stress does not simply kill the germ cells by not meeting their biosynthetic demands. Instead, we found that ER stress initiates a signaling cascade in neurons that regulates germ cell survival non-autonomously. Thus, our findings reveal that germ cell sensitivity to ER stress conditions can be regulated at an organismal level and can be uncoupled from germ cell stress.
To investigate whether ER stress induces apoptosis in the C. elegans germline, we first assessed the number of apoptotic corpses in the gonads of animals treated with tunicamycin, a chemical ER stress inducer which blocks N-linked glycosylation. Apoptotic corpses in the gonad were identified by staining with the vital dye SYTO12 and by their discrete cellularization within the germline syncytium. We found that tunicamycin treatment increased the number of apoptotic germ cells present in wild-type gonads by approximately 3 fold compared to control DMSO treatment from day-1 to day-3 of adulthood (P<0. 001, Figure 1A–B). If indeed the increased number of germline corpses in tunicamycin-treated animals is a consequence of ER stress, then additional manipulations that disrupt ER homeostasis should also increase germ cell apoptosis. tfg-1 encodes a protein that directly interacts with SEC-16 to control COPII subunit accumulation at ER exit sites and is required for the vesicular export of cargo from the ER [34]. We hypothesized that ER homeostasis would be disrupted in tfg-1-deficient animals. To examine the effect of tfg-1 deficiency on ER homeostasis, we assessed the effect of tfg-1 RNAi treatment on the levels of the ER stress response reporter Phsp-4: : gfp [22]. tfg-1 RNAi efficacy was confirmed by the reduction in the animals' body size compared to control RNAi treated animals [35]. We found that treatment with tfg-1 RNAi specifically activated the ER stress response, as it increased the level of the ER stress response reporter without increasing the expression of oxidative stress response, heat shock response or mitochondrial stress response reporters (Figure S1). In terms of germ cell apoptosis, we observed that tfg-1 RNAi consistently increased the number of apoptotic germ cells in the gonad by approximately 4 fold from day-1 to day-3 of adulthood compared to wild-type animals (P<0. 001, Figure 2A, B). A similar 4 fold increase in germ cell apoptosis was observed by scoring germ cell engulfment by neighboring cells that expressed GFP-labeled CED-1, a transmembrane receptor that mediates cell corpse engulfment in C. elegans [36] (Figure 2C). tfg-1 RNAi treatment did not increase the number of SYTO12-labeled cells in the gonads of apoptosis-defective ced-3 (n1286) mutants, confirming that the dye specifically labels apoptotic cells (Figure S2). Together, these results indicate that conditions that disrupt ER homeostasis, including tunicamycin treatment or blocking secretory traffic from the ER, increase apoptosis frequency in the gonad compared to non-stressed animals. We next asked which apoptotic machinery is implicated in ER stress-induced germ cell apoptosis. To this end, we examined mutants deficient in core-apoptotic genes as well as mutants deficient in genes specifically implicated in germ cell apoptosis. This array of apoptosis-related mutants was treated with control or tfg-1 RNAi, and germ cell apoptosis was scored by SYTO12 labeling. As expected, we found that the core apoptosis machinery genes ced-3 and ced-4 [6], [13], [15], [16] were required for germ cell apoptosis in response to ER stress (Figure 3A). Importantly, the cep-1, egl-1 and ced-13 genes, previously implicated in DNA damage-induced apoptosis [9], [17]–[19], were also found to be completely essential for germ cell apoptosis in response to tfg-1 RNAi treatment (Figure 3A). Accordingly, the levels of a CEP-1: : GFP translational fusion transgene driven by the cep-1 promoter were increased within the germ cells of tfg-1 RNAi-treated animals (P<0. 001, Figure 3B). In contrast, pmk-1 and sek-1, previously implicated in oxidative stress-induced and pathogen-induced germ cell apoptosis respectively [7], [8], were dispensable for germ cell apoptosis in response to tfg-1 RNAi treatment (Figure 3C). Thus, the genetic analysis clearly implicated the apoptotic machinery that mediates DNA damage-induced germ cell apoptosis in ER stress-induced germ cell apoptosis as well. The strong changes in CEP-1 levels observed in tfg-1 RNAi-treated animals suggest that ER stress controls CEP-1 activation within the germ cells. One possible explanation for the involvement of genes implicated in DNA damage-induced germ cell apoptosis is that ER stress indirectly damages DNA, which in turn leads to CEP-1 activation and DNA damage-induced germ cell apoptosis. However, whereas a previous study implicated the intestinal kri-1 gene in non-autonomous regulation of ionizing-radiation induced germ cell apoptosis [37], we found that tfg-1 RNAi treatment efficiently induced germ cell apoptosis in kri-1-deficient animals (Figure 3C). This genetic uncoupling between the requirements for ionizing-radiation induced germ cell apoptosis and ER stress-induced germ cell apoptosis suggest that ER stress does not simply induce DNA damage which in turn leads to germ cell apoptosis. To further substantiate this conclusion, we examined directly whether the DNA damage response is activated in the germ cells of ER stressed- animals treated with tfg-1 RNAi. To this end, we followed the nuclear aggregation of HUS-1: : GFP, which encodes a DNA damage checkpoint protein that relocalizes to distinct nuclear foci upon induction of DNA damage [38]. Whereas nuclear HUS-1: : GFP aggregates were clearly observed in the germ cells of DNA-damaged rad-51 RNAi treated animals (P<0. 001 compared to control RNAi), HUS-1: : GFP aggregates were not detected in tfg-1 RNAi treated animals (P = 0. 78 compared to control RNAi, Figure 3D). Thus, ER stress activates CEP-1/p53 to induce germ cell apoptosis without generating DNA damage. We next asked whether any of the canonical ER stress sensing genes was implicated in ER stress-induced germ cell apoptosis. To this end, we examined mutants deficient in the ER stress-response sensor genes that comprise the UPR: ire-1, pek-1 or atf-6. This array of mutants was treated with DMSO or with tunicamycin and germ cell apoptosis was scored by SYTO12 labeling. We found that similarly to its effect in wild-type animals, tunicamycin treatment increased the number of germline corpses in atf-6 and pek-1-deficient animals by approximately 3 fold from day-1 to day-3 of adulthood (P<0. 001, Figure 1A, C). In contrast, ER stress induced by tunicamycin treatment failed to increase germ cell apoptosis in ire-1 mutants (Figure 1A, B). Therefore, the insensitivity of the germline to tunicamycin is unique to ire-1-deficient animals and not seen in animals deficient in other UPR sensors. ire-1 mutants are abnormal in terms of their gonad anatomy and their reproductive capacity: ire-1 mutants have approximately 2 fold less progeny and 2 fold less mitotic germ cells within their proliferative zones compared to ire-1 (+) wild-type animals (P<0. 001, Figures S3A, D). Thus, we wondered whether these abnormalities affected the ability of their germ cells to undergo apoptosis in general or whether they were specifically defective in their ability to undergo apoptosis in response to ER stress. First we assessed germline apoptosis in ire-1 (−) mutants under normal growth conditions, from day-0 (L4) to day-3 of adulthood. At all timepoints, we detected approximately half the amount of germline corpses in ire-1 (−) gonads compared to ire-1 (+) wild-type gonads, as assessed by SYTO12 labeling and by the CED-1: : GFP engulfment marker (Figure 2A–C). The low levels of germ cell apoptosis persisted in ire-1 mutants treated with vps-18 RNAi (Figure 2D), a treatment that impairs germ cell corpse clearance [39]. However, normalization of the number of apoptotic corpses to the number of mitotic germ cells resulted in comparable levels of germ cell corpses in ire-1 mutants and in non-stressed wild-type animals (P = 0. 12, Figure S3C). This indicates that in spite of their reproductive abnormalities, physiological germ cell apoptosis in ire-1 (−) and ire-1 (+) animals is comparable. Next, we assessed stress-induced germline apoptosis in ire-1 mutants. We found that although manipulations that disrupt ER homeostasis fail to increase germ cell apoptosis in ire-1 mutants (Figure 1B, 1D–E), DNA damage and oxidative stress conditions did increase germ cell apoptosis in ire-1 mutants (P<0. 001 compared to non-stressed ire-1 mutants, Figure 1D–E). This resulted in a similar level of germline apoptosis as in stressed wild-type animals upon normalization of the number of apoptotic corpses to the number of mitotic germ cells (P = 0. 44 for DNA damage and P = 0. 42 for oxidative stress, Figure 1D–E). Thus, in spite of the reproductive abnormalities of ire-1 mutants, the germ cells of these mutants undergo stress-induced apoptosis similarly to wild-type animals, however not in response to ER stress. The inability of ire-1 mutants to increase germline apoptosis specifically in response to perturbations in ER homeostasis suggests that IRE-1 may be a critical mediator of ER stress-induced germ cell apoptosis. The most established mode of action of IRE-1 under ER stress conditions is via the activation of the UPR-related transcription factor XBP-1 [22]–[24]. Therefore, if IRE-1 enabled ER stress-induced germ cell apoptosis via its downstream target xbp-1, then the number of germline corpses detected in xbp-1 (−) mutants should remain low under ER stress conditions, similarly to ire-1 (−) mutants. In order to test this, we first examined germ cell apoptosis in xbp-1 (tm2457) null mutants. Surprisingly, in contrast to ire-1 (−) mutants, we consistently detected increased germ cell apoptosis in xbp-1 (−) gonads compared to wild-type gonads under normal growth conditions. A 2. 5 fold increase in the number of germline corpses in xbp-1 (−) mutants was detected by SYTO12 labeling of gonads from day-1 to day-3 of adulthood compared to wild-type animals (P<0. 001, Figure 2A–B). A similar observation was apparent by using the CED-1: : GFP engulfment marker (Figure 2C). The 2. 5 fold increase in the number of germ cell corpses was still apparent in engulfment defective vps-18 RNAi-treated animals (Figure 2D) and upon normalization to the number of mitotic germ cells located in the proliferative zone (Figure S3C). Thus, in contrast to ire-1 mutants and wild-type animals, xbp-1 mutants exhibit a high basal level of germ cell apoptosis. We hypothesized that the increase in germline apoptosis in xbp-1 mutants may be due to perturbed ER homeostasis in these animals [25], [26], If so, then it should be mediated via ire-1, similarly to other ER stress conditions that induce germ cell apoptosis. Accordingly, we found that in an ire-1 (−) background, the xbp-1 mutation did not increase germ cell apoptosis. This observation was consistent along different time points spanning from day-0 (L4) to day-3 of adulthood (Figure 2A–B). This also persisted upon normalization to the number of mitotic germ cells located in the proliferative zone (Figure S3C). Interestingly, the amount of mitotic germ cells of xbp-1; ire-1 double mutants was similar to that of xbp-1 single mutants (P = 0. 072, Figure S3A), whose germ cells were responsive to ER stress-induced apoptosis (Figure S3B–C). This indicates that the reduced amount of mitotic cells in the gonad of ire-1 mutants can be uncoupled from the inability of their germ cells to undergo ER stress-associated apoptosis. The finding that xbp-1 deficiency per se promotes ire-1-dependent ER stress-induced germ cell apoptosis suggests that xbp-1 is dispensable for increasing germ cell apoptosis in response to ER stress. Consistent with this, we found that tunicamycin treatment further increased germ cell apoptosis in xbp-1 mutants (P<0. 001, Figure 1A–B). Altogether, these results lend further support to the notion that ire-1 is a critical signaling molecule in mediating ER stress-induced germline apoptosis, whereas its' downstream canonical target xbp-1 is not. Furthermore, since ER function is compromised both in ire-1 and in xbp-1 deficient mutants [25], [26], the differential ability to induce germ cell apoptosis in these mutants suggests that germ cell apoptosis may be the result of active IRE-1 signaling, rather than simply a consequence of ER dysfunction. In mammalian cells, activation of IRE1 can cell-autonomously activate JNK via the adaptor protein TRAF. Consequently, IRE1-mediated activation of JNK initiates proapoptotic signaling, independently of XBP1 [29]. Thus, we examined whether the C. elegans homologs of TRAF and JNK proteins were required for ire-1/ER stress-induced apoptosis in C. elegans, which is also independent of xbp-1. To this end, trf-1 mutants or mutants deficient in all three C. elegans JNK homologs were treated with control or tfg-1 RNAi. We found that tfg-1 RNAi increased germ cell apoptosis independently of the trf-1 and the JNK-like genes (Figure 3E). Thus, since ER stress can effectively induce germ cell apoptosis in the absence of xbp-1, trf-1 and JNK homologs, the signaling mediated by IRE-1, in this case, must be executed by an alternative xbp-1-independent output of IRE-1. Next, we examined whether ER stress triggers programmed cell death autonomously within the germ cells, or non-autonomously from the soma. To test this, we used tfg-1 RNAi to induce ER stress specifically in the germline or in the soma. To induce ER stress primarily in the germ cells, mutants in the rrf-1 gene, encoding an RNA-directed RNA polymerase (RdRP) homolog required for most somatic RNAi but not for germline RNAi [40], were treated with tfg-1 RNAi. No increase in the amount of germline corpses was observed as a result of tfg-1 RNAi treatment in rrf-1 mutants (P = 0. 19, Figure 4A). To induce ER stress specifically in the soma, mutants in the ppw-1 gene, which is required for efficient RNAi in the germline [41], were treated with tfg-1 RNAi. This resulted in a 4. 5 fold increase in the amount of apoptotic corpses in the gonads (P<0. 001, Figure 4A). Thus, ER stress in the soma, rather than in the germ cells, is sufficient for the induction of germ cell apoptosis. Does germ cell apoptosis occur upon disruption of ER homeostasis in the entire soma or does it occur in response to ER stress in a particular part of the soma? To answer this, ER stress was induced locally in specific somatic tissues. This was achieved by treating animals expressing functional RNAi machinery only in specific tissues with tfg-1 RNAi and assessing germ cell apoptosis in these animals. We found that tfg-1 RNAi treatment did not increase germ cell apoptosis in animals which respond to RNAi only in the intestine, in the muscle, in the hypodermis, in the uterine or in the distal tip cells (P>0. 1 in each one of these strains, Figure 4B). In contrast, tfg-1 RNAi treatment increased germ cell apoptosis by approximately 7 fold in animals which respond to RNAi specifically in the neurons (P<0. 001, Figure 4A). Next, we examined whether ER stress-induced germline apoptosis is under pan-neuronal control or under the control of specific neurons. To this end, we introduced ER stress-inducing tfg-1 RNAi into animals expressing functional RNAi machinery specifically in the cholinergic, glutamatergic, GABAergic, dopaminergic or in a subset of sensory neurons. Importantly, we found that tfg-1 RNAi treatment increased germ cell apoptosis only in animals whose sensory neurons responded to RNAi (Figure 4C). Among the sensory neurons whose exposure to ER stress increased germline apoptosis were the ASI neurons, which have been previously implicated in the regulation of germ cell proliferation and maturation [42]. Hence, we examined whether ER stress in the ASI sensory neurons alone is sufficient for the induction of germ cell apoptosis in the gonad. To this end, we first assessed germline apoptosis in daf-28 (sa191) mutants, which produce a toxic insulin peptide that activates the UPR specifically in the ASI neurons [43]. We found that germ cell apoptosis in the gonads of daf-28 (sa191) mutants was increased by approximately 4 fold compared to wild-type animals (P<0. 001, Figure 4D). Importantly, germ cell apoptosis was not increased in a daf-28 (tm2308) null strain, which is deficient in daf-28 and does not produce the toxic insulin peptide which induces ER stress (P = 0. 15, Figure 4D). tfg-1 RNAi treatment of the two daf-28 mutant strains increased the number of germline corpses in daf-28 (tm2308) null strain (P<0. 001), but did not further increase germline apoptosis in the daf-28 (sa191) strain (P = 0. 09, Figure 4D). tfg-1 RNAi treatment did not alter ASI overall morphology as assessed by the expression pattern of a GFP reporter driven by an ASI-specific promoter (Figure S4A). Together, these findings suggest that expression of the toxic form of DAF-28 and tfg-1-deficiency increase germ cell apoptosis by similar means; most likely by causing ER stress and activating the UPR in the ASI neurons. We have demonstrated that in the absence of the ER stress sensor ire-1, ER stress does not increase germline apoptosis. We further demonstrated that ER stress in the ASI sensory neurons is sufficient to induce germ cell apoptosis. Thus, we next examined whether it is also sufficient to express ire-1 in the soma, and specifically in the ASI neurons, to restore germ cell apoptosis in response to ER stress. To this end, we restored ire-1 expression in the entire soma, pan-neuronally or specifically in the ASI/ASJ neurons of ire-1 (−) mutants. This was achieved using multi-copy ire-1 transgenes under ire-1, rgef-1 and daf-28 promoters respectively. Since the expression of multi-copy transgenes is normally suppressed in germ cells [44], and due to the specificity of their promoters, these transgenes restore ire-1 expression within different parts of the soma but not in the germline. We found that expression of each of these ire-1 transgenes completely restored the increase in germline apoptosis in response to treatment with tfg-1 RNAi (P<0. 001, compare white and black bars within each strain in Figure 5A). Similarly, we restored ire-1 expression in muscle cells and in the PVD and OLL neurons using multi-copy ire-1 transgenes under myo-3 and ser-2 promoters respectively. No increase in germline apoptosis in response to tfg-1 RNAi treatment was apparent in these two transgenic lines compared to control RNAi treatment (P>0. 1, Figure 5A). The fact that not all ire-1 transgenes induced apoptosis supports the notion that ire-1-induced germline apoptosis is not the result of leaky expression of the transgenes in other tissues. Altogether, this implies that not all tissues and not all neurons are involved in the regulation of this process. Next, we asked whether increasing IRE-1 levels to a greater extent may be sufficient for inducing germline apoptosis even in the absence of ER stress. To this end, we overexpressed ire-1 transgenes in various tissues or cells of ire-1 (+) wild-type animals. This was achieved by using multi-copy ire-1 transgenes under ire-1, rgef-1, daf-28 and daf-7 promoters. This is consistent with the interpretation that some activation of IRE-1 is achieved merely by its over-expression, as has been previously observed in yeast and in mammalian cells [45], [46]. We found that this artificial activation of IRE-1 in the soma, pan-neuronally or specifically in the ASI/ASJ neurons of ire-1 (+) animals was sufficient to induce high levels of germ cell apoptosis (P<0. 001, compare white bars of transgenic animals to that of wild-type animals in Figure 5B). No increase in germ-cell apoptosis was observed upon overexpression of an ire-1 transgene in muscle cells or in the AIY neurons in ire-1 (+) animals (P>0. 1, Figure 5B). These findings support the claim that the rescuing activity of the ire-1 transgenes stems from their expression in specific neurons. tfg-1 RNAi treatment of ire-1 (+) animals over-expressing the ire-1 transgenes in the soma, in the neurons and specifically in the ASI/ASJ neurons did not further increase germline apoptosis (P>0. 5, compare white and black bars within the strains, Figure 5B). This suggests that IRE-1 overexpression and tfg-1-deficiency increase germ cell apoptosis by similar means, i. e. by activating IRE-1. Taken together, our data demonstrate that activation of ire-1 specifically in the ASI neurons, either by ER stress in the ASI neurons or by IRE-1 overexpression, can non-autonomously regulate germ cell apoptosis. Furthermore, since over-expression of transgenic IRE-1 is sufficient for its artificial activation in a manner that is independent of ER stress, this further suggests that active IRE-1 signaling in the ASI neurons per se, rather than neuronal ER stress or ER dysfunction, is the cause of germ cell apoptosis.
Understanding the molecular events that regulate the life-death decision of cells is of fundamental importance in cell biology research, cell development, cancer biology and disease biology [47]. In this study, we gained new and fascinating insights into the complex coupling between ER stress in the nerve system and germ cell apoptosis. We report for the first time that germ cells undergo apoptosis in response to ER stress. We find that activation of the ER stress response gene ire-1 is required and sufficient to induce germ cell apoptosis in response to several ER stress-inducing conditions. Strikingly, we find that germ cell fate is regulated non-autonomously by ER stress and/or through IRE-1 activation specifically in the ASI neurons. This implies that ER homeostasis and UPR signaling in the germ cells themselves is not a factor in determining their fate, ruling out the possibility that these apoptotic events are part of a quality control process that removes “stress-damaged” germ cells from the gonad [6], [11], [12]. Furthermore, this assigns a central neuroendocrine role for the ASI neuron pair in coupling between stress sensing and the onset of germ cell apoptosis. This is in addition to other central physiological processes in C. elegans, such as dauer formation [48], [49] and longevity [50], [51], that are also controlled by the sensory ASI neuron pair. Interestingly, another pair of sensory neurons, the ASJ neurons, has been previously implicated in the protection of germ cells from apoptosis under hypoxic conditions [52]. Thus, depending on the stress condition, different neurons can shift germ cell fate from survival to death or vice versa. How might IRE-1 activation in the ASI neurons dictate germ cell survival or death? One possibility is that defects associated with ire-1 deficiency and/or ire-1 activation indirectly abrogate the communication between the neurons and the gonad. However, several lines of evidence undermine this hypothesis: (1) We find that ER stress-induced germ cell apoptosis proceeds normally in animals with a severely defective nervous system (Figure S4B, C). This implies that germline apoptosis does not result from a generic neuronal defect. (2) ire-1 deficiency is associated with germline abnormalities which include a significant reduction in the number of mitotic germ cells and in reduced progeny number. However, these gonad-related defects do not confer generic resistance to stress-induced apoptosis as the germ cells of ire-1 mutants do undergo apoptosis in response to a variety of stresses. Furthermore, a mutation in xbp-1, which improved the reproductive abnormalities of ire-1 mutants, did not restore responsiveness to ER stress induced germ cell apoptosis in ire-1; xbp-1 double mutants, thus uncoupling the two. (3) Whereas the comparison of germ cell apoptosis in ire-1 and wild-type animals may be confined by the basal discrepancy of their reproductive systems, this concern does not exist in the analysis of ire-1 overexpressing strains, whose gonad appears to be normal (P>0. 1 for Pire-1: : ire-1 and Pdaf-7: : ire-1 compared to wild-type animals Figure S3A, D). Similarly, this concern does not exist in the intra-strain comparisons of germline apoptosis within the ire-1 (−) strain under control and stress conditions. If ire-1 misregulation in the ASI neurons does not indirectly abrogate the communication between the neurons and the gonad, how might it dictate germ cell survival or death? IRE-1 is a dual-activity enzyme, bearing both kinase and endoribonuclease activities and a propensity to self-aggregate at the ER membrane in response to ER stress. The most characterized mode of action of IRE-1 is the activation of its downstream transcription factor XBP-1 [53]. Significantly less characterized are XBP-1 independent targets of IRE-1, that include activation of the cell death machinery via JNK/TRAF signaling and degradation of ER-localized mRNAs that encode secreted and membrane proteins in a process called RIDD [29], [30], [32], [54]–[58]. Since we find that ER stress can effectively induce germ cell apoptosis in the absence of xbp-1, trf-1 and JNK homologs, the signaling mediated by IRE-1 in this case may be executed by the RIDD pathway or via a novel, yet undescribed, xbp-1-independent output of IRE-1. We propose that activation of IRE-1 in the neurons (either as a result of ER stress or merely by its over-expression) actively regulates the production of a germ cell regulatory signal. In principle this may be a germ cell proapoptotic signal produced by the neurons upon IRE-1 activation. Alternatively, this may be a germ cell anti-apoptotic signal that is down-regulated by IRE-1 upon its activation. This ASI-regulated signal, whose identity and nature remain to be elucidated, propagates in the animal and affects the gonad where it acts upstream to the p53 homolog cep-1, activating the same apoptotic machinery in the germ cells as the DNA damage response, without inducing DNA damage in the germ cells (Figure 6). This indicates the existence of a new pathway that can activate CEP-1 independently of DNA damage upon activation of neuronal IRE-1. Interestingly, in adult animals, exposure to ER stress or activation of IRE-1 in the soma induce apoptosis only in germ cells, as we did not detect any apoptotic corpses outside of the gonad of these animals. This is in contrast to the developing embryo, where exposure to ER stress can induce apoptosis in the soma [35], [59]. We propose that as the organism completes development, its ability to respond or execute programmed cell death upon exposure to ER stress is maintained in mitotic germ cells while being selectively abrogated in the post-mitotic soma, as has been demonstrated for their ability to execute apoptosis in response to DNA-damage [60]. This resistance of the soma is important in terms of survival of the animal as it prevents cell death of somatic tissues that lack stem cell pools and regenerative capacity, while allowing cell death of immortal germline cells at times of stress. What could be the advantage in diluting the germ cell pool when neurons “feel” ER-stressed (i. e. when IRE-1 is activated naturally by ER stress or artificially by overexpression)? Recent studies demonstrate a tight inverse correlation between germ-cell proliferation and the maintenance of somatic proteostasis and longevity [61]–[65]. This inverse correlation is thought to be due to a limitation of resources shared by the germline and the soma and due to altered metabolic and cellular repair mechanisms in the soma that are enabled upon germ cell loss. Previous studies implicated the nervous system in systemic and hierarchical control of cellular stress responses elsewhere in the soma to maintain organismal homeostasis [66]–[72]. Our data further imply that neurons also have the ability to communicate with the germ cells to promote their death in response to stress in the ER. This, in turn, may orchestrate a proteostasis switch in the soma at the expense of a replenishable germ cell pool in times of stress. This adds a new layer of complexity to our understanding of how protein homeostasis is regulated and coordinated across tissues in multicellular organisms.
For single time-point experiments, the number of apoptotic germ cells was scored in day-2 animals stained with SYTO12 (Molecular Probes) as previously described [6]. For time course experiments, the number of SYTO12-labeled apoptotic corpses per gonad arm was scored in animals from day-0 (L4) to day-3 of adulthood. Where indicated, the average number of apoptotic corpses was normalized to the number of mitotic germ cells within the proliferative zone of the gonads, determined by section analysis of DAPI-stained gonads. Day-1 adult animals were placed in 200 µl of M9 (control) or 10 mM paraquat (oxidative stress) for 1. 5 h at 20°C. After the incubation period, 1 ml of M9 was added to dilute the paraquat. Animals were then transferred to eppendorfs with SYTO12 staining for 4. 5 hrs. Animals were allowed to recover on plates for 40 min. Finally, the animals were mounted and observed under the microscope to determine cell corpse numbers. Gonads of day-1 adults were dissected, fixed, and stained with DAPI as previously described [10]. Bacteria expressing dsRNA were cultured overnight in LB containing tetracycline and ampicillin. Bacteria were seeded on NGM plates containing IPTG and carbenicillin. RNAi clone identity was verified by sequencing. Eggs were placed on plates and synchronized from day-0 (L4). The efficacy of the tfg-1 RNAi was confirmed by the animals' reduced body size [35]. Animals were anaesthetized on 2% agarose pads containing 2 mM levamisol. Images were taken with a CCD digital camera using a Nikon 90i fluorescence microscope. For each trial, exposure time was calibrated to minimize the number of saturated pixels and was kept constant through the experiment. The NIS element software was used to quantify mean fluorescence intensity as measured by intensity of each pixel in the selected area. Error bars represent the standard error of the mean (SEM) of at least 3 independent experiments. P values were calculated using the unpaired Student' s t test. The following lines were used in this study: N2, CF2012: pek-1 (ok275) X, CF2988: atf-6 (ok551) X, CF2473: ire-1 (ok799) II, CF3208: xbp-1 (tm2457) III, SHK62: ire-1 (ok799) II; xbp-1 (tm2457) III, MD701: Plim-7 ced-1: : gfp V, xbp-1 (tm2482) III, CF2185: ced-3 (n1289) IV, MT2547: ced-4 (n1162) III, TJ1: cep-1 (gk138) I, MT8735: egl-1 (n1084n3082) V, FX536: ced-13 (tm536) X, WS1433: hus-1 (op241) I; unc-119 (ed3) III; opIs34, CF2052: kri-1 (ok1251) I, KU25: pmk-1 (km25) IV, AU1: sek-1 (ag1) X, CF3030: kgb-1 (um3) kgb-2 (gk361) jnk-1 (gk7) IV, NS2937: trf-1 (nr2014) III, CF2260: zcIs4[Phsp-4: : gfp] V, CL2166: Pgst-4: : gfp (dvls19) V, CF1553: muIs84 [ (pAD76) Psod-3: : gfp+rol-6], SJ4100: Phsp-6: : gfp (zcIs13) V, CL2070: Phsp-16. 2: : gfp (dvls70) V, SHK57: xbp-1 (tm2457) III; ced-3 (n1286) IV, SHK189: zcIs4[Phsp-4: : gfp] V; Pire-1: : ire-1, NL2098: rrf-1 (pk1417) I, NL2550: ppw-1 (pk2505) I, SHK185: ire-1 (ok799) II; Prgef-1: : ire-1, SHK4: Pire-1: : ire-1, SHK182: Prgef-1: : ire-1, BB22 rde-4 (ne299) III; adr-2 (gv42) III, TG12: cep-1 (lg12501) I; unc-119 (ed4) III; gtIs1 [CEP-1: : GFP+unc-119 (+) ], SHK8: Pmyo-3: : ire-1, SHK14: Pmyo-3: : ire-1; ire-1 (ok799) II, Pser-2: : ire-1; ire-1 (ok799) II, SHK15: Pdaf-28: : ire-1; ire-1 (ok799) II, SHK6: Pdaf-28: : ire-1, SHK237: Pttx-3: : ire-1; Pdaf-7: : gfp, SHK234: Pdaf-7: : ire-1; Pdaf-7: : gfp, daf-28 (tm2308) V, CF2638: daf-28 (sa191) V, VB1605: svls69[Pdaf-28: : daf-28: : gfp], SHK11: ire-1 (ok799) II; svls69[Pdaf-28: : daf-28: : gfp], SHK27: ire-1 (ok799) II; Pire-1: : ire-1; svls69[Pdaf-28: : daf-28: : gfp], SHK60: unc-13 (e51) I, SK7: unc-64 (e246) III; unc-31 (e928) IV and MT6308: eat-4 (ky5) III. The following strains were used for tissue-specific RNAi experiments: TU3401: sid-1 (pk3321) V; Punc-119: : sid-1 (neuron only RNAi), VP303: rde-1 (ne213) V; Pnhx-2: : rde-1 (intestine only RNAi), WM118: rde-1 (ne300) V; Pmyo-3: : rde-1 (muscle only RNAi), NR222: rde-1 (ne219) V; Plin-26: : rde-1 (hypodermis only RNAi), NK640: rrf-3 (pk1426) II; rde-1 (ne219) V; Pfos-1A: : rde-1 (uterine only RNAi), JK4143: rde-1 (ne219) V; Plag-2: : rde-1: : gfp (distal tip cell only RNAi). The following strains were used for neuron-specific RNAi treatments as previously described [73]: XE1581: wpSi10 II [unc-17p: : rde-1: : SL2: : sid-1+Cbr-unc-119 (+) ]; eri-1 (mg366) IV; rde-1 (ne219) V; lin-15B (n744) X - Cholinergic neuron-specific RNAi strain. XE1375: wpIs36 I [unc-47p: : mCherry]; wpSi1 II [unc-47p: : rde-1: : SL2: : sid-1+Cbr-unc-119 (+) ]; eri-1 (mg366) IV; rde-1 (ne219) V; lin-15B (n744) X - GABAergic neuron-specific RNAi strain. XE1582: wpSi11 II [eat-4p: : rde-1: : SL2: : sid-1+Cbr-unc-119 (+) ] II. ; eri-1 (mg366) IV; rde-1 (ne219) V; lin-15B (n744) X - Glutamatergic neuron-specific RNAi strain. XE1474: wpSi6 II [dat-1p: : rde-1: : SL2: : sid-1+Cbr-unc-119 (+) ] II; eri-1 (mg366) IV; rde-1 (ne219) V; lin-15B (n744) X - Dopaminergic neuron-specific RNAi strain, SHK231: sid-1 (pk3321) V; Pche-12: : sid-1 (+); rol-6 (su1006). Prgef-1: : ire-1 - ire-1 cDNA was cloned under the 3. 5 kb rgef-1 (F25B3. 3) promoter and injected at 5 ng/µl with Pmyo-3: : mCherry at 50 ng/µl. Pire-1: : ire-1 - ire-1 cDNA was cloned under the 4. 5 kb ire-1 (C41C4. 4) promoter in the L3691 vector and injected at 25 ng/µl with rol-6 at 100 ng/µl. Pttx-3: : ire-1 - was created by cloning the ire-1 cDNA into a Pttx-3 vector [74] using KpnI/SphI. Pdaf-7: : gfp and Pdaf-7: : ire-1 - daf-7 promoter fragment [49] was cloned into pPD95. 75 (gift from A. Fire, Carnegie Institute) using SphI/XbaI to create daf-7p: : gfp transcriptional fusion. The gfp fragment was then replaced by ire-1 cDNA using XmaI/AflII to make daf-7p: : ire-1. daf-7p: : ire-1 or ttx-3: : ire-1 were injected at 10 ng/µl with daf-7p: : gfp and pRF4 (rol-6) at 20 ng/µl each. ser-2prom-3: : ire-1 - was created by cloning ire-1 cDNA under ser-2prom-3 fragment [75] using XmaI/AflII. ser-2prom-3: : ire-1 was injected at 10 ng/µl with ttx-3: : mCherry at 40 ng/µl. | Cells in the C. elegans germline undergo programmed cell death as part of the normal developmental program and in response to various stresses. Here, we discovered that more germ cells undergo programmed cell death under stress conditions associated with the accumulation of misfolded proteins in the endoplasmic reticulum, a cellular organelle responsible for protein folding and trafficking. Surprisingly, we found that germ cell death is a consequence of stress in neurons rather than in the germ cells themselves. This implies that germ cell death under ER stress conditions is regulated at the organismal level and implicates signaling between tissues. | Abstract
Introduction
Results
Discussion
Materials and Methods | model organisms
cell biology
genetics
biology and life sciences
organisms
research and analysis methods | 2014 | It's All in Your Mind: Determining Germ Cell Fate by Neuronal IRE-1 in C. elegans | 12,136 | 138 |
Subcutaneous immunization delivers antigen (Ag) to local Ag-presenting cells that subsequently migrate into draining lymph nodes (LNs). There, they initiate the activation and expansion of lymphocytes specific for their cognate Ag. In mammals, the structural environment of secondary lymphoid tissues (SLTs) is considered essential for the initiation of adaptive immunity. Nevertheless, cold-blooded vertebrates can initiate potent systemic immune responses even though they lack conventional SLTs. The emergence of lymph nodes provided mammals with drastically improved affinity maturation of B cells. Here, we combine the use of different strains of alymphoplastic mice and T cell migration mutants with an experimental paradigm in which the site of Ag delivery is distant from the site of priming and inflammation. We demonstrate that in mammals, SLTs serve primarily B cell priming and affinity maturation, whereas the induction of T cell-driven immune responses can occur outside of SLTs. We found that mice lacking conventional SLTs generate productive systemic CD4- as well as CD8-mediated responses, even under conditions in which draining LNs are considered compulsory for the initiation of adaptive immunity. We describe an alternative pathway for the induction of cell-mediated immunity (CMI), in which Ag-presenting cells sample Ag and migrate into the liver where they induce neo-lymphoid aggregates. These structures are insufficient to support antibody affinity maturation and class switching, but provide a novel surrogate environment for the initiation of CMI.
Secondary lymphoid tissues (SLTs) are highly organized structures with defined compartments consisting of B and T cell areas. These distinct locations support the rapid circulation and concentration of Ag and the interaction of Ag-presenting cells (APCs) with lymphocytes. Prevailing dogma dictates that only if competent APCs transport Ag into SLTs, an adaptive immune response is initiated; otherwise, the Ag is ignored by the immune system [1]. For the initiation of humoral antibody (Ab) -mediated immunity in mammals, the formation of B cell follicles and germinal centers (GCs) appears to be a prerequisite. The dynamic nature of such GCs, including the interaction of follicular dendritic cells (FDCs) with B cells and Ag, was recently elegantly demonstrated by others [2]. However, in contrast to the B cell-dominated cortex, T cell areas, where T cells encounter mature APCs and their cognate Ag, are structurally ill defined. Whereas intravital confocal microscopy has provided compelling evidence for the capacity of SLTs to host T cell priming [3], definitive data supporting their absolute requirement for the initiation of T cell-mediated immunity (CMI) do not exist. In addition, cold-blooded vertebrates lacking conventional SLTs generate potent immune responses upon immunization. However, in the mammalian system, the apparent immunodeficiency of mice that lack SLTs strongly supports the notion that the initiation of effective immune responses requires the dedicated structures provided by SLTs [4]–[8]. Alymphoplasia (aly/aly) mice are characterized by a complete lack of lymph nodes (LNs) and Peyer' s patches, and structural alterations of the spleen and thymus due to a point mutation in the NFκB-inducing kinase (NIK) [9]. NIK is vital for the initiation of the noncanonical NFκB cascade, which appears to play a discrete role, for instance, in the function of CD40 and lymphotoxin-β receptor (LTβR) signaling in some cell types [10]–[12]. Aly/aly mice display impaired Ab responses and loss of CMI, demonstrated by their inability to reject allogeneic grafts or tumors [4], [13], [14]. The developmental deficits in aly/aly mutants are readily explained by the requirement of NIK in LTβR signaling. LTβR is vital for the development of SLTs, and LTβR−/− mice display similar developmental defects as do aly/aly mice or NIK−/− mice [12], [15]. In this study, we describe that the immunodeficiency of aly/aly mice is not due to the absence of SLTs, but due to the impact of the underlying genetic defect on cellular immunity. Using different strains of alymphoplastic mice and T cell migration mutants in an experimental paradigm in which the site of Ag-delivery is distant from the site of priming and again distant from the site of inflammation, we can detect both TH cell-driven autoimmune disease as well as systemic CTL-mediated antitumor immunity initiated through classical subcutaneous (s. c.) immunization/vaccination independent of SLTs. APCs present at the site of immunization migrate to and select the liver as a natural extra-lymphoid tissue for the initiation of CMI, which we propose to be an evolutionary hard-wired pathway already found in cold-blooded vertebrates. This alternative pathway, undescribed to this day, can potently drive CMI but fails to elicit B cell immunity, indicating that the immunization-induced T cell accumulation within conventional lymphoid organs mainly serves humoral immunity but that CMI can be initiated elsewhere.
We first sought to determine whether LNs are an absolute requirement for the induction of a complex TH cell-driven autoimmune response initiated by the s. c. delivery of auto-Ag. Experimental autoimmune encephalomyelitis (EAE) is a B cell-independent, TH cell-mediated demyelinating autoimmune disease of the central nervous system (CNS) and serves as the animal model for multiple sclerosis (MS). The conversion of TH cells from the naive to effector state is vitally dependent on the structures provided by LNs [5], [6]. Cervical LNs are widely held to constitute the predominant intrinsic priming site for encephalitogenic T cells, based on the observation that these LNs support the expansion of PLP-TcR transgenic (Tg) T cells [16], [17]. However, draining inguinal LNs drive the polyclonal, endogenous T cell population after s. c. immunization with encephalitogenic peptides. To assess the role of SLTs in the transition of TH cells from a naive to effector state (T cell priming), we induced EAE in aly/aly or aly/+ mice by s. c. immunization with myelin oligodendrocyte glycoprotein peptide and complete Freud adjuvant (MOG35–55/CFA). Figure 1A shows that aly/aly mice are completely resistant to EAE compared to aly/+ control mice (the latter developing normal SLTs as NIK is haplosufficient). To verify the notion that pathogenic T cells cannot be raised in aly/aly mice, they were immunized s. c. with MOG35–55. Eleven days postimmunization (dpi), splenocytes were harvested, and MOG35–55-reactive cells were expanded in vitro and subsequently transferred into aly/aly as well as aly/+ recipients. Figure 1B shows that only cells derived from aly/+ donors were able to induce disease regardless of whether the recipients had SLTs (aly/+) or not (aly/aly). In contrast, MOG35–55-reactive T cells derived from aly/aly donors were not pathogenic and did not mediate CNS inflammation. To assess the capacity of LN-less mice to initiate T cell expansion in response to s. c. -delivered Ag, CFSE-labeled TcR Tg T cells (2D2) specific for the encephalitogenic MOG35–55 peptide [18] were adoptively transferred into either aly/aly or aly/+ mice prior to immunization with their cognate Ag. After 4 d, splenocytes were analyzed for T cell expansion by flow cytometry (Figure 1C). Ag-specific T cell proliferation can be observed in aly/aly mice; however, they display slightly delayed kinetics in comparison to aly/+ mice. Similar results were obtained with Ovalbumin (OVA) TcR Tg T cells (OTII) transferred into aly/aly and aly/+ mice (unpublished data), indicating that T cell expansion can be initiated independent of SLTs, whereas efficient effector function is dependent on the microenvironment provided by SLTs. The fact that aly/aly mice do not develop T cell-driven autoimmune disease could be explained by their inability to prime self-reactive T cells (a) due to the lack of dedicated draining LNs [5], [6], or (b) due to a direct impact of the NIK mutation on immune cells [19], [20]. In order to define whether their EAE resistance is due to the lack of LNs or an intrinsic defect of aly/aly mice to prime T cells, we generated a series of bone marrow (BM) -chimeric mice. To restrict the NIK mutation to the hematopoietic system, lethally irradiated aly/+ mice were injected with BM cells from aly/aly donor mice (aly/aly→aly/+). Conversely, to conserve the developmental structural defects, without the NIK lesion of the hematopoietic compartment, aly/aly mice were reconstituted with BM cells of normal aly/+ donors (aly/+→aly/aly). As previously reported, spontaneous development of lymphoid tissues in aly/aly recipients upon reconstitution was expectedly not detected [21]. Surprisingly, we discovered that aly/+→aly/aly BM-chimeras were fully susceptible to EAE after s. c. immunization with MOG35–55 (Figure 2A), clearly demonstrating that s. c. immunization can mount a productive T cell-driven autoimmune response even in the absence of draining LNs. Using the reciprocal approach, by generating aly/+→aly/+ (WT-NIK immune system and normal SLTs) as well as aly/aly→aly/+ BM-chimeras (NIK-deficient immune system and normal SLTs), we found that the NIK mutation lead to EAE resistance, even when the lymphoreticular compartment is unperturbed (Figure 2B). This finding clearly demonstrates that the reported immunodeficiency of aly/aly mice can largely be explained by the requirement of NIK for the initiation of immunity rather than the lack of LNs. In support of this, we found that unmanipulated LTβR−/− mice, which also lack all LNs but have normal NIK function, are also fully susceptible to EAE (Figure 2C). The formation of IFNγ and IL-17–secreting autoreactive T cells has been demonstrated to be a prerequisite for the development of autoimmunity [22]. In aly/aly→aly/+ mice we observed a substantial reduction in IL-17– and IFNγ-producing cells compared to the control mice aly/+→aly/+ (Figure 2D), indicating that the resistance to EAE in the absence of NIK could be related to the function of NIK in T cell polarization. The mechanistic underpinnings of this phenomenon are currently being investigated, but it is clear that the loss of NIK signaling impairs the capacity of aly/aly mice to generate pathogenic TH cells regardless of their structural defects. Given the dogma that in mammals, CMI initiated by s. c. or intramuscular Ag-delivery requires the presence of SLTs, it is feasible that the remaining SLT (i. e. , the spleen) in aly/+→aly/aly BM-chimeras compensates for the absence of LNs. In order to test this notion, we splenectomized aly/+→aly/aly BM-chimeras (aly/+→aly/alyspl) 14 d prior to the induction of EAE. Upon immunization, aly/+→aly/alyspl mice developed EAE with the same disease severity as control mice (Table 1). We noted a slight delay in disease onset when all SLTs are absent, while histopathological analysis of diseased mice revealed no difference between aly/+→aly/+ and aly/+→aly/alyspl mice (Figure S1). In contrast to T cell activation, we found that B cell activation requires the structural environment provided by SLTs. To investigate the impact of immunization on T versus B cell responses, we used Keyhole limpet hemocyanin (KLH) as a model of foreign Ag to elicit delayed-type hypersensitivity (DTH) responses. Aly/aly as well as aly/+ mice were immunized with KLH, and 11 dpi, they were challenged by intradermal injection with KLH into the ear. As illustrated in Figure 2E, both groups were able to mount a solid DTH reaction measured by ear swelling, which was only marginally lower in aly/aly than in aly/+ mice. However, in contrast to ear swelling, which is indicative of CMI, aly/aly mice did not mount Abs against KLH when compared to aly/+ mice, demonstrating that the development of a humoral immune response is ablated in the absence of lymphoreticular structures (Figure 2F). We could reproduce functional DTH responses using other Ags including OVA and MOG35–55 (unpublished data). Similarly, in our EAE paradigm using BM-chimeras, whereas control mice (aly/+→aly/+ and aly/+→aly/+spl) elicit high Ab titers, anti-MOG Abs are virtually absent in mice without LNs (either aly/+→aly/aly or aly/+→aly/alyspl) (Figure 3A). Analysis of isotype subtypes revealed that in splenectomized alymphoplastic mice, elevated anti-MOG IgM could be detected, which has previously been reported [7], [23], [24], whereas class switching to IgG could not be observed (Figure 3B). Taken together, and in agreement with the notion that SLTs are vital for B cell activation, highly organized SLTs are obligatory for the generation of high-affinity Igs and class switching, whereas potent cellular immunity can be induced successfully upon s. c. immunization even in the absence of SLTs. Since the loss of SLTs in aly BM-chimeric mice does not hinder the development of T cell immunity, we wanted to determine at which alternative site T cell priming could take place and to which organ the Ag travels from the site of immunization (s. c.). Therefore, aly BM-chimeras were injected s. c. with yellow green (YG) carboxylate microspheres emulsified in CFA. At 7 dpi, various organs were isolated and analyzed for the presence of fluorescent cells by flow cytometry. Figure 4A shows that in control mice (aly/+→aly/+), fluorescently labeled APCs were exclusively detected in LNs upon s. c. immunization. It was previously shown that the BM has the capacity to drive an enriched population of high-affinity TcR Tg T cells in response to blood-borne Ag [25]. As expected, upon intravenous (i. v.) delivery of Ag, the vast majority of it accumulates in the spleen, BM, and liver, regardless of the presence of SLTs (Figure S2). However, after (s. c.) immunization of aly/+→aly/alyspl BM-chimeras lacking SLTs, APCs carrying fluorescent microspheres migrate primarily to the liver and not the other organs analyzed (thymus, CNS, and gut; unpublished data) (Figure 4A). Only a small amount of Ag reaches the liver when draining SLTs are present. Next, we wanted to determine the means of the Ag transport from the s. c. reservoir to the liver. To determine whether the Ag diffuses to the liver or is actively transported by APCs, aly/+→aly/+ and aly/+→aly/alyspl chimeric mice were separated into two groups. One received YG microspheres/CFA in the left flank and polychromatic red (PR) microspheres/CFA in the right flank. The other group received a mixture of YG- and PR-coupled beads in both flanks (see scheme in Figure 4B). After 7 d, mice were sacrificed, perfused, and a single-cell suspension of livers, LNs, and spleens was generated for cytofluorometric analysis. We found that the mixture of PR/YG-coupled beads generated a large proportion of dual-labeled CD11b as well as CD11c-positive APCs. Conversely, the injection of either PR- or YG-coupled microspheres into each flank revealed merely single-labeled APCs in the liver. The presence of single-labeled cells within the liver strongly suggests that the Ag is delivered to the liver by the migration of APCs initially present at the site of immunization. Passive diffusion of the Ag from the site of immunization via the bloodstream to the liver cannot be fully excluded, but is evidently not the dominant means of Ag delivery. In addition, only a negligible amount of Ag reaches the liver when dedicated SLTs are present (Figure 4). We could also confirm these findings by using soluble FITC painted on shaved flanks (without the adjuvant CFA). Twenty-four hours after FITC skin painting, we found FITC+ APCs primarily in the liver, again supporting the notion that the liver can serve as an alternative Ag-presenting site when draining LNs are not available (Figure 4C). In order to determine whether lymphoid-like structures can be found in the liver, we analyzed the livers of BM-chimeric mice immunized s. c. with MOG35–55/CFA by histology (7 dpi). Livers of aly/+→aly/alyspl BM-chimeras showed massive infiltration of leukocytes in comparison to aly/+→aly/+ control mice (Figure 5). Histological analysis displays dendritic cells (DCs) in close proximity to T cells in the infiltrated periportal areas of the liver, indicative of T cell priming by Ag-laden APCs (Figure 5B). In spite of the stroma' s inability to respond to LTα/β, detailed histological analysis revealed the presence of VCAM and ICAM in the infiltrates as well as B cells (Figure S3) and even the presence of CXCL13 transcripts indicative of aggregates ability to recruit B cells (unpublished data). However, no evidence for GC formation could be obtained (Figure S3). We also transferred TcR Tg T cells from Luciferase-2D2 (Luc-2D2) mice into recipient BM-chimeras and observed the accumulation of Ag-responsive T cells in the liver 2 dpi with MOG35–55/CFA by bioluminescence imaging (Figure 6A). Figure 6B shows that the number of DCs (CD11c+) and adoptively transferred 2D2 T cells (CD4+/Vβ11+) is drastically increased in the liver in mice lacking SLTs. In order to demonstrate that the observed lymphocyte accumulations in the liver can support cell expansion, we injected naive (CD62L+) CD4+ T cells derived from 2D2 Tg mice into aly BM-chimeras and subsequently immunized them with MOG35–55/CFA. At 5 dpi, livers were analyzed for Ag-specific CD4+ T cell proliferation. Even in normal mice, we find a large number of expanded T cells within the liver (Figure 6C), but one could argue that they have immigrated from their initial priming site, the draining LN. However, in the absence of SLTs, the livers of aly/+→aly/alyspl BM-chimeric mice are sufficient to propagate Ag-driven T cell expansion and accumulation. In order to confirm that Ag-specific T cell proliferation occurs in situ in the liver, we administered BrdU intraperitoneally (i. p.) into aly BM-chimeras 7 dpi with MOG35–55/CFA. Thirty minutes after BrdU injections, the mice were sacrificed, and livers were analyzed for proliferating (BrdU+) CD4+ T cells by flow cytometry. Figure 6D and 6E reveal the presence of BrdU+ cells in the livers of both aly/+→aly/+ and aly/+→aly/alyspl BM-chimeras. The number of BrdU+ T cells in the liver is increased in aly/+→aly/alyspl BM-chimeras compared to the controls. The fact that we found such a rapid (30 min) emergence of proliferating T cells even in normal mice in which SLTs are present, indicates that some degree of liver-initiated CMI occurs simultaneously to the priming within draining LNs. In contrast to our findings, which show that mice lacking SLTs do not generate high-affinity Ab-responses, intranasal influenza infection of splenectomized LTα−/− mice reconstituted with wild-type (wt) stem cells, for instance, can initiate the formation of extra-lymphoid follicles within the lung, which support some degree of B cell maturation and Ab secretion [24], [26]. One possible explanation for these contrasting observations regarding Ab production is that in our case, stroma cells such as FDCs cannot signal through LTβR due to the mutation within NIK and that this could be the reason for our inability to observe GC formation and Ab secretion, whereas Moyron-Quiroz et al. [24], [26] used mice in which the stroma compartment can be engaged by LTα/β. To definitively address whether the stroma' s inability to signal through NIK is the reason for the weak B cell response, we obtained LN-deficient LTα−/− mice and reconstituted their hematapoietic system with wt stem cells. The resulting chimeras were splenectomized and lacked all peripheral SLTs (analogous to the aly/+→aly/alyspl). Yet in contrast to aly/+→aly/alyspl, wt→LTα−/−spl chimeras have normal stromal cell function, and FDCs are capable of responding to LTα/β. These mice were immunized s. c. , and the formation of B cell maturation and Ab production was analyzed. Figure 7A demonstrates that these wt→LTα−/−spl chimeras behave exactly like aly/+→aly/alyspl in regards to their inability to generate high Ab titers and to class switch. In a comparative fashion, we analyzed the histological parameters of wt→wt, aly/+→aly/alyspl, and wt→LTα−/−spl chimeras (Figure 7B). Although only alymphoplastic mutants revealed the presence of lymphoid aggregates surrounding periportal areas of the liver, neither FDCs nor PNA-positive clusters could be found, again supporting the notion that the surrogate structures in the liver support T cell function but fail to initiate the formation of GCs needed for Ab-affinity cell maturation and class switching. Lastly, the large number of Ki67+ cells within the liver aggregates again support our conclusion, that active proliferation within the liver can be induced by s. c. immunization (Figure 7B). Although we have demonstrated the development of TH cell-driven autoimmune disease in mice lacking SLTs, we wanted to elucidate whether these mice are also capable of inducing successful CTL immunity. We used the B16. F10 murine melanoma model, which represents a lethal and poorly immunogenic cancer. Irradiated GM-CSF expressing B16. F10 cells are used as s. c. vaccine to initiate potent CD8+-antitumor immunity against live parental B16. F10 tumor cells [27]. We injected irradiated B16. F10-GM-CSF cells s. c. into one flank of aly/+→aly/+ and aly/+→aly/alyspl chimeric mice. At 12 dpi, mice were challenged with parental B16. F10 cells injected into the opposite flank. Figure 8A shows that aly/+→aly/alyspl chimeric mice can elicit potent antitumor CTL responses revealed by the inhibition of tumor growth. Next, we transferred CFSE-labeled MHC class I-restricted OVA-TcR Tg OTI T cells into aly/+→aly/alyspl and aly/+→aly/+ BM-chimeric mice and subsequently injected irradiated B16. F10 cells expressing OVA. At 12 dpi, livers and, in control animals, also spleen and LNs were analyzed by FACS for Ag-specific CD8+ T cell expansion. As demonstrated in Figure 8B, proliferation of CD8+ OTI cells was detected in the liver of mice lacking SLTs. Hence, even under conditions in which the draining LNs are considered a compulsory site hosting the encounter of captured Ag and infiltrating CD8+ T cells, we can detect potent T cell responses, which originate in the liver when SLTs are absent. We next wanted to address the relevance of the liver to serve as an alternative priming site in a setting where LNs are present but T cell migration into LNs is defective. To this end, we analyzed plt/plt (paucity of LN T cells) mice, which display undisturbed B cell zones but severely abrogated T cell zones due to the loss of CCL19 and CCL21, which results in the inhibition of both naive T cell and DC homing into SLTs [28]. We found that plt/plt mice also developed delayed but fulminant EAE after s. c. immunization with MOG35–55/CFA (Figure 8C). Examination of liver sections of immunized plt/plt mice again revealed lymphocyte aggregates consisting mainly of CD4+ T cells and DCs within the liver (Figure 8D).
S. c. immunization instigates a situation in which draining LNs are widely held to be absolutely obligatory for the initiation of adaptive immunity. In the absence of such draining LNs, we found however, that APCs take up the Ag at the site of immunization and subsequently select the liver as an extra-lymphoid environment for the initiation of CMI. These findings are consistent with the propensity of alymphoplastic mice (NIK−/−, LTα−/−, and LTβR−/−) to develop abnormal lymphocytic infiltrates primarily in the liver [15], [29]. The lymphocyte accumulation seen in the liver of naive alymphoplastic mice does not coincide with any overt tissue damage, nor do they develop any secondary sign of hepatic injury (M. Heikenwaelder, Zurich, Switzerland, personal correspondence). Such surrogate structures are evidently not as sophisticated as true SLTs and fail to support B cell priming, but are clearly sufficient to support CMI. Such neo-lymphoid structures in the liver are not restricted to alymphoplastic mouse strains, but can be reproduced in mice in which T cells do not migrate into the LNs (plt/plt). The fact that we observe the rapid emergence of immunization-induced T cell expansion in the liver of normal mice supports the notion that the adult liver provides an efficient niche for the initiation of CMI. Moyron-Quiroz et al. [26] elegantly demonstrated that the lymphoid tissue in the lung (BALT) is sufficient to generate immunity against an infectious agent attacking the lung. In their experimental paradigm, peripheral SLTs are not compulsory for the initiation of protective immunity, and they could even observe some degree of B cell maturation. In our report, however, after s. c. immunization, the local APCs must sample the Ag and then actively migrate to and select the liver as a site for T cell priming, which then is even capable of driving autoimmune responses within the CNS. In our experimental paradigm, the site of Ag deposition, priming, and inflammation are distinct. The liver is thus not like the BALT or the NALT, a site where local immune responses can be initiated, but represents a niche for systemic T cell priming under conditions in which the draining LNs are widely held to be absolutely compulsory. The fact that Ag-laden APCs migrate from the site of immunization to the periportal areas in the liver could be explained by the presence of chemoattractive factors in the liver aggregates observed in SLT mutants. Alternatively, the extensive lymphatic network of the liver makes it an ideal niche for the accumulation of leukocytes as a reservoir when regular SLTs are inaccessible. Although the induction of CMI is not a function traditionally attributed to the adult liver, the fetal liver is a primary lymphoid organ hosting early hematopoiesis. Our findings suggest that the liver has the potential to “remember” its lymphoid function. The phenomenon, that, for instance, food allergies can be transferred by the transplantation of livers from an allergic donor to a previously nonallergic recipient [30], can be explained by our findings. Such transplant-acquired food allergy has only been described for the liver and not for other transplanted organs of the same donor [30]. It has been hypothesized that this occurrence is due to donor-derived allergen-specific lymphocytes residing in the liver. In support of this, Klein and Crispe [31] reported recently that after liver transplantation in a mouse in which Ag presentation was restricted to resident cells of the liver grafts, efficient CD8+ T cell priming can be induced locally in the transplanted liver. The situation also is reminiscent of the effect of immunizations on some cold-blooded vertebrates that are much more primitive than mammals in their SLT organization (i. e. , lacking GCs and showing only minimal affinity maturation). Frog tadpoles (Alytes obstetricans) immunized with rabbit serum in CFA developed a large accumulation of lymphocytes in the liver visible 2–3 wk after injection (L. Dupasquier, Basel, Switzerland, personal correspondence). Interestingly, during evolution, the emergence of RAG was permissive for the development of adaptive immunity in jawed fish [32]. RAG mediates somatic recombination and is required for the formation of both B and T cell receptors, which appear to have emerged simultaneously during evolution. However, whereas the adaptive immune system is well developed in the oldest jawed vertebrates (cartilaginous fish, e. g. , sharks), potent affinity maturation, Ig-class switching, and GC formation are lacking. Class switching only appeared at the time of the divergence of amphibians [33]. The fact that CMI evolved earlier than modern humoral immune responses corroborates our discovery that T cells can function outside of dedicated lymphoreticular structures. In summary, we demonstrate that the structural requirements for the initiation of B and T cell responses differ significantly. We found that B cells are dependent on the topography of dedicated lymphoid tissues, whereas CD4+ as well as CD8+ T lymphocytes retain the capacity to recognize Ag in a structure-independent fashion. This finding has obvious implications for our understanding of adaptive immunity and vaccination. As for the development of autoimmune diseases, our findings show that self-reactive T cells may not need to be primed in tissue-draining LNs, but could occur at the inflammatory site or even in organs distant to the target tissue.
C57BL/6 mice were purchased from Janvier Laboratories. Alymphoplasia (aly/aly) mice were obtained from Clea Laboratories and bred in-house under specific pathogen-free (SPF) conditions. Heterozygous aly (aly/+) mice were used as controls for homozygous aly mice (aly/aly); 2D2 (MOG-TCR Tg) mice were provided by V. Kuchroo (Harvard Medical School, Boston, Massachusetts); LTβR−/− and LTα−/− mice were provided by A. Aguzzi and M. Heikenwalder (University Hospital Zurich, Zurich, Switzerland); and OTII and OTI mice were purchased from Jackson Laboratories. Luciferase (pbActin-Luciferase) Tg mice were obtained from C. Contag (UCSF) and crossed to the 2D2 mice (Luc-2D2). Plt/plt mice were obtained from B. Ludewig (Kantonsspital St. Gallen, Switzerland). All mice were bred in-house under SPF conditions. BM-chimeras were generated as described previously [34]. Mice were splenectomized as described previously [35]. Animal experiments were approved by the Swiss veterinary Office (68/2003,70/2003,10/2006, and 13/2006). MOG35–55 peptide (MEVGWYRSPFSRVVHLYRNGK) was obtained from GenScript. EAE was induced as described previously [34] with the modification that BM-chimeras were generally not boosted with pertussis toxin. For adoptive transfer, MOG-reactive lymphocytes were generated as described [34]. Each time point shown is the average disease score of each group±the standard error of the mean (SEM). Mice were euthanized with CO2, and various organs were removed to isolate leukocytes: For isolating lung cells, lungs were incubated with DNase (0. 5 mg) /Liberase (1 mg/ml) (Roche) for 30 min at 37°C. Spleen, LNs, thymus, and lung were homogenized, and BM cells were isolated by flushing the bones with PBS. Cells were strained through a 100-µm nylon filter (Fisher) and washed. Erythrocytes of whole blood, BM, and spleen were lysed. For isolating hepatic nonparenchymal cells, the liver was incubated with DNase/Liberase for 30 min at 37°C, homogenized, and then centrifuged at room temperature (RT) for 2 min at 50g. The supernatant was then centrifuged at 1,500 rpm for 10 min, and the pellet was resuspended in 30% Percoll (Pharmacia) and centrifuged at 12,000 rpm for 30 min at 4°C. The interphase cells were collected and washed. For isolating intestinal lymphocytes, intestines were opened longitudinally, washed, and then cut into small pieces. Tissues were then incubated with DNase/Liberase and leukocytes were isolated using a percoll gradient as described above. Isolation of CNS lymphocytes has been described previously [35]. Mice were injected i. v. with 20×106 CFSE (carbofluorescein diacetate succinimidyl ester) -labeled (Invitrogen/Molecular Probes) (10 µM) splenocytes obtained from either 2D2, OT-II, or OT-I TcR Tg mice or with 8×106 CFSE-labeled naive CD4+ 2D2 Tg T cells (isolated with CD4+CD62L+ isolation kit from Miltenyi). Mice were subsequently immunized s. c. with 200 µg of MOG35–55. /CFA (Adjuvant complete H37 Ra. . ; DIFCO) (for 2D2), OVA323–339/CFA (for OT-II), or with a 1∶1 mix of irradiated 2×106 B16. F10-GM-CSF/B16. F10-OVA cells (for OT-I). At 4 or 5 dpi (12 dpi for OT-I), mice were sacrificed, and spleen, LNs (if present), and livers were analyzed by fluorescence-activated cell sorting (FACS) for the proliferation of CD4+ T cells using the clonotypic TcR and CFSE fluorescence (2D2: TCR Vα3. 2 Ab; OT-II and OT1: Vα2 Ab). Tissues were freshly snap-frozen in liquid nitrogen. To determine infiltration of inflammatory cells, tissue sections were stained with hematoxylin and eosin (H&E) or with the following mouse-specific Abs as previously described [34]: anti-CD11c (Jackson ImmunoResearch Labs), anti-CD11b (BMA Biomedicals), anti-CD3, anti-CD4, anti-CD19, anti-FDC M1, and anti-Thy1. 1 (BD-Pharmingen), anti-ICAM, anti-VCAM, and anti-CD8 (Serotec). GC cells were stained with peanut agglutinin (PNA; Vector Laboratories). For FACS analysis, the following Abs were used: anti-CD11c, anti-CD4, anti-CD8, anti-CD11b, anti-Vα3. 2, anti-Vα3, and anti-Thy1. 1 (BD-Pharmingen). The cells were analyzed using a FACS-Canto (BD) with Cell-Diva software. Postacquisition analysis was performed using FLOWJO software. To trace the distribution of Ag after immunization, mice were injected s. c. with 200 µl of yellow-green (YG) or polychromatic red (PR) 1. 0-µm microspheres (Polysciences) emulsified in CFA. At 7 dpi, mice were euthanized with CO2, and organs were removed to isolate lymphocytes as described above. Single-cell suspensions were analyzed by FACS for the presence of fluorescein isothiocyanate (FITC+) or PE+ cells. For FITC skin painting, mice were painted on the shaved flanks with 100 µl of 5 mg/ml FITC (Molecular Probes) dissolved in 1∶1 acetone: dibutylphtalate. On day 1, mice were euthanized with CO2, and organs removed and analyzed by FACS as described above. Mice were immunized s. c. with 100 µg/flank of KLH (Sigma) emulsified in CFA. At 11 dpi, mice were challenged by injecting 10 µg/10 µl KLH, PBS into the dorsal surface of the ear. DTH responses were determined by measuring the ear thickness using a caliper micrometer (Mitutoyo) 24 h after challenge, and Δ ear swelling was established by the increase in ear thickness over baseline (prechallenge ear thickness). Plates were coated with 10 µg of rMOG1–121 in 0. 1 M NaHCO3 (pH 9. 6) at 4°C overnight or KLH (Sigma), and blocked with 1% (w/v) bovine serum albumin (BSA). Diluted sera were incubated for 2 h at RT. After washing, peroxidase-conjugated antibodies to mouse immunoglobulins, IgG, IgA, and IgM (Sigma) were added (1∶1,000 diluted) and incubated for 1 h at RT. Plates were washed, and chromogen (Biosource) was added. Absorbance was measured on a microplate reader (450 nm) (Bio-Rad). A total of 2×105 cells were plated in medium containing 10% FCS and 50 µg/ml of MOG35–55 in 96-well plates (Millipore) coated with the capture Ab against either IFNγ or IL-17A [36]. Elispots were revealed as described previously [36] and subsequently analyzed on an Elispot reader (CTL immunospot). To visualize Luc-2D2 cells, mice were injected i. p. with 3 mg of luciferin (Xenogen) prior to bioluminescence imaging using an IVIS100 imaging station (Xenogen). The luminescent image was overlaid on the photographic image. Mice were immunized s. c. with MOG35–55/CFA. At 7 dpi, BrdU (BD Pharmingen) (2. 5 mg) was injected i. p. 30 min before the mice were sacrificed and analyzed for proliferating (BrdU+) CD4+ T cells by flow cytometry with anti-BrdU Ab (eBioscience). Mice were s. c. vaccinated into one flank with irradiated (6,000 rads) 1×106 B16. F10-GM-CSF cells. At day 12 after vaccination, mice were injected with live 2×105 B16. F10-Luc cells into the opposite flank. Each time point shown is the average tumor size of each group±SEM, measured using a caliper. | Lymph nodes (LNs) are believed to be the most important tissues initiating immune responses by facilitating the activation of T and B lymphocytes. Mice lacking such LNs (called alymphoplastic) are severely immune compromised and resistant to immunizations. We discovered that the immune-deficiency of such alymphoplastic mice is actually not caused by the loss of LNs, but rather by the underlying genetic lesion. Surprisingly, mice lacking all lymph nodes can still mount potent T cell-mediated immune responses. We also discovered that T and B cells have completely different structural requirements for their activation/maturation. Whereas B cells rely on LNs to become efficient antibody-producing cells, T cells can be activated successfully outside of such dedicated tissues. So—in the absence of LNs—antigens delivered by immunization are actively transported into the liver where cellular immunity is initiated. The mammalian fetal liver is responsible for the early formation of blood and immune cells, and we propose that the adult liver can still provide a niche for T cell–antigen encounters. During evolution, T and B cells emerged simultaneously, allowing cold-blooded vertebrates (which lack LNs) to launch adaptive immune responses. The development of LNs in mammals coincided with a drastic improvement in antibody affinity maturation, whereas T cells remain LN-independent to this day. | Abstract
Introduction
Results
Discussion
Materials and Methods | immunology/immunomodulation
immunology/immune response
immunology/innate immunity
surgery/transplantation
immunology
evolutionary biology
neuroscience/neurobiology of disease and regeneration
immunology/autoimmunity
immunology/leukocyte activation | 2009 | Neo-Lymphoid Aggregates in the Adult Liver Can Initiate Potent Cell-Mediated Immunity | 10,172 | 328 |
Oncolytic viruses replicate selectively in tumor cells and can serve as targeted treatment agents. While promising results have been observed in clinical trials, consistent success of therapy remains elusive. The dynamics of virus spread through tumor cell populations has been studied both experimentally and computationally. However, a basic understanding of the principles underlying virus spread in spatially structured target cell populations has yet to be obtained. This paper studies such dynamics, using a newly constructed recombinant adenovirus type-5 (Ad5) that expresses enhanced jellyfish green fluorescent protein (EGFP), AdEGFPuci, and grows on human 293 embryonic kidney epithelial cells, allowing us to track cell numbers and spatial patterns over time. The cells are arranged in a two-dimensional setting and allow virus spread to occur only to target cells within the local neighborhood. Despite the simplicity of the setup, complex dynamics are observed. Experiments gave rise to three spatial patterns that we call “hollow ring structure”, “filled ring structure”, and “disperse pattern”. An agent-based, stochastic computational model is used to simulate and interpret the experiments. The model can reproduce the experimentally observed patterns, and identifies key parameters that determine which pattern of virus growth arises. The model is further used to study the long-term outcome of the dynamics for the different growth patterns, and to investigate conditions under which the virus population eliminates the target cells. We find that both the filled ring structure and disperse pattern of initial expansion are indicative of treatment failure, where target cells persist in the long run. The hollow ring structure is associated with either target cell extinction or low-level persistence, both of which can be viewed as treatment success. Interestingly, it is found that equilibrium properties of ordinary differential equations describing the dynamics in local neighborhoods in the agent-based model can predict the outcome of the spatial virus-cell dynamics, which has important practical implications. This analysis provides a first step towards understanding spatial oncolytic virus dynamics, upon which more detailed investigations and further complexity can be built.
Oncolytic viruses replicate selectively in tumor cells and have been explored as a targeted treatment approach against cancers [1], [2], [3], [4], [5], [6], [7], [8], [9], [10], [11], [12], [13], [14], [15]. In principle an oncolytic virus will spread though the tumor cell population and lyse the infected cells, leading to eradication or control of the tumor. Because of the selectivity of such viruses for cancer cells rather than normal human cells, side effects also should be less pronounced than those associated with traditional treatments, such as chemotherapy or ionizing radiation. Oncolytic virus therapy has been explored in the context of several different virus species. While some non-human viruses display natural selectivity for cancer cells in humans [16], modern approaches use genetically engineered viruses to achieve tumor selectivity. The first engineered virus generated in the 1990s was a herpes simplex virus-1 [17]. Engineered adenoviruses have been of major interest in recent clinical trials, especially in the context of head and neck cancer [15]. Indeed the adenovirus H101 (Shangahi Sunway Biotech, Shanghai, China) was approved in China for the treatment of head and neck cancer in combination with chemotherapy [18]. A variety of other virus types has also been explored [19]. However despite initial promising results and observations in the laboratory and clinic, oncolytic viruses have so far failed to demonstrate sustained and reliable treatment success [15]. Besides experimental research, mathematical and computational modeling has increasingly become a tool to study the dynamics of oncolytic viruses. Mathematical models can help us understand the emerging properties of cancer-virus interactions, to interpret experimental results, and to design new experiments. The first mathematical models of oncolytic virus therapy considered ordinary differential equations that described the basic interactions between a replicating virus and a growing population of tumor cells, and also immune responses [20], [21]. Further work extended this type of approach in a number of ways, describing different scenarios and applying models to specific virus-tumor systems [22], [23], [24], [25], [26], [27], [28], [29], [30], [31], [32], [33], [34], [35]. One of the assumptions that is implicit in such modeling approaches is that cells and viruses mix perfectly with each other (mass action). While this might hold true in the context of some in vitro experiments, and while this might be a reasonable approximation of the dynamics occurring in some non-solid tumors, the majority of tumors have intricate spatial structures where cells and viruses do not mix well, but where interactions are limited to local neighborhoods. Hence, to gain a better understanding about the dynamics of oncolytic viruses, spatially explicit models are required. While some spatial modeling studies have been performed and have given rise to interesting results [30], [36], [37], [38], they commonly include, in addition to basic spatial dynamics, one or more additional assumptions that introduce further complexity. We still do not, however, have a good understanding of the basic principles that govern spatially restricted virus spread through a population of target cells, what outcomes can be expected, and what determines those outcomes. Obtaining such basic knowledge is a necessary foundation for building predictive models of virus therapy. This knowledge can be used as a basis for examining the effects of further biological complexities on the outcome of virus therapy, such as immune responses, tumor-microenvironment interactions, cellular heterogeneity, cell-cell interactions, among others. Therefore, the aim of this paper is to study the basic dynamics of virus spread through a spatially arranged population of growing cells in a simple setting. To achieve this, we have constructed an in vitro experimental system in which a fluorescent labeled virus spreads through a target cell population in a two-dimensional geometry, such that an infected source cell can transfer the virus only to target cells in the direct neighborhood. Besides quantifying the number of infected cells, this also allows us to track the emerging spatial patterns over time. We found three distinct patterns of virus spread and determined the frequency of their occurrence. An agent-based model was used to simulate these experiments and to interpret the data. The model can qualitatively reproduce the experimental observations and suggests key parameters that determine the different growth patterns. Using this model, we explore the implications of the observed growth patterns for the long-term outcome of the dynamics, and obtain insights about the conditions required for the virus to drive the target cell population extinct in this setting. This is a first step towards understanding the basic principles of virus spread and the correlates of successful virotherapy in spatially structured cell populations, and provides a basis for more detailed explorations and for the incorporation of other complexities that are relevant for virus-tumor dynamics in vivo.
In order to examine spatial virus spread in a relatively simple setting, we constructed a recombinant adenovirus type-5 (Ad5) that expresses enhanced jellyfish green fluorescent protein (EGFP), AdEGFPuci, and grows on human 293 embryonic kidney epithelial (293) cells [39]. The experiment was set up such that cells are arranged in a two-dimensional layer, and virus spread is most likely to occur to neighboring cells. An agar overlay prevents long-range spread of the virus away from infected cells in the culture medium. This set-up allows us to not only quantify the number of infected cells over time, but also the spatial patterns of infected cells that are formed as the virus population expands. In addition, we used fluorescent markers to visualize the spatial distribution of all cells (infected and uninfected) by generating HEK293-H2BmCherry cells, that stably express the core nuclear histone protein H2B fused to mCherry (a highly photostable, monomeric red fluorescent protein (RFP) ) [40]. Thus, using HEK293-H2BmCherry cells allows us to visualize all the cell nuclei (i. e. , intact cells) in any particular culture. The culture was infected at a very low multiplicity of infection (MOI), such that any area of infection resulted from a single “founder” infected cell. Each culture contained several such founder cells that were sufficiently separated from each other, allowing us to track multiple growth foci across the dish. Details of the experimental procedures are given in the Methods section. The earliest stages of virus growth starting from a single founder infected cell were characterized in detail in a separate study [39]. This gave rise to the interesting observation that while virus extinction was a likely event as long as the number of infected cells in a given area was less than three, spreading virus growth was always observed once the number of infected cells reached three or higher. In the current study, we followed the growth of such spreading infections and characterized the consequent growth patterns. We observe three basic patterns of virus spread, which interestingly occur under identical experimental conditions and even within the same culture. They are shown in Figure 1A and described as follows. (i) In the first pattern, the virus infection spreads rapidly outwards as a ring, leaving no cells behind in the core of the ring (Figure 1A, pattern (i) ). This classic plaque pattern is observed in virus growth experiments. We call this the “hollow ring” structure. In the second and third patters there is viral spread, but it is limited. (ii) In the second case, a “disperse” growth pattern is observed, where the virus population expands as a mixed cluster of infected and uninfected cells (Figure 1A, pattern (ii) ). Finally, the virus population expands as a thinner ring, but in contrast to the first case, uninfected cells are left behind in the core of the ring (Figure 1A, pattern (iii) ). We call this the “filled ring” structure. A limited growth pattern is magnified in Figure 1B, in which uninfected cells are visible within the center of the virus infected population. In the top right panel of Figure 1B, an AdEGFPuci infected (fluorescent) cell is indicated (arrow, inf.), whereas an uninfected cell in the center of the spreading infection does not fluoresce green (arrow, un inf.). The same cells are indicated in the middle right panel of Figure 1B, showing red fluorescence. In the bottom left panel of Figure 1B, images of the top and middle panels are merged; infected cell (arrow, inf.) fluoresces yellow, while the uninfected cell, (arrow, un inf.) remains red. As mentioned the area over which the infection spread remained limited in patterns (ii) and (iii) and persisted throughout the infection (through 19 dpi). In contrast, in pattern (i) the ring of infected cells continued to spread outward as long as there was space; cell clearing in the center of the plaque was apparent at 13 dpi, as shown in Figure 1A. Similar patterns of spreading infection were also seen in Ad293 cells, a HEK293 cell derivative optimized for adenovirus plaque assays. Overall, among 436 scored growth foci, the hollow ring structure was found in 45%, and the limited patterns in 55% of cases. In the following sections, we simulate these experiments with a computational model. We examine conditions for the formation of different patterns and examine implications of the different growth patterns for the ability of the virus to eradicate the target cell population. Our in silico studies of the interactions between oncolytic viruses and cancerous cells rely on the agent-based modeling technique, where each cell is represented as an “agent” occupying a certain position on a grid, and interacting with other cells according to some (probabilistic) rules. Our modeling approach is spatial, that is, it takes into account the spatial distribution of the uninfected and infected cells The model, based on previous work [41], describes target cell-virus dynamics on a two dimensional grid that contains N×N spots. Each spot is either occupied by a cell (infected or uninfected), or it is empty. We model the development of the populations in discrete time. Given the state of the system at time t, a set of rules is applied to each spot, and this gives rise to the state of the system at time t+1. At each time step, the grid is randomly sampled N2 times. If the chosen spot is occupied by an uninfected cell, it can die with a probability D, leaving the spot empty. Alternatively, the cell can reproduce with a probability R, and a destination spot is randomly chosen for the offspring from the set of eight nearest neighboring spots. If the destination spot is empty, the offspring is placed there, otherwise, no reproduction occurs. If the chosen spot contains an infected cell, it can die with a probability A, or attempt to transmit the virus with a probability B. A destination spot is chosen randomly from the eight nearest neighbors, and infection only proceeds when a susceptible cell is present. Infected cells do not reproduce since adenoviruses lock the cell in the S-phase for replication, thus preventing further divisions [34]. The particular formulation of the infection process used in this paper corresponds to virus transmission that is not coupled with cell death, but simulations indicate that transmission coupled to cell death does not qualitatively alter the results reported here. For reference, parameters and their meaning are summarized in Table 1. In this section, we explore the initial virus growth patterns. As starting conditions we assume that the grid is filled with uninfected cells and that a relatively small square of infected cells (30×30 spots) is placed in the middle of the grid (which overall contains 300×300 spots). The emerging growth pattern depends on parameters that influence the rate of virus spread, in particular the probability for an infected cell to die, A, and the probability for an infected cell to transmit the virus, B. The patterns that we observe are presented in Figure 2. In Figures 2a and b, the infected cell population expands as a ring or wave that leaves no cell behind in its core. The two pictures differ in the death probability of infected cells. In Figure 2a, the probability for infected cells to die is relatively low such that during the time frame of the simulation a hollow ring has not yet formed and the infected cell population expands as a relatively solid mass. In Figure 2b, the death probability of infected cells is higher such that during the time frame of the simulation a hollow ring has formed. In Text S1 we derive an approximate growth law for these scenarios. The total number of cells is proportional to, such that for short time-scales (or smaller death rates) the growth is quadratic in time, and for longer times scales (or larger death rates) it is linear in time. This is exactly what is observed. Figure 2a, characterized by smaller values of A, shows a growth law of the infected cell population that is close to quadratic. In Figure 2b, where the death rate is larger, the infected cell population grows linearly once the hollow ring is present. Note that these two scenarios are identical in principle because in Figure 2a, the formation of the hollow ring requires more time (and a larger grid). The higher the death rate of infected cells, the faster the ring is formed, and the faster the growth law changes from square to linear. Lowering the rate of virus spread (decreasing the value of B and increasing the value of A) gives rise to patterns of a different nature. In Figure 2c, uninfected cells are left behind in the core of the expanding ring. When they grow and become infected by virus, a coupled expanding ring of uninfected and infected cells forms. This can occur repeatedly, giving rise to concentric rings. The persistence of cells in the core of the ring is probabilistic in nature, and that is reflected in the growth laws that are observed in multiple runs of the simulation. In cases where uninfected cells are not left behind inside the ring, the infected cell population grows linearly. When concentric rings do occur, the growth becomes quadratic. Finally, no expanding ring structure is formed in Figure 2d because the viral spread kinetics are even slower. Instead, the area of virus growth is characterized by a mix of infected and uninfected cells that expands over time. In this case, quadratic growth of infected cells is observed (see Text S1). Note that if the viral spread kinetics are in the lower end of this spectrum, it is possible to observe a variation of this pattern, shown in the inset of Figure 2d: While the spreading infection leaves uninfected cells behind, the viral spread kinetics are too low to maintain significant numbers of infected cells throughout this area. Most of the infected cells will be at the outer edge of the infection due to a higher density of target cells. In this case, a relatively thin, ring-like structure can be formed, with a large area of uninfected cells remaining in its core. This pattern, however, is temporary. With time, one of two scenarios can be observed. A mixed pattern can be generated, characterized by a large number of uninfected cells and a low number of infected cells, because the virus eventually spreads to the remaining susceptible cells. Alternatively, there is a chance that the virus population goes extinct due to the slow rate of spread. Long term outcomes are discussed further below. Figures 2a and b qualitatively correspond to experimental pattern (i), the hollow ring structure. Figure 2d corresponds to experimental pattern (ii), the disperse growth structure. The pattern shown in the inset of Figure 2d likely corresponds to experimental pattern (iii), where a limited ring with mainly uninfected cells in its core temporarily forms before either developing into a mixed pattern or resulting in virus extinction. According to this interpretation, the experimentally observed patterns (ii) and (iii) are variations of the same theme. The concentric rings observed in the model simulations are not found in the experimental data. This is not surprising because they only occur in relatively narrow parameter regions in the model. In order to go beyond the qualitative comparison of model and data, we fit the model to two sets of experimental data, one showing an expanding hollow ring, and the other the disperse growth pattern. A least squares algorithm (see Text S1) was used to fit the number of infected cells over time, and a relatively good fit was obtained for both cases (Figure 3). The types of spatial patterns that emerged matched the observed ones qualitatively (Figures 4 and 5). While this procedure found best fitting parameter values, their biological meaning remains questionable, since different parameter combinations can give rise to similarly good fits. A more solid validation would require an independent estimation of parameter values, and a subsequent generation of the predicted growth patterns. Due to the complexity of the experimental observations, this is not currently possible and is discussed in detail below. The fitting procedure does, however, indicate that the model is at least consistent with experimental data. Here, we explore the long term dynamics, investigating how the above described patterns play out and correlate with the overall outcome if both the uninfected and infected cell population can expand in space. We seek to define conditions under which the virus can eliminate the target cell population in this system. All simulations are started with a small number of infected cells placed in a compact vicinity into a larger space filled with uninfected cells, which is in turn embedded into an even larger “empty” space (for the exact initial conditions for particular cases, see appropriate figure legends). In contrast to the simulations reported above, here we go beyond the initial virus growth stage, and focus on time-scales where the population of target cells experiences significant changes (grows in size in the absence of infection). The outcomes of this system include extinction of the target cells and thus the virus; extinction of the virus and persistence of the target cells; coexistence of virus and target cells. The dependency of these outcomes on the parameters is shown in Figure 6, which is the result of at least 104 instances of the simulation, where the log10 of all the parameters was varied between −4 and 4. Figures 7 and 8 show corresponding spatial and temporal patterns. We examine the outcomes first in a relatively small 30×30 grid, and subsequently in a larger, 300×300 grid.
We used a newly designed virus, AdEGFPuci, replicating on 293 embrionic kidney epithelial cells in order to study the spatial dynamics of virus growth in a two-dimensional setting in which a source cell transmits the virus to target cells in the immediate vicinity. We found that when the virus is placed into the midst of a target cell population, the initial growth pattern can be divided into two categories. Either the infected cells expand as a wave and eventually form a hollow ring, or the hollow ring is not formed and we observe cells remaining in the core of the ring which can lead to a disperse growth pattern. An agent-based model was used to qualitatively simulate these experiments. The model produced the same types of patterns found in the experiments and model fitting to the experimental data has shown that our description is consistent with observation. The data, together with the model, gave rise to a number of insights. When a ring is formed, the initial virus growth is quadratic and then becomes linear. In the absence of a ring, the entire virus growth is quadratic. These growth patterns are obviously related to and have implications for the dynamics of viral plaque formation, which have been previously investigated mathematically from other angles [53], [54], [55]. Initial virus growth patterns are correlated with the long-term outcomes of the dynamics. If a disperse growth structure is formed, persistence of the target cells at high levels is the only outcome, corresponding to definite treatment failure. On the other hand, the long-term outcome of the dynamics when a ring structure is formed includes extinction or low-level persistence of the target cells. In the best case scenario (which corresponds to parameter region A in our notations), the wave of infected cells catches up with the spreading target cells and drives the cell population extinct, which is the most desirable treatment outcome. Low-level persistence occurs when the virus wave fails to overtake and eliminate the target cell wave, resulting in a non-equilibrium persistence situation that requires larger grids in the model. These outcomes are likely to be relevant to tumors since the number of cells in a tumor is significantly larger than the number of cells present in any of our simulations. Cell persistence in this case, however, occurs at relatively low levels and this can be thought of as controlled persistence. Since a tumor is less likely to cause morbidity or mortality in such a state, this outcome can be considered as partially successful treatment. An important theoretical finding is that the outcome of the spatial system (extinction versus persistence) can be predicted well from the local dynamics, characterized by the interactions among neighboring cells. In our model, this is given by the neighborhood of 3×3 spots. This result suggests that ordinary differential equations, which describe mass action dynamics, can be a valid approach to study the correlates of successful virus therapy even for tumors that exhibit spatial structure. The results of such approaches have to be correctly interpreted to include the notion of local neighborhoods. If interpreted correctly, insights gained from such previous modeling studies are applicable to the treatment of spatial tumors. In addition, this result tells us that simple in vitro experiments (where viruses and cells mix well) can be used to compare the basic replication and spread efficiencies of different viruses, and that this directly correlates with their ability to fight a spatially structured tumor. Of course, the correlates of successful treatment derived from such simple modeling studies are qualitative in nature. That is, it is for example possible to predict the effect of increasing a certain parameter, such as the death rate of infected cells, on the outcome of treatment. However, predicting whether or not virus treatment will result in the eradication/control of tumor cells is a very difficult problem, and so far no modeling approach exists that is suited to perform this task. Not only do we lack sufficient biological information, the same biological processes can be described mathematically in different ways, adding uncertainty to these models. As a next step it will be important to examine whether the global spatial dynamics can still be predicted from local mass-action dynamics in models when further complexity is introduced, including three-dimensional spatial structures, immune responses, physical barriers to spread, or cell populations with differential susceptibility to infection. If our result holds, then relatively simple ordinary differential equation models can be used to guide the exploration of other, more complex modeling approaches, as well as the design of experiments aimed at evaluating candidate viruses. To put this work in the context of the existing literature, we note that experimental research has made much progress in the construction of viruses, in elucidating the molecular biology of virus replication in tumor cells, and in investigating clinical trials and correlates of success for several different types of candidate viruses and different tumors. Yet, the complex and multi-factorial interactions between viruses, their target cells, and other relevant components make it difficult to predict the outcomes of such dynamics, an area where mathematical tools can be of great help to complement experimental analysis, interpret data, and make experimentally testable predictions. In the recent years, there have been several mathematical/computational studies which examined spatially explicit models in more or less complex settings, taking into account an array of relevant factors that might affect the in vivo virus spread through a tumor [30], [34]–[36]. Here, we took a step back and examined the basic spatial dynamics between a virus and its susceptible tumor cells, ignoring more complex factors such as the immune system, tumor vascularization, or physical barriers to virus spread within tumors. In principle, the simpler setting allows for more solid experimental testing and validation of models. As it turns out, the dynamics in such a simplified setting already exhibit very complex behavior, and it is imperative to gain a thorough understanding of such a basic system before venturing on towards more comprehensive scenarios. This paper provides a first step to such an understanding. Due to the complexity of the experimentally observed dynamics, further questions remain that are subject to ongoing investigation. The most puzzling observation was that identical experimental conditions, using the same virus-target cell system, gave rise to different patterns of virus growth. This indicates the existence of so far unidentified factors that influence virus spread in this in vitro system. It is possible that initial events, stochastic in nature, might determine the remaining fate of the virus population. One hypothesis is that infection of cells triggers the production of anti-viral factors by the infected cell, which could induce an anti-viral state in neighboring cells. An example of such a factor could be interferons. Indeed adenovirus infection has been reported to induce interferon beta production [56], and 293 cells respond to interferon beta by activation of interferon-responsive genes (System Biosciences. Interferon Response Detection Kit – User Manual. Mountain View, CA). Cells in the anti-viral state can have both a reduced susceptibility to infection and/or an increased death rate through the induction of apoptosis if they do become infected [57]. In such a setting, it is possible that a race occurs between the spread of the virus and the anti-viral factors to neighboring cells. The population that initially gains the upper hand in this race could determine the emerging pattern of virus growth. The hollow ring structure could be formed if the virus out-runs the anti-viral factors, and the disperse growth pattern could be observed if the virus population fails to do so. While parameter estimates from the model fit to the data remain inconclusive, the fit accounted for the disperse pattern through an increased death rate of infected cells, which might suggest that apoptosis significantly contributes to the anti-viral state. This concept could also explain the observation that the area of infection remains limited when the disperse growth pattern is found in the experiments. In our model, which does not take into account such anti-viral factors, the infection area continues to grow, albeit slowly. Interferons have been shown to influence the development of plaques before in the context of herpes simplex virus [58], and the dynamics have been examined with a similar modeling approach [59]. In this study, however, different dynamics and outcomes were observed. The finding that different patterns can form in identical experimental conditions also blocks our ability to turn our model into a truly predictive one. This would require parameters to be measured independently in the experimental system, and to demonstrate that these parameters yield a satisfactory fit of the model to the data. However, the dynamics are most likely characterized by different parameter combinations in the context of the different observed patterns. Before we fully understand the reasons for the various patterns, parameterization of the model remains an impossible task. Nevertheless, the model analysis presented here does highlight the parameters that determine the different outcomes, which guides our search for the responsible mechanisms. Identifying these mechanisms still requires extensive work which goes beyond the scope of the current paper. We conclude that the dynamics of virus spread in the simplest spatial setting can be very complicated. We interpreted these dynamics with a computational model, and shed light onto the meaning of the different spatial patterns observed. We further found theoretically that mass-action dynamics in local areas can be indicative of the outcome of virus spread in a spatially structured cell population. This suggests that previous insights, gained from the analysis of ordinary differential equations, remain relevant for the spread of viruses in a spatial setting.
Both an agent-based model and a metapopulation model were used to examine the spatial dynamics of virus spread through a population of growing target cells. Analytical and computational methods are described in detail in the Text S1. | Traditional chemotherapy of cancers is characterized by strong side effects, while showing a low success rate in the long term control of tumors. Besides small molecule inhibitors, which have shown great promise, oncolytic viruses present an emerging specific treatment approach. They are engineered viruses that spread from tumor cell to tumor cell, killing them in the process. Non-tumor cells are generally not infected. While clinical trials have given rise to promising results, reliable success remains elusive. Besides experiments, computational approaches provide a valuable tool to better understand the dynamics of virus spread through a growing population or tumor cells. Combining in vitro experimental approaches with computational models, we study the principles of virus spread through a spatially structured population of cells, which is of fundamental importance to understanding virus treatment of solid tumors. We describe different growth patterns that can occur, interpret them, and explore how they relate to the ability of the virus to induce tumor regression. We further define how these spatial dynamics relate to settings where cells and viruses mix more readily, such as in many cell culture experiments that are used to evaluate candidate viruses. | Abstract
Introduction
Results
Discussion
Methods | theoretical biology
biology | 2012 | Complex Spatial Dynamics of Oncolytic Viruses In Vitro: Mathematical and Experimental Approaches | 6,720 | 235 |
A powerful way of gaining insight into biological systems is by creating a nonlinear differential equation model, which usually contains many unknown parameters. Such a model is called structurally identifiable if it is possible to determine the values of its parameters from measurements of the model outputs. Structural identifiability is a prerequisite for parameter estimation, and should be assessed before exploiting a model. However, this analysis is seldom performed due to the high computational cost involved in the necessary symbolic calculations, which quickly becomes prohibitive as the problem size increases. In this paper we show how to analyse the structural identifiability of a very general class of nonlinear models by extending methods originally developed for studying observability. We present results about models whose identifiability had not been previously determined, report unidentifiabilities that had not been found before, and show how to modify those unidentifiable models to make them identifiable. This method helps prevent problems caused by lack of identifiability analysis, which can compromise the success of tasks such as experiment design, parameter estimation, and model-based optimization. The procedure is called STRIKE-GOLDD (STRuctural Identifiability taKen as Extended-Generalized Observability with Lie Derivatives and Decomposition), and it is implemented in a MATLAB toolbox which is available as open source software. The broad applicability of this approach facilitates the analysis of the increasingly complex models used in systems biology and other areas.
Mathematical modelling has become a fundamental tool in present day biology [1], and system identification is one of the key tasks of this process [2]. Building a dynamic model usually requires establishing the values of some unknown parameters, which raises the issue of parameter identifiability [3]. A model is structurally identifiable if it is possible to determine the values of its parameters from observations of its outputs and knowledge of its dynamic equations [4]. While the related concept of practical identifiability refers to quantifying the uncertainty in parameter values when estimated from noisy measurements, structural identifiability does not take into account limitations caused by the quality or availability of experimental data. It is, however, a necessary (a priori) condition for practical identifiability, which, in turn, is a prerequisite for model calibration, also known as parameter estimation [5]. Any identification efforts aimed at estimating unidentifiable parameters will fail, leading to wrong estimates, waste of resources, and possibly misleading model predictions [6]. Furthermore, if structural unidentifiability is mistaken for practical unidentifiability, it may lead to trying to overcome it by investing additional efforts in designing and performing new experiments [7], which will nevertheless be sterile. Hence it is essential to assess the structural identifiability of any unknown parameters in a model before attempting to calibrate it. As stressed in the conclusions of a recent parameter estimation challenge [8], “modelers must avoid creating structurally unidentifiable parameters that can never be estimated”. However, in real applications structural identifiability is seldom checked before performing parameter estimation [9]. This is at least partly due to the computational complexity of the problem: structural identifiability methods generally require symbolic manipulations, which can quickly give rise to long expressions as the system size increases [10]. This is a major challenge in systems biology, as the models constructed are increasingly complex, large [11], and more difficult to identify [8]. However, the development of structural identifiability tools has been lagging behind, and, despite the wide variety of methods developed for this task (some of which have publicly available implementations [12–16]), the analysis of some models remains elusive. For example, although recent improvements in efficiency [16,17] have enabled the analysis of increasingly large rational models (those that can be expressed as fractions of polynomial functions), non-rational systems such as those including trigonometric expressions or Hill-type kinetics (which are common in mechanical and biochemical models, respectively) can currently be analysed only for small sizes. While in certain cases non-rational models can be rewritten in rational form, by introducing additional variables and equations, it is not always possible or convenient to do so. Furthermore, the results obtained for the rational counterpart are not necessarily valid for the original non-rational model in the case of unidentifiability [18]. Recent studies [9,10,19–21] show that, in general, the choice of a structural identifiability method involves trade-offs between generality of the application, computational cost, and level of detail of the results. In conclusion, there is currently a lack of structural identifiability methods of the sufficient generality and robustness to be applied to nonlinear models of general form and realistic size [21,22]. To address this issue, we propose a methodology applicable in principle to any analytic system and geared towards computational efficiency. This method approaches local structural identifiability as a generalized version of observability, a classic concept in systems and control theory [23]. A system is observable if it is possible to determine its internal state from output measurements in finite time. If the model parameters are considered as state variables with zero dynamics, structural identifiability analysis can be recast as a generalization of observability analysis [17,24,25]. In this way it is possible to assess the structural identifiability of nonlinear systems using results from differential geometry [21]. Essentially, identifiability is determined by calculating the rank of a generalized observability-identifiability matrix, which is constructed using Lie derivatives. When this rank test classifies a model as unidentifiable, the procedure determines the subset of identifiable parameters. In some cases it is also possible to find identifiable combinations of the remaining parameters. This approach is directly applicable to many models of small and medium size; larger systems can be analysed using additional features of the method. One of them is decomposition into more tractable submodels, which is performed with a combinatorial optimization metaheuristic as in [26]. Another possibility is to build identifiability matrices with a reduced number of Lie derivatives. In some cases these additional procedures allow to determine the identifiability of every parameter in the model (complete case analysis); when such result cannot be achieved, at least partial results—i. e. identifiability of a subset of parameters—can be obtained. We illustrate the applicability of this method to systems biology models of different types, including genetic, signalling, metabolic, and pharmacokinetic networks. Some of them are non-rational systems exhibiting Hill kinetics, that is, with expressions containing terms of the form k1xn/ (k2 + xn), such as the Goodwin model of transcriptional repression [27], the mitogen-activated protein kinase (MAPK) signalling cascade [28], and the genetic network that controls the circadian clock in Arabidopsis thaliana [29]. Other models analysed here include drug uptake into hepatocytes [19], NF-κB [30] and JAK/STAT [31] signalling pathways, and the central carbon metabolism of Chinese hamster ovary cells [32]. These case studies include models whose identifiability had not been previously determined, and for some of them we found unidentifiabilities that had not been reported before. In those cases, we obtained identifiable reparameterizations by removing redundant parameters and fixing the values of other parameters a priori.
In this work we consider identifiability as an augmented observability property. We begin the description of the approach by defining observability and showing how it can be assessed. A system is (locally) observable at a state x0 if there exists a neighbourhood N of x0 such that every other state x1 ∈ N is distinguishable from x0. Two states x0 ≠ x1 are said to be distinguishable when there exists some input u (t) such that y (t, x0, u (t) ) ≠ y (t, x1, u (t) ), where y (t, xi, u (t) ) denotes the output function of the system for the input u (t) and initial state xi (i = 0,1). The concept of observability was initially formulated by Kalman for linear systems [34], and then extended to the nonlinear case by Hermann and Krener [23]. For a nonlinear system given by Eq (1) it is possible to obtain information about the states x from its outputs y by calculating the derivatives y˙, y¨, …. These differentiations are performed by taking Lie derivatives of the output function g. The Lie derivative of g with respect to f is: L f g (x) = ∂ g (x) ∂ x f (x, u) (3) For a system with n states and m outputs, ∂ ∂ x g (x) is an m × n matrix, and L f g (x) = ∂ g (x) ∂ x f (x, u) is an m × 1 column vector. The ith order Lie derivatives are recursively defined as follows: L f 2 g (x) = ∂ L f g (x) ∂ x f (x, u) ⋯ L f i g (x) = ∂ L f i - 1 g (x) ∂ x f (x, u) (4) Stacking n sub-matrices, we obtain the nonlinear observability matrix: O (x) = ∂ ∂ x g (x) ∂ ∂ x (L f g (x) ) ∂ ∂ x (L f 2 g (x) ) ⋮ ∂ ∂ x (L f n - 1 g (x) ) (5) We can now formulate the Observability Rank Condition (ORC) as follows: if the system given by Eq (1) satisfies rank (O (x 0) ) = n, where O is defined by Eq (5), then it is (locally) observable around x0 [35]. The rank condition provides a result about local observability of any possible state x0. That is, if the matrix is full rank then for every state x0 there exists a neighbourhood N (x0) in which x0 can be distinguished from any other state x*. In other words, every state can be distinguished from its neighbours, but not necessarily from other distant states. In contrast, global observability is a property that must hold for every possible N (x0). The difference is clearly shown with the following example [23]: x ˙ = u, y 1 = cos (x), y 2 = sin (x) (6) While this system satisfies the observability rank condition and is therefore locally observable, it is not globally observable because it is impossible to distinguish between x0 and xk = x0 + 2kπ, for any integer k. We remark that the observability rank condition does not require the assumption of constant inputs u; analytic differentiable input functions can be used [36,37]. As noted in [38], this entails that u can be treated symbolically in rank calculations. While identifiability problems can be addressed by a number of techniques not explicitly related to nonlinear observability, it is possible to consider the parameters p as additional states with trivial dynamics p ˙ = 0 and, in this way, the identifiability problem can be recast in the framework of observability [17,21,24]. Thus, by augmenting the state variable vector so as to include model parameters, x ˜ = [ x, p ], we obtain a generalized observability-identifiability matrix, O I (x ˜): O I (x ˜) = ∂ ∂ x ˜ g (x ˜) ∂ ∂ x ˜ (L f g (x ˜) ) ∂ ∂ x ˜ (L f 2 g (x ˜) ) ⋮ ∂ ∂ x ˜ (L f n + q - 1 g (x ˜) ) (7) With this formulation we can define a generalized Observability-Identifiability Condition (OIC) as follows: if the system given by Eq (1) satisfies rank (O I (x ˜ 0) ) = n + q, it is (locally) observable and identifiable in a neighbourhood N (x ˜ 0) of x ˜ 0. Since we have recast the analysis of structural identifiability as a particular case of observability, the same remark that was made in the preceding subsection about the difference between local and global properties applies here. It has been noted [39] that in certain cases a system may become unreachable for specific values of the initial conditions, leading to the impossibility of determining the values of parameters classified as identifiable by structural identifiability methods. This situation can be detected if rank (O I (x ˜ 0) ) is calculated using a vector of specific initial conditions instead of a generic symbolic vector. Finally, we note that the idea of treating parameters and state variables similarly is also adopted by estimation methods such as extended Kalman filtering [40]. However, the context is different, since the goal of such techniques is to determine the value of states and parameters from data, while structural identifiability analysis aims at establishing whether such estimation is theoretically possible. In practice, checking the aforementioned Observability-Identifiability Condition (OIC) is often computationally inefficient (or even infeasible) because building O I and calculating its rank is a highly demanding, memory-consuming task. Fortunately, sometimes this cost can be decreased by building a smaller matrix. Let us first note that each of the n + q sub-matrices vertically stacked in the generalized observability-identifiability matrix of Eq (7) has dimension m × (n + q), and the full matrix O I has dimensions (m ⋅ (n + q) ) × (n + q). Therefore it may not be necessary to calculate the n + q − 1 Lie derivatives in order to test whether O I is full rank, since full rank may be achieved with a lower number of derivatives. The minimum number of Lie derivatives for which the matrix may be full rank is n d = n + q m - 1 (8) that is, the smallest integer not less than (n + q) /m − 1, where n, q, and m are the numbers of states, parameters, and outputs, respectively. The maximum number of Lie derivatives is also known a priori: derivatives of order higher than n + q − 1 cannot increase the matrix rank [38]. Having lower and upper bounds for the necessary Lie derivatives is an advantage of this methodology compared to, e. g. , power series approaches, for which the maximum number of derivatives is in principle infinite [10]. Our method builds O I recursively. Once nd is reached, addition of a new Lie derivative is followed by calculation of the rank. This process is repeated until the maximum number n + q − 1 is reached, or until adding a new Lie derivative does not increase the matrix rank; in both cases no further derivatives are necessary [38]. At that point, if O I is full rank the corresponding model is observable and identifiable, as seen in the previous subsection. If O I is not full rank, the algorithm proceeds to find identifiable parameters, as explained in the following subsection. Further improvements in the computational burden can be obtained by calculating the rank numerically instead of symbolically. A way in which this can be performed is by replacing the symbolic variables in the O I with prime numbers to minimize the risk of accidental cancellations, which would reduce the rank. If O I is not full rank, the Observability-Identifiability Condition (OIC) does not inform us about which parameters are identifiable and which are not. This can be achieved by realizing that each column of O I corresponds to a parameter-to-output relation (or state-to-output): ∂ ∂ x 1 g (x ˜) ∂ ∂ x 2 g (x ˜) ⋯ ∂ ∂ p q g (x ˜) ∂ ∂ x 1 (L f g (x ˜) ) ∂ ∂ x 2 (L f g (x ˜) ) ⋯ ∂ ∂ p q (L f g (x ˜) ) ⋮ ⋮ ⋮ ⋮ ∂ ∂ x 1 (L f n + q - 1 g (x ˜) ) ∂ ∂ x 2 (L f n + q - 1 g (x ˜) ) ⋯ ∂ ∂ p q (L f n + q - 1 g (x ˜) ) Therefore, if deleting the ith column of the generalized observability-identifiability matrix does not change its rank, then the corresponding ith state (parameter) is non-observable (unidentifiable). This fact can be exploited to determine which of the parameters in an unidentifiable model are identifiable and which are not, using a sequential procedure: after the matrix rank has been calculated and the model has been found to be unidentifiable, each of the columns in O I corresponding to a particular parameter is removed one by one and the rank is recalculated. In this way the identifiability of each of the parameters is evaluated. The procedure outlined in the preceding subsections classifies the model parameters as either identifiable of unidentifiable. A question that naturally follows is: are there combinations of the unidentifiable parameters which are themselves identifiable? If the answer is affirmative, the model can be reparameterized and converted to a structurally identifiable model. However, this is a difficult problem, which few methods can address, and only for models of moderate size. An example is COMBOS [41,42], which is based on differential algebra. Here we suggest an approach based on ideas presented in [43,44] and on the method for finding symmetries proposed by [38]; related work has been recently presented in [45]. The procedure is as follows: if O I is rank-deficient, remove the columns corresponding to identifiable parameters and obtain a reduced submatrix, O U [38]. Then, obtain a basis for the kernel (null space) of this matrix, N (O U) (step 2 in [44]). Its coefficients define one or several partial differential equations whose solution (s) are the identifiable combinations (step 3 in [44]). This procedure is illustrated in the Methods section with the JAK/STAT signalling pathway, for which an identifiable combination of two parameters is found. While this example shows the potential of this procedure, it must be acknowledged that the computational complexity of calculating the kernel of O U limits its applicability to models with a moderate number of unidentifiable parameters. The methodology described in the previous subsections can be used to analyse the identifiability of whole models and, if the model is unidentifiable, of its parameters individually. However, since it relies heavily on symbolic operations, it may be computationally infeasible for large or complex models. It should be noted that the main limiting operations are: The minimum number of derivatives necessary for building O I (x ˜) is given by nd as defined in Eq (8). The limit of what is computationally possible is difficult to quantify a priori, since it depends on the model equations and the machine used in the calculations. As a rule of thumb, analyses involving nd ≥ 10 are infeasible except for very small models. As model size or complexity increases, this upper bound decreases; some examples will be shown in the Results section. One solution is to decompose those models into smaller submodels whose analysis is possible computationally. Thus, we seek to decompose a model M into submodels {M1, M2, …} which require few Lie derivatives for their analysis, that is, they have a small nd. Each submodel Msub includes a subset of the model states, xsub. Its outputs, ysub, are the outputs of M which are functions of at least one state included in xsub. The submodel parameters and inputs are those appearing in the equations of xsub and ysub. There may be states that appear in the equations of xsub or ysub but are not part of xsub; they are considered as additional unknown parameters of Msub. The submodels can be found by optimization as follows. For each submodel Mi we select a subset of the states in M by performing a combinatorial optimization where we minimize nd: min s n d (s) (9) where s = {s1, s2, …, sn} is a binary vector of size n, whose entries sj ∈ {0,1} denote inclusion (sj = 1) or exclusion (sj = 0) of the corresponding state. The combinatorial optimization is performed with the Variable Neighbourhood Search metaheuristic [46]. We carry out n optimizations (one per state); in the jth optimization we force sj = 1, so that each state appears in at least one solution. This, in turn, guarantees that all the parameters will eventually be evaluated. A penalty term is included in the objective function to penalize solutions that have more states than a chosen maximum. Apart from this optimization-based decomposition, it may sometimes be useful to specify a particular submodel in order to explore the identifiability of a specific part of the model. Let us clarify how we can conclude identifiability of a parameter from analysis of a submodel. As an example, consider M to be the model of Arabidopsis thaliana described in the Results section; its equations are given in the Supplementary Information (S1 Text). Let us consider a submodel Msub consisting of two states, xsub = {x1, x7}. The equations of Msub are those that correspond to the states {x1, x7}, that is: { x˙1=n1x6ag1a+x6a−m1x1k1+x1+q1x7u (t), x˙7=p3−m7x7k7+x7− (p3+q2x7) u (t), x1 (0) =0, x7 (0) =0 (10) The outputs of Msub are those outputs of M which are functions of at least one of the states in Msub (in this example, y1 = x1). The parameters and inputs of Msub are those present in Eq (10): respectively, {n1, g1, a, m1, k1, q1, p3, m7, k7, q2} and u. Additionally, we must also include as parameters the states that do not belong to xsub but appear in Eq (10) or in ysub (in this case, x6). Thus in this example the submodel parameters would be {n1, g1, a, m1, k1, q1, p3, m7, k7, q2, x6}. By including states such as x6 as parameters we are considering them as unknown and constant. In contrast, if they were included as inputs to the submodel, we would be implicitly assuming that they provide sufficient excitation for identification purposes. Thus, including them as parameters is a conservative assumption in terms of identifiability. Therefore, if a parameter is classified as identifiable in a submodel under these conditions, it will also be identifiable when considering the whole model. When the nd of the full model is so high that it is not feasible to build O I, one solution is to decompose the model into smaller submodels as described in the previous subsections. Another possibility is to build O I with i < nd derivatives. In this case we know that full rank cannot be achieved, so even if the model is identifiable we will not be able to determine it in this way. However, it may be possible to determine identifiability of at least some of the parameters. Note that this procedure can be helpful exactly in the same circumstances as decomposition. In some cases one approach will be more appropriate than the other, but both can be used to determine the identifiability of different parameters, and may therefore be complementary. Fig 1 shows a diagram of the methodology presented so far. The next sections discuss the types of analyses that can be performed with this methodology and show how to refine the solutions iteratively in order to obtain more complete diagnoses. By assessing identifiability as explained in sections “Assessing the OIC efficiently” and “Determining identifiability of individual parameters” we are performing a “Complete Case Analysis” (CCA): every parameter in the model is either classified as identifiable or as unidentifiable. However, it may not always possible to carry out the aforementioned procedure due to computational limitations, as explained in sections “Decomposing large models into submodels to facilitate their analysis” and “Building O I with less than nd Lie derivatives”, which presented two different alternatives. In certain cases these alternatives can yield incomplete results, that is, they may fail to determine the (un) identifiability of some parameters. For example, this may happen in the following scenarios: The two cases mentioned above will be called “Partial Analyses for Identifiability” (PAI): some parameters are conclusively classified as identifiable, but nothing can be said about the rest. It is also possible to perform similar analyses to guarantee unidentifiability of some parameters, leading to what we will call “Partial Analyses for Unidentifiability” (PAU). In such tests, some parameters are classified as unidentifiable while the analysis of the rest is not conclusive. This can happen in at least two situations: The different types of analyses that can be performed are summarized in Table 1. As shown in the preceding subsection, for some complex problems a complete analysis—that is, classifying all the parameters as identifiable or unidentifiable—may not be possible, at least in a first approach, due to computational limitations. In such cases, one can try to obtain more complete diagnoses by running the procedure iteratively. At each time, the computational cost can be reduced by removing from the augmented state vector, x ˜ = [ x, p ], those parameters that were already found to be identifiable in previous steps. This operation, which leads to a smaller O I matrix, does not alter the result of the identifiability test, because the resulting O I is identical to the one obtained with the original vector x ˜ = [ x, p ] after removing the columns corresponding to identifiable parameters—which results in a decreased rank. Note that this equivalence is made possible by the fact that p ˙ = 0, so the procedure cannot be applied to the model states, since x ˙ ≠ 0. In summary, if a model M is too large to be analysed as a whole—i. e. to directly calculate the rank of its identifiability matrix and perform a complete case analysis (CCA) —identifiability analysis can be approached as follows: The present method has been implemented as a MATLAB toolbox named STRIKE-GOLDD (STRuctural Identifiability taKen as Extended-Generalized Observability using Lie Derivatives and Decomposition). It is an open source tool licensed under the GNU General Public License version 3 (GPLv3). It is freely available from https: //sites. google. com/site/strikegolddtoolbox/ and as supplementary information accompanying this article (S1 File). It requires a MATLAB installation with the Symbolic toolbox. Additionally, to use optimization-based decomposition it is necessary to install the MEIGO toolbox [48]. The usage of the STRIKE-GOLDD software is discussed in detail in the manual (S2 Text); in the following lines we provide a brief description of the key options. The toolbox allows limiting the number of Lie derivatives that are calculated when building O I (x ˜). This is useful to prevent the algorithm from getting stuck in excessively lengthy calculations. To adapt this limit to the computer where the algorithm is running, it is specified as a machine-dependent criterion: the user can set a limit on the time invested in calculating these derivatives by entering it in opts. maxLietime (that is, the algorithm will not calculate the ith + 1 derivative if the time spent in obtaining the ith one was ti > opts. maxLietime). Furthermore, the user can choose what to do if this time limit is reached without O I (x ˜) being full rank: by setting opts. decomp = 0, STRIKE-GOLDD will perform a partial analysis of the whole model with the current O I (x ˜); if opts. decomp = 1, it will decompose the model. It is also possible to enforce the use of decomposition from the start, i. e. without checking whether the time limit is reached, with the option opts. forcedecomp. The submodels can be found by optimization or specified by the user; this choice is made by opts. decomp_user. Another option, opts. numeric, allows deciding whether to calculate rank (O I (x ˜) ) numerically or symbolically. The symbolic calculation is always exact. It is possible to perform a numerical calculation by replacing the symbolic variables with prime numbers. This usually reduces the computational cost, although in some cases it might lead to accidental cancellations that decrease the rank artificially. However, the risk of obtaining a spurious result is low, and it can be minimized by running the procedure several times, since the substitutions are random. In all of the case studies analyzed in the Results section we found agreement between numeric and symbolic rank calculations. Finally, it is possible to assess identifiability for specific values of the system’s initial conditions. As mentioned in subsection “Structural identifiability as augmented observability: the OIC”, this can be useful in order to detect situations in which loss of reachability from particular initial conditions leads to loss of identifiability. Such pathological cases are not detected if rank (O I (x ˜ 0) ) is calculated using a generic symbolic vector of initial conditions. However, they can be tested by setting the option opts. knowninitc = 1 and entering the corresponding vector of initial conditions in the script that creates the model.
In [19], Grandjean and coworkers proposed 18 alternative pharmacokinetic nonlinear compartmental models of the uptake process of Pitavastatin (a drug used to treat hypercholesterolaemia) into hepatocytes. They applied five different methods to analyse their structural identifiability: similarity transformation, differential algebra, Taylor series, and two approaches based on a non-differential input/output observable normal form and an algebraic input/output relationship approach. With these techniques they established the identifiability of most of the models. However, for several model formulations none of the methods was able to produce results. This was the case for two candidate models (with or without pseudo-state assumption) that accounted for drug metabolising within the cell. A diagram of these Pitavastatin uptake models is shown in panel A of Fig 2. The upper part of the panel shows the system’s functional diagram. The lower part shows a graph drawn following the same convention as in [47], in which a directed arrow from A to B indicates that B appears in the dynamic equation of A. This graphical approach was originally proposed to study observability, and hence in [47] only the states were shown in the graphs. Since here we use it for identifiability purposes, we extend it to include both states and parameters (see figure caption for more details). The method presented here determines that both Pitavastatin uptake models (with and without pseudo steady state assumption) are structurally identifiable. The classical model of oscillations in enzyme kinetics proposed by Goodwin in 1965 [27] and shown in panel B of Fig 2 is still the subject of analyses [28]. It was selected by [10] to compare the performance of several structural identifiability methods, considering two different scenarios or variations of the model: when the three states are measured, or when only one of them—the enzyme concentration, x1—can be measured. The latter situation is more realistic, but its analysis is particularly challenging, and none of the methods tested by [10] managed to reach a conclusion due to computational complexity. According to Eq (8), the minimum number of Lie derivatives for which the identifiability matrix may be full rank is nd = 10 for this model. While the subsequent rank calculation is very demanding, the computational cost is substantially reduced by building O I (x ˜) with only 9 Lie derivatives. In this way the method classifies four parameters as identifiable: b, σ, β, δ. Then, removing these parameters from the model as explained in “Iterative refinement of the diagnosis” enables the analysis of the remaining parameters (a, A, α, γ), which are found to be unidentifiable. Thus this model is unidentifiable. It can be made identifiable by considering two parameters as known, one from each of the pairs {A, a} and {α, γ}. For example, if we fix the values of {A, α}, the remaining six unknown parameters in the model are identifiable. An alternative solution is to measure more states, if it is experimentally possible. In this case, if all three states are outputs, the model is structurally identifiable. Measuring only two of the three states, however, increases the number of identifiable parameters but does not render the model fully identifiable. The subsets of unidentifiable parameters for y = {x1, x2}, y = {x1, x3}, and y = {x2, x3} are, respectively, {a, A, γ}, {α, γ} and {a, α}. This model was presented in [28] as an example of a system exhibiting both oscillation and bistability. It is a three-layer signalling cascade with positive and negative feedback loops and Hill nonlinearities, shown in panel C of Fig 2. It has three states, which are the phosphorylated forms (x1, x2, x3), and 14 parameters (k1, k2, k3, k4, k5, k6, s1t, s2t, s3t, K1, K2, n1, n2, α). This system requires that all its three states are measured in order to be identifiable. However, if just one of the states is left unmeasured, some parameters become unidentifiable: if x1 is not measured, k3 and s1t are unidentifiable; if x2 is not measured, k5 and s2t are unidentifiable; and if x3 is not measured, K1, K2, and s3t are unidentifiable. This model was presented by [30] and was used by both [10] and [16] as a benchmark for structural identifiability methods. In the formulation of [10], only 13 parameters are considered unknown. In that case, all of them are identifiable. The general case, in which all 29 parameters are in principle unknown, is more challenging. For this case STRIKE-GOLDD classifies 5 parameters as unidentifiable: c1c, c2c, c3c, c4, and k2, and the remaining as identifiable. Part of this diagnosis can be confirmed by inspection of the connection diagram in the right side of Fig 3, which shows that c1c, c2c, and c3c only appear in the equation of state x15. Since x15 is in turn “disconnected” from the rest of the model (i. e. it does not appear in the equation of any other state), and it is not measured, there is clearly no way of determining its value. Hence x15 is unobservable, and the three parameters associated with it are unidentifiable. In contrast, the unidentifiability of c4 and k2 is by no means apparent from the figure. However, it can be determined with the methodology that they are not only unidentifiable, but related: fixing any of the two renders the other one identifiable. In summary, this 29-parameter model can be converted into a structurally identifiable 25-parameter model by fixing the values of four parameters: c1c, c2c, c3c, and either c4 or k2. This model of the IL13-Induced JAK/STAT signalling pathway was presented in [31] and later used in [20] to benchmark three identifiability analysis methods. The network interaction diagrams are shown in panel A of Fig 4. The results of our method coincide with those reported in [20], that is, five of the 23 parameters are unidentifiable, pu = [θ11, θ15, θ17, θ21, θ22]. Following the procedure outlined in the Methods section, it is possible to find an identifiable combination of unidentifiable parameters. To do this we remove the columns corresponding to identifiable parameters and obtain a reduced submatrix, O U. Calculation of a basis of the kernel of O U yields the following vector: v = [ 0,0, - θ 17 θ 22,0, 1 ] (11) which in turn leads to the following PDE: - θ 17 θ 22 · ∂ Φ ∂ θ 17 + ∂ Φ ∂ θ 22 = 0 ⇒ Φ = θ 17 · θ 22 (12) Thus, Φ = θ17 ⋅ θ22 is an identifiable parameter combination. The methodology does not report any combination involving θ11, θ15, θ21. If, additionally, we fix the value of θ11 a priori, we obtain a structurally identifiable model with 21 unknown parameters. The genetic subnetwork that controls the circadian clock in the plant A. thaliana was modelled in [29]; its diagram is shown in Fig 4. This model uses both Michaelis-Menten and Hill kinetics. Two Hill coefficients of transcription (a, b) were considered as known parameters in the original publication [29]. Although it was argued in [29] that there is evidence that b = 2, coefficient a was fixed to a = 1 without experimental evidence. In [10] it was reported that (for the case of a = 1) no global structural identifiability method was capable of successfully analysing the model; at most, identifiability of five parameters was established. While the choice of a = 1 makes the system rational and reduces the problem dimension, here we consider the more general case in which a is an unknown parameter. According to Eq (8), the minimum number of Lie derivatives for which O I (x ˜) may be full rank is very high for this model (nd = 16). This is the same situation as with the previously analysed Goodwin model, that is, the computational cost of the construction and subsequent rank calculation of O I (x ˜) with nd derivatives is too high. Furthermore, we found that the approach adopted for the Goodwin model—building the matrix with less than nd derivatives—was not successful in the case of this example, at least when performed with few derivatives. Hence we turned to the alternative procedure, i. e. decomposing the model using optimization. In this way, identifiability of ten parameters was established: a, k1, k4, m1, m4, n1, n2, q2, r2, and r4. Removing these parameters from the model decreases the number of required derivatives nd to 12, which is still very high; however, building O I (x ˜) with 9 derivatives reports identifiability of an additional parameter, r1. By performing partial analyses for unidentifiability (PAUs) we confirmed that the model is indeed unidentifiable. This can be remedied in several ways. A possible solution is to measure more states, if it is experimentally feasible. In the model it has been assumed that only mRNA concentrations are measured (i. e. states x1 and x4); however, if protein concentrations (i. e. the remaining states) are also measured, then all the parameters become structurally identifiable. Alternatively, if we assume that only the original outputs can be measured, it is possible to obtain an identifiable reformulation of the model by fixing some parameters. For example, choosing fixed values for the five degradation constants that were not found to be identifiable (k2, k3, k5, k6, k7) yields a structurally identifiable model with 23 parameters. This large-scale model was taken from the BioPreDyn-bench collection [32], where it was included as benchmark B4. It models a batch fermentation process for protein production using Chinese Hamster Ovary cells. Its diagrams are shown in Fig 5. It contains 34 states (which are metabolites present in three compartments: fermenter, cytosol, and mitochondria), of which 13 are measured outputs. Its 32 reactions include protein product formation, the Embden-Meyerhof-Parnas pathway (EMP), the TCA cycle, a reduced amino acid metabolism, lactate production, and the electron transport chain. The reactions are modelled using lin-log kinetics [49], resulting in non-rational equations with 117 unknown parameters. While it was noted in [32] that the parameter estimation results suggested practical identifiability issues, possible deficiencies in structural identifiability were not mentioned. Given the size of this model, its analysis is very challenging. Using decomposition it is possible to classify most of the parameters in the model as identifiable. However, we also found that at least four parameters are structurally unidentifiable: they are the parameters numbered 47,48,55, and 57, which correspond to the following kinetic constants (elasticities): {e54, − e55, − e62, e64}. After inspecting the model, we realised that it is possible to rewrite its dynamic equations in such a way that these parameters appear as (e54 + e55) and (e62 + e64); clearly, the individual parameters appearing in these sums are not identifiable. Thus, we replaced these four parameters with two new ones, en1 = e54 + e55 and en2 = e62 + e64. In this way we obtained a new model with 115 parameters, and confirmed that the newly introduced ones are structurally identifiable. Overall, we determined the identifiability of 97 parameters. While we did not manage to assess the identifiability of the remaining 18, we did find that fixing six of them (e. g. {p28, p72, p77, p101, p105, p115}) results in a structurally identifiable model. This action is slightly conservative, since those parameters can in principle be s. l. i. However, since the model has practical identifiability deficiencies [32] (as is typical of models of this type and size [49]), and given that it would be necessary to perform many Lie derivatives to relate these parameters to the model outputs, it is likely that in practice their values will be difficult to estimate. Therefore, fixing a subset of them appears as a reasonable solution. In summary, we found that: (i) this model is structurally unidentifiable, (ii) there exist two identifiable combinations of parameters, which convert 4 unidentifiable parameters into 2 identifiable ones, (iii) of the remaining 113 parameters, at least 95 are identifiable, and (iv) fixing the values of 6 parameters ensures that the remaining 12 (and the model as a whole) are identifiable.
We have presented a methodology for analysing the structural identifiability of dynamic models described by a system of ordinary differential equations. It builds on concepts and techniques originally presented in the context of nonlinear observability analysis. More specifically, it adopts a differential geometry approach, which is based on building an augmented observability matrix—with the parameters considered as additional state variables—and calculating its rank. This formulation, as opposed to other approaches based on differential algebra, allows handling any analytic models, without requiring them to be in rational or polynomial form. If a model is structurally unidentifiable the method determines the identifiability of each parameter individually, by recalculating the matrix rank after removing the corresponding column. Realising that the structural identifiability analysis of nonlinear dynamic models is a challenging task, and that this difficulty increases rapidly with the problem size, our method is geared towards computational efficiency. To this end it includes several algorithmic developments to facilitate the analysis of models of larger size. One is the possibility of decomposing the model into smaller submodels, which can be found by optimization or specified by the user. Another is the calculation of the matrix rank with a reduced number of Lie derivatives. These alternatives lead in some cases to partial analyses, whose result is only conclusive if a parameter is classified as identifiable, but not as unidentifiable (or vice versa, depending on the type of analysis). In these situations the method also allows for an iterative refinement of the diagnosis: by removing parameters already classified as identifiable, the problem size is reduced and more complete analyses are made possible. To facilitate the application of this methodology, we have provided it as a free MATLAB (The MathWorks, Natick, MA) toolbox called STRIKE-GOLDD (STRuctural Identifiability taKen as Extended-Generalized Observability with Lie Derivatives and Decomposition), available under the GNU General Public License from https: //sites. google. com/site/strikegolddtoolbox/. We expect that STRIKE-GOLDD will contribute to fill the gap between the complexity of current systems biology models and their usability, which can be compromised unless structural identifiability is assessed. We have validated the methodology using a set of nonlinear systems biology models whose size and/or complexity make them challenging case studies. They range from a classic model of enzymatic oscillations with 8 parameters proposed by Goodwin in 1965 [27] to a metabolic model of more than 100 parameters published in 2015 [32]. Interestingly, we found structural identifiability issues even in models of relatively small size, such as the aforementioned Goodwin model. Indeed, the results show that identifiability issues are likely to appear in over-parameterized models (with many parameters per state), specially if only few of their states are available for measurement (in order words, if there are few outputs). A large parameter-to-output ratio also implies that the structural identifiability of the model will be difficult to analyse, because it will be necessary to perform many Lie derivative calculations in order to build the augmented observability matrix, thus incurring a high computational cost. Could this common cause mean that the difficulty in analysing a model is a hint of lack of identifiability? We ask this question because we know that, on the other hand, it is possible to analyse models with many parameters as long as sufficient measurements are available. Among the models analysed here, the JAK/STAT pathway had already been studied [20], and for that case our method confirmed previously reported results. In other cases we established the identifiability of systems that had not been analysed before, such as the mixed feedback MAPK pathway [28] or the model of Pitavastatin hepatic uptake (which had been reported to resist analysis when attempted with other methods, although it was suspected that it was identifiable [19]). Perhaps more interestingly, we also found some unidentifiabilities that had not been previously reported. An example is the Goodwin oscillator [27], for which it was established that half of its parameters are structurally unidentifiable. Despite the relatively small size of this model (3 states and 8 parameters), the fact that it is not a rational system, combined with the high parameter-to-output ratio (given that only one of its states is measured) make it a very challenging problem. Similar issues were found in the NF-κB signalling pathway [30] and in the genetic subnetwork of the circadian clock in Arabidopsis thaliana [29]. In these cases it can be noted that the ratio of unidentifiable parameters is larger in models with a lower ratio of measured outputs. Finally, we also detected unidentifiabilities in a recently presented large-scale dynamic model of metabolism of Chinese Hamster Ovary cells [32] with 117 parameters. We have also shown how to turn unidentifiable models into identifiable ones. With the procedure described in this paper it is sometimes possible to combine several unidentifiable parameters into a single identifiable combination. More often the solution is to reparameterize the model by considering some of the unidentifiable parameters as known constants, fixing them to values that appear reasonable according to available knowledge. In this way the remaining unknown parameters are rendered identifiable. Finally, a model can also be made identifiable by increasing the number of its outputs, if it is experimentally possible to measure more of its states. | Advances in computing power have facilitated the development of increasingly larger dynamic models of biological processes, which usually have many unknown parameters. Often times, such models contain parameters that are structurally unidentifiable, i. e. , they cannot be uniquely determined from experiments. Any parameter estimation algorithm will fail when trying to estimate unidentifiable parameters, leading to waste of resources and possibly wrong model predictions. Hence, it is essential to assess structural identifiability before exploiting a model. However, performing such analysis can be hard, especially as models become increasingly complicated. To address this challenge, we developed a methodology for structural identifiability analysis that aims at generality—it can handle any analytic model written as a set of ordinary differential equations—and computational efficiency—it includes features that facilitate the analysis of large systems. We provide an implementation of the methodology as a MATLAB toolbox called STRIKE-GOLDD. We illustrate its applicability to systems biology models of genetic, signalling, metabolic, and pharmacokinetic networks, showing which of them are unidentifiable and how they can be made identifiable. | Abstract
Introduction
Methods
Results
Discussion | medicine and health sciences
biological cultures
brassica
optimization
model organisms
systems science
mathematics
pharmacology
plants
drug metabolism
mapk signaling cascades
research and analysis methods
arabidopsis thaliana
computer and information sciences
nonlinear systems
cho cells
cell lines
pharmacokinetics
systems biology
signal transduction
plant and algal models
cell biology
biology and life sciences
physical sciences
cell signaling
organisms
signaling cascades | 2016 | Structural Identifiability of Dynamic Systems Biology Models | 11,606 | 252 |
Genetic variants that are associated with common human diseases do not lead directly to disease, but instead act on intermediate, molecular phenotypes that in turn induce changes in higher-order disease traits. Therefore, identifying the molecular phenotypes that vary in response to changes in DNA and that also associate with changes in disease traits has the potential to provide the functional information required to not only identify and validate the susceptibility genes that are directly affected by changes in DNA, but also to understand the molecular networks in which such genes operate and how changes in these networks lead to changes in disease traits. Toward that end, we profiled more than 39,000 transcripts and we genotyped 782,476 unique single nucleotide polymorphisms (SNPs) in more than 400 human liver samples to characterize the genetic architecture of gene expression in the human liver, a metabolically active tissue that is important in a number of common human diseases, including obesity, diabetes, and atherosclerosis. This genome-wide association study of gene expression resulted in the detection of more than 6,000 associations between SNP genotypes and liver gene expression traits, where many of the corresponding genes identified have already been implicated in a number of human diseases. The utility of these data for elucidating the causes of common human diseases is demonstrated by integrating them with genotypic and expression data from other human and mouse populations. This provides much-needed functional support for the candidate susceptibility genes being identified at a growing number of genetic loci that have been identified as key drivers of disease from genome-wide association studies of disease. By using an integrative genomics approach, we highlight how the gene RPS26 and not ERBB3 is supported by our data as the most likely susceptibility gene for a novel type 1 diabetes locus recently identified in a large-scale, genome-wide association study. We also identify SORT1 and CELSR2 as candidate susceptibility genes for a locus recently associated with coronary artery disease and plasma low-density lipoprotein cholesterol levels in the process.
Recent large-scale, genome-wide association studies have now delivered a number of novel findings across a diversity of diseases, including age-related macular degeneration [1–3], heart disease [4,5], host control of HIV-1 [6], type I and II diabetes [7,8], and obesity [9]. However, despite this astonishing rate of success, the major challenge still remains to not only confirm that the genes implicated in these studies are truly the genes conferring protection from or risk of disease, but to elucidate the functional roles that these implicated genes play with respect to disease. Most of the genetic association studies reporting novel, highly replicated associations to disease traits do not provide experimental data supporting the putative functional roles a given candidate susceptibility gene may play in disease onset or progression. Even in cases where susceptibility genes are well studied, with well known functions, nailing down how these genes confer disease susceptibility can linger for years, or even decades, as has been the case for genes like ApoE, an Alzheimer disease susceptibility gene identified more than 15 years ago [10]. Complex networks of molecular phenotypes—gene expression (mRNA, ncRNA, miRNA, and so on), protein expression, protein state, and metabolite levels—respond more proximally to DNA variations that lead to variations in disease-associated traits. These intermediate phenotypes respond to variations in DNA that in turn can induce changes in disease associated traits. Because a majority of single nucleotide polymorphisms (SNPs) detected as associated with disease traits from the recent wave of genome-wide association studies (GWASs) do not appear to affect protein sequence, it is likely that these SNPs either regulate gene activity at the transcript level directly or link to other DNA variations involved in this type of regulatory role. Therefore, to uncover the genetic determinants affecting expression in a metabolically active tissue that is relevant to the study of obesity, diabetes, atherosclerosis, and other common human diseases, we profiled 427 human liver samples on a comprehensive gene expression microarray targeting more than 39,000 transcripts, and we genotyped DNA from each of these samples at 782,476 unique SNPs. The relatively large sample size of this study and the large number of SNPs genotyped provided the means to assess the relationship between genetic variants and gene expression with more statistical power than many previous studies allowed [11–13]. A comprehensive analysis of the liver gene expression traits revealed that thousands of these traits are under the control of well-defined genetic loci, with many of the genes having already been implicated in a number of human diseases. Here we demonstrate directly how integrating genotypic and expression data in mouse and human can provide much-needed functional support for candidate susceptibility genes identified in a growing number of genetic loci that have been identified as key drivers of disease from GWASs. Specifically, we highlight how the gene RPS26 and not ERBB3 is most strongly supported by our data as a susceptibility gene for a novel type 1 diabetes (T1D) locus that was recently identified in a large-scale GWAS [14] and subsequently extensively replicated in a number of cohorts [15]. We also identify SORT1 and CELSR2 as candidate susceptibility genes for a locus recently associated with coronary artery disease [16] and plasma low-density lipoprotein (LDL) -cholesterol levels [17,18].
To identify expression quantitative trait loci (eQTL) that have putative cis and trans [19] regulatory effects on the liver gene expression traits, we tested all expression traits for association with each of the SNPs in the analysis SNP set typed in the HLC. The strongest putative cis eQTL for a given expression trait was defined as the SNP most strongly associated with the expression trait over all of the SNPs typed within 1 megabase (Mb) of the transcription start or stop of the corresponding structural gene. The association p-values were adjusted to control for testing of multiple SNPs and expression traits using two different methods: (1) a highly conservative Bonferroni correction method to constrain the study-wise significance level, and (2) an empirical false discovery rate (FDR) method that constrains the overall rate of false positive events. For cis eQTL, we only test for associations to SNPs that are within 1 Mb of the annotated start or stop site of the corresponding structural gene. To achieve a study-wise significance level of 0. 05, the Bonferroni adjusted p-value threshold was computed as, where Ni denotes the number of SNPs tested for trait i, over all 39,280 expression traits tested. At this threshold, 1,350 expression traits corresponding to 1,273 genes were identified. The Bonferroni adjustment method can be conservative when there is dependence among the expression traits and among the SNP genotypes. Given that strong correlation structures exist among expression traits and among SNP genotypes in a givenlinkage disequilibrium (LD) block, the Bonferroni adjustment may be overly conservative. Therefore, we used an empirical FDR method based on permutations that accounts for the correlation structures among the expression traits and among the SNP genotypes. We constrained the empirically determined FDR to be less than 10% (see Methods). At this level, we identified 3,210 expression traits corresponding to 3,043 genes that were significantly associated with at least one SNP near the corresponding gene region (referred to here as a putative cis eQTL). The full list of association results are provided in Table S2. The magnitude of the effects ranged from SNPs that explained roughly 2% of the in vivo expression variation (p ∼ 0. 003) to those that explained roughly 90% of the expression variation (p < 10−65). Several recent studies have been published that examine the extent of genetic control in blood [20–22], brain [23], and adipose [21] gene expression via genetic association testing. In one of these studies [21], we performed the study on human blood and adipose tissues profiled on the same expression platform as the HLC, providing a straightforward way to compare the extent of cis eQTL overlap between blood, adipose, and liver tissues. In our characterization of blood and adipose tissue eQTL, there were 2,573 and 2,789 expression traits, respectively, represented on the microarray used to profile the HLC samples and that gave rise to cis eQTL. Of these, 752 blood and 881 adipose cis eQTL overlapped the set of 3,210 cis eQTL detected in the HLC. Therefore, in both adipose and blood, roughly 30% of the cis eQTL detected in these tissues were also detected in the HLC, confirming that there is significant, common genetic control between tissues. However, these overlaps also highlight that a majority of cis eQTL detected in one tissue may be specific to that tissue, potentially reflecting the genetic control of tissue-specific biological functions. The significance of the trans eQTL in the HLC were also assessed by the Bonferroni method and by constraining the empirically determined FDR to be less than 10%. In the case of trans eQTL, all 782,476 SNPs were tested for association to each of the 39,280 expression traits. Therefore, the Bonferroni adjusted p-value threshold was computed as 0. 05/ (782,476 × 39,280) = 1. 6 × 10−12. At this threshold, 242 expression traits corresponding to 236 genes were significantly associated with a SNP in trans (referred to here as a trans eQTL). On the other hand, by constraining the FDR to be less than 10%, 491 expression traits corresponding to 474 genes were identified as significantly associated with a SNP in trans. For the FDR-computed cis and trans eQTL signatures, the trans eQTL signature was only 15% the size of the cis eQTL signature, consistent with findings in other human genetics of gene expression studies [12,13]. The smaller trans eQTL signature likely reflects a lack of power to detect the small-to-moderate eQTL effects, given the sample size of this study in the context of testing 782,476 SNPs and profiling 39,280 expression traits. Other studies have noted strong heritability estimates for a majority of the expression traits that, when taken together with the small number of associations detected, suggests that expression in general is a complex trait under the control of many loci [21]. With the more stringent threshold required to constrain the FDR in searching for trans eQTL, the magnitude of the trans eSNP effects (mean R2 = 0. 19) was larger than the cis eSNP effects (mean R2 = 0. 14). In this study we used both the Affymetrix and Illumina genotyping platforms, providing for increased power to detect cis and trans eQTL in the HLC compared to the detections achieved using the Affymetrix and Illumina sets independently [49]. Conditional on the sample size and FDR, the Illumina SNP set provided for roughly 15% more eQTLs than the Affymetrix SNP set, corresponding to a 15% increase in the relative power. This increase in power is primarily due to the higher genetic coverage of the Illumina SNP set compared to the Affymetrix SNP set. Further, given the ∼40,000 expression traits profiled in the HLC, we were able to estimate the genetic coverage of the Illumina and Affymetrix SNP sets on a cohort that is independent of the HapMap CEU (Utah residents with ancestry from Northern and Western Europe) subjects. Interestingly, we observed significantly lower genetic coverage (78%) than previously reported (90%) (electronic database: http: //www. cidr. jhmi. edu/download/HumanHap650Y_info. pdf). Finally, in comparing whether more samples or more SNPs enhanced power most dramatically, we found that a modest increase in sample size (19%) had a more profound impact on the power to detect gene expression associations (a 21% increase in this case) than increasing the genetic coverage. These power and genetic coverage issues are fully detailed in a separate report [49]. The cis and trans eQTLs identified from the first pass analysis provide a significantly reduced set of SNPs on which to focus (∼3,700 versus 782,476). The set of SNPs associated with expression (eSNPs) can be considered a functionally validated set, given that the SNPs in this set have been found to associate with biologically relevant control of gene expression. In fact, many of the gene expression traits associated with eSNPs correspond to genes that have previously been associated with many different human diseases (Table S3). For example, BRCA1, a well-known susceptibility gene for breast cancer, and CFH, a susceptibility gene for age-related macular degeneration identified in one of the first published GWASs, are each strongly associated with an eSNP (p = 9. 73 × 10−17 for BRCA1 and p = 6. 94 × 10−22 for CFH) that falls within 1Mb of the corresponding structural gene (Table S3). Genes associated with drug response are also represented in this set. For example, VKORC1, a gene recently associated with warfarin dosing [24], has liver gene expression values that are significantly associated with an eSNP near the 3′ end of the gene (p = 1. 66 × 10−23). To characterize further the effect that this set of eSNPs has on the liver transcriptional network, we re-analyzed the association results by restricting attention to this panel of SNPs. We again constrained the FDR to be less than 10% with respect to the eSNP set and identified an additional 3,053 expression traits, corresponding to 2,838 genes that were significantly associated with at least one of the eSNPs (Table S2). We assessed the significance of this new set of expression traits by randomly sampling five sets of SNPs from the full set of SNPs typed in the HLC, such that the size and minor allele frequency distribution matched that of the eSNP set. For each of the randomly selected SNP sets, we analyzed the associations between all expression traits and SNPs in this set. The maximum number of associations detected in any of the five sets at a 10% FDR was only 20, and the mean detection rate over all sets was 12. This demonstrates well the biological utility of the eSNP set, given that this set is significantly enriched for SNPs that associate with expression traits, beyond the initial set of expression traits that defined the eSNP set, compared to comparable sets of randomly selected SNPs. A number of eQTL hot spots emerged as well in this full set of expression traits, where a given locus was identified as a hot spot if greater than 20 expression traits linked to a single eSNP at the locus. The significance of these hot spots was assessed by permuting the genotypes and examining the distribution of associations in the permuted sets. In each permutation set, we identified the maximum number of traits associated with a single marker over all markers. The mean of the maximum counts over 10 permutation sets was only 12, compared to a maximum of 283 in the observed data (Table S2). Identifying candidate susceptibility genes in regions associated with disease using the proximity of the candidate genes to SNPs in that region may be misleading a lot of the time. For example, from Table S2, for the 3,210 expression traits giving rise to cis eQTLs, only 627 of the corresponding cis eSNPs are located within the corresponding gene region, whereas 1,282 are located downstream of the 3′ untranslated region (UTR) and 1,301 are located upstream of the 5′ UTR. Further, of the cis eSNPs located up- and downstream of the corresponding genes, 490 and 526, respectively, are >100 kb away. That is, greater than 30% of all cis eSNPs fall greater than 100 kb away from the transcription start and stop sites of the corresponding gene. Therefore, at least for expression traits, the nearest SNP rule for inferring genes given an association finding would result in an unacceptably high miss-call rate. Genes with expression values that are strongly associated with variations in DNA provide a different path to elucidate the gene or genes and their respective functions underlying genetic loci associated with disease in a more objective fashion.
Previous studies on the genetics of gene expression in humans have focused primarily on lymphoblastoid cell lines or other blood-derived samples [13,14,17]. We have provided a large-scale assessment of the genetics of gene expression in human liver, a metabolically active tissue that is critical to a number of core biological processes and that plays a role in a number of common human diseases. After profiling 427 human liver samples on a comprehensive gene expression microarray and genotyping the DNA from these samples at greater than one million SNPs, we identified a significant genetic signature underlying the expression of more than 6,000 genes, with many of these genes already implicated as causal for a number of different diseases, including heart disease, breast cancer, inflammatory bowel disease, age-related macular degeneration, schizophrenia, and Alzheimer disease. This set of data highlights the utility of monitoring molecular phenotypes that underlie the higher order clinical states of a system. Whereas the eQTL data in the human liver cohort is valuable in its own right, when integrated with other GWAS data and with genetics of gene expression and clinical data in segregating mouse populations, there is the potential to directly identify experimentally supported candidate susceptibility genes for disease. We demonstrated directly how genetics of gene expression data can complement multiple GWAS datasets by highlighting SORT1 and CELSR2 as candidate susceptibility genes for CAD and LDL cholesterol levels at a recently identified locus associated with CAD [16]. In this instance, the association to LDL cholesterol levels is novel and based on publicly available GWAS data and a mouse cross designed specifically to study lipid and other metabolic syndrome traits. In addition to the CAD locus, we highlighted RPS26 as a candidate susceptibility gene for T1D from a novel, highly replicated T1D locus on Chromosome 12q13, which was identified in a separate GWAS [15]. Not only was the expression of this gene in the HLC strongly associated with the T1D SNP at this locus, but it was observed to operate in a part of the molecular network that is significantly enriched for genes associated with T1D (like HLA-DRB1), whereas the gene inferred as the most likely susceptibility gene at that locus (ERBB3) [15] was not supported by any of our experimental data. Recent studies have demonstrated that ribosomal proteins may be involved in auto-immune diseases like systemic lupus erythematosus [36]. In addition, recent work has demonstrated a connection between endoplasmic reticulum (ER) stress in the cytoplasm and diabetes, where protein unfolding in response to ER stress is hypothesized to disrupt processes associated with diabetes [37]. Given RPS26' s protein translation role as part of the ribosomal complex on the ER, its association to T1D is particularly intriguing. The unfolded protein response has also been linked to inflammation and oxidative stress [38], hence the putative connection between RPS26 and an auto-immune disease like T1D is worthy of further consideration. Cells with high secretory capacity like pancreatic beta cells are also more likely to be susceptible to ER stress, making the link between RPS26 and T1D even more plausible. In fact, previous work has indicated higher ER stress levels in T1D patients [39]. It is important to note that a lack of association between expression traits in the HLC and disease-associated SNPs is not a valid filter for excluding a gene as a candidate disease susceptibility gene, given that variation in a gene leading to disease may affect protein function and not expression, or it may affect expression in a different tissue or under different environmental conditions. However, the approach of analyzing the genetics of gene expression in human populations does provide a more objective view into the functioning of genes in a given disease-associated region. This view has the potential to lead to higher confidence candidates in the absence of direct functional support for any one gene, which is typically the case in GWASs where the SNPs identified have no known functional role. Given the potential that genetics of gene expression studies have to affect our understanding of common human diseases, generating even larger-scale molecular profiling datasets in segregating populations may provide a path to more rapidly elucidating not only the genetic basis of disease, but the impact the genetic basis of disease has on molecular networks that in turn induce variations in disease associated traits.
The HLC was assembled from a total of 780 liver samples (1–2 g) that were acquired from Caucasian individuals from three independent liver collections at tissue resource centers at Vanderbilt University, the University of Pittsburgh, and Merck Research Laboratories (Table S1). The Vanderbilt samples (n = 504) included both postmortem tissue and surgical resections from organ donors and were obtained from the Nashville Regional Organ Procurement Agency (Nashville, Tennessee), the National Disease Research Interchange (Philadelphia, Pennsylvania), and the Cooperative Human Tissue Network (University of Pennsylvania, Ohio State University, and University of Alabama at Birmingham). The Pittsburgh samples were normal postmortem human liver and were obtained through the Liver Tissue Procurement and Distribution System (Dr. Stephen Strom, University of Pittsburgh, Pittsburgh, Pennsylvania). The University of Pittsburgh samples (n = 211) were all postmortem, as were the Merck samples (n = 65), which collected by the Drug Metabolism Department and reported previously [40]. All samples were stored frozen at −80 °C from collection until processing for RNA and DNA; some samples had been stored for over a decade before being processed for this study. Demographic data varied across centers for these samples and were missing in many cases. In cases where age, sex, or ethnicity data were not available in the patient records, we imputed it from the gene expression and/or genotype data (see below). Of the 780 samples collected, high-quality DNA was isolated on 548 samples, and 517 of these were successfully genotyped on the Affymetrix genotyping platform (see Methods below). Of the 517 successfully genotyped samples, high-quality RNA was isolated and successfully profiled on 427 samples. This set of 427 genotyped and expression profiled samples comprised the HLC. Table S1 gives a summary of the demographics and other annotations on the 427 individuals that were successfully genotyped and expression profiled. All counts and descriptive statistics include the imputed data. All samples and patient data were handled in accordance with the policies and procedures of the participating organizations. C57BL/6J (B6) mice were intercrossed with C3H/HeJ (C3H) mice to generate 321 F2 progeny (161 females, 160 males) for the BXH wild type (BXH/wt). C57BL/6J (B6) mice were intercrossed with Castaneus (CAST) mice to generate 442 F2 progeny (276 females, 166 males) for the BXC cross. All mice were maintained on a 12 h light–12 h dark cycle and fed ad libitum. BXH mice were fed Purina Chow (Ralston-Purina) containing 4% fat until 8 wk of age. From that time until the mice were killed at 20 wk, mice were fed a western diet (Teklad 88137, Harlan Teklad) containing 42% fat and 0. 15% cholesterol. BXC mice were fed Purina Chow until 10 wk of age, and then fed western diet (Teklad 88137, Harlan Teklad) for the subsequent 8 wk. Mice were fasted overnight before they were killed. Their livers were collected, flash frozen in liquid nitrogen, and stored in −80 °C prior to RNA isolation. The BXH cross on an ApoE null background (BXH/apoE) was previously described [41]. Briefly, C57BL/6J ApoE null (B6. ApoE–/–) were purchased from Jackson Laboratory. C3H/HeJ ApoE null (C3H. Apo E–/–) were generated by backcrossing B6. ApoE–/– to C3H for ten generations. F1 mice were generated from reciprocal intercrossing between B6. ApoE–/– and C3H. ApoE–/–, and F2 mice were subsequently bred by intercrossing F1 mice. A total of 334 (169 female, 165 male) were bred, and all were fed Purina Chow containing 4% fat until 8 wk of age, and then transferred to western diet containing 42% fat and 0. 15% cholesterol for 16 wk. Mice were killed at 24 wk, and liver, white adipose tissue, and whole brains were immediately collected and flash-frozen in liquid nitrogen. All procedures of housing and treatment of animals were performed in accordance with Institutional Animal Care and Use Committee regulations. Array design and preparation of labeled cDNA and hybridizations to microarrays for the human liver cohort. RNA preparation and array hybridizations were performed at Rosetta Inpharmatics. The custom ink-jet microarrays used in this study were manufactured by Agilent Technologies and consisted of 4,720 control probes and 39,280 noncontrol oligonucleotides extracted from mouse Unigene clusters and combined with RefSeq sequences and RIKEN full-length cDNA clones (Table S4). Liver samples extracted from the 427 Caucasian individuals were homogenized, and total RNA extracted using TRIzol reagent (Invitrogen) according to manufacturer' s protocol. Three micrograms of total RNA was reverse transcribed and labeled with either Cy3 or Cy5 fluorochrome. Purified Cy3 or Cy5 complementary RNA was hybridized to at least two single microarrays with fluor reversal for 24 h in a hybridization chamber, washed, and scanned using a laser confocal scanner. Arrays were quantified on the basis of spot intensity relative to background, adjusted for experimental variation between arrays using average intensity over multiple channels, and fitted to an error model to determine significance (type I error), as previously described [42]. Gene expression is reported as the mean-log ratio relative to the pool derived from 192 liver samples selected for sex balance from the Vanderbilt and Pittsburgh samples, because the RNA from the Merck samples had been amplified at an earlier date. The error model used to assess whether a given gene is significantly differentially expressed in a single sample relative to a pool composed of a randomly selected subset of samples has been extensively described and tested in a number of publications [42–44]. The age, sex, race, center, alcohol use, drug use, and steatosis variables presented in Table S1 were tested for association to the gene expression traits. Only age, sex, race, and center were significantly associated with the expression traits beyond what would be expected by chance. As a result, all gene expression traits were adjusted for these covariates. The lack of association between the expression traits and alcohol use, drug use, and steatosis was somewhat surprising, but may be due to the sparseness of these data, resulting in a lack of power to detect significant associations. Array design and preparation of labeled cDNA and hybridizations to microarrays for the mouse liver and adipose tissue samples. RNA preparation and array hybridizations were again performed at Rosetta Inpharmatics. The custom ink-jet microarrays used in the BXH/wt, BXH/apoE, and BXC crosses were manufactured by Agilent Technologies. The array used for the BXH/apoE and BXH/wt samples consisted of 2,186 control probes and 23,574 noncontrol oligonucleotides extracted from mouse Unigene clusters and combined with RefSeq sequences and RIKEN full-length cDNA clones (Table S5). The array used for the BXC cross consisted of 39,280 noncontrol oligonuceotides again extracted from the mouse Unigene clusters and combined with RefSeq sequences and RIKEN full-length cDNA clones (Table S6). Mouse adipose and liver tissues from all of the crosses were homogenized, and total RNA extracted using Trizol reagent (Invitrogen) according to manufacturer' s protocol. Three micrograms of total RNA was reverse transcribed and labeled with either Cy3 or Cy5 fluorochrome. Labeled complementary RNA (cRNA) from each F2 animal was hybridized against a cross-specific pool of labeled cRNAs constructed from equal aliquots of RNA from 150 F2 animals and parental mouse strains for each of the three tissues for each cross. The hybridizations for the BXH/apoE cross were performed in fluor reversal for 24 h in a hybridization chamber, washed, and scanned using a confocal laser scanner. The hybridizations for the BXH/wt and BXC crosses were performed to single arrays (individuals F2 samples labeled with Cy5 and reference pools labeled with Cy3 fluorochromes) for 24 h in a hybridization chamber, washed, and again scanned using a confocal laser scanner. Arrays were quantified on the basis of spot intensity relative to background, adjusted for experimental variation between arrays using average intensity over multiple channels, and fitted to a previously described error model to determine significance (type I error) [42]. Gene expression measures are reported as the ratio of the mean log10 intensity (mlratio). DNA isolation. DNA isolation was performed at Rosetta Inpharmatics. DNeasy tissue kits from QIAGEN were used to carry out all DNA extractions. For each liver sample, 20–30 mg of liver was placed in a 1. 5-ml microcentrifuge tube along with 80 μl buffer ATL and 20 μl proteinase K. The contents of each tube were then mixed thoroughly by vortexing, followed by incubation at 55 °C until the tissue was completely lysed. Transcriptionally active tissues such as liver and kidney contain high levels of RNA, which will co-purify with genomic DNA. Because RNA-free genomic DNA was required for processing, 4 μl RNase A (100 mg/ml) was added and mixed by vortexing, followed by incubation for 2 min at room temperature before continuing. Samples were then vortexed and 200 μl buffer AL was added to the sample and mixed thoroughly. After 10 min incubation at 70 °C, 200 μl ethanol (96%–100%) was then added and mixed again. The mixture was placed into the DNeasy Mini column and centrifuged at 6,000g (8,000 rpm) for 1 min. The DNeasy Mini spin column was then placed in a new 2-ml collection tube, and 500 μl buffer AW1 was added, followed by placement in a centrifuge for 1 min at 6,000g (8,000 rpm). The DNeasy Mini spin column was then placed in a new 2-ml collection tube again, and 500 μl buffer AW2 was added and centrifuged for 3 min at 20,000g (14,000 rpm) to dry the DNeasy membrane. Then the DNeasy Mini spin column was placed in a clean 1. 5-ml or 2-ml microcentrifuge tube and 200 μl buffer AE was pipetted directly onto the DNeasy membrane. This was incubated at room temperature for 1 min and then centrifuged for 1 min at 6,000g (8,000 rpm) to elute. Two 200-μl elutions were performed followed by ethanol/sodium acetate precipitation and resuspension of the resultant pellet with TE buffer. Genotyping data from the Affymetrix 500K panel. SNP genotyping was performed with the commercial release of the Affymetrix 500K genotyping array. The genotyping was carried out at the Perlegen genotyping facility in Mountain View, California. Genotyping was attempted on 548 samples. 18 samples were unable to be genotyped because of poor DNA quality, and an additional 13 samples were removed after genotyping because their overall call rate did not exceed the 90% cutoff we required. We then applied SNP-wise quality checks on the 517 samples that were successfully genotyped. The Affymetrix 500K array consisted of 500,568 SNPs in total, 429,545 SNPs provided quality data from the genotyping assay, and we rejected those SNPs with a call rate < 75%, resulting in a final panel of 393,494 SNPs. We further filtered out SNPs with minor allele frequencies < 4% (81,646 SNPs) or SNPs that deviated from Hardy-Weinberg equilibrium (p < 10−4; 1,104 SNPs). The resulting set of 310,744 SNPs were used to carry out tests for association to the liver gene expression traits in the HLC. Genotyping data from the lllumina 650Y panel. SNP genotyping was performed on the same set of samples that were genotyped on the Affymetrix 500K panel using the Sentrix humanHap650Y genotyping beadchip from Illumina. The genotyping was carried out at the Illumina genotyping facility in La Jolla, California. This chip consists of 655,352 tag SNP markers derived from the International HapMap Project (http: //www. hapmap. org) on a single BeadChip, with ∼100,000 Yoruba-specific tag SNPs to provide more comprehensive coverage in African and African-American populations. Genotyping was attempted on 517 samples. A total of 497 samples were genotyped successfully, and 654,069 SNP assays genotyped successfully. The same genotype quality control measures applied to the Affymetrix 500K dataset were applied to Illumina HumanHap 650Y dataset to determine the analysis set. The sample set for analysis (n = 397) was restricted to those identified or imputed as Caucasian. Of the 397 samples we attempted to genotype, 13 failed the Illumina genotyping assay (overall call rate < 75%), resulting in a set of 384 genotyped samples carried forward for the expression analysis. In total, 652,648 SNPs were called, with only two SNPs rejected because the call rate was <75%. We then sequentially removed 94,915 SNPs with MAF <4% and 491 SNPs that deviated from the Hardy-Weinberg equilibrium (p < 10−4). The resulting set of 557,240 SNPs was used to carry out tests for association to the liver gene expression traits in the HLC. A total of 85,508 SNPs were represented in both the Illumina and Affymetrix SNP sets. Therefore, there were 782,476 unique SNPs successfully genotyped in the HLC such that the call rate was greater than 75%, the MAF >4%, and there was not significant deviation from Hardy-Weinberg equilibrium at the 0. 0001 significance level. The sample set for analysis was restricted to the 427 HLC samples that had both genotype and gene expression data available, passed the criteria outlined above and those that were identified as Caucasian, or imputed to be Caucasian when data was missing (see below). Sex confirmation. Sex identifiers were available for most of the liver samples obtained from the three study centers. We independently confirmed the sex of each individual providing a liver sample by two methods. First, we looked for expression of Y-specific genes in the liver gene expression based on three probes representing three distinct transcripts. Second, we scored heterozygosity of X-chromosome markers. We excluded any individual for which there was a discrepancy in any of the three measures of sex in order to ensure a coherent data set for analysis and that we had excluded as many potential cases of annotation or sample-handling errors as possible. For samples where sex was not noted in the records, we imputed the sex call if both the genotype and gene-expression data were concordant. Ethnicity. Ethnicities were confirmed or imputed using STRUCTURE [45]. A panel of 106 autosomal markers was randomly selected from around the genome to be unlinked and ancestry informative. Markers were selected from the HapMap data [46] that were present on the Affy 500K panel such that the minor allele frequency was >0. 05 and the absolute allele frequency difference in the Caucasians and African Americans ∼0. 5, with average minor allele frequency 0. 5 (standard deviation = 12). Several K were tested (K = 1–6) with burn-in 100,000 and 100,000 reps of MCMC before any information was collected. In all cases, the greatest support was for K = 2. Admixture was detected for some individuals in some runs and some individuals were reclassified. For those unknown and reclassified, population reassignment was made if the probability of group membership was >0. 9 for that individual. This resulted in 469 individuals assigned to the Caucasian group, 28 individuals assigned to the African descent or African American group, and 18 individuals assigned as “unknown”. The data set for further analysis was restricted to Caucasian samples. Age. Ages were imputed using the Elastic Net method [47]. This method performs model selection and parameter estimation in a manner that is a combination of ridge-regression and the lasso. The prediction method is also explained in [47]. For computational reasons, λ was set to zero, in which case the Elastic Net method reduces to the lasso method. For most applications, experience demonstrates that the optimal value for λ is zero or quite near zero. Ages were imputed using separate models for each data source, due to evidence of a source effect, and each sex separately. In cases where the sex was missing or the reported sex was different from the sex implied by the expression data, the sex implied by the expression data was used. This was done so that in the case the annotation data and expression data were mismatched, the imputed age would correspond to the data used to predict it. The 5,000 genes with the highest correlation to age were used as potential regressors. Cross-validation was used to select the number of steps in the model selection procedure. The number of predictors in the model was between 67 and 76 for the four different models. The percentage of variation explained in the training set is quite high (97%–99%) for three of the models. For the fourth, the model for Vanderbilt females, the percentage of variation explained was slightly lower, 0. 92. This is a vast improvement over more naïve imputation methods that are used when adjusting for covariates with missing data, where mean values of the nonmissing data are used to fill in the missing values. Very few of the predictors we constructed were common between the different models. Given the number of predictors with high correlation to age, this is not surprising. Nonetheless, within a given data source (i. e. , Pittsburgh or Vanderbilt samples), the male model is a reasonable predictor for the ages of the females and vice-versa. This same trend did not hold for predicting the ages of same-sex individuals across data sources. Expression trait processing. Expression traits were adjusted for age, sex, and medical center. Residuals were computed using rlm function from R statistical package (M-estimation with Tukey' s bisquare weights). In examining the distributions of the mean log ratio measures for each expression trait in the HLC set, we noted a high rate of outliers. As a result, we used robust residuals and nonparametric tests to carry out the association analyses in the HLC. For each expression trait, residual values deviating from the median by more than three robust standard deviations were filtered out as outliers. Genome-wide eQTL association analysis. The Kruskal-Wallis test was used to determine association between adjusted expression traits and genotypes. We chose this nonparametric method because of its robust nature to underlying genetic model and trait distribution. p-Values were computed using nag_mann_whitney (for loci with two observed genotypes) and nag_kruskal_wallis_test (for loci with three observed genotypes) routines from NAG C library (http: //www. nag. co. uk). We used FDR for multiple-test correction. FDR was estimated as the ratio of the average number of eQTLs found in datasets with randomized sample labels to the number of eQTLs identified in the original data set. Since the number of tests was large (∼1,010), we found the empirical null distribution was very stable and three permutation runs were sufficient for convergence to estimate FDR. FDR computation was performed separately for cis (<1 Mb probe to SNP distance) and trans associations resulting in nominal p-value cutoffs of 5. 0 × 10−5 and 1. 0 × 10−8 for cis and trans eQTLs, respectively. Targeted set association analysis. The 3,346 SNPs identified in the first round of analysis as associating with expression traits in cis at an FDR < 0. 1 were picked for a second round of analysis. To assess the significance of the resulting set of expression traits detected as associated with this set of SNPs, sets of randomly selected SNPs of size 3,346 with MAF distributions identical to the original set were generated. All sets of SNPS were then analyzed using the same method described above for genome-wide associations. Identifying differentially expressed genes. To assess whether a gene in a given sample was differentially expressed, we used a previously described and validated error model for testing whether the mean log ratio of the intensity measures between the experiment and reference channels was significantly different from zero [42,43,48]. Based on this error model we obtained p-values for each of the individual gene expression measures in each sample as previously described [33]. We then computed the standard deviation of –log10 of the p-value for each gene expression measure over all samples profiled for a given tissue, and then rank ordered all of the genes profiled in each tissue based on this standard deviation value (rank ordered in descending order). Genes that fall at the top of this rank ordered list can be considered as the most differentially expressed or variable genes in the study. We have previously shown that this type of ordering approach well captures the most active genes in a set of samples [33]. For demonstrating the number of genome-wide significant eQTLs and eSNPs as a function of differential gene expression, we binned the expression traits into quartiles (Q1-Q4) based on the rank-ordered gene list, with each bin containing 10,025 genes and the bins increasing in significance with respect to differential expression, from Q1 to Q4. Visualization of networks. Networks were visualized using the Target Gene Information (TGI) Network Analysis and Visualization (NAV) desktop application developed at Rosetta Inpharmatics. This tool enables rapid, real-time, graphical analysis of pathway network models built from a comprehensive and fully integrated set of public and proprietary interaction databases available through a back-end central database, described in detail in a separate report. Additionally, the TGI NAV tool supports experimentally generated systems biology data such as the statistical associations and causal relationships described here. TGI NAV enables integration and visualization of orthogonal data sets using network models as a framework and facilitates dissection of networks into smaller, functionally significant subnetworks amenable to biological interpretation. To construct the local networks for H2-Eb1, Erbb3, and Rps26, the whole-gene probabilistic causal networks were loaded into the database and the TGI NAV tool was used to extract all edges from this network involving the central gene of interest. In the case of the Erbb3 network, the local network was expanded by extracting all additional edges involving any genes directly connected to Erbb3. Note that while the underlying networks describe causal relationships between transcripts, TGI NAV was used to translate this network into the space of genes using an integrated mapping database that clusters transcripts into gene models utilizing their genomic coordinates. As a result, multiple causal relationships between gene pairs can be observed in cases where multiple transcripts for a single gene were profiled. Visualization properties of nodes (e. g. , color) are specified in TGI NAV either for individual nodes, or in a data-driven manner by associating attributes, such as KEGG pathway membership, with groups of nodes and mapping visualization properties to these attributes.
All microarray data associated with the HLC have been deposited into the Gene Expression Ominbus database under accession number GSE9588. | Genome-wide association studies seek to identify regions of the genome in which changes in DNA in a given population are correlated with disease, drug response, or other phenotypes of interest. However, changes in DNA that associate with traits like common human diseases do not lead directly to disease, but instead act on intermediate, molecular phenotypes that in turn induce changes in the higher-order disease traits. Therefore, identifying molecular phenotypes that vary in response to changes in DNA that also associate with changes in disease traits can provide the functional information necessary to not only identify and validate the susceptibility genes directly affected by changes in DNA, but to understand as well the molecular networks in which such genes operate and how changes in these networks lead to changes in disease traits. To enable this type of approach we profiled the expression levels of 39,280 transcripts and genotyped 782,476 SNPs in 427 human liver samples, identifying thousands of DNA variants that strongly associated with liver gene expression. These relationships were then leveraged by integrating them with genotypic and expression data from other human and mouse populations, leading to the direct identification of candidate susceptibility genes corresponding to genetic loci identified as key drivers of disease. Our analysis is able to provide much needed functional support for these candidate susceptibility genes. | Abstract
Introduction
Results
Discussion
Materials and Methods
Supporting Information | diabetes and endocrinology
computational biology
genetics and genomics
cardiovascular disorders | 2008 | Mapping the Genetic Architecture of Gene Expression in Human Liver | 10,561 | 288 |
To form a percept of the environment, the brain needs to solve the binding problem—inferring whether signals come from a common cause and are integrated or come from independent causes and are segregated. Behaviourally, humans solve this problem near-optimally as predicted by Bayesian causal inference; but the neural mechanisms remain unclear. Combining Bayesian modelling, electroencephalography (EEG), and multivariate decoding in an audiovisual spatial localisation task, we show that the brain accomplishes Bayesian causal inference by dynamically encoding multiple spatial estimates. Initially, auditory and visual signal locations are estimated independently; next, an estimate is formed that combines information from vision and audition. Yet, it is only from 200 ms onwards that the brain integrates audiovisual signals weighted by their bottom-up sensory reliabilities and top-down task relevance into spatial priority maps that guide behavioural responses. As predicted by Bayesian causal inference, these spatial priority maps take into account the brain’s uncertainty about the world’s causal structure and flexibly arbitrate between sensory integration and segregation. The dynamic evolution of perceptual estimates thus reflects the hierarchical nature of Bayesian causal inference, a statistical computation, which is crucial for effective interactions with the environment.
In our natural environment, our senses are exposed to a barrage of sensory signals: the sight of a rapidly approaching truck, its looming motor noise, the smell of traffic fumes. How the brain effortlessly merges these signals into a seamless percept of the environment remains unclear. The brain faces two fundamental computational challenges: First, we need to solve the ‘binding’ or ‘causal inference’ problem—deciding whether signals come from a common cause and thus should be integrated or instead be treated independently [1,2]. Second, when there is a common cause, the brain should integrate signals taking into account their uncertainties [3,4]. Hierarchical Bayesian causal inference provides a rational strategy to arbitrate between sensory integration and segregation in perception [2]. Bayesian causal inference explicitly models the potential causal structures that could have generated the sensory signals—i. e. , whether signals come from common or independent sources. In line with Helmholtz’s notion of ‘unconscious inference’, the brain is then thought to invert this generative model during perception [5]. In case of a common signal source, signals are integrated weighted in proportion to their relative sensory reliabilities (i. e. , forced fusion [3,4, 6–10]). In case of independent sources, they are processed independently (i. e. , full segregation [11,12]). Iin a particular instance, the brain does not know the world’s causal structure that gave rise to the sensory signals. To account for this causal uncertainty, a final estimate (e. g. , object’s location) is obtained by averaging the estimates under the two causal structures (i. e. , common versus independent source models) weighted by each causal structure’s posterior probability—a strategy referred to as model averaging (for other decisional strategies, see [13]). A large body of psychophysics research has demonstrated that human observers combine sensory signals near-optimally as predicted by Bayesian causal inference [2,13–16]. Most prominently, when locating events in the environment, observers gracefully transition between sensory integration and segregation as a function of audiovisual spatial disparity [12]. For small spatial disparities, they integrate signals weighted by their reliabilities, leading to cross-modal spatial biases [17]; for larger spatial disparities, audiovisual interactions are attenuated. A recent functional MRI (fMRI) study showed how Bayesian causal inference is accomplished within the cortical hierarchy [14,16]: While early auditory and visual areas represented the signals on the basis that they were generated by independent sources (i. e. , full segregation), the posterior parietal cortex integrated sensory signals into one unified percept (i. e. , forced fusion). Only at the top of the cortical hierarchy, in anterior parietal cortex, the uncertainty about the world’s causal structure was taken into account and signals were integrated into a spatial estimate consistent with Bayesian causal inference. The organisation of Bayesian causal inference across the cortical hierarchy raises the critical question of how these neural computations unfold dynamically over time within a trial. How does the brain merge spatial information that is initially coded in different reference frames and representational formats? Whereas the brain is likely to recurrently update all spatial estimates by passing messages forwards and backwards across the cortical hierarchy [18–20], the unisensory estimates may to some extent precede the computation of the Bayesian causal inference estimate. To characterise the neural dynamics of Bayesian causal inference, we presented human observers with auditory, visual, and audiovisual signals that varied in their spatial disparity in an auditory and visual spatial localisation task while recording their neural activity with electroencephalography (EEG). First, we employed cross-sensory decoding and temporal generalisation matrices [21] of the unisensory auditory and visual signal trials to characterise the emergence and the temporal stability of spatial representations across the senses. Second, combining psychophysics, EEG, and Bayesian modelling, we temporally resolved the evolution of unisensory segregation, forced fusion, and Bayesian causal inference in multisensory perception.
Combining psychophysics, multivariate EEG pattern decoding, and computational modelling, we next investigated the computational principles and neural dynamics underlying audiovisual integration of spatial representations using a general linear model (GLM) -based wAV and a Bayesian modelling analysis. As shown in Fig 3, both analyses were applied to the spatial estimates that were either reported by participants (i. e. , behaviour, Fig 3B left) or decoded from EEG activity patterns independently for each poststimulus time point (i. e. , neural, Fig 3B right, for further details, see the Methods section and the Fig 3 legend). The GLM-based wAV analysis quantifies the influence of the true auditory and true visual location on (1) the reported or (2) EEG decoded auditory and visual spatial estimates in terms of an audiovisual weight index wAV. The Bayesian modelling analysis formally assessed the extent to which (2) the full-segregation model (s) (Fig 3C, encircled in light blue, red or green), (2) the forced-fusion model (Fig 3C, yellow), and (3) the Bayesian causal inference model (i. e. , using model averaging as decision function, encircled in dark blue; see supporting material S1 Table for other decision functions) can account for the spatial estimates reported by observers (i. e. , behaviour) or decoded from EEG activity pattern (i. e. , neural).
Integrating information from vision and audition into a coherent representation of the space around us is critical for effective interactions with the environment. This EEG study temporally resolved the neural dynamics that enable the brain to flexibly integrate auditory and visual signals into spatial representations in line with the predictions of Bayesian causal inference. Auditory and visual senses code spatial location in different reference frames and representational formats [26]. Vision provides spatial information in eye-centred and audition in head-centred reference frames [27,28]. Furthermore, spatial location is directly coded in the retinotopic organisation in primary visual cortex [29], whereas spatial location in audition is computed from sound latency and amplitude differences between the ears, starting in the brainstem [27]. In auditory cortices of primates, spatial location is thought to be represented by neuronal populations with broad tuning functions [30,31]. In order to merge spatial information from vision and audition, the brain thus needs to establish coordinate mappings and/or transform spatial information into partially shared ‘hybrid’ reference frames, as previously suggested by neurophysiological recordings in nonhuman primates [30,32]. In the first step, we therefore investigated the neural dynamics of spatial representations encoded in EEG activity patterns separately for unisensory auditory and visual signals using the method of temporal generalisation matrices [21]. In vision, spatial location was encoded initially at 60 ms in transient neural activity associated with the early P1 and N1 components and then turned into temporally more stable representations from 200 ms and particularly from 350 ms (Fig 2, upper right quadrant, S2 Fig). In audition, spatial location was encoded by relatively stable EEG activity from 95 ms and particularly from 250 ms, which is associated with the auditory long latency P2 component [22–24] (S3 Fig). Activity patterns encoding spatial location generalised not only across time but also across sensory modalities between 160 and 360 ms. As indicated in Fig 2, SVR models trained on visual-evoked responses generalised to auditory-evoked responses and vice versa (upper left and lower right quadrant, significant cross-sensory generalisation encircled by thick grey line). These results suggest that unisensory auditory and visual spatial locations are initially represented by transient and modality-specific activity patterns. Later, at about 200 ms, they are transformed into temporally more stable representations that may rely on neural sources in frontoparietal cortices that are at least to some extent shared between auditory and visual modalities [22,33,34]. Next, we asked when and how the human brain combines spatial information from vision and audition into a coherent representation of space. The brain should integrate sensory signals only when they come from a common event but should segregate signals from independent events [1,2, 12]. To investigate how the brain arbitrates between sensory integration and segregation, we presented observers with synchronous audiovisual signals that varied in their spatial disparity across trials. On each trial, observers reported either the auditory or the visual location. Our results show that a concurrent yet spatially disparate visual signal biased observers’ perceived sound location towards the visual location—a phenomenon coined spatial ventriloquist illusion [17,35]. Consistent with reliability-weighted integration, this audiovisual spatial bias was significantly stronger when the visual signal was more reliable (Fig 1C left, grey solid versus dashed lines). Furthermore, observers reported different locations for auditory and visual signals, and this difference was even greater for large- relative to small-spatial-disparity trials. This significant interaction between spatial disparity and task relevance indicates that human observers arbitrate between sensory integration and segregation depending on the probabilities of different causal structures of the world that can be inferred from audiovisual spatial disparity. Using EEG, we then investigated how the brain forms neural spatial representations dynamically post stimulus. Our analysis of the neural audiovisual weight index wAV shows that the spatial estimates decoded from EEG activity patterns are initially dominated by visual inputs (i. e. , wAV close to 90°). This visual dominance is most likely explained by the retinotopic representation of visual space that facilitates EEG decoding of space leading to visual predominance (for further discussion, see the Methods section). From about 65 ms onwards, visual reliability significantly influenced wAV (Fig 4A): as expected, the location of the visual signal exerted a stronger influence on the spatial estimate decoded from EEG activity patterns when the visual signal was more reliable than unreliable. By contrast, the signal’s task relevance influenced the audiovisual weight index only later, from about 190 ms (Fig 4B). Thus, visual reliability as a bottom-up stimulus-bound factor impacted the sensory weighting in audiovisual integration prior to top-down effects of task relevance. We observed a significant interaction between task relevance and spatial disparity as the characteristic profile for Bayesian causal inference from about 310 ms: the difference in wAV between auditory and visual report was significantly greater for large- than for small-disparity trials (Fig 4D, Table 2). Thus, spatial disparity determined the influence of task-irrelevant signals on the spatial representations encoded in EEG activity from about 310 ms onwards. A task-irrelevant signal influenced the spatial representations mainly when auditory and visual signals were close in space and hence likely to come from a common event, but it had minimal influence when they were far apart in space. Collectively, our statistical analysis of the audiovisual weight index revealed a sequential emergence of visual dominance, reliability weighting (from about 100 ms), effects of task relevance (from about 200 ms), and finally the interaction between task relevance and spatial disparity (from about 310 ms, Fig 4A–4D). This multistage process was also mirrored in the time course of exceedance probabilities furnished by our formal Bayesian model comparison: The unisensory visual segregation (SegV) model was the winning model for the first 100 ms, thereby modelling the early visual dominance. The audiovisual forced-fusion model embodying reliability-weighted integration dominated the time interval of 100–250 ms. Finally, the Bayesian causal inference model that enables the arbitration between sensory integration and segregation depending on spatial disparity outperformed all other models from 350 ms onwards. Hence, both our Bayesian modelling analysis and our wAV analysis showed that the hierarchical structure of Bayesian causal inference is reflected in the neural dynamics of spatial representations decoded from EEG. The Bayesian causal inference model also outperformed the audiovisual full-segregation (SegV, A) model that enables the representation of the location of the task-relevant stimulus unaffected by the location of the task-irrelevant stimulus. Instead, our Bayesian modelling analysis confirmed that from 350 ms onwards, the brain integrates audiovisual signals weighted by their bottom-up reliability and top-down task relevance into spatial priority maps [36,37] that take into account the probabilities of the different causal structures consistent with Bayesian causal inference. The spatial priority maps were behaviourally relevant for guiding spatial orienting and actions, as indicated by the correlation between the neural and behavioural audiovisual weight indices, which progressively increased from 100 ms and culminated at about 300–400 ms. Two recent studies have also demonstrated such a temporal evolution of Bayesian causal inference in an audiovisual temporal numerosity judgement task [38] and an audiovisual rate categorisation task [39]. The timing and the parietal-dominant topographies of the AV potentials (see S2 and S3 Figs) that form the basis for our spatial decoding (and hence for wAV and Bayesian modelling analyses) closely match the P3b component (i. e. , a subcomponent of the classical P300). Although it is thought that the P3b relies on neural generators located mainly in parietal cortices [40,41], its specific functional role remains controversial [42]. Given its sensitivity to stimulus probability [43–45] and discriminability [46] as well as task context [42,47,48], it was proposed to reflect neural processes involved in transforming sensory evidence into decisions and actions [49]. Most recent research has suggested that the P3b may sustain processes of evidence accumulation [50] that are influenced by observers’ prior [51], incoming evidence (i. e. , likelihood [52]), and observers’ belief updating [53]. Likewise, our supplementary time-frequency analyses revealed that alpha/beta power, which has previously been associated with the generation of the P3b component [54], depended on bottom-up visual reliability between 200 and 400 ms and top-down task relevance between 350 and 550 ms post stimulus (see S5 Fig, S2 Table and S1 Text), thereby mimicking the temporal evolution of bottom-up and top-down influences observed in our main wAV and Bayesian modelling analysis. Yet, our main analysis took a different approach. Rather than focusing on the effects of visual reliability, task relevance/attention, and spatial disparity directly on event-related potentials (ERPs) or time-frequency power, the wAV analysis investigated how these manipulations affect the spatial representations encoded in EEG activity patterns, and the Bayesian modelling analysis accommodated those effects directly in the computations of Bayesian causal inference. Along similar lines, two recent fMRI studies characterised the computations involved in integrating audiovisual spatial inputs across the cortical hierarchy [14,16]: whereas low level auditory and visual areas predominantly encoded the unisensory auditory or visual locations (i. e. , full-segregation model) [55–64], higher-order visual areas and posterior parietal cortices combined audiovisual signals weighted by their sensory reliabilities (i. e. , forced-fusion model) [65–68]. Only at the top of the hierarchy, in anterior parietal cortices, did the brain integrate sensory signals consistent with Bayesian causal inference. Thus, the temporal evolution of Bayesian causal inference observed in our current EEG study mirrored its organisation across the cortical hierarchy observed in fMRI. Fusing the results from EEG and fMRI studies (see caveats in the Methods section) thus suggests that Bayesian causal inference in multisensory perception relies on dynamic encoding of multiple spatial estimates across the cortical hierarchy. During early processing, multisensory perception is dominated by full-segregation models associated with activity in low-level sensory areas. Later audiovisual interactions that are governed by forced-fusion principles rely on posterior parietal areas. Finally, Bayesian causal inference estimates are formed in anterior parietal areas. Yet, although our results suggest that full segregation, forced fusion, and Bayesian causal inference dominate EEG activity patterns at different latencies, they do not imply a strictly feed-forward architecture. Instead, we propose that the brain concurrently accumulates evidence about the different spatial estimates and the underlying causal structure (i. e. , common versus independent sources) most likely via multiple feedback loops across the cortical hierarchy [18,19]. Only after 350 ms is a final perceptual estimate formed in anterior parietal cortices that takes into account the uncertainty about the world’s causal structure and combines audiovisual signals into spatial priority maps as predicted by Bayesian causal inference.
Sixteen right-handed participants participated in the experiment; three of those participants did not complete the entire experiment: two participants were excluded based on eye tracking results from the first day (the inclusion criterion was less than 10% of trials rejected because of eye blinks or saccades; see the Eye movement recording and analysis section for details), and one participant withdrew from the experiment. The remaining 13 participants (7 females, mean age = 22. 1 years; SD = 3. 0) completed the 3-day experiment and are thus included in the analysis. All participants had no history of neurological or psychiatric illnesses, had normal or corrected-to-normal vision, and had normal hearing. All participants gave informed written consent to participate in the experiment. The study was approved by the research ethics committee of the University of Birmingham (approval number: ERN_11_0470AP4) and was conducted in accordance with the principles outlined in the Declaration of Helsinki. The visual (‘V’) stimulus was a cloud of 20 white dots (diameter = 0. 43° visual angle, stimulus duration: 50 ms) sampled from a bivariate Gaussian distribution with vertical standard deviation of 2° and horizontal standard deviation of 2° or 12° visual angle presented on a dark grey background (67% contrast). Participants were told that the 20 dots were generated by one underlying source in the centre of the cloud. The visual cloud of dots was presented at one of four possible locations along the azimuth (i. e. , −10°, −3. 3°, 3. 3°, or 10°). The auditory (‘A’) stimulus was a 50-ms-long burst of white noise with a 5-ms on/off ramp. Each auditory stimulus was delivered at a 75-dB sound pressure level through one of four pairs of two vertically aligned loudspeakers placed above and below the monitor at four positions along the azimuth (i. e. , −10°, −3. 3°, 3. 3°, or 10°). The volumes of the 2 × 4 speakers were carefully calibrated across and within each pair to ensure that participants perceived the sounds as emanating from the horizontal midline of the monitor. In a spatial ventriloquist paradigm, participants were presented with synchronous, spatially congruent or disparate visual and auditory signals (Fig 1A and 1B). On each trial, visual and auditory locations were independently sampled from four possible locations along the azimuth (i. e. , −10°, −3. 3°, 3. 3°, or 10°), leading to four levels of spatial disparity (i. e. , 0°, 6. 6°, 13. 3°, or 20°; i. e. , as indicated by the greyscale in Fig 1A). In addition, we manipulated the reliability of the visual signal by setting the horizontal standard deviation of the Gaussian cloud to a 2° (high reliability) or 14° (low reliability) visual angle. In an intersensory selective-attention paradigm, participants reported either their auditory or visual perceived signal location and ignored signals in the other modality. For the visual modality, they were asked to determine the location of the centre of the visual cloud of dots. Hence, the 4 × 4 × 2 × 2 factorial design manipulated (1) the location of the visual stimulus (−10°, −3. 3°, 3. 3°, 10°; i. e. , the mean of the Gaussian), (2) the location of the auditory stimulus (−10°, −3. 3°, 3. 3°, 10°), (3) the reliability of the visual signal (2°, 14°; SD of the Gaussian), and (4) task relevance (auditory-/visual-selective report), resulting in 64 conditions (Fig 1A). To characterise the computational principles of multisensory integration, we reorganised these conditions into a 2 (visual reliability: high versus low) × 2 (task relevance: auditory versus visual report) × 2 (spatial disparity: ≤6. 6° versus >6. 6°) factorial design for the statistical analysis of the behavioural and EEG data. In addition, we included 4 (locations: −10°, −3. 3°, 3. 3°, or 10°) × 2 (visual reliability: high, low) unisensory visual conditions and 4 (locations: −10°, −3. 3°, 3. 3°, or 10°) unisensory auditory conditions. We did not manipulate auditory reliability, because the reliability of auditory spatial information is anyhow limited. Furthermore, the manipulation of visual reliability is sufficient to determine reliability-weighted integration as a computational principle and arbitrate between the different multisensory integration models (see Bayesian modelling analysis section). On each trial, synchronous audiovisual, unisensory visual, or unisensory auditory signals were presented for 50 ms, followed by a response cue 1,000 ms after stimulus onset (Fig 1B). The response was cued by a central pure tone (1,000 Hz) and a blue colour change of the fixation cross presented in synchrony for 100 ms. Participants were instructed to withhold their response and avoid blinking until the presentation of the cue. They fixated on a central cross throughout the entire experiment. The next stimulus was presented after a variable response interval of 2. 6–3. 1 s. Stimuli and conditions were presented in a pseudo-randomised fashion. The stimulus type (bisensory versus unisensory) and task relevance (auditory versus visual) was held constant within a run of 128 trials. This yielded four run types: audiovisual with auditory report, audiovisual with visual report, auditory with auditory report, and visual with visual report. The task relevance of the sensory modality in a given run was displayed to the participant at the beginning of the run. Furthermore, across runs we counterbalanced the response hand (i. e. , left versus right hand) to partly dissociate spatial processing from motor responses. The order of the runs was counterbalanced across participants. All conditions within a run were presented an equal number of times. Each participant completed 60 runs, leading to 7,680 trials in total (3,840 auditory and 3,840 visual localisation tasks—i. e. , 96 trials for each of the 76 conditions were included in total; apart from the four unisensory auditory conditions that included 192 trials). The runs were performed across 3 days with 20 runs per day. Each day was started with a brief practice run. Stimuli were presented using Psychtoolbox version 3. 0. 11 [69] (http: //psychtoolbox. org/) under MATLAB R2014a (MathWorks) on a desktop PC running Windows 7. Visual stimuli were presented via a gamma-corrected 30” LCD monitor with a resolution of 2,560 × 1,600 pixels at a frame rate of 60 Hz. Auditory stimuli were presented at a sampling rate of 44. 1 kHz via eight external speakers (Multimedia) and an ASUS Xonar DSX sound card. Exact audiovisual onset timing was confirmed by recording visual and auditory signals concurrently with a photodiode and a microphone. Participants rested their head on a chin rest at a distance of 475 mm from the monitor and at a height that matched participants’ ears to the horizontal midline of the monitor. Participants responded by pressing one of four response buttons on a USB keypad with their index, middle, ring, and little finger, respectively. To address potential concerns that results were confounded by eye movements, we recorded participants’ eye movements. Eye recordings were calibrated in the recommended field of view (32° horizontally and 24° vertically) for the EyeLink 1000 Plus system with the desktop mount at a sampling rate of 2,000 Hz. Eye position data were on-line parsed into events (saccade, fixation, eye blink) using the EyeLink 1000 Plus software. The ‘cognitive configuration’ was used for saccade detection (velocity threshold = 30°/sec, acceleration threshold = 8,000°/sec2, motion threshold = 0. 15°) with an additional criterion of radial amplitude larger than 1°. Individual trials were rejected if saccades or eye blinks were detected from −100 to 700 ms post stimulus. Participants’ stimulus localisation accuracy was assessed as the Pearson correlation between their location responses and the true signal source location separately for unisensory auditory, visual high reliability, and visual low reliability conditions. To confirm whether localisation accuracy in vision exceeded performance in audition in both visual reliabilities, we performed Monte-Carlo permutation tests. Specifically, we entered the subject-specific Fisher z-transformed Pearson correlation differences between vision and audition (i. e. , visual–auditory) separately for the two visual reliability levels into a Monte-Carlo permutation test at the group level based on the one-sample t statistic with 5,000 permutations [70]. Continuous EEG signals were recorded from 64 channels using Ag/AgCl active electrodes arranged in a 10–20 layout (ActiCap, Brain Products GmbH, Gilching, Germany) at a sampling rate of 1,000 Hz, referenced at FCz. Channel impedances were kept below 10 kΩ. Preprocessing was performed with the FieldTrip toolbox [71] (http: //www. fieldtriptoolbox. org/). For the decoding analysis, raw data were high-pass filtered at 0. 1 Hz, re-referenced to average reference, and low-pass filtered at 120 Hz. Trials were extracted with a 100-ms prestimulus and 700-ms poststimulus period and baseline corrected by subtracting the average value of the interval between −100 and 0 ms from the time course. Trials were then temporally smoothed with a 20-ms moving window and downsampled to 200 Hz (note that a 20-ms moving average is comparable to a finite impulse response [FIR] filter with a cutoff frequency of 50 Hz). Trials containing artefacts were rejected based on visual inspection. Furthermore, trials were rejected if (1) they included eye blinks, (2) they included saccades, (3) the distance between eye fixation and the central fixation cross exceeded 2°, (4) participants responded prior to the response cue, or (5) there was no response. For ERPs (S2 and S3 Figs), the preprocessing was identical to the decoding analysis, except that a 45-Hz low-pass filter was applied without additional temporal smoothing with a tempora |